Article

Generating pre-test probabilities: a neglected area in clinical decision making.

Centre for Clinical Epidemiology and Biostatistics, Level 3, David Maddison Building, University of Newcastle, Newcastle, NSW 2300, Australia.
The Medical journal of Australia (Impact Factor: 3.79). 06/2004; 180(9):449-54.
Source: PubMed

ABSTRACT To assess the accuracy and variability of clinicians' estimates of pre-test probability for three common clinical scenarios.
Postal questionnaire survey conducted between April and October 2001 eliciting pre-test probability estimates from scenarios for risk of ischaemic heart disease (IHD), deep vein thrombosis (DVT), and stroke.
Physicians and general practitioners randomly drawn from College membership lists for New South Wales and north-west England.
Agreement with the "correct" estimate (being within 10, 20, 30, or > 30 percentage points of the "correct" estimate derived from validated clinical-decision rules); variability in estimates (median and interquartile ranges of estimates); and association of demographic, practice, or educational factors with accuracy (using linear regression analysis).
819 doctors participated: 310 GPs and 288 physicians in Australia, and 106 GPs and 115 physicians in the UK. Accuracy varied from about 55% of respondents being within 20% of the "correct" risk estimate for the IHD and stroke scenarios to 6.7% for the DVT scenario. Although median estimates varied between the UK and Australian participants, both were similar in accuracy and showed a similarly wide spread of estimates. No demographic, practice, or educational variables substantially predicted accuracy.
Experienced clinicians, in response to the same clinical scenarios, gave a wide range of estimates for pre-test probability. The development and dissemination of clinical decision rules is needed to support decision making by practising clinicians.

Download full-text

Full-text

Available from: Lesley Patterson, Jul 07, 2015
0 Followers
 · 
102 Views
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Accident Compensation Corporation of New Zealand Abstract Background An analysis of Accident Compensation Corporation claims shows “inconsistent and inadequate diagnoses” by health care providers. Diagnostic performance is a result of two independent parameters, namely discrimination (accuracy) and decision (bias). Bias is related to the medical practitioner’s perception of the costs and benefits of making one choice over another. Bias may be statistical, sociological, political, biological or psychological in nature. This study investigated the factors that potentially bias diagnostic decision-making by general practitioners and the subjective value placed on these factors by different stakeholder groups in society. Methods Phase 1 of the study used focus groups of standard setters for general practitioners to identify factors that influenced diagnostic decision-making in general practice. These factors were evaluated for importance and desirability using standard Delphi methodology and Rasch analysis. Phase 2 of the study evaluated the importance and desirability of the factors identified in Phase 1 for influencing decision making as judged by significant health care stakeholder groups in New Zealand. Participant response was via questionnaire analysed by the Rasch Model. Results Thirty-nine factors were identified that potentially biased diagnostic decision-making in general practice. The measurements of, particularly, desirability have high reproducibility across stakeholder groups and high positive loading for the first principal component consistent with construct validity. No stakeholder group identifies factors consistent with Bayes’ theorem of diagnostic reasoning as being the only desirable influence on diagnosis. There is considerable categorical homogeneity between the stakeholder groups GP, GPACC, P, RACCSLT and RACCSST. Conclusions The findings of this and other studies challenge the current biomedical paradigm, indicating a less than Bayesian approach to medical decision-making. A social constructivist model, incorporating non-Bayesian factors into the definition of “illness” versus “disease”, may be more representative of reality. A social constructivist model of medicine is incompatible with the current legislatory and administrative framework within which the Accident Compensation Corporation and a number of other medical organisations operate.
  • Source
    The New Zealand medical journal 11/2004; 117(1203):U1095.
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: According to numerous commentators, clinical freedom, the art of medicine, and, by implication, a degree of patient welfare, are threatened by evidence based medicine (EBM). As EBM has developed over the last fifteen years, claims about better evidence for medical treatments, and improvements in healthcare delivery, have been matched by critiques of EBM's reductionism and uniformity, its problematic application to individual patients, and its alleged denial of the continuing need for clinical interpretation, insight, and judgment. Most of these attacks on EBM and defences of clinical freedom fail. They are based on erroneous understandings of the relationships between inductive knowledge, clinical uncertainty, and action. Evidence based medicine is a necessary condition for clinical freedom, not a threat to it, and EBM is not something to be balanced with either clinical experience or patient preferences. The art and science of medicine are more conceptually and practically connected than the defenders of clinical freedom, whatever they conceive that to be, are willing to admit.
    Medical Humanities 07/2005; 31(1):23-30. DOI:10.1136/jmh.2004.000195