Can We Rely on Pharmacy Claims Databases to Ascertain Maternal Use of Medications during Pregnancy?
Authors:
- 9 March 2017
- journal article
- research article
- Published by Wiley in Birth Defects Research
- Vol. 109 (6) , 423-431
Abstract
Background: Administrative databases are increasingly used to measure drug exposure in perinatal pharmacoepidemiology. We aimed to estimate the concordance between records of prescriptions filled in pharmacies and self‐reported drug use during pregnancy.Methods: Data on self‐reported medication use were collected at each trimester of pregnancy among a sub‐sample from the Organization of Teratology Information Specialists Antidepressants in Pregnancy Cohort. Women were eligible if they were Quebec resident and provided their pharmacist's contact information. Maternal self‐reports were compared with prescriptions filled in pharmacies, which are transferred to pharmaceutical services files of Quebec provincial health plan database (Régie de l'asssurance maladie du Québec). Positive and negative predictive values (PPV and NPV) for medications taken chronically (antidepressants, thyroid hormones), acutely (antibiotics), and as needed (antiemetics, asthma medications) were calculated.Results: Among the 93 participants (mean age = 30.2 ± 3.8 years), 41.9% (n = 39) took at least one antidepressant during pregnancy according to self‐reports, and 39.8% (n = 37) according to pharmacy records. Other commonly used drugs were antiemetics (self‐reported 22.6%, pharmacy record 24.7%), antibiotics (20.4%, 16.1%), asthma medications (15.1%, 15.1%), and thyroid hormones (10.8%, 8.6%). PPVs and NPVs were: (1) chronic medication: antidepressants PPV = 100% (95% confidence interval [CI], 100–100%), NPV = 96% (95% CI, 92–100%); thyroid hormones PPV = 100% (95% CI, 100–100%), NPV = 98% (95% CI, 95–100%); (2) Acute medication: antibiotics PPV = 87% (95% CI, 70–100%), NPV = 92% (95% CI, 86–98%); (3) as needed medications: antiemetics: PPV = 78% (95% CI, 62–95%), NPV = 96% (95% CI, 91–100%); asthma: PPV = 33% (95% CI, 3–64%), NPV = 99% (95% CI, 97–100%).Conclusion: The high PPV and NPV validate the use of filled prescription data in large databases as a measure of medication exposure. Birth Defects Research 109:423–431, 2017.Keywords
Funding Information
- Santé (30962)
This publication has 36 references indexed in Scilit:
- Use of macrolides during pregnancy and the risk of birth defects: a population‐based studyPharmacoepidemiology and Drug Safety, 2015
- The Quebec Pregnancy Cohort – Prevalence of Medication Use during Gestation and Pregnancy OutcomesPLOS ONE, 2014
- Validity of congenital malformation diagnostic codes recorded in Québec's administrative databasesPharmacoepidemiology and Drug Safety, 2013
- Evolving knowledge of the teratogenicity of medications in human pregnancyAmerican Journal of Medical Genetics Part C: Seminars in Medical Genetics, 2011
- Maternal medicine: Drug prescription patterns before, during and after pregnancy for chronic, occasional and pregnancy‐related drugs in the NetherlandsBJOG: An International Journal of Obstetrics & Gynaecology, 2006
- Agreement and validity of pharmacy data versus self-report for use of osteoporosis medications among chronic glucocorticoid usersPharmacoepidemiology and Drug Safety, 2006
- Postmarketing surveillance for human teratogenicity: A model approachTeratology, 2001
- Improvement of drug exposure data in a registration of congenital anomalies. Pilot-study: Pharmacist and mother as sources for drug exposure data during pregnancyTeratology, 1999
- Birth Outcomes in Pregnant Women Taking FluoxetineNew England Journal of Medicine, 1996
- Recall of medication during pregnancy: Validity and accuracy of an adjusted questionnairePharmacoepidemiology and Drug Safety, 1995