Screening tests are widely recommended for the early detection of disease among asymptomatic individuals. While detecting disease at an earlier stage has the potential to improve outcomes, screening also has negative consequences, including false positive results which may lead to anxiety, unnecessary diagnostic procedures, and increased healthcare costs. In addition, multiple false positive results could discourage participating in subsequent screening rounds. Screening guidelines typically recommend repeated screening over a period of many years, but little prior research has investigated how often individuals receive multiple false positive test results. Estimating the cumulative risk of multiple false positive results over the course of multiple rounds of screening is challenging due to the presence of censoring and competing risks, which may depend on the false positive risk, screening round, and number of prior false positive results. To address the general challenge of estimating the cumulative risk of multiple false positive test results, we propose a nonhomogeneous multistate model to describe the screening process including competing events. We developed alternative approaches for estimating the cumulative risk of multiple false positive results using this multistate model based on existing estimators for the cumulative risk of a single false positive. We compared the performance of the newly proposed models through simulation studies and illustrate model performance using data on screening mammography from the Breast Cancer Surveillance Consortium. Across most simulation scenarios, the multistate extension of a censoring bias model demonstrated lower bias compared to other approaches. In the context of screening mammography, we found that the cumulative risk of multiple false positive results is high. For instance, based on the censoring bias model, for a high-risk individual, the cumulative probability of at least two false positive mammography results after 10 rounds of annual screening is 40.4.

Background

Whether timeliness of follow-up after abnormal mammography differs at facilities serving vulnerable populations, such as women with limited education or income, in rural areas, and racial/ethnic minorities is unknown.

Methods

We examined receipt of diagnostic evaluation after abnormal mammography using 1998-2006 Breast Cancer Surveillance Consortium-linked Medicare claims. We compared whether time to recommended breast imaging or biopsy depended on whether women attended facilities serving vulnerable populations. We characterized a facility by the proportion of mammograms performed on women with limited education or income, in rural areas, or racial/ethnic minorities.

Results

We analyzed 30,874 abnormal screening examinations recommended for follow-up imaging across 142 facilities and 10,049 abnormal diagnostic examinations recommended for biopsy across 114 facilities. Women at facilities serving populations with less education or more racial/ethnic minorities had lower rates of follow-up imaging (4%-5% difference, P<0.05), and women at facilities serving more rural and low-income populations had lower rates of biopsy (4%-5% difference, P<0.05). Women undergoing biopsy at facilities serving vulnerable populations had longer times until biopsy than those at facilities serving nonvulnerable populations (21.6 vs. 15.6 d; 95% confidence interval for mean difference 4.1-7.7). The proportion of women receiving recommended imaging within 11 months and biopsy within 3 months varied across facilities (interquartile range, 85.5%-96.5% for imaging and 79.4%-87.3% for biopsy).

Conclusions

Among Medicare recipients, follow-up rates were slightly lower at facilities serving vulnerable populations, and among those women who returned for diagnostic evaluation, time to follow-up was slightly longer at facilities that served vulnerable population. Interventions should target variability in follow-up rates across facilities, and evaluate effectiveness particularly at facilities serving vulnerable populations.

Quality assessment is critical for healthcare reform, but data sources are lacking for measurement of many important healthcare outcomes. With over 49 million people covered by Medicare as of 2010, Medicare claims data offer a potentially valuable source that could be used in targeted health care quality improvement efforts. However, little is known about the operating characteristics of provider profiling methods using claims-based outcome measures that may estimate provider performance with error. Motivated by the example of screening mammography performance, we compared approaches to identifying providers failing to meet guideline targets using Medicare claims data. We used data from the Breast Cancer Surveillance Consortium and linked Medicare claims to compare claims-based and clinical estimates of cancer detection rate. We then demonstrated the performance of claim-based estimates across a broad range of operating characteristics using simulation studies. We found that identification of poor performing providers was extremely sensitive to algorithm specificity, with no approach identifying more than 65% of poor performing providers when claims-based measures had specificity of 0.995 or less. We conclude that claims have the potential to contribute important information on healthcare outcomes to quality improvement efforts. However, to achieve this potential, development of highly accurate claims-based outcome measures should remain a priority.

Objective

To develop and validate Medicare claims-based approaches for identifying abnormal screening mammography interpretation.

Data sources

Mammography data and linked Medicare claims for 387,709 mammograms performed from 1999 to 2005 within the Breast Cancer Surveillance Consortium (BCSC).

Study design

Split-sample validation of algorithms based on claims for breast imaging or biopsy following screening mammography.

Data extraction methods

Medicare claims and BCSC mammography data were pooled at a central Statistical Coordinating Center.

Principal findings

Presence of claims for subsequent imaging or biopsy had sensitivity of 74.9 percent (95 percent confidence interval [CI], 74.1-75.6) and specificity of 99.4 percent (95 percent CI, 99.4-99.5). A classification and regression tree improved sensitivity to 82.5 percent (95 percent CI, 81.9-83.2) but decreased specificity (96.6 percent, 95 percent CI, 96.6-96.8).

Conclusions

Medicare claims may be a feasible data source for research or quality improvement efforts addressing high rates of abnormal screening mammography.

Background

Medicare claims data may be a fruitful data source for research or quality measurement in mammography. However, it is uncertain whether claims data can accurately distinguish screening from diagnostic mammograms, particularly when claims are not linked with cancer registry data.

Objectives

To validate claims-based algorithms that can identify screening mammograms with high positive predictive value (PPV) in claims data with and without cancer registry linkage.

Research design

Development of claims-derived algorithms using classification and regression tree analyses within a random half-sample of bilateral mammogram claims with validation in the remaining half-sample.

Subjects

Female fee-for-service Medicare enrollees aged 66 years and older, who underwent bilateral mammography from 1999 to 2005 within Breast Cancer Surveillance Consortium (BCSC) registries in 4 states (CA, NC, NH, and VT), enabling linkage of claims and BCSC mammography data (N=383,730 mammograms obtained from 146,346 women).

Measures

Sensitivity, specificity, and PPV of algorithmic designation of a "screening" purpose of the mammogram using a BCSC-derived reference standard.

Results

In claims data without cancer registry linkage, a 3-step claims-derived algorithm identified screening mammograms with 97.1% sensitivity, 69.4% specificity, and a PPV of 94.9%. In claims that are linked to cancer registry data, a similar 3-step algorithm had higher sensitivity (99.7%), similar specificity (62.7%), and higher PPV (97.4%).

Conclusions

Simple algorithms can identify Medicare claims for screening mammography with high predictive values in Medicare claims alone and in claims linked with cancer registry data.

Multistate models are used to characterize individuals' natural histories through diseases with discrete states. Observational data resources based on electronic medical records pose new opportunities for studying such diseases. However, these data consist of observations of the process at discrete sampling times, which may either be pre-scheduled and non-informative, or symptom-driven and informative about an individual's underlying disease status. We have developed a novel joint observation and disease transition model for this setting. The disease process is modeled according to a latent continuous-time Markov chain; and the observation process, according to a Markov-modulated Poisson process with observation rates that depend on the individual's underlying disease status. The disease process is observed at a combination of informative and non-informative sampling times, with possible misclassification error. We demonstrate that the model is computationally tractable and devise an expectation-maximization algorithm for parameter estimation. Using simulated data, we show how estimates from our joint observation and disease transition model lead to less biased and more precise estimates of the disease rate parameters. We apply the model to a study of secondary breast cancer events, utilizing mammography and biopsy records from a sample of women with a history of primary breast cancer.

Background

This study illustrates alternative statistical methods for estimating cumulative risk of screening mammography outcomes in longitudinal studies.

Methods

Data from the US Breast Cancer Surveillance Consortium (BCSC) and the Nijmegen Breast Cancer Screening Program in the Netherlands were used to compare four statistical approaches to estimating cumulative risk. We estimated cumulative risk of false-positive recall and screen-detected cancer after 10 screening rounds using data from 242,835 women ages 40 to 74 years screened at the BCSC facilities in 1993-2012 and from 17,297 women ages 50 to 74 years screened in Nijmegen in 1990-2012.

Results

In the BCSC cohort, a censoring bias model estimated bounds of 53.8% to 59.3% for false-positive recall and 2.4% to 7.6% for screen-detected cancer, assuming 10% increased or decreased risk among women screened for one additional round. In the Nijmegen cohort, false-positive recall appeared to be associated with subsequent discontinuation of screening leading to overestimation of risk of a false-positive recall based on adjusted discrete-time survival models. Bounds estimated by the censoring bias model were 11.0% to 19.9% for false-positive recall and 4.2% to 9.7% for screen-detected cancer.

Conclusion

Choice of statistical methodology can substantially affect cumulative risk estimates. The censoring bias model is appropriate under a variety of censoring mechanisms and provides bounds for cumulative risk estimates under varying degrees of dependent censoring.

Impact

This article illustrates statistical methods for estimating cumulative risks of cancer screening outcomes, which will be increasingly important as screening test recommendations proliferate.

Background

While Medicare claims are a potential resource for clinical mammography research or quality monitoring, the validity of key data elements remains uncertain. Claims codes for digital mammography and computer-aided detection (CAD), for example, have not been validated against a credible external reference standard.

Methods

We matched Medicare mammography claims for women who received bilateral mammograms from 2003 to 2006 to corresponding mammography data from the Breast Cancer Surveillance Consortium (BCSC) registries in four U.S. states (N = 253,727 mammograms received by 120,709 women). We assessed the accuracy of the claims-based classifications of bilateral mammograms as either digital versus film and CAD versus non-CAD relative to a reference standard derived from BCSC data.

Results

Claims data correctly classified the large majority of film and digital mammograms (97.2% and 97.3%, respectively), yielding excellent agreement beyond chance (κ = 0.90). Claims data correctly classified the large majority of CAD mammograms (96.6%) but a lower percentage of non-CAD mammograms (86.7%). Agreement beyond chance remained high for CAD classification (κ = 0.83). From 2003 to 2006, the predictive values of claims-based digital and CAD classifications increased as the sample prevalences of each technology increased.

Conclusion

Medicare claims data can accurately distinguish film and digital bilateral mammograms and mammograms conducted with and without CAD.

Impact

The validity of Medicare claims data regarding film versus digital mammography and CAD suggests that these data elements can be useful in research and quality improvement.

Purpose

Oral contraceptive (OC) use can occur throughout a woman's reproductive life span with the potential for long-term impacts on health. To assess potential measurement error in prior OC use, this study compared level of agreement between self-reported prior OC use and pharmacy dispensing data in perimenopausal and/or early postmenopausal women.

Methods

The study's 1399 women (ages, 45-59 years) were participants in a population-based case-control study of the association between OC use and fracture risk. Episodes of lifetime self-reported OC use (in months) were collected, by telephone interview, for January 1, 2008 through November 25, 2012. Pharmacy fills, back to 1980, were collected from automated data. Agreement was measured using the prevalence-adjusted and bias-adjusted kappa index.

Results

The number of women with OC pharmacy fills was 11% to 45% higher than those who reported OC use during each time period. Between-measures agreement was better for more recent use. Prevalence-adjusted and bias-adjusted kappa index values ranged from 0.88 (95% confidence interval, 0.85-0.90) within 5 years from the reference date to 0.65 (95% confidence interval, 0.59-0.71) within 15 to 20 years.

Conclusions

For studies designed to assess the long-term effects of OC use, the current results are reassuring in noting moderate agreement between self-reported OC use and pharmacy data for up to 15 to 20 years before the interview.

Objective

The effect of oral contraceptive (OC) use on risk of fracture remains unclear, and use during later reproductive life may be increasing. To determine the association between OC use during later reproductive life and risk of fracture across the menopausal transition, we conducted a population-based case-control study in a Pacific Northwest HMO, Group Health Cooperative.

Methods

For the January 2008 to March 2013 interval, 1,204 case women aged 45 to 59 years with incident fractures, and 2,275 control women were enrolled. Potential cases with fracture codes in automated data were adjudicated by electronic health record review. Potential control women without fracture codes were selected concurrently, sampling based on age. Participants received a structured study interview. Using logistic regression, associations between OC use and fracture risk were calculated as odds ratios (ORs) and 95% confidence intervals (CIs).

Results

Participation was 69% for cases and 64% for controls. The study sample was 82% white; mean age was 54 years. The most common fracture site for cases was the wrist/forearm (32%). Adjusted fracture risk did not differ between cases versus controls for OC use in the 10 years before menopause (OR 0.90, 95% CI 0.74, 1.11); for OC use after age 38 (OR 0.94, 95% CI 0.78, 1.14); for duration of use, or for other OC exposures.

Conclusions

The current study does not show an association between fractures near the menopausal transition and OC use in the decade before menopause or after age 38. For women considering OC use at these times, fracture risk does not seem to be either reduced or-reassuringly-increased.