The data collection strategy was modeled on a method that was previously used by Cure and Van Enk [8]. The hand hygiene and low-level disinfection observations took place at 12 units during the morning shift and 8 units during the evening shift. The study was conducted in inpatient medical and/or surgery, surgical and medical critical, emergency, women and children, and rehabilitation units from May to June 2019 by an observer that spent 1 hour at each unit. The “Observations” indicate the number of times the investigator observed compliant behavior (using hand sanitizer, washing hands, or cleaning equipment) and “Opportunities” correspond to the number of opportunities or instances in which compliant behavior should have been practiced. The compliance rate was then calculated by using the following equation:

If the door to a room was closed or the staff member was out of view, the behavior was not documented.

The usability of hand sanitizer dispensers was measured based on the criteria described elsewhere [8] and comprised the following: (1) easily visible on entry, (2) easy, unobstructed access, (3) close to the point of care, (4) visible from point of care, (5) along the workflow path, (6) close to the entrance or exit, and (7) placed at optimal height (85 to 110 centimeters). A final criterion, (8) visible on exit, was also added. 

Semistructured interviews were modeled using the existing Theoretical Domains Framework [9,10]. This framework was created to help assess the potential factors that may influence the behavior of health care professionals. The interview questions (Multimedia Appendix 1) were derived from a succinct table describing the 12 domain details, which are available elsewhere [9]. In total, 4 staff members were consented and interviewed, 2 each from units with ≥0.80 and ≤0.80 compliance, respectively. The Theoretical Domains Framework is useful for studying the implementation of desired behaviors among health care professionals and developing interventions to alter or improve behaviors [10].

Findings of the quantitative phase and semistructured interviews were shared with each unit’s clinical staff leaders along with an open-ended 8-question survey (Multimedia Appendix 2). The aim of the follow-up survey was to understand leadership perception and insight pertaining to the results of both steps of the study. This survey was answered by 2 leaders representing 4 main hospital units.

An Excel worksheet (Microsoft Corp) was used to provide frequencies, percentages, and average and total compliance rates. MedCalc for Windows Version 15.0 (MedCalc Software) was used for testing the statistical significance of different associations between morning and evening shift compliance, and reporting odds ratios (OR) and 95% confidence intervals. Qualitative analysis software NVivo Version 12 (QSR International) was used to classify responses under each theme and analyze differences in interview responses depending on the aggregate compliance score of the unit that interviewees were associated with.

Verbal informed consent was obtained from the participants of the qualitative survey. Units were informed of the observational part of the study prior to implementation. Although no personal identifiers of respondents and staff were recorded, the data was kept in secured and password-protected computers. The study was designed as a practice-based quality improvement project; hence, institutional review board approval was not required.

The characteristics and activities of the sample of staff observed within each unit are displayed in Table 1. We observed 222 staff members (156 during morning shifts and 66 during evening shifts). Of 526 hand hygiene and 33 low-level disinfection opportunities, compliance was observed 410 (78%) and 17 (51%) times, respectively.

Table 2 and Multimedia Appendix 3 (detailed data) show the compliance scores by each unit. During the day shift, 6 of the 12 observed units had hand hygiene compliance scores less than 0.80 while 5 of the 8 units observed during the night shift had below 0.80 compliance. Low-level disinfection compliance was below 0.80 in 5 of 11 units observed during the day shift and 3 of 5 units observed during the night shift.

We compared the odds of noncompliant behavior in hand hygiene between morning and night shifts by considering the night shift as adverse exposure and assuming that compliance may be lower during evening hours. The odds of noncompliant behavior in hand hygiene were higher among 6 of the 8 observed units; however, this relationship was not statistically significant (Table 3).

The observer determined the usability of hand sanitizer dispensers based on their placement in each unit. The scores were marked down as “0” if the usability factor was not met and “1” if it was adequate. In most units, the dispensers met criteria 2, 5, 6, 7, and 8 but did not meet criteria 1, 3, and 4 (Table 4). All units, except one, had similar usability characteristics; hence, their impact on hand hygiene compliance was not analyzed.

In total, 4 staff members, 2 each from units with morning shift compliance ≥0.80 (units B and E) and ≤0.80 (units D and G) consented to interviews. Using the Theoretical Domains Framework to analyze the results from the interview data, we identified several themes across units and compliance levels. Staff had the knowledge and skills to maintain hand hygiene and low-level disinfection of equipment; however, time shortage while seeing patients can make implementation of these skills difficult:

Despite the difficulty, staff members do think of hand hygiene and low-level disinfection as a standard part of their patient consultations. This indicates that they have the self-standard to adhere to the guidelines. Furthermore, staff had the self-efficacy to complete the steps required to perform these behaviors:

Self-efficacy can become an issue if the patient load is high or time is an issue:

The staff has greater self-efficacy when they see visual cues such as soiled hands. They were familiar with the consequences of not performing the behavior, with one participant stating that “Infection rates would go up and I... would get sick more often.” They also thought that these behaviors are automatic and do not require reminders; however, one respondent said, “you’re [a clinical staff member is] going to forget [to perform low-level disinfection of equipment] sometimes but the key is to try to be consistent.” When specifically asked if these behaviors take up too much time, they generally stated, “No because those wipes [which are used for low-level disinfection of equipment] dry out in like 2 minutes.” However, “if we’re [unit staff members are] busy and they can’t get to it [cleaning equipment using low-level disinfection] then that’s a chore that’s time-consuming.” This suggests that staff members mostly agree that the environment is conducive to performing these behaviors but that, in a given context, it can take up too much time. Other answers around environmental constraints varied and included the following:

Being highly aware of the risks involved with failing to carry out the protocols was the most common answer as to why the work environment motivated staff members to perform a behavior. 

Answers regarding social influences or norms also varied. Some staff members thought that social influences had a role in their hand hygiene behavior, including the following:

Other staff members indicated that social influences do not correlate with their hand hygiene or low-level disinfection of equipment behavior:

In addition, themes surrounding behavioral regulation varied greatly. Staff thought that maintenance of behavior was dependent upon access to supplies:

It also depends on time:

In contrast to behavioral regulation, social norms, and the environmental context domains explored, staff members agreed that emotion is not a factor in hand hygiene or low-level disinfection of equipment behavior. They also agreed that these behaviors are habitual within the given context.

Commenting on whether or not hand sanitizer dispensers located in the hallway, rather than the patient room, would change their hand hygiene habits, a unit G staff member answered, “I don’t think the hand sanitizer is good to tell you the truth.” A unit D staff member responded that “having it outside the room I think you’d [a clinical staff member would] see it more.” These units both had an aggregate compliance score of ≤0.80 for both hand hygiene and low-level disinfection behavior. The unit E interviewee responded with the following:

However, the unit B respondent did not think that any change would occur, stating “I think it would be the same.” Units E and B both had an aggregate compliance score of ≥0.80 for both behaviors.

When asked to comment on the changes that could improve low-level disinfection compliance rates, the 2 units with compliance ≤0.80 stated that access and supplies can help:

Suggestions from the other units with compliance ≥0.80 included education and staffing:

Regarding the availability of the cleaning supplies for low-level disinfection, staff generally believed that access to them is easy; however, there could be more supplies:

Table 5 shows the differences in interview responses given by staff from low- and high-compliance units. Each factor has an aggregate compliance score from 4 units. The scores under each factor were compiled by sorting interviewee responses into 36 relevant categories.

One theme that emerged from this data was that there might be issues with resources or access to supplies needed to perform these behaviors. For example, one respondent said that “resources are not readily available” and that having holders for the wipe containers in patient rooms could be a viable solution to improve rates of low-level disinfection behavior. In addition, the respondent believed another way of improving compliance scores would be to hold team members accountable. When asked about processes that are in place to maintain these behaviors, the respondents referred to the survey readiness audits, suggesting that these are effective in monitoring behavior. Another process that was considered effective is the Stop the Line program, which urges staff members to speak up in situations where they feel that patient safety is at risk. Having more hand sanitizer dispensers in hallways as opposed to primarily in patient rooms was also proposed as a solution to improve hand hygiene compliance.