Areas to check before launching a healthcare app on Infermedica API

💡 Infermedica API is a powerful tool that connects top medical knowledge with the capacities of Artificial Intelligence (AI). Learn more about our medical API.

We've been using Infermedica API to develop white-labeled modules and custom solutions for leading healthcare companies for years. With the growing popularity of our medical API, we've decided to open it for everyone. Gladly, we’ve observed that a growing number of companies and startups are using it as the engine for their applications. They are building amazing single applications, as well as complex healthcare platforms, upon it.

Some examples of solutions using Infermedica API are Microsoft Health Bot, Diagnostikare, and IMC Medical Center.

If you are planning to set up your healthcare application with Infermedica API, make sure that your solution is flawless in all possible areas. The success of your application relies on this.

What all of the companies working with Infermedica API have in common is the thorough verification of their solutions before launching. We strongly encourage and support this approach. Working with health solutions requires top security and the highest performance.

Things that should be checked range from legal requirements to technical and design issues. An example could be the undefined scope of age in the app. Infermedica API can be adjusted to various age groups and this is why the age scope is not set by default. If this parameter remains unset, the user can enter a negative value in the age field, or high numbers that definitely exceed a human’s possible age. This can lead to 4xx errors, as the reasoning engine gets confused with such values.

Another, quite common, example is the adjustment of apps for mobile users, where elements of the application are not adjusted for small screens and they strangely widen instead of narrowing. Correcting this positively influences the user experience.

It is also worth checking all the possibilities that come with Infermedica API. Although using many of them is optional, developers often decide to explore and choose them. One such example is an endpoint responsible for adding questions about related symptoms. It helps us understand the causes of the patient's condition better and suggest a more accurate symptom assessment at the end of the health checkup.

Using additional endpoints or auditing selected areas often improves the accuracy of an app’s analytic capacities and user flow. We have reviewed similar situations and gathered a list that should help everyone deciding to set up their healthcare solution on Infermedica API. The recommendations we have selected are based on the experience we had assisting with numerous API integrations, as well as our team’s UX and technical know-how.

Each of them has several elements worth checking before the deployment of a newly developed application. You can also use our list to optimize the performance of existing applications.

Below, we present a brief description of these areas. You can also jump directly to our practical checklist.

This is probably the most important, and the least technical area that should be verified. It should take place at the very beginning of the project. Checking contractual and legal requirements is very important as it helps to set up an appropriate communication of the tool and its purpose. It is crucial to underline that any symptom assessment or triage solution built upon Infermedica API is no substitute for a doctor’s visit. It must be treated as a preliminary symptom assessment tool and communicated as such.

Additionally, there are some requirements, like Terms of Service, to be embodied, as stipulated in the contract with Infermedica.

Make sure that your application is secure. The 'App key' and 'App ID' should be hidden and protected. Simple checks will make you sure that your credentials are safe and not accessible to others. Additionally, you should make sure the medical information of your application users is safe. This is solely on your side, since the data you send to us is fully anonymized and serves only for statistical purposes.

In every application, the user flow is different but there are some universal rules that help end-users to move fluently in the healthcare application. We have developed a set of rules to help you utilize all the possibilities of our medical API and the endpoints that it is providing. This includes elements like setting the minimal number of requested initial symptoms, an additional question about related symptoms, more triage levels, display of common names, and age range adjusted to the actual target group. The goal is to simplify the process of gathering information about the patient’s state and optimize the number of API calls, as well as maximize the accuracy of the final results.

Infermedica API provides multiple endpoints. Using them is not obligatory but can better the performance of your healthcare application. The endpoints available are designed to increase the medical accuracy and improve User Experience, i.e. by giving additional explanations to the answers given by AI or making sure the sex and age of the user are taken into account when returning questions.

Based on our experience with our own healthcare application (which is called Symptomate) for mobile and desktop users, we came up with the best practices and details you should pay attention to.

Think about your users, their goals, and their needs. We’ve observed that apps that solve real problems are performing the best. Also, pay attention to communication with your users and the responsiveness of your application. Sometimes elements like a progress bar, welcome message, or well-selected icons make a huge difference. Take care about the accessibility of your app. There will be people with disabilities among your users, for sure. In this case, Web Content Accessibility Guidelines (WCAG) 2.1 recommendations are good to start with.

Implementing analytical tools (we recommend Google Analytics) will help you observe the flow of your users and make data-driven decisions.

Download the checklist: Areas to check before launching any solution based on Infermedica API.

Last but not least, there is the Feedback Loop. It is a proven way to check the performance of newly developed applications by comparing the initial symptom assessment from the app with the physician’s opinion. With the Feedback Loop, we get transparent information on the Inference Engine accuracy and confirmation if the engine's suggested list of most probable conditions is accurate, as confirmed by final diagnosis by a doctor. This feedback is useful for you, as well as for Infermedica.

If you wish to set up a Feedback Loop or simply need to run a technical review of your implementation, which includes all areas mentioned above, you can always turn to our technical team. They are always ready to support clients working with Infermedica API. You can do so by contacting your Customer Success Manager or emailing support@infermedica.com.

BL/EN/2020/10/08/3