Development of a decision support tool for primary care management of patients with abnormal liver function tests without clinically apparent liver disease: a record-linkage population cohort study and decision analysis (ALFIE).

摘要: Objectives: to determine the natural history of abnormalities in liver function tests (LFTs), derive predictive algorithms for disease and identify most cost-effective strategies further investigation. Data sources: MEDLINE database from 1966 September 2006, EMBASE, CINAHL Cochrane Library. Methods: population-based retrospective cohort study set primary care Tayside, Scotland, between 1989 2003. Participants were patients with no obvious signs registered a general practitioner (GP). The followed up those an incident batch LFTs subsequent or mortality over maximum 15 years. health technologies being assessed LFTs, viral autoantibody tests, ultrasound biopsy. Measures used epidemiology Tayside (ELDIT) database, time-to-event modelling, derived using Weibull survival model, decision analyses NHS perspective, cost–utility analyses, one-way two-way sensitivity analyses. Results: total 95,977 had 364,194 initial median follow-up 3.7 Of these, 21.7% at least one abnormal test (ALFT) 1090 (1.14%) developed disease. Elevated transaminases strongly associated diagnosed disease, hazard ratios (HRs) 4.23 [95% CI (confidence interval) 3.55–5.04] mild levels 12.67 (95% 9.74–16.47) severe versus normal. For gamma-glutamyltransferase (GGT), these HRs 2.54 2.17–2.96) 13.44 (10.71–16.87) respectively. Low albumin was all cause mortality, 2.65 2.47–2.85) 4.99 4.26–5.84) levels. Sensitivity predicting events 5 years low specificity high. Follow-up time split into baseline 3 months, months 1 year year. All high probability female, methadone use, alcohol dependency deprivation. shorter-term models overall c-statistics 0.85 0.72 outcome respectively, 0.88 0.82 Calibration good Discrimination In retesting dominated referral as option. However, top percentile risk incremental ratio £7588 relative referral. Conclusions: GGT should be included care. If patient has moderate is ALFTs depends on willingness pay NHS. Cut-offs are arbitrary developing aids it important treat LFT results continuous variables