Medication management support in diabetes: a systematic assessment of diabetes self-management apps

Medication adherence, broadly understood as the act of taking medicines as prescribed by the healthcare provider, is important for achieving treatment goals [1]. This is paramount for chronic conditions such as diabetes. However, studies have shown that approximately 33% of oral medications and 38% of insulin for type 2 diabetes (T2D) are not taken/used as prescribed [2, 3] due to forgetfulness, inconvenience, negative treatment beliefs, fear of injections and a myriad of other personal and health system factors [4].

Medication management strategies have been developed and implemented to assist people in adhering to their medications. These strategies include education on disease management, simplification of dosing regimen, counselling, reminders, or a combination of these methods [5, 6]. Digital solutions have also been studied in the past 20 years to assist in medication adherence. Although research has shown that mobile text messaging can double the odds of medication adherence in chronic diseases [7], more successful interventions often involved the use of two-way communication [8] and were tailored to individual needs [9, 10]. This suggests the need for innovation and a combination of measures that go beyond basic reminders to improve medication adherence.

Smartphone apps have gained popularity in diabetes self-management in recent years. Compared to SMS reminders, smartphone apps have the advantage of performing more sophisticated medication management functions such as pill organisation, tracking of medication-taking, information provision, and adherence assessment [6]. With the rise in the number of smartphone users [11, 12] and integration of smartphone apps in daily living [13], a myriad of apps were developed to assist people in adhering to their medications. Despite the large number (approximately 400) of accessible and free apps for medication self-management in the app market in recent years [14], the majority of these apps lacked useful, desirable features for medication adherence [15]. According to national digital health consumer surveys, only 11% of respondents tracking health goals tracked their medications [16]. Medication adherence is also least likely to be tracked in an app (10%) amongst other trackable health-related metrics like physical activity, heart rate and sleep [17].

Currently, the large number of available diabetes management apps provides an opportunity to support medication management, but also represents missed opportunities to improve care for people with diabetes with gaps that fall short of users’ needs [18]. It is unclear if diabetes apps are adequately incorporating medication management strategies and if app features are aligned with best-practice evidence-based recommendations [19]. Improvements in app quality and utility can only be realised if gaps in app features are identified. We constructed a diagram linking good medication management practice with possible app features and systematically assessed and characterised the medication management features in available apps for T2D self-management. We discussed the implications of our findings in relation to diabetes management and provided suggestions to address the identified gaps.

The search terms yielded 3369 Android and 1799 iOS apps. After title and description screening, 191 (5.7%) Android and 160 (8.9%) iOS apps remained for downloading. Apps were further excluded due to their unavailability, exclusive device tie-in requirement, technical issues, not meeting the inclusion criteria (i.e. non-English), and if they were duplicated on the same platform. Restricted apps that received no reply from developers were also excluded from the study. Finally, 143 apps (81 Android, 62 iOS) were downloaded and assessed against the app assessment criteria (Fig. 2).

The search terms “(Diabetes OR Diabetic OR Diabetics) OR (glucose OR glycaemic OR glycemic OR blood sugar OR HbA1c OR A1c) OR insulin” yielded 4876 results from (https://​42matters.​com). After screening the app descriptions for relevance, 351 apps were downloaded for assessment; of which, 208 apps were excluded due to (i) technical issues, (ii) device tie-in, (iii) not available at app assessment, (iv) regional restriction, (v) do not meet inclusion criteria on second screening, and (vi) duplication of apps on the same platform. Finally, 143 apps (81 Android; 62 iOS) were fully assessed against the app assessment criteria in Table 1.

We identified, downloaded and systematically evaluated 143 apps against assessment criteria derived from international medication management guidelines. There were few differences in app features between Android and iOS apps except for a higher proportion of iOS apps having medication-intake documentation and data export features. Most of the assessed apps, including apps with higher downloads (≥ 100,000), have basic logging and tracking features for diabetes medication, but lacked features that could enhance medication adherence and safety. We identified the following gaps in the assessed apps. First, many diabetes apps lacked any form of medication management features, which concurs with a 2017 study which found that only 50% of the highest rated iOS diabetes management apps had medication adherence features [33]. A separate study by our team which employed a similar search strategy also found that only 43% of the accessible diabetes self-management apps had medication management features [34]. This may be attributed to a lack of emphasis given to medication adherence in diabetes management. Amongst the assessed apps with medication management features, a large proportion did not have important features such as a basic reminder feature, the capability to enter medication-taking instructions and medication adherence review. Apps devoid of essential features for enhancing medication adherence are less likely to be useful in helping users adhere to their medications.

Second, less than 10% of the apps provided any information on diabetes medication or allowed the user to communicate with a healthcare professional. Although this feature is more important for users needing to adjust to a new medication therapy, having information on medication will be beneficial as diabetes is likely to progress over time. Third, only two iOS apps prompted the user if complementary medicine was being used. The use of complementary medicine is common in many cultures and can lead to contraindications [35]. Stopping conventional medication in favour of complementary medicine can also lead to ineffective or adverse treatment outcomes. It would be important for app developers to include a cautionary message or features to alert users to potential contraindications. Documentation of allergies is also important to flag possible medication or food-related contraindication but only 3.5% of the assessed apps have this capability.

Fourth, less than 5% of the apps had features that provided any form of motivation to the user. A few apps that encouraged medication adherence had interactive features that could possibly increase the time spent on the app. Sustained app use may increase medication-taking awareness of users who may otherwise not remember to take their medications. Lastly, about 40% of the apps do not allow data export, which can assist the individual or a healthcare provider to review treatment plans and goals.

An explanation for the lack of evidence-based features in health apps may be in the absence of healthcare providers’ involvement in the development of the app [19]. Only 13.6% of the apps for medication adherence were developed with the involvement of a healthcare provider and only 1% of the apps were evidence-based according to another study [14]. Intermittent app use or app intervention failure may sometimes be caused by a lack of useful app features rather than apps being ineffective per se. Many medication adherence app intervention studies focused on reminding the user to take their medications, but the quality of reminder features and alignment with evidence-based recommendations were unclear [36, 37]. Assessing the app against a medication management checklist (as illustrated in Fig. 1) before the intervention will better align the app for its purpose.

While our study focused on users with T2D, 86% of the assessed apps allow users to log and track insulin doses and hence can also be used by people with type 1 diabetes (T1D). However, adherence to insulin is more challenging than to oral medications due to barriers such as fear of injections, embarrassment of injecting in public, concerns over cost and side effects such as hypoglycaemia) [3, 38, 39]. These barriers cannot be overcome solely by the use of a medication management app, although apps could potentially support adherence to insulin therapy via incorporating patient education.

Strengths of this study include that our app assessment criteria were developed referring to evidence-based guidelines and covered a broader scope on medication management compared to other studies [6, 15]. We also assessed free apps and apps requiring payment which were not limited to one country’s app store.

Despite attempts to minimise bias, there were limitations to the study. We were unable to cover the entire spectrum of medication management apps. Instead, we chose to focus on diabetes due to its prevalence and the need for long-term medication management. The assessment may also not reflect the current state of the apps due to constant updates. However, we attempted to cover all apps at a particular time point and believe that our findings remain unchanged as a previous study showed that the quality of apps in terms of alignment with evidence-based guidelines did not improve within a 2-year period [40]. The app assessment criteria for this study were selected based on their perceived usefulness to people with T2D requiring long-term medication management. Other criteria such as focusing on shared decision-making when medications are not taken as intended could be derived for future assessments. Additionally, we did not investigate the app’s ability to flag medication contraindications nor assessed the content of the medication information provided in these apps. Although the app assessment criteria were developed from the perspective of chronic disease management, we believe we covered the app features important for medication management in diabetes. Since diabetes self-management requires additional self-care activities such as blood glucose monitoring, physical activity and diet modifications, other features of the app should be considered in assessing the overall quality of the app.

A minimum standard (i.e. certification or selection of apps using an evidence-based checklist) could be one way to raise the standard of medication management features of apps. The implementation of the NHS digital library and the FDA “precertification” programme for mobile apps are precedents of tools to objectively evaluate apps, although gaps still exist in the app marketplace in meeting patients’ and clinicians’ needs [41]. We also believe that app stores should play a greater role in quality assurance of health/medical apps. In addition to certification, health app developers should take active steps to ensure that their apps meet minimum standards by co-designing apps with potential users and by continuing to upgrade their app. Healthcare providers can take a more active role in participating in app co-design and work with their patients to effectively use an app to manage chronic conditions. Researchers planning for medication adherence app intervention studies should also be aware of the shortcomings of current apps when evaluating the effectiveness of these apps in improving medication adherence.

The list of assessment criteria is non-exhaustive and should be tailored to patient needs and advancements in technology. Given that these assessment criteria may be applicable to medication management of other chronic diseases, future studies can explore apps specific to other chronic diseases to determine if similar gaps exist. Studies can also explore the usability of these apps for better patient experience, and the efficacy of the medication management features in improving medication adherence in different settings.

Our systematic, broad and evidence-based assessment of smartphone apps provides an overview of medication management features of diabetes self-management apps. A large proportion of the apps lacked features that were useful for enhancing medication adherence and safety, such as the capability to enter allergies and medication-taking instructions, functional reminders, information provision and prompts for the usage of complementary medicine. These gaps represent missed opportunities for better app features which can potentially enhance digital medication management in people with T2D. More emphasis should be given to the inclusion and design of medication management features in diabetes apps. Healthcare providers, app developers and researchers should be involved in the co-design of health apps in order to improve their quality and be aware of the shortcomings of current apps when making recommendations about their effectiveness.