Case management versus treatment as usual

There were three RCTs and two observational studies involving comparisons of case management and treatment as usual (AHLES1983; CONRAD1998; COX1998; MCLELLAN1999; PATTERSON1997). AHLES1983 compared case management with treatment as usual (standard aftercare arrangements), where the importance of attending aftercare was emphasised but not enforced. Patients were scheduled for one aftercare session at discharge and aftercare consisted of individual problem-oriented counselling. COX1998 compared case management with treatment as usual (there was no further description of treatment as usual). CONRAD1998 compared two types of residential inpatient care; the experimental group was case managed, whereas the control group participated in the residential care programme without case management.

Two observational were also included in the review. PATTERSON1997 compared the addition of a community psychiatric nurse (CPN) to aftercare with standard hospital care. Standard hospital care consisted of an offer of review appointments every 6 weeks following discharge. Lastly, MCLELLAN1999 compared case management with treatment as usual (no case management). In the standard-care condition, participants received group abstinence-oriented outpatient drug-misuse counselling twice weekly. In the case management condition, participants received a clinical case manager to provide support for housing, medical care, legal advice and parenting classes in addition to the drug counselling programme. For a graphical representation of the data, these two studies were inputted into the forest plots for comparison with the results of the RCTs; however, it should be noted that the outcomes and data were not pooled with the data found in the RCTs.

Case management versus treatment as usual

There was a significant difference in lapse (non-abstinence) at 6-month follow-up, in favour of case management, with a small effect size; however, this effect was not significant at 12-month follow-up. There was a significant difference favouring case management found at 3-, 4- and 5-year follow-up, with the largest effect size occurring at 3-year follow-up and decreasing to a moderate effect size at 4- and 5-year follow-up, respectively. It is important to note that these results are based on one observational study (PATTERSON1997).

On measures of drinking frequency, when considering the number of days drinking any alcohol (in the last 30 days) or mean days of intoxication, there were no significant differences between case management or treatment as usual at either 6-, 9-or 12-month follow-up. Interestingly, there was a significant effect observed at 18-month follow-up in favour of case management (small effect size) based on the results of one study (COX1998).

When considering the number of days using alcohol since the last interview (COX1998), there was a significant difference observed, favouring case management over treatment as usual at all follow-up points (small to moderate effect sizes): 6-, 12-and 18-month follow-up.

Based on the GRADE methodology outlined in Chapter 3, the quality of this evidence is moderate, therefore further research is likely to have an important impact on the confidence in the estimate of the effect (see Table 8).

Due to the heterogeneous nature of studies within case management, it was not possible to combine the outcome data provided across studies. As a result, there are a number of RCTs which add value to the meta-analysis presented. For the purpose of this guideline and to obtain a better overview of the available literature, four RCTs (Chutuape et al., 2001; Gilbert, 1988; Krupski et al., 2009; Sannibale et al., 2003; Stout et al., 1999), which met methodological criteria but did not have outcomes that could be used in meta-analyses for this review, are described below.

Gilbert (1988) conducted an RCT comparing case management, a home visit and treatment as usual for those with alcohol dependence. After receiving inpatient or outpatient treatment, patients were scheduled to be assigned a case manager or have a home visit, which consisted of appointments scheduled not at the hospital but at a convenient location for the patient. Patients in the home visit condition were contacted with follow-up letters to reschedule aftercare appointments. In the traditional treatment (treatment as usual), no active attempts were made to improve attendance at aftercare appointments. On appointment keeping measures, results from an analysis of variance (ANOVA) revealed a significant group by time interaction F = 4.56 (6,240), p <0.01, and post-hoc Tukey's Honestly Significant Difference (HSD) test revealed significant differences between home visit and case manager groups at 6- (p <0.05), 9- and 12-month follow-up (p <0.01). Both active treatment groups showed a decline in appointment keeping rates after the therapists stopped making active attempts to encourage the patient to attend therapy. On drinking outcomes, there were no significant differences between groups at any follow-up point.

Stout and colleagues (1999) conducted an RCT comparing case monitoring versus treatment as usual for those with alcohol dependence. The results indicated a significant difference on percentage of days heavy drinking at 3-year follow-up, where the frequency of heavy drinking was twice as high in the controls as in the case monitored participants. In addition, survival analysis indicated that case monitoring was significantly better at prolonging time to lapse and relapse (p = 0.05), as well as in reducing the severity of the relapse. There was no significant difference between the two groups for time to first heavy drinking day (p = 0.1). It should be noted that 66% of this sample had a comorbid Axis 1 diagnosis.

Chutuape and colleagues (2001) looked at the transition from an assisted-with-drawal programme to aftercare. Participants were randomly assigned to one of three conditions: incentive and escort to aftercare; incentive only; or standard treatment. Standard treatment participants only received referral instructions and were told to go to aftercare following discharge. Results from a logistic regression analysis indicated that aftercare contact rates differed significantly by referral condition (p = 0.001). Post hoc tests indicated that participants in the escort and incentive and incentive only conditions completed intake at aftercare more (p <0.05) than those receiving standard treatment.

When comparing a structured aftercare programme with an unstructured aftercare programme, Sannibale and colleagues (2003) found that structured programmes had a fourfold increase in aftercare attendance (odds ratio [OR] 4.3, 95% CI 1.7 to 11.2) and a reduced rate of uncontrolled substance use at follow-up (OR 0.3, 95% CI 0.1 to 0.9). Furthermore, participants in either aftercare condition relapsed later than those who attended no aftercare programme; however, this significant difference did not emerge for time to lapse.

More recently, Krupski and colleagues (2009) evaluated the impact of recovery support services (including case management) provided through an access to recovery programme in the US for clients undergoing substance-misuse treatment. Standard treatment consisted of ‘chemical dependency treatment’. The comparison group was a multi-modal programme entitled Access to Recovery (ATR), which included a case management component. They found that in comparison with standard care the ATR programme was associated with increased length of stay in treatment and completion of treatment (42.5 days longer). Further, multivariate survival analysis indicated the risk of ending treatment was significantly lower (hazard ratio = 0.58, p <0.05) among the ATR clients.

Definition

For the purposes of this review, stepped care is defined as a system for the organisation and delivery of care to people with harmful or dependent drinking which:

1. provides to the majority, if not all harmful or dependent drinkers, the least restrictive and least costly brief interventions that will be effective for a person's presenting problems
2. has a system of built-in monitoring that ensures that those who have not benefited from the initial intervention will be identified
3. has the referral systems and capacity to ensure that more intensive interventions are provided to those who have not benefited from a low intensity intervention.

Clinical utility

This criterion required the primary use of the assessment tool to be feasible and implementable in a routine clinical care. The tool should contribute to the identification of treatment needs and therefore be useful for treatment planning.

Psychometric data

Reported findings for sensitivity, specificity, area under the curve, positive predictive value, negative predictive value, reliability and validity of the assessment tools were considered. Although sensitivity and specificity are important outcomes in deciding on the usefulness of an assessment tool, particularly for diagnostic purposes, for other clinical purposes reliability and validity are also important. See Chapter 3 for a description of diagnostic test accuracy terms. The tool should be applicable to a UK population, for example by being validated in either a UK population or one that is similar to the UK population.

Tool characteristics and administrative properties

The assessment tool should have well-validated cut-offs in the patient population of interest. Furthermore, and dependent on the practitioner skill-set and the setting, tools were evaluated for the time needed to administer and score them as well as the nature of the training (if any) required for administration or scoring. Lastly, the cost of the tool and copyright issues were also considered.

Case identification/diagnosis

From the review of the literature, using the stipulated inclusion and exclusion criteria and properties outlined above, the GDG identified three tools for case identification in children and young people. These were the AAIS (Mayer & Filstead, 1979), the ADI (Harrell & Wirtz, 1985) and AUDIT (Babor et al., 2001). Both the AAIS and ADI have both been developed for use in an adolescent population. However, the AAIS has not been adequately validated and the ADI, although claiming adequate reliability and validity data, is not routinely used in the UK. As was the case in the review of adult assessment tools, the AUDIT questionnaire was found to be the most appropriate and suitable for use as a case identification/diagnostic instrument. For a review of the psychometric properties and characteristics of the AUDIT, see Section 5.18.1. The need for a revised cut-off in young people using the AUDIT questionnaire was evaluated. Chung and colleagues (2000) recommended modification of the AUDIT so that it is more appropriate to young people. Two studies using representative populations suggest a cut off score of 4 or more (Chung et al., 2000; Santis et al., 2009).

Comprehensive assessment instruments

As part of the systematic review and associated search strategies, a number of clinical interview tools which provide a comprehensive assessment of alcohol misuse in children and young people specifically were identified. These are: the ADI (Winters & Henly, 1993); the CASI-A (Meyers et al., 1995); the CDDR (Brown et al., 1998); the Diagnostic Interview Schedule for Children (DISC; Piacentini et al., 1993); the SCID SUDM (Martin et al., 1995a); the SUDDS-IV (Hoffman & Harrison, 1995); and the Teen Addiction Severity Index (T-ASI; Kaminer et al., 1991). Based on the criteria outlined above, the clinical interview tools that met inclusion criteria and are included in this narrative review are the ADI, DISC and T-ASI (see Table 17 below for characteristics of these tools). The GDG made a consensus-based decision to exclude the CASI-A, CDDR, SCID SUDM and SUDDS-IV from the narrative review because these tools have been developed for the use in an adolescent population over the age of 16 years old only, and hence may be inappropriate for use with children under that age. See Table 16 for characteristics of these excluded tools.

Table 17

Characteristics of clinical interview tools for children and adolescents included in the narrative review.

Use of biological markers

The review of adult alcohol misuse identified that no particular biological markers were of value in achieving a diagnosis of harmful or dependent drinking. Given that clinically significant changes in liver enzymes are rare in adults, even in those with established alcohol dependence (Clark et al., 2001), it seems unlikely that the routine use of such biological markers is of value in children and young people. However, the use of urine analysis or breath testing to determine the presence during treatment and/or assessment of drug or alcohol misuse may be of value in assessing the veracity in the overall assessment, but should not be used as a diagnostic marker.

Consumption

Harmful effects of alcohol use have been found to be influenced by both the amount and pattern of alcohol consumption (Rehm et al., 2004). Assessing typical daily and weekly alcohol consumption and comparing findings with recommended levels of alcohol consumption is therefore a useful starting point.

Individuals may present at different stages of a drinking cycle, so it is important to acknowledge that the absence of current alcohol use does not exclude the patient from being diagnosed with an alcohol-use disorder (WHO, 1992). Therefore an overview of the patient's current drinking status, preferred type of alcohol/brand consumed, the setting in which this occurs and general amount consumed is an important part of a assessment (Edwards et al., 2003; MoCAM; Department of Health, 2006a). Usually the assessment of consumption and frequency relies on the evaluation of self-reported alcohol consumption. Sobell and Sobell (2003) considered previous reviews of the validity and reliability of self-reported alcohol consumption and found enhanced accuracy in individuals who were: (i) alcohol free when interviewed; (ii) given written assurances of confidentiality; (iii) interviewed in a setting that encourages openness and honesty; (iv) asked clearly worded questions; and (v) provided with memory aids to recall drinking (that is, drink diaries), because those interviewed with alcohol in their system tended to underestimate their consumption. Previous reviews support the concept of enquiring about the patient's typical drinking day (Edwards et al., 2003; Rollnick et al., 1999). The notion of focusing on the typical drinking day allows staff to concentrate on what may normally occur in the absence of other factors that may influence large variations in alcohol consumption (such as stress, finances and life events) and which may be misleading. Regular high-level alcohol consumption may indicate tolerance to alcohol that has a significant relationship to alcohol dependence and consequent alcohol withdrawal.

The evolution of the patient's current alcohol consumption over time needs to be considered in order to identify significant patterns of alcohol use that are diagnostically important. In a more detailed assessment, the concept of drinking milestones may help to identify significant drinking-related instances from the time of first drink through to present alcohol consumption patterns. Edwards and colleagues (2003) suggests the inclusion of milestones such as age at first alcoholic drink, first drinking most weekends, first drinking daily and when they commenced drinking at the current levels. Additionally, it is important to document when the patient recognises the following: when they first felt alcohol was a problem; the heaviest period of alcohol consumption; and significant periods where they have experienced being alcohol free. Seeking clarification with regard to typical quantities of alcohol consumed at significant milestones will help establish the development of potential alcohol misuse.

Dependence

People who are dependent on alcohol develop adjustments to alcohol being present or absent in the body. Regular alcohol consumption can result in central nervous system changes that adapt to and compensate for the depressant effects of alcohol on the body. If this adaptation occurs, these changes may also result in the central nervous system being hyper-excited when alcohol levels are reduced, presenting characteristic alcohol withdrawal symptoms. Sensitive exploration of the six individual alcohol dependence criteria will confirm a diagnosis and help the individual to understand and acknowledge the condition that they are experiencing (Edwards et al., 2003). It is generally accepted that a number of aspects of dependence should be covered in a comprehensive assessment, including tolerance, neglecting activities and interests, compulsion, physiological withdrawal and drinking despite problems (Maisto et al., 2003).

Tolerance

Regular alcohol drinkers become tolerant to the central nervous system effects of alcohol (Kalant, 1996). There appears to be a number of individual factors that influence the development of tolerance to alcohol including metabolic, environmental and learned factors (Tabakoff et al., 1986). There is no simple clinical tool to directly measure alcohol tolerance. However, increasing consumption levels and a reduced effect of the same amount of alcohol over time are indicative of tolerance. The effect of blood alcohol concentration on an individual will decrease as tolerance develops (Hoffman & Tabakoff, 1996), but even in tolerant individuals high-level alcohol consumption will still impair functioning and judgement. Nevertheless, people with very high alcohol tolerance will be able to tolerate a high blood alcohol concentration that would be fatal to a non-tolerant individual.

Alcohol withdrawal

Staff will need to understand and recognise the features of alcohol withdrawal to accurately arrive at a diagnosis of alcohol dependence. Alcohol withdrawal symptoms need to be differentiated from other clinical characteristics and conditions that may present similarly.

Alcohol withdrawal symptoms include:

• Tremor
• Nausea
• Sweating
• Mood disturbance including agitation and anxiety
• Disturbed sleep pattern
• Hyperacusis – sensitivity to sound
• Hyperthermia – increased body temperature
• Tachycardia – increased pulse rate
• Increased respiratory rate
• Tactile and/or visual disturbances – itching, burning and so on.

Severe alcohol withdrawal symptoms include:

• Hallucinations – auditory, visual and/or tactile
• Alcohol withdrawal seizures – grand mal-type seizures
• DTs – coarse tremor, agitation, fever, tachycardia, profound confusion, delusions and hallucinations.

People who are moderately or severely alcohol dependent will develop an acute alcohol withdrawal syndrome when they abruptly stop or substantially reduce their alcohol consumption. People who are mildly dependent may experience some milder symptoms of alcohol withdrawal including sweating, nausea, and mild tremor but generally do not require medical treatment. Withdrawal symptoms develop as early as 6 to 8 hours after abrupt reduction or cessation of alcohol intake.

Table 18 provides an illustration of alcohol withdrawal symptoms against a timeline since last drink. The time from last drink to onset of withdrawal symptoms reduces with increasing severity of dependence, such that people who are severely alcohol dependent may begin to experience withdrawal within a few hours of the last drink.

Table 18

Timeline for the emergence of alcohol withdrawal symptoms.

The individual may describe the use of alcohol to avoid or relieve the effects of alcohol withdrawal, which would further demonstrate dependence to alcohol.

Compulsion

An individual's compulsion to consume alcohol is commonly reported when an alcohol-dependent drinker attempts to control or stop use. This has been described as urges or cravings (Drummond & Phillips, 2002). The intensity of craving is highly correlated with the severity of dependence. Certain situations and emotional states (cues) that influence the presence and intensity of alcohol craving, as these may be an important factor in precipitating future drinking episodes (Drummond, 2000). Not everyone who is alcohol dependent reports alcohol craving, and craving per se does not inevitably lead to relapse. However, for some service users it can be an unpleasant and troubling symptom.

Neglecting activities and interests

Individuals who are dependent on alcohol may describe a reduction or change in their participation in activities they hold as important (Edwards & Gross, 1976). As alcohol becomes increasingly more salient to the drinker the need to obtain, consume and/or recover from excessive alcohol consumption takes higher priority in their lives relative to usual obligations and interests. Identifying the salience alcohol has for the individual by exploring past and current interests with them is an important factor to establish.

Drinking despite problems associated with alcohol

Alcohol-related problems can occur in the absence of alcohol dependence (that is, accidents, legal problems and so on). However, a person who is dependent on alcohol may maintain drinking behaviour despite experiencing harmful effects of alcohol such as harm to the liver and depressed mood (Edwards & Gross, 1976). The individual may continue to drink despite criticisms from family, friends and work colleagues. This can be difficult to establish because the individual may not make a connection between their drinking and the consequences, or may be embarrassed about discussing problems related to their drinking. Part of the process of the assessment is to help the individual make these causal connections through motivational interviewing techniques.

Other substances of misuse

The assessment of alcohol misuse is often complicated by the presence of co-occurring conditions; these, along with the implications for assessment, are outlined below.

Comorbid opioid and alcohol dependence

In treatment services for opioid dependency, about a quarter to a third of patients will have problems with alcohol (Department of Health, 2007). In addition, prognosis for this group can be poor with many showing limited changes in drinking behaviour. A recent systematic review about whether alcohol consumption is affected during the course of methadone maintenance treatment concluded that alcohol use is not likely to reduce by just entering such programmes, with most studies reporting no change (Srivastava et al., 2008). In the UK National Treatment Outcome Research Study, 25% of people misusing opiates were drinking heavily (more than 10 units per day) at the start of the study and 4 to 5 years later about a quarter were continuing to do so (Gossop et al., 2003).

Comorbid cocaine and alcohol dependence

Cocaine use is increasing in England (NHS Information Centre & National Statistics, 2009), and comorbid cocaine and alcohol dependence is commonly seen and can be challenging to treat. There is little known in the UK about the level of this comorbidity in alcohol treatment services. In the US Epidemiological Catchment Area study, 85% of cocaine-dependent patients were also alcohol dependent (Regier et al., 1990). In a sample of 298 treatment-seeking cocaine users, 62% had a lifetime history of alcohol dependence (Carroll et al., 1993). In a sample of people in contact with drug treatment agencies (mainly for opiate addiction and in the community abusing cocaine), heavy drinking was common. Those using cocaine powder were more likely to drink heavily than those using crack cocaine (Gossop et al., 2006).

When taken together, cocaine and alcohol interact to produce cocaethylene, an active metabolite with a half-life three times that of cocaine. In addition alcohol inhibits some enzymes involved in cocaine metabolism, so can increase its concentration by about 30% (Pennings et al., 2002). Due to the presence of cocaethylene, which has similar effects to cocaine and a longer half-life, this leads to enhanced effects. For instance, taken together cocaine and alcohol result in greater euphoria and increased heart rate compared with either drug alone (McCance-Katz et al., 1993; see Pennings et al., 2002).

Comorbid alcohol and benzodiazepine dependence

Benzodiazepine use is more common in patients with alcohol misuse than in the general population, with surveys reporting prevalence of around 10 to 20% (Ciraulo et al., 1988; Busto et al., 1983). In more complex patients it can be as high as 40%, which is similar to that seen in psychiatric patients. A proportion of alcohol misusers who take benzodiazepines will be benzodiazepine dependent. For some individuals, their growing dependence on benzodiazepines began when a prescription for withdrawal from alcohol was extended and then repeatedly renewed. For others the prescription may have been initiated as a treatment for anxiety or insomnia, but then was not discontinued in line with current guidelines.

Comorbid alcohol and nicotine dependence

Many patients with alcohol misuse smoke cigarettes, which leads to an extra burden of morbidity and mortality in addition to the alcohol misuse. The prevalence of smoking has been estimated at around 40% in population-based studies of alcohol-use disorder and as much as 80% in people with alcohol dependence who are seeking treatment (Grant et al., 2004b; Hughes, 1995). Comorbidity is higher in men than women and in younger compared with older people (Falk et al., 2006). Comorbid nicotine and alcohol dependence has been comprehensively reviewed recently by Kalman and colleagues (2010).

Motivation and self-efficacy

The assessment of an individual's willingness to engage in treatment can vary considerably and has been the subject of considerable debate. Assessment can be effective as an intervention in itself and has been shown to influence behaviour change (McCambridge & Day, 2008), increasing an individual's confidence towards change that may prompt reductions in alcohol consumption (Rollnick et al., 1999). Being sensitive to the individual's needs, developing rapport and a therapeutic alliance have all been identified as important aspects in the effective engagement of an individual who drinks excessively (Najavits & Weiss, 1994; Raistrick et al., 2006; Edwards et al., 2003). Indeed, there is evidence to suggest that a premature focus on information gathering and completion of the assessment process may have a negative impact on the engagement of the patient (Miller & Rollnick, 2002). Where this approach is adopted, there is some evidence to suggest that initial low levels of motivation are not necessarily a barrier to an effective assessment and the future uptake of treatment (Miller & Rollnick, 2002).

An openness to discussion aimed at understanding a person's reasons for seeking help and the goals they wish to attain has also been positively associated with engagement in assessment and treatment (Miller, 1996). The individual's personal drinking goals can then be acknowledged and used as a basis for negotiation once the assessment is completed (Adamson et al., 2010).

Alcohol-related problems present in a number of different settings, often concurrently (for example, a person may present as depressed in primary care subsequent to a brief admission for acute pancreatitis, both related to excessive alcohol intake). Therefore, effective assessment systems need to be linked to equally effective communication between those involved in the care and treatment of people with alcohol-related problems (Maisto et al., 2003). Sharing of information between agencies should be encouraged to maximise safety and effectiveness of treatment (MoCAM; Department of Health, 2006a).

Aims

Case identification and, following on from that, diagnosis seek to identify individuals who are in need of intensive care-planned treatment because of possible alcohol dependence, those with harmful alcohol use who are in need of or have not responded to brief interventions and those experiencing comorbid problems which may complicate the treatment of the alcohol misuse. Given the overall stepped framework in which the assessment takes place it is anticipated that this level of care would have two main objectives:

1. To identify those individuals who need an intervention (see Chapters 6 and 7) for harmful or alcohol dependence
2. To identify those who may need referral for a comprehensive assessment and/or withdrawal assessment including those who:

   –

   have not responded to an extended brief intervention

   –

   have moderate to severe alcohol dependence or otherwise may need assisted alcohol withdrawal

   –

   those that show signs of clinically significant alcohol-related impairment (for example, liver disease or significant alcohol-related mental health problems).

Settings

Case identification and diagnosis should be available across the whole range of healthcare and related services (for example, GPs, accident and emergency departments, children and families social services, and specialist alcohol treatment agencies).

Method

This level of assessment should consider:

• establishing the probable presence of an alcohol-use disorder
• the level of alcohol consumption (as units¹⁸ of alcohol per day or per week)
• where an alcohol-use disorder is suggested, distinguish harmful drinking from alcohol dependence
• establishing the presence of risks (for example, self-harm, harm to others, medical/mental health emergencies and safeguarding children)
• establishing the capacity to consent to treatment or onward referral
• experience and outcome of previous intervention(s)
• establishing the willingness to engage in further assessment and/or treatment
• establishing the presence of possible co-existing common problems features (for example, additional substance misuse, medical, mental health and social problems)
• determining the urgency of referral and/or an assessment for alcohol withdrawal.

The treatment options that follow immediately on from this initial assessment, with the exception of assisted withdrawal, will focus on harmful or dependent drinking. A significant number of individuals may already have received brief intervention and not benefited from them; if this is the case then the individual will need to be referred for a comprehensive assessment. See Figure 5 for an outline of the care pathway for the case identification and possible diagnosis for adults.

Figure 5. Care pathway: case identification and possible diagnosis for adults (PDF, 54K)

Aims

Assessment of the need for a medically managed withdrawal, the potential risks (for example, DTs or seizures) and the most appropriate setting in which to manage withdrawal. A key factor will be determining whether the withdrawal management should take place in a community, inpatient or residential setting. This section of the guideline should be read in conjunction with the section on planned assisted alcohol withdrawal in this guideline and the reader should also refer to the guideline on the management of unplanned acute withdrawal (NICE, 2010b). It should be noted that assisted withdrawal from alcohol should not be seen as a standalone treatment for alcohol dependence but rather as an often essential initial intervention within a broader care plan including psychosocial or pharmacological therapies to prevent relapse. Specifically, the withdrawal assessment should aim:

1. to identify those individuals who need an assisted withdrawal because of alcohol dependence
2. to identify:
   • the severity of the dependence
   • the level of alcohol consumption
   • the presence of comorbid factors such as substance misuse, severe psychiatric disorders, significant physical illness or disability
   • the availability of personal and social support and housing support
3. to identify in which setting a withdrawal can be most clinically- and cost-effectively and safely managed
4. to determine the urgency with which the assisted withdrawal should be provided
5. to provide sufficient information to properly integrate the assisted withdrawal programme into a wider care plan.

Settings

Withdrawal assessments take place in a number of healthcare settings; the management of those presenting in a state of unplanned withdrawal in acute medical settings is dealt with in NICE (2010b). However, although this guideline's recommendations are focused primarily on the management of planned withdrawal, a number of the recommendation in this guideline will be relevant to the assessment of all individuals who are alcohol dependent and at risk of developing withdrawal symptoms. Primary care, prisons, police custody, general hospitals, secondary care mental health services, and specialist drug and alcohol services are all settings in which the need for a withdrawal assessment may arise. These varied settings mean that the nature of the assessment will vary depending on the resources and skills available in those settings. However, as described in Section 4 of this chapter, there is evidence that assisted withdrawal from alcohol can be safely and effectively delivered in all of those settings provided that an assessment has been performed to determine the most appropriate environment in which to undertake the withdrawal and the regimen required (Maisto et al., 2003). The impact of comorbid conditions and their implications for the choice of withdrawal setting is described more fully in Section 4. A number of reviews (for example, NICE, 2010b; Raistrick et al., 2006) highlight factors that suggest the need for residential or inpatient withdrawal programmes. These include: those who are at high risk¹⁹ of developing alcohol withdrawal seizures or DTs; those with a history of polydrug use; significant cognitive impairment; the homeless; and those with an illness that requires medical/surgical or psychiatric treatment.

Methods

Those who experience a significant degree of alcohol dependence will typically exhibit alcohol withdrawal symptoms 6 to 8 hours after their last drink, with peak effect of alcohol withdrawal symptoms occurring after between 10 to 30 hours (see guideline on management of alcohol-related physical complications; NICE, 2010b). However the onset of withdrawal varies with severity of dependence such that people who are severely dependent will experience withdrawal earlier after stopping drinking than those who are less dependent. Early diagnosis of alcohol dependence will help to initiate proactive management strategies for the individual and reduce risks to the patient.

The NICE guideline on management of alcohol-related physical complications

(NICE, 2010b) reviewed the tools for the assessment and monitoring of patients who are alcohol dependent and at risk of developing alcohol withdrawal. The guideline recommends the use of a validated tool to support clinical judgement in the assessment of alcohol withdrawal. Furthermore, the guideline recommended the use of an assessment tool in situations particularly where staff are less experienced with the assessment and management of alcohol withdrawal. The guideline identified the CIWA-Ar as a valuable tool for measuring alcohol withdrawal symptoms. The guideline also noted that a delay in assessment and treatment of withdrawal of more than 24 hours is associated with greater withdrawal complications.

After establishing the possibility of alcohol misuse it is important to establish first whether or not dependence is present; in all settings this is a two stage process. The first stage involves the identification of those at risk of dependence and withdrawal. The preferred aid to a clinical assessment is the AUDIT questionnaire. An AUDIT score greater than 20 is an indication of likely alcohol dependence and the need for withdrawal assessment (Babor et al., 2001). If it is not possible to complete an AUDIT questionnaire then regular consumption of alcohol of 15 to 20 or greater units per day suggests likely dependence. Although there is no absolute level of daily or weekly alcohol consumption which indicates the likelihood of alcohol dependence, the SADQ score (a measure of the severity of dependence – see above) correlates with high-level alcohol consumption (Stockwell et al., 1979). Others support the view that typical drinks per drinking day is a useful indicator of the severity of alcohol dependence and need for alcohol withdrawal management (Shaw et al., 1998). There are a number of methods to establish alcohol quantity and frequency including direct patient report, drinking diaries and retrospective recording systems (Sobell & Sobell, 2003), although previous reviews have identified that such techniques vary in accuracy (Raistrick et al., 2006). However it should be noted that both AUDIT scores and typical drinks per day should be adjusted for gender (Reinert & Allen, 2007), age (both for older adults [Beullens & Aertgeerts, 2004] and adolescents [McArdle, 2008]) and people with established liver disease (Gleeson et al., 2009).

The second stage involves an assessment of the severity of alcohol dependence. Again, a formal assessment tool is the preferred means to identify the severity of dependence in this guideline. The review of such tools for this guideline revealed that the SADQ (Stockwell et al., 1979 and 1983) has broad clinical utility because it identifies the presence and severity of alcohol dependence, predicts withdrawal severity and acts as a useful guide for the quantity of medication to be prescribed during alcohol withdrawal.

Identification of need for assisted alcohol withdrawal

Identification of withdrawal should be based on careful assessment of the pattern, frequency and intensity of drinking. The limited data available for review, the evidence from adults and the greater vulnerability of young people to the harmful effects of alcohol led the GDG to conclude that there should be a significant reduction in the threshold for young people for initiating withdrawal management. The threshold that has been established for adults of an AUDIT score of more than 20, an SADQ score of more than 20 or the typical consumption of 15 units per day is not appropriate for adolescents. In adolescents, binge drinking is common (defined as more than five units of alcohol on any one occasion) and a pattern of frequent binge drinking (for example, a pattern of two or more episodes of binge drinking in a month) or an AUDIT score of 15 should alert the clinician to possible dependence and trigger a comprehensive assessment. The presence of any potential withdrawal symptoms should be taken seriously and a comprehensive assessment initiated. A range of factors including age, weight, previous history of alcohol misuse and the presence of co-occurring disorders will also influence the threshold for initiating a comprehensive assessment and withdrawal management. See Figure 6 for a summary of the care pathway for withdrawal assessment.

Figure 6. Care pathway: withdrawal assessment (PDF, 60K)

Aims

A brief triage assessment should be undertaken when an individual first contacts a specialist alcohol service, and it has the aim of developing an initial plan of care (MoCAM; Department of Health, 2006a). Failure to identify clinical and/or social priorities may result in an individual being directed to inappropriate services or lost to any form of care. Typically, people presenting for a triage assessment may be harmful drinkers who have not benefitted from an extended brief intervention (see NICE, 2010a) and/or those with an AUDIT score of more than 20, or have been referred to or have self-referred to a specialist alcohol services.

A brief triage assessment is not simply a brief assessment of alcohol misuse only. The focus is equally on the management risk, identification of urgent clinical or social problems to be addressed and accessing the most appropriate pathways of care for alcohol misuse. The triage assessment therefore incorporates the common elements of assessment identified above with the aim of establishing the severity of the individual's problems, the urgency to action required and referral to the most appropriate treatment interventions and service provider.

Specifically the triage assessment should establish:

• the need for emergency or acute interventions, for example referral to accident and emergency for an acute medical problem or to a crisis team for a mental health emergency
• presence and degree of risk of harm to the person or others (including children) due to alcohol or substance misuse and related problems (medical, mental health, social and criminal)
• the appropriate alcohol treatment intervention(s) and setting(s) for the problems assessed
• an appropriate level of communication and liaison with all those involved in the direct care and management of the individual
• the need for a further comprehensive specialist assessment (see Section 5.25.5 below)
• the need to agree follow-up plans.

Settings

All specialist alcohol services (including those that provide combined drug and alcohol services) should operate a triage assessment according to agreed local procedures. This level of assessment is not intended to be a full assessment of an individual's needs on which to base a care plan. The triage assessment should identify immediate plans of care through the use of standardised procedures to ensure that all clinically significant information and risk factors are captured in one assessment. Incorporating tools and questionnaires as an adjunct to the clinical interview will help improve consistency of decision making.

Methods

The triage assessment should include:

• alcohol use history including:

  –

  typical drinking; setting, brand, and regularity

  –

  alcohol consumption using units of alcohol consumed on a typical drinking day

  –

  features of alcohol dependence

  –

  alcohol-related problems

  –

  adjunctive assessment tools (including the AUDIT and SADQ) to inform the assessment of risk and the immediate and future clinical management plan

• co-occurring problems (medical, mental health, substance misuse, social and criminal)
• risk assessment
• readiness and motivation to change.

Risk assessment

The increasing importance of risk assessment in the clinical decision-making process has led to a number of tools being developed to systematically screen for high-risk problems and behaviours which draw on a common framework for risk assessment systems in mental health (Department of Health, 2006a). In the NHS, it is expected that local protocols are agreed that specify the elements and tools for risk assessment to be applied (MoCAM; Department of Health, 2006a). Establishing these protocols and standards will also identify the competencies required for the collation and interpretation of risk to develop a risk management plan.

The risk assessment process should review all aspects of the information collected during the clinical interview and, where appropriate, consider results from investigations, questionnaire items, correspondence and records, and liaison with other professionals as well as family and carers, to formulate an assessment of risks to the individual, to others and to the wider community. The risk assessment should consider the interaction between comorbid features to arrive at an informed opinion of the severity of risk and the urgency to act.

MoCAM (Department of Health, 2006a) identifies that risk assessment should consider the following domains:

• risks associated with alcohol use or other substance misuse (such as physical damage or alcohol poisoning)
• risk of self-harm or suicide
• risk of harm to others (including risk of harm to children and other domestic violence, harm to treatment staff and risk of driving while intoxicated)
• risk of harm from others (including being a victim of domestic abuse)
• risk of self-neglect
• safeguarding-children procedures (must be included).

Where risks are identified, a risk management plan that considers monitoring arrangements, contingency plans and information sharing procedures needs to be developed and implemented (MoCAM; Department of Health, 2006a). Guidance developed for those working with patients with mental health problems indicates that the most effective risk assessments and risk management plans are developed by multidisciplinary teams and in collaboration between health and social care agencies (Department of Health, 2007).

Urgency to act

The urgency to act will be linked to the severity and level of risks identified from all the information gathered and should consider:

• The individual's intentions to carry out acts of self-harm or harm to others
• The state of distress being experienced by the individual
• The severity of comorbid medical or mental health conditions and the sudden deterioration of the individual's presentation
• The safeguarding needs of children/young people.

Aims

A comprehensive assessment should be undertaken where an individual experiences significant comorbidity, moderate or severe alcohol dependence or presents a high level of risk to themself or others. This group will often require structured and/or intensive interventions and is often involved with multiple agencies. Those presenting with complex problems will require their care to be planned and co-ordinated.

The comprehensive assessment aims:

• to determine the exact nature of problems experienced by the individual across multiple domains
• to specify needs to inform development of a care plan
• to identify planned outcomes to be achieved and methods to assess these outcomes

Settings

Comprehensive assessment is undertaken by specialist alcohol services that provide Tier 3 and 4 interventions, although some Tier 2 services with sufficiently experienced staff may also offer comprehensive assessments as outlined by MoCAM (Department of Health, 2006a).

Methods

Comprehensive assessment should not be seen as a single event conducted by one member of the multidisciplinary team, although coordination of the assessment process may bring real benefit (see Section 5.3 for a review of care coordination and case management). The complex nature of the problems experienced by an individual with long-standing alcohol misuse or dependence suggests that the comprehensive assessment may need to be spread across a number of appointments and may typically involve more than one member of the multidisciplinary team. A range of expertise will often be necessary to assess the nature of problems. Comprehensive assessment may require specific professional groups to undertake tasks such as physical examination, prescribing needs, social care needs, psychiatric assessment and a formal assessment of cognitive functioning. Specialist alcohol services conducting comprehensive assessments therefore need to have access to: GPs and specialist physicians, addiction psychiatrists, nurses, psychologists and specialist social workers.

The comprehensive assessment should include an in-depth consideration and assessment of the following domains:

• alcohol use and related consequences

  –

  alcohol consumption

  –

  alcohol dependence

  –

  alcohol-related problems

• motivation
• self-efficacy
• co-occurring problems

  –

  substance misuse

  –

  physical health history and problems

  –

  mental health history and problems

  –

  social functioning and problems

• risk assessment
• treatment goals
• assessment of the service user's capacity to consent to treatment
• formulation a plan of care and risk management plan.

Breath/blood alcohol level

Alcohol is excreted in the breath and its concentration in the breath is correlated with blood alcohol concentration. On average it takes approximately 1 hour to eliminate one unit of alcohol from the body; however, the elimination rate varies between individuals and is more rapid in people who are alcohol dependent than those who are not (Allen et al., 2003; Ugarte et al., 1977). Breath alcometers reliably measure the breath alcohol concentration in a non-invasive way. Blood/breath alcohol concentration may be a useful part of the clinical assessment in the following areas:

• Although self report has been found to be a reliable indicator of levels of alcohol consumption in treatment-seeking populations, patients with alcohol in their system at the time of assessment are more likely to underestimate their levels of alcohol consumption (Sobell & Sobell, 2003).
• Clinicians have a responsibility to discuss drink-driving concerns with patients and their responsibilities in reporting this to the Driver and Vehicle Licensing Agency (DVLA, 2010). Service users who have driven on the way to an assessment and who are over the legal limit for driving (80 mg/100 ml) need to be advised not to drive until they are legally able to do so.

Blood investigations

There are a number of biomarkers suggested to be clinically useful in the assessment of alcohol-related physical harm (Allen et al., 2003), monitoring of clinical outcome and as a motivational enhancement strategy (Miller et al., 1992). However, in people who are seeking treatment for alcohol misuse, biomarkers do not offer any advantage over self-report measures in terms of accuracy of assessing alcohol consumption (Allen et al., 2003; Sobell & Sobell, 2003), and are less sensitive and specific than the AUDIT in screening for alcohol misuse (Drummond & Ghodse, 1999).

Gamma-glutamyl transferase (GGT) has a sensitivity of 50 to 70% in the detection of high levels of alcohol consumption in the last 1 to 2 months and a specificity of 75 to 85% (Drummond & Ghodse, 1999). Reasons for false positive results include hepatitis, cirrhosis, cholestatic jaundice, metastatic carcinoma, treatment with simvastatin and obesity.

Mean corpuscular volume has a sensitivity of 25 to 52% and specificity of 85 to 95% in the detection of alcohol misuse. It remains elevated for 1 to 3 months after abstinence. Reasons for false positives include vitamin B12 and folate deficiency, pernicious anaemia, pregnancy and phenytoin (Allen et al., 2003; Drummond & Ghodse, 1999).

Carbohydrate-deficient transferrin (CDT) has greater specificity (80 to 98%) than other biomarkers for heavy alcohol consumption, and there are few causes of false positive results (severe liver disease, chronic active hepatitis) (Schwan et al., 2004). However, routine CDT monitoring is not routinely available and there remains some debate about how best to measure it. Evidence suggests that the test is less sensitive in women (Anton & Moak, 1994). CDT increases and recovers more rapidly than GGT in response to a drinking binge, increasing within 1 week of onset of heavy drinking, and recovery typically occurs in 1 to 3 weeks compared with 1 to 2 months for GGT (Drummond & Ghodse, 1999).

Advantages of blood investigations as part of the initial assessment include:

• screening for alcohol-related physical conditions that may need further investigation and onward referral
• providing baseline measures of alcohol-related damage (in some patients) against which to measure improvement and act as motivational enhancement strategy
• objective measurement of outcome, particularly when combined (for example, CDT and GGT; Allen et al., 2003) and in conjunction with other structured outcome measures (Drummond et al., 2007).

Hair and sweat analysis

As alcohol is rapidly excreted from the body, there is currently no reliable or accurate way of measuring alcohol consumption in the recent past and the mainstay of outcome measurement is self-report (Sobell & Sobell, 2003). This is less useful for regulatory monitoring purposes and so there is a growing interest by manufacturers in the design of biomarkers for recent alcohol consumption. Studies to date focus on hair and skin sweat analysis, but there is currently a lack of evidence to recommend their use in routine clinical care (Pragst & Balikova, 2006).

Assessment of alcohol-related physical harm

The assessment of alcohol-related physical harm is an important component of a specialist service (Edwards et al., 2003). The aims of such an assessment are to:

• identify physical health problems that require immediate attention and onward referral to appropriate acute medical care
• identify physical health problems that are a consequence of the alcohol misuse and require monitoring, and potential future referral

The relationship between alcohol-related physical health problems and level of alcohol consumption is complex (Morgan & Ritson, 2009), as is the presence of physical signs in relation to underlying pathology. Consequently, patients presenting with longstanding, severe alcohol dependence may have few overt physical signs, but have significant underlying organ damage (for example, liver disease). Others may present with significant symptoms (for example, gastritis) or signs (for example, hypertension) that may resolve without active treatment once the service user abstains.

Liver/gastrointestinal problems

Alcohol-related liver disease often develops ‘silently’ over a 10 to 15 year period and blood tests of liver function (alanine amino transferase [ALT]) may only become abnormal at quite advanced stages of disease, so a test that is within the normal range does not exclude liver damage (Prati et al., 2002). Equally, raised ALT may be the result of induction of liver enzymes by alcohol rather than an indication of liver pathology (Drummond & Ghodse, 1999). Other laboratory test results including GGT and serum aspartate amino transferase (AST) may be raised in people who misuse alcohol, but do not necessarily indicate the presence of significant organ damage as a result of enzyme induction (Bagrel et al., 1979). Patients with signs of severe (decom-pensated) liver disease (for example, presenting with jaundice, fluid retention, spontaneous bruising or hepatic encephalopathy) will need urgent specialist medical care from a hepatology service. Symptoms of anorexia, nausea, vomiting and diarrhoea, and malabsorption syndromes are common in people who misuse alcohol. In many cases these symptoms resolve with treatment of the underlying alcohol misuse, but people with significant pain or evidence of gastrointestinal blood loss will need referral for further investigation.

Cardiovascular

Alcohol has a dose-related effect on blood pressure in addition to blood pressure being elevated during alcohol withdrawal (Xin et al., 2001). Patients who present with hypertension or who are already prescribed anti-hypertensive medication will need to have this reviewed as treatment progresses.

Neurological

Wernicke's encephalopathy classically presents with a triad of symptoms (ataxia, confusion and ophthalmoplegia), but in practice this triad only occurs in a minority of cases (Thomson & Marshall, 2006). Given the severity of brain damage (Wernicke–Korsakoff syndrome [WKS]) that may occur if the condition is untreated, clinicians need to have a high index of suspicion particularly in those patients who are malnourished and have any of the following clinical signs: ataxia, ophthalmoplegia, nystagmus, acute confusional state, or (more rarely) hypotension or hypothermia. Patients presumed to have a diagnosis of WE will need immediate treatment or onward referral (NICE, 2010b).

Symptoms of peripheral neuropathy are common (30 to 70%) in people who misuse alcohol (Monteforte et al., 1995). The symptoms are predominantly sensory (although muscle weakness is also seen) and include numbness, pain and hyperaesthesia in a ‘glove and stocking’ distribution, primarily in the legs. Symptoms should be monitored and will require referral if they do not improve with alcohol abstinence.

Depression and anxiety

Although many symptoms of depression or anxiety may be directly attributable to alcohol misuse, many people still reach the threshold for a diagnosis of a psychiatric disorder. For instance, 85% of patients in UK alcohol treatment services had one or more comorbid psychiatric disorders including 81% with affective and/or anxiety disorders (34% severe depression; 47% mild depression; 32% anxiety) and 53% had a personality disorder (Weaver et al., 2003). Such high levels of comorbidity are not surprising given that the underlying neurobiology of depression or anxiety and alcoholism have many similarities, particularly during withdrawal (Markou & Koob, 1991). In addition there are shared risk factors because twin studies reveal that the presence of one disorder increases the risk of the other disorder (Davis et al., 2008).

There is a high prevalence of comorbidity between anxiety and alcohol misuse, both in the general and clinical populations. Anxiety disorders and alcohol dependence demonstrate a reciprocal causal relationship over time, with anxiety disorders leading to alcohol dependence and vice versa (Kushner et al., 1990). Panic disorder and generalised anxiety disorder can emerge during periods of alcohol misuse; however, the association with obsessive-compulsive disorder is less robust.

Social phobia and agoraphobia often predate the onset of alcohol misuse. The prevalence of social anxiety ranges from 8 to 56% in people who misuse alcohol, which makes it the most prevalent psychiatric comorbidity. People who are alcohol dependent and have comorbid social anxiety disorder show significantly more symptoms of alcohol dependence, higher levels of reported depression, and greater problems and deficits in social support networks as compared with alcohol dependent patients without social anxiety (Thevos et al., 1999).

The relationship between alcohol and depression is also reciprocal in that depression can increase consumption, but depression can also be caused or worsened by alcohol misuse (Merikangas et al., 1996).

Sleep disorders

Sleep disorders (commonly insomnia) increase the risk of alcohol misuse and also contribute to relapse (Brower, 2003; Krystal et al., 2008). Whilst many people believe that alcohol helps them sleep, this is not the case. Although onset of sleep may be reduced after drinking alcohol, disruption to sleep patterns occur later in the night such as rapid eye-movement rebound and increased dreaming, as well as sympathetic arousal (Krystal et al., 2008). Abstinence may reveal a sleep disorder that the person has not been entirely aware of because they have always used alcohol to sleep. Insomnia is also a prominent feature of both acute and protracted alcohol withdrawal syndromes, the latter of which can last for 3 to 12 months.

Eating disorders

There is substantial evidence that alcohol misuse and eating disorders commonly co-occur (Sinha & O'Malley, 2000). In specialist alcohol inpatient treatment, the prevalence of eating disorders can be as high as 40%. Commonly, an eating disorder exists together with other psychiatric disorders such as depression. In people with an eating disorder, up to half have been reported to misuse alcohol (Dansky et al., 2000). A number of studies have found the strongest relationship for bulimia nervosa, followed by patients suffering from binge eating disorder and eating disorder not otherwise specified (EDNOS) (Gadalla & Piran, 2007). No association has been reported between anorexia nervosa and alcohol misuse. In a study of European specialist eating-disorder services, alcohol consumption was higher in patients with EDNOS and bulimia nervosa than anorexia nervosa, but a greater lifetime prevalence of alcohol misuse was not found (Krug et al., 2009).

Psychosis

Patients with psychotic disorders (including schizophrenia and bipolar disorder) are vulnerable to the effects of alcohol and at increased risk of alcohol misuse (Weaver et al., 2003). Approximately 50% of patients requiring inpatient psychiatric treatment for these disorders will also misuse alcohol (Barnaby et al., 2003; Sinclair et al., 2008). However, a smaller proportion of patients will present without a diagnosis of an underlying psychotic or mood disorder, which will need to be identified as part of a comprehensive assessment. For a more thorough review of this area see the NICE guideline on psychosis and substance misuse (NICE, 2011a).

Self-harm and suicide

There is a significant but complex association between alcohol misuse and self-harm and suicide. Approximately 50% of all patients presenting to hospital following an episode of self-harm have consumed alcohol immediately before or as part of the act of self-harm (Hawton et al., 2007). The mortality by suicide in patients who present following an episode of self-harm is significantly increased in the next 12 months (66 times that of the general population) (Zahl & Hawton, 2004) and this risk remains high after many years (Owens et al., 2002). However, recent data from a long-term follow-up study suggests that the mortality of self-harm patients appears to be caused by alcohol-related conditions as much as suicide (Sinclair et al., 2009). For people whose self-harm occurs only when intoxicated, abstinence from alcohol was recognised as the effective intervention (Sinclair & Green, 2005). Alcohol dependence has been shown to increase the risk of suicide by five to 17 times, with the RR being greatest in women (Wilcox et al., 2004).

Cognitive impairment

Prolonged cognitive deficits are seen in 50 to 80% of people with alcohol dependence who have undergone assisted alcohol withdrawal (Bates et al., 2002). Cognitive impairments frequently improve significantly once abstinence has been achieved and so should be reassessed after 2 to 3 weeks of abstinence (Loeber et al., 2009).

A number of assessment tools can be used to assess cognitive function in people who misuse alcohol have been identified. These include the Mini-Mental State Examination (MMSE; Folstein et al., 1975); the Cognitive Capacity Screening Examination (CCSE; Jacobs et al., 1977); the Neuropsychological Impairment Scale (NIS; O'Donnell & Reynolds, 1983); and the Cognitive Laterality Battery (CLB; Gordon, 1986).

The MMSE (Folstein et al., 1975) is a cognitive screening instrument widely used in clinical practice and has been established as a valid and reliable test of cognitive function (Folstein et al., 1975). It measures orientation, registration, short-term memory, attention and calculation, and language. A score of 17 or less is considered to be severe cognitive impairment, 18 to 24 to be mild to moderate impairment and 25 to 30 to be normal or borderline impairment. It has the advantage of being brief, requiring little training in administration and interpretation, free to use, and is designed to assess specific facets of cognitive function (Small et al., 1997). The MMSE has been found to have high sensitivity for detecting moderate to severe cognitive impairment as well as satisfactory reliability and validity (see Nelson and colleagues [1986] for a review). The MMSE can be utilised as a brief screening tool as well as for assessing changes in cognitive function over time (Brayne et al., 1997).

It must be noted, however, that the MMSE has been found to be sensitive to educational level in populations where educational levels are low (Escobar et al., 1986; Liu et al., 1994). Therefore, the cut-offs used to identify cognitive impairment may need to be adjusted for people who misuse alcohol with few years of formal education (Crum et al., 1993; Cummings, 1993). Furthermore, the MMSE has been criticised for not being sensitive enough for those in various cultures where the education levels are low and participants may fail to respond correctly to specific items (Escobar et al., 1986; Iype et al., 2006; Katzman et al., 1988; Liu et al., 1994). Because of this, it is necessary (and often practised) to amend and adjust aspects of the MMSE to increase applicability to a particular cultural setting. For example, a Hindi version of the MMSE, the Hindi Mental State Examination (Ganguli et al., 1995) was designed to address some of the cultural problems with the MMSE and to make it more applicable to an Indian cultural setting.

Most research evaluating the accuracy, reliability and validity of the MMSE has been in the assessment of age-related cognitive impairment and dementia, whereas research in the field of alcohol and substance misuse is limited. However, the MMSE has been utilised in substance-misuse research (Smith et al., 2006). Additionally, it has been highlighted that the MMSE mainly assesses verbal cognitive function and is limited in assessing non-dominant hemisphere skills and executive functions (Bak & Mioshi, 2007). This could lead to frontal-dysexecutive and visuospatial symptoms going undetected.

The CCSE (Jacobs et al., 1977) was designed to screen for diffuse organic mental syndromes. The CCSE has 30 items that provide information on the areas of orientation, digit span, concentration, serial sevens, repetition, verbal concept formulation and short-term verbal memory. A score of less than 19 has been suggested as indicative of organic dysfunction (Haddad & Coffman, 1987; Hershey et al., 1987; Jacobs et al., 1977). As with most cognitive screening instruments, the CCSE has been studied extensively in people with dementia (Nelson et al., 1986). It has been found to have adequate reliability and validity in detecting cognitive impairment (Foreman, 1987; Villardita & Lomeo, 1992). However, the CCSE has been found to be sensitive to age and education (Luxenberg & Feigenbaum, 1986; Omer et al., 1983) and has been found to have a high false-negative rate and hence low sensitivity (Nelson et al., 1986; Schwamm et al., 1987). Furthermore, Gillen and colleagues (1991) and Anderson and Parker (1997) reported that the CCSE did not adequately distinguish between cognitively impaired and non-impaired people who misuse substances.

The NIS is a 50-item scale that has been designed to identify brain damage. The reliability and validity of the NIS has been previously reported in normal and neuropsychiatric populations (O'Donnell et al., 1984a and 1984b), as well as having a sensitivity of between 68 and 91% and a specificity of between 43 and 86% (O'Donnell et al., 1984b). Errico and colleagues (1990) further reported predictive validity, and test–retest reliability in a sample of people undergoing assisted alcohol withdrawal.

The CLB was developed to measure visuospatial and verbosequential functioning, with tests administered on a sound/sync projector, and takes 80 minutes for administration. However, the CLB has been reported to have limited clinical utility in the assessment of cognitive function in an alcohol-dependent population (Errico et al., 1991).

Addenbrooke's Cognitive Evaluation (ACE; Mathuranath et al., 2000) was developed as a brief test of key aspects of cognition which expanded on the MMSE by assessing memory, language and visuospatial abilities in greater depth, as well as including assessment of verbal fluency. ACE is designed to be sensitive to the early stages of dementia (Mioshi et al., 2006) and was found to detect dementia earlier and discriminate more between different subtypes of dementia than the MMSE (Mathuranath, 2000). To comprehensively assess cognitive impairment, ACE can be used in a cognitive test battery along with tests which assess other cognitive domains (Lezak, 1995; Spreen & Strauss, 1998), such as the Trail-Making Test, Part B (Army Individual Test Battery, 1944) or the Block Design subtest of the revised Wechsler Adult Intelligence Scale (Wechsler, 1981). Mioshi and colleagues (2006) developed a revised version of the test, ACE-R, which addressed previously-identified weaknesses in the original test and made it easier to administer. Bak and colleagues (2007) found that ACE-R has a good specificity for the detection of dementia (94%) with a specificity of 89% (at a cut-off score of 88/100). The ACE-R is administered as a bedside test, takes approximately 16 minutes to complete and consists of five sections each designed to assess a specific cognitive domain (Mioshi et al., 2006). Although the ACE-R takes longer to administer than some other tests for cognitive impairment, such as the MMSE, it has been found to have a high level of patient acceptability and can be administered without specialist training (Larner, 2007). The ACE and ACE-R tests are published in 19 languages, although no evidence was found regarding the effect of cultural or educational differences on testing outcomes.

The ACE has been used in screening for cognitive dysfunction in people who misuse alcohol (Gilchrist & Morrison, 2004), although no research in this field using ACE-R could be identified. Additionally, research into the efficacy and sensitivity of ACE-R in assessing substance induced cognitive impairment is negligible. It has been suggested, however, that it is possible to extrapolate the validity of the ACE as an instrument to assess age-related cognitive impairment and apply it in assessing cognitive impairment in people who misuse alcohol. The increased sensitivity of ACE in relation to the MMSE may mean that it is subtle enough to identify people who misuse alcohol that have mild cognitive impairment who are able to function successfully in the community but have a history of non-engagement with alcohol services (personal communication, Ken Wilson, October, 2010).

Childhood abuse

A history of physical and/or sexual abuse is common in patients seeking treatment for alcohol misuse, particularly women (Moncrieff et al., 1996). Patients identified with childhood trauma who wish for further intervention should be referred to appropriate services once they have reached a degree of stability in terms of their alcohol use (guideline on PTSD; NCCMH, 2005).

Family and relationships

Relationships with partners, parents, children and significant others are often affected by alcohol misuse (Copello et al., 2005). Families and carers also suffer significantly in their own right, with an increased incidence of mental disorder (Dawson et al., 2007). Involvement of partners or family can help identify the needs of the help seeking individual. The prevalence of alcohol misuse in the victims and perpetrators of domestic violence provides an important rationale for the exploration of these issues.

Employment

The status of the individual's occupation is significant in terms of the individual's ability to remain economically active. Past employment history may indicate the individual's capacity to obtain and retain employment. Employment history might provide insights into factors that maintain the individuals drinking status that need to be explored. Those assessing employed individuals will need to consider potential risks to the person, colleagues and the public because of excessive drinking (for example, when the individual has responsibility for the safety of others).

Criminality and offending

Assessment of criminality and offending behaviour should encompass a number of factors including the presence and onset of criminal activity, the severity of offending behaviour, the relationship between offending and alcohol consumption and/or alcohol withdrawal, and the presence of violence and aggressive behaviour, and hence risk to others. Liaison with criminal justice services is necessary to ensure that appropriate co-ordination of care, and effective communication and information-sharing protocols, are in place.

Fitness to drive

For people who misuse alcohol and continue to drive a motor vehicle, clinical staff have a duty to advise the individual that it is the duty of the license holder or license applicant to notify the DVLA of any medical condition that may affect safe driving. There are circumstances in which the license holder cannot or will not notify the DVLA. Clinicians will need to consult the national medical guidelines of fitness to drive in these circumstances (DVLA, 2010).

What outcome should be measured?

Assessment of alcohol consumption (for example, intensity and frequency of drinking) is a basic component of outcome monitoring. For example, Emrick (1974) states that monitoring abstinence post-treatment is a significant predictor of psychosocial functioning. Alcohol-related problems or harm have also been suggested to be important in outcome monitoring. Longabaugh and colleagues (1994) state that outcome measurement should contain a range of assessment domains and include life functioning (such as physical health and social functioning). Alcohol problems are the only assessment domain significantly associated with drinking outcome measures (PDA, DDD, first drink) (Project MATCH Research Group, 1997 and 1998). This suggests that alcohol-related problem outcome measures should be assessed in addition to alcohol consumption.

How should outcome be measured?

When selecting a suitable tool for outcome monitoring there are a number of factors that need to be considered, as suggested in a review of by Raistrick and colleagues (2006). An outcome monitoring tool should be:

• universal and not constrained by any particular substance or social group
• have proven validity and reliability, and have published psychometric properties
• sensitive to change
• easily readable and in a neutral language
• either practitioner-completed, self completed, or a combination of both
• appropriate for the clinical population.

The outcome measure that is applicable to all tiers of services is assessing the level of alcohol consumption by interviewing the patient about their quantity and frequency of alcohol consumption, but the use of a formal measure will improve the reliability and validity of measurement (Sobell et al., 1979). The most valid and reliable measures of alcohol consumption include a diary method to obtain drinking data (Sobell & Sobell, 2003). However, measures such as the Timeline Followback questionnaire (Sobell & Sobell, 2003) are more feasible to administer in the research setting rather than a routine clinical setting. Some clinical services in the UK use prospective weekly drinking diaries that are self-completed by service users, but their reliability and validity is unknown.

A number of assessment tools have been designed specifically for outcome measurement in addiction treatment. They all measure multiple domains of functioning, but their comprehensiveness, utility and specificity to alcohol treatment vary. The most widely used tools for outcome measurement are the Addiction Severity Index (ASI), AUDIT, the Maudsley Addiction Profile (MAP), the Christo Inventory for Substance Misuse Services (CISS), the Comprehensive Drinker Profile (CDP), the Routine Evaluation of the Substance-Use Ladder of Treatments (RESULT) and the Treatment Outcomes Profile (TOP). The GDG evaluated the clinical utility of these tools in alcohol treatment on the extent to which each tool has sufficient validity and reliability data in an alcohol dependent population and if the tool has high usability (that is, it is easy to read and understand, it does not require extensive training for use and it is brief). Table 19 describes the characteristics of the outcome measurement tools identified.

Table 19

Characteristics of routine outcome monitoring tool.

The GDG excluded the ASI because it was excluded in the earlier review of primary outcome tools and also is too lengthy for use as an outcome monitoring tool. The CISS, CDP, MAP and RESULT were also excluded because they have not been adequately validated in an alcohol-dependent clinical sample in the UK. Lastly, the TOP is primarily used in a drug misusing population with only limited psychometric data for alcohol-dependent clinical samples. The protocol for reporting TOP states explicitly that ‘the reporting of the TOP for adult primary alcohol users is not required’ (National Treatment Agency for Substance Misuse, 2010) and therefore the TOP is not being applied in routine practice.

Based on these criteria, a GDG consensus-based decision was made that the AUDIT has the greatest utility as a routine outcome monitoring tool to evaluate drinking-related outcomes. The AUDIT questionnaire is already widely used. It contains several relevant drinking domains in addition to alcohol consumption (problems and dependence). The time taken to complete the AUDIT (less than 2 minutes) also lends itself to use in routine practice. The AUDIT-C (Bush et al., 1998) is a three-item version of the AUDIT which measures only alcohol consumption; that is, frequency of drinking, quantity consumed on a typical occasion and the frequency of heavy episodic drinking (six or more standard drinks on a single occasion). Bush and colleagues (1998) reported that the AUDIT-C performed better than the full AUDIT in detecting heavy drinking and was just as effective as the full AUDIT in identifying active alcohol misuse or dependence. The study also found that using a cut-off of 3 out of a possible 12 points, the AUDIT-C correctly identified 90% of active alcohol ‘abuse’/‘dependence’ and 98% of patients drinking heavily. However, other studies have reported that a cut-off of 5 or more for men and 4 or more for women results in the optimal sensitivity and specificity for detecting any alcohol-use disorders (Dawson et al., 2005b; Gual et al., 2002). In addition, the AUDIT-C has been found to be equally effective in detecting alcohol-use disorders across ethnic groups (Frank et al., 2008). However, it should be noted that the AUDIT-C has been reported to have a high false positive rate when used as a screening tool (Nordqvist et al., 2004). Nevertheless, the ease of use and the already established relationship between frequency/quantity of drinking and alcohol misuse/dependence give the AUDIT-C credence for the use of outcome monitoring.

The APQ has been widely used in alcohol treatment outcome studies as a measure of alcohol-related problems in the UK (for example, Drummond, 1990; Drummond et al., 2009; UKATT Research Team, 2005). Furthermore, it is quick and easy to administer. Therefore, the APQ can be used to measure alcohol-related problems in conjunction with a drinking-related outcome tool such as the AUDIT-C. However, the ten-item AUDIT still has the advantage of measuring a wider range of domains in one simple validated questionnaire and therefore more readily lends itself to routine clinical outcome monitoring.

When should outcome be measured?

Most changes in drinking behaviour and the largest reduction in severity of drinking occurs in the first 3 months of treatment (Babor et al., 2003; Weisner et al., 2003). Initial benefits in drinking-related outcomes may be more apparent at 3 months, but other non-drinking domains such as social functioning and global health may need longer to show improvements following treatment. Because there is a high attrition rate in many alcohol services this can result in poor response rates in routine outcome monitoring. This underlines the importance of routine session-by-session measurement and the utility of a brief measure such as the AUDIT or a prospective weekly drinking diary. The latter requires considerable cooperation of service users and is of unknown reliability and validity.

A framework for assessment for children and young people with alcohol problems

As with the adult assessment, the use of any assessment tool needs to be set in context. The context here is that all children who are beyond initial identification should be offered an assessment within specialist CAMHS. Although recommendations are made below for the use of specific measures to assess the nature and extent of the alcohol misuse and related problems, it was also the view of the GDG that the assessment should take place in the context of a comprehensive overall assessment of the mental health, educational, and social care needs of the children and young people, in line with current best practice (Department for Education and Skills, 2007). In common with good assessment practice in CAMHS services, the involvement of parents, carers and others (for example, schools) is an essential part of any assessment. It should also be noted that parents not only have a key role as informants, advisors and participants in the process of assessment, but they also have a key role to play in the development of any future treatment plans. It is therefore important that wherever possible they are involved from the beginning.

The overall structure of assessment (at least for the assessment of alcohol misuse) is provided by the assessment tools reviewed above. However, whatever assessment tool is used from both the child and adult literature, (Harrington et al., 1999, and see Chapter 5) suggest that the following domains need to be considered as part of any assessment of alcohol related problems in children and young people:

• Alcohol use – consumption, dependence features and associated problems
• Comorbid substance misuse – consumption, dependence features and associated problems
• Motivation
• Self-efficacy
• Other problem domains

  –

  Physical history and problems

  –

  Mental health and problems

  –

  Social functioning

  –

  Educational attainment and attendance

  –

  Peer relationships

  –

  History of abuse and trauma

  –

  Family functioning and relationships

• Risk assessment
• Developmental needs of the young person
• Treatment goals
• Obtaining consent to treatment
• Formulation of a care plan and risk management plan.

An additional point to bear in mind is the use of further informants. For example, in terms of the assessment of consumption, the use of other informants such as parents, carers or schools may assist in detailing the history of consumption, and in clarifying the level and veracity of use.

As was identified in the introduction, the presentation of alcohol misuse or dependence does not typically follow the pattern seen in adults. More often, a pattern of binge drinking is observed that is often accompanied by drug misuse. It is important, therefore, to detail both the pattern of drinking and the comorbid drug misuse. It should also be noted that adolescents may have lower prevalence of withdrawal symptoms along with a lower tolerance. Both of these factors may contribute to continued high alcohol intake, particularly of binge drinking, with consequent serious implications for psychological and physical health but without the ‘warning signs’ of emerging withdrawal symptoms.

History of trauma and abuse

It has already been noted that comorbidity of substance misuse is significantly higher in adolescents who misuse alcohol. It is also important to note that alcohol misusing adolescents have a significant increased rate of physical abuse (by a factor of 6 to 12) and a significant increased rate of sexual abuse (by up to a factor of 20) (Clark et al., 1997a). Given that it is possible that these histories may have a significant aetiological role in the development of alcohol misuse, it is important that these issues are part of assessment. It is also likely that a history of trauma has an impact for the likely comorbidity, for example the existence of PTSD (Clark et al., 2003), and also that it may be associated with poor response to treatment and the need for more complex treatment interventions.

Content of the clinical assessment

The literature review identified a number of components of a structured clinical interview. These included assessment of the current extent and history of drinking, associated potential for withdrawal, the likelihood of withdrawal, the need for review of associated physical health problems, the examination of mental health and the impact of alcohol on social, personal, occupational and educational functioning. It also identified that the impact of alcohol on the family would be an important issue also to be considered. Considerable emphasis on the literature reviewed was placed on the importance of engaging people with alcohol-related problems in treatment and negotiating appropriate goals. It is clear from the literature that for people who are moderately and severely dependent drinkers, the initial goal should be one of abstinence. For others who are harmful and mildly dependent drinkers, it may be possible to consider a reduction in drinking as a reasonable treatment goal. However, past history of unsuccessful attempts to moderate drinking should be born in mind when making these assessments.

There is little evidence that indicates the identification and assessment methods needed for assisted withdrawal in children and young people. Therefore, the GDG makes a consensus-based decision to extrapolate from the review of the adult literature and combine this with expert opinion. The group concluded that a comprehensive assessment and an assessment for assisted withdrawal should be offered to all children and young people with established binge drinking, an AUDIT score of more than 15 and those who consume more than five units per day, but this decision should also take into consideration other factors such as age, weight, previous history of alcohol misuse and the presence of co-occurring disorders.

The review of formal assessment measures also considered a number of measures of motivation (readiness to change). It was not felt by the group that the quality of these measures (in part because of impracticality of these measures which were designed primarily for use in research) warranted their use in standard clinical care. However, a consideration of a patient's readiness and/or motivation for change is a vital part of assessment.

Physical investigations

An awareness of and inquiry into the nature of commonly presenting physical health problems with alcohol misuse are important. This guideline, and other related NICE guidelines (NICE, 2010a and 2010b), considered the value of biomarkers; for example, liver function tests as indicators for diagnosis of alcohol-related disorders. From the reviews conducted for this and the other NICE guidelines it was concluded that these measures have insufficient sensitivity and specificity compared with validated assessment methods such as the AUDIT. However, for people with specific physical health problems, for those whom regular feedback on a particular biological marker may act as a motivational tool and those for whom pharmacological treatments may require liver function tests (for example, naltrexone and disulfiram), these measures may have an important part to play in the ongoing treatment and management of alcohol misuse. No evidence was identified in this or the other NICE guidelines (2010a and 2010b) to support the use of other biomarkers (for example, hair analysis) for routine clinical use in assessment or outcome monitoring of alcohol misuse.

Assessment of comorbid substance misuse

It is recognised that smoking, drinking and drug taking behaviours cluster together (Farrell et al., 2001) and that excessive drinkers with high AUDIT scores are more likely to have used drugs in the past (Coulthard et al., 2002). Therefore, the evidence suggests that co-existing substance misuse should be assessed. Clinical assessment should include the type of drug and its route of administration, the quantity and the frequency with which it is used.

Assessment of comorbid mental health problems

Mental health problems which co-exist with alcohol misuse can have a significant impact, both on the treatment and long-term outcome of the alcohol-related problem. However, depression and anxiety can often develop as a consequence of alcohol misuse. At assessment there is no reliable way of determining whether a comorbid mental health problem is primary or secondary to alcohol misuse. This means that symptoms of comorbid mental disorders need to be monitored throughout the course of assessment and treatment. A common presentation in alcohol misuse is suicidal ideation. This needs to be assessed and actively managed as part of an overall risk management process. The GDG considered that as, a minimum, the re-assessment of common mental disorders should occur 3 to 4 weeks following abstinence from alcohol. At this point, consideration may be given to treatment. NICE guideline for the management of these disorders should be consulted.

Cognitive impairment

Cognitive impairment is present in most people who misuse alcohol presenting for treatment. These impairments, which may be transitory, are, however, often missed in the initial assessment. The evidence reviewed suggested that the MMSE has reasonable validity as an initial identification tool and should be supplemented with specific questions to detect duration extent or functional impairment. There is also evidence to suggest that the ACE-R has good sensitivity for diagnosing mild cognitive impairment. However, it does not assess all aspects of cognitive function and should be used as a part of a specialist comprehensive assessment test battery in conjunction with an executive test such as the Block Design subtest of the Wechsler Adult Intelligence Scale (WAIS-III) (Wechsler, 1945/1997) or the Trail-Making Test, Part B (Army Individual Test Battery, 1944). It is not possible to conduct an effective cognitive assessment in people who misuse alcohol who are actively drinking. Unless there is evidence of gross cognitive impairment that may require further and immediate investigation, the GDG took the view that adequate assessment of cognitive impairment is best left until 3 to 4 weeks following abstinence from alcohol. At this point if significant cognitive impairment persists, it should be subject to more formal assessment including conducting a more detailed history and neuropsychological testing. Those patients presenting acutely with a confused state and significant memory loss may be suffering from WE and should be assessed and treated accordingly (see NICE, 2010b).

Organisation and delivery of assessment

The evidence for the organisation and delivery of the range of assessment was reviewed. This included a review of the currently recommended assessment systems in England and, in particular, the MoCAM stepped-care framework. This approach begins with an initial case identification/diagnostic assessment. Here, the emphasis is on brief assessments that can be administered by staff in a range of services in healthcare and related settings. There is good evidence from the assessment tools reviewed above that scores on measures such as the AUDIT and SADQ provide a useful indication of the appropriate level of intervention needed. There is also evidence that people who misuse alcohol can be assessed in a relatively brief triage assessment. The guideline also reviewed the evidence for the factors to be considered in a withdrawal assessment. Finally, the indications for and content of comprehensive assessment was reviewed. In summary the GDG felt that a stepped approach to assessment in line with that set out in MoCAM (Department of Health, 2006a) was appropriate.

Outcome monitoring

The GDG reviewed the evidence for the use of routine outcome monitoring. A range of assessment tools were considered as part of the overall view of assessment tools. Although these measures are effective at identifying the presence or severity of the disorder, most were felt unsuitable or impractical for routine outcome measurement. The evidence suggested that the AUDIT questionnaire provides a valid, reliable and feasible method to monitor outcome in routine clinical care. Prospective weekly drinking diaries, whilst widely used in clinical services, are of unknown reliability and validity. The routine use of breath alcohol concentration measurement was not supported by the evidence either in initial assessment or routine outcome monitoring, although it has a useful place in monitoring abstinence in the context of an assisted withdrawal programme. The GDG therefore favoured the AUDIT (specifically the first three questions from the questionnaire with subsequent questions only used for 6-month follow-up) as a routine measure, but recognised that in some services, especially Tier 3 and Tier 4 specialist services, an additional, more detailed assessment measure may also be used. The GDG also favoured the APQ as an outcome monitoring tool when assessing alcohol-related problems.

Assessment tools

The review of assessment tools identified a number of measures which had sufficiently robust psychometric properties to be used in routine clinical care. In addition to these factors, the GDG also used its expert knowledge to assess the benefit and feasibility of their use in routine care. As a case identification tool (that is, one that would indicate whether or not further treatment was required) the AUDIT questionnaire is the most appropriate instrument. On occasions where the AUDIT questionnaire was not available and/or not practical, then a simple typical daily alcohol consumption measure could also be used as an indicator of potential need for treatment. For people suspected of having alcohol dependence, the use of the SADQ or the LDQ were supported by the GDG as effective instruments to measure the severity of alcohol dependence in order to guide further management. For assessing the extent of problems associated with alcohol misuse, the APQ was identified as meeting all the necessary criteria. In addition, on the basis of the NICE guideline on the management of alcohol-related physical complications review (NICE, 2010b), the CIWA-Ar was judged to be the most appropriate instrument to measure alcohol withdrawal symptoms.

Content of the clinical assessment and the organisation and delivery of assessment systems

It is important to recognise that the use of individual assessment tools alone, such as those identified above, does not constitute a comprehensive assessment. The evidence suggested that, in addition to a historical and recent history of drinking, the associated physical and mental health problems, and the impact on health and social and economic problems, should also be assessed. This section also identified the importance of the impact on family (including, importantly, children). It is also important to recognise that a key aspect of effective assessment is the process of engaging people and identifying treatment goals. For example, determining whether abstinence, which is the initial preferred goal for moderately and severely dependent drinkers, or moderation of alcohol consumption is the preferred goal. The GDG therefore decided to provide detail on the content of the range of assessment domains. The GDG also reviewed the evidence for the organisation and delivery of assessment systems and supported the established system recommended within MoCAM (Department of Health, 2006a). This may require additional specialist assessment resources and systems to ensure that individuals have the capacity and competency to deliver these assessments.

Physical investigations

The review for this guideline (based in significant part on parallel work undertaken on other NICE guidelines; NICE, 2010b) established that physical investigations (in particular, blood tests including measures of liver function) are not sufficiently sensitive or specific measures for routine use in specialist alcohol services. However, biomarkers can have added benefit as motivational tools by providing feedback on progress and in assessing suitability for some pharmacological interventions (for example, naltrexone and disulfiram). The GDG also considered that the measurement of breath alcohol is a useful, objective part of the clinical monitoring in the management of assisted alcohol withdrawal.

Assessment of comorbid substance misuse

The presence of comorbid substance misuse is associated with poorer outcomes for those with alcohol misuse and the GDG reviewed evidence on this along with the recommendation in the NICE (2007a) guideline on psychosocial management of substance misuse. It was agreed that assessment of comorbid drug misuse should therefore be a part of routine assessment of alcohol misuse. Consideration should be given to the use of biological testing (for example, of urine or saliva samples) as part of a comprehensive assessment of drug misuse, but clinicians should not rely on it as the sole method of diagnosis and assessment.

Assessment of comorbid mental health problems

Comorbid mental health problems are a common presentation in people who misuse alcohol. It is important that this is assessed at initial presentation. However, it should be noted that for many people, symptoms of, for example, depression and anxiety may remit following 3 to 4 weeks of abstinence from alcohol. It is therefore often not appropriate or necessary to instigate treatment for the disorder at the point of the initial assessment. Careful monitoring and reassessment of mental health symptoms following abstinence are an important part of the assessment procedure. Treatment of mental health disorders that persist beyond 3 to 4 weeks after abstinence should be considered.

Routine outcome monitoring

Routine outcome monitoring is an essential part of any effective healthcare system provision. The AUDIT questionnaire was identified as the most reliable and feasible measure for routine outcome monitoring. Prospective drinking diaries are of unknown reliability and validity. The APQ was also identified as beneficial for the assessment of alcohol-related problems when monitoring treatment outcome.

Competence of staff

It is essential that clinicians performing assessments of people who misuse alcohol should be fully competent to do so.

Children and young people

Due to the lack of sufficient evidence specifically for children and young people, the GDG decided to adopt a modified version of the assessment framework adopted for adults. As with the adult review the GDG favoured the use of the AUDIT tool as a case identification/screening tool and this is consistent with the approach adopted the NICE prevention and brief intervention guideline (NICE, 2010a) However, the GDG decided to adjust the threshold for AUDIT in light of evidence that this increased the sensitivity for adolescent alcohol misuse. For a more comprehensive assessment, the GDG recommended two possible assessment tools and the integration of any assessment of alcohol misuse into a comprehensive assessment of the needs of the child or young person.

Identification and assessment

Assessment in all specialist alcohol settings

Community settings

In a community setting, a person undergoing assisted withdrawal lives in their own accommodation throughout the treatment. A spectrum of treatment intensity is also possible. Day hospital treatment (sometimes known as ‘partial hospitalisation’; see Mee-Lee et al., 2001) may involve the patient attending a treatment facility for up to 40 hours per week during working hours, Monday to Friday, and returning home in the evening and weekends. This facility may be located within an inpatient or residential rehabilitation unit, or may be stand-alone. It is likely to be staffed by a multidisciplinary team, with input from medical and nursing staff, psychologists, occupational therapists, social workers, counsellors, and other staff specialising in debt, employment or housing issues. Other community assisted withdrawal programmes may invite the patient to attend for appointments with a similar range of multidisciplinary staff, but at a much lower frequency and intensity (for example, once or twice a week), or they may be provided by GPs often with a special interest in treating alcohol-related problems. Alternatively, staff may visit the patient in their own home to deliver interventions. Between these two options are most intensive community-based options, where an increased frequency of community visits and some limited use of office or team-based treatment may form part of an intensive community programme.

Inpatient and residential settings

In inpatient and residential settings, the service user is on-site for 24 hours a day for the duration of assisted withdrawal. Inpatient and residential settings encompass a spectrum of treatment intensity. At one end lie specialist units within either acute medical or psychiatric hospitals, dedicated to the treatment of alcohol or drug problems (known as ‘inpatient units’). Such units have specialist medical and nursing input available 24 hours a day, and are staffed by a multidisciplinary team that may also include psychologists, occupational therapists, social workers, counsellors, and other staff specialising in debt, employment or housing issues. At the other end of the spectrum are facilities usually known as ‘residential rehabilitation’ units, which are usually provided by the non-statutory sector and not sited within hospital premises. Although the goal of such units is usually the provision of longer-term treatment (3 to 12 months) aimed at enhancing the patient's ability to live without using alcohol, increasingly they also provide an initial period of assisted withdrawal. Such units may also have access to medical and nursing input over the full 24-hour period, but this is usually at a lower level of intensity and more likely to utilise non-specialist staff (for example, GPs). Such units are more likely to adopt a ‘social model’ rather than a ‘medical model’, and may be staffed by both professionals and individuals in recovery. In addition, a number of prisons may offer a high level of medical supervision including, where necessary, admission to the hospital wing of the prison.

Studies comparing different settings for assisted withdrawal

Apart from the Hayashida and colleagues (1989) study, the studies discussed above were observational in design and participants were only matched for severity of alcohol dependence. Furthermore, although these studies indicated that it is feasible for assisted withdrawal to take place in a community setting for a severely dependent population, it is probable that a number of patients with significant comorbidities and previous history of seizures where excluded. As these patients form a significant proportion of those who are referred to and receive inpatient or residential assisted withdrawal, caution is needed when interpreting these results.

Only one randomised trial (Hayashida et al., 1989), conducted in a US Department of Veterans Affairs (VA) medical centre, compared the effectiveness and safety of inpatient (n = 77) and outpatient (n = 87) assisted withdrawal. Patients with serious medical or psychiatric symptoms, predicted DTs and a very recent history of seizures were excluded from this study. The authors reported that more inpatients than outpatients completed assisted withdrawal. However, inpatient treatment was significantly longer and more costly than outpatient treatment. Additionally, both groups had similar reductions in problems post-treatment when assessed at 1- and 6-month follow-up. Although abstinence was statistically significantly higher for the inpatient group at 1-month follow-up, these differences were not observed at 6-month follow-up. The authors concluded that outpatient assisted withdrawal should be considered for people with mild-to-moderate symptoms of alcohol withdrawal.

Stockwell and colleagues (1991) compared a retrospective inpatient sample (n = 35) with a group receiving home-based assisted withdrawal (n = 41). The two samples were matched for age, sex and drinking severity. Patients undertaking home-based assisted withdrawal were severely dependent (SADQ score = 28.7; mean alcohol consumption 174.6 units per week) and had a high level of alcohol-related problems (APQ score = 4.6). The authors reported that home-based assisted withdrawal was as safe and effective for a severely dependent population as inpatient care. However, the matched inpatient sample did not include anyone with severe alcohol withdrawal syndrome or physical or psychiatric symptoms and, therefore, is not representative of an inpatient population.

Bartu and Saunders (1994) also compared people undertaking home-based assisted withdrawal (n = 20) with patients in an inpatient specialist unit (n = 20). Patients were matched for age, sex, presence of social support, absence of medical complications and severity of withdrawal symptoms. It was reported that home-based assisted withdrawal was as beneficial as inpatient assisted withdrawal. It should be noted, however, that the matched inpatient sample was not representative of a typical inpatient, who may be severely dependent and have several complications.

Parrott and colleagues (2006) compared alcohol-focused outcomes and cost of residential (n = 54) with any day (n = 49) settings for assisted withdrawal in the UK and reported similar alcohol-focused outcomes (PDA and DDD). This paper mainly discusses cost implications and is reviewed in the health economic section (Section 5.29.7).

In a comparison between home-based assisted withdrawal (n = 29) and day hospital services (n = 36) in severely dependent patients, Allan and colleagues (2000), in a UK-based study, evaluated the types of patients selected for home-based assisted withdrawal, its safety and efficacy, and patient satisfaction and involvement in further treatment. Participants in both groups were severely dependent (two thirds had a SADQ score of over 30), although the day hospital group drank significantly more at baseline (home-based group = 178 units, day hospital group = 194 units in the week before assisted withdrawal). Furthermore, although both groups had alcohol-related problems, as assessed by the APQ, the day hospital group had significantly more severe problems and social instability. The authors reported that there were no significant differences between the groups in the proportion of participants who completed assisted withdrawal, complication rates (which were low) and uptake of treatment post withdrawal. However, it should be noted that this study did not match participants in both settings but aimed to assess the characteristics of the patients who use home-based and day hospital assisted withdrawal.

Prospective studies evaluating outpatient assisted withdrawal

Further studies assessing the treatment outcomes and characteristics of patients in various settings were identified from the literature search. These studies were open prospective studies and aimed to evaluate the safety and efficacy of outpatient assisted withdrawal. Feldman and colleagues (1975) evaluated an outpatient treatment programme for alcohol withdrawal (n = 564). The authors reported that only 47% required outpatient assisted withdrawal and 19% required inpatient assisted withdrawal. Outpatient assisted withdrawal was successful and had a low dropout rate of 14%. However, the authors attributed this success to the early involvement of the family, the use of withdrawal medication and involvement in peer group therapeutic activity. The results of an earlier study reflected these findings (Alterman et al., 1988). The investigators reported that ambulatory assisted withdrawal was relatively successful for mild-to-moderate alcohol withdrawal symptomatology.

Soyka and Horak (2004) assessed the efficacy and safety of outpatient assisted withdrawal in a German open prospective study. Alcohol dependent participants were excluded if they presented with severe alcohol-related disorders, such as seizures or psychosis, or major psychiatric and medical comorbidity. Some participants referred to the treatment clinic had to be admitted for inpatient care (n = 348), leaving 331 patients to be treated in an outpatient setting. The study reported very high completion rates (94%) for patients in an outpatient assisted withdrawal programme. Furthermore, outpatient assisted withdrawal was associated with increased participation in further treatment (91% of initial sample). Soyka and Horak (2004) additionally found that of those who completed assisted withdrawal successfully, all entered either motivationally- or psychotherapy-based treatment.

Stinnett (1982) evaluated the effectiveness and safety of 116 participants referred for outpatient assisted withdrawal in an alcoholism treatment centre. Fifty per cent completed treatment, and 89% of these completers went on to continue with follow-up rehabilitation treatment. Collins and colleagues (1990) assessed the efficacy of a UK-based outpatient alcohol withdrawal programme. Of those deemed suitable for outpatient assisted withdrawal (n = 76; 44% of all referrals), 79% successfully completed the treatment. These patients were severely alcohol dependent (91% had an SADQ score greater than 30). However, not all studies have reported such favourable completion rates. For example, in a severely dependent sample of 26 patients (77% with a SADQ score greater than 31), Drummond and Chalmers (1986) reported that only 23% of patients completed assisted withdrawal and 19% attended a follow-up 1 month later.

In a UK-based RCT, Alwyn and colleagues (2004) evaluated the addition of a brief psychological intervention to GP-managed home-based assisted withdrawal. The psychological intervention consisted of five 30-minute sessions with motivational, coping skills and social support approaches. The study reported that both the control and the psychological intervention group (total n = 91) showed significant improvements in drinking outcomes from baseline to follow-up (3- and 12-month) indicating that home-based assisted withdrawal was effective. In addition, the psychological intervention group showed significantly greater improvements than the control group at 12-month follow-up. These results suggest that there is benefit in adding a brief psychological intervention to assisted withdrawal.

From the patients' perspective, it has been suggested that gains made in inpatient assisted withdrawal may not be easily transferable to the patient's home and social environment (Bischof et al., 2003). Undertaking assisted withdrawal in a home or outpatient setting enables the patient to retain important social contacts that may facilitate their attempts to achieve abstinence as well as subsequent rehabilitation. Patients can continue in employment (if appropriate) and be in a familiar environment with family support, which may help to minimise stress and anxiety and help to motivate them. It has also been suggested that the home environment is also less stigmatising than an inpatient setting for assisted withdrawal (Allen et al., 2005). In a study assessing patients' perceptions and fears of alcohol withdrawal, Allen and colleagues (2005) found that patients were fearful and concerned about the psychiatric residential setting for assisted withdrawal and expressed feelings of stigmatisation associated with being in an ‘institutional’ setting. The authors also reported no difference in patient satisfaction between a home and outpatient setting for assisted withdrawal. Additionally, patient satisfaction with outpatient assisted withdrawal services have also been found to be high when administered in an intensive day programme (Strobbe et al., 2004). Stockwell and colleagues (1990) found that three quarters of patients preferred their home as the setting for assisted withdrawal, and two fifths and one third were unwilling to undergo withdrawal in, respectively, a psychiatric hospital and a general hospital. The patients also emphasised the importance of support from the nurse supervising their assisted withdrawal, the breathalyser, medications, telephone support service and the involvement of supporters, familiar surroundings, privacy and confidentiality, and being able to stay with their family.

Another factor that may be relevant to the provision of home or outpatient assisted withdrawal is availability of treatment capacity. An early report (Stockwell et al., 1986) revealed that in the Exeter Health Authority, GPs arranged as many home-based assisted withdrawals as hospital-based. However, of the home-based assisted withdrawals, two fifths were unsupervised. Approximately a third of GPs were reluctant to take medical responsibility for home-based assisted withdrawal, but of those who were happy to, they reported a preference for this setting. Winters and McGourty (1994) also surveyed GPs in Chester and Ellesmere Port. Approximately 60% reported that they provided home-based assisted withdrawal from their practices. However, 10% believed specialist help was required. Additionally, they reported that unsuccessful home-based assisted withdrawal was usually due to lack of support at weekends and lack of patient motivation. Over 20% of Northumberland GPs reported carrying out home-based assisted withdrawals in the last year (Kaner & Masterson, 1996). Similar to Winters and McGourty (1994), most GPs stressed the importance of having daily supervision as well as more information about the process of assessing patients for suitability for home-based assisted withdrawal.

Medical or psychiatric comorbidities

For the majority of people who misuse alcohol, outpatient or home-based assisted withdrawal appears to be safe, viable and effective (see above). However, for a minority of patients, a non-residential setting for assisted withdrawal may be inappropriate or unsafe. An inpatient setting may be more appropriate for the management of moderate to severe withdrawal symptoms such as DTs and seizures, comorbid medical, surgical and psychiatric problems (for example, suicidal ideation), pregnancy, or if the patient is not able to take medication by mouth (Bischof et al., 2003; Blondell et al., 2002; Blondell, 2005; Dukan et al., 2002; Ferguson et al., 1996; Kraemer et al., 1997; Saitz & O'Malley, 1997). There is evidence to suggest that a history of multiple prior episodes of assisted withdrawal may lead to an increased risk of seizures and withdrawal problems (Booth & Blow, 1993; Brown et al., 1988; Lechtenberg & Worner, 1990), and so a number of previous unsuccessful attempts at outpatient assisted withdrawal may also suggest the need for referral to an inpatient setting. Dependence on drugs can increase the risks associated with withdrawal and also the duration and severity of withdrawal symptoms, therefore patients with comorbid drug misuse disorders may require treatment in an inpatient setting. In addition, Pettinati and colleagues (1993) found that those with high psychiatric comorbidity and/or poor social support benefited more from inpatient than outpatient treatment.

Children and young people

No evidence evaluating the safety and efficacy of different settings for withdrawal management in children and young people was identified. The GDG therefore drew from the adult literature with special consideration to the additional problems often associated with alcohol misuse in children and young people (for example, problems with school, family, crime and mental health). A significant concern of the GDG for children and young people was with the identification of potential dependence and subsequent withdrawal. In formulating recommendations about the appropriate setting for assisted withdrawal, the GDG considered that the safety issues concerning assisted withdrawal might differ for children (aged 10 to 15 years) and young people (aged 16 to 18 years).

Older people

The GDG did not identify any clinical evidence evaluating the efficacy and safety of different settings for assisted withdrawal specifically for older people. However, research suggests that older patients (aged 60 years and above) are more at risk of cognitive and functional impairment during withdrawal and hence should be considered for inpatient care (Kraemer et al., 1997).

Homeless patients

Homeless patients requiring assisted withdrawal may also require inpatient care unless other shelter and accommodation can be arranged. For example, in a large study assessing the effectiveness of an ambulatory assisted withdrawal programme in the VA system in the US (Wiseman et al., 1997), half of the patients were homeless. The study reported that 88% of patients successfully completed assisted withdrawal and 96% of these successful completers were referred for further treatment on either an inpatient or an outpatient basis. However, the programme provided supported housing for the homeless during the period of assisted withdrawal. Although low socioeconomic status and homelessness may make outpatient assisted withdrawal more challenging, they are not necessarily contraindications for treatment failure and hence should be assessed on a more detailed individual basis. O'Connor and colleagues (1991) reported that socially disadvantaged people were not at an increased risk of unsuccessful assisted withdrawal in an outpatient setting.

Systematic literature review

The literature search identified only one economic study that assessed the cost effectiveness and cost utility of different settings for assisted withdrawal (Parrott et al., 2006). The study evaluated two UK-based withdrawal programmes for people dependent on alcohol. The first intervention was a 10-day assisted withdrawal in a 22-bed facility in Manchester staffed by mental health nurses with support from a local GP. The first part of the intervention involved managing withdrawal safely whilst the second part involved social care interventions. The second intervention was a brief hospitalisation programme based at a Newcastle NHS inpatient unit. This involved 3-day inpatient assisted withdrawal, if required, followed by attendance at a day programme. Patients in this programme were also given counselling based on cognitive-behavioural principles, including motivational intervention prior to structured interventions aimed at abstinence or moderate drinking. Both programmes were compared with no intervention rather than with each other because baseline data was compared with clinical and economic outcome data collected at 6 months after implementation. The economic analysis adopted a societal perspective. It included costs to the NHS, other alcohol treatment services, social services and the criminal justice system. The outcome measures used were QALYs for the cost-utility analysis and reduction in units of alcohol per day and reduction in percentage of drinking days in the cost-effectiveness analysis. QALYs were estimated using EQ-5D scores obtained from participants in the study.

In the cost-effectiveness analysis, the cost per unit reduction in alcohol consumption was £1.87 in the Manchester sample and £1.66 in the Newcastle sample. The cost per reduction of one drink per day was £92.75 in the Manchester sample and £22.56 in the Newcastle sample. The cost per percentage point reduction in drinking was £30.71 in the Manchester sample and £45.06 in the Newcastle sample. In the cost-utility analysis, the ICER per QALY gained, compared with no intervention, was £65,454 (£33,727 when considering treatment costs only) in the Manchester sample and £131,750 (£90,375 treatment costs only) in the Newcastle sample. Overall, the authors concluded that both alcohol withdrawal programmes improved clinical outcomes at a reasonable cost to society. However QALY differences were not significant over 6 months, with both ICERs well above current NICE thresholds for cost-effectiveness.

The validity of the study results is limited by the absence of a non-treatment group for both alcohol withdrawal programmes as changes in clinical outcomes may have occurred without the interventions. Also, the within-group before-and-after study design meant that time-dependent confounding variables could not be controlled for. Data for each programme were collected from single centres, which may limit generalisability of the study findings to other UK centres. The small patient sample-size in both centres and substantial loss to follow-up also limits the robustness of the analysis. It should be noted that patients in the two centres were different in terms of severity of dependence, the number and severity of alcohol-related problems, and socioeconomic status, and therefore direct comparison of costs and outcomes associated with each intervention is not appropriate. No sensitivity analyses were performed to test the robustness of the base case results.

Summary of existing economic evidence

The findings of Parrott and colleagues (2006) suggest that both programmes may be cost effective in terms of reduction in alcohol consumption rather than QALYs gained. The settings, costs reported and measure of benefit adopted in the study make this study directly applicable. However, the effectiveness evidence is not without limitations: the comparator of ‘no treatment’ may not be relevant and the robustness of the results was not fully explored in sensitivity analyses.

Cost analysis of assisted withdrawal in different settings

The cost effectiveness of assisted withdrawal across different settings was considered by the GDG as an area with potentially significant resource implications. As previously discussed, clinical evidence was derived from studies with different designs and therefore it was not possible to synthesise the clinical data in order to conduct a formal economic evaluation. Nevertheless, existing clinical evidence does not suggest that the effectiveness of home-based or outpatient assisted withdrawal attempted in outpatient/home settings differs significantly from that of assisted withdrawal provided in inpatient/residential settings. Therefore, a simple cost analysis was undertaken to estimate costs associated with assisted withdrawal that are specific to the setting in which assisted withdrawal is provided.

Three different assisted withdrawal settings were considered in the cost analysis: inpatient/residential, outpatient and home-based. The healthcare resource-use estimates for each setting were based on descriptions of resource use in studies included in the systematic literature review of clinical evidence. Information was mainly sought in studies conducted in the UK, as clinical practice and respective resource use described in these studies is directly relevant to the guideline context. After reviewing the relevant literature it was decided to utilise resource-use estimates reported in Alwyn and colleagues (2004), which were then adapted according to the expert opinion of the GDG to reflect current routine clinical practice within the NHS. The estimated resource use was subsequently combined with national unit costs to provide a total cost associated with provision of assisted withdrawal in the three settings assessed. Unit costs were derived from national sources (Curtis, 2009; Department of Health, 2010) and reflected 2009 prices. It should be noted that the cost estimates reported below do not include the cost of drugs administered to people undergoing assisted withdrawal. However, such a cost is expected to be similar across all assisted withdrawal settings and therefore its omission is unlikely to significantly affect the relative costs between different options assessed.

Inpatient/residential assisted withdrawal

According to Alwyn and colleagues (2004), inpatient/residential assisted withdrawal lasts for 2 weeks and requires an extra outpatient visit. The GDG estimated that inpatient assisted withdrawal may last longer, between 2 and 3 weeks. The unit cost of NHS adult acute mental health inpatient care is £290 per patient day (Department of Health, 2010). The unit cost of hospital outpatient consultant drug and alcohol services is £85 per face-to-face contact for a follow-up visit (Department of Health, 2010). By combining the above resource-use estimates with the respective unit costs, the total cost of inpatient/residential assisted withdrawal is estimated to range between £4,145 and £6,175 per person treated.

Outpatient assisted withdrawal

Outpatient assisted withdrawal is estimated to require six outpatient attendances (Alwyn et al., 2004). The unit cost of a face-to-face contact with hospital outpatient consultant drug and alcohol services is £181 for the first visit and £85 for each follow-up visit (Department of Health, 2010). By combining these data, the total cost of outpatient assisted withdrawal is estimated at £606 per person treated.

Home-based assisted withdrawal

Alwyn and colleagues (2004) estimated that home-based assisted withdrawal requires six CPN home visits lasting 30 minutes each. The GDG were of the opinion that the first of these visits should be replaced by an outpatient visit to alcohol consultant services, so that appropriate assessment is carried out before starting assisted withdrawal. Moreover, the GDG advised that the travel time of the healthcare professional providing home-based assisted withdrawal should be taken into account. Considering that home visits often take place in remote areas, the GDG estimated that the travelling time of the healthcare professional staff was likely to range between 1 and 2 hours per home visit. The unit cost of a face-to-face contact with outpatient consultant drug and alcohol services is £181 for the first visit (Department of Health, 2010). The unit cost of a CPN is not available for 2009. The total cost of home-based assisted withdrawal was therefore based on the unit cost of community nurse specialists (Band 6), as this type of healthcare professional is expected to provide home-based assisted withdrawal. The unit cost for community nurse specialists is £35 per working hour and £88 per hour of patient contact (Curtis, 2009). This unit cost includes salary (based on the median full-time equivalent basic salary for Agenda for Change (AfC) Band 6 of the January to March 2009 NHS Staff Earnings estimates for qualified nurses), salary oncosts, capital and revenue overheads, as well as qualification costs. The unit cost per working hour was combined with the estimated travelling time, while the unit cost per hour of patient contact was combined with the estimated total duration of home visiting. A £4 travel cost was assumed for each visit. By combining all the above data, the total cost of home-based assisted withdrawal was estimated to range between £596 and £771.

Summary

The cost analysis indicates that, provided that the different assisted withdrawal settings have similar effectiveness, then outpatient and home-based assisted withdrawal are probably less costly (and thus potentially more cost effective) than inpatient assisted withdrawal, resulting in an estimated cost saving of approximately £3,400 to £5,600 per person treated.

Fixed-dose regimen

A fixed-dose (FD) regimen involves starting treatment with a standard dose determined by the recent severity of alcohol dependence and/or typical level of daily alcohol consumption, followed by reducing the dose to zero usually over 7 to 10 days according to a standard protocol.

Table 21 gives an example of a titrated FD regimen (Ghodse et al., 1998; South West London and St George's Mental Health NHS Trust, 2010). Note that due to the gradual rate of reduction, with higher starting doses, the duration of treatment is longer than with lower starting doses. A common error in management of alcohol withdrawal is too rapid reduction of chlordiazepoxide, which can result in emergence or re-emergence of severe alcohol withdrawal symptoms. Another common error is too low a starting dose of chlordiazepoxide. This can be avoided by taking account of typical daily alcohol consumption and/or SADQ score in determining the starting dose. In addition, the response to FD withdrawal regimes should be monitored using a validated tool such as the CIWA-Ar (Sullivan et al., 1989) and the dose of medication adjusted upwards or downwards accordingly in the early stages of withdrawal. In severe alcohol dependence the doses of chlordiazepoxide required may exceed the British National Formulary (BNF) prescribing range. It is more clinically effective to increase the dose of chlordiazepoxide to adequately control alcohol withdrawal symptoms than to add another type of medication (for example, haloperidol).

Table 21

Titrated fixed-dose chlordiazepoxide protocol for treatment of alcohol withdrawal (Ghodse et al., 1998; South West London and St George's Mental Health NHS Trust, 2010).

The first dose of medication should be given before withdrawal symptoms begin to emerge. Delay in initiating chlordiazepoxide treatment can result in withdrawal symptoms either becoming difficult to control or the emergence of complications such as DTs or seizures. Therefore, in people with severe alcohol dependence, the first dose should be given before the breath alcohol concentration falls to zero, as withdrawal will emerge during the falling phase of breath alcohol concentration. The more severe the alcohol dependence, the earlier withdrawal symptoms emerge after last alcohol intake. Some people who are severely alcohol dependent can experience withdrawal with a blood alcohol concentration of 100 mg per 100 ml or more.

Symptom-triggered regimen

A symptom-triggered (ST) approach involves tailoring the drug regimen according to the severity of withdrawal and complications the patient is displaying. The patient is monitored on a regular basis and pharmacotherapy is administered according to the patient's level of withdrawal symptoms. Pharmacotherapy only continues as long as the patient is displaying withdrawal symptoms and the administered dose is dependent on the assessed level of alcohol withdrawal. Withdrawal symptoms are usually assessed by clinical assessment including observation and interview, and/or with the use of a validated withdrawal measurement tool such as the CIWA-Ar. See Table 22 for an example of an ST dosing regimen (NICE, 2010b).

Table 22

Examples of symptom-triggered and front-loaded dosing regimens for treating alcohol withdrawal with chlordiazepoxide.

Front-loading regimen

A front-loading regimen involves providing the patient with an initially high dose of medication and then using either an FD or ST dosing regimen for subsequent assisted withdrawal. See Table 22 for an example of a front-loading dosing regimen (NICE, 2010b).

Medication use and duration of treatment

The results of most studies favoured the use of ST over FD regimens for outcomes assessing medication use and duration of treatment. The ST approach resulted in lower medication needed (Daeppen et al., 2002; Day et al., 2004; DeCaroulis et al., 2007; Lange-Asschenfeldt et al., 2003; Reoux & Miller, 2000; Saitz et al., 1994; Sullivan et al., 1991; Wasilewski et al., 1996; Weaver et al., 2006), lower frequency of administration (Daeppen et al., 2002; Reoux & Miller, 2000) and a shorter duration of treatment (Daeppen et al., 2002; Day et al., 2004; Lange-Asschenfeldt et al., 2003; Reoux & Miller, 2000; Saitz et al., 1994). However, not all studies assessing these outcomes reported results favouring an ST approach. Sullivan and colleagues (1991) and Jaeger and colleagues (2001) found no difference between ST front-loading and FD front-loading regimens in terms of length of stay, and Jaeger and colleagues (2001) reported no significant difference between groups in total dose of medication required. Hardern and Page (2005) found no difference in dose administered and length of stay between ST and regular FD regimens.

Severity of withdrawal symptoms

DeCaroulis and colleagues (2007) reported significantly less time to reach symptom control in the ST protocol group when compared with an FD regimen. Saitz and colleagues (1994) found no difference between an ST and FD regimen in time taken from admission to achieving a CIWA-Ar score of less than 8. Manikant and colleagues (1993) and Day and colleagues (2004) also found no significant difference in severity of withdrawal (using the CIWA-Ar) between an ST front loading and an FD regimen.

Rates of complications or adverse effects

Jaeger and colleagues (2001) reported significantly fewer episodes of DTs in the ST regimen group when compared with routine care but found no difference in overall complication rates. Other studies, however, reported no difference between ST and other FD regimens/routine care in rates of complications and adverse effects (for example, incidence of seizures, DTs and hallucinations) (Lange-Asschenfeldt et al., 2003; Reoux & Miller, 2000; Saitz et al., 1994; Sullivan et al., 1991). In Wasilewski and colleagues' (1996) study, although patients in the ST front loading group had a significantly shorter duration of delirium than the FD group, no significant difference was observed in somatic disorders and abnormalities. Additionally, Day and colleagues (2004) did not find a significant difference between ST front loading and FD regimens in self-reported adverse symptoms.

Other outcomes

Other outcomes, including patient satisfaction, discharge against medication advice, use of co-medication and protocol errors, were reported in the reviewed studies.

Daeppen and colleagues (2002)²² and Sullivan and colleagues (1991) reported that there were no significant differences in patient comfort level between groups, and Day and colleagues (2004) reported no significant difference between ST front loading and FD regimens in terms of patient satisfaction. Two studies (Saitz et al., 1994; Sullivan et al., 1991) reported no difference between ST and FD regimens in terms of rates of discharge against medical advice, and Lange-Asschenfeldt and colleagues (2003) found no difference in use of co-medication. Weaver and colleagues (2006) reported significantly more protocol errors in the ST group as opposed to the FD regimen group.

Symptom-triggered assisted withdrawal in a general medical setting

The studies reviewed above are not likely to be reflective of patients with complex problems who typically are admitted to a general hospital ward for medical treatment and present with withdrawal symptoms (that is, they are undergoing unplanned withdrawal) (Hecksel et al., 2008). For example, although the Jaeger and colleagues' (2001) study found fewer episodes of DTs in the ST regimen group, patients were excluded from the study if they presented with medical comorbidities. In a general admissions unit, this in effect would exclude any post-surgical patients (Hecksel et al., 2008). Additionally, Reoux and Miller (2000) excluded any patients with complex medical histories, and Sullivan and colleagues (1991) did not take into account medical comorbidity. Therefore, Hecksel and colleagues (2008) suggested that in these studies, which have assessed an ST approach in a non-specialist general medical setting, patients that are most likely to develop complications such as DTs have not been investigated using the CIWA-Ar tool and therefore some uncertainty about its value with this population remains (Ferguson et al., 1996).

The majority of the ST studies were conducted in addiction specialist inpatient settings or psychiatric hospitals, which have highly trained specialist staff familiar with the ST dosing regimen and methods (Daeppen et al., 2002; Day et al., 2004; Lange-Asschenfeldt et al., 2003; Manikant et al., 1993; Reoux & Miller, 2000; Saitz et al., 1994; Wasilewski et al., 1996). When dosing regimens were compared in non-alcohol specialist settings (that is, in general hospital medical wards), extensive training was delivered to staff (Jaeger et al., 2001; Sullivan et al., 1991; Weaver et al., 2006). For example, in the study by Sullivan and colleagues, training was delivered over a 6-month period with the assistance of clinical nurse specialist in alcohol and substance misuse. In the Hardern and Page (2005) study, a retrospective audit compared the use of an ST regime (which had been introduced in the medical admissions unit) with regular fixed dosing. However, nurses who were trained to use the scoring tool were frequently unavailable when the patient was admitted. This is reflective of the competing demands on staff in a non-addiction treatment setting. This variability can also be observed in different non-specialist departments such as emergency departments (Kahan et al., 2005).

Nurses, whether in a specialist unit, psychiatric ward, general medical ward or in the community, play a vital role in successful assisted withdrawal. Stockwell and colleagues (1990) found both patients and family members rated the support from community nurses as more important than medication for assisted withdrawal. Nursing staff in specialist addiction treatment centres are highly skilled and trained in all aspects of the medical management of alcohol withdrawal (Cooper, 1994), and have a working knowledge of current working practices and liaise with other staff and services (Choudry, 1990). This may well have an impact on the efficacy of the ST programmes in the studies above.

Most physicians and nurses working in general medical wards are not specialists in the management of alcohol dependence. This is a concern because the first point of contact for many alcohol-dependent people is not a specialist addiction unit, but more usually a general physician in a non-specialist treatment setting such as a general medical ward (O'Connor & Schottenfeld, 1998). Nurses in general medical practice may also lack specialised knowledge and education about addiction and assisted withdrawal (Coffey, 1996; Happell & Taylor, 1999; Ryan & Ottlinger, 1999). Even if training were provided, the obstacles to ensuring comprehensive training in a general medical setting also needs consideration (Moriarty et al., 2010; Schumacher et al., 2000).

Bostwick and Lapid (2004) reported on the use of an ST approach by psychiatrists at the Mayo Clinic in Rochester, Minnesota. A CIWA-Ar controlled protocol was not effective in managing alcohol withdrawal and patients deteriorated with use of an ST approach. In these specific cases reported by Bostwick and Lapid (2004), patients were assumed to be presenting with pure alcohol withdrawal syndrome. However, because no thorough clinical interview was utilised and the patients could not communicate effectively, medical staff did not ascertain whether the apparent presenting alcohol withdrawal symptoms were a result of other acute medical conditions such as sepsis, pain and shock. In another study of admissions in Mayo Clinics, Hecksel and colleagues (2008) found that half of patients receiving ST assisted withdrawal did not meet criteria using the CIWA-Ar. The investigators reported that 44% of patients given this protocol had not been drinking and 23% were unable to communicate effectively. Surprisingly, of those who could communicate, 64% were not currently drinking but were still receiving ST medication. Again, and reflective of Bostwick and Lapid's (2004) study, medical histories were overlooked by physicians with a slight hint at alcohol use in the patient's history informing a decision to use this approach. Physicians also regularly assumed that autonomic hyperactivity and psychological distress were a result of alcohol withdrawal and hence a high CIWA-Ar score was attained, resulting in unnecessary benzodiazepine treatment. The investigators concluded that in patients with a history of alcohol dependence who are likely to develop adverse effects (DTs and seizures), a CIWA-Ar-based ST approach is not appropriate and a more patient-centred, personalised approach to medication management that goes beyond the CIWA-Ar is needed. Furthermore, in medical and surgical patients with a history of drinking, the ST approach to medication management has not been proven. Bostwick and Lapid (2004) and Hecksel and colleagues (2008) also conclude that an ST approach is not appropriate for patients with complex medical and surgical comorbidities, and hence may not be suitable for many patients presenting with alcohol withdrawal syndrome in a general medical setting.

Medication not appropriate for use in a setting without 24-hour monitoring

The use of certain medications for assisted withdrawal may not be appropriate in non-residential settings such as an outpatient clinic or the patient's home. Outpatient medication should be administered orally, have low potential for misuse or overdose, and have few side effects (O'Connor et al., 1994).

Contraindications for benzodiazepines and chlormethiazole in non-residential settings identified in the literature are set out below.

Benzodiazepines

Although long-acting benzodiazepines (such as chlordiazepoxide and diazepam) are preferred for patients with alcohol withdrawal syndrome, short-acting benzodiazepines (such as oxazepam) may be preferred in those for whom over-sedation must be avoided, in people with liver disease who may not be able to metabolise long-acting agents efficiently, and in people with chronic obstructive pulmonary disease (COPD) (Blondell, 2005; Mayo-Smith et al., 2004). However, apart from patients with liver failure and those with COPD (who may well be managed as inpatients; see above), short-acting benzodiazepines may not be suitable for outpatient assisted withdrawal due to the risk of breakthrough seizures (Mayo-Smith, 1997). Furthermore, short-acting benzodiazepines (such as oxazepam and halazepam) may have a greater potential for misuse than benzodiazepines (such as diazepam, chlordiazepoxide, alprazolam and lorazepam) (Griffiths & Wolf, 1990; McKinley, 2005; Soyka & Horak, 2004).

Chlormethiazole

Chlormethiazole is used in inpatient care as it has a short half-life (Majumdar, 1990). However, it requires close medical supervision and is therefore not recommended for non-residential settings such as outpatient clinics, patients' homes and prisons. Furthermore, it is addictive (although this is unlikely to develop in the short time period of an assisted withdrawal) and, more importantly, it can have fatal consequences in overdose resulting from coma and respiratory depression, especially when taken with alcohol (Gregg & Akhter, 1979; Horder, 1978; McInnes et al., 1980; McInnes, 1987; Stockwell et al., 1986).

Children and young people

The use of the same drug regimens as for adults should be used, with doses appropriately adjusted for age and alcohol consumption. The evidence for favouring either ST or FD regimens with children and young people remains uncertain as there are no trials which have investigated this issue. Nevertheless whichever regimen is chosen there is a clear requirement for very close monitoring of withdrawal symptoms. Given the uncertainty identified in this guideline about the capacity of staff to manage symptom triggered withdrawal, where symptoms are easily identifiable it was suggested that a cautious approach to the management of symptoms in young people is a fixed dose regimen but with very close symptom monitoring using a validated rating scale such as the CIWA-Ar.

Older people

As noted earlier, older people can have higher levels of physical comorbidity, cognitive impairment, a lower capacity to metabolise alcohol and medications, and be in receipt of a larger number of medications than younger people. In addition, older people can be more frail and prone to accidents and falls. Therefore it is prudent to have a lower threshold for admission for inpatient assisted alcohol withdrawal in older people who misuse alcohol. Further, doses of benzodiazepines may need to be reduced in older people compared with guidelines for younger adults.

Assessment and interventions for assisted alcohol withdrawal

[Refer to 5.31.1.19–5.31.1.22 for assessment for assisted withdrawal in children and young people]

5.31.1.1.

For service users who typically drink over 15 units of alcohol per day and/or who score 20 or more on the AUDIT, consider offering:

• an assessment for and delivery of a community-based assisted withdrawal, or
• assessment and management in specialist alcohol services if there are safety concerns (see 5.31.1.5) about a community-based assisted withdrawal.

5.31.1.2.

Service users who need assisted withdrawal should usually be offered a community-based programme, which should vary in intensity according to the severity of the dependence, available social support and the presence of comorbidities.

• For people with mild to moderate dependence, offer an outpatient-based withdrawal programme in which contact between staff and the service user averages 2–4 meetings per week over the first week.
• For people with mild to moderate dependence and complex needs²³, or severe dependence, offer an intensive community programme following assisted withdrawal in which the service user may attend a day programme lasting between 4 and 7 days per week over a 3-week period.

5.31.1.3.

Outpatient-based community assisted withdrawal programmes should consist of a drug regimen (see 5.31.1.7–5.31.1.18) and psychosocial support including motivational interviewing (see 6.24.1.1).

5.31.1.4.

Intensive community programmes following assisted withdrawal should consist of a drug regimen (see 7.15.1.1–7.15.1.3) supported by psychological interventions including individual treatments (see 7.15.1.1–7.15.1.3), group treatments, psychoeducational interventions, help to attend self-help groups, family and carer support and involvement, and case management (see 5.11.1.2).

5.31.1.5.

Consider inpatient or residential assisted withdrawal if a service user meets one or more of the following criteria. They:

• drink over 30 units of alcohol per day
• have a score of more than 30 on the SADQ
• have a history of epilepsy, or experience of withdrawal-related seizures or delirium tremens during previous assisted withdrawal programmes
• need concurrent withdrawal from alcohol and benzodiazepines
• regularly drink between 15 and 20 units of alcohol per day and have:

  –

  significant psychiatric or physical comorbidities (for example, chronic severe depression, psychosis, malnutrition, congestive cardiac failure, unstable angina, chronic liver disease) or

  –

  a significant learning disability or cognitive impairment.

5.31.1.6.

Consider a lower threshold for inpatient or residential assisted withdrawal in vulnerable groups, for example, homeless and older people.

Drug regimens for assisted withdrawal

5.31.1.7.

When conducting community-based assisted withdrawal programmes, use fixed-dose medication regimens²⁴.

5.31.1.8.

Fixed-dose or symptom-triggered medication regimens²⁵ can be used in assisted withdrawal programmes in inpatient or residential settings. If a symptom-triggered regimen is used, all staff should be competent in monitoring symptoms effectively and the unit should have sufficient resources to allow them to do so frequently and safely.

5.31.1.9.

For service users having assisted withdrawal, particularly those who are more severely alcohol dependent or those undergoing a symptom-triggered regimen, consider using a formal measure of withdrawal symptoms such as the CIWA-Ar.

5.31.1.10.

Prescribe and administer medication for assisted withdrawal within a standard clinical protocol. The preferred medication for assisted withdrawal is a benzodiazepine (chlordiazepoxide or diazepam).

5.31.1.11.

In a fixed-dose regimen, titrate the initial dose of medication to the severity of alcohol dependence and/or regular daily level of alcohol consumption. In severe alcohol dependence higher doses will be required to adequately control withdrawal and should be prescribed according to the Summary of Product Characteristics (SPC). Make sure there is adequate supervision if high doses are administered. Gradually reduce the dose of the benzodiazepine over 7–10 days to avoid alcohol withdrawal recurring²⁶.

5.31.1.12.

Be aware that benzodiazepine doses may need to be reduced for children and young people²⁷, older people, and people with liver impairment (see 5.31.1.13).

5.31.1.13.

If benzodiazepines are used for people with liver impairment, consider one requiring limited liver metabolism (for example, lorazepam); start with a reduced dose and monitor liver function carefully. Avoid using benzodiazepines for people with severe liver impairment.

5.31.1.14.

When managing withdrawal from co-existing benzodiazepine and alcohol dependence increase the dose of benzodiazepine medication used for withdrawal. Calculate the initial daily dose based on the requirements for alcohol withdrawal plus the equivalent regularly used daily dose of benzodiazepine²⁸. This is best managed with one benzodiazepine (chlordiazepoxide or diazepam) rather than multiple benzodiazepines. Inpatient withdrawal regimens should last for 2–3 weeks or longer, depending on the severity of co-existing benzodiazepine dependence. When withdrawal is managed in the community, and/or where there is a high level of benzodiazepine dependence, the regimen should last for longer than 3 weeks, tailored to the service user's symptoms and discomfort.

5.31.1.15.

When managing alcohol withdrawal in the community, avoid giving people who misuse alcohol large quantities of medication to take home to prevent overdose or diversion²⁹. Prescribe for installment dispensing, with no more than 2 days' medication supplied at any time.

5.31.1.16.

In a community-based assisted withdrawal programme, monitor the service user every other day during assisted withdrawal. A family member or carer should preferably oversee the administration of medication. Adjust the dose if severe withdrawal symptoms or over-sedation occur.

5.31.1.17.

Do not offer clomethiazole for community-based assisted withdrawal because of the risk of overdose and misuse.

5.31.1.18.

For managing unplanned acute alcohol withdrawal and complications including delirium tremens and withdrawal-related seizures, refer to NICE clinical guideline 100.

Special considerations for children and young people who misuse alcohol – assessment and referral

5.31.1.19.

If alcohol misuse is identified as a potential problem, with potential physical, psychological, educational or social consequences, in children and young people aged 10–17 years, conduct an initial brief assessment to assess:

• the duration and severity of the alcohol misuse (the standard adult threshold on the AUDIT for referral and intervention should be lowered for young people aged 10–16 years because of the more harmful effects of a given level of alcohol consumption in this population)
• any associated health and social problems
• the potential need for assisted withdrawal.

5.31.1.20.

Refer all children and young people aged 10–15 years to a specialist child and adolescent mental health service (CAMHS) for a comprehensive assessment of their needs, if their alcohol misuse is associated with physical, psychological, educational and social problems and/or comorbid drug misuse.

5.31.1.21.

When considering referral to CAMHS for young people aged 16-17 years who misuse alcohol, use the same referral criteria as for adults (see 5.26.1.8–5.26.1.18).

5.31.1.22.

A comprehensive assessment for children and young people (supported if possible by additional information from a parent or carer) should assess multiple areas of need, be structured around a clinical interview using a validated clinical tool (such as the Adolescent Diagnostic Interview [ADI] or the Teen Addiction Severity Index [T-ASI]), and cover the following areas:

• consumption, dependence features and patterns of drinking
• comorbid substance misuse (consumption and dependence features) and associated problems
• mental and physical health problems
• peer relationships and social and family functioning
• developmental and cognitive needs, and educational attainment and attendance
• history of abuse and trauma
• risk to self and others
• readiness to change and belief in the ability to change
• obtaining consent to treatment
• developing a care plan and risk management plan.

Why this is important

Alcohol misuse is an increasingly common problem in children and young people. However, diagnostic instruments are poorly developed or not available for children and young people. In adults there is a range of diagnostic and assessment tools (with reasonable sensitivity and specificity, and reliability and validity) that are recommended for routine use in the NHS to both assess the severity of the alcohol misuse and to guide treatment decisions. No similar well-developed measures exist for children and young people, with the result that problems are missed and/or inappropriate treatment is offered. The results of this study will have important implications for the identification and the provision of effective treatment in the NHS for children and young people with alcohol-related problems.

Residential units versus outpatient treatment

Of the 14 included trials, three involved a comparison of residential units versus outpatient treatment. RYCHTARIK2000A compared a residential unit versus an outpatient setting; CHAPMAN1988 compared a 6-week inpatient programme with a 6 week outpatient programme. WALSH1991 compared compulsory inpatient treatment versus compulsory attendance at AA; this study was atypical in that the sample consisted of workers at an industrial plant in the US who were part of an employee assistance programme and whose jobs were at risk should they fail to attend treatment. A 3-week period of residential treatment was followed by a year of job probation, during which attendance at AA meetings at least three times per week, sobriety at work and weekly checks with the programme staff were compulsory if the person wanted to keep their job. The outpatient treatment group were referred and offered an escort to a local AA meeting, which they were advised to continue attending at least three times a week for a year. They were treated in the same way as participants in the residential group for the following year.

Residential units versus day hospital

Of the 14 included trials, seven (BELL1994; LONGABAUGH1983; MCKAY1995; MCLACHLAN1982; RYCHTARIK2000A; WEITHMANN2005; WITBRODT2007) involved a comparison of residential rehabilitation units versus day hospital. All seven trials had a 28-day length of stay in treatment. Both MCKAY1995 and WITBRODT2007 looked at day hospital versus residential rehabilitation treatment, with the populations being split into a self-selected arm and a randomised arm.

Day hospital versus outpatient treatment

Two trials out of the 14 involved a comparison of day hospital versus outpatient treatment (MORGENSTERN2003; RYCHTARIK2000A).

Residential unit versus residential unit

Of the 14 included trials, one (KESO1990) involved a comparison of two different types of residential treatment, assessing the efficacy of two different therapeutic approaches. The Kalliola programme was based on the Hazelden or Minnesota model, with a focus on AA 12-step principles with abstinence as the designated treatment goal, whereas the Jarvenpaa programme was a more traditional approach to residential rehabilitation without the focus on AA 12-step principles.

Short versus long duration inpatient treatment

Three of the 14 trials involved a comparison of different lengths of admission to inpatient treatment. MOSHER1975 compared a 9-day versus a 30-day inpatient stay. STEIN1975 compared a 9-day residential inpatient stay with a 9-day stay with an additional 25 days of residential rehabilitative care. PITTMAN1972 compared a group receiving 7 to 10 days of inpatient care only with 3 to 6 weeks of inpatient care with an additional option of further outpatient aftercare.

Residential unit versus outpatient treatment

Residential unit treatment was no more effective than an outpatient treatment in maintaining abstinence or in reducing the number of DDD at 3-month follow-up (RYCHTARIK2000A). Furthermore, there was no significant difference observed between treatment in a residential unit and a day hospital in reducing the number of participants drinking more than 60 g of alcohol per drinking day at 6-month follow-up (CHAPMAN1988).

A residential unit setting was significantly more effective than an outpatient setting in increasing the number of participants abstinent at 2-year follow-up in only one study (WALSH1991). This study population was atypical and is unlikely to be representative of patients attending UK alcohol treatment services, and the study included treatment elements that would be difficult to replicate in the UK.

Based on the GRADE method outlined in Chapter 3, the quality of this evidence is moderate and further research is likely to have an important impact on the confidence in the estimate of the effect and may change the estimate (for further information, see Table 27).

Residential unit versus day hospital

On measures of alcohol consumption, there was no significant difference between a residential unit and a day hospital on DDD day at 3-month follow-up. At 6-month follow-up, there was a significant difference between the two groups favouring day hospital treatment on mean number of drinking days, based on the results of the MCKAY1995 study. This effect did not remain at 12-month follow-up, however there was a trend (p = 0.08) slightly favouring day hospital treatment. It should be noted that this study had both a randomised and self-selected sample, and, because inclusion into this analysis was restricted to RCTs, only the randomised population was used. However, the results from the self-selected sample parallel the results from the randomised arm. The self-selected participants did not do any better on drinking outcomes than those who were randomly assigned at 6- or 12-month follow-up. Any differences that did emerge from the self-selected group, tended to favour the partial hospitalisation group (day hospital), as found in the randomised sample.

On rates of relapse or lapse to alcohol at 6 and 12 months post-treatment, there were no significant differences between residential unit and day hospital treatment. Additionally, there were no significant differences in the number of participants drinking daily at 6-month follow-up (LONGABAUGH1983), or in the PDA at 3-month follow-up (RYCHTARIK2000A).

One study found that more participants were retained in treatment in the residential setting than the day hospital setting (BELL1994). However, this study included a mixture of participants with primary drug and alcohol misuse problems, and so the results may not be representative of individuals presenting to an alcohol treatment service.

Based on the GRADE methodology outlined in Chapter 3, the quality of this evidence is moderate and further research is likely to have an important impact on the confidence in the estimate of the effect and may change the estimate (for further information see Table 28).

Day hospital versus outpatient treatment

A day hospital was not found to be any more effective than a less intensive outpatient setting in terms of PDA or DDD at 3-month follow-up. However, it is important to consider that the MORGENSTERN2003 study contained a mixture of both primary drug and alcohol users, so these results may not be generalisable to the wider population presenting for treatment of alcohol problems.

Based on the GRADE methodology outlined in Chapter 3, the quality of this evidence is moderate to high and further research is likely to have an important impact on the confidence in the estimate of the effect and may change the estimate (for further information, see Table 29).

Residential unit versus residential unit

When analysing two different therapeutic approaches to residential treatment, no difference was found between the two different residential treatment models (Kalliola and Jarvenpaa) on reducing the number of participants who relapsed between 4- and 12-month follow-up.

Based on the GRADE methodology outlined in Chapter 3, the quality of this evidence is moderate and further research is likely to have an important impact on the confidence in the estimate of the effect and may change the estimate (for further information, see Table 30).

Short duration versus longer duration level (inpatient)

There was no significant difference between a 21-day inpatient stay and an extended 9-day inpatient stay at reducing the number of participants consuming alcohol post-treatment, or at 3- or 6-month follow-up (MOSHER1975). A longer duration in an inpatient setting was no more effective in preventing lapse (non-abstinence) than a shorter duration in an inpatient setting. No effect remained at 6-, 7-, 10- and 13-month follow-up.

Based on the GRADE methodology outlined in Chapter 3, the quality of this evidence is moderate and further research is likely to have an important impact on the confidence in the estimate of the effect and may change the estimate (for further information, see Table 31).

Randomised controlled trials

There are several additional studies that were well-conducted trials but could not be included in meta-analyses and did not evaluate the treatment settings as defined above. These studies nevertheless found similar results that support this meta-analysis. Chick (1988) compared simple advice with amplified advice (simple advice plus one session of motivational interviewing) with extended treatment, which included the offer of further outpatient appointments, inpatient, or day treatment. There were no differences between the advice groups or the extended treatment on abstinence outcomes at 2-year follow-up, nor on drinking frequency outcomes. There were no significant differences found on alcohol consumed in 7 days prior to follow-up, frequency of drinking over 200 g per day in the past year, period of abstinence in the past year, or on other measures such as employment or marital situation. Edwards and Guthrie (1966) assigned participants to an average of 9 weeks of inpatient or outpatient treatment, and found no significant differences on measures of drinking at 6- and 12-month follow-up. Lastly, Eriksen (1986a) assigned 17 ‘alcoholics’ who were post-assisted withdrawal to either immediate inpatient treatment or a 4-week waitlist control. Results indicated no significant differences between groups on outcomes of days drinking, or on other outcomes such as sick leave or institutionalisation.

Predictor studies

Even in the absence of overall differences in treatment outcomes between residential and outpatient settings, it is possible that certain types of patients derive differential benefits or harms from being treated in these alternative settings. This is the central issue in matching patients to optimal treatment approaches. Relatively few of the above studies report differential outcome based on patient characteristics, but a reasonably consistent picture does emerge – although it should be pointed out this is often based on post hoc analysis of non-randomised populations and so should be treated with caution. The GDG considered this issue, the main evidence points of which are summarised below; in doing so the GDG drew on the existing systematic review developed by the Specialist Clinical Addiction Network (Specialist Clinical Addiction Network, 2006) for the consensus statement on in-patient treatment.

The most commonly studied predictor variables in the treatment of alcohol dependence have been measures of problem severity and social stability. More severe and less socially stable patients who misuse alcohol seem to fare better in inpatient or more intensive treatment (possibly outpatient based), whereas among married patients with stable accommodation, fewer years of problem drinking and less history of treatment, outpatient (and less intensive) treatment yields more favourable outcomes than inpatient treatment (Kissin et al., 1970; McLellan et al., 1983; Orford et al., 1976; Smart, 1977; Stinson, 1979; Willems, 1973). When heterogeneous populations of people who misuse alcohol are averaged together, the consistent finding is of comparable (or better) outcomes from outpatient as opposed to residential treatment (McLellan et al., 1983). Moos and colleagues (1999) found in an effectiveness trial of inpatient treatment of different theoretical orientations within the VA treatment system that longer lengths of stay were associated with better outcomes. Likewise, in Project MATCH, patients who received inpatient treatment prior to 12 weeks of outpatient care had better drinking outcomes than those who went directly into outpatient care (Project MATCH Research Group, 1997).

Systematic literature review

One study was identified in the systematic search of the economic literature that considered the cost effectiveness of different settings for rehabilitation treatment for people with an alcohol-use disorder (alcohol dependence or harmful alcohol use) (Pettinati et al., 1999). Details of the methods used for the systematic search of the economic literature are described in Chapter 3.

The study by Pettinati and colleagues (1999) assessed the cost-effectiveness of inpatient versus outpatient treatment of people with alcohol dependence. Both inpatient and outpatient treatment programmes followed the same multi-modal clinical approach based on the traditional 12-step programme of AA. This involved individual, marital, family and group counselling provided in the intensive treatment period, including 4 weeks of inpatient and 6 weeks of outpatient treatment. The primary difference between the inpatient and outpatient programmes was the amount of treatment hours and attendance at support groups. Inpatients attended educational and therapy sessions from 9 a.m. to 5 p.m., and attended an AA meeting in the evening, whilst outpatients were expected to attend individual and/or group sessions approximately one to two evenings a week, and AA meetings on the evenings that they did not attend therapy sessions as well as a family educational programme at the weekends. The study population consisted of 173 patients with a formal diagnosis of alcohol dependence but no other substance dependence. The primary outcome measure used in the study was the probability of returning to significant drinking over 12 months. This was defined as three or more alcoholic drinks in one sitting, admission to an inpatient or detoxification centre or incarceration due to alcohol-related disorders. A US healthcare payer perspective was adopted for the analysis. Resource use and cost items included the total number of treatment service hours attended during the intensive treatment programme each week via interviews with the subject. A weighted cost-to-charge ratio was applied to the billing charges for services to adjust for geographic- or institution-specific charges.

Rather than calculate ICERs, the authors presented cost-effectiveness ratios by dividing treatment costs by the probability of returning to significant drinking. For treatment responders, the inpatient:outpatient cost-effectiveness ratio was calculated for the 3-month follow-up as 4.5:1, at the 6-month follow-up as 5.3:1 and at the 12-month follow-up as 5.6:1. For treatment responders, the mean (SD) cost of successfully completing inpatient treatment was $9,014 ($2,986) versus $1,420 ($619), (p <0.01); a ratio of 6.5:1. The validity of the study findings are limited it was based on a non-randomised study design within the US healthcare system which may not be generalisable to the UK setting. Only the costs of treatment were included in the cost analysis, with no consideration of any subsequent healthcare and very little detail was given by the authors on resource use and cost estimation. Finally, health outcomes, which were not formally combined with cost differences to compute ICERs.

Cost analysis of rehabilitation treatment in different settings

The cost effectiveness of rehabilitation treatment for people with an alcohol-use disorder in different settings was considered by the GDG as an area with potentially significant resource implications. A formal economic evaluation comparing different rehabilitation settings was not attempted due to time constraints and problems in synthesising relevant clinical evidence. Nevertheless, a cost analysis was undertaken to estimate costs associated with rehabilitation treatment of people who misuse alcohol in different settings in the UK. The results of this analysis were considered by the GDG alongside the findings of the clinical effectiveness review, to make a judgement regarding the cost effectiveness of different settings for rehabilitation treatment.

Two different settings for rehabilitation treatment were considered in the analysis: residential settings and day hospital (partial hospitalisation) settings. The healthcare resource-use estimates for each setting were based on descriptions of resource use in studies included in the systematic literature review of clinical evidence. Studies conducted in the UK were limited in this review. Therefore, resource-use estimates from studies conducted outside the UK were refined using the expert opinion of the GDG to reflect current routine clinical practice within the NHS. The estimated resource use was subsequently combined with national unit costs to provide a total cost associated with rehabilitation treatment in the three settings assessed. Unit costs were derived from national sources (Curtis, 2009; Department of Health, 2010) and reflected 2009 prices.

Residential treatment unit

The duration of treatment in this setting has been reported to vary from 4 weeks (Sannibale et al., 2003) to 60 days (Zemore & Kaskutas, 2008). Both studies were conducted outside the UK. The GDG estimated that residential treatment typically lasts 12 weeks (3 months) in the UK setting. No unit costs for residential treatment for people with an alcohol-use disorder provided within the NHS are available. Residential units for people who misuse drugs/alcohol provided by the voluntary sector cost £808 per resident week (Curtis, 2009). By combining estimated duration of residential treatment with the respective unit cost, the total cost of residential rehabilitation treatment is estimated at £9,696.

Day hospital treatment

According to Zemore and Kaskutas (2008) and McKay and colleagues (1995), the duration of rehabilitation treatment taking place in day hospitals ranges between 2 and 4 weeks. The GDG considered 4 weeks to be a reasonable duration of day hospital rehabilitation in the UK. McKay and colleagues (1995) reported that participants in their study attended a day hospital 5 days per week. The GDG estimated that frequency of attendance in day hospital rehabilitation should be between 5 and 7 days per week. UK unit costs of such services are not available. The NHS unit cost of mental health day care is £102 per attendance (Department of Health, 2010). However, this facility is likely to provide, on average, non-specialist services and therefore this unit cost is expected to be somewhat lower than the cost of a day hospital rehabilitation service. On the other hand, Parrott and colleagues (2006) reported a local unit cost of a day hospital assisted withdrawal and rehabilitation service for people who are alcohol dependent of £129 per day (uplifted from the originally reported cost of £109 per day in 2004 prices, using the Hospital and Community Health Services pay and prices inflation indices provided in Curtis, 2009). Using the range of these two unit costs and combining them with the estimated resource use, the total cost of a day hospital rehabilitation treatment for people who misuse alcohol is estimated to range from £2,040 (for a 5-day per week programme, using the lower unit cost) to £3,612 (for a 7-day per week programme, using the higher unit cost).

Summary

The cost analysis indicates that, as expected, day hospital treatment is less costly than residential rehabilitation.

Interventions to promote abstinence and relapse prevention

5.32.11.1.

For people with alcohol dependence who are homeless, consider offering residential rehabilitation for a maximum of 3 months. Help the service user find stable accommodation before discharge.

5.32.11.2.

For all children and young people aged 10–17 years who misuse alcohol, the goal of treatment should usually be abstinence in the first instance.

Why this is important

Many people, in particular those with severe problems and complex comorbidities, do not benefit from treatment and/or lose contact with services. One common approach is to offer intensive residential rehabilitation and current policy favours this. However, the research on the effectiveness of residential rehabilitation is uncertain with a suggestion that intensive community services may be as effective. The interpretation of this research is limited by the fact that many of the more severely ill people are not entered into the clinical trials because some clinicians are unsure of the safety of the community setting. However, clinical opinion is divided on the benefits of residential rehabilitation, with some suggesting that those who benefit are a motivated and selfselected group who may do just as well with intensive community treatment, which is currently limited in availability. Given the costs associated with residential treatment and the uncertainty about outcomes, the results of this study will have important implications for the cost effectiveness and provision of alcohol services in the NHS.