Healthcare App Intelligence for Physicians

Arriving at diagnosis, the key point of every treatment process, physicians and nurses use large-screen medical applications, such as EHR/EMR and others, to record test and examination results and diagnostic information.

Usually, to help doctors enter needed data easily and quickly, desktop applications support the autosuggestion feature. Physicians type in several first characters and then pick the relevant item from the list.

Sadly, mobile equivalents of the large-screen solutions can’t offer similar time-saving convenience.

Smaller screens and virtual keyboards make typing slower and increase the chance of typos, which is unacceptable for recording medical data. Scrolling and choosing one relevant option among thousands of examination results, medications and diagnoses is even worse.

Now, how can we handle this convenience problem effectively?

Decision Support

Decision support, enabled at the stage of medical app development, works as a physician’s mobile care council.

For example, when a doctor enters the patient’s symptoms, the app suggests selecting from the narrowed list of the possible diagnoses, where the most relevant ones appear on top.

Apart from increasing convenience, this mobile healthcare decision support allows physicians to reaffirm diagnoses and treatment as well as consider additional tests and checks to reduce medical errors.

Right, it might seem a little bit too sci-fi, but we have several examples to show you how intelligent healthcare mobile apps can be. Let’s imagine a patient with typical chickenpox (varicella) symptoms – a headache, stomachache, fever and fluid-filled blisters. They also suffer from vomiting, chills and skin rash in the locations uncommon for chickenpox.

A physician enters the patient’s data into the mobile app and gets reminded that chickenpox is not the only possible option, as the symptoms also point at necrotizing fasciitis, which can quickly worsen the patient’s condition. The physician can prescribe a CT scan and/or skin tissue biopsy to decide on the accurate diagnosis and treatment.

Decision Support Implementation

Now, off to more prosaic questions. How to enrich a care application with decision support and not spend major cash on it?

The best thing is that you don’t have to be a Mayo-level provider to get your hands on this solution. The current technology simplifies the creation of smart autosuggestions lists thanks to implementing the following two intelligence techniques:

  • User behavior analysis
  • Machine learning

The mobile app remembers a user’s searches and choices and then offers relevant options according to the established patterns and dependencies. When the user’s behavior analysis gets together with the sets of symptoms entered, this alliance creates a pool of consistent alternatives.

Machine learning is unstoppable, and invisible for regular user, thus, the application’s suggestions evolve constantly. The more physicians use these healthcare mobile apps, the more mature and relevant machine learning becomes. Say, if a solution gathers information from 10 physicians instead of one, it learns 10-times faster.

Even if your knowledge base is empty, in 1-2 months the app will suggest options for the common cases. Handling rare cases effectively will take up to a year. To accumulate experience across the organization, we also recommend storing the machine learning outputs on a server, at the mobile app’s backend.


God is in the details. At first thought, changing a small, visible part of a healthcare app to allow a faster and easier entering of patients’ data (diagnoses, treatment, examination results and more) can’t have a big impact. But dig deeper, and you’ll discover the inner meaningful layer.

Apart from saving time and providing convenience, mobile apps enriched with smart autosuggestion techniques serve as a tool to reaffirm particular diagnoses, test results, checks and treatment.

Doctors can lessen diagnostic doubts, highlight possible care gaps and select needed tests and checks precisely, thus reducing the chance of a medical error. This is how a small detail leads to accurate and quality care delivery.

Andrei Khomushka is Head of Android Development Department at ScienceSoft.

About The Author