Abstract
Objective: Population-based identification of patients with a myocardial infarction is limited to patients presenting to hospital with an acute event. We set out to determine if adding physician billing data to hospital discharge data would result in an accurate capture of patients who have had a myocardial infarction.

Methods: We performed a retrospective chart abstraction of 969 randomly selected adult patients using data abstracted from primary care physicians on an electronic medical record in Ontario, Canada, as the reference standard.

Results: An algorithm of 3 physician billings in a one-year period with at least one being by a specialist or within a hospital or emergency room plus one hospital discharge abstract performed with a sensitivity of 80.4% (95% CI: 69.5-91.3), specificity of 98.0% (95% CI: 97.1-98.9), positive predictive value of 69.5% (95% CI: 57.7-81.2), negative predictive value of 98.9% (95% CI: 98.2% to 99.6%) and kappa statistic of 0.73 (95% CI: 0.63-0.83).

Conclusion: Using a combination of hospital discharge abstracts and physician bill- ing data may be the best way of assessing trends of MI occurrence over time since it increases the capture of MI beyond those patients who have been hospitalized.

SUMMARY

Purpose: To determine the validity of pregnancy variables recorded in administrative databases of Quebec using patient medical charts as the gold standard among asthmatic pregnant women.

Methods: Three administrative databases were linked and provided information on maternal, pregnancy and infant characteristics for 726 pregnant asthmatic women who delivered in 1990–2000. Algorithms were developed to measure variables that were not recorded directly in the databases or to minimize the number of missing values for variables recorded in two or more databases. Medical file data were collected by two trained research nurses in 43 hospitals. The validity of categorical variables was assessed with sensitivity, specificity, predictive positive values (PPVs) and predictive negative values (PNVs), whereas the validity of continuous variables was assessed with Pearson correlation using the medical chart as the gold standard.

Results: The sensitivity of the sex of the baby, previous live birth and previous pregnancy ranged from 0.97 to 0.99. Corresponding figures were 0.92–0.98 for specificity. We also found high correlation coefficients, ranging from 0.875 to 0.999 for the length of gestation, dates of last menstruation and delivery, maternal age and birth weight.

Conclusion: Pregnancy-related variables recorded in administrative databases or derived from algorithms based on two or more databases were found to be highly valid as compared to the medical chart among asthmatic women.

key words — validation; pregnancy-related variables; administrative databases; medical chart; asthma

Objective. To assess the association between the initiation of anti–tumor necrosis factor ’ (anti-TNF’) therapy and the risk of serious bacterial infections in routine care.

Methods. This was a cohort study of patients with rheumatoid arthritis (RA) in whom specific disease- modifying antirheumatic drugs (DMARDs) were initiated. Patients were Medicare beneficiaries ages 65 years and older (mean age 76.5 years) who were concurrently enrolled in the Pharmaceutical Assistance Contract for the Elderly provided by the state of Pennsylvania. A total of 15,597 RA patients in whom a DMARD was initiated between January 1, 1995 and December 31, 2003 were identified using linked data on all prescription drug dispensings, physician services, and hospitalizations. Initiation of anti-TNF’ therapy, cytotoxic agents other than methotrexate (MTX), noncytotoxic agents, and glucocorticoids was compared with initia- tion of MTX. The main outcome measure was serious bacterial infections that required hospitalization.

Objective. To develop the Self-Administered Comorbidity Questionnaire (SCQ) and assess its psychometric properties, including the predictive validity of the instrument, as reflected by its association with health status and health care utilization after 1 year.

Methods. A cross-sectional comparison of the SCQ with a standard, chart abstraction-based measure (Charlson Index) was conducted on 170 inpatients from medical and surgical care units. The association of the SCQ with the chart-based comorbidity instrument and health status (short form 36) was evaluated cross sectionally. The association between these measures and health status and resource utilization was assessed after 1 year.

Results. The Spearman correlation coefficient for the association between the SCQ and the Charlson Index was 0.32. After restricting each measure to include only comparable items, the correlation between measures was stronger (Spearman r

OBJECTIVE — Accurate information about the magnitude and distribution of diabetes can inform policy and support health care evaluation. We linked physician service claims (PSCs) and hospital discharge abstracts (HDAs) to determine diabetes prevalence and incidence.

RESEARCH DESIGN AND METHODS — A retrospective cohort was constructed us- ing administrative data from the national HDA database, PSCs for Ontario (population 11 million), and registries carrying demographics and vital statistics. All HDAs and PSCs bearing a diagnosis of diabetes (ICD9-CM 250) were selected for 1991–1999. Two previously reported algorithms for identification of diabetes were applied as follows: “1-claim” (any HDA or PSC showing diabetes) and “2-claim” (one HDA or two PSCs within 2 years showing diabetes). Incident cases were defined as individuals who met the criteria for diabetes for the first time after at least 2 years of observation. For validation, diagnostic data abstracted from primary care charts (n

Summary. Because primary data collection can be expensive, researchers are increasingly using information collected in medical administrative databases for scientific purposes. This information, however, is typically collected for reasons other than research, and many such databases have been shown to contain substantial proportions of misclassification errors. For example, many administrative databases contain fields for patient diagnostic codes, but these are often missing or inaccurate, in part because physician reimbursement schemes depend on medical acts performed rather than any diagnosis. Errors in ascertaining which individuals have a given disease bias not only prevalence estimates, but also estimates of associations between the disease and other variables, such as medication use. We attempt to estimate the prevalence of osteoarthritis (OA) among elderly Quebeckers using a government administrative database. We compare a naive estimate relying solely on the physician diagnoses of OA listed in the database to estimates from several different Bayesian latent class models which adjust for misclassified physician diagnostic codes via use of other available diagnostic clues. We find that the prevalence estimates vary widely, depending on the model used and assumptions made. We conclude that any inferences from these databases need to be interpreted with great caution, until further work estimating the reliability of database items is carried out.

Key words: Administrative databases; Bayesian latent class models; Diagnosis; Misclassification; Preva- lence; Robustness; Sensitivity; Specificity.

Two distinct subtypes of hip fracture, transcervical and pertrochanteric, can be distinguished on the basis of the anatomical location of the injury. While the epidemiology of hip fractures has been well described, typically, little or no distinction is made between these subtypes. The objective of this study was to compare and contrast age- and sex-specific rates of transcervical and pertrochantenc fractures in Quebec, Canada. The data for this study were obtained from a database containing records of all persons discharged from all hospitals in Quebec from 1981 to 1992. Rates of hip fracture were calculated by using the population aged 50 years and older as the denominator, and changes in rates over time were assessed using Poisson regression. There were no statistically significant trends in the changes in rates over time (i e., 95 percent confidence intervals overlapped the null value). Among women below age 70 years, transcervical fractures were more common, whereas among older women, pertrochantenc fractures predominated. Among men, pertrochanteric fractures predominated at all ages. There was a marked seasonal vanation in the occurrence of all hip fractures combined: Compared with the summer months, the relative nsk of all hip fractures during the winter was 1.32 (95 percent confidence interval 1.28-1.36). The results of this study indicate that the two subtypes of hip fracture, transcervical and pertrochanteric, have different patterns of occurrence, suggesting different risk factor profiles. Clearly, a multidisciplinary research approach is needed before it will be possible to untangle the complex relation between the metabolic processes occurring at the level of the individual and the distribution of the disease in the population.

SUMMARY

The goal of this study was to evaluate the validity of asthma diagnoses recorded in the Medical Services (physician billing) database of the Canadian province of Quebec. The predictive positive value (PPV) and predictive negative value (PNV) of two operational definitions of asthma based on diagnoses recorded in the database were evaluated. Patients 16–80 years old treated by a respiratory or a family physician in 2002 were selected from the database. The diagnosis derived from the Medical Services database was compared to the diagnosis written in the patient’s medical chart. The PPV and PNV of the first operational definition based on one asthma diagnosis or more recorded in the database over a 1-year period were found to be 0.75 and 0.96 for respiratory physicians and 0.67 and 0.99 for family physicians, for patients 16–44 years old. The PPV increased to 0.78 for family physicians and to 0.77 for respiratory physicians when the second operational definition based on two diagnoses of asthma or more was used. Results tended to be lower for 45–80 years old patients. We conclude that diagnoses recorded in the Medical Services database of Quebec are valid to identify patients with asthma.

ABSTRACT
Medicaid claims and eligibility data, particularly when linked to other sources of patient-level and contextual information, represent a powerful and under-used resource for health services research on the use and outcomes of prescription drugs. However, their effective use poses many methodological and inferential challenges. This article reviews strengths, limitations, challenges, and recommended strategies in using Medicaid data for research on the initiation, continuation, and outcomes of prescription drug therapies. Drawing from published research using Medicaid data by the investigators and other groups, we review several key validity and methodological issues. We discuss strategies for claims-based identification of diagnostic subgroups and procedures, measuring and modeling initiation and persistence of regimens, analysis of treatment disparities, and examination of comorbidity patterns. Based on this review, we discuss “best practices” for appropriate data use and validity checking, approaches to statistical modeling of longitudinal patterns in the presence of typical challenges, and strategies for strengthening the power and potential of Medicaid datasets. Finally, we discuss policy implications, including the potential for the research use of Medicare Part D data and the need for further initiatives to systematically develop and optimally use research datasets that link Medicaid and other sources of clinical and outcome information.

KEYWORDS
Medicaid; prescription drugs; methodology; claims; outcomes; adherence

Often, data on important confounders are not available in cohort studies. Sensitivity analyses based on the relation of single, but not multiple, unmeasured confounders with an exposure of interest in a separate validation study have been proposed. In this paper, the authors controlled for measured confounding in the main cohort using propensity scores (PS’s) and addressed unmeasured confounding by estimating two additional PS’s in a validation study. The ‘‘error-prone’’ PS exclusively used information available in the main cohort. The ‘‘gold standard’’ PS additionally included data on covariates available only in the validation study. Based on these two PS’s in the validation study, regression calibration was applied to adjust regression coefficients. This propensity score cali- bration (PSC) adjusts for unmeasured confounding in cohort studies with validation data under certain, usually untestable, assumptions. The authors used PSC to assess the relation between nonsteroidal antiinflammatory drugs (NSAIDs) and 1-year mortality in a large cohort of elderly persons. ‘‘Traditional’’ adjustment resulted in a hazard ratio for NSAID users of 0.80 (95% confidence interval (CI): 0.77, 0.83) as compared with an unadjusted hazard ratio of 0.68 (95% CI: 0.66, 0.71). Application of PSC resulted in a more plausible hazard ratio of 1.06 (95% CI: 1.00, 1.12). Until the validity and limitations of PSC have been assessed in different settings, the method should be seen as a sensitivity analysis.

KEYWORDS: bias (epidemiology); cohort studies; confounding factors (epidemiology); epidemiologic methods; propensity score calibration; research design