The Canadian Assisted Reproductive Technologies Register (CARTR) Plus Database: A Validation Study

Abstract

Background: Maternal and fetal complications of pregnancy, such as ectopic pregnancy, preterm birth, placenta previa, and preeclampsia, are known to occur with greater frequency among women who conceive using assisted reproductive technology (ART) compared with those who conceive naturally. Despite the increased relative risks, since the absolute incidence of these conditions remains low researchers often rely on large cohorts to evaluate complications following the use of ART. Unfortunately, large prospective cohort studies are both time-consuming and expensive to conduct, therefore, administrative databases and registry databases are often used instead. These sources are excellent sources of data for research purposes as they are relatively inexpensive, easily accessible and are collected on a large population scale. Routinely-collected data are generally not collected with the intention of performing research and, therefore, may introduce information bias due to variable misclassification. To provide insight into data quality and inform the extent to which this may be a concern, database validation studies are highly recommended. Notwithstanding increasing utilization of administrative databases and registries in research investigating pregnancy outcomes of fertility treatments, there is a paucity of validation studies in the literature for these routinely-collected data. The objective of this thesis was to perform two studies, a systematic review of database validation studies among fertility populations and a validation study of the Canadian Assisted Reproductive Technologies Register. Methods: Study 1: We conducted a systematic review to identify validation studies of databases that contain routinely-collected data in populations using ART. In addition to searching Medline, Embase and CINAHL for relevant literature, we also examined webpages of international ART surveillance programs and databases. Database managers for national ART registries were contacted to obtain unpublished reports of data quality assurance practices. Screening of articles was performed in two steps by two independent reviewers. Data from relevant articles were extracted and results were synthesized qualitatively. Study 2: Using patient chart reabstraction for in vitro fertilization cycles performed in 2015, we then performed a validation study of the Canadian national ART registry, CARTR Plus. Clinics spanning Canada were recruited to participate. We selected twenty-five data elements from CARTR Plus that were deemed clinically relevant. These data were reabstracted from a random sample of patient charts at each of the participating clinics, (considered the reference standard), and compared to those in the database. We calculated agreement between the two data sources using sensitivity, specificity, positive and negative predictive values, kappa and intraclass correlation coefficients. Results:Nineteen studies met the inclusion criteria for the systematic review, of which one was a validation of a national registry. Seven studies used an ART database to validate another administrative database or maternal questionnaires, while four studies used either maternal questionnaire or an administrative database to validate an element within a registry. Prevalence estimates were generally not reported well for the data variable of interest. In the CARTR Plus validation study, agreement between reabstracted data and the CARTR Plus was excellent for most of the investigated variables. There were a few variables –– namely follicle stimulating hormone (FSH) level, clinical reason for treatment cycle, oocyte origin and elective embryo transfer –– that had moderate agreement. Conclusion: National fertility databases and registries are used for research and feedback to clinics and government and, therefore, the accuracy of these data is essential. Our validation study provides valuable information for Canadian research studies using the database, and can serve as a guide for other databases to perform quality assurance projects.