Data Source

Registry

Type of Study

Cross-section, regular

Data gathering method

Registries

Registries, based on hospital patient files

Ease

Data cannot be downloaded directly.

Conditions of access

Formal agreements (e.g. fees, research contracts, institutional agreements, etc.)

Access to data requires: (1) Access fee (2) Official authorisation (CCS) (3) Authorisation of register holder is required www.dutchhospitaldata.nl

Time elapsed between application and reception of data

Months

Type of available data (e.g. anonymised microdata, aggregated tables, etc.)

Anonymised microdata

Formats available

SPSS (default)

Language issues

Documentation is only available in Dutch.

Years of collection, reference years, and sample sizes

Data was first collected between 1981 and 2010 (data collected in 2011 should be available in April/May 2013); data collection is ongoing and is scheduled to continue at some point in the future. It was nearly an integral observation, but in recent years 13-18% of the discharges are missing.

First year of collection

1981

Stratification if applicable

Not applicable

Base used for sampling

Not applicable

Geographical coverage and breakdowns

All hospitalisations in short-stay general and university hospitals. (Semi)-private clinics are excluded.

Age range

No age restrictions

Statistical representativeness

Representative for all hospitalisations in short stay general and university hospitals.

Coverage of main and cross-cutting topics

Detailed study of: Hospitalisation; frequency, diagnosis.

Further Reading

de Bruin, A., Kardaun, J., Gast, F., et al. "Record linkage of hospital discharge register with population register: Experiences at Statistics Netherlands". Statistical Journal of the United Nations ECE 21 (2004): 23-32.

Possibility of linkage among databases

Linkage is possible, but linkage to key LMR is limited, therefore not all discharges can be linked. Data can be longitudinally linked with auxiliary file of uniquely linkable persons based on population register.

Entry errors if applicable

Missing registrations of hospitalizations, unknown diagnosis (from 2006 onwards; 18% in 2011). Missing registrations are imputed to make a complete file, but the imputed records cannot be linked.

Breaks

Change in ICD-version from 2012 onwards (ICD10). Linkable data from 1995 onwards.

Consistency of terminology or coding used during collection

Yes, except for ICD9-ICD10 change from 2012 onwards

Strengths: The data source provides long time series and high coverage (up to 2005). Weaknesses: The registry is missing registrations from 2006 onwards.