ORIGINAL ARTICLES, p 1176, 1187
As a long-time advocate of patient-reported outcome measures (PROMs) in routine clinical practice, the publication in this issue of BJD of not just one but two systematic reviews of potential measures for use in skin cancer is most welcome.[1, 2] While both reviews highlight several of the challenges in selecting a suitable instrument, it is their differences that provide a window into other often overlooked questions. The first is, what is the point or reason for introducing PROMs alongside well-established clinician-reported outcomes? Neither review really addresses whether the aim is to enhance the clinical management of individual patients, to provide a means of comparing the outcomes (and by implication, the performance) of different providers (practitioners, hospitals, Trusts), to determine the cost-utility of treating skin cancer (i.e. comparing this with other clinical areas in which healthcare resources might be spent), or all three. This is important as it will determine the type of PROM that will be needed. Having said that, it is a question that needs to be answered by policy-makers rather than researchers.
The second question is, when is a PROM good enough or fit-for-purpose? There is a sense, in particular in the review by Bates et al., that they are seeking perfection; but there is the danger that the perfect can be the enemy of the good. While both reviews mention the existing national PROMs programme for elective surgery in England (not the U.K.), they implicitly credit the PROMs being used with stronger measurement properties than is the case (having been instrumental in determining their selection I can say that!). While it is important that any measure be sufficiently reliable, valid and responsive, progress in introducing PROMs in dermatology could be held back by the use of too demanding standards.
My third observation is that although both reviews identified the same four candidate disease-specific PROMs (Facial Skin Cancer Index, Dermatology Life Quality Index, Skindex, Skin Cancer Quality of Life Impact Tool), their recommendations differed. While Bates et al. want to go back to the drawing board and develop a new PROM (albeit based on existing items from the four PROMs), Gibbons et al. are more pragmatic and suggest further testing of the existing ones before making a choice. My inclination is to support the latter approach. While none of the four is perfect, I suspect at least one of them will prove adequate for helping patients and clinicians manage the condition and be sufficiently responsive for comparing providers. And given that the EQ-5D has become established (until something better comes along) as the generic PROM of choice in the English National Health Service, I would include it in any draft questionnaires being piloted. Only in that way will we learn whether or not it is suitable to detect the impact of skin cancers on generic quality of life.
The final thought is that it would be wise for those interested in taking this forward to discuss it with those involved in the current piloting of PROMs for other cancers. Hopefully the latter will have some suggestions as how best to interpret PROMs data in conditions in which, unlike most elective surgery, patients may have no symptoms or disabilities at the time of their diagnosis so the option of comparing a patient's health before and after treatment is not available. Despite all these challenges, I await progress with interest.