General discussion 231 7 - Overdetection is the identification of abnormalities that are unlikely to cause harm, for instance because of absent or slow progress, or spontaneous recovery. Examples include full-body scanning, which may reveal non-progressive tumours (known as ‘incidentalomas’) or growing cultures of saliva in self-limiting upper airway tract infections. - Overdefinition occurs when the threshold for a risk factor is lowered without evidence of net benefit, or when the definition of a disease is expanded to include people with ambiguous or very mild symptoms. Examples include the definition of hypertension (which is <130/80 according to the American College of Cardiology and the American Heart Association, and <140/90 according to the European Society of Hypertension) [41, 42], diagnosing pre-diabetes [43], or Alzheimer’s disease [44]. Overdiagnosis can lead to labelling (including stigmatisation) and overtreatment, which can have negative physical, mental, social, and financial impact on patients [40]. Rates of overdiagnosis exist for various conditions: - In Australia, estimates suggest that overdiagnosis occurs in 18% of all cancer diagnoses in women and 24% in men. The most commonly affected types are renal, thyroid, melanoma, breast, and prostate cancer [45]. - A recent meta-analysis found that using the LLN (lower limit of normal) definition resulted in overdiagnosis of COPD in an average of 48% of cases, with outliers above 60% in primary care. However, when using the GOLD criteria, the prevalence of COPD overdiagnosis was significantly lower. Overdiagnosis was also found to be associated with inappropriate treatment [46]. - Approximately 5% of adult patients who self-report a penicillin allergy are truly allergic to penicillins [47]. Overreporting of penicillin allergy leads to the prescription of non-first choice antibiotics, which can contribute to the development of antimicrobial resistance. Implementing the test-management pathway in healthcare policymaking may reduce overtesting, overdiagnosis, and consequently overtreatment. This involves evaluating the net benefit of testing on people-important outcomes in guideline development. This approach aligns with updated guidance for systematic reviewers and guideline developers [19, 21, 48, 49]. Conclusions The research conducted in this thesis identified challenges and proposed suggestions to overcome these challenges, including the suggestion to focus on those elements of the test-management pathway that drive the decision of whether or not to recommend
RkJQdWJsaXNoZXIy MTk4NDMw