Linking Australian Stroke Clinical Registry data with Australian government Medicare and medication dispensing claims data and the potential for bias. (5th April 2021)

• Record Type: Journal Article
• Authors: Andrew, Nadine E.; Cadilhac, Dominique A.; Sundararajan, Vijaya; Thrift, Amanda G.; Anderson, Phil; Lannin, Natasha A.; Kilkenny, Monique F.
• Is Part Of: Australian and New Zealand journal of public health. Volume 45:Number 4(2021)
• Page Start: 364
• Page End: 369
• Publication Date: 2021-04-05
• Subjects: data linkage -- record linkage -- health data -- clinical registry -- stroke -- primary care -- medication -- bias; Public health -- Australia -- Periodicals; Public health -- New Zealand -- Periodicals; Medical care -- Australia -- Periodicals; Medical care -- New Zealand -- Periodicals; 362.10993
• DOI: 10.1111/1753-6405.13079
• Languages: English
• ISSNs: 1326-0200

Abstract

Abstract: Objective : We aim to report the accuracy of linking data from a non‐government‐held clinical quality registry to national claims data and identify associated sources of systematic bias. Methods: Patients with stroke or transient ischaemic attack admitted to hospitals participating in the Australian Stroke Clinical Registry (AuSCR) were linked with Medicare and medication dispensings through the Australian Medicare enrolment file (MEF). The proportion of registrants in the datasets was calculated and factors associated with a non‐merge assessed using multivariable analyses. Results : A total of 17, 980 AuSCR registrants (January 2010 – July 2014) were submitted for linkage (median age 76 years; 46% female; 67% ischaemic stroke); the proportion merged was 97% MEF, 93% Medicare and 95% medication dispensings. Data from registrants born in Asia were less likely to link with the MEF (adjusted Odds Ratio [aOR]: 0.20; 95%Confidence Interval [CI]: 0.15, 0.27). Data for those aged 85‐plus compared to those under 65 years were less likely to merge with Medicare (aOR 0.25; 95%CI:0.21, 0.30) but more likely to merge with dispensing claims data (aOR: 2.15 (95%CI:1.71, 2.69). Implications for public health : Linkage between the AuSCR, a national clinical quality registry and Commonwealth datasets was achieved and potential sources of bias were identified.