Application of the Capability, Opportunity, Motivation and Behavior (COM-B) model to identify predictors of two self-reported hand hygiene behaviors (handwashing and hand sanitizer use) to prevent COVID-19 infection among U.S. adults, Fall 2020

Data on multiple demographic characteristics were collected when respondents agreed to participate in the survey panel. Of those, gender, age, income, education, race/ethnicity, and geographic region (based on U.S. Census regions) were included in our analyses as statistical controls.

To assess hand hygiene precautions taken in response to COVID-19, respondents were asked, “What, if any, precautions are you taking to prevent coronavirus?”. They were then provided a list of precautions, presented randomly, and asked to endorse those they engaged in. We present data on the following hygiene-related behaviors from the list of precautions: washing hands often with soap and water and using alcohol-based hand sanitizer (hereafter referred to as handwashing and hand sanitizer use) (Table 1).

To assess the COM-B constructs of capability, opportunity and motivation to engage in hand hygiene behaviors to prevent coronavirus infection, respondents were presented with a series of statements about each hand hygiene behavior and asked whether they agreed with the statements. Items were developed based on the definitions of the COM-B constructs, example hand hygiene items from The Behaviour Change Wheel: A Guide to Designing Interventions [34], and from the COM-B Hand Hygiene Behavior Questionnaire [35]. For handwashing and hand sanitizer use, respondents responded to six statements about capability (e.g., I am capable of frequently washing my hands/using hand sanitizer to prevent coronavirus), four statements about opportunity (e.g., Frequent handwashing/frequently using hand sanitizer is a habit in my everyday life), and four statements about motivation (e.g., Frequent handwashing/Frequent use of hand sanitizer is a good way to protect myself from coronavirus) (Table 2).

We used results from exploratory factor analysis and Cronbach’s alpha to produce internally consistent, single factor COM-B constructs for the two hand hygiene behaviors. Based on the results of the analyses, we created a capability composite measure using five of the six capability items, an opportunity composite measure using two of the four opportunity items, and a motivation composite measure using all four motivation items (Table 2). Standardized Cronbach’s alpha are located in Table 2 and were as follows: Capability (Handwashing α = 0.82; Hand Sanitizer α = 0.82), Opportunity (Handwashing α = 0.74; Hand Sanitizer α = 0.70), Motivation (Handwashing α = 0.79; Hand Sanitizer α = 0.83). To create the composite measures, we summed the number of items respondents endorsed in each group of items. For example, if a respondent endorsed 2 of the 4 motivation items for handwashing, their score would be 2. The items that did not fit into the composite measures were treated individually. To aid interpretability, several items were reverse scored (Table 2).

We also assessed other aspects of respondents’ motivation including perceived susceptibility to illness, perceived severity of illness, and perceived behavioral control. To assess these constructs, respondents were presented with a series of statements and asked if they agreed with them. One statement focused on perceived susceptibility (I am concerned about my own risk of infection with coronavirus), two focused on perceived severity (e.g., Getting sick with coronavirus can be life threatening), and three focused on perceived behavioral control (e.g., It is impossible to prevent coronavirus infection [reversed]). Exploratory factor analyses of the items indicated that these items should be treated individually. For interpretability’s sake, several items were reverse scored (Table 3).

We calculated weighted frequencies and means for the hand hygiene behaviors (Table 1), as well as measures of capability, opportunity, and motivation (Table 2). We then conducted a series of weighted multiple logistic regressions using backward variable selection based on Schwartz Bayesian criterion to create models for the hand hygiene behaviors. We used the capability, opportunity, and motivation measures as explanatory variables and controlled for six demographic variables often associated with hygiene behaviors (age, gender, race/ethnicity, education, income, region, Table 1) [7, 10, 11, 16]. Results are presented as adjusted odds ratios along with the 95% confidence intervals and probability values. All analyses were conducted using SAS 9.4. This activity was reviewed by the U.S. Centers for Disease Control and Prevention (CDC) and was conducted consistent with applicable federal law and CDC policy.¹

This study identified how components of the COM-B Model (capability, opportunity, motivation) were related to hand hygiene behaviors. The COM-B model states that individuals must have both the psychological and physical capability to engage in a specific behavior [32]. This includes knowledge and understanding of the behavior and physical skill to do the behavior. Although high proportions of respondents agreed with the capability statements concerning handwashing and hand sanitizer use, there is still room for improvement. For example, 24% of respondents reported they were not capable of washing hands frequently, and 35% reported they were not capable of using hand sanitizer. This study was not designed to assess factors that impact capability to engage in hand hygiene behaviors; future work should do so. Additionally, 33% and 38% of respondents did not know when they should wash their hands or use hand sanitizer, respectively, and 26% and 36% of respondents did not know how to adequately wash their hands or use hand sanitizer. Both of these behaviors are not novel, and U.S. adults were already engaging in them during the COVID-19 pandemic [7‐15]. Additionally, there is evidence that levels of knowledge about the use of handwashing as a prevention strategy is high among U.S. adults [34, 38]. However, the present study’s results suggest that health communication efforts may need to not only provide information about the importance of hand hygiene behaviors, but also provide information about when to engage in these behaviors, and the steps needed to correctly engage in them.

The composite measure of capability was associated with both hand hygiene behaviors assessed in this study, although the relationship was stronger for hand sanitizer use than for handwashing. The finding of association between capability and hand hygiene is consistent with studies conducted both before and during the pandemic. In a pre-COVID-19 systematic review, White et al. (2020) found that knowledge of when to wash hands and having access to supplies were positively associated with handwashing [39]. A study using the COM-B model to identify predictors of a suite of hygiene behaviors during the COVID-19 pandemic also found a relationship between capability and engaging in hygiene behaviors [22]. These results suggest that efforts to promote hand hygiene during the COVID-19 pandemic among U.S. adults should consider addressing aspects of individual capability to engage in hand hygiene behaviors.

Opportunity in the COM-B model refers to the physical and social opportunities for individuals to engage in a specific behavior and includes factors such as having time or resources to do the behavior, cues to engage in the behavior, and interpersonal influencers such as social cues and norms [32]. Our data suggest that respondents had the physical opportunity of time to engage in hand hygiene; over 95% reported that lack of time did not make it hard to engage in frequent handwashing and hand sanitizer use. In contrast, however, most (60%) respondents said that visual cues (visual cues and seeing other people engage in hand hygiene) did not remind them to engage in hand hygiene behaviors. Indeed, the composite variable for opportunity, which included two items assessing visual cues, had a negative association with handwashing and was not associated with hand sanitizer use. This finding contrasts with work conducted before the COVID-19 pandemic that found that being cued by handwashing facilities tended to be positively associated with handwashing [39]. It is unclear why, in this study, visual cues were associated with respondents being less likely to report frequent handwashing behavior; additional studies are needed to further understand the relationship between cues and handwashing behaviors, including studies that assess frequency of observed handwashing behavior rather than self-reported behavior. Respondents’ perceptions of frequent handwashing and hand sanitizer use as habits in everyday life was the only opportunity measure associated with respondents’ self-reported handwashing and hand sanitizer use behaviors. The finding of positive associations between habits and hand hygiene is consistent with findings from other studies and suggests that supporting the formation of hand hygiene habits may be useful in efforts to promote hand hygiene during the COVID-19 pandemic (particularly so, given the relatively low percentages of respondents reporting hand hygiene habits) [40, 41]. Although habit was associated with self-reported hand hygiene behavior, only 38% of our respondents reported that frequent hand sanitizer use was a habit, and less than two-thirds reported that frequent handwashing was a habit in everyday life (62%). To promote hand hygiene behaviors, future communication approaches may want to focus on providing the public with strategies and tips to incorporate hand hygiene behaviors in daily routines. These strategies could include self-monitoring, planning, and behavior cues to foster habit formation [42].

Overall, however, our results suggest that when designing health communication, education, and behavior change interventions to promote hand hygiene behavior, a focus on the construct of opportunity may not be as useful as focusing on other constructs of the COM-B model.

Of the three COM-B constructs evaluated in this study, motivation was most strongly associated with likelihood to report frequent handwashing and hand sanitizer use. The COM-B model posits that an individual must have enough motivation to engage in a particular behavior rather than engage in other behaviors or to not engage in any behaviors at all [32]. Our results support those of Miller et al. (2020) who also found that, overall, motivation was strongly associated with engaging in hygiene-related behaviors during the COVID-19 pandemic, with motivation more strongly associated with hygiene behavior than capability and opportunity [22].

Our data on motivation measures suggest that the majority of U.S. adults are motivated to engage in hand hygiene behaviors—they believe that frequent handwashing is a good way to protect themselves from coronavirus, and to a lesser extent feel this way about hand sanitizer. Perceived efficacy of recommended behaviors and outcome expectancies have been found to be associated with increased hand hygiene behavior during the COVID-19 pandemic [17, 26] and during previous pandemics [43], and with general determinants of hand hygiene behavior [39]. Of note, perceptions of effectiveness of mitigation behaviors have been found to be associated with knowledge about COVID-19 [44], suggesting that efforts to address this aspect of motivation should include a focus on increasing baseline knowledge of COVID-19.

Motivational measures of perceived susceptibility, severity, and behavioral control were associated with one or both of the hand hygiene behaviors. These findings support other research showing significant relationships between these constructs and hand hygiene in the COVID-19 pandemic [8, 10, 23, 26], during other pandemics [43], and in general [39]. Data collected during the COVID-19 pandemic shows significant variability in individuals’ perceptions of susceptibility and severity [8, 10, 23, 26]. Given the multiple factors that influence susceptibility to and severity of COVID-19, particularly before the availability of vaccines, when these data were collected, this variability is perhaps not surprising. However, it does indicate that future efforts to increase prevention behaviors may need to focus on clear communication concerning these concepts. Researchers may also wish to explore factors that influence perceived susceptibility to and severity of COVID-19. Indeed, it would likely be useful to identify factors that influence all the constructs of the COM-B model.

This study is subject to several limitations. First, although the data were weighted to be nationally representative and computers and internet access were provided to respondents who needed them, it is possible that results may not be fully representative of the U.S. adult population (e.g., those uncomfortable with the internet may have been less likely to participate). Second, the items developed for this study were not piloted prior to use. Correspondingly, composite measures were created post hoc and may lack construct validity. Third, the data might be affected by social desirability bias, which might have resulted in overreporting of agreement with statements assessing COM-B constructs and of socially desirable behaviors, such as hand hygiene behaviors. Indeed, self-reported handwashing rates are consistently higher than observed handwashing rates [45]. Finally, data for this study were cross-sectional and as such we cannot make causal statements about the directionality of the relationships between hand hygiene behaviors and COM-B constructs.