Study Design

Four non-randomized studies^1,3,5,11 were included in this report. Two of the non-randomized studies were published in 2014,^1,11 while one each was published in 2011,³ and 2008.⁵

Country of Origin

Of the four non-randomized studies, one was a retrospective study from the USA,¹¹ and another was a pre- and post-intervention cohort study from Israel.¹ There were two before-and-after intervention prospective studies from the USA.^3,5

Patient Population

One non-randomized study¹¹ included in this report had both adult and pediatric patients, including those receiving specialized services for trauma, burn, neurosurgery, cardiothoracic surgery, transplant, and oncology. Details of patients’ characteristics were not provided.

In one study,¹ the similarity in some characteristics and treatments among patient groups hospitalized during and before the period of the intervention were reported. For example, the age of patients admitted before the intervention ranged from 18 to 90 years (mean ± standard deviation [SD]: 57 ± 19), while those hospitalized during the intervention were between 18 to 83 years old (49.7 ± 22). The patients were mostly immobile and majority received tube feeding during the periods being compared. In the period before the intervention 25.9% of the patients had pressure sores and 30.4% received steroid treatments, compared with 16.7% of patients with pressure sores and 19.4% who received steroid treatment during the intervention period.

One of the prospective studies³ was conducted among patients in a neonatal intensive care unit (NICU) while the other prospective study⁵ reported infection rates in patients admitted to five wards “with the highest incidence of CDAD” (Clostridium difficile-associated disease) as well as among the hospital-wide patients. No other details of patients characteristics were provided in this study.⁵

Interventions and Comparators

One prospective study⁵ included in this report used hydrogen peroxide vapor (HPV) as the decontamination agent and compared infection rates during a 9 months period of intervention (June 2005 through March 2006) to a parallel pre-intervention period of 9 months (June 2004 through March 2005).

One study¹¹ examined infection rates during a period of 22 months when a pulsed xenon UV disinfection (UVD) technique was used in a hospital compared to the preceding 30 month period before the UVD technology was in use. Another study³ used enhanced UV germicidal irradiation (eUVGI) in a NICU and reported before-and-after intervention results for four consecutive 6-month periods (6 months before and 18 months after the introduction). No reason was given for the uneven period lengths. A newsletter concerning the study stated that the eUVGI (also called Pathogen Control System)³ “integrates standard UVGI emitters with MERV15 air filters in such a manner as to predictably destroy harmful viruses, bacteria and fungi at a pre-determined efficiency within a given HVAC’s airstream.”¹⁵ One study¹ compared the infections rates in a ward during a 6-month hospitalization period when biocidal copper impregnated linens were being used to a parallel 6-month hospitalization period when ordinary linens were used.

In all the included studies, non-manual disinfection interventions were used as adjunct to traditional cleaning and disinfection protocols that already existed in the hospital facilities. Cleaning prior to application removes organic matter that reduces the effectiveness of the non-manual room disinfections intervention systems.

It is noteworthy that within the HP- and UV-based technologies, there are differences in available systems that impact efficacy and suitability, and necessitate trade-off between time and effectiveness. For example, VHP systems (also known as aerosolized hydrogen peroxide systems) usually deliver pressure-generated aerosol containing 5 to 6% HP and <50 part per million (ppm) silver via a unidirectional nozzle, and they have a typical recommended dose of 6 ml/M³ for hospital rooms.² Following exposure, the aerosol is left to decompose naturally without any active aeration system.² On the other hand, HPV systems achieve a homogeneous distribution throughout an enclosed area by delivering a heat-generated vapor of 30 to 35% w/w aqueous HP through a high velocity air stream. HPV systems have modules to measure the concentration of HP, and the temperature and relative humidity in the enclosure, with some systems having the technology to hold a steady HP concentration throughout the exposure period. Following exposure, HPV systems catalyze the breakdown of HP vapor to oxygen and water vapor using an aeration unit.²

Non-manual technologies based on UV light also vary. Ultraviolet-C (UVC) systems use specifically designated wavelengths (254 nm range) and deliver targeted doses for vegetative bacteria (for example, 12,000 µWs/cm²) or for spores (22,000 to 36,000 µWs/cm²) on surfaces, while the pulse xenon UV systems emit broad spectrum UV in short pulses and have relatively short cycle time.² Ultraviolet germicidal irradiation (UVGI) is an air purification technology produced by mercury vapour lamp with a predominant wavelength within the UVC bandwidth of the electromagnetic spectrum.^3,16 The UVGI is a specialized system installed through upper room fixtures and the lamps can be placed inside mechanical ventilation systems.¹⁶ Air currents rapidly carry pathogens into the UVGI energy beam located well above the occupants’ heads, which destroys their DNA, interfering with replication, and inactivating them.¹⁶ By locating the UVGI in the upper part of rooms, occupants are protected from direct UV irradiation while the system works safely and effectively to interrupt the transmission of airborne infectious diseases.^1,16

The literature search did not find any studies which evaluated steam cleaning, ozone disinfection, high-intensity narrow-spectrum light, or polycationic and light activated surfaces as decontamination technologies that met the specified inclusion criteria for this report (Table 1). Also, no studies meeting the inclusion criteria were found evaluating anti-microbial coating or sharkskin-like surfaces as disinfection techniques. A review¹³ has briefly discussed some of these technologies and may help explain in part why they are not in use currently. Some of limitations of the steam technology are the risk of hazards to switches, computers, and electrical appliances, as well as increased risk of slips and falls due to residual moisture.¹³ Furthermore, while careless handling of steam increases the risk of burns and scalds for nearby persons, including patients, the temperature of steam at delivery may rapidly dissipate depending upon the type and conductivity of exposed surfaces,¹³ with potential for reduced effectiveness at inactivating pathogens. For ozone, its toxicity, limited effectiveness against bacterial spores and fungi, and potentially corrosive effect on materials (metals and rubber) commonly found in hospital equipment were identified as some limitations. Application of surface technologies in general are limited by insufficient information on durability and whether antimicrobial activity is affected by factors such as humidity, temperature, cleaning frequency, and/or the presence of an organic load.¹³ There are also concerns over possible toxicity, resistance, and allergenic properties, in addition to uncertainty about their relative contribution toward hand contamination and the risk of cross-transmission as a consequence. Moreover, the sites, surfaces, and clinical equipment in patient areas which could be coated with an antimicrobial product are currently unknown.¹³ The HINS light technology also requires further work to investigate any benefits on HAI rates, although, according to the review, one study has evaluated its overall effect for decontaminating the clinical environment.¹³

Outcomes

One non-randomized study¹¹ had incidence rates of hospital acquired multidrug resistant organisms (MDROs) and C. difficile as outcomes, and another study¹ reported general hospital acquired infection (HAI) rates in a long-term care ward. One study³ measured changes in tracheal colonization and prevalence of ventilator-associated pneumonia (VAP) among intubated NICU patients. Tracheal colonization was defined by the investigators using an airway microbial load index (MLI) which quantified each pathogen per patient sample on a scale of 1 to 4 for rare, few, moderate, or heavy growth. Patients whose tracheal aspirates showed no growth were assigned a zero.³ Cultures from the environment and intubated NICU patients’ tracheas were obtained before eUVGI installation and over the next 12 months. Episodes of VAP, number of antibiotic courses, and antibiotic days, among other outcomes, were compared between the pre- and post-eUVGI time periods. Another study⁵ measured new C. difficile-associated disease (CDAD) cases, both hospital-wide and in five rooms designated as high-incidence CDAD wards.

Although the included studies also measured colony forming units (CFU), these are not discussed because the focus of this report is on clinical outcomes such as infection prevention/reduction following disinfection by the non-manual room disinfection interventions of interest.

Reporting

The objectives and the main study outcomes of each of the four non-randomized studies were clearly defined, and the non-manual disinfection techniques being evaluated were specified.^1,3,5,11 All the studies reported percentage reductions in the specific incidence of hospital-acquired infections they had pre-specified as outcome of interest. Although they also indicated degree of statistical significance with P-values, because the confidence intervals (CI) were not reported the level of certainty of the reported outcomes in these studies is unknown. However, one of the studies¹¹ also reported rate ratios with corresponding 95% CI. Only two studies^1,3 provided any information on patient characteristics, making it difficult to evaluate the potential for confounding. One study¹ reported some demographic and medical conditions of the hospitalized patients before and during the intervention, and another study³ reported the demographic profile of NICU and high-risk cohort. One study³ reported procedure and outcome determination in sufficient detail to facilitate replicability. Another study⁵ reported trends in rates of antimicrobial and proton pump inhibitor (PPI) use, which are known to be risk factors for infections, in both the pre-intervention and the intervention phases of the study. Overall there was similarity between hospitalized patients groups in the two study periods, with respect to the use these medications.

External validity

Each of the non-randomized studies was conducted in a single hospital, or hospital department. This may limit the extent to which the findings of the study can be generalizable. The hospitals where each of the studies took place provided specialized services or had logistics and staff that may not be commonly found is other health care facilities.

The study that evaluated the pulsed xenon UVD¹¹ took place in a tertiary care hospital that offers full services to adult and pediatric patients including specialized services for trauma, burn, neurosurgery, cardiothoracic surgery, transplant, and oncology. The broad scope of patients and services suggests commonality with many healthcare facilities. However, all pediatric rooms were single occupancy, while most adult patient rooms outside of the intensive care units were double occupancy. Patients with MDROs or C. difficile received care in a private room or semiprivate room with the other bed blocked from occupancy, or they may have been cohorted with another patient who harbored the same organism.¹¹ It is reasonable to expect that such measures could contribute to minimizing dissemination of pathogens and spread of infection in the hospital. Thus, it is uncertain whether the same extent of success with UVD could be replicated in hospitals without sufficient room to allow this sort of occupancy arrangement.

One study¹ was conducted in a severe head injury long-term care ward and it is unclear whether its findings would be generalizable to other health care settings. One study³ was conducted in the NICU of a university-affiliated regional perinatal center, and the study benefited from environmental sample collection services provided by a research-based company. Another study⁵ was conducted in a university-affiliated hospital and focused on a particular strain of C. difficile which has enhanced virulence properties – the North American pulsed-field (NAPI) strain. Therefore, the generalizability of findings of these studies is unknown.

Internal validity

The adequacy of sample sizes to detect differences in effect of the interventions was not discussed in any of the studies. However, the non-manual room disinfection methods were used over at least 6 months and/or repeatedly for many cycles to allow sufficient data to be collected for analysis. The pre- and post-intervention comparisons used by all the non-randomized studies^1,3,5,11 are subject to possible clinical care and/or environmental changes over time, and there is no way of evaluating the extent to which such variations, if they occurred, influenced the reported outcomes of the studies. In all the studies,^1,3,5,11 it was not reported whether or not cleaning staff were aware of the use of the non-manual room disinfection procedures. It is reasonable to expect that knowledge of the investigation could influence behavioral change among housekeeping staff to increased or decreased intensity of cleaning which could impact the outcomes of the studies.

In one study,¹¹ the UVD system was used a high number of times (11,389 times) following discharge cleaning of contact precautions rooms and other high-risk area during the study period. This reduced the probability that the results were due to chance, although the study was conducted in a single institution. However, in this study,¹¹ the UVD was used exclusively at a setting recommended to inactivate C. difficile spores which is higher than the setting require for vegetative forms of C. difficile. Thus, we are unable to tell how effective the UVD would be against vegetative form of C. difficile using the appropriate recommended setting. However, this may not be problematic since disinfection against the transmission C. difficile associated disease usually targets both the spores and vegetative forms. Although there were several initiatives to optimize environmental disinfection during both the UVD and pre-intervention periods of the study,¹¹ they were not adjusted for in the analysis despite being potential confounding factors. On the other hand, although investigators indicated the use of a more sensitive diagnostic test (a change from C. difficile cytotoxin A + B enzyme immunoassay to polymerase chain reaction) increased overall test positivity from 10% to 13% during the study,¹¹ C. difficile infection rates decreased during UVD which supports the effectiveness of the intervention for this purpose.

In another study,¹ data from parallel periods before and after the intervention were analyzed using rigorous statistical methods to compare the differences between the two populations in terms of patient medical characteristics, treatments, and nosocomial infections.¹ The period before the non-manual room disinfection was introduced had more patients with pressure sores or who were using steroidal treatments. Pressure sores may suggest very ill patients on admission for a prolonged period, who may be more susceptible to infection by reason of longer exposure and/or reduced immunity. Steroid use has also been liked to reduced immunity and infections. However, the effect of possible potential cofounders, such as patient age, gender, mobility, presence of sores, steroid administration, tracheostomy, urinary catheter, and inhalation treatments, on the differences found in fever days, use of antibiotics, and rates of hospital-acquired infections, was analyzed and accounted for using multivariate analysis of covariance (ANCOVA). According to the authors, the study was designed to use parallel periods to minimize seasonal variations between the study periods. However, no mechanisms were described to show how variations would be detected if they occurred. Thus, it is unknown if seasonal changes affected the reported outcomes and to what extent. Furthermore, although the study used data from two 6-month parallel periods, the number of patients who were hospitalized before and during the intervention was relatively low (57 and 51, respectively). It is therefore, uncertain whether the sample size was enough to detect clinically relevant differences between the two periods.

In one of the studies,³ the periods before and after intervention were neither parallel or equal in length. Data was collected for the 6-month period before the installation of the non-manual room disinfection intervention, and for four 6-month periods after the intervention was installed. Therefore, although the data from the duration of the intervention show reduced VAP gains compared to the 6 months before the intervention, the influence of seasonal variations and changes in clinical care over time on the reported findings cannot be ruled out.

In another study,⁵ the incidence of nosocomial CDAD was investigated hospital-wide and in five high-incidence wards before and after HPV decontamination. Patient groups hospitalized during the two study periods had similar levels of treatment with antibiotics and PPIs (both of which are known to be risk factors for C. difficile transmission) without statistically significant differences which could influence the reported HAIs. Although an analysis to examine the effects of antimicrobial medication and PPI use on outcomes was done, details about patient characteristics, medical history and other potential confounding factors were not provided. To distinguish between patients with hospital-acquired CDAD and patients who were infected before hospital admission, nosocomial CDAD case diagnosis was limited to patients with a positive C. difficile toxin test result for a test obtained more than 72 hours after admission.

Funding Support

One of the non-randomized studies¹¹ reported no conflict of interest. One study¹ was funded in part by the company that developed the copper oxide in linen technology that was being studied. In addition one of the investigators was the Chief Medical Scientist of the company. In one study,³ the investigators received an eUVGI system, which was the non-manual disinfection technology being investigated, along with installation as an in-kind contribution from the manufacturer who also provided environmental sample collection services for the study. Another study⁵ received price discounts for HPV decontamination services from the manufacturer of the HPV technology under study, and two out of 10 investigators received salary, at least in part, from the same manufacturer.

What is the clinical effectiveness of non-manual room disinfection methods for infection prevention in health care facilities?

One retrospective study¹¹ reported a reduction in the incidence of hospital-acquired MDROs and C. difficile from 2.67 cases per 1,000 patient-days in the 30-month period before UVD to 2.14 cases per 1,000 patient-days during the 22 months when UVD was used. This represented a decrease of 20% with a rate ratio of 0.80 (95% confidence interval [CI]: 0.73 to 0.88; P < 0.001).

A cohort study¹ found the use of biocidal copper-impregnated linens reduced hospital-acquired infection (HAI) rate per 1,000 hospitalization-days by 24% (P < 0.05) compared with the use of ordinary linens. The use of biocidal copper impregnated linens also reduced the number of days patients had fever (body temperature >38.5 °C) per 1000 hospitalization-days by 47% (P < 0.01), and total number of days of antibiotic administration per 1000 hospitalization-days by 32.8% (P < 0.0001) compared to the use of ordinary linens. Expectedly, there was a reported cost saving (approximately 27%, data not provided) as a result of reductions in antibiotics use, HAI-related treatments, X-rays, disposables, labor, and laundry expenses during the period when biocidal copper impregnated linens were used.

In one prospective study,³ fewer NICU patients were found to be colonized following eUVGI, and tracheal microbial loads decreased by 45% (P = 0.004). The percentage of patients who had little or no tracheal colonization (MLI≤1) was 44% post-eUVGI compared with 14% pre-eUVGI. In addition, VAP rates among high risk cohorts, defined as infants with less than 30 weeks gestation who were ventilated for 14 weeks or more, declined significantly (P = 0.04) from 74% at baseline (n = 31) to 44% (n = 18) at 18 months. The overall antibiotic usage fell by 62% (P = 0.013) while episodes of VAP per patient also decreased significantly (P = 0.04) from 1.2 to 0.4 for the period before and during the 12 months of eUVGI, respectively.

Another prospective study⁵ reported that the incidence of nosocomial CDAD in the five high-incidence wards reduced from 2.28 cases per 1,000 patient-days in the pre-intervention period to 1.28 cases per 1,000 patient-days during a similar time period for the intervention (P = 0.047). For the hospital-wide incidence of CDAD, although HPV intervention resulted in lower incidence, the difference between the pre-intervention and the intervention periods reached the level of significance only when the analysis was limited to months when an epidemic strain was known to be present. During that time, 0.88 versus 1.89 cases per 1,000 patient-days (P = 0.047) was reported for the intervention and pre-intervention periods, respectively.

What is the comparative clinical effectiveness of non-manual room disinfection methods for infection prevention in health care facilities?

The literature search did not find any study with a direct or indirect comparison of non-manual room disinfection methods for the prevention of infection in health care facilities..

What is the cost-effectiveness of non-manual room disinfection methods for infection prevention in health care facilities?

The literature search did not produce any studies on the cost-effectiveness of non-manual room disinfection methods for infection prevention in health care facilities.

What is the comparative cost-effectiveness of non-manual room disinfection methods for infection prevention in health care facilities?

The literature search for this report did not identify any studies on the comparative cost-effectiveness of non-manual room disinfection methods for infection prevention in health care facilities.

What are the evidence-based guidelines regarding non-manual room disinfection methods for infection prevention in health care facilities?

The literature search for this report did not identify evidence-based guidelines regarding non-manual room disinfection methods for infection prevention in health care facilities.