Homogeneity and standardization in hand hygiene compliance

To the Editor—Healthcare-associated infection (HAI) is a major problem for patient safety and hand hygiene is recommended as one of the most effective strategies for preventing HAI. ^1,2 In 2004, WHO launched the global hand hygiene campaign to improve hand hygiene practices, which included 5 indications for hand hygiene: before patient contact, before an aseptic task, after body fluid exposure risk, after patient contact, and after contact with patient surroundings. ³

Monitoring healthcare workers’ adherence to recommended hand hygiene practices is considered an important element of an effective hand hygiene program. Because many factors (eg, observation bias, selection bias, information bias and the Hawthorne effect) can occur during monitoring, increasing attention has been given to reduction of these biases to achieve more accurate measurement of hand hygiene compliance. ^4–6 Notably, overall hand hygiene compliance observed in different studies should not be compared directly, due to heterogeneity among the studies. That is, the overall hand hygiene compliance should not be directly compared when the proportion of observations conducted among each of the 5 indications for hand hygiene differs among the studies.

Taking the following research studies as an example (Table 1), hand hygiene indications 1 and 2 in study 1 show lower compliance than those in study 2. However, study 1 (53.8%) exhibits a higher overall compliance than study 2 (46.3%). The main reason for this inconsistency is the difference in the proportion of various indications for hand hygiene between these 2 studies.

Table 1. Example of Different Proportions of Hand Hygiene Indications

Accordingly, we believe that homogeneity and standardization should be considered not only at the design stage of every hand hygiene monitoring scheme but also at the time of compliance analysis. We also recommend that these factors should be added to the new version of the WHO Guidelines on Hand Hygiene in Health Care.

Homogeneity in design

To allow accurate comparison of hand hygiene compliance, researchers should ensure that each hand hygiene indication has the same proportion of observed opportunities during each study period. For example, in each study period, 40% of total observations are conducted in relation to indication 1, 20% of observations are conducted in relation to indication 4, and 40% of observations are conducted in relation to indication 5. Using this scheme, all observational samples will be homogeneous and the overall compliance rates will be comparable because the proportion in each sample is equal for the same indication. When publishing compliance results, it is necessary to report the total number of hand hygiene opportunities and actions regarding the 5 individual indications as well as total hand hygiene opportunities.

Standardization in comparison

When hygiene compliance from different studies is compared, standardization is needed. Direct and indirect standardization can be applied, and the use of these 2 standardized methods should depend on the availability of data. Direct standardization can be carried out using the following formula:

(1) $$p' = {{{N_1}{p_1} + {N_2}{p_2} + \ldots + {N_i}{p_i}} \over {{N_1} + {N_2} + \ldots + {N_i}}} = {{\sum {N_i}{p_i}} \over {\sum {N_i}}}$$

The numerator of p′ may be recognized as the number of standardized actions and the denominator of p′ is the overall standardized opportunity number. N _i represents the standardized opportunity number (the sum of opportunity numbers of every corresponding hand hygiene indication in all studies), and p _i is the respective original compliance of each hand hygiene indication. In this example, the values of N ₁, N ₂, p ₁, and p ₂ were 500 (ie, 300 + 200), 700 (ie, 100 + 600), 60.0%, and 35.0% in study 1, respectively. Therefore, the overall standardized compliance should be calculated as 45.4% for study 1 and 50.4% for study 2 using the direct standardization method, which corrected for the effect of having different proportions of observations performed among the various hand hygiene indications during the 2 studies.

Additionally, when the total number of hand hygiene actions (r) in all studies and the opportunity number of every indication (n _i ) are available but compliance is missing or the opportunity number is small in each indication, the indirect method, based on the following relationship, is appropriate:

(2) $$p' = P{r \over {\sum {n_i}{P_i}}}$$

Here, P is defined as the overall hand hygiene compliance published in a reference study, and P _i is the reference compliance of each hand hygiene indication obtained from the same study. Hence, n _i P _i is the expected action number of each hand hygiene indication.

In summary, a threat to meaningful hand hygiene compliance measurement is bias, which includes selection, observation, and confounding biases. If hand hygiene compliance is compared between healthcare settings or over time, homogeneity of measurement and standardization of results should be considered. Another essential component of meaningful hand hygiene compliance measurement is an appropriate sample size, as described in the WHO recommendations. ⁷