Characteristics of General Practice Research Database 2005

CHARACTERISTICS OF GENERAL PRACTICE RESEARCH DATABASE 2005

Table 27.1 lists the main characteristics of the GPRD in 2005. It should be recognised that GPs use their computers primarily to create electronic medical records for the purpose of managing their patients. However, contributing GPs are provided with record-ing guidelines that define what information should be recorded electronically so making the research under-taken in GPRD more valid:

•   Demographics, including the patient’s age and sex.

•   Medical diagnosis, including free-text comments.

•   All prescriptions and immunisations as given in primary care.

•   Referrals to hospitals or specialists.

•   Laboratory results, including microbiology.

•   Treatment outcomes, including hospital discharge reports where patients are referred to hospital for treatment.

•   Key patient information, e.g. smoking status, height and weight.

•   Date and cause of death.

•   Pregnancy-related information.

Following receipt and processing of a data collec-tion, the GPRD Group provide feedback reports to the contributing practice on the completeness of data in key areas (e.g. date and cause of death and patient registration details), to enable practices to address any deficiencies they have with their recording. In addi-tion, the quality of recording across the entirety of data contributed by a practice is assessed by means of the ‘up to standard’ audit that assesses the complete-ness, continuity and plausibility of data recording in key areas, in accordance with the recording guide-lines issued to practices. Where data quality is found to be acceptable, the practice is judged to be ‘up to standard’ and marked as such in the database; this marker can be used to identify those practices where data recording is considered by the GPRD Group to be of sufficient quality for research purposes.

In April 2004, the Quality Outcomes Framework (QOF) was introduced into UK general practice. This framework provides incentives to practices for the provision of high quality care that naturally involves improved data documentation. Data from the prac-tice records are submitted to and analysed by the Quality Management and Analysis System (QMAS), a national IT system, that supports the QOF payment process. Achievement is measured against indica-tors in four domains; most importantly, the clinical domain, which focuses on a range of indicators in 11 key disease areas. Given the key role of practice records in supplying the data needed to assess prac-tice achievement/performance, there is now even more emphasis for practices to ensure that their records are complete, particularly in areas related to QOF indica-tors. This can only be of benefit to research.

The GPRD is a unique public health research tool, which has been used widely for drug safety stud-ies and many other types of pharmacoepidemiolog-ical research. There are now over 500 GPRD-based peer-reviewed publications, nearly 2000 peer-review impact points and an ever expanding international user base. Numerous independent validation studies have confirmed a high level of completeness and validity of the data in the GPRD. A large study recently examined the validity of the computerized diagnoses of autism in the GPRD. Anonymised copies of all relevant avail-able clinical reports, including GP’s notes, consultant, speech therapy and educational psychologists reports, were evaluated for 318 subjects with a diagnosis of autism recorded in their electronic general prac-tice record. For 294 subjects (92.5%), the diagnosis of pervasive developmental disorder was confirmed after review of the records, providing evidence that the positive predictive value of a coded diagnosis of autism recorded in the GPRD is high (Fombonne et al., 2004). Another study compared the distribution of the cause of death in GPRD to national mortality statistics and concluded that they were broadly simi-lar. This provides further evidence that the GPRD population is broadly representative of the general population (Shah and Martinez, 2004).

Recent research includes a case-series analysis of the risks of myocardial infarction and stroke after common vaccinations and naturally occurring infec-tions. It found that there was no increase in the risk of myocardial infarction or stroke in the period after influenza, tetanus or pneumococcal vaccination. However, the risks of both events were substantially higher after a diagnosis of systemic respiratory tract infection and were highest during the first 3 days, suggesting that acute infections are associated with a transient increase in the risk of vascular events (Smeeth et al., 2004). A study by Martinez compared the risk of non-fatal self-harm and suicide in patients n taking selective serotonin reuptake inhibitors (SSRIs) with that of patients taking tricyclic antidepressants. No evidence was found that the risk of suicide or non-fatal self-harm in adults prescribed SSRIs was greater than in those prescribed tricyclic antidepres-sants (Martinez et al., 2005).

Most of the drug safety research in the GPRD has concerned the estimation of relative rates, i.e. the rate of outcomes in exposed patients divided by that in control patients. But relative rates do not convey the public health importance of a safety issue. Large rela-tive rates for rare events may not be of major concern, whereas small relative rates for frequent events may potentially have large implications. An example for this may be the cardiovascular risk of selective cyclooxygenase-2 inhibitors, which may have affected a large number of patients. Recent research developed methods to estimate, in the GPRD, individual long-term probabilities specific for a patient’s age, sex and clinical characteristics. This was done for estimating the long-term risk of fracture in patients using oral glucocorticoids. As an example, it was found that a woman aged 65 years with rheumatoid arthritis, low body mass index (BMI) and a previous history of frac-ture and falls, who used 15 mg glucocorticoids daily, would have a 5-year fracture risk of 47% (a man with similar history, 30.1%) (van Staa et al., 2005). This approach to quantify individualised long-term probabilities can help to better quantify the risks and benefits associated with a treatment.