Development and validation of an algorithm for identifying urinary retention in a cohort of patients with epilepsy in a large US administrative claims database

Purpose The aim of this study was to develop and validate an insurance claims‐based algorithm for identifying urinary retention (UR) in epilepsy patients receiving antiepileptic drugs to facilitate safety monitoring. Methods Data from the HealthCore Integrated Research DatabaseSM in 2008–2011 (retro...

Full description

Saved in:
Bibliographic Details
Published in:Pharmacoepidemiology and drug safety Vol. 25; no. 4; pp. 413 - 421
Main Authors: Quinlan, Scott C., Cheng, Wendy Y., Ishihara, Lianna, Irizarry, Michael C., Holick, Crystal N., Duh, Mei Sheng
Format: Journal Article
Language:English
Published: England Blackwell Publishing Ltd 01.04.2016
Wiley Subscription Services, Inc
Subjects:
algorithm
Algorithms
Anticonvulsants - adverse effects
Anticonvulsants - therapeutic use
Confidence intervals
Databases, Factual - statistics & numerical data
Epilepsy
Epilepsy - drug therapy
Humans
Medical Records
pharmacoepidemiology
Predictive Value of Tests
Prospective Studies
Retrospective Studies
United States
Urinary incontinence
urinary retention
Urinary Retention - diagnosis
Urinary Retention - epidemiology
Urinary Retention - etiology
epilepsy
algorithm
pharmacoepidemiology
urinary retention
ISSN:1053-8569, 1099-1557
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Purpose The aim of this study was to develop and validate an insurance claims‐based algorithm for identifying urinary retention (UR) in epilepsy patients receiving antiepileptic drugs to facilitate safety monitoring. Methods Data from the HealthCore Integrated Research DatabaseSM in 2008–2011 (retrospective) and 2012–2013 (prospective) were used to identify epilepsy patients with UR. During the retrospective phase, three algorithms identified potential UR: (i) UR diagnosis code with a catheterization procedure code; (ii) UR diagnosis code alone; or (iii) diagnosis with UR‐related symptoms. Medical records for 50 randomly selected patients satisfying ≥1 algorithm were reviewed by urologists to ascertain UR status. Positive predictive value (PPV) and 95% confidence intervals (CI) were calculated for the three component algorithms and the overall algorithm (defined as satisfying ≥1 component algorithms). Algorithms were refined using urologist review notes. In the prospective phase, the UR algorithm was refined using medical records for an additional 150 cases. Results In the retrospective phase, the PPV of the overall algorithm was 72.0% (95%CI: 57.5–83.8%). Algorithm 3 performed poorly and was dropped. Algorithm 1 was unchanged; urinary incontinence and cystitis were added as exclusionary diagnoses to Algorithm 2. The PPV for the modified overall algorithm was 89.2% (74.6–97.0%). In the prospective phase, the PPV for the modified overall algorithm was 76.0% (68.4–82.6%). Upon adding overactive bladder, nocturia and urinary frequency as exclusionary diagnoses, the PPV for the final overall algorithm was 81.9% (73.7–88.4%). Conclusions The current UR algorithm yielded a PPV > 80% and could be used for more accurate identification of UR among epilepsy patients in a large claims database. Copyright © 2016 John Wiley & Sons, Ltd.
Bibliography:ArticleID:PDS3975
ark:/67375/WNG-SL9GWLSX-N
istex:B3B96B57A772DAB7593EB50F8E4A5296EDDA3CF4
ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ObjectType-Article-2
ObjectType-Feature-1
content type line 23
ObjectType-Undefined-3
ISSN:1053-8569
1099-1557
DOI:10.1002/pds.3975