Premium
Failed validation of risk prediction model for intervention in renal colic patients after emergency department evaluation
Author(s) -
Dean Tanya,
Crozier Jack,
Klim Sharon,
Kelly AnneMaree
Publication year - 2016
Publication title -
anz journal of surgery
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.426
H-Index - 70
eISSN - 1445-2197
pISSN - 1445-1433
DOI - 10.1111/ans.13109
Subject(s) - medicine , renal colic , emergency department , confidence interval , retrospective cohort study , cohort , surgery , pathology , alternative medicine , psychiatry
Background It has been reported that three criteria (size of calculus ≥6 mm, visual analogue scale pain score at discharge ≥2 cm and location above mid‐ureter; the P apa criteria) were sensitive for predicting patients who require intervention (surgery or lithotripsy) within 28 days of index emergency department ( ED ) visit for ureteric colic. It was suggested that absence of these criteria identified a group for whom early follow‐up may not be needed. No validation has been reported. We aimed to validate these criteria. Methods Retrospective cohort study of patients with clinical presentation of ureteric colic and radiologically proven renal tract stones. Data collected included demographics, clinical features, features of the stone, imaging results and 28‐day outcome. Outcome of interest was performance of the P apa criteria for prediction of urological intervention by clinical performance analysis. We also undertook a post hoc analysis to identify predictors of urological intervention for the group overall and for the subgroup discharged from ED . Results Two hundred and twenty‐four patients were studied (median age 49, 79% male) with 75 (33%) requiring urological intervention within 28 days. The presence of any of the P apa criteria had sensitivity for urological intervention of 83.9% (95% confidence interval ( CI ) 71.2–91.9%) with specificity of 47.7% (95% CI 38.9–56.6%), positive predictive value of 40.9% (95% CI 31.9–50.4%) and negative predictive value of 87.3% (95% CI 76.8–93.7%). Nine patients with no P apa criteria had intervention: 12.7% (95% CI 6.8–22.4%). Conclusion The P apa criteria are not sufficiently accurate to determine which patients require intervention or a subgroup who do not need specialist urological follow‐up.