Friday, September 26, 2008
Physician - patient disagreement about performance status
Cancer has a paper looking at the prognostic implication of physician-patient disagreement about performance status. The data are taken from a several multi-center therapy trials involving ~1600 patients with metastatic lung or colorectal cancer in which patients and physicians independently rated performance status (using ECOG and/or Karnofsky scores). For some reason these data are from trials done in 1987-1990. It's unclear to me why such old data were used...no trials since then employed both physician and patient assessment of functional status?
Patients had a median survival of ~9 months. About 60% of patients 'disagreed' with their physician's assessment of their performance status; overall physicians overestimated patients' function. However, patients with the worse physician-rated functional status were more likely to rate their function as being better. Disagreement in functional status was associated with poorer prognosis (by ~1-2 months) and this survived multivariate analysis at least for Karnofsky (they adjusted for age, sex, work status, and depression).
I couldn't find anything in the paper to dissuade me that what was happening was that the patients' assessment of their functional status was likely more accurate than physicians' and that since overall patients swung more pessimistically than physicians, those patients with divergent assessments of functional status tended to be, in fact, sicker and so had a worse prognosis. Anyway, it raised the question for me of how we actually go about assessing performance status? How are people trained in doing this? Are there evidenced based ways of assessing it that we should all be adopting? My training in it, while decent I think (I hope!), was relatively informal....