129 Overdiagnosis via computer suggestions for follow up testing: time to hit restart. (20th August 2018)

• Record Type: Journal Article
• Authors: Xu, Eric; Mei, Junyi; Capote, Karina; Cembrowski, George
• Is Part Of: BMJ evidence-based medicine. Volume 23:Supplement 2(2018)
• Page Start: A61
• Page End: A61
• Publication Date: 2018-08-20
• Subjects: Evidence-based medicine -- Periodicals; 616.005
• DOI: 10.1136/bmjebm-2018-111070.129
• Languages: English
• ISSNs: 2515-446X

Abstract

Abstract : Objectives: Not infrequently, patient laboratory printouts suggest followup tests. In reviewing the report of a 46 year old British Columbia female's ferritin of 86 ug/L, we were puzzled by the comment: 'Possible Iron Deficiency' for ferritins between 50 and 100 ug/L based on 2010 guidelines. This comment medicalizes the patient and she thinks she has iron deficiency; the comment also initiates the cycle of further testing with the end result being the over-ordering of iron-related tests in an individual presumably without a 'low'akers in the Canadian and British Columbia health care systems, starting at the level of the ministers of health down to the clinical laboratory owners and medical directors. We need a curtailment of these automated comments followed by a medical economics study of the utility of such algorithmic testing. Method: In the neighboring province of Alberta, which lacks such ferritin guidelines, we collected 4 months of ferritin and iron results generated in a large urban referral laboratory. We then derived the prevalence of low, normal and high iron for varying ferritins and age groups of female and male patients. Results: The association of ferritin to low iron levels is only evident in Albertan women and men with low ferritins. For women, the highest risk of low iron stores is associated with ferritins less than 15 ug/L (26% incidence). For males, increased risks of reduced iron stores are associated with ferritin levels less than 50 ug/L (between 9% to 44% incidence of iron deficiency). For higher ferritins, there appears to be little relationship between higher ferritin and low iron stores (between 3% and 5% incidence). Conclusions: The interpretative comment for ferritins of 50 to 100 ug/L is misleading as these normal ferritin levels are not diagnostic of decreased iron. The British Columbia guidelines state that screening for iron deficiency is not indicated in the general population. The interpretive comment implies that regardless of the ferritin value, low iron levels should be ruled out. Such interpretive comments do not make sense and result in increased profitability for today's highly efficient referral laboratories. Similar 'diagnostic aids' exist in many clinical information systems and induce overtesting and overdiagnosis. In the spirit of patient-centric medicine and cost reduction, it is incumbent on the laboratorian to search for, and eliminate such non-value added interpretive comments.