Insights from and limitations of data linkage studies: analysis of short-stay urgent admission referral source from routinely collected Scottish data

Abstract

Introduction This study identified the referral source for urgent short-stay admissions (SSAs) and compared characteristics of children with SSA stratified by different referral sources. Methods Routinely acquired data from urgent admissions to Scottish hospitals during 2015–2017 were linked to data held by the three referral sources: emergency department (ED), out-of-hours (OOH) service and general practice (GP). Results There were 171 039 admissions including 92 229 (54%) SSAs. Only 171 (19%) of all of Scotland’s GP practices contributed data. Among the subgroup of 10 588 SSAs where GP data were available (11% all SSA), there was contact with the following referral source on the day of admission: only ED, 1853 (18%); only GP, 3384 (32%); and only OOH, 823 (8%). Additionally, 2165 (20%) had contact with more than one referral source, and 1037 (10%) had contact with referral source(s) on the day before the admission. When all 92 229 SSAs were considered, those with an ED referrer were more likely to be for older children, of white ethnicity, living in more deprived communities and diagnosed with asthma, convulsions or croup. The odds ratio for an SSA for a given condition differed by referral source and ranged from 0.07 to 1.9 (with reference to ED referrals). Conclusion This study yielded insights and potential limitations regarding data linkage in a healthcare setting. Data coverage, particularly from primary care, needs to improve further. Evidence from data linkage studies can inform future intervention designed to provide safe integrated care pathways.