Sticky apps, not sticky hands: A systematic review and content synthesis of hand hygiene mobile apps

Background and significance

According to the World Health Organization,¹ good hand hygiene is one of the most simple but effective means of preventing the transmission of harmful microbes and acquirement of healthcare-associated infections. The severe acute respiratory syndrome coronavirus 2, causing the COVID-19 (coronavirus disease 2019) pandemic, further reenforces the need for good hand hygiene practice, both within healthcare settings and in everyday life.²

Daily good hand hygiene practice should include washing hands with clean water and soap before and after eating, handling raw meat or vegetables, changing a nappy, and treating a wound and after using the toilet and touching animals, and whenever hands feel or look dirty.³ In healthcare settings, the World Health Organization (WHO) recommendations for 5 Moments for Hand Hygiene include hand washing before and after touching a patient, after (risk to) exposure to bodily fluids, before clean or aseptic procedures, and after touching patient surroundings.⁴ The WHO 11 steps is an example of guidance on the correct hand hygiene technique.⁵ This includes wetting hands with water, applying soap, rubbing palms, interlacing fingers and then rubbing palms, cleaning the back of fingers by interlocking fingers, rotational rubbing of thumbs, rinsing hands with water, drying them thoroughly, and using a towel to turn off the faucet. Other techniques (eg, WHO 6 steps) combine or exclude some of the steps (eg, turning off the faucet).

Despite the presence of well-established guidelines, adherence to best practice hand hygiene by the general public and patients is suboptimal. Evans et al⁶ found that 90% of 100 adults did not wash their hands appropriately after handling raw chicken. In addition, microorganisms of fecal origin have been found on the hands of one-quarter of adults in the general population, indicating low compliance of hand hygiene after using the toilet.⁷ Randle et al⁸ observed a 56% and 57% hand hygiene compliance rate for hospitalized patients and their visitors, respectively.

Studies on hand hygiene in health and care staff showed similar room for improvement: home care (community) nurses were found to have an average adherence rate of 45.6%⁹ and nurses working in intensive care units (ICUs) of 52%.¹⁰ A systematic review of literature pertaining to student nurses concluded that hand hygiene knowledge and compliant behavior was low to moderate.¹¹ Another review synthesized 61 studies in ICU settings and found that mean hand hygiene compliance by staff was 59.6%.¹² However, this differed by type of ICU, geographic region, and healthcare workers (nursing staff 43.8% and physicians 32.6%). Erasmus et al¹³ also reported a higher adherence rate for nurses compared with physicians (48% and 32%, respectively) in their review on healthcare workers, with overall median compliance rate being 40%. This may be because doctors have been shown to perceive hand hygiene as important for protecting themselvesbut not for protecting their patients,¹³ as reflected by better compliance after having direct contact with a patient (55.1%) than before (38.4%).¹⁴

While barriers to hand hygiene in general adults have not been extensively researched, many factors explain why health and care providers do not hand wash appropriately.^15–18 These barriers are related to awareness, familiarity, agreement, self-efficacy, outcome expectancy, the inability to overcome the inertia of previous practice, and external barriers.¹⁹ For example, a lack of knowledge of hand hygiene protocols can be explained as a lack of awareness and familiarity with guidelines. Additionally, skepticism and the belief that gloves are sufficient align with the lack of agreement with best practice guidelines and of outcome expectancy. External barriers may include limited hand hygiene training and training that is too theoretical and passive.^14,20

Interventions aiming to improve hand hygiene compliance in health and care staff and general adults have been evaluated and include monitoring and providing performance feedback, electronic reminders, workshops, educational campaigns, and using cues (eg, visual posters or examining disinfected hands under ultraviolet lights).^18,21–24 These mostly address barriers related to lack of awareness, overcoming inertia of previous practice, and improving self-efficacy.

Interactive and engaging digital interventions may help to overcome some of the barriers to hand hygiene. Primarily, mobile apps can be more affordable than traditional hand hygiene interventions (estimatedto be £146-£3035 when accounting for staff time and resources).²³ When compared with automated monitoring systems,^25,26 smartphone apps can monitor instances within communities and provide feedback on technique. In addition, they do not require significant financial investment and can provide educational material, which is recommended in adjunct to monitoring systems.²⁶ Interventions delivered on mobile devices can also be creative, less time consuming, use multiple behavior change strategies, and be continuously available (eg, at home), so they may reinforce habitual behavior to help overcome engrained previous practice and change perceptions or attitudes (eg, skepticism).²⁷ However, many digital interventions experience poor uptake, and very few are “sticky” (ie, enjoyable to use and engaging). While it is unknown as to what exactly makes technologies sticky,²⁸ some expected characteristics include being intuitive, accessible, engaging, motivational, personalizable, and providing triggers, such as external cues or reminders.^29,30 Smartphones naturally incorporate some of these aspects. For example, mobile apps are easily accessible, making them ideal for behavior change by using techniques such as self-monitoring, goal setting, increasing awareness or familiarity, and improving self-efficacy.

Previous research has already compared which technologies may help to improve hand hygiene compliance³¹ and has identified that mobile apps may be an appropriate new way of addressing suboptimal hand hygiene.^32,33 However, the app market is saturated with health apps, with over 318 000 available in 2018 and an additional 200 added daily.³⁴ Many of these have not been formally evaluated nor developed with experts, creating concern about the extent to which they distribute reliable health information and, in the case of hand hygiene, encourage good practice.

Search strategy and inclusion criteria

While the target audience of this review was intended to be healthcare providers, an initial scoping search highlighted that few apps specified their intended users. App markets do also not enable for categorization of target users, unless they are children. Therefore, we included apps aimed at adults in general.

We identified apps on November 19, 2020, from 2 UK app markets (App Store and Google Play) using the keywords hand hygiene and hand washing. Apps that focused on hand hygiene (hand washing or sanitation) and targeted at adults were included in the review. Those aimed at children, groups or organizations (rather than individuals), and particular health hygiene issues other than hand hygiene (eg, wearing masks to protect against COVID-19) or those not in English were excluded.

App selection

A 2-step screening process was conducted. First, 1 author (N.G.) screened the app markets for relevant apps by looking at their titles and descriptions. Duplicates between the app markets were removed. Second, 3 authors (N.G., S.M.A., A.J.S.S.) downloaded the apps to confirm their eligibility. Two authors reviewed the apps in the Google Play store (N.G. and S.M.A.) and 2 in the App store (N.G. and A.J.S.S.). The iPhone SE (iOS 13.6.1), iPhone 7 (iOS 14.2), Samsung Galaxy J6 (Android v8.0 Oreo), and Samsung Galaxy S20 FE (Android v10.0) devices were used. Some apps required additional devices, so we also used the Xiaomi Mi Smart Band 4 and the Destek V5 virtual reality headset.

Data extraction

Two authors independently assessed the functionality and quality of each app and recorded relevant information on a data extraction sheet in Excel. This included descriptive information, content, technical and functional features, and quality (see Supplementary Table 1 for a summary of the data extraction items).

The descriptive information included the name of the app, the version number, the developer, the market(s) in which the app was available, the cost, whether the app was affiliated with a professional body, the average user rating, and the number of ratings.

We explored content by identifying the purpose of each app and common behavior change techniques it employed. The behavior change techniques included self-monitoring, feedback, goal setting, social support, reward or threat, prompting practice, or personalization.^41–43 These were the most commonly reported behavior change strategies used in mobile apps, as identified by Taj et al⁴⁴ and Dugas et al.⁴⁵ Other content-related information included whether the app mentioned aligning with any best practice hand hygiene guidelines and the comprehensiveness of the content.

To assess the comprehensiveness of the content, we developed common criteria from reputable sources and guidelines on hand hygiene^3–5,46,47 and determined whether or not these were present in each app. Similar methods were used by Lalloo et al⁴⁸ and Grainger et al.³⁷ The 4 criteria were (1) context (eg, when to hand wash as in the WHO 5 Moments for Hand Hygiene, or for the general public after sneezing or using the toilet, and before eating), (2) use of soap or hand sanitizer (when soap is not available), (3) specific overall timing and length of hand wash (eg, 15-60 seconds), and (4) skill (eg, the WHO 11 steps to hand washing).

We also extracted information on technical and functional features. Technical features included whether the app had a privacy strategy (eg, policy, login, 2-factor authentication) and whether it used innovative features. We considered features innovative if they used any of the following technologies: geolocation tracking (eg, GPS), virtual or augmented reality, virtual assistants (eg, Siri), smartwatches (eg, Wear OS or Apple Watch), personal profiles, or gamification (eg, scores, competition or characters). Other information included whether it worked offline or specifically asked the user to work in the background or enable push notifications.

As in some other mobile app reviews,^49,50 functionality was assessed using the IMS Institute for Healthcare Informatics functionality score.⁵¹ It consists of 7 criteria and 4 subcategories (see Table 1), rated 1 if they apply or 0 if they do not apply. We removed the item “evaluate data” because it pertains to whether the entered data can be evaluated by patient or provider, provider or administrator, or patient or caregiver, which was not relevant to the apps in our review. A total score was generated by summing the items, with a maximum score of 10. Reviewers resolved disagreements through discussion.

In line with previous mobile app reviews,^35,37–39 quality was assessed using the 19-item Mobile App Rating Scale (MARS).⁵² Items were rated on a 5-point Likert scale (1 = inadequate, 2 = poor, 3 = acceptable, 4 = good, and 5 = excellent), with items categorized into 4 dimensions: engagement, functionality, aesthetics, and information quality. Each author independently rated each dimension and calculated an overall mean score. A mean score between both reviewers was then calculated. A question in the scale asks, “Has the app being trialed/tested?,” which we answered by searching for empirical published literature that evaluated the app. This included evaluations for effectiveness, acceptability, usability, or satisfaction.

MARS also provided an optional subscale for subjective quality. However, as consistent with previous reviews, this scale was omitted to ensure that the quality assessments were consistent and objective across the reviewers.^35,38,39 Reviewers watched the MARS training video on YouTube⁵³ prior to rating.

Data synthesis

We synthesized the data descriptively. Where applicable, we generated descriptive summary statistics. The highest-scoring apps across the functionality and quality measures were identified and cross-compared, in order to identify the best overall apps.

To assess interrater reliability of the MARS scores, we calculated the intraclass correlation coefficients (ICCs) on all MARS items, using a 2-way mixed-effects, average-measures model with absolute agreement.⁵⁴ For the IMS Institute for Healthcare Informatics functionality score, we calculated Cohen’s kappa, as ratings were categorical (0 = no, 1 = yes). All statistical analyses were conducted using IBM SPSS Statistics (version 23; IBM, Armonk, NY).

Descriptive characteristics

Most of the apps (n = 86 of 90, 96%) were free to download, with 11 offering in-app purchases. The costs for paid apps ranged from £0.89 to £3.99. At the time of this review, 24 apps had been rated by between 1 and 249 users. Ratings ranged from 2.5 to 5 stars (out of 5). Most of the apps (n = 86 of 90, 96%) were not affiliated with a professional health or medical body, and only 4 said that they had been developed in partnership with infection prevention and control (IPC) expert groups, hospitals, universities, or behavioral scientists. Supplementary Table 2 lists and summarizes the included apps.

Hand hygiene content

The purpose of the included apps was to educate, instruct, or inform users on hand hygiene (eg, importance or skill) (n = 70 of 90, 78%), remind them to practice hand hygiene (n = 62 of 90, 69%), track their hand washing instances (eg, through a log) (n = 27 of 90, 30%), provide feedback (eg, on their performance or technique) (4 of 90, 4%), or convince them (through hypnosis) (n = 1 of 90, 1%).

Apps used a range of established behavior change techniques, with an average of 2 techniques per app (range, 0-5). Thirteen (14%) did not use any of the behavior change techniques, 8 (9%) used 1 technique, 45 (50%) used 2 techniques, 13 (14%) used 3 techniques, 8 (9%) used 4 techniques, and 3 apps (3%) used 5 techniques. Figure 2 presents the frequency of each technique.

Across the apps, the most commonly used technique was personalization (n = 70 of 90, 78%) whereby users received personalized feedback or could change settings within the app. This included setting their reminder times, frequencies, sounds or music, notifications, or setting the language and location. The second most common technique was prompting hand hygiene practice (n = 62 of 90, 69%), as evident in all of the reminder apps.

Self-monitoring (n = 28 of 90, 31%) was often used in the apps that intended to track hand washing instances or performance, and often included presenting a log or hand hygiene history that enabled users to monitor their progress. Some (n = 8 of 90, 9%) of the apps also used goal-setting strategies such as by calculating the number of hand washing instances that should be achieved within a specified timeframe. This goal was then presented to the user and they could monitor their instances against it.

Feedback (n = 5 of 90, 6%) was used less often but included giving users a summary of their performance (eg, in games) or of their technique. Similarly, reward or threat were less commonly used (n = 9 of 90, 10%) but were usually employed in games through scores (as reward) or loss of points or “lives/hearts” (as a threat). Reward was also used when goals were met, such as by enabling the user to move up a level or change a function in the app. Very few (n = 2 of 90, 2%) of the apps facilitated social support (eg, encouraging users to share their data or progress to social media).

Although many of the apps intended to educate or inform users, not many drew on best practice hand hygiene guidelines. Specifically, only 20% (n = 18 of 90) named a guideline. Steps for how to practice hand hygiene were named the most (n = 15), compared with guides on when to hand wash (n = 7) or general guidelines for IPC (n = 2). Guidelines for practicing hand hygiene included the WHO 6, 10, or 11 Steps to Hand Washing (n = 8); Centers for Disease Control and Prevention (CDC) guides for hand washing (n = 6); or the NHS 11 Steps to Hand Wash (n = 1). Guidelines for when to hand wash mostly included the WHO 5 Moments for Hand Hygiene (n = 5), a general WHO guide for when to hand wash (n = 1), and a CDC guide for when to hand wash (n = 1). The 2 IPC guidelines named were the CDC guide for Infection Control and the WHO general guideline for preventing the transmission of diseases.

Four criteria were used to assess the comprehensiveness of the hand hygiene content. Of the included apps, 14 (16%) did not mention any criteria, 26 (29%) mentioned 1 criterion, 9 (10%) apps mentioned 2 criteria, 28 (31%) mentioned 3 criteria, and 13 (14%) mentioned all 4 criteria. Length of practice was included the most (n = 63 of 90, 70%), followed by use of soap or sanitizer (n = 54 of 90, 60%), skill (n = 47 of 90, 52%), and context (n = 15 of 90, 17%).

Technical features

Of the apps, 83% (n = 75 of 90) worked offline and 7% (n = 6 of 90) asked the user whether it could work in the background. Although 62 of the apps sent reminder notifications to users, 47 (78%) of these asked users to enable push notifications. Some included a privacy policy (n = 19 of 90, 21%) or a login of password (n = 3 of 90, 3%). However, none offered 2-factor authentication and most (n = 70 of 90, 78%) did not include any privacy strategies.

Innovative means of engagement were used in 42% (n = 38 of 90) of the apps, including geolocation or movement-based reminders (n = 13 of 90, 14%), gamification (n = 10 of 90, 11%), personal profiles (n = 10 of 90, 11%), interfaced with smartwatches (n = 5 of 90, 6%), virtual assistants (eg, Siri) (n = 4 of 90, 4%), augmented reality (n = 3 of 90, 3%), and virtual reality (n = 1 of 90, 1%). Table 3 provides examples of how apps implemented these features.

Functionality

Functionality was assessed using 10 items from the IMS Institute for Healthcare Informatics functionality score. There was substantial agreement between the 2 raters’ scores of the Apple App Store apps (κ = .76; 95% confidence interval, .67-.85; P < .0000). Similarly, there was substantial agreement between the 2 raters in the Google Play Store (κ = .80; 95% confidence interval, .74-.86; P < .0000). Absolute agreement was reached through discussion.

Figure 3 shows that the most common functions were the ability to inform users on hand hygiene (n = 62 of 90, 69%), provide reminders or alerts (eg, push notifications) (n = 61 of 90, 68%), and record data (eg, hand washing scores) (n = 34 of 90, 38%). Overall, apps had an average of 2.4 (out of 10) functions (range, 0-8) (see Supplementary Table 2). Only 1 app (SureWash Hand Hygiene Pro) had 8 functions, followed by 3 apps (Tork VR Clean Hands Training, SureWash Hand Hygiene, and Remember Washing Your Hands—Handwashing Reminder) that had 7 functions. Three apps (Wash Stuff, Hand Wash Timer—Wash Your Hands For 20 Seconds, and Wash Hands Timer) had none of the functions. These apps tended to be simple hand wash timer or countdown apps.

Quality

All apps were independently evaluated by 2 raters using MARS. The authors who reviewed the apps in the App Store had good interrater reliability (ICC = .88; 95% confidence interval, .86-.90). Interrater reliability between the authors who reviewed the apps in the Google Play market was also good (ICC = .88; 95% confidence interval, .87-.90).

The mean overall quality score across the apps was 3.2 ± 0.5). Only 68% (n = 61 of 90) met the minimum acceptability score of 3.⁵² The best apps were rated 4.6 (SureWash Hand Hygiene; SureWash Hand Hygiene Pro) and the worst rated was 2.4 (Hand Wash Timer—Wash Your Hands for 20 Seconds). Mean scores for engagement, functionality, aesthetics, and dimensions were 2.7 ± 0.7, 4.2 ± 0.5, 3.3 ± 0.7, and 2.6 ± 0.6, respectively. Most apps (n = 84 of 90, 93%) scored highest on the functionality dimension and lowest on information quality (n = 50 of 90, 56%). Only 2 of the apps (SureWash Hand Hygiene and SureWash Hand Hygiene Pro) had been formally trialed or tested,^55,56 but not in robust randomized controlled trial experiments. We reviewed the SureWash apps individually, as they are 2 different versions. Supplementary Table 3 summarizes the MARS scores for each app.

Overall app quality

The apps that had the highest IMS functionality and MARS scores were cross-compared. The highest-performing apps and their defining characteristics are highlighted in Table 4 and include SureWash Hand Hygiene Pro, SureWash Hand Hygiene, Tork VR Clean Hands Training, and Remember Washing Your Hands—Handwashing Reminder. Figure 4 provides images of each of these apps.

Design and research recommendations

Research is warranted to assess the feasibility and effectiveness of hand hygiene apps (especially within healthcare settings), including when and how they “work” in improving self-efficacy, knowledge of guidelines, and compliance to best practice. This should be considered before additional effort is devoted to producing better hand hygiene apps. In addition, robust research and development of theory is needed to determine which innovative methods of engagement create “sticky” apps. This includes in-app methods (eg, gamification or augmented reality), external phone-based tools (eg, virtual assistants, geolocation or movement tracking), or additional external tools (eg, virtual reality headsets or smartwatches, such as in Samyoun et al).⁶² Future research should explore the impact of these methods on engagement, as well as what level and duration of engagement leads to changed hand hygiene behavior or increased awareness of the correct technique. This may differ between types of apps. For example, compared with educational apps, users may need to engage with reminder apps more frequently and for longer to change their behavior.

Currently, there are few apps that contain hand hygiene content that aligns with recommended best practice. There is therefore great opportunity to include IPC experts in developing apps, something that only few apps in this review mentioned. Collaboration between IPC experts and app developers will likely yield apps that contain high-quality, accurate, and comprehensive content. At the same time (or when including experts is not practical), the content may be derived from best practice guidelines. In addition to education, reminding and feedback, future apps should also target attitude, intention, and self-efficacy, to promote effective and sustained behavior change. These should use an adaptive “Just-in-Time” approach that accounts for time and context.⁶³ For example, providing reminders when individuals reach a high-risk area (eg, workplace) or enabling users to postpone or disregard notifications to avoid intervention fatigue.

Strengths and limitations

Unlike other app reviews,^32,36 we included apps that required additional devices (eg, smartwatches) or used innovative features (eg, virtual assistants). This enabled us to also explore innovative methods of engagement, even though we had to exclude some apps that depended on additional devices that were not compatible with the ones we had available. Another strength included using the IMS Institute for Healthcare Informatics functionality score, unlike most other app reviews. While the MARS⁵² focused on navigation, performance, gestural design, and ease of use, the IMS Institute for Healthcare Informatics functionality score⁵¹ focused on the broader scope of function in regard to abilities. This includes the ability to inform, instruct, record, display, guide, remind, and communicate information.

The main limitation of the study was that only apps available in the United Kingdom were included. This means that it is possible that hand hygiene apps developed and released privately or by researchers were not included. Apps reported in academic literature (but not currently available in app stores), released overseas, or in languagesother than English were also excluded.