A G Randolph1, G H Guyatt, W S Richardson. 1. Department of Critical Care Medicine, The Hospital for Sick Children, University of Toronto, ON, Canada.
Abstract
OBJECTIVES: Counseling critically ill patients and their families about what the future is likely to hold requires accurate prognostic information. Our goal is to teach clinicians how to find and critically appraise prognostic studies that examine homogeneous populations. CLINICAL EXAMPLE: An article describing the outcomes of a group of children who are in a prolonged, persistent vegetative state. RECOMMENDATIONS: The validity of prognostic studies is increased when: a) the sample of patients is representative; b) patients are homogeneous with respect to prognostic risk; c) follow-up is sufficient to minimize the possibility that the missing patients could alter the interpretation of the results; and d) health outcomes are evaluated, using objective and unbiased criteria. The likelihood of these outcomes over time and the precision around these probability estimates should be easily understandable. Before using the results of these studies to counsel patients and families, practitioners should ensure that the patients in the study and their management are similar to the patient in question, and that follow-up of the subjects is sufficiently long. CONCLUSIONS: The criteria outlined in this article may assist clinicians in interpreting articles describing the prognosis of patients with similar clinical conditions.
OBJECTIVES: Counseling critically ill patients and their families about what the future is likely to hold requires accurate prognostic information. Our goal is to teach clinicians how to find and critically appraise prognostic studies that examine homogeneous populations. CLINICAL EXAMPLE: An article describing the outcomes of a group of children who are in a prolonged, persistent vegetative state. RECOMMENDATIONS: The validity of prognostic studies is increased when: a) the sample of patients is representative; b) patients are homogeneous with respect to prognostic risk; c) follow-up is sufficient to minimize the possibility that the missing patients could alter the interpretation of the results; and d) health outcomes are evaluated, using objective and unbiased criteria. The likelihood of these outcomes over time and the precision around these probability estimates should be easily understandable. Before using the results of these studies to counsel patients and families, practitioners should ensure that the patients in the study and their management are similar to the patient in question, and that follow-up of the subjects is sufficiently long. CONCLUSIONS: The criteria outlined in this article may assist clinicians in interpreting articles describing the prognosis of patients with similar clinical conditions.