• 
• 

Three Part Question

For [patients admitted to hospital] does staff use of [ alcohol-based hand rub] reduce the risk of [hospital acquired infection]?

Clinical Scenario

You have noticed the appearance of numerous dispensers for alcohol based hand gel within your emergency department. This is accompanied by a poster campaign to make staff aware of the importance of hand hygiene. You wonder what evidence lies behind the campaign.

Search Strategy

Ovid MEDLINE (1950 - October Week 2 1950)
EMBASE (1980 - 2007 wk 42)
Cochrane Database of Systematic Reviews (3rd quarter 2007)
Cochrane Central Registrar of Controlled Trials (3rd quarter 2007)
CINAHL (1982 - wk3 October 2007)
[alcohol gel.mp OR alcohol rub.mp OR alcohol hand gel.mp OR alcohol hand rub.mp OR spirigel.mp OR alcogel.mp OR sanitizer.mp OR degerming agent.mp OR antiseptic.mp] AND [hospital acquired infection.mp OR hospital associated infection.mp OR healthcare associated infection.mp OR health care associated infection.mp OR hai.mp OR hcai.mp OR mrsa.mp OR methicillin-resistant staph$.mp OR multi-resistant staph$.mp OR clostridium.mp OR coli.mp OR nosocomial.mp

Search Outcome

588 papers were found. 9 papers were considered relevant to the posed question.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Pittet, D; Hugonnet, S; Harbarth, S;et al 2000 Switzerland	Nosocomial infection rates in large, acute-care teaching hospital in Geneva.	Infection identified by trained infection-control nurses and monitored since 1994, hand hygiene program introduced in 1995.	Hospital wide nosocomial infection prevalence	16.9% in 1994 vs. 9.9% in 1998 (p=0.04)	Have mentioned that there were significant changes in the choice and usage of different antibiotlc types during this time period but have not looked at any other potentially confounding factors over this time period e.g. bed occupancy or length of hospital stay.
			MRSA infection incidence per 10,000 patient days	2.16 to 0.93 episodes (p<0.01)
			MRSA bacteraemia per 10,000 patient days	0.74 to 0.24 episodes
Fendler, E; Hammonda, B; Lyons M; et al 2002 USA	257 bed extended care hospital with 265 employees.	2nd and 3rd floors of the hospital provided with alcohol hand gel. Ground floor hygiene regime remained unchanged.	Infection rates per 1000 patient days	2.27 in alcohol gel floors vs. 3.19 in the control area.	Patient groups different: 2nd and 3rd floor patients had sub-acute medical conditions and required intensive staff assistance. There was a higher turnover of patients. Patients in the remaining units were patients with long-term medical illness.
Rao, G; Osman, J; Osman, C; et al 2002 UK	Patients at a 600 bed teaching hospital in the UK.	Looked at the incidence of hospital-acquired methicillin-resistant Staphylococcal aureus (HAMRSA) and Clostridium difficile associated diarrhoea (CDAD) in 2000-1 following the introduction of bedside Spirigel compared with the incidences in the prior year. Defined HAMRSA as patients who were negative for MRSA when screened on admission but were positive on subsequent screening and patients who were not screened and then found to be positive for MRSA at any point after 3 days from admission. Did not actually use the incidence of HAMRSA but the proportion of HAMRSA over the total incidence of MRSA.	Percentage of total MRSA cases thought to be hospital acquired	Nearly 50% in 1999-2000 versus 39% in 2000-2001.	Only provided data on percentage of HAMRSA over MRSA not the raw figures so cannot tell if actual incidence of MRSA or HAMRSA was increasing or decreasing. No data provided on how many patients found to be positive for MRSA were actually screened on admission.
			Average incidence of CDAD	11.5 per 1000 patients admitted in 1999-2000 vs. 9.5 per 1,000 patients admitted in 2000-2001 (p=0.2)
MacDonald, A; Dinah, F; MacKenzie, D; Wilson, A 2003 UK	Plastic surgical unit of 600 bed district general hospital.	Monitored patients considered to be newly infected with MRSA for one year prior to introduction of alcohol hand gel and one year following this introduction. Decision as to whether the infection was new made by the infection control team on the basis of screening results and whether the positive swab was carried out 48h after admission.	Incidence of new MRSA infection	23 out of 1211 patients vs. 11 out of 1261 patients in the following year. (p<0.05)	Subjective decision made as to whether or not the MRSA was acquired as an in-patient. Data not provided on the total MRSA incidence for the studied time periods. Introduction of alcohol gel accompanied by a campaign for cleaner hands.
King, S 2004 UK	Patients in a 28 bed surgical ward in a UK district general hospital.	Observation of incidence of MRSA and Clostridium difficile infections in the ward in a 3 month period following introduction of alcohol hand gel.	Incidence of MRSA	4 patients during trial period vs. mean of 8 (range 5-10) in the same 3 month period over the previous 4 years.	Very small trial. No power calculation. No information provided about the patient numbers.
			Incidence of Clostridium difficile	5 cases over the trial period vs. no cases in the previous 12 months.
Ng, P; Wong, H; Lyon D; et al 2004 Hong Kong	Very low weight infants admitted to a neonatal intensive care unit over a 6 year period.	There was a new policy of hand hygiene introduced half-way through the study period where staff had to wear gloves and use alcohol gel for all procedures. The incidence of sepsis for the periods before and after the intervention were compared.	Incidence of gram +ve infection (including coagulase negative staphylococcus).	7.5 vs. 3.0 infections per 1000 patient days.	Patients in the first group were of lower gestational age, weighed less, were ventilated for longer and spent longer on supplemental oxygen compared with the group following the intervention. Other potential confounding factors e.g. antibiotic policy or length of stay on the unit not discussed. Not clear when study was conceived i.e. was the data prior to the intervention obtained retrospectively and the data following the intervention obtained prospectively?
			Incidence of gram negative infections	2.3 vs. 0.9 per 1000 patient days.
			Incidence of MRSA	20 cases out of 161 patients in the first group vs. 2 cases per 176 patients in the second group.
			Incidence of NEC requiring surgery	5 cases in the first group vs. 4 cases in the second group.
			Incidence of infection related deaths	4 cases from the first group vs. 2 cases in the second group.
Kew Lai, K; Fontecchio, S; Melvin, Z; et al 2006 USA	Patients in two hospitals associated with the same university and under the care of the same infection control team. The University campus hospital is a 388 bed unit with 6 ICUs, the Memmorial campus is a 275 bed hospital with 3 ICUs.	The use of alcohol was introduced to the University Campus hospital but not to the Memorial Campus hospital. Incidence of MRSA was monitored at both hospitals for 5 months before and after the introduction of the alcohol hand gel. Incidence of vancomycin resistant enterococcus was monitored at the Medical ICU ward at the University Campus hospital before and after the introduction of the alcohol hand gel. Additional screening was carried out on patients in wards with 2 or more cases of MRSA at the University Campus hospital to allow the effects of clustering to be calculated (i.e. the more patients who have MRSA, the more likely other patients are to be infected.)	Incidence of MRSA at the University campus	56 cases over the first 5 months (1.26 cases per 1,000 pt days ) vs. 40 cases over the last 6 months (0.75 cases per 1,000 patient days) p=0.037	Not clear how modeling was done to account for the effects of clustering. Some data is provided regarding the incidence of VRE but I cannot see how the calculation was performed to account for a reduction of incidence from 12 to 3 cases per 1,000 pt days. I calculate a reduction from 12 to 8 cases per 1,000 pt days.
			Incidence of MRSA at the Memmorial campus	14 cases over first 5 months ( 0.34 per 1,000 patient days) vs. 24 cases over the last 6 months (0.49 cases per 1,000 pt days) not significant.
			Incidence of MRSA at University campus allowing for effects of clustering	No significant change following the intervention
			Incidence of VRE (colonisation or infection at >72h from admission) at the University campus	29 cases (12 per 1,000 pt days) prior to intervention and 19 cases following intervention (3 per 1,000 pt days) p<0.001
Hilburn, J; Hammond, B; Fendler, E; et al 2003 USA	Patients at a 498 bed acute care facility with 1700 employees.	Baseline data collected for 6 months then data collected for 10 months following the introduction of hand gel.	Infection rate (ratio of infections per 1000 patient days)	For first 6 months was 8.2%, for next 10 months was 5.3%.	Different time periods compared at different times of the year. No justification provided for time periods studied. Little raw data provided, although it was explained how the prevalence of infection would be calculated the data was not provided.
Mahamat, A; MacKenzie, F; Brooker, K; et al 2007 UK	Looked at 2 hospitals in NE Scotland. The Intervention Hospital (IH) is a 200 bed DGH while the Control Hospital (CH) is a 300 bed hospital for the elderly and elective orthopaedic surgery.	Monitored the rates of MRSA in in-patients over a five year period and looked at the effects of several interventions on infection rates in the IH vs. the CH where these changes were not applied. Both hospitals introduced alcohol hand gel at the same time but the other interventions were only at the IH.	Calculated decrease in %MRSA due to the introduction of alcohol hand gel	21% reduction in the IH and 30% reduction in the CH.	Not explained why use % MRSA (i.e. cases of MRSA / cases of SA*100) rather than incidence of MRSA. Incidence of MRSA continued to increase at both hospitals over the time period studied.

Comment(s)

Hospital acquired infection (HAI) has been recognised as an important cause of mortality and morbidity for centuries. Recently it has been the focus of many articles in the media and in the UK the government has set national targets to reduce the incidence of certain infections. Various measures have been introduced to hospitals in order to decrease the incidence of HAIs, particularly methicillin-resistant staphylococcus aureus. Most hospitals have introduced various types of alcohol hand rubs at convenient sites to encourage hand hygiene. The use of alcohol gel is promoted as it can be made more widely available, is quicker and is less irritating to skin than hand washing. It is effective at killing most bacteria though is less effective for spores (e.g. clostridium). The evidence for how the promotion and provision of alcohol hand rub translates into reduction of infection transmission is largely from longitudinal interventional studies. There are no randomised controlled trials. The studies that have used a control group have generally used 2 different populations. This form of research is extremely open to confounding factors which change over the time period given e.g. antibiotic use (types and quantities); length of stay in hospital; promotional campaigns; media campaigns; use of barrier nursing etc. The open nature of the trial also allows bias. The method used for calculating infection incidence tends to be the number of HAIs divided by the number of patients seen over a given time period calculated as the number of infections per patient days. This method does not take into account the length of stay for individual patients. As most of the studies only count infections detected 72h after admission as HAIs patients staying for shorter periods will not be included. Other research suggests that the risks of acquiring infection as an in-patient increase with length of stay, peaking at around 18-21 days following admission(1). It would be better to describe the risk of acquiring infection in a way that incorporates information on the length of stay. Only one study included interpretation of data to allow for clustering, i.e. the more patients with an infectious condition in an environment the more risk there is of transmission, the converse also being true. This is particularly pertinent when comparing populations with a different prevalence of infection. All of the included studies are positive for the outcomes that they have looked at. The translation of these findings into a definite clinical benefit is less clear but it is likely that the incorporation of alcohol hand rub into good hygiene practice with regular hand washing and the use of gloves will reduce transmission of infection between patients. The evidence to confirm this theory is limited.

Clinical Bottom Line

The use of alcohol gel or rubs to improve hand hygiene in the hospital setting is likely to reduce the risk of hospital-acquired infection.

Level of Evidence

Level 2 - Studies considered were neither 1 or 3.

References

1. Pittet, D; Hugonnet, S; Harbarth, S;et al Effectiveness of a hospital-wide programme to improve compliance Lancet 2000; 356: 1307-12
2. Fendler, E; Hammonda, B; Lyons M; et al The impact of alcohol hand sanitizer use on infection rates in an extended care facility American Journal of Infection Control 2002; 30:226-33
3. Rao, G; Osman, J; Osman, C; et al Marketing hand hygiene in hospitals - a case study Journal of Hospital Infection 2002; 50:42-47
4. MacDonald, A; Dinah, F; MacKenzie, D; Wilson, A Performance feedback of hand hygiene, using alcohol gel as the skin decontaminant, reduces the number of inpatients newly affected by MRSA and antibiotic costs Journal of Hospital Infection 2004; 56: 56-63
5. King, S Provision of alcohol hand rub at the hospital bedside: a case study Journal of Hospital Infection 2004:56: S10-12
6. Ng, P; Wong, H; Lyon D; et al Combined use of alcohol hand rub and gloves reduces the incidence of late onset infection in very low birthweight infants Archives of Diseases of Childhood 2004; 89:336-40
7. Kew Lai, K; Fontecchio, S; Melvin, Z; et al Impact of Alcohol-Based, Waterless Hand Antiseptic on the Incidence of Infection and Colonization with Methicillin-Resistant Staphylococcus Aureus and Vancomycin-Resistant Enterococci Infection Control and Hospital Epidemiology 2006;27:1018-21
8. Hilburn, J; Hammond, B; Fendler, E; et al Use of alcohol hand sanitizer as an infection control strategy in an acute care facility American Journal of Infection Control 2003; 31:109-16
9. Mahamat, A; MacKenzie, F; Brooker, K; et al Impact of infection conrol interventions and antibiotic use on hospital MRSA: a multivariate interrupted time-series ananlysis International Journal of Antimicrobial Agents 2007; 30:169-76
10. Tess B; Glenister H; Rodrigues L; Wagner M Incidence of Hospital-Acquired Infection and Length of Hospital Stay European Journal of Clinical Microbiology and Infectious Disease Feb 1993; (12) p81-86