Reliability of such instruments needs to be provedBMJ 1996; 313 doi: https://doi.org/10.1136/bmj.313.7057.626b (Published 07 September 1996) Cite this as: BMJ 1996;313:626
- Danny Ruta,
- Andrew Garratt
- Senior lecturer Department of Epidemiology and Public Health, University of Dundee, Dundee DD1 9NL
- Research fellow Department of Health Sciences, University of York, York
EDITOR,—Charlotte Paterson should be commended for piloting her patient generated outcome measure in a primary care setting.1 The “measure yourself medical outcome profile” (MYMOP) has potential for routine use, not least because it is simple to use. Also, the authors' tests of validity and responsiveness indicate that it correlates with health status and perceived improvements in health. Three important questions remain unanswered, however. Firstly, does the MYMOP really measure outcomes that matter to patients—that is, that have meaning and relevance in the context of their daily lives? Secondly, does it make sense to combine three quite distinct health dimensions—symptoms, activity, and wellbeing—to produce a single index? Thirdly, is the MYMOP reliable enough to be used in assessing individual patients?
The MYMOP is adapted from the patient generated index, which was designed to measure quality of life.2 This index derives from Calman's definition of quality of life as “the extent to which our hopes and ambitions are matched by experience.”3 Patients are asked to list the five most important areas of life that are affected by their condition, to rate how badly they are affected, and, finally, to weight the relative importance of potential improvements in these areas. Although the MYMOP is promoted as a patient centred measure, it imposes three professionally derived dimensions on the patient and fails to take account of the relative importance that he or she attaches to these disparate domains. It is unclear if the patient centredness of the original measure has been compromised.
An instrument that is valid must be reliable, but validity is a relative concept. The methods used by researchers in assessing validity are crude, particularly when applied to quality of life measures. The fact that the MYMOP seems to be “fairly” valid does not mean that it is precise enough to detect real differences in an individual over repeated administrations. It may well satisfy the reliability requirements for assessments of individual patients, but this needs to be shown.
In summary, we would recommend the following methods to anyone wishing to develop the MYMOP approach: in depth interviews with patients who have completed the instrument to assess qualitatively its face and content validity; a content analysis of the symptoms and activities that patients include in their MYMOP; reliability testing; and correlation of scores obtained with the MYMOP with those obtained with a quality of life measure such as the patient generated index.