The breast cancer detection rate is a benchmark measure of screening mammography quality, but its computation requires linkage of mammography interpretive performance information with cancer incidence data. A Medicare claims-based measure of detected breast cancers could simplify measurement of this benchmark and facilitate mammography quality assessment and research.
To validate a claims-based algorithm that can identify with high positive predictive value (PPV) incident breast cancers that were detected at screening mammography.
Development of a claims-derived algorithm using classification and regression tree analyses within a random half-sample of Medicare screening mammography claims followed by validation of the algoritm in the remaining half-sample using clinical data on mammography results and cancer incidence from the Breast Cancer Surveillance Consortium (BCSC).
Female fee-for-service Medicare enrollees age 68 years and older who underwent screening mammography from 2001 to 2005 within BCSC registries in four states (CA, NC, NH, and VT), enabling linkage of claims and BCSC mammography data (N=233,044 mammograms obtained by 104,997 women).
Sensitivity, specificity, and PPV of algorithmic identification of incident breast cancers that were detected by radiologists relative to a reference standard based on BCSC mammography and cancer incidence data.
An algorithm based on subsequent codes for breast cancer diagnoses and treatments and follow-up mammography identified incident screen-detected breast cancers with 92.9% sensitivity (95% CI: 91.0%-94.8%), 99.9% specificity (95% CI: 99.9%-99.9%), and a PPV of 88.0% (95% CI: 85.7%-90.4%).
A simple claims-based algorithm can accurately identify incident breast cancers detected at screening mammography among Medicare enrollees. The algorithm may enable mammography quality assessment using Medicare claims alone.
Breast Cancer Screening; Mammography; Validation Studies; Medicare; Quality Assessment
Accurate identification of hepatocellular cancer (HCC) cases from automated data is needed for efficient and valid quality improvement initiatives and research. We validated HCC ICD-9 codes, and evaluated whether natural language processing (NLP) by the Automated Retrieval Console (ARC) for document classification improves HCC identification.
We identified a cohort of patients with ICD-9 codes for HCC during 2005–2010 from Veterans Affairs administrative data. Pathology and radiology reports were reviewed to confirm HCC. The positive predictive value (PPV), sensitivity, and specificity of ICD-9 codes were calculated. A split validation study of pathology and radiology reports was performed to develop and validate ARC algorithms. Reports were manually classified as diagnostic of HCC or not. ARC generated document classification algorithms using the Clinical Text Analysis and Knowledge Extraction System. ARC performance was compared to manual classification. PPV, sensitivity, and specificity of ARC were calculated.
1138 patients with HCC were identified by ICD-9 codes. Based on manual review, 773 had HCC. The HCC ICD-9 code algorithm had a PPV of 0.67, sensitivity of 0.95, and specificity of 0.93. For a random subset of 619 patients, we identified 471 pathology reports for 323 patients and 943 radiology reports for 557 patients. The pathology ARC algorithm had PPV of 0.96, sensitivity of 0.96, and specificity of 0.97. The radiology ARC algorithm had PPV of 0.75, sensitivity of 0.94, and specificity of 0.68.
A combined approach of ICD-9 codes and NLP of pathology and radiology reports improves HCC case identification in automated data.
Hepatocellular cancer; natural language processing
To conduct a systematic literature review to determine if there were any intervention strategies that had any measurable effect on acute-care hospitalizations among community-dwelling adults with dementia.
Studies were identified by a professional research librarian and content experts.
Participants were diagnosed with dementia, severity ranging from mild to severe, and were recruited from health care and community agencies.
A study met the inclusion criteria if it: (a) was published in English; (b) included a control or comparison group; (c) published outcome data from the intervention under study; (d) reported hospitalization as one of the outcomes; (e) included community-dwelling older adults; and (f) enrolled participants with dementia. Ten studies met all inclusion criteria.
Of the 10 studies included, most assessed health services use (ie, hospitalizations) as a secondary outcome. Participants were recruited from a range of health care and community agencies, and most were diagnosed with dementia with severity ratings ranging from mild to severe. Most intervention strategies consisted of face-to-face assessments of the persons living with dementia, their caregivers, and the development and implementation of a care plan. A significant reduction in hospital admissions was not found in any of the included studies, although 1 study did observe a reduction in hospital days.
The majority of studies included hospitalizations as a secondary outcome. Only 1 intervention was found to have an effect on hospitalizations. Future work would benefit from strategies specifically designed to reduce and prevent acute hospitalizations in persons with dementia.
older adults; hospitalizations; systematic review; intervention studies; patient care management; case management; dementia; Alzheimer disease
Electronic health information routinely collected during healthcare delivery and reimbursement can help address the need for evidence about the real-world effectiveness, safety, and quality of medical care. Often, distributed networks that combine information from multiple sources are needed to generate this real-world evidence.
We provide a set of field-tested best practices and a set of recommendations for data quality checking for comparative effectiveness research (CER) in distributed data networks.
Explore the requirements for data quality checking and describe data quality approaches undertaken by several existing multi-site networks.
There are no established standards regarding how to evaluate the quality of electronic health data for CER within distributed networks. Data checks of increasing complexity are often employed, ranging from consistency with syntactic rules to evaluation of semantics and consistency within and across sites. Temporal trends within and across sites are widely used, as are checks of each data refresh or update. Rates of specific events and exposures by age group, sex, and month are also common.
Secondary use of electronic health data for CER holds promise but is complex, especially in distributed data networks that incorporate periodic data refreshes. The viability of a learning health system is dependent on a robust understanding of the quality, validity, and optimal secondary uses of routinely collected electronic health data within distributed health data networks. Robust data quality checking can strengthen confidence in findings based on distributed data network.
Comparative effectiveness research; distributed research network; data quality
Approximately 15% of the 4 million annual U.S. births occur in rural hospitals.
To 1) measure differences in obstetric care in rural and urban hospitals and 2) examine whether trends over time differ by rural-urban hospital location
Research Design and Subjects
Retrospective analysis of hospital discharge records for all births in the 2002-2010 Nationwide Inpatient Sample, a 20% sample of U.S. hospitals (N=7,188,972 births; 6,316,743 in urban hospitals, 837,772 in rural hospitals).
Rates of low-risk cesarean (full-term, singleton, vertex pregnancies; no prior cesarean), vaginal birth after cesarean (VBAC), non-indicated cesarean, and non-indicated labor induction.
In 2010, low-risk cesarean rates in rural and urban hospitals were 15.5% and 16.1%, respectively, and non-indicated cesarean rates were 16.9% and 17.8%. VBAC rates were 5.0% in rural and 10.0% in urban hospitals in 2010. Between 2002 and 2010, rates of low-risk cesarean and non-indicated cesarean increased, and VBAC rates decreased in both rural and urban hospitals. Non-indicated labor induction was less frequent in rural versus urban hospitals in 2002 (AOR=0.79 [0.78-0.81]), but increased more rapidly in rural hospitals from 2002-2010 (AOR=1.05 [1.05-1.06]). In 2010, 16.5% of rural births were induced without indication (12.0% of urban births).
From 2002-2010, cesarean rates rose and VBAC rates fell in both rural and urban hospitals. Non-indicated labor induction rates rose disproportionately faster in rural vs. urban settings. Tailored clinical and policy tools are needed to address differences between rural and urban hospitals.
rural health; urban health; obstetrics; cesarean; labor induction
Growing numbers of reproductive-age U.S. women with chronic physical disabilities (CPD) are becoming pregnant. Little is known about the health conditions causing their CPD.
To identify health conditions causing CPD among reproductive-age women and specifically among currently pregnant women.
Cross-sectional, nationally-representative National Health Interview Survey data from 2006–2011.
6,043 civilian, noninstitutionalized women ages 18–49 with CPD
NHIS asks about various movement difficulties and their underlying causes and about current pregnancy. We used responses from 8 movement difficulty and other questions to identify women with mobility difficulties caused by chronic physical health conditions.
Among women with CPD, 2.0% report current pregnancy, with pregnancy rates falling monotonically as CPD severity rises. Regardless of pregnancy, 20.8% report 2 causes for their CPD, and 12.7% report 3+ causes; the most common causes are arthritis, back or neck problems, and other musculoskeletal conditions. Compared with nonpregnant women, currently pregnant women report fewer causal conditions: 15.8% report 2 causes and 8.0% 3+ causes; back or neck problems are reported most frequently, followed by musculoskeletal problems and arthritis. Multivariable logistic regression analyses predicting current pregnancy controlling for age category and individual common causes of CPD found that no cause was significantly associated with higher or lower adjusted odds of pregnancy.
Almost one-quarter of currently pregnant women with CPD report more than one cause for their disability. Further research will need to examine how obstetrical needs vary for depending on the cause of a pregnant woman’s disability.
disability; pregnancy; mobility; comorbidities; National Health Interview Survey
Translational research is increasingly important as academic health centers transform themselves to meet new requirements of NIH funding. Most attention has focused on T1 translation studies (bench to bedside) with considerable uncertainty about how to enhance T2 (effectiveness trials) and especially T3 (implementation studies).
To describe an innovative example of a T3 study, conducted as partnership research with the leaders of a major natural experiment in Minnesota to improve the primary care of depression.
All health plans in the state have agreed on a new payment model to support clinics that implement the well-evidenced collaborative care model for depression in the DIAMOND Initiative (Depression Improvement Across Minnesota: Offering a New Direction). The DIAMOND Study was developed in an ongoing partnership with Initiative leaders from seven health plans, 85 clinics, and a regional quality improvement collaborative to evaluate the implementation and its impacts on patients and other stakeholders. We agreed upon a staggered implementation, multiple baseline research design, utilizing the concepts of practical clinical trials and engaged scholarship and have collaborated on all aspects of conducting the study, including joint identification of patient and clinic survey recipients.
Complex study methods have worked well through 20 months because of the commitment of all stakeholders to both the Initiative and study. Over 1,500 patient subjects have been recruited from health plan information delivered weekly and 99.7% of 316 physicians and administrators from all participating clinical organizations have completed Study surveys.
Partnership research can greatly facilitate translational research studies.
Although prevention of complications in diabetes requires careful control over many years, little is known about which patients persistently fail to get recommended care.
To determine the frequency and correlates of persistent long-term gaps in diabetes care.
Patient surveys and reviews of medical records were used to assess preventive care services for diabetes among 8392 patients who were continuously enrolled in 10 US managed care plans from 1999 to 2002. Demographic and socioeconomic characteristics, access to care, social support, and mental and physical health were determined by interview. Five preventive care services of diabetes care (testing of hemoglobin A1c, cholesterol, and albuminuria, dilated eye exams, and foot exams) were assessed by survey and chart abstraction for a 3-year period (1999–2002). We defined a “persistent lapse” as a participant’s missing a preventive care service for the entire 3 years.
In all, 70% of patients had no persistent lapses, 22% had 1, 6% had 2, and 2% had ≥ 3. Persistent lapses occurred most often for lipid testing (11.6%), microalbuminuria testing (9.7%), and eye exams (9.0%), but less frequently for foot exams (6.9%) and A1c tests (4.2%). In multivariate analyses, the odds of a persistent lapse in care was 42% higher for young (age 18–44) than middle aged persons and 26% higher among lean than very obese persons. In addition, the odds of a persistent lapse was 26% higher for those of low income, 29% higher among employed persons, 18% higher for smokers, 27% higher in those with fewer than 5 years of diabetes than those with > 15 years, and 42% higher for persons with zero or 1 comorbid conditions (compared to ≥ 3). In addition, non-Hispanic blacks were particularly likely to miss lipid tests (15.3%) and those not taking medications were especially likely to miss foot exams (7.1%), A1c tests (10.6%), and proteinuria tests (10.8%). Sex, education, marital status, family demands, transportation, trust in physicians, and mental health were not associated with lapses in care.
Even in an insured cohort, 3 in 10 participants had 1 or more persistent lapses in diabetes care. Patients with lower income, younger age, having fewer co-morbidities, taking fewer medications and poor health behaviors are particularly vulnerable to persistent lapses in care and a group who warrant targeted interventions to improve preventive diabetes care.
Critics argue that expanding health insurance coverage through Medicaid may not result in improved access to care. The ACA provides reimbursement incentives aimed at improving access to primary care services for new Medicaid beneficiaries; however, there are no such incentives for specialty services. Using the natural experiment of Medicaid expansion in New York State in October 2001, we examined whether Medicaid expansion increased access to common musculoskeletal procedures for Medicaid beneficiaries.
From the State Inpatient Database for NY State, we identified 19–64 year old patients who received; lower extremity large joint replacement, spine procedures and upper/lower extremity fracture/dislocation repair from January 1998–December 2006. We used interrupted time series analysis to evaluate the association between Medicaid expansion and trends in the relative and absolute number of Medicaid beneficiaries who received these musculoskeletal procedures.
Prior to Medicaid expansion, we observed a slight but steady temporal decline in the proportion of musculoskeletal surgical patients who were Medicaid beneficiaries. Following expansion this trend reversed and by 5 years after Medicaid expansion, the proportion of musculoskeletal surgical patients who were Medicaid beneficiaries was 4.7 percentage points (95% CI 3.9, 5.5) higher than expected based on the pre-expansion time trend.
Medicaid expansion in NY State significantly improved access to common musculoskeletal procedures, for Medicaid beneficiaries.
access; Medicaid; specialist; specialty; surgical
Among physicians who perform endoscopic retrograde cholangiopancreatography (ERCP), the relationship between procedure volume and outcome is unknown.
Quantify the ERCP volume-outcome relationship by measuring provider-specific failure rates, hospitalization rates and other quality measures.
16,968 ERCPs performed by 130 physicians between 2001-2011, identified in the Indiana Network for Patient Care (INPC)
Physicians were classified by their average annual INPC volume and stratified into low (<25/year) and high (≥25/year). Outcomes included failed procedures, defined as repeat ERCP, percutaneous transhepatic cholangiography or surgical exploration of the bile duct ≤ 7 days after the index procedure, hospitalization rates, and 30-day mortality.
Among 15,514 index ERCPs, there were 1,163 (7.5%) failures; the failure rate was higher among low (9.5%) compared to high volume (5.7%) providers (p<0.001). A second ERCP within 7 days (a subgroup of failure rate) occurred more frequently when the original ERCP was performed by a low (4.1%) versus a high volume physician (2.3%, p=0.013). Patients were more frequently hospitalized within 24 hours when the ERCP was performed by a low (28.3%) vs. high volume physician (14.8%, p=0.002). Mortality within 30 days was similar (low – 1.9%, high – 1.9%). Among low volume physicians and after adjusting, the odds of having a failed procedure decreased 3.3% (95% CI 1.6-5.0%, p<0.001) with each additional ERCP performed per year.
Lower provider volume is associated with higher failure rate for ERCP, and greater need for post-procedure hospitalization.
ERCP quality; outcomes; gastroenterology
Examination of at least 12 lymph nodes in the staging of colon cancer (CC) was recommended by the National Comprehensive Cancer Network in 2000; however, rates of an adequate examination remain low. This study compares the impact of the hospital contextual variance against that of the operating surgeon on delivery of an adequate lymph node examination.
Retrospective analysis of California Cancer Registry data for all CC operations (2001–2006). Hierarchical models predicted the adequacy of lymph node examination as a function of patient, surgeon, and hospital characteristics. Models were created using penalized quasi-likelihood approximation with 2nd order Taylor linearization as implemented in MLwiN 2.15.
25,606 resections involving 3,376 surgeons operating in 346 hospitals were analyzed. Half of cases had an adequate exam. Hierarchical models showed the median odds of an adequate exam associated with the hospital context ((MORhosp 2.05; 95% CI 1.9–2.2) was much higher than that associated with the surgeon (MORsurg 1.34; 95% CI 1.2–1.4)). Hospital characteristics teaching and high volume predicted higher odds of an adequate examination. There was no association with hospital revenue.
Approximately half of patients undergoing surgery for CC received an adequate lymph node examination. Hospital contextual factors had a stronger association with receipt of an adequate exam than surgeon factors. Our results suggest that quality improvement initiatives and incentives should be targeted at the hospital level to achieve the highest impact. Furthermore, we have identified non-teaching and low volume settings as rational targets for these efforts.
Use of local therapy for prostate cancer may increase because of the perceived advantages of new technologies such as intensity-modulated radiotherapy (IMRT) and robotic prostatectomy.
To examine the association of market-level technological capacity with receipt of local therapy.
Patients with localized prostate cancer who were diagnosed between 2003 and 2007 (n=59,043) from the Surveillance Epidemiology and End Results (SEER) – Medicare database.
We measured the capacity for delivering treatment with new technology as the number of providers offering robotic prostatectomy or IMRT per population in a market (hospital referral region). The association of this measure with receipt of prostatectomy, radiotherapy, or observation was examined with multinomial logistic regression.
For each 1,000 patients diagnosed with prostate cancer, 174 underwent prostatectomy, 490 radiotherapy, and 336 were observed. Markets with high robotic prostatectomy capacity had higher use of prostatectomy (146 vs. 118 per 1,000 men, p=0.008) but a trend towards decreased use of radiotherapy (574 vs. 601 per 1,000 men, p=0.068), resulting in a stable rate of local therapy. High versus low IMRT capacity did not significantly impact use of prostatectomy (129 vs. 129 per 1,000 men, p=0.947) and radiotherapy (594 vs. 585 per 1,000 men, p=0.579).
Although there was a small shift from radiotherapy to prostatectomy in markets with high robotic prostatectomy capacity, increased capacity for both robotic prostatectomy and IMRT did not change the overall rate of local therapy. Our findings temper concerns that new technology spurs additional therapy of prostate cancer.
intensity-modulated radiotherapy; prostate cancer treatment; robotic prostatectomy; technology
We assess whether posttraumatic stress disorder (PTSD) varies in prevalence, diagnostic criteria endorsement, and type and frequency of traumatic events (PTEs) among a nationally representative U.S. sample of 5071 non-Latino whites, 3264 Latinos, 2178 Asians, 4249 African Americans, and 1476 Afro-Caribbeans.
PTSD and other psychiatric disorders were evaluated using the World Mental Health-Composite International Diagnostic Interview (WMH-CIDI) in a national household sample that oversampled ethnic/racial minorities (n=16,238) but was weighted to produce results representative of the general population.
Asians have lower prevalence rates of probable lifetime PTSD while African Americans have higher rates as compared to non-Latino whites, even after adjusting for type and number of exposures to traumatic events, and for sociodemographic, clinical and social support factors. Afro-Caribbeans and Latinos seem to demonstrate similar risk to non-Latino whites, adjusting for these same covariates. Higher rates of probable PTSD exhibited by African Americans and lower rates for Asians, as compared to non-Latino whites, do not appear related to differential symptom endorsement, differences in risk or protective factors or differences in types and frequencies of PTEs across groups.
There appears to be marked differences in conditional risk of probable PTSD across ethnic/racial groups. Questions remain about what explains risk of probable PTSD. Several factors that might account for these differences are discussed as well as the clinical implications of our findings. Uncertainty of the PTSD diagnostic assessment for Latinos and Asians requires further evaluation.
Posttraumatic Stress Disorders across racial and ethnic minority groups; diagnosis
The incidence of lung cancer cases among HIV-infected individuals is increasing with time. It is unclear whether HIV-infected individuals receive the same care for lung cancer as immunocompetent patients due to comorbidities, the potential for interaction between antiretroviral agents and cancer chemotherapy, and concerns regarding complications related to treatment or infection.
The objective of this study was to assess the effect of HIV infection on receipt of guidance-concordant care, and its impact on overall survival among non-small cell lung cancer (NSCLC) Medicare beneficiaries.
Matched case-control design where each HIV patient was matched by age group, gender, race and lung cancer stage at diagnosis with 20 controls randomly selected among those who were not HIV-infected.
Medicare beneficiaries diagnosed with NSCLC between 1998 and 2007, qualified for Medicare based on age, and were 65 years of age or older at the time of the lung cancer diagnosis. HIV infection status was based on Medicare claims data. A total of 174 HIV cases and 3480 controls are included in the analysis.
Odds ratios (OR) for receiving guidance concordant care; hazard ratios (HR) for overall survival
HIV infection was not independently associated with the receipt of guidance concordant care. Among stage I/II patients, median survival times were 26 and 43 months, respectively, for those with and without HIV infection (OR=1.48, P=0.021).
HIV infection was not associated with receipt of guidance concordant care, but reduced survival in early stage patients.
Comorbidity adjustment is an important component of health services research and clinical prognosis. When adjusting for comorbidities in statistical models, researchers can include comorbidities individually or through the use of summary measures such as the Charlson Comorbidity Index or Elixhauser score. We examined the conditions under which individual versus summary measures are most appropriate.
We provide an analytic proof of the utility of comorbidity summary measures when used in place of individual comorbidities. We compared the use of the Charlson and Elixhauser scores versus individual comorbidities in prognostic models using a SEER-Medicare data example. We examined the ability of summary comorbidity measures to adjust for confounding using simulations.
We devised a mathematical proof that found that the comorbidity summary measures are appropriate prognostic or adjustment mechanisms in survival analyses. Once one knows the comorbidity score, no other information about the comorbidity variables used to create the score is generally needed. Our data example and simulations largely confirmed this finding.
Summary comorbidity measures, such as the Charlson Comorbidity Index and Elixhauser scores, are commonly used for clinical prognosis and comorbidity adjustment. We have provided a theoretical justification that validates the use of such scores under many conditions. Our simulations generally confirm the utility of the summary comorbidity measures as substitutes for use of the individual comorbidity variables in health services research. One caveat is that a summary measure may only be as good as the variables used to create it.
Screening for fecal occult blood can be effective in reducing colorectal cancer mortality only if positive tests are appropriately followed up with complete diagnostic evaluation (i.e., colonoscopy or flexible sigmoidoscopy with double contrast barium enema) and treatment.
To examine whether rates of complete diagnostic evaluation following a positive fecal occult blood test (FOBT) have improved over time after the implementation of tracking systems and physician guidelines within a large integrated health care organization.
From 1993 to 2005, 8513 positive FOBTs were identified on 8291 enrollees aged 50–79 of a large health care system. Automated records were used to identify repeat FOBTs, colonoscopy, flexible sigmoidoscopy, and double-contrast barium enema within one year after the positive FOBT. National rates of complete diagnostic evaluation were estimated from the 2005 National Health Interview Survey.
In this integrated health care organization, the percentage of positive FOBTs followed by complete diagnostic evaluation within one year increased from 57%–64% in 1993–1996 to 82%–86% from 2000–2005. Use of repeat FOBT following a positive FOBT decreased from 28–31% in 1993–1996 to 6–11% in 2000–2005. Based on the National Health Interview Survey, only 52% of positive FOBTs from 2000–2005 were followed by complete diagnostic evaluation nationally.
Adherence to recommendations for complete diagnostic evaluation following a positive FOBT has greatly improved over time in an integrated group medical practice. Through the use of tracking systems and screening guidelines, it may be possible to reach levels of follow-up that are comparable to those observed in randomized trials.
Fecal occult blood test; colorectal cancer screening; diagnostic follow-up
Supplemental Digital Content is available in the text.
Risk prediction models have been developed to identify those at increased risk for emergency admissions, which could facilitate targeted interventions in primary care to prevent these events.
Systematic review of validated risk prediction models for predicting emergency hospital admissions in community-dwelling adults.
A systematic literature review and narrative analysis was conducted. Inclusion criteria were as follows; Population: community-dwelling adults (aged 18 years and above); Risk: risk prediction models, not contingent on an index hospital admission, with a derivation and ≥1 validation cohort; Primary outcome: emergency hospital admission (defined as unplanned overnight stay in hospital); Study design: retrospective or prospective cohort studies.
Of 18,983 records reviewed, 27 unique risk prediction models met the inclusion criteria. Eleven were developed in the United States, 11 in the United Kingdom, 3 in Italy, 1 in Spain, and 1 in Canada. Nine models were derived using self-report data, and the remainder (n=18) used routine administrative or clinical record data. Total study sample sizes ranged from 96 to 4.7 million participants. Predictor variables most frequently included in models were: (1) named medical diagnoses (n=23); (2) age (n=23); (3) prior emergency admission (n=22); and (4) sex (n=18). Eleven models included nonmedical factors, such as functional status and social supports. Regarding predictive accuracy, models developed using administrative or clinical record data tended to perform better than those developed using self-report data (c statistics 0.63–0.83 vs. 0.61–0.74, respectively). Six models reported c statistics of >0.8, indicating good performance. All 6 included variables for prior health care utilization, multimorbidity or polypharmacy, and named medical diagnoses or prescribed medications. Three predicted admissions regarded as being ambulatory care sensitive.
This study suggests that risk models developed using administrative or clinical record data tend to perform better. In applying a risk prediction model to a new population, careful consideration needs to be given to the purpose of its use and local factors.
risk prediction model; emergency hospital admission; community-dwelling adults
Medication adherence may be a proxy for healthy behaviors and other factors that affect outcomes. Prior studies of the association between placebo adherence and health outcomes have been limited primarily to men enrolled in clinical trials and cardiovascular disease outcomes. We examined associations between adherence to placebo and the risk of fracture, coronary heart disease, cancer, and all-cause mortality in the two Women’s Health Initiative (WHI) hormone therapy randomized trials.
Postmenopausal women randomized to placebo with adherence measured at least once were eligible for analysis. Time-varying adherence was assessed by dispensing history and pill counts. Outcome adjudication was based on physician review of medical records. Cox proportional hazards models evaluated the relation between high adherence (≥80%) to placebo and various outcomes, referent to low adherence (<80%).
A total of 13,444 postmenopausal women were under observation for 106,066 person-years. High placebo adherence was inversely associated with most outcomes including hip fracture (HR 0.50, 95% CI 0.33–0.78), myocardial infarction (HR 0.69, 95% CI 0.50–0.95), cancer death (HR 0.60, 95% CI 0.43–0.82) and all cause mortality (HR 0.64, 95% CI 0.51–0.80) after adjustment for potential confounders. Women with low adherence to placebo were 20% more likely to have low adherence to statins and osteoporosis medications.
In the WHI clinical trials, high adherence to placebo was associated with favorable clinical outcomes and mortality. Until the healthy behaviors and/or other factors for which high adherence is a proxy can be better elucidated, caution is warranted when interpreting the magnitude of benefit of medication adherence.
fracture; adherence; placebo; compliance; mortality; myocardial infarction; malignancy; osteoporosis
Emergency department (ED) crowding is linked with poor quality of care and worse outcomes, including higher mortality. With the growing emphasis on hospital performance measures, there is additional concern whether inadequate care during crowded periods increases a patient’s likelihood of subsequent inpatient admission. We sought to determine if ED crowding during the index visit was associated with these “bounceback” admissions.
We used comprehensive, non-public, statewide ED and inpatient discharge data from the California Office of Statewide Health Planning and Development from 2007 to identify index outpatient ED visits and bounceback admissions within seven days. We further used ambulance diversion data collected from California local emergency medical services agencies to identify crowded days using intra-hospital daily diversion hour quartiles. Using a hierarchical logistic regression model, we then determined if patients visiting on crowded days were more likely to have a subsequent bounceback admission.
We analyzed 3,368,527 index visits across 202 hospitals, of which 596,471 (17.7%) observations were on crowded days. We found no association between ED crowding and bounceback admissions. This lack of relationship persisted in both a discrete (high/low) model (OR 1.01, 95% CI 0.99, 1.02) and a secondary model using ambulance diversion hours as a continuous predictor (OR 1.00, 95% CI 1.00, 1.00).
Crowding as measured by ambulance diversion does not have an association with hospitalization within 7 days of an ED visit discharge. Therefore, bounceback admission may be a poor measure of delayed or worsened quality of care due to crowding.
A systematic scan of the disparities intervention literature will allow researchers, providers, and policymakers to understand which interventions are being evaluated to improve minority health and which areas require further research.
We systematically categorized 391 disparities intervention articles published between 1979 and 2011, covering 11 diseases. We developed a taxonomy of disparities interventions using qualitative theme analysis. We identified the tactic, or what was done to intervene; the strategy, or a group of tactics with common characteristics; and the level, or who was targeted by the effort.
The taxonomy included 44 tactics, nine strategies, and six levels. Delivering education and training was the most common strategy (37%). Within education and training, the most common tactics were education about disease (14%) and self-management (11%), while communication-skills training (3%) and decision-making aids (1%) were less frequent. The strategy of actively engaging the community through tactics such as community health workers and outreach efforts accounted for 6.5% of tactics. Interventions most commonly targeted patients (50%) and community members who were not established patients of the intervening organization (32%). Interventions targeting providers (7%), the microsystem (immediate care team) (9%), organizations (3%), and policies (0.1%) were less common.
Disparities researchers have predominantly focused on the patient as the target for change; future research should also investigate how to improve the system that serves minority patients. Areas for further study include interventions that engage the community, educational interventions that address communication barriers, and the impact of policy reform on disparities in care.
disparities; quality of care; race; intervention; equity
There is much interest in understanding how using bundled primary care payments to support a patient-centered medial home (PCMH) affects total medical costs.
Research Design and Subjects
We compare 2008–2010 claims and eligibility records on about 10,000 patients in practices transforming to a PCMH and receiving risk-adjusted base payments and bonuses, with similar data on approximately 200,000 patients of non-transformed practices remaining under fee-for-service reimbursement.
We estimate the treatment effect using difference-in-differences, controlling for trend, payer type, plan type, and fixed effects. We weight to account for partial-year eligibility, use propensity weights to address differences in exogenous variables between control and treatment patients, and use the Massachusetts Health Quality Project (MHQP) algorithm to assign patients to practices.
Estimated treatment effects are sensitive to: control variables, propensity weighting, the algorithm used to assign patients to practices, how we address differences in health risk, and whether/how we use data from enrollees who join, leave or change practices. Unadjusted PCMH spending reductions are 1.5% in year 1 and 1.8% in year 2. With fixed patient assignment and other adjustments, medical spending in the treatment group appears to be 5.8% (p=0.20) lower in Year 1 and 8.7% (p=0.14) lower in Year 2 than for propensity-weighted, continuously-enrolled controls; the largest proportional two-year reduction in spending occurs in laboratory test use (16.5%, p=0.02).
Although estimates are imprecise due to limited data and quasi-experimental design, risk-adjusted bundled payment for primary care may have dampened spending growth in three practices implementing a PCMH.
Patient-centered medical home; payment systems; primary care; risk adjustment; Medicare; Medicaid
In non-experimental comparative effectiveness research using healthcare databases, outcome measurements must be validated to evaluate and potentially adjust for misclassification bias. We aimed to validate claims-based myocardial infarction algorithms in a Medicaid population using an HIV clinical cohort as the gold standard.
Medicaid administrative data were obtained for the years 2002–2008 and linked to the UNC CFAR HIV Clinical Cohort based on social security number, first name and last name and myocardial infarction were adjudicated. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated.
There were 1,063 individuals included. Over a median observed time of 2.5 years, 17 had a myocardial infarction. Specificity ranged from 0.979–0.993 with the highest specificity obtained using criteria with the ICD-9 code in the primary and secondary position and a length of stay ≥ 3 days. Sensitivity of myocardial infarction ascertainment varied from 0.588–0.824 depending on algorithm. Conclusion: Specificities of varying claims-based myocardial infarction ascertainment criteria are high but small changes impact positive predictive value in a cohort with low incidence. Sensitivities vary based on ascertainment criteria. Type of algorithm used should be prioritized based on study question and maximization of specific validation parameters that will minimize bias while also considering precision.
Multiple clinical practice guidelines exist for breast and cervical cancer screening, and differ in aggressiveness with respect to the recommended frequency and target populations for screening.
To determine (1) US primary care physicians’ (PCPs) perceptions of the influence of different clinical practice guidelines; (2) the relationship between the number, aggressiveness, and agreement of influential guidelines and the aggressiveness of physicians’ screening recommendations; and (3) factors associated with guideline perceptions.
Research Design and Methods
A nationally representative sample of 1212 PCPs was surveyed in 2006–2007. Cross-sectional analyses examined physicians’ perceptions of the influence of different breast and cervical cancer screening guidelines, the relationship of guideline perceptions to screening recommendations in response to hypothetical vignettes, and the predictors of guideline perceptions.
American Cancer Society and American College of Obstetricians and Gynecologists guidelines were perceived as more influential than other guidelines. Most physicians (62%) valued multiple guidelines, and conflicting and aggressive rather than conservative guideline combinations. The number, aggressiveness, and agreement of influential guidelines were associated with the aggressiveness of screening recommendations (P < 0.01)—which was highest for physicians valuing multiple-aggressive, lowest for physicians valuing multiple-conservative, and intermediate for physicians valuing multiple-conflicting, single, and no guidelines. Obstetrician/gynecologists specialty predicted valuation of aggressive guidelines (P < 0.001).
PCPs’ perceptions of cancer screening guidelines vary, relate to screening recommendations in logically-consistent ways, and are predicted by specialty and other factors. The number, aggressiveness, and agreement of valued guidelines are associated with screening recommendations, suggesting that guideline multiplicity is an important problem in clinical decision-making.
clinical practice guidelines; perceptions; cancer screening; primary care physicians
Most data regarding medical care for cancer patients in the United States comes from SEER-linked Medicare analyses of individuals aged 65 years or older and typically excludes Medicare Advantage enrollees.
To assess the accuracy of chemo- and hormone therapy treatment data available via the Cancer Research Network’s Virtual Data Warehouse (VDW).
Retrospective, longitudinal cohort study. Medical record-abstracted, tumor registry-indicated treatments (“gold standard”) were compared to VDW-indicated treatments derived from HMO pharmacy, electronic medical record, and claim-based data systems.
Enrollees aged 18 years and older diagnosed with incident breast, colorectal, lung or prostate cancer from 2000 through 2008.
Sensitivity, specificity, and positive predictive value (PPV) were computed at six-and 12-months post-cancer diagnosis.
Approximately 45% of all cancer cases (total N=23,800) were aged 64 years or younger. Overall chemotherapy sensitivity/specificities across the three health plans for incident breast, colorectal, lung, and prostate cancer cases were 95%/90%, 95%/93%, 93%/93%, and 85%/77%, respectively. With the exception of prostate cancer cases, overall PPV ranged from 86% to 89%. Small variations in chemotherapy data accuracy existed by cancer site and data source, while greater variation existed in hormone therapy capture across sites.
Strong concordance exists between gold standard tumor registry measures of chemotherapy receipt and CRN VDW data. HMO VDW data can be used for a variety of studies addressing patterns of cancer care and comparative effectiveness research that previously could only be conducted among elderly SEER-Medicare populations.
cancer; chemotherapy; tilization
In 2006, dual-eligible nursing home residents were randomly assigned to a Medicare Part D prescription drug plan (PDP). Subsequently, residents not enrolled in qualified plans at the start of the next year were re-randomized. PDPs vary in generosity through differences in medication coverage and utilization management. Therefore, residents’ assigned plans may be relatively more or less generous for their particular drugs. The impact of generosity on residents’ medication use and health outcomes is unknown.
Using 2005–2008 data, we estimate logistic regression models of the impact of coverage and utilization management on the risk for medication changes and gaps in use, hospitalizations, and death among elderly nursing home residents using one of six drug classes, adjusting for patient characteristics.
Few current medication users faced non-coverage of their drug (0.4%–8.7%) or prior authorization or step therapy requirements if the drug was covered (1.1%–37.4%). After adjusting for individual-level covariates, residents with non-covered drugs were more likely than residents with covered drugs to change medications in most classes studied (e.g., for 2006 angiotensin receptor blocker users, the adjusted average probability of medication change was 0.35 when uncovered vs. 0.11 when covered). Those subject to prior authorization or step therapy were more likely to change in a subset of classes. There were no statistically-significant differences in rates of hospitalization or death after correcting for multiple comparisons.
The Part D benefit’s special protections for nursing home residents may have ameliorated the health impact of coverage limits on this frail elderly population.
Part D; plan generosity; nursing home; dual eligible