Swedish Agency for Health Technology Assessment and Assessment of Social Services
Can prognostic studies be trusted?
Certain critical basic questions must always be asked about research findings related to prognoses concerning people with a particular problem or medical condition.
Share:
Reading time approx. 2 minutes
Published:
Publication type:
- Have the investigators studied a representative sample, where all individuals were in the same phase, such as in the early phase of an illness? Ideally, all subjects were included at the same early stage, such as at the very onset of the problem.
- Have enough of those who were to be included in the study been followed for a sufficiently long period of time? Follow-up must be long enough to allow important outcomes to be identified. If there is attrition from the study, researchers must analyse the underlying reasons and check that these individuals did not differ in any essential way from those who remained.
- Were outcomes registered using an objective or ‘blinded’ approach? If measurement entails subjective assessments, ideally the assessors should have no information about study participants that may introduce bias, a process known as blinded assessment.
- If it is known that people with certain characteristics have a better or worse projection than others, have the researchers adjusted for such factors in their analysis? In the case of health, age or phase of illness may be of significance for outcome.
- Can a difference in the way results are reported yield a different picture? For example, five-year survival is often reported for cancer, i.e. the number of people in the group who are still alive at five-year follow-up. However, those who die may do so early or late during that interval. A survival curve with several measurement points can provide more information.
- How precise is the number reflecting prognosis, according to the 95-percent confidence interval of the assessment? The more observations, the narrower the interval and the more precise the result. At a late stage in the study, assessment figures tend to become less precise than in the beginning due to participant attrition.
- Is the reported prognosis valid and useful here and now? Or does the current situation deviate too much from the studied conditions? Is there any evidence that interventions improve the prognosis? • RL
Paraphrased from: Centre for evidence-based medicine (CEBM), University of Oxford. Critical appraisal of prognostic studies. Worksheet. Downloaded July 2020 from https://www.cebm.net/wp-content/uploads/2018/11/Prognosis.pdf
Page published