General Practice Research Database for prescribing analysis

Patterns of respiratory drug use in the Lombardy region of Italy, 1995-1997

OBJECTIVE

To examine the patterns of respiratory drugs use in the Lombardy region of Italy, and to quantify the period prevalence of different drug-dose combinations defined by reference to international guidelines for prescription in asthma and chronic obstructive pulmonary disease.

DESIGN

Analysis of records from an anonymised population-based database of dispensed prescriptions.

SETTING

Lombardy, 1995-1997.

METHODS

Analyses were made of prescription records of patients who had at least one respiratory drug dispensed during the study period. Prescriptions for the same individual were linked using a unique identifier.

MAIN OUTCOMES

Number of patients using respiratory drugs; number of visits to the pharmacy; defined daily doses.

MAIN RESULTS

1,387,004 or 15.5% of the resident Lombardy population had at least one dispensed prescription of a respiratory drug in the three years of study. The annual percentage increased from 6.6% in 1995 to 7.9% in 1997. 808,786 patients (58.3% of respiratory drug users) had only one visit to the pharmacy during this period, while the 25.1% of respiratory drug-user (3.9% of the resident population) who had three or more visits to the pharmacy in three years, accounted for 86.1% of all respiratory drug use in terms of defined daily doses. When combinations of prescribed drugs were examined in the light of published guidelines, we estimated that <2.8%> of the population had patterns of drug prescription appropriate for mild intermittent or persistent forms of asthma/chronic obstructive pulmonary disease, and <1.1%> had prescription patterns appropriate for moderate or severe disease.

CONCLUSIONS

The definition of specific drug use patterns by individual can provide prevalence measures that are broad proxies for varying categories of respiratory disease. However, a very high proportion of people who receive respiratory drugs do so for one-off or very infrequent episodes of illness. Analysis of the infrequent and more regular users of respiratory drugs may be useful for comparative research.

The General Practice Research Network: the capabilities of an electronic patient management system for longitudinal patient data

PURPOSE

To evaluate the potential of a general practice research database derived directly from de-identified electronic medical records to provide national prescribing data in Australia. To observe the utilisation of a computer-based patient management system over time. To evaluate the impact of the research network participation on data quality in participants' electronic records.

METHOD

A random sample of 297 general practitioners (GPs) from 128 practices provided longitudinal patient data from electronic medical records (using Medical Director software) retrospectively from 1 January 1999 to April 2002. The General Practice Research Network (GPRN) database contains approximately 600,000 patients, representing over 4 million prescriptions from 4 million encounters. The quality and representativeness of data for prescribing, morbidity and software usage were evaluated by comparison with National data.

RESULTS

Older GPs (> 55) were under-represented, perhaps due to lower computer usage rates, but patient visits were similar to the national distribution. Over time, there were increases in: the quality of prescribing data (with reason for visit/prescription being compulsorily recorded); recording of non-prescribing visits; and the use of other features of the electronic patient management software.

CONCLUSIONS

Data derived from electronic general practice records is of sufficient quality to be used to provide national prescribing estimates and has potential value for pharmacoepidemiology and population health monitoring. Such longitudinal data has previously been unavailable in Australia. Monitoring of software usage demonstrates the evolution of the Australian GP user and will be increasingly useful in assessing and improving the quality of electronic medical records.

Ratio of inhaled corticosteroid to bronchodilator treatment and asthma hospitalisation

BACKGROUND

Previous research based on aggregated data has led to conflicting interpretations of the relationship between the corticosteroid:bronchodilator (C:B) ratio and outcome measures.

OBJECTIVES

To assess whether the C:B ratio is associated with hospital contact for asthma at individual patient level.

METHODS

The study was a retrospective multivariate analysis, using data from the U.K. General Practice Research Database from 1993 to 1996. The subjects were 3465 asthma-diagnosed patients receiving bronchodilator and corticosteroid medication. The main outcome measures were asthma-related hospital contacts.

RESULTS

There was an inverse association between the C:B ratio and hospital contact after controlling for age. The odds ratio for the C:B ratio was 0.87 (95 % CI 0.73-0.98) and 1.04 (95% CI 1.01-1.07) for five-year agebands among patients aged five years and over. There was no systematic relationship between the C:B ratio and hospital contacts for patients aged under five years.

CONCLUSION

The results of this study show that higher C:B ratios are associated with lower levels of hospital contacts at patient level, although there are exceptions possibly linked to disease severity. For patients under five years, the ratio may not be a good outcome measure, perhaps owing to the difficulty in diagnosing asthma or poor compliance.