Julie M. Legler

Development of a comorbidity index using physician claims data

Important comorbidities recorded on outpatient claims in administrative datasets may be missed in analyses when only inpatient care is considered. Using the comorbid conditions identified by Charlson and colleagues, we developed a comorbidity index that incorporates the diagnostic and procedure data contained in Medicare physician (Part B) claims. In the national cohorts of elderly prostate (n = 28,868) and breast cancer (n = 14,943) patients assessed in this study, less than 10% of patients had comorbid conditions identified when only Medicare hospital (Part A) claims were examined. By incorporating physician claims, the proportion of patients with comorbid conditions increased to 25%. The new physician claims comorbidity index significantly contributes to models of 2-year noncancer mortality and treatment received in both patient cohorts. We demonstrate the utility of a disease-specific index using an alternative method of construction employing study-specific weights. The physician claims index can be used in conjunction with a comorbidity index derived from inpatient hospital claims, or employed as a stand-alone measure.

Latent Variable Models for Mixed Discrete and Continuous Outcomes

We propose a latent variable model for mixed discrete and continuous outcomes. The model accommodates any mixture of outcomes from an exponential family and allows for arbitrary covariate effects, as well as direct modelling of covariates on the latent variable. An EM algorithm is proposed for parameter estimation and estimates of the latent variables are produced as a by-product of the analysis. A generalized likelihood ratio test can be used to test the significance of covariates affecting the latent outcomes. This method is applied to birth defects data, where the outcomes of interest are continuous measures of size and binary indicators of minor physical anomalies. Infants who were exposed in utero to anticonvulsant medications are compared with controls.

The role of prostate-specific antigen (PSA) testing patterns in the recent prostate cancer incidence decline in the United States.

Objectives: Trends in first-time and later PSA procedure rates are ascertained using longitudinal data from a population-based cohort. These trends are compared to trends in prostate cancer incidence to determine the role of PSA in the recent decline in prostate cancer incidence.Methods: Medicare data were linked with tumor registry data from the National Cancer Institutes Surveillance, Epidemiology, and End Results (SEER) Program. A 5 percent random sample (n=39985) of Medicare beneficiaries from the SEER areas without a previous diagnosis of prostate cancer as of January 1, 1988 was followed through 1994. Trends in first-time PSA use were distinguished from those of second or later for men without diagnosed prostate cancer.Results: Trends in the rate of first-time PSA procedures track closely with trends in prostate cancer incidence rates, increasing until 1992 and decreasing thereafter. Similar patterns were observed by race and age group. Geographic variability in the dissemination of PSA screening was observed, yet the association between testing and incidence remained. Men in the cohort had a 4.7 percent chance of being diagnosed within three months of an initial PSA test, with the percentage falling for subsequent tests.Conclusions: It is informative to distinguish first from later tests when assessing the effect of the diffusion of a test in a population. Taking this approach was useful in illuminating the role of PSA testing in a reversal of a long-term increase in prostate cancer incidence rates.

Prostate-specific antigen testing in black and white men: an analysis of medicare claims from 1991–1998

OBJECTIVES To describe the trends in prostate-specific antigen (PSA) use and associated cancer detection among black and white Medicare beneficiaries older than 65 years during the calendar period from January 1991 through December 1998. METHODS Medicare claims data were linked with cancer registry data from the Surveillance, Epidemiology and End Results program of the National Cancer Institute. Data from a 5% random sample of men without a diagnosis of prostate cancer were combined with data from prostate cancer cases diagnosed during the calendar period from 1991 to 1998. PSA tests conducted after a diagnosis of prostate cancer were excluded. RESULTS PSA use has stabilized among white men, reaching an annual rate of 38% by 1995 and remaining at this level through 1998. The annual rate of use among black men reached 31% by 1998, but was still increasing at that time. By 1996, at least 80% of tests in both blacks and whites were second or later tests. By the end of 1996, 35% of white men and 25% of black men were undergoing testing at least biannually or more frequently. In 1996, 83% of diagnoses in whites and 77% in blacks were preceded by a PSA test. CONCLUSIONS Older black men lag slightly behind older white men in their use of the PSA test; however, annual testing rates in blacks have yet to stabilize. In both race groups, an overwhelming majority of diagnoses are associated with a PSA test, whether for screening or diagnostic purposes. Regular screening rates in blacks are substantially lower than in whites, but the regular screening rates are relatively low in both race groups. Should PSA screening prove efficacious, efforts to promote regular use among both black and white men will likely be needed.

Efficiency and Power of Tests for Multiple Binary Outcomes

Abstract Global tests provide a useful tool for comparing two or more groups with respect to multiple correlated outcomes. We adapt and compare the performance of tests that have been suggested for use with multiple continuous outcomes to the case of multiple binary outcomes. Comparisons and guidelines are based on asymptotic relative efficiencies (AREs) and simulations. These results are illustrated using an application from teratology. We extend the work of Lefkopoulou and Ryan to include general M-group comparisons alternatives where group effects may differ for each outcome. A concise form for this general class of score tests is derived. To compute the AREs for this class of tests, we devise a useful characterization of the alternative space based on multivariate polar coordinates. Our findings indicate that the common outcome effect tests are efficient for a remarkably large range of circumstances. A simple formula applies to compute the maximum number of unaffected outcomes that can be included i...

Validation study of retrospective recall of disease-targeted function: results from the prostate cancer outcomes study.

OBJECTIVES This was an ancillary methodological study within the Prostate Cancer Outcomes Study (PCOS) to assess the validity of 6-month retrospective recall of prediagnostic disease-targeted function among men diagnosed with prostate cancer. METHODS A convenience sample of 133 prostate cancer cases were administered a baseline questionnaire shortly after diagnosis that asked about prediagnostic urinary, sexual, and bowel function. They were surveyed again concerning the same items 6 months later and asked to recall their prediagnostic function. Reports of prediagnostic function obtained at baseline and 6 months are compared, as are measures of change derived from these reports. Percent agreement and weighted kappas are calculated to measure the extent of agreement. RESULTS Over 70% of the men reported prediagnostic functioning at the highest level on 12 of 17 survey items. For each of these items, recall at 6 months was identical to the baseline survey response for > or =69% of the men. The values of the weighted kappas for changes computed with baseline reports (prospective) and changes computed with 6-month recall (retrospective) ranged from 0.396 to 0.919 for the 17 individual items. Intraclass correlations for the retrospective versus prospective changes in the multi-item function scores were 0.828 for urinary, 0.618 for bowel, and 0.692 for sexual function. CONCLUSIONS At baseline, men recently diagnosed with prostate cancer report few disease-related problems before diagnosis, and a high percentage of men recall this accurately 6 months later. There is reasonably high agreement between baseline and 6-month estimates of prediagnostic function and between prospective and retrospective measures of change over 6 months.

Latent Variable Models for Teratogenesis Using Multiple Binary Outcomes

Abstract Multiple outcomes are commonly measured in the study of birth defects. The reason is that most teratogens do not cause a single, uniquely defined defect, but rather result in a range of effects, including major malformations, minor anomalies, and deficiencies in birth weight, length and head circumference. The spectrum of effects associated with a particular teratogen is sometimes described as a “syndrome.” In this article we develop a latent variable model to characterize exposure effects on multiple binary outcomes. Not only does the method allow comparisons of control and exposed infants with respect to multiple outcomes, but it also provides a measure of the “severity” of each childs condition. Data from a study of the teratogenic effects of anticonvulsants illustrate our results.

Core Symptoms That Discriminate Premenstrual Syndrome

OBJECTIVE To identify core symptoms that discriminate premenstrual syndrome (PMS) in prospective daily diary ratings and determine the association of these symptoms with functional impairment. METHODS The study analyzed prospective daily symptom ratings and functional impairment data provided by 1081 women who requested PMS treatment at an academic medical center. The data were obtained before any treatment procedures. A random-split sample design provided separate developmental and validation datasets. Logistic regression was used to identify a reduced set of symptoms that best discriminated PMS. The results were validated in a separate dataset. Optimal cutoff points in the symptom scores were identified for clinical use. RESULTS Statistical modeling identified 6 symptoms that discriminated PMS and not PMS as well as 17 symptoms in daily diary ratings. The identified core symptoms included anxiety/tension, mood swings, aches, appetite/food cravings, cramps, and decreased interest in activities. The area under the curve (AUC) was 0.84 in both models. The sums of the premenstrual symptom scores also discriminated PMS and not PMS and correctly classified 84%-86% of the cases. CONCLUSIONS Six symptoms rated in daily diaries discriminate between PMS and not PMS among women seeking treatment and are significantly associated with functional impairment. The findings suggest that the burden of daily diaries to confirm PMS can be reduced to a smaller number of symptoms that distinguish the patients who meet this requirement. Results also support the concept that a clinical diagnosis of PMS can be developed around a core symptom group.

Social and behavioral interventions to increase breast cancer screening.

Over the past 10 years, breast cancer screening has increased greatly among women in the United States. Among women ages 50 and older, more than 56% reported having received a clinical breast exam and mammogram (X-ray images of the breast) in the 2 years preceding the 1994 National Health Interview Survey (NHIS), compared with 25% in 1987 (US. Department of Health and Human Services [DHHS], 1990). Mammography rates doubled or tripled for most ethnic groups over that period. In the past decade, the progress that has been made in breast cancer screening is striking. Although it may be impossible to demonstrate a causal relationship, the increase in screening rates surely is a result of a concerted effort by researchers, clinicians, and practitioners. The investments made by the National Institutes of Health (NIH), the Centers for Disease Control and Prevention, the American Cancer Society, and other organizations to conduct research, promote screening, and deliver services have played an important role in the increase in breast cancer screening. In this review, we examine briefly, and with more emphasis on breadth

Assessing uncertainty in microsimulation modelling with application to cancer screening interventions

Microsimulation is fast becoming the approach of choice for modelling and analysing complex processes in the absence of mathematical tractability. While this approach has been developed and promoted in engineering contexts for some time, it has more recently found a place in the mainstream of the study of chronic disease interventions such as cancer screening. The construction of a simulation model requires the specification of a model structure and sets of parameter values, both of which may have a considerable amount of uncertainty associated with them. This uncertainty is rarely quantified when reporting micro-simulation results. We suggest a Bayesian approach and assume a parametric probability distribution to mathematically express the uncertainty related to model parameters. First, we design a simulation experiment to achieve good coverage of the parameter space. Second, we model a response surface for the outcome of interest as a function of the model parameters using the simulation results. Third, we summarize the variability in the outcome of interest, including variation due to parameter uncertainty, using the response surface in combination with parameter probability distributions. We illustrate the proposed method with an application of a microsimulator designed to investigate the effect of prostate specific antigen (PSA) screening on prostate cancer mortality rates.