Intent-to-treat analysis of cluster randomized trials when clusters report unidentifiable outcome proportions.

BACKGROUND: Cluster randomized trials are designed to evaluate interventions at the cluster or group level. When clusters are randomized but some clusters report no or non-analyzable data, intent-to-treat analysis, the gold standard for the analysis of randomized controlled trials, can be compromised. This article presents a very flexible statistical methodology for cluster randomized trials whose outcome is a cluster-level proportion (e.g. proportion from a cluster reporting an event) in the setting where clusters report non-analyzable data (which in general could be due to nonadherence, dropout, missingness, etc.). The approach is motivated by a previously published stratified randomized controlled trial called, "The Randomized Recruitment Intervention Trial (RECRUIT)," designed to examine the effectiveness of a trust-based continuous quality improvement intervention on increasing minority recruitment into clinical trials (ClinicalTrials.gov Identifier: NCT01911208). METHODS: The novel approach exploits the use of generalized estimating equations for cluster-level reports, such that all clusters randomized at baseline are able to be analyzed, and intervention effects are presented as risk ratios. Simulation studies under different outcome missingness scenarios and a variety of intra-cluster correlations are conducted. A comparative analysis of the method with imputation and per protocol approaches for RECRUIT is presented. RESULTS: Simulation results show the novel approach produces unbiased and efficient estimates of the intervention effect that maintain the nominal type I error rate. Application to RECRUIT shows similar effect sizes when compared to the imputation and per protocol approach. CONCLUSION: The article demonstrates that an innovative bivariate generalized estimating equations framework allows one to implement an intent-to-treat analysis to obtain risk ratios or odds ratios, for a variety of cluster randomized designs.

Design of a cluster-randomized minority recruitment trial: RECRUIT.

BACKGROUND: Racial/ethnic minority groups remain underrepresented in clinical trials. Many strategies to increase minority recruitment focus on minority communities and emphasize common diseases such as hypertension. Scant literature focuses on minority recruitment to trials of less common conditions, often conducted in specialty clinics and dependent on physician referrals. We identified trust/mistrust of specialist physician investigators and institutions conducting medical research and consequent participant reluctance to participate in clinical trials as key-shared barriers across racial/ethnic groups. We developed a trust-based continuous quality improvement intervention to build trust between specialist physician investigators and community minority-serving physicians and ultimately potential trial participants. To avoid the inherent biases of non-randomized studies, we evaluated the intervention in the national Randomized Recruitment Intervention Trial (RECRUIT). This report presents the design of RECRUIT. Specialty clinic follow-up continues through April 2017. METHODS: We hypothesized that specialist physician investigators and coordinators trained in the trust-based continuous quality improvement intervention would enroll a greater proportion of minority participants in their specialty clinics than specialist physician investigators in control specialty clinics. Specialty clinic was the unit of randomization. Using continuous quality improvement, the specialist physician investigators and coordinators tailored recruitment approaches to their specialty clinic characteristics and populations. Primary analyses were adjusted for clustering by specialty clinic within parent trial and matching covariates. RESULTS: RECRUIT was implemented in four multi-site clinical trials (parent trials) supported by three National Institutes of Health institutes and included 50 associated specialty clinics from these parent trials. Using current data, we have 88% power or greater to detect a 0.15 or greater difference from the currently observed control proportion adjusting for clustering. We detected no differences in baseline matching criteria between intervention and control specialty clinics (all p values > 0.17). CONCLUSION: RECRUIT was the first multi-site randomized control trial to examine the effectiveness of a trust-based continuous quality improvement intervention to increase minority recruitment into clinical trials. RECRUIT's innovations included its focus on building trust between specialist investigators and minority-serving physicians, the use of continuous quality improvement to tailor the intervention to each specialty clinic's specific racial/ethnic populations and barriers to minority recruitment, and the use of specialty clinics from more than one parent multi-site trial to increase generalizability. The effectiveness of the RECRUIT intervention will be determined after the completion of trial data collection and planned analyses.

The Impact of the Share 35 Policy on Racial and Ethnic Disparities in Access to Liver Transplantation for Patients with End Stage Liver Disease in the United States: An Analysis from UNOS Database.

BACKGROUND: The Share 35 policy was instituted in June 2013 by the United Network for Organ Sharing (UNOS) in order to reduce death on liver transplant waiting list. The effect of this policy on racial and ethnic disparities in access to liver transplantation has not been examined. METHODS: A total of 14,585 adult patients registered for liver transplantation between 2012 and 2015 were identified from UNOS database. Logistic and proportional hazards models were used to model the effects of race and ethnicity on access to liver transplantation. Stratification on pre- and post-Share 35 periods was performed to compare the first 18 months of Share 35 policy to an equivalent time period before. RESULTS: Comparison of the pre- and post-Share 35 periods showed significantly decreased time on waiting list and increased numbers of minorities having access to liver transplantation. Hispanic recipients still experienced significantly longer waiting time (HR: 0.69, 95% CI: 0.53-0.88) before they received liver transplantation after Share 35 policy took effect. CONCLUSION: The Share 35 policy did not lead to improved access to liver transplantation among minorities but eliminated the previously observed racial and ethnic disparities in transplant rates as well as shortened the waiting time.

Racial/Ethnic Disparity in Utilization of General Dental Care Services Among US Adults: Medical Expenditure Panel Survey 2012.

OBJECTIVE: The objective of this study was to examine racial/ethnic disparity and associated factors in general dental care utilization among US adults. METHODS: Data were adults 18-85 years old (N = 27,394) from the 2012 Medical Expenditure Panel Survey (MEPS). The outcome was the receipt of general dental care, measured by (1) whether the individual has ever had a general dental visit and (2) the number of general dental visits the individual has had during the past year. Race/ethnicity was the primary interest. Logistic regressions and negative binomial regressions were conducted using STATA version 12 to assess the effect of race/ethnicity on the receipt of general dental care both independently and adjusted for other demographic and socioeconomic factors. RESULTS: Non-Hispanic blacks (odds ratio (OR) 0.39, 95 % confidence intervals (CI) 0.37-0.43), Hispanics (0.34, 0.43-0.37), and other minorities (0.61, 0.56-0.68) were less likely to report general dental visits both independently and adjusted for other demographic and socioeconomic factors, compared with non-Hispanic whites. General dental visits were more likely to be observed among individuals who were female, married, native speakers, living in a metropolitan statistical area, and with dental insurance, and the number of visits was increasing with age, educational level, and family income. CONCLUSION: The race/ethnicity disparity in the utilization of general dental care still existed. Policy makers and dental care providers should promote dental insurance coverage and language support programs, and increase the diversity of dental professionals among minorities to encourage their visits to dentists.

Associations between hematopoietic growth factors and risks of venous thromboembolism, stroke, ischemic heart disease and myelodysplastic syndrome: findings from a large population-based cohort of women with breast cancer.

PURPOSE: To determine the risk of venous thromboembolism (VTE), stroke, ischemic heart disease, and myelodysplastic syndrome (MDS) in association with the receipt of colony-stimulating factors (CSFs) and/or erythropoiesis-stimulating agents (ESAs) in women with breast cancer. METHODS: We studied 77,233 women with breast cancer aged ≥65 in 1992-2009 from the Surveillance, Epidemiology, and End Results-Medicare linked data with up to 19 years of follow-up. RESULTS: Incidence of VTE increased from 9 cases in women receiving no chemotherapy and no CSFs/ESAs to 22.79 cases per 1,000 person-years in those receiving chemotherapy with CSFs and ESAs. Women with chemotherapy who received both CSFs and ESAs (adjusted hazard ratio and 95 % confidence interval 2.01, 1.80-2.25) or received ESAs without CSFs (2.03, 1.74-2.36) were twice as likely to develop VTE than those receiving no chemotherapy and no CSFs/ESAs, whereas those receiving CSF alone without ESA were 64 % more likely to have VTE (1.64, 1.45-1.85). Risk of MDS was significantly increased by fivefold in patients receiving ESA following chemotherapy. CONCLUSIONS: Receipts of CSFs and ESAs were significantly associated with an increased risk of VTE in women with breast cancer. Use of ESAs was significantly associated with substantially increased risks of MDS. These findings support those of previous studies.

Effects of Health Insurance on Tumor Stage, Treatment, and Survival in Large Cohorts of Patients with Breast and Colorectal Cancer.

To examine the impact of health insurance status on tumor stage at diagnosis, treatment rendered, and overall survival, we identified 52,566 breast cancer patients and 34,316 colorectal cancer patients aged 20 or older in 2007-2010 from Texas Cancer Registry. Those aged younger than 65 years without health insurance coverage had significantly higher risks of mortality than those with private health insurance regardless of tumor stage, chemotherapy, or surgery for colorectal cancer. However, in patients younger than 65 years with breast cancer, the risk of mortality was not significantly higher for those who received chemotherapy or cancer-directed surgery in patients without insurance coverage compared with those with private health insurance. In Medicare beneficiaries aged 65 years or older, risk of mortality was not significantly different between those with Medicare only and those with additional private health insurance, except an increased mortality in patients without chemotherapy for breast and colorectal cancer and in those without receiving surgery for colorectal cancer.

Risks of Venous Thromboembolism, Stroke, Heart Disease, and Myelodysplastic Syndrome Associated With Hematopoietic Growth Factors in a Large Population-Based Cohort of Patients With Colorectal Cancer.

PURPOSE: To determine the relationship between the receipt of colony-stimulating factors (CSFs) with erythropoiesis-stimulating agents (ESAs) and the risk of developing venous thromboembolism (VTE), stroke, heart disease, and myelodysplastic syndrome (MDS) in patients with colorectal cancer. METHODS: We studied 80,925 patients diagnosed with colorectal cancer at age ≥ 65 years in 1992-2009 from the nationwide 16 areas of the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked data. Cumulative incidence and the time to events Cox hazard regressions were used to explore the risks of outcomes in association with the receipt of CSFs and ESAs. RESULTS: Patients who received chemotherapy (CT) with both CSF and ESA were 58% more likely to develop VTE than those who received CT without CSF and ESA (hazard ratio, 1.58; 95% confidence interval, 1.43-1.76). The risk of stroke appeared to be not associated with the use of CSF and ESA, whereas the risk of heart disease was only significantly elevated in those patients who did not receive CT but received ESA. The risk of acute myeloid leukemia or MDS was significantly increased 4- to 9-fold in patients who received ESA, regardless of receipt of CT or CSF. CONCLUSION: The use of ESAs was significantly associated with a substantially increased risk of MDS in patients with colorectal cancer. The use of CSFs and ESAs was also significantly associated with a moderately increased risk of VTE and a slightly elevated risk of heart disease.

Increasing utilization and predictors of hematopoietic growth factors in patients diagnosed with colorectal cancer: findings from a large national population-based cohort in the USA, 1992-2009.

None of previous studies has provided the detailed pattern, variation, and temporal trend in the use of growth factors for patients with colorectal cancer. The aim of the study was to examine the temporal trend and predictors of receiving hematopoietic growth factors in a large nationwide and population-based cohort of patients with colorectal cancer in the USA from 1992 to 2009. We studied 50,768 patients diagnosed with colorectal cancer at age 65-89 years in 1992-2009 in the Surveillance, Epidemiology and End Results areas who received chemotherapy as part of initial therapy within 12 months of diagnosis according to Medicare data. Growth factors were identified for colony-stimulating factors (CSFs) and for erythropoiesis-stimulating agents (ESAs). Overall, 16.3% received CSFs and 26.5% received ESAs with an increase from 0.8 and 1.5% in 1992 to 29.4 and 14.1% in 2009, respectively. Compared with patients diagnosed in 1992-1994, those diagnosed in 1995-1997 were >2 times more likely to receive CSFs and ESAs, whereas patients diagnosed recently in 2007-2009 were >22 times and 4 times to receive CSFs and ESAs, respectively. Gender, marital status, comorbidity scores, geographic area, year of diagnosis, tumor stage, number of lymph nodes, and risk profile for febrile neutropenia were statistically significant predictors of using CSFs and ESAs. There were substantial temporal and geographic variations in the use of hematopoietic growth factors in patients with colorectal cancer following chemotherapy. More studies would be needed to explore the effectiveness of hematopoietic growth factors in preventing and treating neutropenia, anemia, and infection.