Outcome Measures in Practice: Choosing Them and Actually Using Them

Why standardised measures often stay unused, how to choose one that fits the caseload, and how to fit administration into a session without losing clinical time.

Outcome MeasuresPROMsAssessment

The case for standardised outcome measures is not disputed. What is disputed, in practice, is whether there is room for them: surveys repeatedly find that a large share of clinicians do not use them routinely, and the reasons given are consistent — uncertainty about which measure to pick, and **not enough time in the session**.

Two kinds of measure, two different jobs

The distinction matters because the two are often treated as interchangeable, and they are not.

Clinician-measuredPatient-reported
SourceYour measurementThe patient’s own report
ExampleRange of motion, strength, functional taskPain intensity, function questionnaires
StrengthObjective, comparable across ratersCaptures what the patient actually experiences
Blind spotSays nothing about impact on daily lifeNot independent of mood, expectation, context

A patient whose shoulder range improved by 25° but who still cannot wash their hair has improved on one measure and not the other. Both readings are correct; using only one of them makes the discrepancy invisible.

How to choose without reading the validation literature

The obstacle most often reported is not scepticism but method: clinicians are rarely taught how to judge whether a measure is appropriate for their setting. Four practical questions get most of the way there.

  1. **Does it match the population you actually see?** A measure validated in post-surgical athletes behaves differently in a sedentary, older caseload.
  2. **Does it move when your patients change?** A measure that is stable across the range you treat cannot show your effect. This is responsiveness, and it is the property most often overlooked.
  3. **Is there a published minimal important difference?** Without one, you can see a number change but cannot say whether the change matters.
  4. **Can it be completed in a few minutes?** A measure that needs fifteen minutes of session time will be abandoned within a month, however good it is.

Examples by category

Searches for outcome measures overwhelmingly ask for examples, and the reason is practical: the abstract case is settled, the open question is which instrument to reach for. The table below groups the commonly used families rather than listing every validated tool.

CategoryWhat it capturesCommonly used examples
Pain intensitySymptom severity at a point in timeVisual analogue scale, numeric rating scale
Region-specific functionFunction in the affected areaShoulder, knee, ankle and spine questionnaires
General function and participationImpact on daily life across conditionsGeneric function and quality-of-life instruments
Fear and confidenceBeliefs that shape movement behaviourKinesiophobia and self-efficacy scales
Performance-basedObserved task performanceTimed mobility tasks, balance tests
Activity levelWhere the patient sits relative to their own baselineActivity rating scales

The last row is easily overlooked and often the most informative in sport: knowing a patient’s pre-injury activity level turns a score from an absolute figure into a distance from their own normal.

Fitting administration into real sessions

The time problem is real but it is usually a *scheduling* problem rather than a duration problem. Most patient-reported measures take a few minutes and do not need clinical supervision — which means they do not need to happen during the session at all.

  • **Fix the points, not the intervals.** Intake, each re-assessment, discharge. Decided once, applied to everyone.
  • **Move completion out of the room** where the measure allows it — waiting time, or before the appointment.
  • **Keep the same version every time.** A shortened or reworded questionnaire is a different instrument and its scores are not comparable with the original.
  • **Record who completed it and when.** A score entered from memory at the end of a session is not the same data point.

Reading a score change

A change in score is not automatically a clinical improvement. Two thresholds separate the three things a difference can mean:

ChangeWhat it means
Smaller than measurement errorIndistinguishable from noise — no conclusion
Larger than error, smaller than the minimal important differenceReal but not necessarily meaningful to the patient
At or above the minimal important differenceA change the patient is likely to notice

Reporting a two-point move on a questionnaire as progress, when the published minimal important difference is eight, overstates the result — to the patient and to yourself.

Where outcome measures fit alongside clinical measurement

They do not replace clinician-measured findings; they answer a question those findings cannot reach. The most informative situation is when the two disagree — a joint that measures better while the patient reports no functional gain is a signal to look again at the goal rather than the technique.

For recording measurements so that both kinds of data stay comparable over time, see Physiotherapy Assessment Documentation.

What are outcome measures in physiotherapy?
Standardised tools that quantify change in function, symptoms or participation over time. They fall into two groups: clinician-measured findings such as range of motion or functional tasks, and patient-reported measures such as pain scales and function questionnaires.
What are examples of outcome measures in physiotherapy?
They group into families: pain intensity scales, region-specific function questionnaires, general function and quality-of-life instruments, fear and self-efficacy scales, performance-based tests such as timed mobility tasks, and activity rating scales. Choosing one per family that fits your caseload is more workable than collecting one per diagnosis.
What are common functional outcome measures used in physiotherapy?
Two kinds are used side by side. Performance-based measures record an observed task — timed mobility, balance, single-leg tasks. Patient-reported function questionnaires record how the person experiences the same domain. They answer different questions and are most informative when they disagree.
How do I choose an outcome measure?
Check four things: that it was validated in a population resembling your caseload, that it is responsive to the changes you expect to produce, that a minimal important difference has been published for it, and that it can be completed in a few minutes. A measure failing the last point tends to be abandoned regardless of its properties.
How many outcome measures should I use?
Fewer than most clinicians assume. Two or three used consistently across a caseload produce comparable data and become habit; a different instrument per condition produces a set that cannot be compared and is administered inconsistently.
How much does a score have to change to matter?
At least more than the measurement error of the instrument, and ideally at or above its published minimal important difference. Below that threshold a change may be real without being noticeable to the patient, and reporting it as progress overstates the result.
When should outcome measures be administered?
At fixed points rather than variable intervals: intake, each re-assessment and discharge. Deciding the points once means the timing is not renegotiated each visit, and most patient-reported measures can be completed outside the session.

Sources

Related reading

← All articlesStart the free trial