Greater expectations in a cancer trial: Absolute more than relative survival increases, community more than academic clinicians

There is no consensus on how the difference between control and experimenta l outcome rates, the clinically important difference, should be estimated w hen designing a clinical trial. We sought to determine whether community an d academic clinicians had different perceptions as to what would constitute a clinically important increase in survival, when asked to respond in abso lute or relative terms, before a trial was started rather than when the res ults were already known. A telephone survey of 25 practicing Canadian oncol ogists was performed. Questions were asked as to the importance of acceptab le and minimally acceptable improvements in survival for a hypothetical tri al of pancreatic cancer where the baseline survival was expected to be betw een 2 and 8 months. Responses were sought for absolute (additional months) or relative gains (percent improvement) in survival. The mean absolute addi tional survival expectations corresponded to at least a doubling of baselin e survival and tended to be greatest when the prognosis was poorest (p = 0. 06). Relative expectations for improved survival varied with baseline survi val (p < 0.001). When improvement in survival was requested in relative ter ms, the median expected improvement was 25%. This is highly significantly d ifferent than when survival improvements were requested in absolute terms ( p < 0.0001). Median absolute survival expectations were greater for communi ty as compared with academically affiliated physicians (p = 0.046). We foun d that physicians are inconsistent in their interpretation of qualitative d ata. What constitutes a potentially clinically important treatment effect d iffers whether viewed in relative or absolute terms before the performance of a trial. Expectations were greatest when the prognosis was poorest and d iffered between community and academic physicians.