Characterization of Apps and Other e-Tools for Medication Use: Insights Into Possible Benefits and Risks

Background In the past years, an enormous increase in the number of available health-related applications (apps) has occurred, from approximately 5800 in 2011 to over 23,000 in 2013, in the iTunes store. However, little is still known regarding the use, possible effectiveness, and risks of these applications. In this study, we focused on apps and other e-tools related to medicine use. A large subset of the general population uses medicines and might benefit from tools that aid in the use of medicine. Objective The aim of the present study was to gain more insight into the characteristics, possible risks, and possible benefits of health apps and e-tools related to medication use. Methods We first made an inventory of apps and other e-tools for medication use (n=116). Tools were coded by two independent researchers, based on the information available in the app stores and websites. Subsequently, for one type of often downloaded apps (aimed at people with diabetes), we investigated users’ experiences using an online questionnaire. Results Results of the inventory show that many apps for medication use are available and that they mainly offer simple functionalities. In line with this, the most experienced benefit by users of apps for regulating blood glucose levels in the online questionnaire was “information quick and conveniently available”. Other often experienced benefits were improving health and self-reliance. Results of the inventory show that a minority of the apps for medication use has potentially high risks and for many of the apps it is unclear whether and how personal data are stored. In contrast, online questionnaire among users of apps for blood glucose regulation indicates that they hardly ever experience problems or doubts considering reliability and/or privacy. Although, respondents do mention to experience disadvantages of use due to incomplete apps and apps with poor ease of use. Respondents not using app(s) indicate that they might use them in the future if reliability of the apps and instructions on how to use them are more clear. Conclusions This study shows that for apps and e-tools related to medicine use a small subset of tools might involve relatively high risks. For the large group of nonmedical devices apps, risks are lower, but risks lie in the enormous availability and low levels of regulation. In addition, both users and nonusers indicated that overall quality of apps (ease of use, completeness, good functionalities) is an issue. Considering that important benefits (eg, improving health and self-reliance) are experienced by many of the respondents using apps for regulating blood glucose levels, improving reliability and quality of apps is likely to have many profits. In addition, creating better awareness regarding the existence and how to use apps will likely improve proper use by more people, enhancing the profits of these tools.


Supplementary data B: Additional information coding of selected apps for medication use Explanation and criteria of codes
All apps were individually assessed for all the codes described below. For some codes, groups of apps are described for which a coding was applicable. For example, when certain functionality was always coded with a certain benefit, but apps were always assessed individually. During the coding process 14 apps were excluded from the original selection because too little information was provided in the app stores, resulting in a total of 116 coded tools. Codes were performed by L. van Kerkhof (PhD, Biomedical Sciences) and C. de Jong (Msc, European Studies) and supervised by I. Hegger (PharmD).

Characteristics
Functionalities: Codes were based on pilot searches and pilot codes.
Users: Codes were based on Nictiz Whitepaper' Orde in the world of eHealth' [18]. Tools were coded with 'patient and healthcare professional' when the interaction between both was clearly an intended use of the tool. For example, when it was clearly described that the tool should be used to send information to you healthcare professional. If this was not clearly stated tools were coded as 'intended user -patient' Technical domain: Codes were based on Nictiz Whitepaper 'Order in the world of eHealth' [18].
Downloads: Downloads were obtained from the information provided with the apps from the Google Play store. For apps from the iTunes Store download numbers were not available.

Benefits
Self-reliance: Does the tool potentially enhance a patient's self-reliance? Apps that are able to make a patient more independent of a health care professional or others. Apps with the functionality of 'improving therapy adherence' and 'monitoring effects or side effects' were coded with 'self-reliance -yes'.
Improving health: Does the tool potentially improve health? Examples of apps coded with 'improve health -yes' are apps with the functionality of 'improving therapy adherence' or 'monitoring effects or side effects'. Apps that only provide information are not considered to improve health in this question, nor are dosage calculators for healthcare professionals.
Lowering health care costs: Has the tool the potential to lower health care costs? Tools coded with functionality 'improving therapy adherence' 'or the benefit of 'enhancing self-reliance' are considered to potentially lower health care costs.
Improving self-management: Does the tool potentially improve self-management? Self-management was considered when the tool helped users actively participate in the decision-making of their own therapy. This mainly included, but not exclusively, apps with the functionality of 'monitoring effects or side effects'

Risks
Medical device: Coding of medical device and class was based on EU regulations [7] and the Nictiz Whitepaper: 'Medical apps, is CE-mark required?' [23]. Apps aimed at monitoring diabetes were only coded as class II medical device when it was clearly stated that calculations (medication dose) were made by the app.
Data upload: Codes regarding data upload were coded with 'no' or 'yes' if this was specifically mentioned in the tools description. In addition, 'yes' was coded when it was clear that data was transferred when using the tool in a normal way.
Healthcare professional involved in obtaining the tool or during use of the tool: Tools were coded with 'yes' on this question when it was clearly stated that a healthcare professional should be involved or when a code was required to obtain or use the tool.
Replacement of health care professional: Is the goal of the tool to (partially) replace a health care professional or can this happen by accident? Tools were coded with 'yes' on this question when tools could replace healthcare professionals. Specifically, tools with the functionality of only 'providing information' were not coded with 'yes'.
Can incorrect use or incorrect design of the tool result into decisions with a large impact on the users health? Four different codes were used for this question: 'yes and realistic', 'yes, but not realistic', 'no' and 'not assessable'. The first two codes were used to distinguish between theoretical risks that might occur and an estimation of these risks actually happening. For example, when a healthcare professional is involved risks were considered less likely to happen considering the education of healthcare professionals, while when only patients are involved risks might be more likely to occur. What are disadvantages that you experience by using apps for regulating blood glucose levels?

Open question
To what extend do you experience the following risks by the apps that you use? Open question   indicated to use these apps and have answered this question (n = 28). This question was not answered by 9 respondents and 39 respondents indicated not to use these apps.