Appendix 2. Controlled studies of the effect of handwashing on transmitting respiratory infections

First author

Nature of study

Pertinent results

Pertinent conclusions

Carabin (1)

Randomized, controlled trial in 47 daycare centers (for children <5 years of age) in Quebec, Canada. Randomization was by center after stratification by incident rate of respiratory infection. Intervention was increased handwashing in children and staff by a single staff training session. Outcomes were upper respiratory tract infections and diarrheal disease in children (measured coliform contamination but no viral microbiology, winters of 1996 and 1997).

Outcome measures were recorded in intervention and control groups in each center in the autumn of 1996 (before intervention) and 1997 (after intervention). Compliance was measured and showed that the intervention had been carried out. Both groups had a decrease in respiratory infections and diarrheal disease; however, intervention groups experienced greater and significantly reduced rates after intervention than control centers. The reduction in upper respiratory infections was 25%, but little effect on diarrheal illnesses was seen. Environmental contamination (with coliforms) was reduced in both groups during the intervention, which suggests spillover of the intervention.

Handwashing reduced the incidence of upper respiratory infections in children <5 years of age.

Dyer (2)

10-week cross-over intervention study among 420 schoolchildren (5–12 years of age) in California compared handwashing and enhanced supervised handwashing and use of a hand sanitizer. Outcome measures were absences due to infectious diseases (no microbiology, early spring 1998).

School absences due to infectious diseases during the enhanced handwashing period were 42% lower than in the ordinary period. For absences due to gastrointestinal disease and respiratory infections, the reductions were 29% and 50%, respectively. The effect was consistent in both periods of the trial, and all reductions were significant.

Enhancing handwashing and use of hand sanitizers among children in school reduces infection.

Falsey (3)

US intervention study in 3 eldercare homes with a historical control period. Intervention was to get staff to wash their hands between clients (residents) (virologic studies, winter 1995/1996).

In 3 preintervention years, rates of respiratory infection in the elderly were 14.5, 12.8, and 10.4, respectively, per 100 person-months, and rates declined significantly (to 5.7) in the intervention year. The equivalent rates for staff were 21.0, 13.9, 11.3, and 9.5, respectively, with no significant decline. Virologic testing indicated only 37 influenza isolations among 392 illness episodes during the 4 years. No change in specific viruses could account for the decline in year 4.

Staff handwashing seemed to be associated with reduced incidence of respiratory infection in the elderly but not in staff; however, the use of a historical control period can be misleading.

Larson (4)

Randomized, double-blind, controlled trial in 238 families in an American city compared the effect of antibacterial and conventional soaps and other products. Outcomes were self-reported symptoms (no microbiology, 48 weeks in an unstated year).

Most symptoms were respiratory. No significant differences were seen in runny noses, fever, cough, or sore throat between intervention and control families.

No advantage to using antibacterial versus conventional washing materials was found in this industrialized country setting.

Luby (5)

Community-cluster, randomized, controlled trial in urban setting (Karachi, Pakistan) compared handwashing promotion in all family members with outcomes of diarrheal disease and lower respiratory tract infections (no microbiology, 12 months, 2002–2003).

Children <15 years of age in intervention clusters had lower incidences of cough and breathing difficulty compared with a control (no intervention) group. Children <5 years of age had lower rates of pneumonia, diarrhea, and impetigo in the intervention versus control groups. No advantage of using antibacterial versus ordinary soap was seen.

In this study in a developing country, handwashing had a significant effect in protecting children against respiratory infections of unknown cause. Although most infections would be viral, only a small proportion might be expected to be due to influenza virus.

Ponka (6)

Open-clustered, unrandomized intervention study in daycare centers in urban setting (Helsinki, Finland) involved 60 centers with 228 controls. The intervention involved training in increased handwashing among children and staff plus other hygiene measures, including cleaning surfaces and toys, toileting hygiene, excluding ill persons, and some instruction of parents. Outcomes were effect on absences due to infections (no microbiology, winter and spring 1999–2000).

For children <3 years of age, intervention centers had significantly fewer absences due to all infections and respiratory infections in the intervention period compared with a baseline period. The crude percentage reduction in absences due to upper respiratory infections was 39% and that due to all infections was 32%. No such effect was seen in the control centers, and no effect was seen in children 3–6 years of age in either intervention or control centers.

An effect of a combination of hygiene measures was seen but only in young children, and handwashing was only 1 measure.

Roberts (7)

Randomized, controlled trial in 23 of 26 daycare centers in an Australian city involved 11 intervention centers and 12 control centers. Compliance (handwashing and wiping children's noses) was measured (no microbiology, 1996).

A significantly lower number of episodes of illness was seen in children <2 years of age, with no significant effect in older children or all children. Rates of absence were lower in the intervention centers, but the difference was not significant. Where compliance was measured against illness rates, a 17% reduction in colds was seen in younger (<24 months) children with no effect in older children.

The study did not support the hypothesis that infection rates could be reduced by handwashing, although this finding could be due to poor compliance with the intervention.

Ryan (8)

Large observational study using a historical control period was undertaken before, during, an after a handwashing intervention among military recruits in the United States. Outcome measures were compliance rates, reported illness, and outpatient and hospitalization rates (limited microbiology, streptococcal cultures; 1996–1998).

A 45% reduction in reported outpatient (primary care) consultations was seen for respiratory infections, with no effect on hospitalization. Those complying with the intervention had a significantly lower rate of reported respiratory infections than those not complying (3.2 vs. 4.7 episodes per recruit).

Although the intervention had statistically significant effects, this finding must be interpreted cautiously because of the use of a historical control period.

Uhari (9)

Randomized, controlled trial in daycare centers in Helsinki, Finland, compared handwashing promotion in staff, children (<5 years of age), siblings (outside the nursery), and parents. Outcomes were all infections and absences (no microbiology, 15 months in 1991 and 1992).

A small but significant difference was seen in all infections and symptoms attributable to respiratory infections (rhinitis and cough) in children (lower in the intervention group). Infection rates were also lower in the staff, but the article does not mention respiratory versus other infections. Parents of children in the intervention groups missed less time from work because of less illness among their children, but no difference was seen in parental or sibling illnesses.

In this study in a well-resourced country, handwashing had a significant effect on protecting children against respiratory infections of unknown cause. No measurable benefit was seen in protecting families against background-level infectious disease by intervening with their children in nurseries. That finding does not exclude an effect during an outbreak or pandemic.

White (10)

Randomized, controlled trial in 4 university residence halls (430 students total) in the United States compared handwashing promotion based around an alcohol-based hand sanitizer (2 halls) versus no intervention. Hand sanitizers were available in both groups but not promoted in the control group (no microbiology, autumn 2001).

Somewhat greater handwashing and far greater use of hand sanitizers were seen in intervention than control residences. Intervention groups had 20% less illness overall and lower rates of all respiratory symptoms (including sore throats, stuffy noses, fever, cough).

In this small study, handwashing and use of a hand sanitizer seemed to protect against respiratory illnesses. No conclusion could be drawn about the additional value of the sanitizer.