Health apps often score badly

A health app has nine features. If apps score well on these features, they provide reliable information about conditions and how to deal with them. In that case, doctors and other professionals can recommend them to their patients, while the costs can be included in the Dutch Health Insurance Act. However, a recent study in Health Affairs reveals that even highly regarded apps score poorly on these nine features.

The nine features of a health app are presented in a concise fashion below:

  1. A description of the group of patients for whom the app is intended.

  2. The functionalities of the app: only education or also support of behaviour change with the help of virtual rewards? Access to a digital patient platform to exchange experiences?
  3. The possibility for users to rate the app. This usually happens with a five star rating system.

  4. Added medical value: would a doctor recommend the app to a friend or colleague?
  5. User-friendliness of the app, for example translation of medical jargon into laymans terms.
  6. Issuing warnings of imminent danger when a user enters specific information. For example, when a patient with diabetes enters very low glucose values, does the app respond with the correct, urgent advice?
  7. The presence of a privacy policy that indicates what the website does with the data it collects from its users.
  8. The possibility of sharing data through a safe connection with, for example, ones own GP.
  9. The costs of the use of the app: payment for downloading, annual subscription fee, purchase of additional software within the app?

A research group from Boston, affiliated with the Harvard Medical School, used these nine features to assess the quality of health apps. In December 2016, they published about it in Health Affairs,¹ a leading American magazine.


How does their study work? The researchers collected apps that had already received high ratings from experts, professional magazines, patient organisations or medical professional organisations. They were only interested in apps for patient groups with serious conditions and a high level of healthcare consumption. Their search ultimately yielded 137 apps: for diabetes (25), hypertension (24), obesity (19), depression (17), stroke (14), tobacco, alcohol or drug addiction (9), all conditions simultaneously (11) and other conditions (18). In a digital attachment, they mention the names and scores of most of these apps. The authors note that there are only very few rated apps for frequently occurring conditions such as arthritis (4) and cancer (8).


Of all apps, 63 are available for both I-phones and I-pads and Android smart phones. Most apps were able to record and archive information entered by users. Most of them provided information and were able to send their users reminders and alerts. Very few apps were able to give advice based on the information given by users. They also rarely supported user networks and rarely rewarded behaviour change with ratings or compliments. The user ratings (feature 3) of these apps were entirely out of line with the researchers assessment of their added medical value and user-friendliness.

Health danger

Of the 121 apps that were able to save information given by users, only 28 gave the right response in times of health danger (feature 6). Of all these apps, 88 had a privacy policy as described in feature 7. The following emerged about the security of these apps: 66 apps could only share information through regular email and 17 through text message. The remaining 54 apps offered a secure option. Finally, almost all of the 137 apps were free in all respects (feature 9).

Low score

The researchers note that even these 137, highly rated, apps score low on the nine mentioned features. The article is concluded by James Madara, the general director of the American Medical Association. The AMA is comparable to the Dutch KNMG. He concludes that many apps fail to make good on their promises.


The researchers are against additional regulations, which already exist for medical equipment and medication. That would be too complicated. They do plead for product labelling, similar to the way food products are labelled, which indicates exactly what the features of a health app are. They suggest using the nine features of the study as a starting point.

The Netherlands

This concludes the discussion of this article. By way of commentary, I would like to add that the study is cheap to conduct. It is worth repeating, for example by Dutch patient organisations or as part of the annual e-health monitor carried out by Nictiz and the NIVEL. In addition, the aforementioned list of apps includes a few very good ones. These can easily be translated into Dutch, used in our country and recommended by doctors.

¹ Singh K, K. Drouin, LP Newmark, Many mobile health apps target high needs, high cost populations, but gaps remain, Health Affairs, 35, No 12 (2016): 2310 -2318

Leave a Reply

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>