Tag: Dose Response Curve

The Data Response Curve (In Honor of the Dose Response Curve)

Data Response Curve

All medical students learn about the dose response curve in pharmacology lectures. The dose-response curve informs us of how we should dose a medication in the context of its efficacy and its toxicity. Too little medicine won’t have the desired effect, and too much medicine can be toxic.

In the era of digital health, data have become the new “big pharma,” and we are facing the emergence of a data-response curve in which access to too little data is inactionable, and access to too much data can be overwhelming. Digital health devices abound today, and has enabled quantification of nearly every health and wellness metric imaginable. Sadly, in our exuberance about these new sources of data, we often conflate “more data” with “better data.”

In the era in which data have become the booming commodity of exchange in healthcare, we describe an emerging data-response curve. Large data sets can be at best clarifying or at worst self-contradictory. Too little data on the data-response curve, as with medication dosing, can be insufficient for effective action or decision-making. Too much data can be toxic to the user such as the physician, leading to poor decisions or worse, to analysis paralysis.

We live in a world of exploding data, and we need to be thoughtful. Medicine is a people business in need of data, not a data business on need of people. In reductive form, all humans make decisions based on inputs (data) from their environment and on individual analysis of the data in the form of non-linear, non-quantifiable perception. When we as doctors, policy-makers, or simply as human beings receive these inputs, one of three things happens: 1. We make a decision to do something (for example a treatment decision based on abnormal data), 2. We make a decision to do nothing (for example, normal data which we believe requires no action), or 3. We need more data in order to make an informed decision to do item #1 or item #2. More does not necessarily equal better data. More data is simply more.

Better data are actionable data wrapped in the context of the patient and the patient’s condition. Imagine each piece of objective data connected to concurrent subjective data, and surfaced in the context of a specific condition relevant to the patient. HealthLoop enables patients to generate contextual objective data married to their subjective symptoms and served to a clinician in an actionable context. High signal and low noise are the digital health equivalents of on the dose-response curve of a favorable therapeutic window.

See where some folks live on the Chart. Where do you live?


Forgotten Password?