Research aims

The aim of this stage was to describe the development and psychometric evaluation of a questionnaire designed to identify and measure factors that influence the identification of individuals at HR for psychosis in primary care. This informed the design of the LEGS educational intervention to help GPs and primary care physicians detect these individuals.

Methods for data collection

Following standard TPB guidelines²⁸ a 106-item preliminary questionnaire was constructed using a semistructured discussion group with eight GPs to elicit commonly held beliefs about identifying HR individuals. The questionnaire was distributed to 400 GPs in 38 practices across 12 counties in England, not including Cambridgeshire and Peterborough where we intended to run the trial.

Analysis

A polytomous graded response model²⁹ was used to identify redundant items and assess the validity of the questionnaire. Factor analysis was used to assess the structural conformity of the final questionnaire with the TPB. Cronbach’s alphas were calculated to determine the reliability of the final questionnaire. Path analysis was conducted to assess the ability of the TPB’s constructs to predict intention and reveal the percentage of variance explained by intention.

Key findings

Indirect measures were well constructed and adequately covered the breadth of the measured construct. Items within all direct measures measured the corresponding construct satisfactorily. The alpha values confirmed improvement for each of the constructs in the reduced version of the questionnaire with the exception of intention, which remained the same. The final instrument consisted of 73 items and showed acceptable reliability (α = 0.77–0.87) for all direct measures. All of the direct measures of the TPB significantly predicted intention, accounting for 35% of the variance. Subjective norm (perceived professional influences) was the strongest predictor of intention. GPs had positive intentions and attitudes towards identifying individuals at HR for psychosis. The foremost motivational factors for GPs were their perceptions of whether or not other GPs identify HR individuals and whether significant others (e.g. patients, colleagues, health-care system) approved or disapproved of identification.

Successes

Theory underpinned the design of all components of the educational intervention: the understanding of the GPs’ behaviour, the development of the measures and the attempt to change behaviour. We confirmed the feasibility, reliability and acceptability of a TPB-based questionnaire to identify GPs’ beliefs and intentions concerning the identification of individuals at HR for psychosis.

Theory-based interventions provide an understanding of what works and thus are a basis for developing better theory across different contexts, populations and behaviours. They could, and should, be used more in the NHS where innovation often requires education and behaviour change by staff.

Challenges

The average time taken to complete the questionnaire was 16 minutes. Most of the declining GPs and some of the participating GPs mentioned that the length of the questionnaire was off-putting.

Limitations

The response to the pilot was very low: only 82 (20.5%) GPs returned questionnaires. The cost and time investments were high for such a low return.

Recommendations for future research

The recommendation that an original TPB questionnaire is developed every time a new behaviour is studied,²⁸ or the same behaviour is studied in a new population, suggests that similar methodology can be used to help GPs in the identification of other pathologies and in a variety of mental health settings. Further application of the TPB in the NHS has already been mentioned.

Methods for data collection

The information collated from the pilot questionnaire identified specific barriers that we targeted with strategies designed to change clinical practice with respect to identifying HR individuals in primary care.

Key findings

This theory-based information could be important for improving the efficiency of referral pathways and contributing to a reduction in the DUP. We had a clear structure on which to base our low-intensity leaflet intervention and our high-intensity face-to-face and video approaches.

Successes

Mapping theoretical constructs to behaviour change techniques provided a clear framework for process analysis and increased the ability of the intervention to accomplish the desired outcome of motivating GPs to identify individuals at HR.

It was invaluable to have a guide to the different behaviour change techniques, with definitions that addressed different behavioural determinants linked to theoretical constructs. This allowed us to select the most appropriate intervention strategies for GPs.

Challenges

The mapping of theoretical constructs to behaviour change techniques was complex and time-consuming. However, this systematic approach ensured that the behaviour change techniques and delivery methods targeted the theoretical determinants of GPs’ behaviour directly.

We learnt from the pilot that individual clinicians have very different levels of knowledge about psychosis and mental health in general. Therefore, it was important to ‘pitch’ the presentation at the right level so as not to be condescending but at the same time ensure a basic level of understanding. Achieving this balance proved extremely difficult, especially given the time restraints of the 60-minute sessions. Trying to explain the complexities of the at-risk concept in a concise way but also making it educationally appropriate for both GPs and practice nurses took many drafts. Again, it was helpful to have comments from GP colleagues regarding the presentation to guide us before we approved the final version.

Limitations

Ensuring that the leaflet was specific enough to capture all possible at-risk symptoms [attenuated symptoms, family vulnerability and BLIPS (Brief Limited Intermittent Psychotic Symptoms)] without being too sensitive and producing numerous ‘false-positive’ referrals was a dilemma. Despite utilising many sources of information, including GPs practising outside the trial area, the resulting leaflet proved a little too sensitive.

Recommendations for future research

Few studies have used the TPB to predict intention to take part in an intervention. Such an application could provide valuable information about how best to recruit GPs into future studies.

Successes

The PCRN was useful in terms of accessing surgeries for participation history and informing them about the study. It provided us with the Research Information Sheet for Practices (RISP). The focus of this was provision of information about the practical implications of the research for the participating GPs (What, Where, How often, etc.). This proved more useful than the trial information sheet, which, although detailed and well structured, did not cover enough of the practical details that were required to help practices reach a decision about whether or not to participate.

Having an out-of-area GP review and critique the educational materials proved invaluable for establishing the appropriate pitch and tone and reviewing the content of the intervention.

Three dedicated research and liaison practitioners (RLPs) were specifically recruited for the trial [one man, two women; mean age 45.5 years, standard deviation (SD) 4.7 years]. All were experienced mental health professionals (one psychologist, one nurse, one social worker). They acted as facilitators between secondary and primary mental health services as it is proposed that this is a fundamental role in helping individuals and teams to understand what they need to change and how they need to change it, to translate evidence into practice. Each RLP was responsible for delivering the high-intensity intervention to the surgeries within one of the three geographical areas in Cambridgeshire covered by the trial. The RLPs found that showing empathy (understanding the nature of school/practice life) was central in building a relationship with the teachers/GPs. Face-to-face meetings at the point of consent facilitated this.

A beneficial approach was conveying the research as an important medium through which problems that were relevant for a GP’s daily practice could be understood and solutions to the problems could be generated. Flexibility when arranging presentations (i.e. offering more than one session to accommodate all staff) was important for optimising participation.

Convincing the practice leads that participation in the trial would benefit individual GPs, the practice as a whole and most importantly the patients was a key factor in gaining consent to participate in the research. Another successful strategy included emphasising that, by taking part in the trial, GPs could potentially be saving themselves and the practice time because the intervention would allow them to quickly and accurately judge whether or not a young person required a specialist assessment of symptoms. The GPs could see how this would benefit everyone involved.

The GPs were also concerned about what would happen to the individuals who they referred to CAMEO. Assurance that all those identified as being at risk would be invited into the PAATH study for 2 years of mental state monitoring and easy access to a CAMEO psychiatrist if there was any concern about symptoms deteriorating also helped them see the benefits of participating in the trial. We were also able to emphasise that patients without a diagnosis of HR would have a thorough mental health assessment and would then receive appropriate referral on to other services.

Challenges

The trial did not directly involve patients; therefore, it was assumed that only the agreement of practices in the high-intensity arm would be needed for the distribution of leaflets and for their participation in the educational sessions. However, despite discussions with previous members of the Cambridgeshire 1 Research Ethics Committee (REC) about this matter, the committee stipulated that formal consent was required from all invited surgeries, regardless of which arm of the trial they were assigned to. This led to an unexpected long delay in the roll-out of the trial, with contacts, sometimes visits, needing to be arranged with > 100 practices, ultimately resulting in our being granted an 18-month no-cost extension to our programme by the NIHR. The upside was more time to develop the theory-based interventions prior to the trial beginning.

This is a relevant point to place a general comment about our programme. We had to manage challenging situations, a protracted ethical review and subsequent adjustments to our protocol. All of these were ultimately beneficial (see in particular the PAATH study in Work package 5), other than the requirement to gain consent from practices to take part in the cRCT. That reduced our sample size but allowed a PAU comparator in those practices that did not consent, and we retained sufficient power to reject our null hypothesis and confirm the hypothesis of doubling referrals with uncanny accuracy. In retrospect, we would have benefited from a Programme Steering Group as is now required by the NIHR, but we drew on valuable advice from the Central Commissioning Facility and from the late Professor Helen Lester, who inspired some of our programme and was generous with her advice.

The recruitment process was lengthy and at times extremely frustrating. Busy GPs were difficult to contact directly but practice managers were good liaison intermediaries. However, many of the practice managers were very protective of the GPs’ time and were occasionally more negative about the likelihood of the practice participating than were the GPs when we eventually spoke to them. It took many attempts to persuade some practice managers to facilitate discussion of the research trial at team meetings, despite the fact that information had already been sent to the lead GPs within the practices.

The time commitment required to participate in the trial was a key issue for GPs. Reassurance that participation involved essentially just the 2 hours of educational sessions over the whole 2 years of the trial, for which the practice would receive income, was helpful in motivating them to participate.

Recommendations for future research

As has become evident in much health research, the process of ethical review can be protracted and frustrating for researchers. GPs’ negative attitudes, concerns and ambivalent feelings should be elicited and addressed with recruitment strategies.

We relied on the assumption that lead GPs would read the trial information sheet, discuss it with their colleagues and decide whether or not they wanted to participate. It became apparent during the trial that this was not always the case, as many GPs with other demands on their time did not know about our trial. A more advantageous approach could be to advertise the trial with individual GPs prior to gaining consent.

Research aims

Our main aim was to test the null hypothesis that, in terms of the effectiveness and cost-effectiveness of detecting individuals at HR aged 16–35 years, a theory-based educational intervention for primary care was not different from a postal information campaign co-ordinated from an office in a secondary care-based EIS (CAMEO). The journal, Trials, where we published the protocol, insisted on this null construction, which is not easy to follow.

Formulated in a positive manner, this cRCT compared two different approaches to liaising with primary care to increase detection and early referral of people at HR to a specialist early-intervention team for young people with psychosis. We predicated the sample size and power on a doubling of HR and FEP referrals by the high-intensity intervention.

Methods for data collection

General practices were randomly allocated into two groups to establish which is the most effective and cost-effective way to identify people at HR for psychosis. One group received postal information about the local EIS, including how to identify young people who may be in the early stages of a psychotic illness. The second group received the same information plus an additional ongoing theory-based educational intervention with dedicated liaison practitioners to train clinical staff at each site.

The primary outcome was count data per practice site on the number of HR referrals to a county-wide specialist EIS (CAMEO). This was conducted over a 2-year period. All referrals during the duration of the trial were assessed clinically by the study team and stratified into those who met criteria for HR or FEP according to the CAARMS¹¹ (true positives) and those who did not fulfil such criteria (false positives).

Analysis of the effectiveness of the intervention

Given that the main outcome (referrals per practice) was count data, the yield, our primary statistical approach was Poisson regression. Results were adjusted for surgery size and the number of GPs working in each site was considered as a covariate in the model. We also employed Pearson’s chi-squared test and Fisher’s exact test to compare demographic characteristics of the general practices. All of the analyses were performed using the statistical package R (version 3.0.0; the R Foundation for Statistical Computing, Vienna, Austria).

Analysis of the cost-effectiveness of the intervention

Decision-analytic modelling was used to investigate the cost-effectiveness of the high- and low-intensity interventions compared with PAU. A decision tree was constructed in Microsoft Excel 2013 to model the care pathways of the young people in the trial and assess the costs and effects over 2 years associated with the two active interventions and PAU. The costs of (a) the high- and low-intensity interventions, (b) diagnosing referrals who did not meet criteria for HR or FEP (false positives), (c) diagnosing and treating identified HR and FEP cases (true positives) and (d) the subsequent treatment of HR and FEP cases who were not identified (false negatives) were included.

Key findings on the effectiveness of the intervention

The intervention succeeded in raising awareness of potential psychotic symptoms. Between 22 December 2009 and 7 September 2010, 54 of 104 eligible practices provided consent and between 16 February 2010 and 11 February 2011 these practices were randomly allocated to the interventions (28 to the low-intensity intervention and 26 to the high-intensity intervention); the remaining 50 practices constituted the PAU group. Two high-intensity practices were excluded from the analysis. In the 2-year intervention period, high-intensity practices referred more FEP cases than low-intensity practices [mean (SD) 1.25 (1.2) for high intensity vs. 0.7 (0.9]) for low intensity; IRR 1.9, 95% CI 1.05 to 3.4; p = 0.04], although the difference was not statistically significant for individuals at HR of psychosis [mean (SD) 0.9 (1.0) for high intensity vs. 0.5 (1.0) for low intensity; IRR 2.2, 95% CI 0.9 to 5.1; p = 0.08]. For HR and FEP cases combined, high-intensity practices referred both more true-positive [mean (SD) 2.2 (1.7) for high intensity vs. 1.1 (1.7) for low intensity; IRR 2.0, 95% CI 1.1 to 3.6; p = 0.02] and more false-positive [mean (SD) 2.3 (2.4) for high intensity vs. 0.9 (1.2) for low intensity; IRR 2.6, 95% CI 1.3 to 5.0; p = 0.005] cases. Most of these (68%) were referred on to appropriate services. Referral patterns did not differ between low-intensity and PAU practices (Figure 6).

FIGURE 6

Comparison of the distribution of referrals by intervention group. (a) HR; (b) FEP; (c) true positives; and (d) false positives. Lower hinge = 25th centile; upper hinge = 75th centile; line = median; filled (more...)

Key findings on the cost-effectiveness of the intervention

Details of the quantitative economic results and how this part of the trial was conducted can be found at www.thelancet.com/cms/attachment/2035390784/2050868157/mmc1.pdf (accessed 19 January 2016).

Total cost per true-positive referral in the 2-year follow-up was £26,785 in high-intensity practices, £27,840 in low-intensity practices and £30,007 in PAU practices. The lower cost was attributable to fewer false negatives (HR and FEP cases that are not identified), which are assumed to be associated with treatment costs at a later point. The high-intensity intervention was the most cost-effective strategy.

Interpretation

This intensive intervention to improve liaison between primary and secondary care for people with early signs of psychosis was clinically effective and cost-effective. Increasing the resources aimed at managing the primary–secondary care interface provides clinical and economic value in this setting.

Successes

In the south of the county, mental health services were cohesive and the early-intervention team was well established. Generally speaking, the GPs in the south of the county were much better informed about the CAMEO team and what it could offer patients. As a result, they were more open to participating in research connected to psychosis.

The GPs were willing to discuss their patients in detail with each other at the two educational sessions. As a general rule, the GPs responded well to the facts and figures used to illustrate the main points within both the presentations and the DVD. Comments were made about how useful the DVD would be for training GPs and disseminating the information to colleagues who were unable to attend the presentations. As a ‘continuing professional development’ session, many of the GPs and practice nurses were positive about the educational intervention that they attended.

Notably, as with PsyMaptic (see Work package 3), NHS England is currently using these results in its strategy to implement waiting time standards and targets for EISs nationally.

Challenges

General practitioners’ perceptions of the poor relationships between primary care and secondary mental health services was a barrier to participation and the building of relationships with the RLPs. In the north of the county, EISs were relatively new and historically there had been difficult relationships between primary care and local mental health services. Attitudes in the north of the county towards the liaison aspect of the trial and the potential outcome of referrals to CAMEO were rather negative in comparison with attitudes in the south. During some of the educational interventions, several GPs were quite adversarial and extremely critical of mental health services in general. However, the process of the study appeared to help this and the GPs had, after all, agreed to take part.

It took GPs between 15 and 30 minutes to complete the TPB questionnaire. Many of them found it arduous and complained that it was too lengthy. The RLPs had to work hard to justify the displeasure that some of the GPs felt at having to complete such a long questionnaire. Much effort was made to explain why such a questionnaire was being used and how much of an essential part of the research it was. However, this remained the most consistent criticism of the interventions delivered.

During the 60-minute sessions it was a challenge to deliver the presentation, complete questionnaires and leave some time for questions at the end. Some GPs were very keen to discuss particular patients, whether or not they had HR symptoms. The numbers of clinicians who did attend the sessions was always fewer than the numbers practising at the individual surgeries.

During the first educational intervention, some GPs commented that they had not seen any patients with suspected HR symptoms since receiving the leaflet and were rather dismissive of how relevant this all was to their everyday experiences of young people and their problems. To address this, the importance of identifying the early signs of psychosis was emphasised in the DVD by the chief investigator of the trial, Professor P Jones, Head of Psychiatry at Cambridge University. He used the example of ‘crushing chest pain’ as an analogy for the importance and urgency of identifying and treating psychotic symptoms because of the devastating, long-term effects of untreated psychosis on the individual. This aspect of the DVD did generate discussion with some groups of GPs and provided an opportunity to emphasise again how the long-term effects of a psychotic illness should be avoided at all costs.

To facilitate further reinforcement of the symptoms that GPs should be looking for in young people, practices in the high-intensity group were given specific details about the symptoms of the HR patients who they referred. These were provided in the initial assessment feedback document and in the updates for each of the follow-up assessments. GPs rarely commented on this more detailed feedback during the second educational intervention.

There were mixed responses to the DVD. Some GPs found it very helpful; however, some found it to be repetitive because of similar material in the ‘revision’ presentation.

Limitations

A variety of reasons were given for non-attendance at the sessions. These ranged from being on duty to sickness or being on annual leave. Our strategy to address this was to offer practices multiple visits. This facilitated the maximum number of clinicians attending each session. At some surgeries GPs could not stay for the whole session because of clinical commitments; therefore, it is possible that some GPs missed vital information. Another consequence was that not all of the surgeries had a comparable educational experience and this could have had some effect on the results of the trial. We used an intention-to-treat approach for the analysis.

General practitioners rarely telephoned RLPs to discuss particular patients or symptoms, despite being encouraged to do so at every opportunity during the course of the trial. We suspect that their decisions to refer did not depend on the minutiae and finer points of psychopathology that interest secondary mental health care. If the number of RLPs was scaled down or RLPs were removed from the high-intensity intervention it could have been more cost-effective; however, this is conjecture.

Sufficient copies of the laminated version of the leaflet were sent to each practice that each GP could receive his or her own copy before the date of the intervention. When we asked the GPs about their experiences of using the leaflet it was rather disappointing to discover that many of them had not used it and in some practices had not seen it at all. In some cases this seemed to be an administrative error; the leaflets may have arrived at the practice but had not been distributed to the individual GPs. Some of the lead GPs took responsibility for this; in other practices it was the responsibility of the practice manager. Clearly, in some cases our leaflets remained in the postal noise, not being recognised as signal.

Recommendations for future research

Many of the GPs were doubtful about being able to refer accurately. They considered that many of the signs and symptoms that could indicate risk are also present in other mental health illnesses such as anxiety, depression or obsessive–compulsive disorder. This issue was revealed in the pilot study; therefore, we had focused on encouraging the GPs to keep their ‘HR radar’ on when interviewing a young person. We specified that they should use the leaflet to help them ask the relevant ‘probing’ questions, which would guide their referral practice. Their lack of confidence, despite these strategies, is an issue that needs to be addressed in future research.

During the course of the trial many GPs expressed that (1) they would rather have had an electronic version of the leaflet or (2) specific at-risk symptoms should have been included in their nationwide, web-based illness identification tool, which is available to all GPs. This would have been a good idea.

The trial did not measure how long the intervention effect endures or calculate the optimal number of sessions required. This information is important to enable a balance between intervention effectiveness and cost-effectiveness while maintaining identification of HR and FEP individuals.

Methods for data collection

An elicitation phase revealed beliefs underlying teachers’ motivations to detect HR students and informed the construction of a preliminary 114-item questionnaire incorporating all constructs outlined in the TPB. To define the determinants of teachers’ intention to identify HR students, 75 teachers from secondary and further education institutions in 12 counties surrounding Cambridgeshire completed the questionnaire. A psychometric model of item response theory was used to identify redundant items and produce a reduced questionnaire of 44 items that would be acceptable to teachers.

Analysis

A psychometric evaluation of the questionnaire was conducted. The polytomous graded response model was used to examine the validity items within direct and indirect measures and to inform decisions regarding the removal of items. Cronbach’s alpha coefficient was used to assess the internal consistency of the direct measures of attitude, subjective norm and PBC. The relationship between intention and the indirect and direct measures was investigated using path analysis.

Key findings

The average time taken to complete the TPB questionnaire was 20 minutes. The majority of the teachers (63%) reported never having attended any kind of mental health training during their careers.

Indirect measures were well constructed and adequately covered the breadth of the measured construct with the exception of PBC. Items within all direct measures measured the corresponding construct satisfactorily. Only one item within direct subjective norms, direct PBC, showed a factor validity of < 0.5. Cronbach’s alphas for the reduced questionnaire showed acceptable internal consistency.

Perceived behavioural control was the strongest predictor of intention, followed by attitude. Subjective norm did not predict intention. Collectively, the direct measures explained 37% of the variance of intention to identify HR for psychosis. Mean scores for direct measures were just above the mid-scale score for intention and attitude and just below the mid-scale score for subjective norm and PBC.

Teachers considered identifying students at HR for psychosis a worthwhile behaviour and would attempt identification during the school day and believed that their peers or superiors might not approve of them identifying at-risk students. The greatest source of social pressure came from the senior management team within 16+ educational institutions. Teachers’ confidence and control over identification were low.

Increasing awareness and understanding of mental health issues emerged as the most important source of personal positive beliefs. The lack of access to information, knowlege and resources could hinder teachers‘ identification behaviour. Teachers’ perceptions of how confident they are that they are capable of identification and how much control they have over identification were prominent motivational factors.

Our questionnaire proved to be reliable, with the analysis supporting the predictive power of intention within the TPB model.

Successes

We have confirmed the feasibility, reliability and acceptability of a TPB-based questionnaire to identify teachers’ beliefs and intentions concerning the identification of individuals at HR of psychosis.

Limitations

Despite strenuous efforts, the response rate to our questionnaire was very low: only 75 teachers (9.5%) returned the questionnaire of the 793 invited. External validity could have been undermined if respondents differed systematically from non-respondents, which they probably did. The potential limitations of using self-report measures must be considered when interpreting the results as acquiescence and social desirability may have been influencing factors.

Recommendations for future research

Objective measures of behaviour should be incorporated in future research to avoid reliance on self-report.

The questionnaire length would be a limiting factor for TPB studies with teachers; therefore, options for reducing the number of items should be explored in future research.

Successes

Work with ‘consultant’ teachers from outside the trial area was very important so that the language used about symptoms could be assessed and deemed to be at an appropriate level for non-medical staff.

Challenges

Teaching teachers was quite a daunting prospect for the RLPs. Achieving the balance between an appropriately academic language for a group of professional teachers and recognising that many teachers would have very little or no knowledge at all about psychosis required many revisions of the presentation material. Describing psychosis and HR symptoms had to be approached differently for the teaching staff as they had no clinical knowledge. Success in presenting this kind of intervention required the open acknowledgement of differing levels of knowledge and understanding about mental health problems. As both nurses and teaching staff would be present at the educational sessions, the needs of these two different professional roles needed to be addressed within the presentations.

It was important not to present the material as all encompassing but rather as an introduction to the subject, covering all of the essential information. Therefore, further information was made available through a recommended reading list, references and a paper copy of the presentation. This material was included in packs that were distributed at the end of the presentations.

Teachers were not able to refer directly to a secondary mental health team but had to refer through school and college nursing staff. It was necessary to contact the nurses’ local area team prior to the roll-out of the intervention to ensure that the referral procedure was in accordance with existing referral practices. This also prepared the nursing staff for the possible increase in number of referrals. Each individual institution’s referral procedures were respected, rather than dictating a county-wide protocol for the study.

It was also necessary to stress the importance of discussing any referral with the student, to establish that consent had been given for information sharing.

Successes

The consent visit to 16+ educational institutions to discuss the trial seemed to considerably improve awareness of the EIS and the importance of not missing what could be the signs and symptoms of a developing mental illness in students who were not performing to expected levels. At the initial contact with 16+ educational institutions the response from most of the head teachers was extremely positive. The majority could see the importance of identifying students early to prevent serious illness from developing.

Many of the teachers were interested in the content of the presentations, asking pertinent questions. Positive recurring themes were:

• the opportunity to discuss students who they recognised could be HR cases
• a genuine enlightenment by the discussion of symptoms
• awareness, for the first time, that there is a service that specialises in the assessment and treatment of young people with psychosis.

The nurses and pastoral support staff felt more confident when talking to students because of the leaflet and what they had learnt from the presentations.

Challenges

As stated in the previous section, many of the teachers were interested in the content of the presentations, asking pertinent questions. Negative recurring themes, that are themselves findings, included:

• the teachers held strong views that identifying HR symptoms in students was not part of a teacher’s role
• a reluctance to take on this extra ‘pastoral’ responsibility
• a feeling that the content of the presentations was too medical and health related
• a level of genuine anxiety about whether or not they should even attempt to identify HR symptoms.

Attempting to speak to the head teachers of some 16+ educational institutions proved difficult, with one school telling us that ‘we don’t have that kind of problem in this school!’ Conversely, other head teachers expressed an eagerness to be in the high-intensity group so that their staff could receive the high-intensity intervention. Many of the head teachers had not heard of the CAMEO service although others had through previous experience of having one or two students with psychotic disorders in their college.

Head teachers expressed a desire to have a ‘youth-focused’ service to which to refer students when necessary. Therefore, it was necessary to stress that referrals would be appropriate only for those who may be at risk of developing psychosis, not for those with any mental health problems generally. Having said that, our primary care trial suggests that this course of action may have benefits.

It became apparent that there were going to be few opportunities within the timetables of 16+ educational institutions to allow staff to be present all together for the presentations. The RLPs were able to reassure the head teachers that they could be as flexible as required to deliver the intervention. Solutions included delivering the intervention out of teaching hours or during lunch breaks or having the option of multiple sessions (one of the interventions was delivered at 0830 in the morning to accommodate as many teachers as possible).

Distributing and collecting the questionnaires at the educational sessions was challenging as the groups consisted of up to 50 teachers. At some of the larger groups extra research staff helped to facilitate this aspect of the presentations.

These larger numbers also meant that waiting for all of the teachers to complete the questionnaire became a problem, with small groups of teachers chatting among themselves before the rest had finished. This became quite disruptive and it took some time to retrieve the attention of those attending in order to continue with the second part of the presentation. Completing the questionnaire elicited some protesting, but it was stressed that this was very much a part of the research and essential to measuring the change in referral behaviour before and after the interventions.

Many teachers raised the subject of substance misuse-related psychosis. They were interested to know whether or not the use of substances would increase the likelihood of psychotic symptoms and whether or not this would be an exclusion criterion for referral to CAMEO. We reassured them that students who may have symptoms following alcohol or substance use should not be excluded from referral for assessment by CAMEO.

Limitations

There were several hundred teachers at some of the colleges and this made it almost impossible to know whether or not the leaflets had been distributed effectively to all staff. It is possible that a proportion of the teaching staff did not see, or have regular access to, the leaflet.

Recommendations for future research

Leaflet distribution should be monitored and verified in future interventions with 16+ educational institutions. Identifying and fitting in research with the rhythm of teachers’ professional development (e.g. ‘Baker days’) would be useful. That said, as with the GPs, teachers’ time is increasingly pressured and so finding innovative ways to generate research-based knowledge is a challenge.

We have completed the collection of data from the teachers involved in the LEGS trial. We plan to analyse these data over the next 6 months.