When surgeons talk about the outcome after Dupuytren surgery, we refer to the narrow surgical target. Our main surgical target is to correct a joint contracture. Our patients also want us to achieve this target with the least loss to them, without any complication that may worsen them, in the shortest time and in the most efficient way.

As surgeons our training and focus is on measuring impairment in range and strength. We assess joint range using goniometry and hand strength using dynamometry. For Dupuytren surgery the outcome is joint range, our outcome is measured using finger goniometry (Engstrand et al. 2012; McVeigh et al. 2016). One study showed that the error rate is 3° but another study showed a 4° intra-rater difference while the difference is far greater between raters. This does need accurate and unbiased goniometry.

But we still do not know simple things. We do not know what degree of correction helps the patient. We do not know the minimal clinically important improvement in the joint contracture, which will benefit the patient. So although we know the target is to improve the range, we do not know how much improvement is going to make a difference to our patients.

Our main concern for the patient is that the joint contractures will recur, but there is no clear definition of recurrence used in published literature. There is now consensus that 20° progression from 6 to 12 weeks to 1 year defines a recurrence (Dias 2015; Felici et al. 2014; Peimer et al. 2013).

Whatever we measure and the closer you look at the method of measurement, we find that the measurement can be flawed. It is easy to assume that we have chosen the correct measurement tool, that it is accurate and lets us assess what is important.

It is easy, when presenting or publishing a paper, to use sophisticated statistical techniques that discount the errors in or relevance of measurements. As clinicians and researchers, we easily accept the “validated” outcome measures, but rarely question whether these measures are truly appropriate for the disorder we are treating and whether the measure is likely to respond to the intervention we are proposing.

Our measurements must help us advise our patient, and this involves asking patients what they think. This requires patients to rate the outcome of intervention. As Berwick said, we need to establish whether our treatment has helped the patient “as they see it”.

The best way to do that is to get patients to assess their own outcome. So using patient-rated outcome measures (PROM) asks the patients to judge the value of our intervention. Have we helped them? We ask the patient a series of questions to gauge their views about their own health (Devlin and Appleby 2010). The patient must be completely independent of you as the clinician. The clinician must not prompt the patient or translate questions as this can promote bias in their responses. PROMs allow us to answer the questions: “Can we be sure that spending on health is justified by the output that it produces? And “Are scarce resources being used in a way that maximises their value to patients and society?” A PROM is a clinical endpoint which measures whether the patient is helped or harmed noting how “a patient feels, functions and survives”.