Development of web applications for self-diagnosis

Development of web applications for self-diagnosis

February 9, 2016

You only need to do type “online diagnosis” into any search engine to come up against a long list of web applications which will supposedly help to identify an illness. Nowhere, however, will you find enthusiastic reviews of a particular website that has actually given an accurate diagnosis. Why so? We will take a look at the main difficulties of attempting to create an online diagnosis tool.

The question of motivation

By analysing the diagnostic tools we can find online, we may judge that few developers actually intend to solve real issues for their users. The purpose of most of these applications is to drive traffic and, ultimately, to monetise the project. There are also resources developed as PR resources to improve a clinic's image.

The usefulness of such medical web applications is judged first and foremost by the question of whether the program can answer the user's questions: What is the problem? What kind of doctor should the patient see? What tests should be carried out? How dangerous is the issue? Suppose that on the basis of the location of a headache the online diagnostic tool has come to the conclusion that there is a 25% chance of hypertension and a 10% chance of a migraine without aura. Would this make you feel better? These are the results from a real algorithm of one of the best diagnostic tools.

The main issue with online diagnostic tools is that the developers very rarely ask themselves which problems their project is supposed to solve. They are not interested in this, believing that there are enough hypochondriacs to bring traffic to the site. It doesn't matter that on clicking the link they find nothing more than a glorified toy. After all, only a doctor can provide an adequate diagnosis and administer proper treatment, so why bother?

At the same time, to create something useful one has to deal with the following questions:

  • How can we help patients before they go to the doctor?
  • How can we help when a doctor's consultation is unavailable?
  • Why might medical assistance be inaccessible?

This information will serve as the starting point before we begin to design the project.

Stupidity is worse than treason

It is understandable why a programmer should endeavour to make an application simple: the less the user has to click the mouse with furrowed brows, trying to identify the character of a skin rash, the better. This has led to a problem common to most applications, whereby the algorithm attempts a diagnosis on the basis of just one or two symptoms. The doctor, on the contrary, looks at a combination of symptoms when making a diagnosis; and the symptoms themselves have certain nuances to be kept in mind. There isn't just pain or a rash. Discharge has a source, colour, consistency and smell.

It is impossible to create truly effective software which turns an electronic medical reference book into an interactive experience. There is no way to design an algorithm for diagnostics which works independently of a doctor, a person who actually knows what they are doing. Simplicity would have to be sacrificed, and we can provide it only on the level of interface.

When does the application help?

Let's return to the question of usefulness. It stands to reason that a doctor is indispensable in the diagnosis and in prescribing treatment. However, the patient doesn't only face the question of how to treat the symptoms. Before meeting a doctor the patient should try to eliminate the main symptoms. It wouldn't do any harm to find some over-the-counter medicine to alleviate the condition. For instance, if a patient has aching joints, a prompt to use Ibuprofen instead of Analgin would be appropriate.

In such situations first aid instructions would be of great use, especially when the problem is not a particularly common one. You will rarely find information on administering emergency care to someone having a heart attack or stroke on TV or in medical brochures. Such applications might also help when a person is unable to decide how to act. Should they call an ambulance or just rest in bed to convalesce? What kind of doctor should be used? Sometimes one might simply need such seemingly obvious information as an ambulance phone number, which will vary depending on the region and network provider, and can often be hard to find.

A web application providing the best possible answers to patient queries will be the most popular. As becomes clear from all that has been outlined, this will need serious investment.

Why undertake it?

Primary web diagnostics will be helpful to clinics which attract patients online. If a patient gets answers to some questions before seeing a specialist they will tend to be more receptive to a doctor who clarifies the situation. But private medicine as a rule is not ready to spend the money needed to create a truly useful web application with a multistage diagnostic algorithm and professional recommendations.

At the same time, a quality web application for self-diagnosis would see great demand from ordinary users, since a healthy way of life and control over one's body is becoming a huge trend. Whoever is first to enter the market as an online diagnosis professional capable of providing real help to people has an opportunity not only to capture the market, but also to receive state support, since the problem of interactive diagnostics in public healthcare also needs to be developed.