Current Areas of Investigation

CURRENT AREAS OF INVESTIGATION

DRUG SAFETY RESEARCH

Numerous drug safety studies have been completed in MEMO. For example, the cohort study design has been used to evaluate the risk profile of non-steroidal anti-inflammatory drugs (NSAIDs). Although the increased risk of upper gastrointestinal complica-tions associated with NSAID use is well established (Hawkey, 1990), the large number of study subjects and the additional information available in MEMO have allowed more detailed investigations. For exam-ple, a cohort study among 78 191 patients newly exposed to NSAIDs and 78 207 unexposed compara-tors showed that there was an increased risk only among patients without a history of upper gastroin-testinal events (McMahon et al., 1997). Another study in 50 000 subjects investigated the risk with dura-tion of use, and found that it remained constant with continuous exposure (MacDonald et al., 1997) in contrast to previous findings (Carson et al., 1987).

The case–control method is an efficient study design requiring fewer subjects than cohort stud-ies. This is an important consideration when a study involves validating information by checking the original medical notes of patients. The case–control design has been used in a range of studies investi-gating the adverse effect profile of topical NSAIDs. These studies found that oral NSAIDs, but not topical NSAIDs, are implicated in hospitalisation for upper gastrointestinal haemorrhage and perforation (Evans et al., 1995b), acute renal failure (Evans et al., 1995a) and acute colitis (Evans et al., 1997b), but that they are unlikely to be associated with acute appendicitis (Evans et al., 1997c).

The case–crossover design was employed in a study examining the risks of road traffic accidents associ-ated with benzodiazepine use (Barbone et al., 1998). This design is suitable for the evaluation of tran-sient risks, and because cases are used as their own controls, problems of confounding can be dealt with neatly.

DRUG UTILISATION RESEARCH

MEMO is able to produce detailed drug utilisa-tion data, broken down by age, sex, date, day of week prescribed, prescriber, generic or proprietary dispensing, co-prescribing, acute prescribing and/or repeat prescribing, dose and duration. One impor-tant dimension is the audit of GP prescribing in the population, although GP-specific data are analysed anonymously and individual GPs are never identified. For example, one study identified rare instances of potentially hazardous co-prescribing of -antagonists and β-agonists to patients in Tayside likely to have asthma or chronic obstructive airways disease, by linking the dispensed prescribing database to hospital admission records (Hayes et al., 1996). The process-ing of prescribing data according to the demographic characteristics of prescribing GPs has also yielded some useful insights into the characteristics of ‘good’ prescribers. For example, a difference in the prescrib-ing of antibiotics was seen between GP registrar train-ing and non-training practices (Steinke et al., 2000a).

VARIATION IN PRESCRIBING AND MEDICATION COMPLIANCE

Prescribing may vary by patient factors that are inde-pendent of need or disease severity. For example, the variation of use of hormone replacement ther-apy by socio-economic status independent of need (Evans et al., 1997a). Compliance to labelled medi-cation direction or therapy is a related issue. By assessing how patients collect dispensed medication, in terms of numbers of prescriptions dispensed and intervals between them, and linking to outcome data sets, patient compliance or non-compliance to medi-cation can be studied. For example, a study in diabetes showed that adolescents in Tayside who have ‘brittle’ diabetes are often non-compliant with insulin (Morris et al., 1997b).

PHARMACOECONOMICS AND HEALTH RESOURCE USE

Pharmacoepidemiology studies often have a phar-macoeconomic analysis ‘attached’ to the protocol. Both methods have specific objectives that are clearly defined and apparently independent. Pharmacoeco-nomic analyses have become more widely used over the past 10 years. Their primary use is for selecting more efficient drugs; in other words, those exhibit-ing a better relationship between acquisition cost and therapeutic effects and/or economic benefits. Pharma-coeconomic studies use the tools of clinical pharma-cology, epidemiology and economics to obtain data on the effects (beneficial or harmful) of drugs and the costs of treatment alternatives.

MEMO has the ability to identify the drug, type of medication (either generic or proprietary), strength, amount and directions for use and therefore can accu-rately cost the medication for cost analyses. For example, a comparison of the use and cost of self-monitoring reagent strips and patterns of drug use by type 1 and type 2 diabetics was investigated by Evans et al. (1999, 2000). Both studies found a difference between the diabetes type and the cost of medication and health resource use.

CHRONIC DISEASES EPIDEMIOLOGY AND AUDIT

Diabetes Audit and Research in Tayside, Scotland (DARTS)

The MEMO/DARTS collaboration is a joint initia-tive of the Department of Medicine and MEMO at the University of Dundee, together with the Diabetes Units at three Tayside Health Care Trusts (Ninewells Hospital and Medical School, Dundee; Perth Royal Infirmary and Stracathro Hospital, Brechin) and all Tayside GPs with an interest in diabetes care. They have combined their expertise to create the Diabetes Audit and Research in Tayside, Scotland (DARTS) initiative (Morris et al., 1997a). It has been in opera-tion since 1995, continually developing and gathering data from the population base of Tayside.

The MEMO/DARTS collaboration has used elec-tronic record-linkage of information to create a robust clinical information system of all patients with type 1 and type 2 diabetes in Tayside whether they attend primary or secondary care. The DARTS database has information from many different sources including: patients attending hospital diabetes clinics, dispensed prescriptions for diabetes-related medica-tion and monitoring equipment, patients discharged from hospital, patients attending a community-based mobile diabetic eye screening facility, glycosy-lated haemoglobin and plasma glucose results from the regional biochemistry database, and information collected from case records of patients in every general practice in Tayside. The register has been used for pharmaco-epidemiologic research (Morris et al., 1997b,c).

Epidemiology of Liver Disease in Tayside (ELDIT)

The Epidemiology of Liver Disease in Tayside (ELDIT) study group has registered and validated a group of patients with potential and definite liver disease in Tayside for research purposes only. This disease register has a range of liver diseases that affect the whole organ including viral hepatitis (A, B and C) (Steinke et al., 2000b), autoimmune hepati-tis, alcoholic liver disease (Steinke et al., 2000c), primary biliary cirrhosis and hepatocellular carci-noma (Weston et al., 2000) and complications of liver disease like ascites. The ascertainment of liver disease by electronic record-linkage was maximised because of the unique integration of multiple sources of data to create a patient-specific information system. The specificity of virology, immunology and biochemistry tests increases the completeness of the data. Accurate incidence and prevalence rates of liver disease and its complications are used to ensure that hepatology services run effectively and efficiently.

Heart-disease, Evidence-based Audit and Research in Tayside, Scotland (HEARTS)

The latest addition to MEMO’s disease management databases is the HEARTS database of cardiovascular disease in Tayside. This is a regional collaborative effort to support improvements in clinical care, educa-tion and research in cardiovascular disease and to provide GPs with information that will be useful for audit and clinical governance purposes. The database contains information on high-risk patient populations like those who have suffered a myocardial infarction (MI) and those who have undergone coronary angio-plasty or artery bypass grafting (CABG). The database includes a variety of other cardiovascular diseases. For example, those with angina pectoris, peripheral vascu-lar disease, ischaemic stroke, cardiac failure, hyper-tension and those undergoing primary prevention for cardiovascular disease. The aims of HEARTS are to identify and determine the risk factors of cardiovas-cular disease from a population base and to evaluate and determine whether medications are optimised in these patients. This information is fed back in various ways to GP practices in an effort to support them in improving care. HEARTS also provides high quality epidemiological data for research, understanding and care of similar patients and their families.