Record-linkage for pharmacovigilance in Scotland

Identifying key health system components associated with improved outcomes to inform the re-configuration of services for adults with rare autoimmune rheumatic diseases: a mixed-methods study

BACKGROUND

Adults with rare autoimmune rheumatic diseases face unique challenges and struggles to navigate health-care systems designed to manage common conditions. Evidence to inform an optimal service framework for their care is scarce. Using systemic vasculitis as an exemplar, we aimed to identify and explain the key service components underpinning effective care for rare diseases.

METHODS

In this mixed-methods study, data were collected as part of a survey of vasculitis service providers across the UK and Ireland, interviews with patients, and from organisational case studies to identify key service components that enable good care. The association between these components and patient outcomes (eg, serious infections, mortality) and provider outcomes (eg, emergency hospital admissions) were examined in a population-based data linkage study using routine health-care data obtained from patients with antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis from national health datasets in Scotland. We did univariable and multivariable analyses using Bayesian poisson and negative binomial regression to estimate incident rate ratios (IRRs), and Cox proportional hazards models to estimate hazard ratios (HRs). People with lived experiences were involved in the research and writing process.

FINDINGS

Good care was characterised by service components that supported timely access to services, integrated care, and expertise. In 1420 patients with ANCA-associated vasculitis identified from national health datasets, service-reported average waiting times for new patients of less than 1 week were associated with fewer serious infections (IRR 0·70 [95% credibility interval 0·55-0·88]) and fewer emergency hospital admissions (0·78 [0·68-0·92]). Nurse-led advice lines were associated with fewer serious infections (0·76 [0·58-0·93]) and fewer emergency hospital admissions (0·85 [0·74-0·96]). Average waiting times for new patients of less than 1 week were also associated with reduced mortality (HR 0·59 [95% credibility interval 0·37-0·93]). Cohorted clinics, nurse-led clinics, and specialist vasculitis multi-disciplinary team meetings were associated with fewer serious infections (IRR 0·75 [0·59-0·96] for cohorted clinics; 0·65 [0·39-0·84] for nurse-led clinics; 0·72 [0·57-0·90] for specialist vasculitis multi-disciplinary team meetings) and emergency hospital admissions (0·81 [0·71-0·91]; 0·75 [0·65-0·94]; 0·86 [0·75-0·96]). Key components were characterised by their ability to overcome professional tensions between specialties.

INTERPRETATION

Key service components associated with important health outcomes and underpinning factors were identified to inform initiatives to improve the design, delivery, and effectiveness of health-care models for rare autoimmune rheumatic diseases.

FUNDING

Versus Arthritis.

Novel linkage approach to join community-acquired and national data

BACKGROUND

Community optometrists in Scotland have performed regular free-at-point-of-care eye examinations for all, for over 15 years. Eye examinations include retinal imaging but image storage is fragmented and they are not used for research. The Scottish Collaborative Optometry-Ophthalmology Network e-research project aimed to collect these images and create a repository linked to routinely collected healthcare data, supporting the development of pre-symptomatic diagnostic tools.

METHODS

As the image record was usually separate from the patient record and contained minimal patient information, we developed an efficient matching algorithm using a combination of deterministic and probabilistic steps which minimised the risk of false positives, to facilitate national health record linkage. We visited two practices and assessed the data contained in their image device and Practice Management Systems. Practice activities were explored to understand the context of data collection processes. Iteratively, we tested a series of matching rules which captured a high proportion of true positive records compared to manual matches. The approach was validated by testing manual matching against automated steps in three further practices.

RESULTS

A sequence of deterministic rules successfully matched 95% of records in the three test practices compared to manual matching. Adding two probabilistic rules to the algorithm successfully matched 99% of records.

CONCLUSIONS

The potential value of community-acquired retinal images can be harnessed only if they are linked to centrally-held healthcare care data. Despite the lack of interoperability between systems within optometry practices and inconsistent use of unique identifiers, data linkage is possible using robust, almost entirely automated processes.

Causes of death after emergency general surgical admission: population cohort study of mortality

Background

A substantial number of patients treated in emergency general surgery (EGS) services die within a year of discharge. The aim of this study was to analyse causes of death and their relationship to discharge diagnoses, in patients who died within 1 year of discharge from an EGS service in Scotland.

Methods

This was a population cohort study of all patients with an EGS admission in Scotland, UK, in the year before death. Patients admitted to EGS services between January 2008 and December 2017 were included. Data regarding patient admissions were obtained from the Information Services Division in Scotland, and cross-referenced to death certificate data, obtained from the National Records of Scotland.

Results

Of 507 308 patients admitted to EGS services, 7917 died while in hospital, and 52 094 within 1 year of discharge. For the latter, the median survival time was 67 (i.q.r. 21–168) days after EGS discharge. Malignancy accounted for 48 per cent of deaths and was the predominant cause of death in patients aged over 35 years. The cause of death was directly related to the discharge diagnosis in 56.5 per cent of patients. Symptom-based discharge diagnoses were often associated with a malignancy not diagnosed on admission.

Conclusion

When analysed by subsequent cause of death, EGS is a cancer-based specialty. Adequate follow-up and close links with oncology and palliative care services merit development.

Multimorbidity in Antineutrophil Cytoplasmic Antibody-Associated Vasculitis: Results From a Longitudinal, Multicenter Data Linkage Study

OBJECTIVE

Antineutrophil cytoplasmic antibody-associated vasculitis (AAV) is considered a chronic, relapsing condition. To date, no studies have investigated multimorbidity in AAV nationally. This study was undertaken to characterize temporal trends in multimorbidity and report excess health care expenditures associated with multimorbidities in a national AAV cohort from Scotland.

METHODS

Eligible patients with AAV were diagnosed between 1997 and 2017. Each patient was matched with up to 5 general population controls. Linked morbidity and health care expenditure data were retrieved from a Scottish national hospitalization repository and from published national cost data. Multimorbidity was defined as the development of ≥2 disorders. Prespecified morbidities, individually and together, were analyzed for risks and associations over time using modified Poisson regression, discrete interval analysis, and chi-square test for trend. The relationship between multimorbidities and health care expenditure was investigated using multivariate linear regression.

RESULTS

In total, 543 patients with AAV (median age 58.7 years [range 48.9-68.0 years]; 53.6% male) and 2,672 general population controls (median age 58.7 years [range 48.9-68.0 years]; 53.7% male) were matched and followed up for a median of 5.1 years. AAV patients were more likely to develop individual morbidities at all time points, but especially <2 years after diagnosis. The highest proportional risk observed was for osteoporosis (adjusted incidence rate ratio 8.0, 95% confidence interval [95% CI] 4.5-14.2). After 1 year, 23.0% of AAV patients and 9.3% of controls had developed multimorbidity (P < 0.0001). After 10 years, 37.0% of AAV patients and 17.3% of controls were reported to have multimorbidity (P < 0.0001). Multimorbidity was associated with disproportionate increases in health care expenditures in AAV patients. Health care expenditure was highest for AAV patients with ≥3 morbidities (3.89-fold increase in costs, 95% CI 2.83-5.31; P < 0.001 versus no morbidities).

CONCLUSION

These findings emphasize the importance of holistic care in patients with AAV, and may identify a potentially critical opportunity to consider early screening.

Characterizing infection in anti-neutrophil cytoplasmic antibody-associated vasculitis: results from a longitudinal, matched-cohort data linkage study

OBJECTIVES

Infection exerts a major burden in ANCA-associated vasculitis (AAV), however, its precise extent and nature remains unclear. In this national study we aimed to longitudinally quantify, characterize and contextualize infection risk in AAV.

METHODS

We conducted a multicentre matched cohort study of AAV. Complementary data on infections were retrieved via data linkage with the population-based Scottish microbiological laboratory, hospitalization and primary care prescribing registries.

RESULTS

A total of 379 AAV patients and 1859 controls were followed up for a median of 3.5 years (interquartile range 1.9-5.7). During follow-up, the proportions of AAV patients with at least one laboratory-confirmed infection, severe infection and primary care antibiotic prescription were 55.4%, 35.6% and 74.6%, respectively. The risk of infection was higher in AAV than in matched controls {laboratory-confirmed infections: incidence rate ratio [IRR] 7.3 [95% confidence interval (CI) 5.6, 9.6]; severe infections: IRR 4.4 [95% CI 3.3, 5.7]; antibiotic prescriptions: IRR 2.2 [95% CI 1.9, 2.6]}. Temporal trend analysis showed that AAV patients remained at a higher risk of infections throughout the follow-up period, especially year 1. Although the Escherichia genus was the most commonly identified pathogen (16.6% of AAV, 5.5% of controls; P < 0.0001), AAV patients had the highest risk for Herpes [IRR 12.5 (95% CI 3.7, 42.6)] and Candida [IRR 11.4 (95% CI 2.4, 55.4)].

CONCLUSION

AAV patients have up to seven times higher risk of infection than the general population and the overall risk remains significant after 8 years of follow-up. The testing of enhanced short- to medium-term prophylactic antibiotic regimes should be considered.

Predicting the chances of having a baby with or without treatment at different time points in couples with unexplained subfertility

STUDY QUESTION

Can we develop a prediction model that can estimate the chances of conception leading to live birth with and without treatment at different points in time in couples with unexplained subfertility?

SUMMARY ANSWER

Yes, a dynamic model was developed that predicted the probability of conceiving under expectant management and following active treatments (in vitro fertilisation (IVF), intrauterine insemination with ovarian stimulation (IUI + SO), clomiphene) at different points in time since diagnosis.

WHAT IS KNOWN ALREADY

Couples with no identified cause for their subfertility continue to have a realistic chance of conceiving naturally, which makes it difficult for clinicians to decide when to intervene. Previous fertility prediction models have attempted to address this by separately estimating either the chances of natural conception or the chances of conception following certain treatments. These models only make predictions at a single point in time and are therefore inadequate for informing continued decision-making at subsequent consultations.

STUDY DESIGN, SIZE, DURATION

A population-based study of 1316 couples with unexplained subfertility attending a regional clinic between 1998 and 2011.

PARTICIPANTS/MATERIALS, SETTING, METHODS

A dynamic prediction model was developed that estimates the chances of conception within 6 months from the point when a diagnosis of unexplained subfertility was made. These predictions were recomputed each month to provide a dynamic assessment of the individualised chances of conception while taking account of treatment status in each month. Conception must have led to live birth and treatments included clomiphene, IUI + SO, and IVF. Predictions for natural conception were externally validated using a prospective cohort from The Netherlands.

MAIN RESULTS AND THE ROLE OF CHANCE

A total of 554 (42%) couples started fertility treatment within 2 years of their first fertility consultation. The natural conception leading to live birth rate was 0.24 natural conceptions per couple per year. Active treatment had a higher chance of conception compared to those who remained under expectant management. This association ranged from weak with clomiphene to strong with IVF [clomiphene, hazard ratio (HR) = 1.42 (95% confidence interval, 1.05 to 1.91); IUI + SO, HR = 2.90 (2.06 to 4.08); IVF, HR = 5.09 (4.04 to 6.40)]. Female age and duration of subfertility were significant predictors, without clear interaction with the relative effect of treatment.

LIMITATIONS, REASONS FOR CAUTION

We were unable to adjust for other potentially important predictors, e.g. measures of ovarian reserve, which were not available in the linked Grampian dataset that may have made predictions more specific. This study was conducted using single centre data meaning that it may not be generalizable to other centres. However, the model performed as well as previous models in reproductive medicine when externally validated using the Dutch cohort.

WIDER IMPLICATIONS OF THE FINDINGS

For the first time, it is possible to estimate the chances of conception following expectant management and different fertility treatments over time in couples with unexplained subfertility. This information will help inform couples and their clinicians of their likely chances of success, which may help manage expectations, not only at diagnostic workup completion but also throughout their fertility journey.

STUDY FUNDING/COMPETING INTEREST(S)

This work was supported by a Chief Scientist Office postdoctoral training fellowship in health services research and health of the public research (ref PDF/12/06). B.W.M. is supported by an NHMRC Practitioner Fellowship (GNT1082548). B.W.M. reports consultancy for ObsEva, Merck, and Guerbet. None of the other authors declare any conflicts of interest.

Evaluating privacy-preserving record linkage using cryptographic long-term keys and multibit trees on large medical datasets

Background

Integrating medical data using databases from different sources by record linkage is a powerful technique increasingly used in medical research. Under many jurisdictions, unique personal identifiers needed for linking the records are unavailable. Since sensitive attributes, such as names, have to be used instead, privacy regulations usually demand encrypting these identifiers. The corresponding set of techniques for privacy-preserving record linkage (PPRL) has received widespread attention. One recent method is based on Bloom filters. Due to superior resilience against cryptographic attacks, composite Bloom filters (cryptographic long-term keys, CLKs) are considered best practice for privacy in PPRL. Real-world performance of these techniques using large-scale data is unknown up to now.

Methods

Using a large subset of Australian hospital admission data, we tested the performance of an innovative PPRL technique (CLKs using multibit trees) against a gold-standard derived from clear-text probabilistic record linkage. Linkage time and linkage quality (recall, precision and F-measure) were evaluated.

Results

Clear text probabilistic linkage resulted in marginally higher precision and recall than CLKs. PPRL required more computing time but 5 million records could still be de-duplicated within one day. However, the PPRL approach required fine tuning of parameters.

Conclusions

We argue that increased privacy of PPRL comes with the price of small losses in precision and recall and a large increase in computational burden and setup time. These costs seem to be acceptable in most applied settings, but they have to be considered in the decision to apply PPRL. Further research on the optimal automatic choice of parameters is needed.

What Can Big Data Offer the Pharmacovigilance of Orphan Drugs?

The pharmacovigilance of drugs for orphan diseases presents problems related to the small patient population. Obtaining high-quality information on individual reports of suspected adverse reactions is of particular importance for the pharmacovigilance of orphan drugs. The possibility of mining "big data" to detect suspected adverse reactions is being explored in pharmacovigilance generally but may have limited application to orphan drugs. Sources of big data such as social media may be infrequently used as communication channels by patients with rare disease or their caregivers or by health care providers; any adverse reactions identified are likely to reflect what is already known about the safety of the drug from the network of support that grows up around these patients. Opportunities related to potential future big data sources are discussed.

Adherence to Oral Glucose-Lowering Therapies and Associations With 1-Year HbA1c: A Retrospective Cohort Analysis in a Large Primary Care Database

OBJECTIVE

The impact of taking oral glucose-lowering medicines intermittently, rather than as recommended, is unclear. We conducted a retrospective cohort study using community-acquired U.K. clinical data (Clinical Practice Research Database [CPRD] and GoDARTS database) to examine the prevalence of nonadherence to treatment for type 2 diabetes and investigate its potential impact on HbA1c reduction stratified by type of glucose-lowering medication.

RESEARCH DESIGN AND METHODS

Data were extracted for patients treated between 2004 and 2014 who were newly prescribed metformin, sulfonylurea, thiazolidinedione, or dipeptidyl peptidase 4 inhibitors and who continued to obtain prescriptions over 1 year. Cohorts were defined by prescribed medication type, and good adherence was defined as a medication possession ratio ≥0.8. Linear regression was used to determine potential associations between adherence and 1-year baseline-adjusted HbA1c reduction.

RESULTS

In CPRD and GoDARTS, 13% and 15% of patients, respectively, were nonadherent. Proportions of nonadherent patients varied by the oral glucose-lowering treatment prescribed (range 8.6% [thiazolidinedione] to 18.8% [metformin]). Nonadherent, compared with adherent, patients had a smaller HbA1c reduction (0.4% [4.4 mmol/mol] and 0.46% [5.0 mmol/mol] for CPRD and GoDARTs, respectively). Difference in HbA1c response for adherent compared with nonadherent patients varied by drug (range 0.38% [4.1 mmol/mol] to 0.75% [8.2 mmol/mol] lower in adherent group). Decreasing levels of adherence were consistently associated with a smaller reduction in HbA1c.

CONCLUSIONS

Reduced medication adherence for commonly used glucose-lowering therapies among patients persisting with treatment is associated with smaller HbA1c reductions compared with those taking treatment as recommended. Differences observed in HbA1c responses to glucose-lowering treatments may be explained in part by their intermittent use.

Protocol of the Febuxostat versus Allopurinol Streamlined Trial (FAST): a large prospective, randomised, open, blinded endpoint study comparing the cardiovascular safety of allopurinol and febuxostat in the management of symptomatic hyperuricaemia

INTRODUCTION

Gout affects 2.5% of the UK's adult population and is now the most common type of inflammatory arthritis. The long-term management of gout requires reduction of serum urate levels and this is most often achieved with use of xanthine oxidase inhibitors, such as allopurinol. Febuxostat is the first new xanthine oxidase inhibitor since allopurinol and was licensed for use in 2008. The European Medicines Agency requested a postlicensing cardiovascular safety study of febuxostat versus allopurinol, which has been named the Febuxostat versus Allopurinol Streamlined trial (FAST).

METHODS AND ANALYSIS

FAST is a cardiovascular safety study using the prospective, randomised, open, blinded endpoint design. FAST is recruiting in the UK and Denmark. Recruited patients are aged over 60 years, prescribed allopurinol for symptomatic hyperuricaemia and have at least one additional cardiovascular risk factor. After an allopurinol lead-in phase where the dose of allopurinol is optimised to achieve European League against Rheumatism (EULAR) urate targets (serum urate <357 µmol/L), patients are randomised to either continue optimal dose allopurinol or to use febuxostat. Patients are followed-up for an average of 3 years. The primary endpoint is first occurrence of the Anti-Platelet Trialists' Collaboration (APTC) cardiovascular endpoint of non-fatal myocardial infarction, non-fatal stroke or cardiovascular death. Secondary endpoints are all cause mortality and hospitalisations for heart failure, unstable, new or worsening angina, coronary or cerebral revascularisation, transient ischaemic attack, non-fatal cardiac arrest, venous and peripheral arterial vascular thrombotic event and arrhythmia with no evidence of ischaemia. The primary analysis is a non-inferiority analysis with a non-inferiority upper limit for the HR for the primary outcome of 1.3.

ETHICS AND DISSEMINATION

FAST (ISRCTN72443728) has ethical approval in the UK and Denmark, and results will be published in a peer reviewed journal.

TRIAL REGISTRATION NUMBER

FAST is registered in the EU Clinical Trials Register (EUDRACT No: 2011-001883-23) and International Standard Randomised Controlled Trial Number Register (ISRCTN No: ISRCTN72443728).

Prediction of liver disease in patients whose liver function tests have been checked in primary care: model development and validation using population-based observational cohorts

OBJECTIVE

To derive and validate a clinical prediction model to estimate the risk of liver disease diagnosis following liver function tests (LFTs) and to convert the model to a simplified scoring tool for use in primary care.

DESIGN

Population-based observational cohort study of patients in Tayside Scotland identified as having their LFTs performed in primary care and followed for 2 years. Biochemistry data were linked to secondary care, prescriptions and mortality data to ascertain baseline characteristics of the derivation cohort. A separate validation cohort was obtained from 19 general practices across the rest of Scotland to externally validate the final model.

SETTING

Primary care, Tayside, Scotland.

PARTICIPANTS

Derivation cohort: LFT results from 310 511 patients. After exclusions (including: patients under 16 years, patients having initial LFTs measured in secondary care, bilirubin >35 μmol/L, liver complications within 6 weeks and history of a liver condition), the derivation cohort contained 95 977 patients with no clinically apparent liver condition. Validation cohort: after exclusions, this cohort contained 11 653 patients.

PRIMARY AND SECONDARY OUTCOME MEASURES

Diagnosis of a liver condition within 2 years.

RESULTS

From the derivation cohort (n=95 977), 481 (0.5%) were diagnosed with a liver disease. The model showed good discrimination (C-statistic=0.78). Given the low prevalence of liver disease, the negative predictive values were high. Positive predictive values were low but rose to 20-30% for high-risk patients.

CONCLUSIONS

This study successfully developed and validated a clinical prediction model and subsequent scoring tool, the Algorithm for Liver Function Investigations (ALFI), which can predict liver disease risk in patients with no clinically obvious liver disease who had their initial LFTs taken in primary care. ALFI can help general practitioners focus referral on a small subset of patients with higher predicted risk while continuing to address modifiable liver disease risk factors in those at lower risk.

Syndromic surveillance of surgical site infections--a case study in coronary artery bypass graft patients

OBJECTIVE

There is a wealth of data routinely collected and stored by healthcare facilities, which are not consistently exploited for surveillance of healthcare associated infections (HCAI). Syndromic surveillance has not yet been widely applied to HCAI. This study aimed to create syndromic surveillance for surgical site infections (SSI) following coronary artery bypass graft (CABG) procedures.

METHODS

A cohort of CABG patients from Imperial College Healthcare NHS Trust was investigated. Data from the local Patient Administration System, Laboratory Information Management System, radiology department, cardiac registry and Health Protection Agency SSI surveillance were linked. This data was explored for biological markers and proxies of infection, which were used to develop syndromic surveillance algorithms; sensitivity analysis was used to determine the best algorithms.

RESULTS

303 patients were included, with a SSI incidence of 6.6%. Wound culture requests, raised platelet and fibrinogen levels were all found to be good indicators of SSI. Two algorithms were generated, one to detect all SSI (sensitivity: 90%; specificity: 93.8%) and one to detect organ space infections specifically (sensitivity: 100%; specificity: 98.5%).

CONCLUSION

Data which is routinely collected and stored in healthcare facilities can be used for syndromic surveillance of SSI, allowing for an efficient surveillance system without the need for resource intensive data collection.

Views of healthcare professionals to linkage of routinely collected healthcare data: a systematic literature review

Objective

To review the literature on the views of healthcare professionals to the linkage of healthcare data and to identify any potential barriers and/or facilitators to participation in a data linkage system.

Methods

Published papers describing the views of healthcare professionals (HCPs) to data sharing and linkage were identified by searches of Medline, EMBASE, SCOPUS, CINAHL, and PsychINFO. The searches were limited to papers published in the English language from 2001 to 2011.

Results

A total of 2917 titles were screened. From these, 18 papers describing the views of HCPs about data linkage or data sharing of routinely collected healthcare data at an individual patient level were included. Views were generally positive, and potential benefits were reported. Facilitators included having trust in the system including data governance, reliability, and feedback. Some negative views, identified as barriers were also expressed including costs, data governance, technical issues, and privacy concerns. Effects on the physician–patient relationship, and workload were also identified as deterrent.

Discussion

From the published literature included in this review, the views of HCPs were in general positive towards data sharing for public health purposes. The identification of barriers to contributing to a data linkage system allows these to be addressed in a planned data linkage project for pharmacovigilance. The main barriers identified were concerns about costs, governance and interference with the prescriber–patient relationship. These would have to be addressed if healthcare professionals are to support a data linkage system to improve patient safety.

The utility of liver function tests for mortality prediction within one year in primary care using the algorithm for liver function investigations (ALFI)

Background

Although liver function tests (LFTs) are routinely measured in primary care, raised levels in patients with no obvious liver disease may trigger a range of subsequent expensive and unnecessary management plans. The aim of this study was to develop and validate a prediction model to guide decision-making by general practitioners, which estimates risk of one year all-cause mortality in patients with no obvious liver disease.

Methods

In this population-based historical cohort study, biochemistry data from patients in Tayside, Scotland, with LFTs performed in primary care were record-linked to secondary care and prescription databases to ascertain baseline characteristics, and to mortality data. Using this derivation cohort a survival model was developed to predict mortality. The model was assessed for calibration, discrimination (using the C-statistic) and performance, and validated using a separate cohort of Scottish primary care practices.

Results

From the derivation cohort (n = 95 977), 2.7% died within one year. Predictors of mortality included: age; male gender; social deprivation; history of cancer, renal disease, stroke, ischaemic heart disease or respiratory disease; statin use; and LFTs (albumin, transaminase, alkaline phosphatase, bilirubin, and gamma-glutamyltransferase). The C-statistic for the final model was 0.82 (95% CI 0.80–0.84), and was similar in the validation cohort (n = 11 653) 0.86 (0.79–0.90). As an example of performance, for a 10% predicted probability cut-off, sensitivity = 52.8%, specificity = 94.0%, PPV = 21.0%, NPV = 98.5%. For the model without LFTs the respective values were 43.8%, 92.8%, 15.6%, 98.1%.

Conclusions

The Algorithm for Liver Function Investigations (ALFI) is the first model to successfully estimate the probability of all-cause mortality in patients with no apparent liver disease having LFTs in primary care. While LFTs added to the model's discrimination and sensitivity, the clinical utility of ALFI remains to be established since LFTs did not improve an already high NPV for short term mortality and only modestly improved a very low PPV.

Prevalence estimates of diagnosed viral hepatitis B, liver condition outcomes and hospitalization costs: a population record-linkage study in Tayside, Scotland

We estimated prevalence and incidence of liver condition outcomes, and costs to the health service of diagnosed hepatitis B virus (HBV) in Tayside, UK. HBV patients were identified from electronic virology data between 1989 and 2003. The health resource costs of HBV for surface antigen-positive (HBsAg+) patients and HBV (HBsAg+ or immune, i.e. recovered) patients were calculated. A total of 633 patients had HBV (275 HBsAg+), and were more likely to be male (62% vs. 48%), older (mean age 42.6 vs. 39 . 2 years) and deprived than the general population. The prevalence of immune individuals increased steadily. Post-HBV diagnosis, 24% of immune and 13% of HBsAg+ patients were diagnosed with a liver condition. The median cost per immune patient (£3023) was greater than per HBsAg+ patient (£1498) (P=0.02). While increasing prevalence of immune HBV patients highlights an increase in screening and treatment, the costs associated with this group are high.

Record linkage for pharmacoepidemiological studies in cancer patients

BACKGROUND

An increasing need has developed for the post-approval surveillance of (new) anti-cancer drugs by means of pharmacoepidemiology and outcomes research in the area of oncology.

OBJECTIVES

To create an overview that makes researchers aware of the available database linkages in Northern America and Europe which facilitate pharmacoepidemiology and outcomes research in cancer patients.

METHODS

In addition to our own database, i.e. the Eindhoven Cancer Registry (ECR) linked to the PHARMO Record Linkage System, we considered database linkages between a population-based cancer registry and an administrative healthcare database that at least contains information on drug use and offers a longitudinal perspective on healthcare utilization. Eligible database linkages were limited to those that had been used in multiple published articles in English language included in Pubmed. The HMO Cancer Research Network (CRN) in the US was excluded from this review, as an overview of the linked databases participating in the CRN is already provided elsewhere. Researchers who had worked with the data resources included in our review were contacted for additional information and verification of the data presented in the overview.

RESULTS

The following database linkages were included: the Surveillance, Epidemiology, and End-Results-Medicare; cancer registry data linked to Medicaid; Canadian cancer registries linked to population-based drug databases; the Scottish cancer registry linked to the Tayside drug dispensing data; linked databases in the Nordic Countries of Europe: Norway, Sweden, Finland and Denmark; and the ECR-PHARMO linkage in the Netherlands. Descriptives of the included database linkages comprise population size, generalizability of the population, year of first data availability, contents of the cancer registry, contents of the administrative healthcare database, the possibility to select a cancer-free control cohort, and linkage to other healthcare databases.

CONCLUSIONS

The linked databases offer a longitudinal perspective, allowing for observations of health care utilization before, during, and after cancer diagnosis. They create new powerful data resources for the monitoring of post-approval drug utilization, as well as a framework to explore the (cost-)effectiveness of new, often expensive, anti-cancer drugs as used in everyday practice.

Adherence with statins in a real-life setting is better when associated cardiovascular risk factors increase: a cohort study

BACKGROUND

While the factors for poor adherence for treatment with statins have been highlighted, the impact of their combination on adherence is not clear.

AIMS

To estimate adherence for statins and whether it differs according to the number of cardiovascular risk factors.

METHODS

A cohort study was conducted using data from the main French national health insurance system reimbursement database. Newly treated patients with statins between September 1 and December 31, 2004 were included. Patients were followed up 15 months. The cohort was split into three groups according to their number of additional cardiovascular risk factors that included age and gender, diabetes mellitus and cardiovascular disease (using co-medications as a proxy). Adherence was assessed for each group by using four parameters: (i) proportion of days covered by statins, (ii) regularity of the treatment over time, (iii) persistence, and (iv) the refill delay.

RESULTS

16,397 newly treated patients were identified. Of these statin users, 21.7% did not have additional cardiovascular risk factors. Thirty-one percent had two cardiovascular risk factors and 47% had at least three risk factors. All the parameters showed a sub-optimal adherence whatever the group: days covered ranged from 56% to 72%, regularity ranged from 23% to 33% and persistence ranged from 44% to 59%, but adherence was better for those with a higher number of cardiovascular risk factors.

CONCLUSIONS

The results confirm that long-term drug treatments are a difficult challenge, particularly in patients at lower risk and invite to the development of therapeutic education.

Validation of the National Health Insurance Research Database with ischemic stroke cases in Taiwan

OBJECTIVE

The National Health Insurance Research Database (NHIRD) is commonly used for pharmacoepidemiological research in Taiwan. This study evaluated the validity of the database for patients with a principal diagnosis of ischemic stroke.

STUDY DESIGN AND METHODS

This cross-sectional study compares records in the NHIRD with those in one medical center. Patients hospitalized for ischemic stroke in 1999 were identified from both databases. The discharge notes, laboratory data, and medication orders during admission and the first discharge visit were reviewed to validate ischemic stroke diagnoses and aspirin prescribing in the NHIRD. Agreement between the two databases in comorbidities of ischemic stroke diagnosis was evaluated using ICD-9 codes.

RESULTS

Three hundred and seventy two cases were identified from the NHIRD; among them, 364 cases (97.85%) were confirmed as ischemic stroke by radiology examination and clinical presentation. Among these confirmed cases, 344 (94.51%) were assigned 'ischemic stroke' as the principal diagnosis in the NHIRD. The overall agreement of comorbid diagnoses between the databases was 48.39%. The PPV for selected conditions also varied widely, from 0.50 for fracture to 1.00 for colon cancer. The accuracy of recorded aspirin prescriptions was higher in first post-discharge visits (PPV = 0.94) than during hospitalization (PPV = 0.88).

CONCLUSION

The accuracy of the NHIRD in recording ischemic stroke diagnoses and aspirin prescriptions was high, and the NHIRD appears to be a valid resource for population research in ischemic stroke.

Health outcomes following liver function testing in primary care: a retrospective cohort study

BACKGROUND

patients who present with abnormal liver function tests (LFTs) in primary care and no obvious symptoms can be difficult to manage.

OBJECTIVE

The objective is to follow-up a cohort of liver function tested patients to determine their outcome.

METHODS

This population-based retrospective cohort study was conducted in Tayside, Scotland, from 1989 to 2003. Subjects were patients with no clinically obvious liver disease at initial liver function testing in primary care. Main outcomes were diagnosed liver disease and mortality. Record linkage of databases ascertained risk factors and outcomes. Measures of performance were calculated and Weibull regression analysis from initial LFT date was performed on all outcomes by level of abnormality.

RESULTS

In total, 95 977 patients had 364 194 incident initial LFTs, with median follow-up 3.7 years. A total of 21.7% had at least one abnormal LFT and 1108 (1.15%) developed liver disease. Elevated transaminase was strongly associated with diagnosed liver disease, hazard ratio (HR) = 4.23 (95% confidence interval 3.55, 5.04) for mild levels and HR = 12.67 (95% CI 9.74, 16.47) for severe levels versus normal. For gamma-glutamyl transferase, these hazards were 2.54 (95% CI 2.17, 2.96) and 13.44 (95% CI 10.71, 16.87), respectively. Low albumin was strongly associated with all-cause mortality, HR = 2.65 (95% CI 2.47, 2.85) for mild levels and HR = 4.99 (95% CI 4.26, 5.84) for severe levels. Sensitivity for predicting events over 5 years was low and specificity high.

CONCLUSIONS

All LFTs were predictive markers for liver disease as well as general ill health, although sensitivity was poor. Most patients with abnormal LFTs had no later formal diagnosis of liver disease within the study period. The time taken to develop liver disease in these patients provides opportunity to intervene.

Development of a decision support tool to facilitate primary care management of patients with abnormal liver function tests without clinically apparent liver disease [HTA03/38/02]. Abnormal Liver Function Investigations Evaluation (ALFIE)

BACKGROUND

Liver function tests (LFTs) are routinely performed in primary care, and are often the gateway to further invasive and/or expensive investigations. Little is known of the consequences in people with an initial abnormal liver function (ALF) test in primary care and with no obvious liver disease. Further investigations may be dangerous for the patient and expensive for Health Services. The aims of this study are to determine the natural history of abnormalities in LFTs before overt liver disease presents in the population and identify those who require minimal further investigations with the potential for reduction in NHS costs.

METHODS/DESIGN

A population-based retrospective cohort study will follow up all those who have had an incident liver function test (LFT) in primary care to subsequent liver disease or mortality over a period of 15 years (approx. 2.3 million tests in 99,000 people). The study is set in Primary Care in the region of Tayside, Scotland (pop approx. 429,000) between 1989 and 2003. The target population consists of patients with no recorded clinical signs or symptoms of liver disease and registered with a GP. The health technologies being assessed are LFTs, viral and auto-antibody tests, ultrasound, CT, MRI and liver biopsy. The study will utilise the Epidemiology of Liver Disease In Tayside (ELDIT) database to determine the outcomes of liver disease. These are based on hospital admission data (Scottish Morbidity Record 1), dispensed medication records, death certificates, and examination of medical records from Tayside hospitals. A sample of patients (n = 150) with recent initial ALF tests or invitation to biopsy will complete questionnaires to obtain quality of life data and anxiety measures. Cost-effectiveness and cost utility Markov model analyses will be performed from health service and patient perspectives using standard NHS costs. The findings will also be used to develop a computerised clinical decision support tool.

DISCUSSION

The results of this study will be widely disseminated to primary care, as well as G.I. hospital specialists through publications and presentations at local and national meetings and the project website. This will facilitate optimal decision-making both for the benefit of the patient and the National Health Service.

Evaluating medicines: let's use all the evidence

The current drug regulatory system is outdated and relies primarily on a process of premarketing evaluation, followed by periodic reviews of reported adverse events. While long-term medicine use for chronic conditions is now commonplace, current drug evaluation systems do not incorporate the comprehensive evidence accruing over time in clinical practice. Good quality, routinely collected data on medicines use are now available in some countries. Consistent with international opinion, we propose an expanded and integrated system of medicines regulation for Australia, based on a surveillance system that improves safety monitoring by complementing existing systems, making best use of routinely collected data, and leveraging the power of information technology. Australia is well placed to pilot such a model system.

The Influence of Primary Care Prescribing Rates for New Drugs on Spontaneous Reporting of Adverse Drug Reactions

INTRODUCTION

Adverse drug reaction (ADR) reporting makes a vital contribution to pharmacovigilance, although the factors that influence the reporting rate remain unclear. The aim of this study was to investigate whether the variation in the rate of reporting of suspected ADRs in different regions of Scotland was explained by differences in local prescribing practice and to quantify the extent of this influence.

METHODS

Population and primary care prescribing data were obtained for ten geographical areas based on the 15 administrative regions of the National Health Service in Scotland. All reports of suspected ADRs received from within Scotland for 2000 and 2001 were available from the regional monitoring centre (Committee on Safety of Medicines, Scotland). The primary analysis was based on 14 medications that appeared in the 'top ten' list for the frequency of reported ADRs for either year. Reporting rates for each area were expressed both in terms of population (reports per million people) and in terms of estimated exposure to those medications in primary care (reports per 1000 prescriptions). For each analysis, the Pearson correlation coefficient between reporting and prescribing data was calculated using SPSS software.

RESULTS

The 'top ten' medications accounted for 1715 of 2817 (60.9%, 95% CI 59.1, 62.7) ADR reports but only 2.2 million out of a total of 128 million primary care prescriptions (1.7%). Although there was a 3-fold geographical variation in the per-population ADR reporting rate, there was a close correlation between local reporting of ADRs and prescribing of the index medications (p = 0.66, p = 0.04, respectively). This implies that 44% of the observed variation in reporting rate can be attributed to variation in prescribing within the same population.

DISCUSSION

Spontaneous ADR reporting in Scotland over the 2 years studied was highly concentrated on a small number of medications that were under intensive surveillance. Although there was a 3-fold variation in reporting rates from individual geographic areas when corrected for the size of the population, primary care prescribing data showed nearly half of this local variation in reporting rates could be explained by differences in prescribing. This study highlights the importance of considering prescribing practice when interpreting spontaneous ADR reporting data.

The annual incidence of diabetic complications in a population of patients with Type 1 and Type 2 diabetes

AIM

The DARTS diabetes register was used to determine incidence rates of diabetes and related complications in 1997.

METHODS

The diabetes register records detailed clinical information for all patients diagnosed with diabetes in Tayside, Scotland. The study population included patients who were alive and registered with a Tayside GP for the duration of 1997 or who died in Tayside during this time. Patients who had diabetes prior to 1997, those who developed diabetes in 1997, and those who developed diabetic complications in 1997, were identified.

RESULTS

In the Tayside population of 385 774 at the start of 1997, there were 942 and 6632 patients with Type 1 and Type 2 diabetes, with a further 29 and 744 patients diagnosed in 1997. The incidence rates (with 95% confidence intervals) of diabetic complications per 1000 patients with Type 1 and Type 2 diabetes, respectively, were: angina 8.8 (4.5-17.3) and 38.4 (33.4-44.2); myocardial infarction 8.6 (4.4-16.9) and 21.9 (18.4-25.9); cerebrovascular accident 1.1 (0.3-6.0) and 14.2 (11.6-17.5); lower extremity amputation 3.2 (1.2-9.4) and 3.1 (2.1-4.8); peripheral vascular disease 5.5 (2.4-12.8) and 13.6 (11.0-16.8); registered blindness 1.1 (0.3-5.9) and 1.6 (0.9-2.9); end-stage renal failure 6.4 (3.0-13.8) and 5.0 (3.6-7.0). Mortality was 14.6 per 1000 (9.6-25.7) in Type 1 diabetes and 50.0 per 1000 (45.1-55.3) in Type 2 diabetes.

CONCLUSION

This study provides baseline figures for rates of diabetic complications for Type 1 and Type 2 diabetes, and confirms the increased burden of macrovascular disease in Type 2 diabetes.

Effect of ibuprofen on cardioprotective effect of aspirin

Treatment with ibuprofen might limit the cardioprotective effects of aspirin. We aimed to assess whether patients with known cardiovascular disease who take low-dose aspirin and ibuprofen have increased risk of cardiovascular mortality. We studied 7107 patients who were discharged after first admission for cardiovascular disease between April, 1989, and April, 1997, and who were prescribed low-dose aspirin (<325 mg/day) and survived for at least 1 month. Compared with those who used aspirin alone, patients taking aspirin plus ibuprofen had an increased risk of all-cause mortality (adjusted hazard ratio 1.93, 95% CI 1.30-2.87, p=0.0011) and cardiovascular mortality (1.73, 1.05-2.84, p=0.0305). Our finding lends support to the hypothesis that ibuprofen may interact with the cardioprotective effects of aspirin, at least in patients with established cardiovascular disease.

Structure, functions, and activities of a research support informatics section

The authors describe a research group that supports the needs of investigators seeking data from an electronic medical record system. Since its creation in 1972, the Regenstrief Medical Records System has captured and stored more than 350 million discrete coded observations on two million patients. This repository has become a central data source for prospective and retrospective research. It is accessed by six data analysts--working closely with the institutional review board--who provide investigators with timely and accurate data while protecting patient and provider privacy and confidentiality. From January 1, 1999, to July 31, 2002, data analysts tracked their activities involving 47,559 hours of work predominantly for physicians (54%). While data retrieval (36%) and analysis (25%) were primary activities, data analysts also actively collaborated with researchers. Primary objectives of data provided to investigators were to address disease-specific (35.4%) and drug-related (12.2%) questions, support guideline implementation (13.1%), and probe various aspects of clinical epidemiology (5.7%). Outcomes of these endeavors included 117 grants (including 300,000 US dollars per year salary support for data analysts) and 139 papers in peer-reviewed journals by investigators who rated the support provided by data analysts as extremely valuable.

Methods and systems to detect adverse drug reactions in hospitals

Detection of adverse drug reactions (ADRs) in hospitals offers the chance to detect serious ADRs resulting in hospitalisation and ADRs occurring in hospitalised patients, i.e. patients with high comorbidity and receiving drugs that are administered only in hospitals. The most commonly applied methods involve stimulated spontaneous reporting of doctors and nurses, comprehensive collection by trained specialists and, more recently, computer-assisted approaches using routine data from hospital information systems. The different methods of ADR detection used result in different rates and types of ADRs and, consequently, in different drug classes being responsible for these ADRs. Another factor influencing the results of surveys is the interpretation of the term ADR, where some authors adhere to the strict definition of the World Health Organization and many others include intended and unintended poisoning as well as errors in prescribing and dispensing, thus referring to adverse drug events. Depending on the method used for screening of patients, a high number of possible ADRs and only few definite ADRs are found, or vice versa. These variations have to be taken into account when comparing the results of further analyses performed with these data. ADR rates and incidences in relation to the number of drugs prescribed or patients exposed have been calculated in only a few surveys and projects, and this interesting pharmacoepidemiological approach deserves further study. In addition, the pharmacoeconomic impact of ADRs, either resulting in hospitalisation or prolonging hospital stay, has been estimated using different approaches. However, a common standardised procedure for such calculations has not yet been defined. Although detection of ADRs in hospitals offers the opportunity to detect severe ADRs of newly approved drugs, these ADRs are still discovered by spontaneous reporting systems. The prospects offered by electronic hospital information systems as well as implementation of pharmacoepidemiological approaches increases the possibilities and the value of ADR detection in hospitals.

Record-linkage methodology for prescribing research

Record-linkage refers to the linking together of data relating to the same individual from separate source files. In this paper, we discuss ways in which the technique can enhance observational prescribing research in large populations. We draw upon the work of the Medicines Monitoring Unit (MEMO), University of Dundee, to illustrate its contribution to prescribing research.

Effect of diuretics on fetal growth: A drug effect or confounding by indication? Pooled Danish and Scottish cohort data

AIMS

The diabetogenic effect of diuretics, as well as the indication for prescribing them, may impact on fetal growth. We analysed whether the purchase of prescription drugs for diuretics during pregnancy was associated with measures of fetal growth.

METHODS

During 1991-98 all women who purchased prescription drugs for diuretics during pregnancy were identified in the Northern Jutland Prescription Database (NJDP), Denmark, and in the Medicines Monitoring Unit's Database (MEMO), Scotland. Information on birth weight and gestational age was obtained from the Danish Birth Registry, the Danish Hospital Discharge Registry and the Scottish Tayside Neonatal Database. Information on diabetes, hypertension and prepregnancy weight were obtained by hospital record review in a sample of women in the Danish cohort. Women who did not purchase prescription diuretics during pregnancy were used as a reference group in both cohorts.

RESULTS

Danish women who purchased prescription loop diuretics during pregnancy gave birth to infants with higher birth weights than women who did not use diuretics; mean difference 104.7 g (95% CI; 2.6, 206.9). However, the high prevalence of diabetes (10.3%) among Danish women who purchased prescription loop diuretics during pregnancy might explain this result. Both the Danish and the Scottish women who purchased prescription diuretics during their pregnancy were at increased risk of preterm delivery (< 37 completed weeks); ORs: 1.8 (CI; 1.2, 2.7)NJDP, 1.9 (CI; 0.9, 4.3)MEMO. The proportion of hypertension among women who purchased prescription thiazides was 15.8%, and the risk of having an infant with a birth weight (BW) < 2500 g was increased; ORs: 2.6 (CI; 1.4, 5.0)NJDP, 2.4 (CI; 0.8, 7.8)MEMO.

CONCLUSIONS

Prescribing diuretics during pregnancy was associated with differences in birth weight and incidence of preterm delivery. Confounding by indication may explain the findings.

Design issues for drug epidemiology

Despite the difficulties involved in designing drug epidemiology studies, these studies are invaluable for investigating the unexpected adverse effects of drugs. The aim of this paper is to discuss various aspects of study design, particularly those issues that are not easily found in either textbooks or review papers. We have also compared and contrasted drug epidemiology with the randomized controlled trial (RCT) wherever possible. Drug epidemiology is especially useful in the many situations where the RCT is not suitable, or even possible. The study base has to be defined before the appropriate cohort of subjects is assembled. If all of the cases are identified, then a referent sample of controls may be assembled by random sampling of the study base. If all of the cases cannot be assembled, a hypothetical secondary base may need to be created. Preferably, only new-users of the drug should be included, and the risk-ratio will be different for acute users and chronic users. Studies will usually only be possible when researching the unintended effects of drugs. It is difficult to study efficacy because of confounding by indication. In occasional circumstances it may be possible to study efficacy (examples are given). Discussion of the dangers of designing with generalisability in mind is provided. Additionally, the similarities in study design between drug epidemiology and the RCT are discussed in detail, as well as the design-characteristics that cannot be shared between the two methods.

Pharmacovigilance in perspective

Pharmacovigilance is more than spontaneous reporting alone, and the evaluation of marketed medicines is more than just pharmacovigilance. The positioning of a drug usually takes place during the years following introduction, when worldwide experience has accumulated. Originally a modest appendix of drug regulation, pharmacovigilance has become a major activity. The provision of the information needed for the evaluation of the benefits and risks of drugs is in the first place a scientific challenge. In addition, there are important ethical, logistical, legal, financial and commercial constraints. Good pharmacovigilance practice needs to be developed to ensure that data are collected and used in the right way and for the right purpose. Pharmacovigilance, and more generally the study of the benefits and risks of drugs, plays a major role in pharmacotherapeutic decision-making, be it individual, regional, national or international. In addition, pharmacovigilance is becoming a scientific discipline in its own right. A variety of changes are taking place in the complex system of drug development, regulation and distribution. Pharmacovigilance should be proactive in monitoring their possible consequences.