Keep Your Validated Instrument. Surround It With Real-World Data.

If you're running an R01 around a PHQ-9, an SF-36, a QOLIE, an mFARS, or a KCCQ, you already made a good decision. A validated instrument has been tested to capture a specific construct and to do it consistently across patients and sites. Years of psychometric work stand behind every item. That rigor is the entire reason it's defensible to a study section and to a regulator. So I want to be clear up front: keep it.

The pitch you hear from a lot of new tools is the opposite. Swap your "outdated" questionnaire for our richer, smarter instrument. Maybe one day, once a new approach has earned the same trust. Today, the lower-risk path is to build on the instrument you already have.

The gap isn't validity, it's cadence

A validated instrument has two real limits, and neither one is whether it measures what it claims.

The first is cadence. It fires episodically, at scheduled visits, maybe every three or six months. Between those visits it goes silent. That silence is where most of a patient's life happens, and where most of the signal lives. A flare, a bad week, a slow recovery, a medication that stopped working in March and showed up as a worse score in June. The instrument records the June number. It tells you nothing about March.

The second is scope. A validated instrument captures exactly what it was built to capture and nothing else. The PHQ-9 measures depression severity. It doesn't know the patient moved cities, started a night shift, or spent two weeks in heavy wildfire smoke. The score moves. The instrument can't tell you why.

What complementing looks like

Keep the validated instrument on its cadence. Keep the score it produces and the endpoint you built from it. Then surround it with continuous real-world data between visits: short daily logs, wearables, automatic location-linked context like air quality, pollen, and weather, and patient-submitted photos and video for things a number can't hold, like a rash, a gait, a tremor.

This is additive, and the distinction matters. Your validated instrument still tells you what changed. The real-world stream tells you why. The June score drops, and now you can see the patient had three flares in May that lined up with a pollen spike and a missed refill. Same endpoint, same regulatory standing, far more context behind it.

That's what we do at Forma. We administer the patient's own validated instrument, often with better completion because the surrounding ePRO experience is lighter and built for a phone, and we add the continuous stream around it. It comes out as one longitudinal, analysis-ready dataset instead of a questionnaire export and a pile of disconnected device files. You can see it in how we run studies like Friedreich's ataxia, where the formal scale anchors the data and the between-visit signal fills in the rest.

If you're designing the protocol

Here's how to layer continuous data without touching your primary endpoint.

1. Leave the validated instrument exactly as is. Same items, same cadence, same scoring. Your endpoint and its analysis plan don't move.
2. Add the real-world stream as secondary and exploratory data. Daily logs, wearables, environmental context, media. It surrounds the instrument; it doesn't replace any item in it.
3. Time-align everything to the instrument's schedule so each formal score has its surrounding context attached. That alignment is what lets the stream explain the measure instead of sitting next to it.
4. Pre-specify how you'll use the continuous data. Hypothesis-generating, covariate, or context for interpretation. Naming its role up front keeps your confirmatory analysis clean and keeps reviewers comfortable.

The instinct to enrich your data is right. You don't have to give up the validated instrument to act on it. Keep the rigor you already have, and spend your effort on the part that's actually missing: the days between the visits.