Background There is increasing use of electronic healthcare data for health research. Identifying and linking secondary electronic health (eHealth) records could offer insight into common non-communicable diseases and healthcare utilisation in specific populations. Secondary health systems in the United Kingdom (UK) are unique for recording Outpatient, Inpatient and Accident and Emergency (A&E) visits. These records, provided by England, Scotland and Wales offer a variety of parameters such as admission/discharge date, diagnosis (ICD10 codes), treatment/procedure undertaken (OPCS codes) and the cost of treatment. However, health records do not distinguish specific occupational groups, such as the military. In this study we have undertaken a data linkage of the King’s Centre for Military Health Research (KCMHR) cohort to eHealth records, identifying use of secondary care in serving and ex-serving personnel.
Methods The KCMHR cohort study comprises 9,990 serving and ex-serving military personnel, 8,602 of whom consented to record linkage. eHealth records for each region were obtained by extracting unique patient identifiers (National Health Service (NHS) number, forename, surname, sex and date of birth). Record linkage for Scotland and Wales engaged pseudo-anonymised fuzzy matching algorithms (probabilistic match rate at 90% and above) to identify and match military personnel. England requires a valid NHS for deterministic linkage.
Results An overall matching rate of 76.66% (n = 6,336) was achieved. However, there were clear regional discrepancies. Scotland had a matching rate of 14.22% (n = 1,223), Wales 8.13% (n = 899) and England 60.7% (n = 5221). Further, differences for those who provide an NHS number were observed. 79.95% (n = 6,877) of participants provided an NHS number, resulting in a matching rate of 83.25% (n = 5,725), whereas for those who did not have an NHS number a matching rate of 35.42% (n = 611) was obtained. A total of 61,558 eHealth episodes of care were identified and extracted.
Conclusion Linking the cohort to multiple regional eHealth records has identified discrepancies in the matching processes, creating challenges for researchers. England, Scotland and Wales obtain, store and process eHealth records using different methodologies. Scotland and Wales utilise participant demographics without the need of a NHS number whereas matching in England requires an NHS number. We believe this requirement could result in under reporting of prevalence, as an NHS number is not required upon admission. The eHealth records show promise for identifying common physical health conditions in military personnel. However, further work is required to identify synergy and overcome regional variations.