(152) Representativeness of the PHARMO Electronic Health Records (EHRs) from Primary Care

Publication Number: 483

Background: Real world evidence obtained from observational data plays an increasing role in healthcare decision making. Representativeness of this observational data is pivotal to provide valid and generalisable study results and conclusions.

Objectives: The aim was to assess the representativeness of the General Practitioner (GP) Data of the PHARMO Data Network of the general Dutch population.

Methods: A cross-sectional study was performed to compare demographic characteristics (sex, age and urbanisation), medication use, diagnoses and deaths. The year 2018 was used as index year, because this was the most recent year all sources had all data available. PHARMO’s general practitioner (GP) Data comprises data from EHRs, including information on diagnoses, laboratory test results, referrals to specialists and healthcare product/drug prescriptions. Data on the Dutch population was obtained from Statline, the electronic database of Statistics Netherlands (CBS), which offers publicly available data. The CBS obtains demographic and vital statistics from the population register of all Dutch municipalities, medication statistics from the National Health Care Institute of the Netherlands, which receives data about reimbursed drugs, and data regarding diagnosis from the Nivel Primary Care Database. The standardised difference (SD) was used to compare the proportions between the Dutch population and PHARMO’s GP population. An absolute SD >0.2 was considered a difference.

Results: On January 1st, 2018, 3,466,321 persons were included in PHARMO’s GP population (mean age: 41.6 years, 49.7% males). The sex and age distribution was similar to the Dutch population. Regarding medication use, only the pharmacological subgroups ‘viral vaccines’ and ‘hormonal contraceptives for systemic use’ differed (SD >0.2); use in PHARMO’s GP population was higher compared to the Dutch population. These differences were explained by the fact that ‘viral vaccines’ and ‘hormonal contraceptives for systemic use’ are not completely reimbursed via the basic health insurance. No differences were observed regarding diagnoses. Mortality was lower in PHARMO’s GP population than in the Dutch population, which is in line with expectations. Persons who die are more often older and no longer under the care of a GP (i.e. institutionalised).

Conclusions: The GP Data of the PHARMO Data Network are representative of the Dutch population regarding demographic characteristics and provided primary care in terms of medication use and diagnoses. Medication use was more difficult to compare as national data does not include non-reimbursed medication and the GP Data of the PHARMO Data Network do include information about non-reimbursed medication.