The Nordic prescription databases as a resource for pharmacoepidemiological research--a literature review.

PURPOSE: All five Nordic countries have nationwide prescription databases covering all dispensed drugs, with potential for linkage to outcomes. The aim of this review is to present an overview of therapeutic areas studied and methods applied in pharmacoepidemiologic studies using data from these databases. METHODS: The study consists of a Medline-based structured literature review of scientific papers published during 2005-2010 using data from the prescription databases in Denmark, Finland, Iceland, Norway, and Sweden, covering 25 million inhabitants. Relevant studies were analyzed in terms of pharmacological group, study population, outcomes examined, type of study (drug utilization vs. effect of drug therapy), country of origin, and extent of cross-national collaboration. RESULTS: A total of 515 studies were identified. Of these, 262 were conducted in Denmark, 97 in Finland, 4 in Iceland, 87 in Norway, and 61 in Sweden. Four studies used data from more than one Nordic country. The most commonly studied drugs were those acting on the nervous system, followed by cardiovascular drugs and gastrointestinal/endocrine drugs. A total of 228 studies examined drug utilization and 263 focused on the effects and safety of drug therapy. Pregnant women were the most commonly studied population in safety studies, whereas prescribers' adherence to guidelines was the most frequent topic of drug utilization studies. CONCLUSIONS: The Nordic prescription databases, with their possibility of record-linkage, represent an outstanding resource for assessing the beneficial and adverse effects of drug use in large populations, under routine care conditions, and with the potential for long-term follow-up.

Prescribing indicators: what can Canada learn from European countries?

BACKGROUND: Drug therapy can improve patients' quality of life and health outcomes; however, underuse, overuse and inappropriate use of drugs can occur. Systematic examination of potential opportunities for improving prescribing and medication use is needed. OBJECTIVE: To convene a diverse group of stakeholders to learn about and discuss advantages and limitations of data sources, tools and methods related to drug prescribing indicators; foster methods to assess safe, appropriate and cost-effective prescribing; increase awareness of international organizations who develop and apply performance indicators relevant to Canadian researchers, practitioners and decision-makers; and provide opportunities to apply information to the Canadian context. METHODS: Approximately 50 stakeholders (health system decision-makers, senior and junior researchers, healthcare professionals, graduate students) met June 1-2, 2009 in Halifax, Canada. Four foundational presentations on evaluating quality of prescribing were followed by discussion in pre-assigned breakout groups of a prepared case (either antibiotic use or prescribing for seniors), followed by feedback presentations. RESULTS: Many European countries have procedures to develop indicators for prescribing and quality use of medicines. Indicators applied in diverse settings across the European Union use various mechanisms to improve quality, including financial incentives for prescribers. CONCLUSION: Further Canadian approaches to develop a system of Canadian prescribing indicators would enable federal/provincial/territorial and international comparisons, identify practice variations and highlight potential areas for improvement in prescribing, drug use and health outcomes across Canada. A more standardized system would facilitate cross-national research opportunities and enable Canada to examine how European countries use prescribing indicators, both within their country and across the European Union.

Differences in antibiotic prescribing in three university hospitals in the Baltic region revealed by a simple protocol for quality assessment of therapeutic indications.

OBJECTIVE: To identify inexpensive and simple quality parameters for the surveillance of antibiotic use in hospital settings. METHODS: A modified point-prevalence study was conducted in three university hospitals in Huddinge, Sweden, Riga, Latvia, and Vilnius, Lithuania. Each ward was visited once during May in the year 2002. All patients receiving antibiotics were identified and their medical records were reviewed by the authors according to the same protocol. Only data from corresponding departments were evaluated and compared. RESULTS: The prevalence of antibiotic use was 35%, 25% and 24% in Huddinge, Riga and Vilnius, respectively. Almost 2/3 of antibiotics were prescribed for treatment and 1/3 for either surgical or medical prophylaxis. Parenteral administration was significantly more common in Riga and Vilnius than in Huddinge. The most commonly prescribed antibiotics were cephalosporins and fluoroquinolones. Prescription of antibiotics for different diagnoses showed large variation between and within hospitals. The first or second generation cephalosporins were prescribed in most cases of surgical prophylaxis. The duration of surgical prophylaxis exceeded one day in 57%, 63% and 87% of cases in Huddinge, Riga and Vilnius, respectively. All antibiotics in Huddinge, and all except five in Riga were supplied by the hospital pharmacy. Antibiotics bought by patients and donated made up 41% of prescribed antibiotics in Vilnius. CONCLUSION: This point-prevalence survey using a simple and inexpensive method for benchmarking demonstrated quantitative and qualitative differences in the use of antibiotics between three university hospitals in the Baltic region, differences that now calls for explanations to their rationality. We suggest that the choice of an antibiotic, rates of intravenously administered treatment and duration of surgical prophylaxis are examples of suitable indicators of rational antibiotic use within a hospital but that comparison of such rates between hospitals is less meaningful.

Antimicrobial use at a university hospital: appropriate or misused? A qualitative study.

OBJECTIVE: To evaluate the quality of antimicrobial drug use in a university hospital medical department (Department of Medicine, University Hospital Rijeka, Croatia) with 279 hospital-beds in wards containing patients from endocrinology, gastroenterology, hematology, clinical immunology, cardiology and coronary care unit, nephrology and pulmonology sections of the hospital. METHODS: The appropriateness of antimicrobial treatment for all in-patients in the Department of Medicine was assessed in a prospective, longitudinal survey carried out during a 21-week period using Kunin's criteria where Categories I and II indicate "appropriate therapy", Categories III and IV indicate major deficiency in the choice or use of antimicrobials. Category V indicates unjustified antimicrobial administration. RESULTS: During the study period, a total of 438 patients were treated with antimicrobials in the Department of Medicine. Of these, 159 (36%) received antimicrobials appropriately (Category I and II), 180 (41%) needed antimicrobials (Category III and IV) but they should have been prescribed differently. The main reason for inappropriate antimicrobial treatment was the wrong choice of antimicrobials (broad-spectrum where a narrow spectrum antibiotic would have been sufficient). In the case of 99 patients (23%) an indication for antimicrobial therapy did not exist (Category V). CONCLUSION: The main reason for suboptimal use of antimicrobials was the over-prescribing of broad-spectrum antimicrobials. This situation should be corrected e.g. by changes in the post-graduate medical teaching program.

Benchmarking antimicrobial drug use at university hospitals in five European countries.

A point-prevalence survey of five European university hospitals was performed to benchmark antimicrobial drug use in order to identify potential problem areas in prescribing practice and to aid in establishing appropriate and attainable goals. All inpatients at the university hospitals of Rijeka (Croatia), Tartu (Estonia), Riga (Latvia), Vilnius (Lithuania) and Karolinska-Huddinge (Sweden) were surveyed for antimicrobial drug use during a single day. The frequency of antimicrobial drug use was 24% in Rijeka, 30% in Tartu, 26% in Riga, 14% in Vilnius and 32% in Huddinge. Surgical patients were treated with antimicrobial agents more often than medical patients in Riga (53% vs. 31%), Tartu (39% vs. 26%) and Vilnius (54% vs. 25%). Two-thirds of patients in Rijeka, Tartu, Riga and Vilnius, and fewer than half of the patients in Huddinge, received antimicrobial agents intravenously. Broad-spectrum antimicrobial agents were used most commonly in Rijeka. The prevalence of nosocomial infections treated with antibiotics was 9% at Huddinge, and 3-5% at the other centres. Benchmarking antimicrobial drug use at five university hospitals identified differences and problem areas. The high rates of intravenous administration, poor compliance with guidelines, and prolonged surgical prophylaxis were general problems that deserved specific attention at all centres. A change in prescription practices may reduce unnecessary drug use and decrease antimicrobial resistance.

Healthcare associated infections in university hospitals in Latvia, Lithuania and Sweden: a simple protocol for quality assessment.

Surveillance of healthcare associated infections is an overlooked parameter of good clinical practice in most healthcare institutions, due to the workload demanded in the absence of adequate IT-systems. The aim of the present study was to investigate whether a simple protocol could be used to estimate the burden of healthcare associated infections in three university hospitals in Huddinge in Sweden, Riga in Latvia and Vilnius in Lithuania and form the basis for initiating a long term follow up system. The medical records of all patients receiving antibiotics were reviewed according to a standardised protocol, focusing on the indications for the drugs and on the frequency of hospital acquired infection (HAI) in a point-prevalence survey. Only comparable specialties were included. The proportion of patients treated with antibiotics (prophylaxis not included) were 63/280 (22%) in Huddinge, 73/649 (11%) in Riga and 99/682 (15%) in Vilnius. The proportion of admitted patients treated for a HAI were 15%, 3% and 4%, respectively, (both comparisons Huddinge versus other centres P <0.001). Surgical site infections were most common, followed by infections with an onset more than 2 days after admission without any of the other registered risk factors present. Our inexpensive and simple method showed that healthcare associated infections were a significant problem among patients admitted to Huddinge. The figures obtained can be used for further discussion and form a baseline for follow up at the local level. The comparison of figures between centres was far less relevant than the process the study created.

Healthcare associated infections in university hospitals in Latvia, Lithuania and Sweden: a simple protocol for quality assessment.

Surveillance of healthcare associated infections is an overlooked parameter of good clinical practice in most healthcare institutions, due to the workload demanded in the absence of adequate IT-systems. The aim of the present study was to investigate whether a simple protocol could be used to estimate the burden of healthcare associated infections in three university hospitals in Huddinge in Sweden, Riga in Latvia and Vilnius in Lithuania and form the basis for initiating a long term follow up system. The medical records of all patients receiving antibiotics were reviewed according to a standardised protocol, focusing on the indications for the drugs and on the frequency of hospital acquired infection (HAI) in a point-prevalence survey. Only comparable specialities were included. The proportion of patients treated with antibiotics (prophylaxis not included) were 63/280 (22%) in Huddinge, 73/649 (11%) in Riga and 99/682 (15%) in Vilnius. The proportion of admitted patients treated for a HAI were 15%, 3% and 4%, respectively, (both comparisons Huddinge versus other centres P<0.001). Surgical site infections were most common, followed by infections with an onset more than 2 days after admission without any of the other registered risk factors present. Our inexpensive and simple method showed that healthcare associated infections were a significant problem among patients admitted to Huddinge. The figures obtained can be used for further discussion and form a baseline for follow up at the local level. The comparison of figures between centres was far less relevant than the process the study created.

Differential effects of olfactory toxicants on olfactory regeneration.

The aim was to study the long-term response in the olfactory mucosa of NMRI mice after exposure to the olfactory toxicants dichlobenil (a herbicide) or methimazole (an antithyroid drug). Three and six months after exposure to dichlobenil (2x or 1 x 25 mg/kg i.p.), the dorsomedial part of the olfactory region showed a respiratory metaplasia with abundant invaginations and a fibrotic lamina propria. In contrast, 3 months after exposure to a toxic dose of methimazole (2 x 50 mg/kg i.p.), the olfactory neuroepithelium and lamina propria had been restored. To study the regenerative events, we used an antibody derived against growth-associated protein 43 (GAP-43), which stains immature neurons. To study epithelial differentiation and horizontal basal cells (HBCs) we used an antibody derived against some cytokeratins. Two weeks after methimazole treatment, there was a marked increase of GAP-43-stained cells in the whole olfactory region, which correlated with the observed regeneration at that time. Two weeks after dichlobenil treatment, the damaged atypical epithelium in the olfactory region showed a distinct keratin staining of basal and columnar cells whereas GAP-43-stained cells were not found. Despite a transient increase of GAP-43-stained cells in the border zone between damaged and undamaged olfactory mucosa, an expansion of a normal neuroepithelium into the damaged olfactory region was not detected in the dichlobenil-treated mice. An intact lamina propria is suggested as a prerequisite for repopulation of the neuroepithelium after toxicant-induced injury.

[How many prescription drugs does a general practitioner handle? A prescription database study]. / Hvor mange forskellige laegemidler håndterer en praktiserende laege? En receptdatabase-undersøgelse.

INTRODUCTION: General practitioners must know of treatments for a broad range of diseases, but it may be difficult to keep abreast of the effects and adverse reactions of a large spectrum of drugs, and a good quality of prescribing implies the use of a limited number of analogue products. The aims were to investigate the number of different drugs prescribed per practice and to analyse the influence of practice characteristics on this number. MATERIAL AND METHODS: The study comprised 173 general practices (99 single and 74 group practices in the County of Funen, Denmark. All prescription registered on the Odense Pharmaco-epidemiological Database (OPED) were analysed. The main outcome measure was the number of different drugs prescribed per dispensing unit. RESULTS: The number of different drugs prescribed per dispensing unit varied almost fourfold (range: 102 to 381) and four practice characteristics were able to predict 74% of this variation. Practices with several doctors, a high number of patients listed per doctor, a high percentage of elderly patients, and a heavy workload showed the highest number of different drugs prescribed. CONCLUSION: There is almost a fourfold variation in the number of different drugs used in general practice, and three-quarters of this variation is associated with factors related to the practice.

Antiepileptic drug therapy and its management in sudden unexpected death in epilepsy: a case-control study.

PURPOSE: Because frequent seizures constitute a major risk factor for sudden unexpected death in epilepsy (SUDEP), the treatment with antiepileptic drugs (AEDs) may play a role for the occurrence of SUDEP. We used data from routine therapeutic drug monitoring (TDM) to study the association between various aspects of AED treatment and the risk of SUDEP. METHODS: A nested case-control study was based on a cohort consisting of 6,880 patients registered in the Stockholm County In Ward Care Register with a diagnosis of epilepsy. Fifty-seven SUDEP cases, and 171 controls, living epilepsy patients, were selected from the cohort. Clinical data including data on TDM were collected through medical record review. RESULTS: The relative risk (RR) of SUDEP was 3.7 (95% CI, 1.0-13.1) for outpatients who had no TDM compared with those who had one to three TDMs during the 2 years of observation. RR was 9.5 (1.4-66.0) if carbamazepine (CBZ) plasma levels at the last TDM were above and not within the common target range (20-40 microM). High CBZ levels were associated with a higher risk in patients receiving polytherapy and in those with frequent dose changes. Although the subgroup of patients with high CBZ levels was small (six cases of 33 with CBZ therapy), and the result should be interpreted with caution, no similar associations were demonstrated for phenytoin plasma levels and risk of SUDEP. No association was found between SUDEP risk and within-patient variation in AED levels over time. CONCLUSIONS: Polytherapy, frequent dose changes, and high CBZ levels as identified risk factors for SUDEP all point to the risks associated with an unstable severe epilepsy. It is unclear whether high CBZ levels per se represent a risk factor or just reflect other unidentified aspects of a severe epilepsy. Our results, however, prompt further detailed analyses of the possible role of AEDs in SUDEP in larger cohorts and suggest that reasonable monitoring of the drug therapy may be useful to reduce risks.

[Feedback on prescribing profiles at a primary health center. Important element in quality assurance of drug prescription]. / Aterföring av förskrivningsprofiler vid en vårdcentral. Viktigt inslag i kvalitetssäkringen av läkemedelsförskrivningen.

DU90%--the number of drugs that account for 90% of DDDs--and adherence to guidelines in this segment were proposed by the Swedish Medical Quality Council (MKR) as indicators for assessing the quality of drug prescribing. We tested these indicators at a primary health care (PHC) center in Stockholm. Bar-coded prescriptions purchased at pharmacies were compared with the guidelines issued by the regional drug committee. The data were presented and discussed at the PHC center. Although the DU90% method neither examines the appropriateness of use nor provides outcome data, it was shown to be an inexpensive, flexible and simple method for assessing the general quality of drug prescribing.

[Quality assurance of drug prescription in primary health care. A new database software makes the drug therapy surveillance easier]. / Kvalitetsarbete kring läkemedelsförskrivning i primärvården. Nytt databasprogram underlättar uppföljning av läkemedelsbehandling.

Quality assurance of drug prescription is a pre-requisite for rational drug use. From 22 health-care centres in the south-western area of the Stockholm County Council region, drug-prescription data were obtained from the patients' computerised medical recordings. This could be done with the aid of a specially designed database program. The drug-prescription data from the 22 health-care centres were collected and compiled in a central unit. Thereafter the results were brought back to the health-care centres, in which the quality assurance of drug prescription could be started.

A study of benzodiazepine users claiming drug-induced psychiatric morbidity.

To ascertain whether benzodiazepines (BZDs) cause behavioural side effects other than those already known or trigger latent mental illness processes, the medical records of 32 patients who had submitted claims for BZD-induced side effects to the Swedish Pharmaceutical Insurance (SPI) claims office during 1985-92 were re-assessed by two psychiatrists on the basis of the DSM-IV system to determine whether a psychiatric syndrome existed before, during, or after the prescription of BZDs. Of the 32 case reports entered into the study, 22 concerned women; mean age at the time of SPI claim submission was 47 years, and the mean duration of BZD treatment was 11.7 years. The BZD dosages were in most cases within recommended therapeutic limits. Twenty-five claimants had also been using other psychotropic drugs. In 28 of the 32 cases a psychiatric syndrome in accordance with DSM-IV was present before the start of BZD treatment-in most cases an anxiety syndrome or anxious personality trait. In the other four cases no diagnosis could be given owing to paucity of information. In the 12 cases of claimants who fulfilled BZD dependence criteria at some point during treatment, withdrawal syndromes might have affected the clinical picture to some extent. Four claimants received additional post-BZD-treatment Axis I diagnoses, but in all four cases the new diagnoses were clearly related to the pre-existing symptom pattern. Thus, apart from the dependence on the drug there was no evidence of psychiatric morbidity caused by the BZDs. The symptoms reported by the claimants as evidence of BZD-induced psychiatric morbidity seem in most cases to have been a feature of pre-existing psychopathology which became more manifest after discontinuation of BZD treatment.

Marked differences in antibiotic use and resistance between university hospitals in Vilnius, Lithuania, and Huddinge, Sweden.

Antibiotic use and antimicrobial resistance was compared between Vilnius and Huddinge University hospitals. Drug use data were expressed in number of defined daily doses/100 bed-days; antimicrobial resistance were given as percentages of resistant isolates. Thirty-five and 48 different antibiotic drugs were used in Vilnius and Huddinge, respectively. The overall consumption of antibiotics was 15 DDD/100 bed-days in Vilnius and 43 DDD/100 bed-days in Huddinge. Benzylpenicillin, ampicillin, and aminoglycosides were the major antibiotics in Vilnius; beta-lactamase-resistant penicillins, cefalosporins, and quinolones in Huddinge. In Vilnius, gentamicin made up one-quarter of the use. Staphylococcus aureus and Gram-negative isolates from wounds and blood were more resistant to gentamicin in Vilnius. S. aureus was more often methicillin resistant in Vilnius than in Huddinge. There was no S. aureus-resistant to vancomycin in either hospital. The vancomycin-resistant enterococci made up from 4% to 10% in Vilnius hospital, but they were not detected in Huddinge hospital (0%). The majority of Streptococcus pneumoniae isolates were sensitive to benzylpenicillin in both hospitals. The higher resistance of microorganisms to some antibiotics in Vilnius may be explained by heavy use of few antibiotics. Lower level of hygiene procedures, sampling bias, and other methodological issues may also have contributed. Guidelines for antibiotic use and hygienic procedures are now under development in Vilnius.

Deviations from evidence-based prescribing of non-steroidal anti-inflammatory drugs in three European regions.

OBJECTIVE: We examined to what extent the evidence of the relative gastrointestinal toxicity with non-steroidal anti-inflammatory drugs (NSAIDs) was implemented in clinical practice in Bologna, Italy, Funen, Denmark, and Stockholm, Sweden, areas with accurate computerised information on prescriptions purchased by defined populations. METHODS: We ranked each NSAID by purchased volume in defined daily doses during September 1996 and compared it with the ranking of gastrointestinal complications from a meta-analysis of controlled epidemiological studies published between 1986 and 1994. We restricted our comparison to those NSAIDs that accounted for 90% of the use and within this DU90% segment we determined the proportion of "high risk" (azapropazone, ketoprofen, piroxicam) and "low risk" (ibuprofen, diclofenac) drugs with respect to gastrointestinal toxicity. RESULTS: In Funen, Denmark, we found the best NSAID profile (63% low risk/11% high risk) while Bologna, Italy, had the other extreme (26% low risk/38% high risk), with Stockholm, Sweden, in between (43% low risk/20% high risk). CONCLUSION: Our study suggests that factors other than evidence-based medicine had a dominating impact on the use of prescription NSAIDs in 1996.

Transfer of dopamine in the olfactory pathway following nasal administration in mice.

PURPOSE: The aim of the study was to investigate whether dopamine is transferred along the olfactory pathway to the brain following nasal administration to mice. METHODS: [3H]-Dopamine was administered nasally or intravenously to female mice. Brain tissue samples were excised and the radioactive content was measured. The precise localisation of dopamine radioactivity in the brain was studied using autoradiography. The presence of dopamine or its metabolites in the olfactory bulb and mucosa was ascertained using thin layer chromatography (TLC). RESULTS: After administration of [3H]-dopamine into the right nostril, the amount of dopamine in the right bulb increased with time until. after 4 h, it was 27 times higher than in the left bulb. Among the other brain tissue samples, significantly higher amount of radioactivity was detected in the lateral olfactory tract. Radioactivity in the right olfactory bulb was shown by autoradiography to be selectively located in the peripheral layers 1 to 4 h after administration. Selective uptake of radioactivity was not seen in other regions of the brain. TLC data indicated that approximately 75% and 10% of the radioactivity in the olfactory bulb and mucosa, respectively, coeluted with dopamine. CONCLUSIONS: The results indicate that unchanged dopamine is transferred into the olfactory bulb following nasal administration of [3H]-dopamine.

Wide variation in the number of different drugs prescribed by general practitioners. A prescription database study.

OBJECTIVES: To investigate the number of different drugs prescribed per dispensing unit and to analyse the influence of practice characteristics on this number. DESIGN: Register analysis based on the Odense Pharmacoepidemiological Database (OPED). SETTING: General practices in the County of Funen, Demnark. SUBJECTS: 173 general practices (99 single-handed and 74 group practices). MAIN OUTCOME MEASURES: Number of different drugs prescribed per dispensing unit. RESULTS: The number of different drugs prescribed per dispensing unit varied nearly fourfold (range 102-381) and four practice characteristics were able to predict 74% of this variation. Practices with several doctors, a high number of patients listed per doctor, a high percentage of elderly individuals, and a heavy workload showed the highest number of different drugs prescribed. CONCLUSION: As the quality of drug prescribing is associated with the use of a limited number of drugs, it is suggested that GPs should agree on a formulary containing the most essential drugs in primary health care.

Workplace drug testing (WDT) likely to increase in Europe. Report from the First European Symposium on WDT including selected abstracts.

Will it take a series of drug-related accidents that have already occurred in the USA before workplace drug testing (WDT) becomes accepted in Europe as a preventive measure? Currently, the development of WDT in most European countries lags some 10-15 years behind that in the USA. Labour authorities in Europe now ought to take initiatives to demand a mandatory programme for accrediting drug analytical laboratories for WDT. Companies should realise that illicit drug use is no longer only a problem at street corners, and that having a testing system in place is important, not just for public health, but also for their reputations as responsible societal actors. Improved networking among police and regulatory authorities is required to keep pace with the rapid appearance and dissemination of new substances of abuse. European research collaboration, including the newly formed European Workplace Drug Testing Group, is needed to assess the impact of drug-testing policies on accidents and other outcome variables, and thereby to convince the general public and politicians that drug testing is beneficial and necessary. A 1993-1994 survey of quality analysis in some 200 European laboratories reported from Institut Municipal d'Investigació Medica (IMIM), Spain, showed good agreement between nominal and found concentrations but that only 10% of the laboratories could both screen, identify and quantify samples. Experiences from Italy show that proficiency testing schemes lead to improved accuracy of results. These were some major conclusions of the First European Symposium on Drug Testing held at Huddinge University Hospital in Stockholm, Sweden, 30 March to 1 April 1998, organised by Karolinska Institute, with participants from 22 countries.

Lack of adherence to lipid-lowering drug treatment. A comparison of utilization patterns in defined populations in Funen, Denmark and Bologna, Italy.

AIMS: The objective was to explore differences in lipid-lowering drug (LLD) prescribing in Italy and Denmark. METHODS: We used two geographical areas with computerized drug prescription records in defined populations, one in Funen, Denmark with 500 000 inhabitants, the other in Bologna, Italy with 400 000 inhabitants. Prescriptions for patients who had purchased a LLD from 1994 until 1996 were retrieved as well as coprescriptions of antidiabetic and cardiovascular drugs as markers for diabetes and cardiovascular disease. Only patients surviving and remaining in the area were included. Compliance was defined as percentage of DDDs purchased divided by the number of days within the time window. The limit between good and poor compliance was set at 82%. RESULTS: In Bologna, LLD consumption measured in DDD increased by 41% and in Funen by 129%. Annual prevalence increased from 36.9 to 46.3 users/1000 inhabitants from 1994 to 1996 and from 3.2 to 6.6 users/1000 inhabitants in Bologna and Funen, respectively. From 1995 to 1996, the incidence of use decreased slightly in Bologna from 19.3 to 18. 8/1000 inhabitants/year, whereas in Funen the incidence increased from 1.8 to 2.3/1000 inhabitants/year. In Bologna 48% and in Funen 91% of users persisted with treatment for 2 years or longer. In Bologna, 7% and in Funen 45% were good compliers. In Bologna, 61% and in Funen, 72% received other drugs indicating cardiovascular or diabetic comorbidity. CONCLUSIONS: Patterns of use differed substantially between the two areas. In contrast with Funen, where long-term use was common, Bologna LLD use was sporadic. Based on a higher rate of coprescription, LLDs seemed to be used for secondary prevention to a higher extent in Funen than in Bologna. In Funen it appeared that the correct patients, but an insufficient number of them, were being treated adequately according to guidelines. The higher discontinuation rate of lipid lowering drugs in the Bologna area indicates that a large proportion of patients use these drugs for too short a period of time to benefit from treatment. Since society's health care resources are limited it is difficult to justify public funding of these medications without at the same time giving appropriate attention to these problems.