Carrie N. Klabunde

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.

Overview of the SEER-Medicare data: content, research applications, and generalizability to the United States elderly population.

Background. The Surveillance, Epidemiology and End Results (SEER)-Medicare–linked database combines clinical information from population-based cancer registries with claims information from the Medicare program. The use of this database to study cancer screening, treatment, outcomes, and costs has grown in recent years. Research Design. This paper provides an overview of the SEER-Medicare files for investigators interested in using these data for epidemiologic and health services research. The overview includes a description of the linkage of SEER and Medicare data and the files included as part of SEER-Medicare. The paper also describes the types of research projects that have been undertaken using the SEER-Medicare data. The overview concludes with a comparison of selected characteristics of elderly persons residing in the SEER areas to the US total aged. Results. The paper identifies a number of potential uses of the SEER-Medicare data. The comparison of the elderly population in SEER areas to the US total shows that in the SEER areas there are a lower percentage of white persons and individuals living in poverty, and a higher percentage of urban-dwellers than the US total. Elderly persons in the SEER regions also have higher rates of HMO enrollment and lower rates of cancer mortality. Conclusions. The SEER-Medicare data are a unique resource that can be used for a variety of health services research projects. Although there are some differences between the elderly residing in the SEER areas and the US total, the SEER-Medicare data offer a large population-based cohort that can be used to longitudinally track care for persons over the course of cancer diagnosis, treatment, and follow-up.

Patterns of Colorectal Cancer Screening Uptake among Men and Women in the United States

Objective: The purpose of this report is to examine (a) gender-specific correlates of colorectal cancer test use using recent national data from 2003 and (b) patterns of colorectal cancer screening by gender and test modality over time. Methods: We analyze data from the 1987, 1992, 1998, 2000, and 2003 National Health Interview Surveys. Our sample consists of men and women ≥50 years never diagnosed with colorectal cancer and who reported a recent fecal occult blood test and/or endoscopy. Results: In 2003, both men and women reported higher rates of colonoscopy (32.2% and 29.8%, respectively) than use of FOBT (16.1% and 15.3%, respectively) or sigmoidoscopy (7.6% and 5.9%, respectively). Men reported higher use of endoscopy than women if they had a usual source of health care, had talked to a general doctor, and had two to five visits to the doctor in the past year. Men and women 65 years and older had higher rates of any recommended colorectal cancer test (55.8% and 48.5%, respectively) than persons 50 to 64 years (males, 41.0%; females, 31.4%). Use of colorectal cancer tests also was higher among both genders if they were not Hispanic, had higher educational attainment, were former smokers, had health insurance or a usual source of care, or if they talked to a general doctor. Recent use of colorectal cancer tests has increased since 2000 for both women and men largely due to increased use of colonoscopy. Conclusions: Colorectal cancer testing is increasing for both men and women, although the prevalence of testing remains higher in men. Our data support previous findings documenting socioeconomic disparities in colorectal cancer test use. Access barriers to screening could be particularly difficult to overcome if colonoscopy becomes the preferred colorectal cancer screening modality. (Cancer Epidemiol Biomarkers Prev 2006;15(2):389–94)

Barriers to colorectal cancer screening: A comparison of reports from primary care physicians and average-risk adults

Background:Barriers to colorectal cancer (CRC) screening are not well understood. Objectives:We sought to compare barriers to CRC screening reported by primary care physicians (PCPs) and by average-risk adults, and to examine characteristics of average-risk adults who identified lack of provider recommendation as a major barrier to CRC screening. Research Design:This was a comparative study using data from the 1999–2000 Survey of Colorectal Cancer Screening Practices and the 2000 National Health Interview Survey (NHIS). Subjects:We recruited nationally representative samples of PCPs (n= 1235) from the SCCSP and average-risk adults (n = 6497) from the NHIS. Measures:We measured barriers to CRC screening identified by PCPs and average-risk adults who were not current with screening. Results:Both PCPs and average-risk adults identified lack of patient awareness and physician recommendation as key barriers to obtaining CRC screening. PCPs also frequently cited patient embarrassment/anxiety about testing and test cost/lack of insurance coverage, but few adults identified these as major barriers. Of adults not current with testing, those who had visited a doctor in the past year or had health insurance were more likely to report lack of physician recommendation as the main reason they were not up-to-date compared with their counterparts with no doctor visit or health insurance. Only 10% of adults not current with testing and who had a doctor visit in the past year reported receiving a screening recommendation. Conclusions:A need exists for continued efforts to educate the public about CRC and the important role of screening in preventing this disease. Practice-based strategies to systematically prompt health care providers to discuss CRC screening with eligible patients also are required.

Colorectal Cancer Test Use from the 2005 National Health Interview Survey

Background: Screening is effective in reducing colorectal cancer mortality. Recommended colorectal cancer screening options include a home fecal occult blood test (FOBT) or colorectal endoscopy (sigmoidoscopy or colonoscopy). Past surveys have indicated that colorectal cancer screening prevalence in the United States is low. The purpose of this analysis was to determine the prevalence of colorectal cancer test use in the United States by various factors and to examine reasons for not having a colorectal cancer test. Methods: Data on respondents ages ≥50 years from the 2005 National Health Interview Survey (n = 13,269) were analyzed. The proportion of the U.S. population that had home FOBT within the past year or endoscopy within the past 10 years was examined by sociodemographic, health-care access, and other health-related factors. Reported reasons for not having FOBT or endoscopy were also analyzed. Results: The age-standardized proportion of respondents who reported FOBT within the past year and/or endoscopy within the past 10 years was 50.0% [95% confidence interval (95% CI), 48.8-51.2]. Colorectal cancer testing rates were particularly low among people without health-care coverage (24.1%; 95% CI, 19.2-29.7) or without a usual source of health care (24.7%; 95% CI, 20.8-29.0). The most commonly reported reason for not having a colorectal cancer test was “never thought about it.” Conclusions: In 2005, about half of Americans ages ≥50 years did not have appropriate colorectal cancer testing. Increased efforts to expand health-care coverage or to provide colorectal cancer tests to people without health-care coverage are needed to increase colorectal cancer screening. (Cancer Epidemiol Biomarkers Prev 2008;17(7):1623–30)

A National Survey of Primary Care Physicians' Methods for Screening for Fecal Occult Blood

Context Authorities recommend at-home fecal occult blood tests (FOBTs), using 2 samples from 3 consecutive bowel movements, as one of several options for colorectal cancer screening. Content The authors asked a national sample of 1147 primary care physicians about their colorectal cancer screening practices. Approximately 33% reported using only a single in-office stool sample for FOBT. To follow up positive results, approximately 30% said that they repeated the FOBT, and about the same percentage recommended sigmoidoscopy rather than total colon examination. Implications Many primary care physicians in the United States may use inadequate methods to screen for colorectal cancer. The Editors The fecal occult blood test (FOBT) is the least expensive and simplest of the tests recommended in national guidelines for colorectal cancer screening (1-6). Nevertheless, implementation requires several decisions that determine the effectiveness of screening. These include the brand of FOBT to use, whether to give patients the multiple-specimen home kit or to conduct in-office FOBT with a single specimen obtained during a digital rectal examination, and selection of a follow-up test for patients with positive results (7). While randomized, controlled trials have shown that FOBT reduces colorectal cancer incidence and mortality (8-13), the effectiveness and cost-effectiveness of FOBT screening in general clinical practice depend on the degree to which health professionals follow recommended testing guidelines. National colorectal cancer screening and clinical practice guidelines (1-6, 14) recommend performing FOBT screening with the home test, the method used in the randomized, controlled trials. To our knowledge, no studies have examined the efficacy of in-office FOBT, and a recent study showed that the in-office test is substantially less sensitive than the home test (15). Patients with positive results on FOBT should have an examination of the entire colon and rectum (2-7). Repeating FOBT before diagnostic work-up to increase test specificity is not recommended because of the significant probability of finding colorectal cancer or a large adenoma among patients with a positive test result (7). We report on the practice of FOBT in the United States as ascertained in the national Survey of Colorectal Cancer Screening Practices (SCCSP) in Health Care Organizations. This survey asked primary care physicians about how they used FOBT in their practices (16). To determine whether physician responses were consistent with those of the public, we also examined data from the nationally representative sample of the U.S. population that completed the 2000 National Health Interview Survey (NHIS) (17). Methods Survey of Colorectal Cancer Screening Practices The National Cancer Institute conducted the SCCSP in collaboration with the Centers for Disease Control and Prevention and the Centers for Medicare & Medicaid Services. The researchers used the American Medical Associations Physician Masterfile to form a nationally representative sample of primary care physicians, including general practitioners, family practitioners, general internists, and obstetriciansgynecologists. The survey took place between November 1999 and April 2000. Eligible respondents were physicians 75 years of age or younger who had an active license to practice medicine and whose major professional activity was patient care. Physicians who were retired; in residency training; or involved in full-time teaching, research, or administration were excluded. Obstetriciansgynecologists were included in the sample because they provide preventive services for many women in the United States (18). Respondents were offered a choice of response modes (mail, fax, telephone, or secure Internet Web site). Ninety percent responded by mail. Additional details of the sampling and survey methods have been described elsewhere (16). Physicians were asked how many times they ordered or performed colorectal cancer screening with FOBT for asymptomatic average-risk patients. Respondents who ordered or performed FOBT at least once during a typical month were also asked the brand of test they used, whether they used in-office FOBT, home FOBT, or both (Complete a single FOBT card in the office during a digital rectal exam, Give or mail patients a set of three FOBT cards to complete at home, Both of the above, or Other [describe]) for most of their patients, and whether they had a mechanism to ensure that patients who were given or mailed home FOBT kits completed and returned them. Physicians were asked which procedure or procedures they usually recommended to healthy, average-risk patients as an initial follow-up test after positive results on FOBT: repeated FOBT, flexible sigmoidoscopy, colonoscopy, double-contrast barium enema (DCBE), or other (describe). Respondents were allowed to check more than 1 procedure, indicated in this report by test a/test b (for example, sigmoidoscopy/colonoscopy). We analyzed follow-up of positive FOBT results in 2 steps (Figure). Among the 1120 physicians who performed FOBT at least once per month and responded to the follow-up question, we looked first at whether they recommended repeating the FOBT. We then examined which tests were recommended for the diagnostic work-up. At this second step, we did not include respondents who only reported recommending repeated FOBT (n= 151) and those who only indicated that they referred patients to another physician (specialist or primary care) for follow-up of positive FOBT results (n= 44). Figure. Two-step analysis of follow-up of positive results on fecal occult blood tests ( FOBTs ). The American Cancer Society and the GI Consortium have provided explicit guidelines on method of specimen collection and follow-up of positive test results (2, 3). We considered physicians to be influenced by guidelines if they responded that at least 1 of these guidelines was very influential in their screening recommendations. Respondents were also asked to comment on the capacity to perform colonoscopy in their geographic practice area and whether, in the past 3 years, they had completed continuing medical education courses on colorectal cancer screening. NHIS The NHIS, conducted by the Centers for Disease Control and Preventions National Center for Health Statistics, is an in-person survey that collects health behavior and sociodemographic information from the civilian, noninstitutionalized U.S. population by using multistage sampling. The 2000 NHIS Cancer Control Module included questions on the use of cancer screening. The final response rate for the module was 72% (17). A total of 11365 men and women 50 years of age or older responded to questions about FOBT use. Respondents were asked whether they had ever had a blood stool test using a home test kit, the time since their last test, and whether they had ever had a home blood stool test with abnormal results. Those responding affirmatively were asked what additional tests or surgery they had because of these results. After the series of questions related to home FOBT, respondents were asked if they had ever had an in-office test and how much time had passed since their last test. Respondents were not asked about abnormal results on in-office tests. Statistical Analysis We conducted univariate and multivariate analyses of data from the SCCSP. For the univariate analysis, we estimated percentages and confidence intervals for each level of an explanatory variable. We performed a Pearson chi-square test to test the association between each explanatory variable and the response variable for Tables 1 and 2. For the multivariate analysis, we used logistic regression and computation of predictive margins to estimate the probability of each outcome variable when we controlled for all other independent variables in Tables 1, 2, and 3. These adjusted percentages (that is, predictive margins) are a type of direct standardization that averages the predicted values from the logistic regression models over the covariate distribution in the population (19). These adjusted percentages do not reflect actual use in the population but allow comparisons across categories of the variables included in the models. They are easier to interpret than odds ratios and do not require designating one of the groups as the referent group. Table 1. Percentage of Physicians Who Use In-Office Fecal Occult Blood Tests Table 2. Percentage of Physicians Who Repeat Fecal Occult Blood Tests after Abnormal Results Table 3. Percentage of Physicians Who Recommend Alternative Diagnostic Work-up after Positive Results on Fecal Occult Blood Tests We did a Wald chi-square test to test the association between each explanatory variable and the response variable while adjusting for all other variables in the table. To permit generalization of the results of both the univariate and multivariate analyses to the U.S. population of practicing primary care physicians, we used sampling weights to account for the probability of selection and nonresponse. We used the SAS statistical package, version 9.0 (SAS Institute, Inc., Cary, North Carolina) (20), and SUDAAN, version 8.0.2 (Research Triangle Institute, Research Triangle Park, North Carolina) (21), to compute prevalence rates, predictive margins, confidence intervals, and P values. We conducted univariate analyses of the NHIS data, using SUDAAN to compute point estimates and confidence intervals to account for the complex survey design (17). Role of the Funding Sources The funding sources, the National Cancer Institute and the Centers for Disease Control and Prevention, had a role in the design, conduct, and reporting of the study. Results Survey of Colorectal Cancer Screening Practices Description of Respondents Of the 1718 eligible physicians, 1235 responded to the survey (overall response rate, 72%). Personal and practice characteristics of these respondents have been described el

Trends in Colorectal Cancer Test Use among Vulnerable Populations in the United States

Background: Evaluating trends in colorectal cancer (CRC) screening use is critical for understanding screening implementation, and whether population groups targeted for screening are receiving it, consistent with guidelines. This study examines recent national trends in CRC test use, including among vulnerable populations. Methods: We used the 2000, 2003, 2005, and 2008 National Health Interview Survey to examine national trends in CRC screening use overall and for fecal occult blood test (FOBT), sigmoidoscopy, and colonoscopy. We also assessed trends by race/ethnicity, educational attainment, income, time in the United States, and access to health care. Results: During 2000 to 2008, significant declines in FOBT and sigmoidoscopy use and significant increases in colonoscopy use and in the percentages of adults up-to-date with CRC screening occurred overall and for most population subgroups. Subgroups with consistently lower rates of colonoscopy use and being up-to-date included Hispanics; people with minimal education, low income, or no health insurance; recent immigrants; and those with no usual source of care or physician visits in the past year. Among up-to-date adults, there were few subgroup differences in the type of test by which they were up-to-date (i.e., FOBT, sigmoidoscopy, or colonoscopy). Conclusions: Although use of CRC screening and colonoscopy increased among U.S. adults, including those from vulnerable populations, 45% of adults aged 50 to 75—or nearly 35 million people—were not up-to-date with screening in 2008. Impact: Continued monitoring of CRC screening rates among population subgroups with consistently low utilization is imperative. Improvement in CRC screening rates among all population groups in the United States is still needed. Cancer Epidemiol Biomarkers Prev; 20(8); 1611–21. ©2011 AACR.

Differences Between Primary Care Physicians’ and Oncologists’ Knowledge, Attitudes and Practices Regarding the Care of Cancer Survivors

BackgroundThe growing number of cancer survivors combined with a looming shortage of oncology specialists will require greater coordination of post-treatment care responsibilities between oncologists and primary care physicians (PCPs). However, data are limited regarding these physicians’ views of cancer survivors’ care.ObjectiveTo compare PCPs and oncologists with regard to their knowledge, attitudes, and practices for follow-up care of breast and colon cancer survivors.Design and SubjectsMailed questionnaires were completed by a nationally representative sample of 1,072 PCPs and 1,130 medical oncologists in 2009 (cooperation rate = 65%). Sampling and non-response weights were used to calculate estimates to reflect practicing US PCPs and oncologists.Main MeasuresPCPs and oncologists reported their 1) preferred model for delivering cancer survivors’ care; 2) assessment of PCPs’ ability to perform follow-up care tasks; 3) confidence in their knowledge; and 4) cancer surveillance practices.Key ResultsCompared with PCPs, oncologists were less likely to believe PCPs had the skills to conduct appropriate testing for breast cancer recurrence (59% vs. 23%, P < 0.001) or to care for late effects of breast cancer (75% vs. 38%, P < 0.001). Only 40% of PCPs were very confident of their own knowledge of testing for recurrence. PCPs were more likely than oncologists to endorse routine use of non-recommended blood and imaging tests for detecting cancer recurrence, with both groups departing substantially from guideline recommendations.ConclusionThere are significant differences in PCPs’ and oncologists’ knowledge, attitudes, and practices with respect to care of cancer survivors. Improving cancer survivors’ care may require more effective communication between these two groups to increase PCPs’ confidence in their knowledge, and must also address oncologists’ attitudes regarding PCPs’ ability to care for cancer survivors.

Trends and black/white differences in treatment for nonmetastatic prostate cancer.

OBJECTIVES Controversy and uncertainty surround use of radical prostatectomy, radiation therapy, and conservative symptomatic management in treating elderly men with nonmetastatic prostate cancer. Prior studies have demonstrated variations in use of these therapies by patient age, race, and geographic region. This study examined trends in treatment for nonmetastatic prostate cancer in black and white men aged 65 and older during the period 1986 to 1993. The study also explored factors related to use of initial therapies in these men. METHODS A cohort of 52,915 men (48,410 white; 4,505 black) obtained from the linked SEER-Medicare dataset was used in an observational design. Various sociodemographic and clinical measures were incorporated in the analysis. RESULTS For both races, use of aggressive therapy had increased with time, although this trend appears to be slowing. Black men were less likely to undergo radical prostatectomy than were white men, but use of radiation therapy did not differ markedly by race. High socioeconomic status and a lack of comorbid conditions were among the factors predictive of aggressive therapy receipt. The relation between race and receipt of aggressive therapy was dependent on whether prostate cancer was detected by transurethral resection of the prostate. Sociodemographic and clinical characteristics explained approximately half the difference between black men and white men in radical prostatectomy use. CONCLUSIONS This study documents racial differences and changing practice patterns in the treatment of nonmetastatic prostate cancer in elderly men. Further research is required to more fully understand reasons for racial differences, as well as to promote rational use of health care resources.

In search of the perfect comorbidity measure for use with administrative claims data : Does it exist?

Background:Numerous measures of comorbidity have been developed for health services research with administrative claims. Objective:We sought to compare the performance of 4 claims-based comorbidity measures. Research Design and Subjects:We undertook a retrospective cohort study of 5777 Medicare beneficiaries ages 66 and older with stage III colon cancer reported to the Surveillance, Epidemiology, and End Results Program between January 1, 1992 and December 31, 1996. Measures:Comorbidity measures included Elixhauser’s set of 30 condition indicators, Klabunde’s outpatient and inpatient indices weighted for colorectal cancer patients, Diagnostic Cost Groups, and the Adjusted Clinical Group (ACG) System. Outcomes included receipt of adjuvant chemotherapy and 2 year noncancer mortality. Results:For all measures, greater comorbidity significantly predicted lower receipt of chemotherapy and higher noncancer death. Nested logistic regression modeling suggests that using more claims sources to measure comorbidity generally improves the prediction of chemotherapy receipt and noncancer death, but depends on the measure type and outcome studied. All 4 comorbidity measures significantly improved the fit of baseline regression models for both chemotherapy receipt (baseline c-statistic 0.776; ranging from 0.779 after adding ACGs and Klabunde to 0.789 after Elixhauser) and noncancer death (baseline c-statistic 0.687; ranging from 0.717 after adding ACGs to 0.744 after Elixhauser). Conclusions:Although some comorbidity measures demonstrate minor advantages over others, each is fairly robust in predicting both chemotherapy receipt and noncancer death. Investigators should choose among these measures based on their availability, comfort with the methodology, and outcomes of interest.