How And Why: Running a Clinical Trial On A Digital Health App
There is no shortage of passion in healthtech. Every healthtech startup promises to improve patient care, and expects this improved care to translate into revenue and traction. In reality, incremental increases in the quality of care only translates into revenue and traction when measured and validated in the clinical setting.
Why Bother?
Clinically validated applications have staying power in the market. These are the applications that will build a secure technology layer around a patient, and enable digital medicine to be “a tap away”. If we expect the healthcare system to pay for applications at price points on par with conventional drugs and therapies (an expectation bolstered by the investments made into the space), then we need to hold ourselves to the same standard of efficacy.
How Do You Get Clinical Metrics At An Early Stage?
There are several steps a startup can take to commit to collecting early clinical data.
- Decide how you want to make patients better. Do you want to patients to reduce their chance of heart attacks? Do you want to reduce the risk of patients falling?
- Find a true clinical metric that connects to this goal. The key part here is a direct and proven connection. For example, lowering cholesterol reduces the chance of a heart attack, thus, cholesterol is a reasonable clinical metric to measure.
→ Remember: Clinical Metrics Are Biology Based
Meaningful clinical metrics are often hard to improve, such as HbA1c, weight, BMI, and cholesterol. These are indisputable hard facts about a patient that cannot change overnight. In mental health applications, the clinical metrics are the clinician’s clinical judgement and assessment of a patient’s mental health, rather than any kind of patient reported survey data on its own.
→ Engagement Is Not A Clinical Metric
Because a hospital is not FarmVille or CandyCrush. There is no doubt capturing patient attention is necessary. Yet, there is insufficient evidence correlating time spent using health apps, with health improvement. For instance, let’s examine the popular medication adherence tools. These apps may help patients to remember to take their medication, if we assume that if they are opening the app, they are also taking their medications. That logic appears sound, until we realize how many times we open an application and dismiss the notifications. Patient engagement does not equal use, and neither does opening the app.
→ Patient Reported Data Is Not A Clinical Metric
Patient reported data is amazing for many reasons- it provides a layer of insight only the patient can provide, and otherwise difficult to gather. Yet, it is not a substitute for clinical metrics, nor is it often times, reliable.
3. Challenge the metric. Find as many confounds as you can. For example, cholesterol is genetic, so can be a difficult metric to lower uniformly. Go back to step 2. Go back to step 1. This is the tough part, but instrumental and worth the effort.
4. Commit to a metric. Once you pick your key metric, stick with it. That is the point of true scientific rigor.
5. Find a way to run your trial. This is tricky, but easier if you form a meaningful partnership with a clinician who can help you run a scientifically sound trial. They want to discover new tools and technologies to help their patients, and they want a way to make sure that the technology they suggest to patients works. There is a way to make this a win-win process, and with a wide variety of organizations supporting research in this space, a way to make this cost efficient as well.
6. Be results agnostic! This is very difficult, but it is the entire point. You want to build something that creates value and improves patient health. If your product isn’t achieving the results it needs to, it will catch up to you. Learn from the results of the trial. Iterate. Figure out what you did actually impact.