Health Databases in Saskatchewan

HEALTH DATABASES IN SASKATCHEWAN

Saskatchewan is a province in western Canada with a stable population of about 1 million people, or about 3.2% of the total population of Canada. The province provides a publicly funded health system for its residents, who are each assigned a Health Services Number upon registration that uniquely identifies that person, and which is captured in records of health service utilization, enabling the linkage of computer databases. Only a very small percentage (less than 1%) of the population of Saskatchewan is excluded from the health registry. Prescription plan coverage excludes about 9% of the population, primarily Indi-ans, who are covered by another government agency. Hospital services and most physician services are available to all persons in the health registry. The population registry captures demographic and cover-age data on every member of the eligible population, including gender, marital status, date of birth and date of death.

Drugs covered by the drug plan are listed in the Saskatchewan formulary; non-formulary drugs are generally not covered. The drugs listed are intended for outpatient use, although the database includes prescriptions to residents of long-term care facil-ities. The formulary is updated semi-annually; as of July 2004, more than 3500 drug products were listed (Downey et al., 2005). The drug database contains information from September 1975, with an 18-month hiatus in 1987–88 when data were incom-plete. The database includes patient, prescriber, phar-macy and cost information. Drug information includes pharmacologic-therapeutic classification, using the AHFS classification system, active ingredient, generic and brand names, strength and dosage form, drug manufacturer, date and quantity dispensed. Unavail-able is information on non-formulary drug use, over-the-counter drugs, use of professional samples and in-hospital drugs. The database also does not provide information about the dosage regimen prescribed, the reason the drug was prescribed, or patient compliance. Approximately 8.4 million prescription claims were processed by the drug plan in fiscal year 2002–03 (Downey et al., 2005).

Data from hospitalizations, including day surg-eries, include up to three discharge diagnoses (ICD-9 codes), up to three procedures, an accident code (ICD-9 external cause code), admission and discharge dates, and attending physician and surgeon (where applicable). Procedures are coded using the Cana-dian Classification of Diagnostic, Therapeutic and Surgical Procedures. There is a lag time of about 6 months from date of discharge to the date when hospital data are available electronically. In 2001–02 there were approximately 140 000 inpatient separa-tions (discharges, transfers or deaths) of adults and children (Downey et al., 2005).

Physician services data are obtained from claims, and include diagnoses (three-digit ICD-9 codes) and procedures (coded from a fee-for-service payment schedule established by the Health Registry and the provincial medical association). These data are limited, however, in that diagnostic data are given only to support the claim for payment, and only one three-digit ICD-9 code is recorded per visit.

Linkage can be made to the Saskatchewan cancer registry, which is required to record all persons diagnosed with cancer, including non-melanoma skin cancers and in situ cancers, and suspected as well as confirmed cancers. A lag time of 6 months exists from date of diagnosis to availability of the data.

Vital statistics data are also maintained by Saskatchewan Health; all birth, death, stillbirth and marriage data are collected. Although cause of death is initially coded as received on a death registration form, it is updated if an autopsy diagnosis is received. The underlying cause of death is recorded electroni-cally as well, and is defined as the disease or injury that initiated the sequence of events that led to death.

Other information available includes institutional long-term care and home care services, mental health services that cover both inpatient psychiatric care and community-based outpatient care, alcohol and drug abuse treatment data and microbiologic and biochem-ical laboratory records.

Hospital medical records are retrievable after the appropriate approvals are obtained, with patient iden-tifiers removed from the record. Hospital record retrieval rates often exceed 95%. Outpatient record retrieval has not approached that level of success. Information on potentially important confounders are only available in patient records or through direct patient contact.