The PROMs physio keeps skipping.

Ask a busy physiotherapist how their patient is doing and the answer comes back fluent. “Much better. Pain's about a four now. Walking comfortably for twenty minutes. We'll do another two sessions and discharge.” Ask the same clinician to show you the number on day one, day fourteen, day twenty-eight, and the answer often comes back as a search through last month's notes.

The patient is improving. The clinician knows it. The patient knows it. What's missing is the line that proves it.

This is the quiet shape of outcome tracking in most physiotherapy clinics. The clinician does the work. The work just isn't visible.

Why the PROM never quite gets done

It is not because anyone is against measuring outcomes. Every physiotherapist trained in the last twenty years has been taught about ODI for low back, KOOS for knee, DASH for upper limb. The instruments exist. The clinicians know them. They just don't use them consistently, and the reason is structural.

Each step in tracking outcomes costs something. Picking the right instrument for the presentation costs a minute of thought. Getting the patient to fill it in costs five minutes of session time. Scoring it correctly costs another two. Recording the score in a way that will be comparable next month costs whatever the EMR makes you do.

Multiply that by twenty patients in a day. The system quietly collapses into “the patient says they're better” and the PROM never quite gets done. A 2023 systematic review on patient adherence to PROM completion in clinical care found that completion rates are typically not well documented or explained, and that the biggest enablers are administration method, patient capacity to engage, and clinician engagement with the resulting data.¹ A Dutch primary-care physiotherapy study found that fewer than two thirds of patients completed standard outcome measures like the VAS, Patient-Specific Complaints questionnaire, Quebec Back Pain Disability Scale, or Neck Disability Index.²

What you actually lose

The first loss is the non-responder. A patient who isn't improving on the standard plan needs a tactic change, a re-screen, or a referral. Without a measured trend, the non-response often isn't caught until week six. With a tracked PROM, it shows up by week two.

The second loss is the patient's own confidence. Recovery often doesn't feel linear to the person living through it. A printed graph showing the ODI score halved over a month is a different conversation from “trust me, you're doing well”. The patient leaves the visit believing the work is working, because they can see it.

The third loss is the clinician's own evidence base. Every physiotherapist accumulates a quiet sense of what works. A systematic PROM record turns that sense into something they can show, defend, and improve. Most physios do better work than their notes suggest. Tracking is how the work becomes visible to the person doing it.

The right PROM is not a long list

You don't need to memorise forty-seven instruments. You need a short list of trusted PROMs that cover the body regions your clinic actually treats. ODI or Roland-Morris for low back. KOOS or Lysholm for knee. DASH or QuickDASH for the upper limb. NDI for neck. LEFS for broader lower extremity work. Tampa for kinesiophobia. That covers the majority of what walks through most clinic doors.

The art is matching the right instrument to the right clinical question. Are you measuring disability, function, fear, quality of life, return to sport? Each PROM is built to answer a specific question, and using the wrong one is how clinics end up with data that doesn't mean what they think it means.

Where AI helps, narrowly

The useful role for AI here is not to interpret the PROM. The clinician does that. The useful role is to remove the friction that stops the PROM from happening in the first place.

That means auto-selecting the right instrument based on the presentation, so the clinician doesn't have to remember which knee PROM the literature prefers for this kind of patient. It means scheduling reassessment at clinically meaningful intervals automatically, so the follow-up isn't left to memory. It means scoring the questionnaire the moment the patient submits it, so the number is on the chart before the next visit starts.

And it means linking the PROM score to the diagnosis it was tracking, so the trend is visible at a glance the next time the clinician opens the patient record. Not a number floating in last month's free-text note. A line on a chart, tied to a condition, comparable visit by visit.

What this looks like in Healui

We built outcome tracking around a simple principle: the system should do the parts that aren't clinical. The clinician should do the parts that are.

The AI auto-selects from 47 validated PROMs based on the diagnosis. The questionnaire goes to the patient via WhatsApp or the clinic tablet. It scores itself. The score lands on the chart tied to the condition. The next visit, the clinician sees the trend, not a list of numbers to interpret manually. When the trend stops moving, the system flags it.

None of this is glamorous. It is the unglamorous administrative work that, when removed, gives the clinician back the cognitive space to actually look at the patient.

The honest closing

Outcome tracking is one of those topics in physiotherapy where everyone agrees in principle and nobody does it well in practice. The gap isn't belief. It is bandwidth.

Close the bandwidth gap and the tracking happens, naturally. Close it badly and you add another data-entry chore on top of an already full afternoon. The difference is whether the system is built to help the clinician, or to extract data from them.