Using secondary data sources for pharmacoepidemiology and outcomes research

Diagnostic validity and solute-corrected prevalence for hyponatremia and hypernatremia among 1 813 356 admissions

Background and hypothesis

We aimed to evaluate the diagnostic validity of the International Classification of Diseases, 10th Revision (ICD-10) codes for hyponatremia and hypernatremia, using a database containing laboratory data. We also aimed to clarify whether corrections for blood glucose, triglyceride, and total protein may affect the prevalence and the diagnostic validity.

Methods

We retrospectively identified admissions with laboratory values using a Japanese hospital-based database. We calculated the sensitivity, specificity, and positive/negative predictive values of recorded ICD-10-based diagnoses of hyponatremia (E87.1) and hypernatremia (E87.2), using serum sodium measurements during hospitalization (<135 and >145 mmol/l, respectively) as the reference standard. We also performed analyses with corrections of sodium concentrations for blood glucose, triglyceride, and total protein.

Results

We identified 1 813 356 hospitalizations, including 419 470 hyponatremic and 132 563 hypernatremic cases based on laboratory measurements, and 18 378 hyponatremic and 2950 hypernatremic cases based on ICD-10 codes. The sensitivity, specificity, positive predictive value, and negative predictive value of the ICD-10 codes were 4.1%, 99.9%, 92.5%, and 77.6%, respectively, for hyponatremia and 2.2%, >99.9%, 96.5%, and 92.8%, respectively, for hypernatremia. Corrections for blood glucose, triglyceride, and total protein did not largely alter diagnostic values, although prevalence changed especially after corrections for blood glucose and total protein.

Conclusions

The ICD-10 diagnostic codes showed low sensitivity, high specificity, and high positive predictive value for identifying hyponatremia and hypernatremia. Corrections for glucose or total protein did not affect diagnostic values but would be necessary for accurate prevalence calculation.

Stroke Research Using Administrative Claims Database in Japan: A Narrative Review

AIMS

Although administrative claims databases have recently been used for clinical research in Japan, no detailed description of their utilization in stroke research is available. We reviewed stroke studies using the Diagnosis Procedure Combination (DPC), the National Database of Health Insurance Claims and Specific Health Checkups (NDB), and several commercial databases sourced from social health insurance associations, focusing on their applications and limitations.

METHODS

Original articles on stroke published by April 2024 using the DPC, NDB, and commercial databases were identified in Ovid MEDLINE. The characteristics of each database were compared in terms of comprehensiveness, traceability, baseline information, and outcome assessment in stroke research.

RESULTS

A total of 114 studies were included (83 for DPC, 6 for NDB, and 25 for commercial databases). The number of stroke studies using administrative databases in Japan is still approximately 10 per year, although there is a slowly increasing trend. The DPC database was utilized for short-term outcome studies because of its detailed baseline and outcome information, although the inability to track patients once they changed facilities limits their use in long-term studies. The NDB database is potentially useful for long-term studies because of its comprehensiveness and traceability, but difficulties in data access restrict its usage. The most commonly used commercial database utilizes baseline information on lifestyle and blood test data, although the lack of coverage for those over 75 years old may limit its generalizability.

CONCLUSIONS

Administrative claims databases are beginning to be used in stroke research in Japan but are not yet fully utilized. Researchers need to understand their applications and limitations.

Assessing the Benefits and Harms of Pharmacotherapy in Older Adults with Frailty: Insights from Pharmacoepidemiologic Studies of Routine Health Care Data

The objective of this review is to summarize and appraise the research methodology, emerging findings, and future directions in pharmacoepidemiologic studies assessing the benefits and harms of pharmacotherapies in older adults with different levels of frailty. Older adults living with frailty are at elevated risk for poor health outcomes and adverse effects from pharmacotherapy. However, current evidence is limited due to the under-enrollment of frail older adults and the lack of validated frailty assessments in clinical trials. Recent advancements in measuring frailty in administrative claims and electronic health records (database-derived frailty scores) have enabled researchers to identify patients with frailty and to evaluate the heterogeneity of treatment effects by patients' frailty levels using routine health care data. When selecting a database-derived frailty score, researchers must consider the type of data (e.g., different coding systems), the length of the predictor assessment period, the extent of validation against clinically validated frailty measures, and the possibility of surveillance bias arising from unequal access to care. We reviewed 13 pharmacoepidemiologic studies published on PubMed from 2013 to 2023 that evaluated the benefits and harms of cardiovascular medications, diabetes medications, anti-neoplastic agents, antipsychotic medications, and vaccines by frailty levels. These studies suggest that, while greater frailty is positively associated with adverse treatment outcomes, older adults with frailty can still benefit from pharmacotherapy. Therefore, we recommend routine frailty subgroup analyses in pharmacoepidemiologic studies. Despite data and design limitations, the findings from such studies may be informative to tailor pharmacotherapy for older adults across the frailty spectrum.

A Review of Studies Using Japanese Nationwide Administrative Claims Databases

BACKGROUND

Administrative claims databases are increasingly being used worldwide for research purposes. We reviewed original published articles that used one of the four nationwide administrative claims databases in Japan: the National Database of Health Insurance Claims and Specific Health Checkups (NDB), NDB Open Data, the JMDC Claims Database, and the Diagnosis Procedure Combination (DPC) database.

METHODS

Studies published from January 2010 to July 2022 using the JMDC and DPC databases, and from January 2013 to July 2022 using the NDB and NDB Open Data were identified using PubMed. The number of original articles was divided into 19 fields. The annual growth rate of the number of studies was calculated using the four databases.

RESULTS

Overall, 1047 studies were included (95 for the NDB, 31 for the NDB Open Data, 222 for the JMDC database, and 699 for the DPC databases). Studies using one of these four databases increased from around 2010, and the average annual growth rate was approximately 41% from 2010 to 2021. DPC database studies had a higher proportion of articles on surgery (19.2%), urology (3.0%), neurosurgery (6.2%), anesthesiology (1.9%), and emergency medicine (14.0%), whereas the NDB and JMDC data had higher proportions of those regarding internal medicine.

CONCLUSIONS

Since 2010, these four databases have increasingly attracted attention, and the number of studies using them has grown rapidly. Our review suggests that each has unique features, and researchers should understand the database characteristics to operate their studies.

Rationale and performances of a data-driven method for computing the duration of pharmacological prescriptions using secondary data sources

The assessment of the duration of pharmacological prescriptions is an important phase in pharmacoepidemiologic studies aiming to investigate persistence, effectiveness or safety of treatments. The Sessa Empirical Estimator (SEE) is a new data-driven method which uses k-means algorithm for computing the duration of pharmacological prescriptions in secondary data sources when this information is missing or incomplete. The SEE was used to compute durations of exposure to pharmacological treatments where simulated and real-world data were used to assess its properties comparing the exposure status extrapolated with the method with the “true” exposure status available in the simulated and real-world data. Finally, the SEE was also compared to a Researcher-Defined Duration (RDD) method. When using simulated data, the SEE showed accuracy of 96% and sensitivity of 96%, while when using real-world data, the method showed sensitivity ranging from 78.0 (nortriptyline) to 95.1% (propafenone). When compared to the RDD, the method had a lower median sensitivity of 2.29% (interquartile range 1.21–4.11%). The SEE showed good properties and may represent a promising tool to assess exposure status when information on treatment duration is not available.

Validation of algorithms for selecting rheumatoid arthritis patients in the Tuscan healthcare administrative databases

Validation of algorithms for selecting patients from healthcare administrative databases (HAD) is recommended. This PATHFINDER study section is aimed at testing algorithms to select rheumatoid arthritis (RA) patients from Tuscan HAD (THAD) and assessing RA diagnosis time interval between the medical chart date and that of THAD. A population was extracted from THAD. The information of the medical charts at the Rheumatology Unit of Pisa University Hospital represented the reference. We included first ever users of biologic disease modifying anti-rheumatic drugs (bDMARDs) between 2014 and 2016 (index date) with at least a specialist visit at the Rheumatology Unit of the Pisa University Hospital recorded from 2013 to the index date. Out of these, we tested four index tests (algorithms): (1) RA according to hospital discharge records or emergency department admissions (ICD-9 code, 714*); (2) RA according to exemption code from co-payment (006); (3) RA according to hospital discharge records or emergency department admissions AND RA according to exemption code from co-payment; (4) RA according to hospital discharge records or emergency department admissions OR RA according to exemption code from co-payment. We estimated sensitivity, specificity, positive and negative predicted values (PPV and NPV) with 95% confidence interval (95% CI) and the RA diagnosis median time interval (interquartile range, IQR). Two sensitivity analyses were performed. Among 277 reference patients, 103 had RA. The fourth algorithm identified 96 true RA patients, PPV 0.78 (95% CI 0.70-0.85), sensitivity 0.93 (95% CI 0.86-0.97), specificity 0.84 (95% CI 0.78-0.90), and NPV 0.95 (95% CI 0.91-0.98). The sensitivity analyses confirmed performance. The time measured between the actual RA diagnosis date recorded in medical charts and that assumed in THAD was 2.2 years (IQR 0.5-8.4). In conclusion, this validation showed the fourth algorithm as the best. The time interval elapsed between the actual RA diagnosis date in medical charts and that extrapolated from THAD has to be considered in the design of future studies.

Pharmacoepidemiological methods for computing the duration of pharmacological prescriptions using secondary data sources

PURPOSE

In pharmacoepidemiology, correctly defining the exposure period of pharmacological treatment is a challenging step when information on the time in treatment is missing or incomplete.

METHODS

In this review, we describe several methods for defining exposure to pharmacological treatments using secondary data sources that lack such information.

RESULTS AND CONCLUSION

Several methods for assessing the duration of redeemed prescriptions and combining them into temporal sequences are available. We present a set of considerations to make researchers aware of the potentials and pitfalls of these methods that may aid in minimizing biases in research using these methods. Additionally, we highlight that, to date, there is no one-size-fits-all solution. Thus, the choice of method should be based on their area of applicability combined with a careful mapping to the research scenario under investigation.

Accuracy of Japanese claims data in identifying diabetes-related complications

PURPOSE

To evaluate the accuracy of various claims-based definitions of diabetes-related complications (coronary artery disease [CAD], heart failure, cerebrovascular disease and dialysis).

METHODS

We evaluated data on 1379 inpatients who received care at the Niigata University Medical & Dental Hospital in September 2018. Manual electronic medical chart reviews were conducted for all patients with regard to diabetes-related complications and were used as the gold standard. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of each claims-based definition associated with diabetes-related complications based on Diagnosis Procedure Combination (DPC), International Classification of Diseases, Tenth Revision (ICD-10) codes, procedure codes and medication codes were calculated.

RESULTS

DPC-based definitions had higher sensitivity, specificity, and PPV than ICD-10 code definitions for CAD and cerebrovascular disease, with sensitivity of 0.963-1.000 and 0.905-0.952, specificity of 1.000 and 1.000, and PPV of 1.000 and 1.000, respectively. Sensitivity, specificity, and PPV were high using procedure codes for CAD and dialysis, with sensitivity of 0.963 and 1.000, specificity of 1.000 and 1.000, and PPV of 1.000 and 1.000, respectively. DPC and/or ICD-10 codes + medication were better for heart failure than the ICD-10 code definition, with sensitivity of 0.933, specificity of 1.000, and PPV of 1.000. The PPVs were lower than 60% for all diabetes-related complications using ICD-10 codes only.

CONCLUSION

The DPC-based definitions for CAD and cerebrovascular disease, procedure codes for CAD and dialysis, and DPC or ICD-10 codes with medication codes for heart failure could accurately identify these diabetes-related complications from claims databases.

Use of the French healthcare insurance database to estimate the prevalence of exposure to potential drug-drug interactions

PURPOSE

Drug-drug interactions (DDIs) require monitoring in an aging population with increasing polypharmacy exposure. We aimed to estimate the prevalence of exposure to potential DDIs using the French healthcare insurance system database, for six DDIs with various clinical relevance: angiotensin II receptor blockers or angiotensin-converting enzyme inhibitors and nonsteroidal anti-inflammatory drugs (ARBs-ACEIs + NSAIDs), antiplatelet agents and NSAIDs (AAP + NSAIDs), serotonergic drugs and tramadol (SD + T), statins and macrolides (S + M), oral anticoagulant and NSAIDs (OAC + NSAIDs), and colchicine and macrolides (C + M).

METHODS

We used exhaustive healthcare data from a 1/97th random sample of the population covered by the French health insurance system (EGB) between 2006 and 2016. Exposure to a DDI was defined as overlapping exposure to two interacting drugs. The prevalence of exposure was estimated by year.

RESULTS

Prevalence of exposure in 2016 was estimated at 3.7% for ARBs-ACEIs + NSAIDs, 1.5% for AAP + NSAIDs, 0.76% for SD + T, 0.36% for S + M, 0.24% for AOC + NSAIDs, and 0.02% for C + M. In 26% to 58% of episodes of exposure, the two interacting drugs were prescribed by the same physician and dispensed by the same pharmacy the same day. Between 2006 and 2016, the yearly prevalence was increasing for SD + T and for DDIs involving NSAIDs, and it was decreasing for those involving macrolides.

CONCLUSION

Exposures to potential DDIs in France are not uncommon with a high proportion resulting from a co-prescription by the same physician. Monitoring the prevalence of exposure to DDIs is needed to implement prevention measures. Administrative data enable this surveillance in large and representative cohorts.

Refill Adherence Measures and Its Association with Economic, Clinical, and Humanistic Outcomes Among Pediatric Patients: A Systematic Review

Although refill adherence measures (RAMs) are widely reviewed on their use among adult patients, existing reviews on adherence among children have only focused on self-report measures and electronic monitoring. Hence, this systematic review aims to examine the use of RAMs and their association with economic, clinical, and humanistic outcomes (ECHO) among pediatric patients. A literature search was conducted in Pubmed, Embase, CINAHL, and PsycINFO. Studies published in English involving subjects aged ≤18 years were included if RAMs were analyzed with ECHO. Of the 35 included studies, the majority (n = 33) were conducted in high-income countries. Asthma was the most common condition (n = 9) studied. Overall, 60.6% of 33 clinical outcomes reported among 22 studies was positive (improved clinical outcomes with improved adherence), while 21.9% of 32 economic outcomes reported among 16 studies was positive (reduced healthcare utilization or cost outcomes with improved adherence). Only four studies evaluated the relationship of adherence with 11 humanistic outcomes, where the majority (72.7%) were considered unclear. RAMs are associated with ECHO and can be considered for use in the pediatric population. Future studies could explore the use of RAMs in low-income countries, and the association of RAMs with quality of life.

Using text mining to extract depressive symptoms and to validate the diagnosis of major depressive disorder from electronic health records

BACKGROUND

Many studies have used Taiwan's National Health Insurance Research database (NHIRD) to conduct psychiatric research. However, the accuracy of the diagnostic codes for psychiatric disorders in NHIRD is not validated, and the symptom profiles are not available either. This study aimed to evaluate the accuracy of diagnostic codes and use text mining to extract symptom profile and functional impairment from electronic health records (EHRs) to overcome the above research limitations.

METHODS

A total of 500 discharge notes were randomly selected from a medical center's database. Three annotators reviewed the notes to establish gold standards. The accuracy of diagnostic codes for major psychiatric illness was evaluated. Text mining approaches were applied to extract depressive symptoms and function profiles and to identify patients with major depressive disorder.

RESULTS

The accuracy of the diagnostic code for major depressive disorder, schizophrenia, and dementia was acceptable but that of bipolar disorder and minor depression was less satisfactory. The performance of text mining approach to recognize depressive symptoms is satisfactory; however, the recall for functional impairment is lower resulting in lower F-scores of 0.774-0.753. Using the text mining approach to identify major depressive disorder, the recall was 0.85 but precision was only 0.69.

CONCLUSIONS

The accuracy of the diagnostic code for major depressive disorder in discharge notes was generally acceptable. This finding supports the utilization of psychiatric diagnoses in claims databases. The application of text mining to EHRs might help in overcoming current limitations in research using claims databases.

Suspected adverse drug reactions reported for Brazilian children: cross-sectional study

OBJECTIVE

To assess spontaneous reports of suspected adverse drug reactions in children aged 0-12 years from the Brazilian Health Regulatory Agency between 2008 and 2013.

METHODS

A cross-sectional study on suspected adverse drug reactions reports related to medicines and health products in children was carried out for a six-year period (2008-2013). Year of report, origin of report by Brazilian state, gender, age, suspected drug, adverse reaction description and seriousness were included in the analysis. The data obtained was compared to the number of pediatric beds in health services and to global data from the VigiBase (World Health Organization).

RESULTS

A total of 3330 adverse drug reactions were reported in children in Brazil in the investigated period (54% were in boys). About 28% of suspected adverse drug reactions reports involved 0 to 1-year-old children. Almost 40% of reports came from the Southeast region. Approximately 60% were classified as serious events. There was death in 75 cases. Nearly 30% of deaths involved off-label use; 3875 medicines (465 active substances) were considered suspected drugs. Anti-infective (vancomycin, ceftriaxone, oxacillin, and amphotericin), nervous system (metamizole) and alimentary tract and metabolism medicines were more frequent in reports.

CONCLUSIONS

The distribution of suspected adverse drug reactions reports by sex and age group corresponded to the profile of children hospitalized in Brazil. Data about seriousness and medicines reported may be useful to encourage regulatory actions and improve the safe use of medicines in children.

Real-world data on healthcare resource consumption and costs before and after kidney transplantation

End-stage renal disease (ESRD) is increasing worldwide as a consequence of population aging and increasing chronic illness. Treatment consists mostly of dialysis and kidney transplantation (KTx), and KTx offers advantages for life expectancy and long-term cost reductions compared with dialysis. This study uses the administrative database of the Lombardy Region to analyze the costs of a cohort of patients with ESRD receiving KTx, covering a time period of 24 months before transplant to 12 months after. During 2011, 276 patients underwent kidney transplantation (8.7% preemptive and 91.3% non-preemptive). In the period before transplantation, the main cost driver was dialysis (66.6% for the period from -24 to -12 months and 73.8% for the period from -12 to 0 months), while in the 12 months after KTx, the most relevant cost was surgery. The total cost -24 to -12 months pre-KTx was 35 049.2€; the cost -12 to 0 months was 36 745.9€; and the cost 12 months after KTx was 43 805.8€. Non-preemptive patients showed much higher costs both pre- and post-KTx than preemptive patients. This study highlights how KTx modifies the resource consumption and costs composition of patients with ESRD vs those undergoing dialysis treatment and how KTx may be economically beneficial, especially preemptive intervention.

Bias in pharmacoepidemiologic studies using secondary health care databases: a scoping review

BACKGROUND

The availability of clinical and therapeutic data drawn from medical records and administrative databases has entailed new opportunities for clinical and epidemiologic research. However, these databases present inherent limitations which may render them prone to new biases. We aimed to conduct a structured review of biases specific to observational clinical studies based on secondary databases, and to propose strategies for the mitigation of those biases.

METHODS

Scoping review of the scientific literature published during the period 2000-2018 through an automated search of MEDLINE, EMBASE and Web of Science, supplemented with manually cross-checking of reference lists. We included opinion essays, methodological reviews, analyses or simulation studies, as well as letters to the editor or retractions, the principal objective of which was to highlight the existence of some type of bias in pharmacoepidemiologic studies using secondary databases.

RESULTS

A total of 117 articles were included. An increasing trend in the number of publications concerning the potential limitations of secondary databases was observed over time and across medical research disciplines. Confounding was the most reported category of bias (63.2% of articles), followed by selection and measurement biases (47.0% and 46.2% respectively). Confounding by indication (32.5%), unmeasured/residual confounding (28.2%), outcome misclassification (28.2%) and "immortal time" bias (25.6%) were the subcategories most frequently mentioned.

CONCLUSIONS

Suboptimal use of secondary databases in pharmacoepidemiologic studies has introduced biases in the studies, which may have led to erroneous conclusions. Methods to mitigate biases are available and must be considered in the design, analysis and interpretation phases of studies using these data sources.

Post-marketing studies of pharmaceutical opioid abuse-deterrent formulations: a framework for research design and reporting

BACKGROUND AND AIMS

Opioid formulations with properties to deter abuse (abuse-deterrent formulations; ADFs) have been developed as one response to the prescription opioid 'epidemic'. As for all medicines, ADFs undergo evaluation of safety and efficacy prior to registration for marketing. However, reduced extra-medical use (the primary intended outcome of ADFs and reason for their introduction) can only be established in post-marketing observational studies, comparing them to opioid formulations without abuse-deterrent properties. This has implications for various features of study design and analysis. We discuss proposals for the design, conduct, governance and reporting of post-marketing studies on the effectiveness of pharmaceutical and opioid ADFs.

METHODS

A review of current guidance documents, public work-shops and forums and our own experience with post-marketing studies of ADFs.

RESULTS AND CONCLUSIONS

Research questions for post-marketing studies on ADFs of opioids should reasonably be framed around detecting any probable intended or unintended clinical and/or meaningful changes in specific aspects of extra-medical use (e.g. injection use) and harms. Outcomes reported by prevalence and frequency of occurrence and disaggregated by specific product and route of administration can illustrate the magnitude of ADF impact. We argue that a multi-faceted approach is required, using data from both general population and sentinel high-risk cohorts and from primary and secondary data sources. The comparator (historical non-ADF formulation of that opioid, equivalent current generic or similar opioid product), duration of monitoring and analytical approach require justification and should be sufficient to add weight to conclusions of causality. To maximize transparency, we recommend explicit declarations of funding and conflict of interest, establishment of an advisory committee, publication of study protocol and access to study results.

Linkage of national soil quality measurements to primary care medical records in England and Wales: a new resource for investigating environmental impacts on human health

Background

Long-term, low-level exposure to toxic elements in soil may be harmful to human health but large longitudinal cohort studies with sufficient follow-up time to study these effects are cost-prohibitive and impractical. Linkage of routinely collected medical outcome data to systematic surveys of soil quality may offer a viable alternative.

Methods

We used the Geochemical Baseline Survey of the Environment (G-BASE), a systematic X-ray fluorescence survey of soil inorganic chemistry throughout England and Wales to obtain estimates of the concentrations of 15 elements in the soil contained within each English and Welsh postcode area. We linked these data to the residential postcodes of individuals enrolled in The Health Improvement Network (THIN), a large database of UK primary care medical records, to provide estimates of exposure. Observed exposure levels among the THIN population were compared with expectations based on UK population estimates to assess representativeness.

Results

Three hundred seventy-seven of three hundred ninety-five English and Welsh THIN practices agreed to participate in the linkage, providing complete residential soil metal estimates for 6,243,363 individuals (92% of all current and former patients) with a mean period of prospective computerised medical data collection (follow-up) of 6.75 years. Overall agreement between the THIN population and expectations was excellent; however, the number of participating practices in the Yorkshire & Humber strategic health authority was low, leading to restricted ranges of measurements for some elements relative to the known variations in geochemical concentrations in this area.

Conclusions

The linked database provides unprecedented population size and statistical power to study the effects of elements in soil on human health. With appropriate adjustment, results should be generalizable to and representative of the wider English and Welsh population.

Electronic supplementary material

The online version of this article (10.1186/s12963-018-0168-2) contains supplementary material, which is available to authorized users.

The Challenges and Opportunities of Pharmacoepidemiology in Bone Diseases

Pharmacoepidemiology is used extensively in osteoporosis research and involves the study of the use and effects of drugs in large numbers of people. Randomized controlled trials are considered the gold standard in assessing treatment efficacy and safety. However, their results can have limited external validity when applied to day-to-day patients. Pharmacoepidemiological studies aim to assess the effect/s of treatments in actual practice conditions, but they are limited by the quality, completeness, and inherent bias due to confounding. Sources of information include prospectively collected (primary) as well as readily available routinely collected (secondary) (eg, electronic medical records, administrative/claims databases) data. Although the former enable the collection of ad hoc measurements, the latter provide a unique opportunity for the study of large representative populations and for the assessment of rare events at relatively low cost. Observational cohort and case-control studies, the most commonly implemented study designs in pharmacoepidemiology, each have their strengths and limitations. However, the choice of the study design depends on the research question that needs to be answered. Despite the many advantages of observational studies, they also have limitations. First, missing data is a common issue in routine data, frequently dealt with using multiple imputation. Second, confounding by indication arises because of the lack of randomization; multivariable regression and more specific techniques such as propensity scores (adjustment, matching, stratification, trimming, or weighting) are used to minimize such biases. In addition, immortal time bias (time period during which a subject is artefactually event-free by study design) and time-varying confounding (patient characteristics changing over time) are other types of biases usually accounted for using time-dependent modeling. Finally, residual "uncontrolled" confounding is difficult to assess, and hence to account for it, sensitivity analyses and specific methods (eg, instrumental variables) should be considered.

Effect modification, interaction and mediation: an overview of theoretical insights for clinical investigators

We revisited the three interrelated epidemiological concepts of effect modification, interaction and mediation for clinical investigators and examined their applicability when using research databases. The standard methods that are available to assess interaction, effect modification and mediation are explained and exemplified. For each concept, we first give a simple “best-case” example from a randomized controlled trial, followed by a structurally similar example from an observational study using research databases. Our explanation of the examples is based on recent theoretical developments and insights in the context of large health care databases. Terminology is sometimes ambiguous for what constitutes effect modification and interaction. The strong assumptions underlying the assessment of interaction, and particularly mediation, require clinicians and epidemiologists to take extra care when conducting observational studies in the context of health care databases. These strong assumptions may limit the applicability of interaction and mediation assessments, at least until the biases and limitations of these assessments when using large research databases are clarified.

Use of pharmacy records to measure treatment adherence: a critical review of the literature

The current frame of reference on adherence to pharmacotherapy includes a set of behaviors experienced by the user, with observation of the detailed and continuous history of the use of each dose of the medication. Indicators based on pharmacy records have been used to measure adherence. The current review aimed to identify and describe indicators based on pharmacy records and to discuss their adequacy and limitations for measuring adherence. An exploratory literature review was conducted in three databases using the terms "adherence", "pharmacy records/administrative data", and "measure" to compose the descriptors for the selection of 81 articles and the elaboration of a chart with the denomination, sources, methods for calculation, description, and interpretation of the operational and referential meaning of 14 indicators. Given the most recent taxonomy for adherence proposed in the literature, we concluded that the indicators can be useful for identifying patients with medication-seeking behavior-related problems and analysis of persistence. The distance between supply-related events and difficulties in treatment follow-up can influence an analysis based exclusively on the use of these indicators.

The RECORD reporting guidelines: meeting the methodological and ethical demands of transparency in research using routinely-collected health data

Routinely-collected health data (RCD) are now used for a wide range of studies, including observational studies, comparative effectiveness research, diagnostics, studies of adverse effects, and predictive analytics. At the same time, limitations inherent in using data collected without specific a priori research questions are increasingly recognized. There is also a growing awareness of the suboptimal quality of reports presenting research based on RCD. This has created a perfect storm of increased interest and use of RCD for research, together with inadequate reporting of the strengths and weaknesses of these data resources. The REporting of studies Conducted using Observational Routinely-collected Data (RECORD) statement was developed to address these limitations and to help researchers using RCD to meet their ethical obligations of complete and accurate reporting, as well as improve the utility of research conducted using RCD. The RECORD statement has been endorsed by more than 15 journals, including Clinical Epidemiology. This journal now recommends that authors submit the RECORD checklist together with any manuscript reporting on research using RCD.

'Big data' in mental health research: current status and emerging possibilities

PURPOSE

'Big data' are accumulating in a multitude of domains and offer novel opportunities for research. The role of these resources in mental health investigations remains relatively unexplored, although a number of datasets are in use and supporting a range of projects. We sought to review big data resources and their use in mental health research to characterise applications to date and consider directions for innovation in future.

METHODS

A narrative review.

RESULTS

Clear disparities were evident in geographic regions covered and in the disorders and interventions receiving most attention.

DISCUSSION

We discuss the strengths and weaknesses of the use of different types of data and the challenges of big data in general. Current research output from big data is still predominantly determined by the information and resources available and there is a need to reverse the situation so that big data platforms are more driven by the needs of clinical services and service users.

Overview of drug data within French health insurance databases and implications for pharmacoepidemiological studies

In this study, we aimed to provide an updated overview of drug data contained in the French health insurance database (SNIIRAM) and its associated national representative sample (EGB). This study identified most common problems concerning drug data: (i) change in level of coverage of drugs of interest (drug no more eligible for reimbursement or no more prescription-only), (ii) break in patients' eligibility (in connection with change of healthcare plan or patients' identifier), and (iii) technical and regulatory issues. We provide a brief checklist to enable a structured identification of these issues. The impact of gap in drug data availability on study validity will depend on the research question, drug, setting, and population of interest. The French health insurance database and associated sample are valuable resources for pharmacoepidemiological research. There is a need to pursue further methodological and validation studies to promote accurate and transparent use of French health insurance databases for pharmacoepidemiology.

Achievements, challenges, and future perspectives of epidemiologic research in restless legs syndrome (RLS)

In the 20 years since the initial consensus on a common definition for restless legs syndrome (RLS), over 600 scientific reports on epidemiological aspects of RLS have been published. Most are descriptive and address important issues such as prevalence, familial patterns, comorbidities, and quality of life. While the establishment of prospective cohort studies and the use of secondary data sources are rather new to RLS research, both options significantly broaden the possibilities for analysis of disease risk factors. These two options, as well as the inclusion of a broader phenotyping of individual patients, have great potential to elucidate etiologic factors for RLS and expand knowledge about this common disorder. This article summarizes achievements in the area of RLS epidemiology, describes current challenges, and highlights future perspectives in the field.

Scaling up health knowledge at European level requires sharing integrated data: an approach for collection of database specification

Computerized health care databases have been widely described as an excellent opportunity for research. The availability of "big data" has brought about a wave of innovation in projects when conducting health services research. Most of the available secondary data sources are restricted to the geographical scope of a given country and present heterogeneous structure and content. Under the umbrella of the European Innovation Partnership on Active and Healthy Ageing, collaborative work conducted by the partners of the group on "adherence to prescription and medical plans" identified the use of observational and large-population databases to monitor medication-taking behavior in the elderly. This article describes the methodology used to gather the information from available databases among the Adherence Action Group partners with the aim of improving data sharing on a European level. A total of six databases belonging to three different European countries (Spain, Republic of Ireland, and Italy) were included in the analysis. Preliminary results suggest that there are some similarities. However, these results should be applied in different contexts and European countries, supporting the idea that large European studies should be designed in order to get the most of already available databases.

Healthcare costs of the progression of chronic kidney disease and different dialysis techniques estimated through administrative database analysis

BACKGROUND

Chronic kidney disease (CKD) progression is associated with significant comorbidities and costs. In Italy, limited evidence of healthcare resource consumption and costs is available. We therefore aimed to investigate the direct healthcare costs in charge to the Lombardy Regional Health Service (RHS) for the treatment of CKD patients in the first year after starting hemodialysis and in the 2 years prior to dialysis.

METHODS

Citizens resident in the Lombardy Region (Italy) who initiated dialysis in the year 2011 (Jan 1 to Dec 31) were selected and data were extracted from Lombardy Regional databases on their direct healthcare costs in the first year after starting dialysis and in the 2 years prior to it was analyzed. Drugs, hospitalizations, diagnostic procedures and outpatient costs covered by RHS were estimated. Patients treated for acute kidney injury, or who died or stopped dialysis during the observational period were excluded.

RESULTS

From the regional population (>9,700,000 inhabitants), 1067 patients (34.3 % females) initiating dialysis were identified, of whom 82 % underwent only hemodialysis (HD), 13 % only peritoneal dialysis (PD) and the remaining 5 % both treatments. Direct healthcare costs/patient were € 5239, € 12,303 and € 38,821 (€ 40,132 for HD vs. € 30,444 for PD patients) for the periods 24-12 months pre-dialysis, 12-0 months pre-dialysis, and in the first year of dialysis, respectively.

CONCLUSIONS

This study highlights a significant economic burden related to CKD and an increase in direct healthcare costs associated with the start of dialysis, pointing to the importance of prevention programs and early diagnosis.

The Australian Pharmaceutical Benefits Scheme data collection: a practical guide for researchers

Background

The Pharmaceutical Benefits Scheme (PBS) is Australia’s national drug subsidy program. This paper provides a practical guide to researchers using PBS data to examine prescribed medicine use.

Findings

Excerpts of the PBS data collection are available in a variety of formats. We describe the core components of four publicly available extracts (the Australian Statistics on Medicines, PBS statistics online, section 85 extract, under co-payment extract). We also detail common analytical challenges and key issues regarding the interpretation of utilisation using the PBS collection and its various extracts.

Conclusions

Research using routinely collected data is increasing internationally. PBS data are a valuable resource for Australian pharmacoepidemiological and pharmaceutical policy research. A detailed knowledge of the PBS, the nuances of data capture, and the extracts available for research purposes are necessary to ensure robust methodology, interpretation, and translation of study findings into policy and practice.

Case Definitions for Chronic Rhinosinusitis in Administrative Data: A Systematic Review

Background

The use of administrative data for pharmacoepidemiology research on chronic rhinosinusitis (CRS) has become increasingly popular. Although large sample sizes and ease of accessibility have made electronic health data an attractive data option, the risk of inaccurate cohort identification can lead to biased outcomes.

Objectives

The objectives of this systematic review were to (1) report current case definitions for CRS used in administrative data base research, and (2) define the various administrative data bases used for CRS research.

Methods

Medical literature data bases were searched from the date of their inception to February 1, 2015. Included studies were publications that obtained CRS-specific data from a health records data base. Studies were excluded if they evaluated a non-CRS cohort, failed to use or report an international classification of disease (ICD) code in the case definition, or published in a non-peer-reviewed journal.

Results

Of the 27 studies that met inclusion criteria, 8 different CRS case definitions were identified and 13 administrative data bases were evaluated. Of the 8 different CRS case definitions identified, only one was validated. The most commonly used CRS case definition was the ICD-9 473.x code alone.

Conclusion

To optimize the accuracy of pharmacoepidemiologic research for CRS that used administrative data, it is important to apply appropriate case definitions for CRS. Various nonvalidated CRS case definitions are currently being used in administrative data base research. There is a need to develop a generalizable and validated ICD-based CRS case definition to increase the accuracy of future pharmacoepidemiologic research.

Overt hepatic encephalopathy in Italy: clinical outcomes and healthcare costs

PURPOSE

Hepatic encephalopathy (HE) is a recurrent severe complication of progressive hepatic cirrhosis. The aim of this study is to evaluate the average annual direct healthcare costs for the treatment of patients with overt HE in Italy.

PATIENTS AND METHODS

This retrospective, observational study analyzed information from the database of ARNO Observatory. Patients with at least one hospitalization due to overt HE in the period from January 1, 2011 to December 31, 2011, were selected and observed during the year following the hospitalization. Costs for drugs, diagnostic and therapeutic procedures, and hospitalizations were estimated from the Italian National Health Service perspective.

RESULTS

Out of a population of 2,678,462 subjects, 381 patients were identified, of whom, 21.5% died during the first hospitalization and 5.8% during the follow-up; the survival rate was 72.7% at the end of the observation period. The direct healthcare costs per patient amounted to €13,393/year (15,295 USD) (88% for hospitalizations, 8% for drugs, and 4% for diagnostic procedures). During the follow-up, 42.5% of patients had at least one rehospitalization due to HE. Patients readmitted for HE had an average annual cost of €21,272 (24,293 USD), almost doubled if compared to patients without readmissions (€12,098 [13,816 USD]).

CONCLUSION

This analysis showed that patients with HE had relevant direct healthcare costs, in which hospitalizations were the most important cost drivers.

Cardiovascular safety of sulphonylureas: over 40 years of continuous controversy without an answer

More than 40 years after publication of the University Group Diabetes Program trial, the cardiovascular safety of sulphonylureas is still contentious. Although several hypotheses linking sulphonylureas to adverse cardiovascular effects exist, none provide conclusive evidence. Adding to the controversy, current clinical trials and observational studies provide inconsistent, and sometimes conflicting, evidence for the cardiovascular effects of sulphonylureas. Overall, observational evidence suggests that an increased risk of adverse cardiovascular outcomes is associated with sulphonylureas; however, these data may be subject to residual confounding and bias. Although evidence from randomized controlled trials has suggested a neutral effect, the majority of these studies were not specifically designed to assess the effect of sulphonylureas on adverse cardiovascular event risk. Current ongoing large clinical trials may provide some clarity on the cardiovascular safety of sulphonylureas, but the results are not expected for several years. With the continued uncertainties concerning the cardiovascular safety of all antidiabetic drugs, a clear answer with regard to sulphonylureas is warranted. The objectives of the present article were to provide an overview of the controversy surrounding sulphonylurea-related cardiovascular effects, to discuss the limitations of the current literature, and to provide recommendations for future studies aiming to elucidate the true relationship between sulphonylureas and adverse cardiovascular effects in people with type 2 diabetes.

The REporting of Studies Conducted Using Observational Routinely-Collected Health Data (RECORD) Statement: Methods for Arriving at Consensus and Developing Reporting Guidelines

Objective

Routinely collected health data, collected for administrative and clinical purposes, without specific a priori research questions, are increasingly used for observational, comparative effectiveness, health services research, and clinical trials. The rapid evolution and availability of routinely collected data for research has brought to light specific issues not addressed by existing reporting guidelines. The aim of the present project was to determine the priorities of stakeholders in order to guide the development of the REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement.

Methods

Two modified electronic Delphi surveys were sent to stakeholders. The first determined themes deemed important to include in the RECORD statement, and was analyzed using qualitative methods. The second determined quantitative prioritization of the themes based on categorization of manuscript headings. The surveys were followed by a meeting of RECORD working committee, and re-engagement with stakeholders via an online commentary period.

Results

The qualitative survey (76 responses of 123 surveys sent) generated 10 overarching themes and 13 themes derived from existing STROBE categories. Highest-rated overall items for inclusion were: Disease/exposure identification algorithms; Characteristics of the population included in databases; and Characteristics of the data. In the quantitative survey (71 responses of 135 sent), the importance assigned to each of the compiled themes varied depending on the manuscript section to which they were assigned. Following the working committee meeting, online ranking by stakeholders provided feedback and resulted in revision of the final checklist.

Conclusions

The RECORD statement incorporated the suggestions provided by a large, diverse group of stakeholders to create a reporting checklist specific to observational research using routinely collected health data. Our findings point to unique aspects of studies conducted with routinely collected health data and the perceived need for better reporting of methodological issues.

Chronic Kidney Disease: Evolution of Healthcare Costs and Resource Consumption from Predialysis to Dialysis in Piedmont Region, Italy

This study aims at assessing the evolution in healthcare costs for chronic kidney disease (CKD) patients through the analysis of administrative databases of Piedmont region, Italy. This is a retrospective, observational study, for which patients undergoing at least one dialysis for CKD in the period of June 1, 2010–May 31, 2011 were selected. Two subpopulations were evaluated: patients incident-to-dialysis observed for the 12 months preceding dialysis entrance (PreD) and “established” dialysis patients (at least 120 dialyses/year) observed for 12 months (EstD). Overall, 1,059 PreD and 2,018 EstD patients were selected. The average yearly cost per PreD patient accounted for 11,123€ ± 15,095€ (75% hospitalizations, 17% drugs, and 8% diagnostic/therapeutic procedures). The average yearly cost per EstD patient accounted for 53,764€ ± 14,685€ (59% dialysis, 21% diagnostic/therapeutic procedures, 13% hospitalizations, and 6.7% drugs). Among EstD population, hemodialysis patients cost 56,049€ ± 13,473€ per year, whereas peritoneal dialysis patients cost 34,978€ ± 10,847€ per year. The significant difference in expenditure between predialysis and dialysis suggests that prevention, early diagnosis, and the consequent possible delay of dialysis entrance could lead to important savings for healthcare services, as well as a better global health status for patients.

Resource consumption and healthcare costs of acute coronary syndrome: a retrospective observational administrative database analysis

The objective of this study was to estimate resource consumption and direct healthcare costs of patients with a first hospitalization for acute coronary syndrome (ACS) in 2008 in the Piedmont Region, Italy. Subjects hospitalized with a first episode of ACS in 2008 were selected from the regional hospital discharge database. All hospitalizations, drug prescriptions, and outpatient episodes of care in the 12 months following discharge were considered to estimate resource consumption and direct healthcare costs from the Piedmont Regional Health Service perspective. The analysis was carried out separately for ST-elevation myocardial infarction (STEMI), non-ST-elevation myocardial infarction (NSTEMI), and unstable angina (UA) populations. In the accrual period, 7765 subjects (1.75‰ of the total population) were hospitalized for ACS (64.2% men). The average age was 66.5 for men and 75.4 for women. The average in-hospital mortality was 6.5% (n = 508). The total ACS population was classified as: STEMI 45.2%, NSTEMI 29.4%, and UA 25.4%. The average yearly costs per patient alive at the end of follow-up (n = 6851) were 14,160.8&OV0556; (18,678.7 USD): 83.9% for inpatient admissions [11,881.2&OV0556; (15,671.8 USD)], 9.3% for drugs [1311.6&OV0556; (1730.1 USD)], 5.0% for diagnostic and therapeutic procedures and outpatient visits [708.2&OV0556; (934.1 USD)], and 1.8% for 1-day hospital stays [259.8&OV0556; (342.7 USD)]. The average yearly direct healthcare costs by ACS event were 14,984.5&OV0556; (19,765.2 USD) for STEMI, 14,554.1&OV0556; (19,197.4 USD) for NSTEMI, and 12,481.5&OV0556; (16,463.6 USD) for UA. In each subpopulation, costs were significantly higher for men than for women. ACS imposes a significant burden in terms of morbidity and mortality and generates major public health service costs.

The intriguing future of pharmacoepidemiology

Over the years more and more data have become available in the constantly growing databases on prescription drugs. This has facilitated the development of pharmacoepidemiology, a dynamic research field that has undergone a more rapid development than many other research areas. There are several reasons why pharmacoepidemiology will remain recognized as one of the most dynamic and challenging research areas of clinical pharmacology. The data explosion in modern society will continue, and observational studies aimed at assessing the value of medicines will be increasingly requested by payers, professionals and patients. Future studies in pharmacoepidemiology must include specialist drugs used in the hospital setting and also be designed to address obstacles hindering the delivery of effective medicines to the patient. Pharmacoepidemiological methods may also be valuable tools to address new challenges, such as the environmental impact of medicines. A potential threat is that the increasing amounts of data available in registries may add fuel to the debate on confidentiality. The too strict application of privacy rules might hinder the further development of pharmacoepidemiology.

Changing prescribing patterns of type 2 diabetes medications from 2002-2010: an electronic health record-based evaluation

BACKGROUND

The implementation of electronic health records (EHRs) may support evaluations of health care delivery, such as the prescription of newly approved medications, to adults with diabetes. We aimed to evaluate prescribing patterns of thiazolidinediones and novel glucose-lowering drug classes using electronic prescribing data contained in an outpatient EHR from 2002-2010.

METHODS

We identified adults with type 2 diabetes seen from 2002-2010 who were newly prescribed rosiglitazone (ROSI), pioglitazone (PIO), or a novel glucose-lowering drug class (other). The annual number of new prescriptions and their relative percentages (per 1000 patients) were calculated.

RESULTS

From 2002-2010, 6209 patients with type 2 diabetes were newly prescribed 8858 eligible medications. In 2006, ROSI and PIO accounted for 44% and 37% of new prescriptions, respectively. After 2007, the relative percentage of new ROSI prescriptions declined more rapidly than PIO prescriptions, falling to 7% and 47% of peak levels, respectively, by 2010. By 2010, the relative percentages of new ROSI, PIO, and other prescriptions were 2%, 18%, and 80%, respectively.

CONCLUSIONS

Evaluations of EHR data represent a cost-effective method for evaluating diabetes medications with new Food and Drug Administration warnings or indications. Validation of demographic and clinical data will expand the scope of EHR-based evaluations of health care delivery and outcomes for adults with diabetes.

Real-world impact of reminder packaging on antihypertensive treatment adherence and persistence

BACKGROUND

Patient medication adherence is multidimensional and poses significant concerns to health care professionals. One aspect of adherence is a patient forgetting to take their prescribed medication, which may be improved with reminder packaging (RP). The objective of this analysis was to assess the impact of RP on patient adherence to antihypertensive therapy.

METHODS

This retrospective, propensity score-matched study evaluated patients switching to a single-pill combination of valsartan-hydrochlorothiazide in RP compared with patients remaining on the combination without reminder packaging (non-RP). Patients receiving combination therapy between April 1, 2009 and July 31, 2010 were eligible for inclusion. Patients were propensity score-matched on baseline adherence and background demographic variables, including comorbidities. Medication possession ratio, proportion of days covered, time to refill, and time to discontinuation were evaluated as primary measures of subsequent adherence and persistence.

RESULTS

In a total of 9266 matched patients (4633 participants in both cohorts), adherence was significantly higher in the RP cohort compared with patients in the non-RP cohort (medication possession ratio, RP 80% versus non-RP 73%; proportion of days covered, RP 76% versus non-RP 63%; both P < 0.001). Refill timing was 10 days for RP patients versus 16 days for non-RP patients (P < 0.001). Similar trends were observed with respect to time to discontinuation (RP 196 days, non-RP 174 days; P < 0.001). A higher proportion of RP patients remained on therapy compared with non-RP patients, with patients in the RP group being 17% less likely to discontinue therapy compared with patients in the non-RP group (hazards ratio 0.833; 95% confidence interval 0.793-0.875).

CONCLUSION

This real-world assessment of differences in adherence and persistence rates demonstrated that patients receiving RP were more adherent and persistent with their treatment regimens.

Correlation versus causation? Pharmacovigilance of the analgesic flupirtine exemplifies the need for refined spontaneous ADR reporting

Annually, adverse drug reactions result in more than 2,000,000 hospitalizations and rank among the top 10 causes of death in the United States. Consequently, there is a need to continuously monitor and to improve the safety assessment of marketed drugs. Nonetheless, pharmacovigilance practice frequently lacks causality assessment. Here, we report the case of flupirtine, a centrally acting non-opioid analgesic. We re-evaluated the plausibility and causality of 226 unselected, spontaneously reported hepatobiliary adverse drug reactions according to the adapted Bradford-Hill criteria, CIOMS score and WHO-UMC scales. Thorough re-evaluation showed that only about 20% of the reported cases were probable or likely for flupirtine treatment, suggesting an incidence of flupirtine-related liver injury of 1∶ 100,000 when estimated prescription data are considered, or 0.8 in 10,000 on the basis of all 226 reported adverse drug reactions. Neither daily or cumulative dose nor duration of treatment correlated with markers of liver injury. In the majority of cases (151/226), an average of 3 co-medications with drugs known for their liver liability was observed that may well be causative for adverse drug reactions, but were reported under a suspected flupirtine ADR. Our study highlights the need to improve the quality and standards of ADR reporting. This should be done with utmost care taking into account contributing factors such as concomitant medications including over-the-counter drugs, the medical history and current health conditions, in order to avoid unjustified flagging and drug warnings that may erroneously cause uncertainty among healthcare professionals and patients, and may eventually lead to unjustified safety signals of useful drugs with a reasonable risk to benefit ratio.

A pharmacoepidemiologic analysis of the impact of calendar packaging on adherence to self-administered medications for long-term use

BACKGROUND

Calendar blister packaging (CBP) that incorporates a day or date feature is a simple medication packaging technology that is designed to improve medication adherence and persistence.

OBJECTIVE

This study was conducted to assess the effect of a new calendar packaging technology on prescription refill adherence and persistence for daily, self-administered, long-term medication use.

METHODS

Anonymized pharmacy dispensing data from a large US mass merchandiser were analyzed. This retrospective cohort study included people aged 18 to 75 years who filled prescriptions for oral lisinopril or enalapril (control group) at a study pharmacy during 1 year before and after the switch of lisinopril packaging from vials to CBP. Cohorts were stratified into new and prevalent medication users. We used linear and logistic regression modeling and propensity score matching to assess the impact of CBP on refill adherence, using medication possession ratio (MPR) and proportion of days covered (PDC), and persistence using length of therapy (LOT).

RESULTS

Our sample comprised 76,321 new users and 249,040 prevalent users. Across all user, medication, and packaging groups, the mean unadjusted LOT decreased in the follow-up year, possibly due to economic recession. The LOT decline was attenuated in the CBP cohort. After adjustment for covariates, CBP use in new and prevalent medication users was associated with significantly higher LOT and PDC but not MPR. The odds of achieving PDC ≥80% were higher by 15% in new users (odds ratio [OR] = 1.15; 95% CI, 1.09-1.21) and 12% in prevalent users (OR = 1.12; 95% CI, 1.09-1.15) who switched to CBP, compared with continued vial use.

CONCLUSIONS

CBP of medication prescribed for daily, self-administered, long-term use was associated with modest improvement in prescription refill adherence and persistence. An adherence strategy of even small effect size that is broadly implemented on a population level could significantly leverage therapeutic effect and provide substantial cumulative public health benefit. Clinical benefit, or harm, associated with use of CBP should be investigated. Usability assessments of CBP in patient subgroups may provide insight about differential impact on adherence and persistence.

Electronic health records: Implications for drug discovery

Electronic health records (EHRs) have increased in popularity in many countries. Pushed by legal mandates, EHR systems have seen substantial progress recently, including increasing adoption of standards, improved medical vocabularies and enhancements in technical infrastructure for data sharing across healthcare providers. Although the progress is directly beneficial to patient care in a hospital or clinical setting, it can also aid drug discovery. In this article, we review three specific applications of EHRs in a drug discovery context: finding novel relationships between diseases, re-evaluating drug usage and discovering phenotype-genotype associations. We believe that in the near future EHR systems and related databases will impact significantly how we discover and develop safe and efficacious medicines.

Principles of drug literature evaluation for observational study designs

Clinical pharmacists must often interpret and use data derived from observational studies when making clinical decisions and recommendations, answering drug information questions, and educating patients about drug therapy. This is the third article in a three-part series intended to enhance clinical pharmacists' understanding of methods frequently used in such research and their application to patient care. Principles for critical evaluation of published observational studies are presented. Issues of bias and confounding that must be taken into account when assessing the data derived from observational studies are explained. Finally, examples of applications in pharmacy practice are presented.

Ascertainment of warfarin and aspirin use by medical record review compared with automated pharmacy data

PURPOSE

Automated pharmacy databases are increasingly available for assessing medication use, but research on the validity of these data is incomplete. This study aimed to measure agreement on warfarin and aspirin use between medical records and automated pharmacy data among patients with newly detected atrial fibrillation (AF).

METHODS

Patients with newly detected AF (n = 1953) were previously identified in a cohort study at Group Health (GH) in Washington State. Medical records were reviewed for information on risk factors and medication use, as well as clinical care during the 6 months after AF onset. Medication data were also obtained from the GH pharmacy database. We determined the sensitivity, specificity, and positive predictive value (PPV) as measures of the validity of the GH pharmacy database as compared with medical records for warfarin and aspirin use during the first 6 and 3 months after AF onset. We also calculated the κ statistic.

RESULTS

For warfarin use, in comparison with the medical record review, the sensitivity, specificity, and PPV for the GH pharmacy database were excellent, and agreement was almost perfect in the 3- and 6-month periods after AF onset (κ = 0.92 and 0.93, respectively). For aspirin use, the GH pharmacy database had low sensitivity but high specificity, and agreement was only fair for these two periods (κ = 0.28 and 0.31, respectively).

CONCLUSIONS

The GH pharmacy database is a valuable source of data for pharmacoepidemiologic research on warfarin use among patients with AF. However, the database cannot be recommended for assessment of aspirin use. Copyright © 2010 John Wiley & Sons, Ltd.