Digitizing diagnosis: a review of mobile applications in the diagnostic process

Introduction

The right to diagnose is a powerful tool which has contributed to the authority of the medical profession for many decades. However, the diagnostic process has been confronted by recent changes in the practice of medicine. Doctors are no longer the only professionals who have authority to diagnose, and access to information has led to changes in the diagnostic relationship [1]. From the 1970s, patient consumerist movements have called for lay people to be provided with greater access to medical information. The ideal of the ‘engaged patient’ has recently emerged, charging lay people with the responsibility for taking greater control of their health states and medical care [2, 3]. With these changes, there has also been movement towards self-diagnosis.

Technologies have also played an integral role in shaping and altering diagnostic practices. The very latest of those to enter the field of medical diagnosis are those using digital technologies as part of the move towards digital health provision. These include mobile software applications (‘apps’) designed for smartphones and tablet computers. It is on these devices that this article focuses. Little scholarly attention has been paid to the place of apps in the diagnostic process and what their role has been in changing the social dynamic around, as well as the nature of, diagnosis. Yet apps designed to assist both health care practitioners and members of the lay public have proliferated in recent years. The implications of these new technologies for health care, public health, and health promotion are profound. These devices, their apps and related software, and the online platforms to which they connect, including social media sites, offer not only ready access to medical and health information on the internet but also new ways of monitoring, measuring, visualizing, and experiencing the human body.

Given the novelty of contemporary digital health technologies such as apps, the apparent benefits they promise and their unintended consequences have not been sufficiently investigated [4–7]. We argue that apps should be viewed as sociocultural artefacts and worthy of social analysis [6]. The study here reported seeks to critically analyze apps designed for the medical diagnosis. Following an overview of the role played by technology in medical diagnosis, we describe the number and nature of available apps designed to streamline or putatively improve diagnosis. We then move on to an analysis of their ethical, social and cultural implications. In our focus on digitized diagnosis, we bring together the sociology of medical diagnosis [1, 8, 9] with a critical perspective on digital health technologies [3, 6, 7]. To our knowledge, no other research has been published that has attempted to adopt this approach to analyzing medical apps.

Much has been written about the ways in which medical technologies have transformed, and indeed created, new diagnostic categories [1, 10–12]. Many diseases are now diagnosed by technology rather than by symptom. Today, institutions and individuals commonly use digital tools. Diagnosticians will search for articles on Medline and entries in Google; they may use specialized search engines, electronic journal article summaries, diagnostic software or online calculators.

Medical and health apps have emerged as important contributors to the digital health ecology. The first smartphones able to install third-party apps were released in 2008, followed in 2010 by the development of tablet computers also capable of downloading these apps. Millions of apps have now been published [13] including over 100,000 medical and health apps [14]. These apps include those directed specifically at medical professionals, providing detailed anatomical information and visuals, digital versions of medical textbooks and dictionaries, training videos and diagrams and drug prescribing information. Apps have been designed for this market that can monitor and measure many physical functions and symptoms (such as testing glucose, blood pressure, heart, kidney and lung function and so on), sometimes using hardware add-ons to smartphones that transform them into medical devices. Many more apps are marketed to lay people to provide medical and health information or assist them in self-monitoring their bodily functions and activities. This includes devices that can measure body weight, blood pressure, energy expended, steps taken, sleep patterns and brain activity. Others are directed at assisting patients to store and access their medical records and treatment regimens and track medical appointments, access medical trials, seek appropriate medical care, engage in patient support networks, share information on their condition with friends and family, monitor and self-manage chronic conditions such as diabetes and high blood pressure and to record their symptoms and manage post-treatment care after treatment for acute conditions such as cancer.

Medical practitioners have begun to use apps and other digital technologies as part of their practice [5, 15]. At least one American medical school is requiring its students to use iPads using a variety of apps [16]. Medical and health apps are used by growing numbers of lay people in a world in which smartphone ownership is becoming increasingly common. In late 2012 a Pew Research Center survey found that 85% of adults in the US owned a mobile phone. Fifty-three per cent of these were smartphones, and one fifth of smartphone users had used their phone to download a health-related app. The most popular of these apps were related to monitoring exercise, diet and weight [17]. Another Pew survey found that one in three of the American adults surveyed had reported using online resources to self-diagnose or diagnose another individual [18]. Online diagnostic and risk calculator tools have been available to lay people for some years now, but such tools as smartphone apps are ever-more accessible and ready-to-hand.

These affordances, however, also offer potential risks concerning data privacy and security and the accuracy of the information that apps contain. Apps can offer many benefits for medical professionals and patients and lay people alike. Several writers and researchers in medicine and public health have drawn attention to such benefits [19–23]. However, other commentators have noted that there are numerous significant potential harms that are posed by the use of apps for medical and health-related purposes, including conflict of interest and transparency, ethical, and privacy issues, as well as those relating to accuracy of content, healthcare delivery and the doctor-patient relationship [5, 16, 24–26]. Despite the progressive and rapidly expanding entry of digital health technologies such as apps into healthcare service provision [3, 7, 27], researchers in the health services literature have yet to fully investigate these topics. No social research has yet been published on medical diagnosis apps. A broader view of the range of apps available and their possible role in either the support or hindrance of the diagnostic process is germane to the current growing culture of digital technology in diagnosis.

Methods

To determine what kinds of medical diagnosis apps for smartphones are being offered to health professionals and the public, we undertook a content analysis of such apps that were available in the Google and Apple apps stores in mid-December 2013. We focused on these stores alone, because they are by far the largest providers of apps, well above any other app stores. We searched for all apps using the phrase “medical diagnosis” in their titles or descriptions. To obtain a sample, we reviewed all of the smartphone apps listed under that search term in the Apple App Store (n=88), and then matched the sample by reviewing the same number of apps from Google Play that were listed under this search term (the first 88 listed). This resulted in a total of 176 apps from both stores. We then reviewed each app and excluded those which were not in the English language or focused on veterinary diagnosis, diagnosis based on alternative medicine, or nursing diagnosis (see Table 1). (While the expression “nursing diagnosis” uses the same terminology as medical diagnosis, it is a different process. According to the American Nursing Association, “The nursing diagnosis is the nurse’s clinical judgment about the client’s response to actual or potential health conditions or needs”) [28]. Following this, from an original pool of 176, a total of 131 were retained for analysis. Sixty-four came from the Apple App Store, 59 from Google Play, and eight were found in both. One third of these apps were free to download, with the others available for purchase at prices starting at <1 and going as high as 80 US dollars or more.

We then reviewed the online descriptions provided on the app stores for each of these apps. We sought to summarize the content and approach of each app, but also to iteratively identify common themes or assumptions and thematic orientations of the apps as these became apparent in the sample as a whole. Each app was analyzed for such features as its claims to medical authority and the ways in which medical knowledge, lay people and medical practitioners were represented. We also reviewed material provided by the app developers on their websites (if provided), examining this content in relationship to further claims of medical authority or information about evidence base. Following agreement about how the apps should be analyzed, both authors conducted separate reviews and conferred about their findings. Any disagreements concerning interpretation of the apps content were reconciled by discussion between the authors.

Results and discussion

We began our analysis by identifying topic categories for the medical diagnosis apps among our sample of 131 apps. They are as follows:

1. Effecting diagnoses: 57 apps

2. Medical books, journals, or other publications about diagnosis in app format: 24 apps

3. Diagnosis coding aids: 20 apps

4. Textbook and teaching tools for medical education in diagnosis in app format: 15 apps.

Fifteen apps either did not fit into any of these categories or did not provide enough details to determine their content, and were coded as “Other” (see Table 2).

Of the total sample of apps we analyzed, 36 targeted lay people, 93 targeted professionals, and two targeted both. We determined the target by the narrative provided by the developers, who generally specified the intended audience by using such terms as “helps physicians” or “make a differential diagnosis of your patient” if directed at medical professionals, or “helps you analyze your symptoms” if designed for lay people.

It was evident from reviewing the app descriptions provided on Google Play (Apple App Store does not provide exact download figures) that the popularity of the apps ranged widely. Several of the apps in our sample had been downloaded hundreds of thousands or even millions of times, while others had attracted only few downloads. One mobile health journalist estimated that there could be as many as ten thousand health-related apps specifically targeting medical professionals [29]. While we cannot know from these figures how or even whether downloaded apps are used once placed on users’ smartphones or mobile devices, these suggest that users are attracted to at least some diagnosis apps in high numbers.

Conclusions

Diagnosis apps can bring tremendous benefit to a diagnostician. Particularly in high pressure or in field settings, they enable the clinician to access information that otherwise would stay on the shelves. Furthermore, with the benefit of digital indexing, they make information easier to look up and source. In the case of less common disorders, the app and its indexing potential provide access to diagnostic definitions which might otherwise be hidden in old medical school memories. These affordances can appeal to healthcare professionals as offering a very appealing diagnostic adjunct.

However our research suggests that they should be used with great caution by both lay people and practitioners in the context of evidence-based practice. The lack of verifiable information provided about the evidence or expertise used to develop these apps is of major concern. The apps are of very variable quality, ranging from those that appear to have the support and input of distinguished medical experts, specialty groups or medical societies to those that offer little or nothing to support their knowledge claims. While at one end of the spectrum we can see apps as a delivery system for information which has been subject to the conventional forms of academic review, at the other extreme, we see apps developed by entrepreneurs with interests in many topics outside medicine, with little input from medical sources, or with inadequate information to ascertain what the sources might be. The lack of information provided by many app developers also raises questions about how users can determine the presence of conflicts of interest and commercial interests that might determine content.

The sheer number and constant proliferation of medical apps in general pose difficulties for regulatory agencies to maintain oversight of their quality and accuracy. Such guidelines are important to protect patient safety and prevent psychological harm from erroneous diagnoses. An app developed by a pharmaceutical company or medical device manufacturer can suggest diagnoses which will result in the use of the company’s products, and there is currently no control over or regulation thereof. App developers can promise benefits and outcomes with little possibility of being exposed if the information they provide is incorrect or misleading.

In the face of these issues, it has been contended that the content of all medical apps should be externally peer-reviewed by medical professionals to ensure quality and safety.⁵ The relevant regulatory bodies in many countries have not yet established guidelines and regulations for overseeing medical apps, and are struggling with how to deal with the vast numbers of new apps flooding the market. At the time of writing, NHS England was working on a bilateral framework with the US Food and Drug Administration (FDA), which in September 2013 released a set of guidelines by which medical apps are to be regulated in that country. The guidelines note that many medical apps would not come under its jurisdiction because they are not defined as medical “devices” and would not pose a risk to patients if they do not operate as intended. Under the FDA guidelines, however, medical diagnosis apps fit the definition of medical “devices”: “When the intended use of a mobile app is for the diagnosis of disease or other conditions [emphasis added], or the cure, mitigation, treatment, or prevention of disease, or is intended to affect the structure or any function of the body of a man [sic], the mobile app is a device” [30].

More broadly, medical apps such as those we examined pose challenges to traditional concepts of how medical education should be undertaken and diagnoses made; and for the doctor-patient relationship. The diagnosis is potentially a point of contention between patient and doctor because of the fact that it is so important to validating the patient’s complaint and allocating of resources (sick leave, treatment, and so forth). Alternatively, it can stigmatize, so whoever can name a disorder has power to shape the patient’s narrative, identity and future. The authority to diagnose resides traditionally with the professional. The diagnosis app, however, challenges this authority, by the pre-diagnostic work it supports. While the app developers, for the most part, are careful to postulate to lay users that their apps are not intended to replace the doctor, they nonetheless guide the patient to shape the diagnosis she will present to the clinician and thus potentially redistribute power in the doctor-patient relationship [31].

The diagnosis app raises not only the specter of shifting locus of authority in the doctor patient encounter, but also questions about agendas concealed behind the algorithmic applications; about quality control and credentials. They approach diagnosis from a necessarily binary system of information, effacing the fine detail of the individual case in the classificatory framework. Digital health technologies such as mobile apps have the potential, therefore, of configuring new domains of health, illness and medical practice. The decision-making pathways of diagnosis apps – their algorithmic authority – delimit the ways in which bodily signs, symptoms and sensations are reported, understood, interpreted and treated.

Apps also differ from previous diagnostic technologies in the open access of the lay public to their use. Many medical apps are designed for medical education as well as to assist qualified practitioners in their daily work routines. As such, they contain highly detailed and specialized medical information that previously was not available to lay people. This in itself is a major change in the ways in which medical information has been distributed. Medical apps designed for healthcare professionals can be downloaded and used by anyone. Some now include specialized sensors or attachments that facilitate monitoring of the body and the identification of medical conditions that in past times were the preserve of members of the medical profession. This openness of access disturbs professional dominance and expertise. We would suggest, therefore, that digital technologies such as medical apps have the potential for challenging medical control over knowledge about health and illness further than any other technology from the past.

Our study suggests that much more research needs to be undertaken into the content of medical apps and how they are actually used (or for that matter, resisted). Further research that directs attention at how both healthcare practitioners and lay people are using these apps would elucidate how these issues are being experienced in the healthcare setting and the implications for healthcare services.