Mulugeta Gebregziabher | Medical University of South Carolina (original) (raw)

Papers by Mulugeta Gebregziabher

American Journal of Nephrology, Dec 2, 2016

justed hazards ratio (aHR) 1.65, 95% CI 1.38-1.97, p < 0.0001). There was a significant interacti... more justed hazards ratio (aHR) 1.65, 95% CI 1.38-1.97, p < 0.0001). There was a significant interaction between appointment and medication non-adherence; those with appointment and medication non-adherence were at very high risk of graft loss (aHR 4.18, 95% CI 3.39-5.15, p < 0.0001), compared to those with only appointment non-adherence (aHR 1.39, 95% CI 0.97-2.01, p = 0.0766) or only medication non-adherence (aHR 2.44, 95% CI 2.11-2.81, p < 0.0001). Conclusion: These results demonstrate that non-adherence to health care appointments is a significant and independent risk factor for graft loss.

Chest, Dec 1, 2013

Background: Lung cancer (LC) is the leading cause of cancer-related death for veterans cared for ... more Background: Lung cancer (LC) is the leading cause of cancer-related death for veterans cared for by the US Veterans Health Administration. The LC burden among veterans is almost double that of the general population. Before implementation of an LC screening program, we set out to assess the role of beliefs and attitudes toward LC screening among veterans. Methods: Veterans presenting to the Ralph H. Johnson VA Medical Center were invited to complete a self-administered survey. The survey comprised questions about demographics, smoking status, health status, and knowledge about LC and willingness to be screened. Responses from veteran ever and never smokers were compared. Results: A total of 209 veterans completed the survey. Smokers were signifi cantly (P , .05) more likely than never smokers to be less educated, have a lower income, and report poorer health. Smokers were more likely than never smokers to have two or more comorbidities, which trended toward signifi cance (P 5 .062). Smokers were more likely to have been told by a physician that they were at high risk for LC and to believe that they were at risk. Nearly all veterans surveyed (92.8%) would have a CT scan for LC screening, and 92.4% would have surgery for a screendetected LC. Conclusions: Veterans are overwhelmingly willing to undergo screening for LC, and it seems that participation will not be a barrier to implementation of an LC screening program. The mortality benefi t of LC screening, however, may not be generalizable to the veteran population because of a higher number of comorbid conditions.

Social Science & Medicine, Aug 1, 2018

A common characteristic of patients seen at the Veterans Health Administration (VHA) is a high nu... more A common characteristic of patients seen at the Veterans Health Administration (VHA) is a high number of concurrent comorbidities (i.e. multimorbidity). This study (i) examines the magnitude and patterns of multimorbidity by race/ethnicity and geography; (ii) compares the level of variation explained by these factors in three multimorbidity measures across three large cohorts. We created three national cohorts for Veterans with chronic kidney disease (CKD:n=2,190,564), traumatic brain injury (TBI:n=167,954) and diabetes-mellitus (DM:n=1,263,906). Multimorbidity was measured by Charlson-Deyo, Elixhauser and Walraven-Elixhauser scores. Multimorbidity differences by race/ethnicity and geography were compared using generalized linear models (GLM). Latent class analysis (LCA) was used to identify groups of conditions that are highly associated with race/ethnic groups. Differences in age (CKD,74.5, TBI,49.7, DM,66.9 years), race (CKD,80.9%, TBI,76.4%, DM,63.8% NHW) and geography (CKD,64.4%, TBI,70%, DM,70.9% urban) were observed among the three cohorts. Accounting for these differences, GLM results showed that risk of multimorbidity in non-Hispanic blacks (NHB) with CKD were 1.16 times higher in urban areas and 1.10 times higher in rural areas compared to non-Hispanic whites (NHW) with CKD. DM and TBI showed similar results with risk for NHB, 1.05 higher in urban areas and 0.97 lower in rural areas for both diseases. Overall, our results show that (i) multimorbidity risk was higher for NHB in urban areas compared to rural areas in all three cohorts; (ii) multimorbidity risk was higher for Hispanics in urban areas compared to rural areas in the DM and CKD cohorts; and (iii) the highest overall multimorbidity risk of any race group or location exists for

Annals of Epidemiology, Feb 1, 2013

The association between glycated hemoglobin (HbA1c), medication use/adherence, and mortality stra... more The association between glycated hemoglobin (HbA1c), medication use/adherence, and mortality stratified by race/ethnicity was examined in a national cohort of veterans with type 2 diabetes. Methods: A total of 892,223 veterans with diabetes in 2002 were followed through 2006. HbA1c category was the main exposure (i.e., HbA1c <7%, HbA1c 7%e8% [reference], HbA1c 8%e9%, and HbA1c >9%). Covariates included age, sex, marital status, rural/urban residence, geographic region, number of comorbidities, and diabetes medication use/adherence (i.e., adherent, medication possession ratio 80%; nonadherent; and nonusers). HbA1c and medication use/adherence varied over time, and Cox regression models accounting for time-varying variables were used. Results: In nonmedication users, HbA1c greater than 9% predicted higher mortality risk relative to HbA1c of 7%e8% in non-Hispanic whites (hazard ratio [HR], 1.55; 95% confidence interval [95% CI], 1.43e1.69), non-Hispanic blacks (NHB) (HR, 1.58; 95% CI, 1.34e1.87), and Hispanics (HR, 2.22; 95% CI, 1.75e2.82). In contrast, in nonadherent medication users, HbA1c less than 7% predicted higher mortality risk in NHB (HR, 1.12; 95% CI, 1.05e1.20), whereas HbA1c greater than 9% only predicted mortality in non-Hispanic whites (HR, 1.11; 95% CI, 1.06e1.16). In adherent medication users, HbA1c less than 7% predicted higher mortality in NHB (HR, 1.18; 95% CI, 1.07e1.31), whereas HbA1c greater than 9.0% predicted higher mortality risk across all race/ethnic groups. Conclusion: We found evidence for racial/ethnic differences in the association between glycemic control and mortality, which varied by medication use/adherence.

Health equity, Oct 1, 2018

Purpose: Traumatic brain injury (TBI) is a signature injury among the U.S. veterans. Hispanic U.S... more Purpose: Traumatic brain injury (TBI) is a signature injury among the U.S. veterans. Hispanic U.S. veterans diagnosed with TBI have been found to have higher risk-adjusted mortality. This study examined the adjusted association of geographic location with all-cause mortality in 114,593 veterans diagnosed with TBI between January 1, 2000 and December 31, 2010, and followed through December 31, 2014. Methods: National Veterans Health Administration (VHA) databases containing administrative data including International Classification of Diseases, 9th Revision (ICD-9) codes, sociodemographic characteristics, and survival were linked. TBI was identified based on ICD-9 codes. Cox proportional hazards regression methods were used to examine the association of time from first TBI ICD-9 code to death with geographic location, after adjustment for TBI severity, race/ethnicity, other sociodemographic characteristics, military factors, and Elixhauser comorbidities. Results: Relative to urban mainland veterans with a median survival of 76.4 months, veterans living in the U.S. territories had a median survival of 69.1 months, whereas rural mainland veterans had a median survival of 77.1 months, and highly rural mainland veterans had a mean survival of 77.6 months. The final model adjusted for race/ethnicity, TBI severity, sociodemographic, military, and comorbidity covariates showed that residing in the U.S. territories was associated with a higher risk of death (hazard ratios = 1.24; 95% confidence interval 1.15-1.34) relative to residing on the U.S. mainland. The race/ethnicity disparity previously found for the U.S. veterans diagnosed with TBI seems to be accounted for by living in the U.S. territories. Conclusion: The study shows that among veterans with TBI, mortality rates were higher in those who reside in the U.S. territories, even after adjustment. Previous documented higher mortality among Hispanic veterans seems to be explained by residing in the U.S. territories. The VA has a mission of ensuring equitable treatment of all veterans, and should investigate targeted policies and interventions to improve the survival of the U.S. territory veterans diagnosed with TBI.

Annals of Pharmacotherapy, Feb 6, 2019

Background: Dual health system use may provide increased access to physicians, medications, and o... more Background: Dual health system use may provide increased access to physicians, medications, and other health care resources but may also increase the complexity and coordination of medication regimens. Thus, it is important to elucidate the impact of dual use on medication adherence. Objective: The objective of this study was to evaluate the impact on medication adherence for veterans with dual health care system use (VA and Medicare) when obtaining prescription antihyperglycemic medications to treat diabetes. Methods: This was a longitudinal cohort study using VA and Medicare data from 2006 to 2010. Medication adherence was estimated by calculating annualized drug class-level proportion of days covered (PDC), where PDC >80% was considered adherent. Generalized linear models were used for estimations, accounting for correlation over time. Results: In total, 254 267 veterans with diabetes were included, with 71 057 (27.9%) defined as pharmacy system dual users. Mean age was 77.5 years, and nearly all had multiple comorbidities (mean count 10.2). During follow-up, 75% of VA-only users were deemed adherent to diabetes prescriptions, compared with 63% of dual users. In adjusted models, dual prescription benefit use from VA/Medicare was associated with 39% lower odds of medication adherence (odds ratio [OR] = 0.61; 95% CI = 0.60-0.61). Medication adherence significantly worsened with each additional diabetes medication (OR = 0.65; 95% CI = 0.64-0.65) and significantly decreased over time (OR = 0.95 per year; 95% CI = 0.95-0.96). Conclusion and Relevance: These data suggest that veterans utilizing VA and Medicare to obtain diabetes prescriptions are significantly less likely to be adherent.

Journal of General Internal Medicine, Aug 16, 2014

BACKGROUND: Multimorbidity presents a significant public health challenge, but regional, rural/ur... more BACKGROUND: Multimorbidity presents a significant public health challenge, but regional, rural/urban, and racial/ethnic differences in patterns of multimorbidity in diabetes are poorly understood. OBJECTIVE: To describe patterns of multimorbidity in medical and mental health by regional, rural/urban, and racial/ethnic variation in patients with type 2 diabetes mellitus. DESIGN: Retrospective cohort study from 2002 through 2006 PARTICIPANTS: A national cohort of 892,223 veterans with diabetes MAIN MEASURES: Multimorbidity was the main outcome defined as: the measure of multimorbidity and two categorical outcomes, with pattern of medical and mental health comorbidities combined and separately. KEY RESULTS: Among patients, 52% had 2+ comorbidities, 33% had a single comorbidity, and 14% had no comorbidity; 13.9% had both medical and mental health comorbidities, 70.3% had medical only, and 1.5% had mental health only. The odds of having 3+ comorbidities were nearly fourfold greater in patients 75 years and older relative to patients younger than 50 years (OR=3.95 [95% CI: 3.84, 4.06]). Compared to non-Hispanic whites, the odds of 3+ comorbidities among non-Hispanic blacks were 1.67 times greater (95% CI: 1.63, 1.71). Hispanics were more likely to have a mental health comorbidity alone (OR=1.20 [95% CI: 1.13, 1.28]) than non-Hispanic whites. For patients living in rural areas, the odds were higher of having 3+ comorbidities (OR=1.21 [95% CI: 1.19, 1.23]) and of having both medical and mental health comorbidities (OR=1.15 [95% CI: 1.13, 1.17]) compared to urban dwellers. CONCLUSIONS: Among individuals with diabetes, traditionally disadvantaged groups, including non-Hispanic blacks and rural patients, appear to bear the greatest burden and risk of multimorbidity. Significantly greater odds with increasing number of comorbidities were seen by race/ethnicity, rural residence, and geographic region.

Annals of Pharmacotherapy, Feb 1, 2011

BACKGROUND: Medication adherence, a critical component of glycemic control for patients with type... more BACKGROUND: Medication adherence, a critical component of glycemic control for patients with type 2 diabetes, differs by race/ethnicity. However, few studies have examined regional and rural/urban differences in medication adherence and whether racial/ethnic differences persist after controlling for these differences. OBJECTIVE: To examine regional, rural/urban, and racial/ethnic differences in medication adherence in a national sample of veterans with type 2 diabetes. METHODS: We performed a cohort study of a national sample of veterans with diabetes (N = 690,968) receiving prescriptions for insulin or oral hypoglycemic agents in 2002. Patients were followed until death, loss to follow-up, or through December 2006. We calculated the annual medication possession ratio (MPR) for each veteran across 4 groups of medication users: individuals using (1) insulin only, (2) oral hypoglycemic agents only, (3) insulin combined with hypoglycemic agents, and (4) insulin or oral hypoglycemic agents (primary analysis). RESULTS: In longitudinal models for the primary analysis, adjusting for relevant covariates and time trends, MPR was significantly lower among non-Hispanic blacks (NHBs), Hispanics, and individuals with other/missing/unknown race/ethnicity (6.07%, 1.76%, and 2.83% lower, respectively) relative to non-Hispanic whites (NHWs). MPR was also 2.0% higher in rural versus urban veterans and 1.28% higher in the mid-Atlantic, 2.04% higher in the Midwest, and 0.76% lower in the West, relative to the South. There was a significant race/ethnicity and urban/rural interaction. In NHWs and NHBs, MPR was 1.91 % and 2.00% higher, respectively, in rural versus urban veterans; in contrast, in Hispanics, MPR was 1.0% lower in rural veterans relative to urban veterans. CONCLUSIONS: In a national longitudinal cohort of veterans with type 2 diabetes, we found significant regional, rural/urban, and racial/ethnic differences in MPR. Rural/urban residence modified the effect of race/ethnicity on MPR. Recognition of these differences can enable clinicians to better allocate resources and target quality improvement programs.

Brain Injury, Nov 7, 2018

Background: The U.S. Veterans Health Administration (VHA) provides depression treatment to vetera... more Background: The U.S. Veterans Health Administration (VHA) provides depression treatment to veterans with Traumatic Brain Injury (TBI). VHA costs of comorbid TBI-depression were estimated by Operation Enduring Freedom/Operation Iraqi Freedom (OEF/OIF) status over 14 years. Methods: VHA-USING veterans with TBI DIAGNOSED IN 2000-2010 were followed through FY2014. TBI severity was determined using the Department of Defense criteria. Depression was identified by the Elixhauser algorithm. Generalized linear and seemingly unrelated regression models were used to estimate the impact of depression on annual per veteran and total VHA inpatient, outpatient, and pharmaceutical costs, by OEF/OIF status.

Annals of Transplantation, Feb 16, 2018

Research reported in this publication was supported by the National Institute of Diabetes and Dig... more Research reported in this publication was supported by the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health under Award Number K23DK099440 Background: Disparities research has traditionally focused on patient-level variables to ascertain predominant risk factors driving differences in outcomes for African-American (AA) kidney transplant recipients. Our objectives were to determine the magnitude and impact of transplant center variability for graft outcome disparities. Material/Methods: This was a retrospective cohort study analyzing 25 years of U.S. national transplant registry data at both the patient and center levels using univariate descriptive statistics and multivariable modeling. Results: A total of 257,024 recipients from 191 centers were analyzed; AAs represented 31.1% of recipients. After adjusting for baseline characteristics, AAs had 42% higher risk of graft loss (aHR 1.42, 95% CI 1.39 to 1.45; p<0.001). Center variability for graft outcome disparities in AAs was significant (race*center interaction term p<0.05), with the aHRs ranging from 0.5 to 4.9; 46% of centers demonstrated a non-statistically significant disparity (aHR p>0.05) and 25% of centers had a large AA disparity (aHR >1.75). In a more recent transplant time period (2000-14), overall racial disparities decreased but center-level disparities increased in variability. Center-level factors significantly associated with increasing disparity included higher acute rejection rates, fewer transplants per year, longer length of stay, lower use of calcineurin inhibitors (CNI), and lower living donor rates. Conclusions: There is evidence of significant center-level variability in graft outcome disparities for AA kidney recipients. Further, there appears to be a number of center-level factors associated with this variability, including acute rejection rates, CNI use, number of transplants per year, and, in recent years, low living donor rates.

Journal of Affective Disorders, 2017

Background: Little evidence exists regarding the costs of telemedicine, especially considering ch... more Background: Little evidence exists regarding the costs of telemedicine, especially considering changes over time. This analysis aimed to analyze trajectory of healthcare cost before, during, and after a behavioral activation intervention delivered via telepsychology and same-room delivery to elderly Veterans with depression. Methods: 241 participants were randomly assigned into one of two study groups: behavioral activation for depression via telemedicine or via same-room treatment. Patients received 8 weeks of weekly 60-min individual sessions of behavioral activation for depression. Primary outcomes were collected at 12-months. Inpatient, outpatient, pharmacy, and total costs were collected from VA Health Economics Resource Center (HERC) datasets for FY 1998-2014 and compared between the two treatment groups. Generalized mixed models were used to investigate the trajectories over time. Results: Overall cost, as well as, outpatient and pharmacy cost show increasing trend over time. Unadjusted and adjusted trajectories over time for any cost were not different between the two treatment groups. There was a significant overall increasing trend over time for outpatient (p < 0.001) and total cost (p < 0.001) but not for inpatient (p=0.543) or pharmacy cost (p=0.084). Limitations: Generalizability to younger, healthier populations may be limited due to inclusion criteria for study participants. Conclusion: Healthcare costs before, during, and after intervention did not differ between the telemedicine and in-person delivery methods. Outpatient costs accounted for most of the increasing trend of cost over time. These results support policies to use both telehealth and in-person treatment modalities to effectively and efficiently provide high quality care.

Transplantation, Jul 1, 2016

Background. Socioeconomic status (SES) is a significant determinant of health outcomes and may be... more Background. Socioeconomic status (SES) is a significant determinant of health outcomes and may be an important component of the causal chain surrounding racial disparities in kidney transplantation. The social adaptability index (SAI) is a validated and quantifiable measure of SES, with a lack of studies analyzing this measure longitudinally or between races. Methods. Longitudinal cohort study in adult kidney transplantation transplanted at a single-center between 2005 and 2012. The SAI score includes 5 domains (employment, education, marital status, substance abuse and income), each with a minimum of 0 and maximum of 3 for an aggregate of 0 to 15 (higher score → better SES). Results. One thousand one hundred seventy-one patients were included; 624 (53%) were African American (AA) and 547 were non-AA. African Americans had significantly lower mean baseline SAI scores (AAs 6.5 vs non-AAs 7.8; P < 0.001). Cox regression analysis demonstrated that there was no association between baseline SAI and acute rejection in non-AAs (hazard ratio [HR], 0.92; 95% confidence interval [95% CI], 0.81-1.05), whereas it was a significant predictor of acute rejection in AAs (HR, 0.89; 95% CI, 0.80-0.99). Similarly, a 2-stage approach to joint modelling of time to graft loss and longitudinal SAI did not predict graft loss in non-AAs (HR, 1.01; 95% CI, 0.28-3.62), whereas it was a significant predictor of graft loss in AAs (HR, 0.23; 95% CI, 0.06-0.93). Conclusions. After controlling for confounders, SAI scores were associated with a lower risk of acute rejection and graft loss in AA kidney transplant recipients, whereas neither baseline nor follow-up SAI predicted outcomes in non-AA kidney transplant recipients.

Diabetes Care, Mar 21, 2011

OBJECTIVE-We performed a retrospective analysis of a national cohort of veterans with diabetes to... more OBJECTIVE-We performed a retrospective analysis of a national cohort of veterans with diabetes to better understand regional, geographic, and racial/ethnic variation in diabetes control as measured by HbA 1c. RESEARCH DESIGN AND METHODS-A retrospective cohort study was conducted in a national cohort of 690,968 veterans with diabetes receiving prescriptions for insulin or oral hypoglycemic agents in 2002 that were followed over a 5-year period. The main outcome measures were HbA 1c levels (as continuous and dichotomized at $8.0%).

Kidney International, Oct 1, 2016

Disparities in outcomes for African American (AA) kidney transplant recipients have persisted for... more Disparities in outcomes for African American (AA) kidney transplant recipients have persisted for 40 years without a comprehensive analysis of evolving trends in the risks associated with this disparity. Here we analyzed U.S. transplant registry data, which included adult Caucasian or AA solitary kidney recipients undergoing transplantation between 1990 and 2009 comprising 202,085 transplantations. During this 20-year period, the estimated rate of 5-year graft loss decreased from 27.6% to 12.8%. Notable trends in baseline characteristics that significantly differed by race over time included the following: increased prevalence of diabetes from 2001 to 2009 in AAs (5-year slope difference: 3.4%), longer time on the waiting list (76.5 more days per 5 years in AAs), fewer living donors in AAs from 2003 to 2009 (5-year slope difference: -3.36%), more circulatory death donors in AAs from 2000-09 (5-year slope difference: 1.78%), and a slower decline in delayed graft function in AAs (5-year slope difference: 0.85%). The absolute risk difference between AAs and Caucasians for 5-year graft loss significantly declined over time (-0.92% decrease per 5 years), whereas the relative risk difference actually significantly increased (3.4% increase per 5 years). These results provide a mixed picture of both promising and concerning trends in disparities for AA kidney transplant recipients. Thus, although the disparity for graft loss has significantly improved, equity is still far off, and other disparities, including living donation rates and delayed graft function rates, have widened during this time.

Global Journal of Health Science, Jul 19, 2015

Objective: Primary: To examine Veterans Administration (VA) utilization and other potential media... more Objective: Primary: To examine Veterans Administration (VA) utilization and other potential mediators between racial/ethnic differentials and mortality in veterans diagnosed with traumatic brain injury (TBI).

JACCP: journal of the American College of Clinical Pharmacy, Jul 3, 2018

Introduction: Health disparities in African-American (AA) kidney transplant recipients compared w... more Introduction: Health disparities in African-American (AA) kidney transplant recipients compared with non-AA recipients are well established. Cardiovascular disease (CVD) risk control is a significant mediator of this disparity. Objective: To assess the efficacy of improved medication safety, CVD risk control, and racial disparities in kidney transplant recipients. Methods: Prospective, pharmacist-led, technology-aided, 6-month interventional clinical trial. A total of 60 kidney recipients with diabetes and hypertension were enrolled. Patients had to be at least one-year post transplant with stable graft function. Primary outcome measured included hypertension, diabetes, and lipid control using intent-to-treat analyses, with differences assessed between AA and non-AA recipients. Results: The participants mean age was 59 years, with 42% being female and 68% being AA. Overall, patients demonstrated improvements in blood pressure <140/90 mmHg (baseline 50% vs. end of study 68%, p=0.054) and hemoglobin A1c <7% (baseline 33% vs. end of study 47%, p=0.061). AAs demonstrated a significant reduction from baseline in systolic blood pressure (−0.86 mmHg per month, p=0.026), which was not evident in non-AAs (−0.13 mmHg per month,

Medicine, Jun 1, 2016

The aim of the study was to examine whether depression impacts medication nonadherence (MNA) over... more The aim of the study was to examine whether depression impacts medication nonadherence (MNA) over time and determine if race has a differential impact on MNA in patients with type 2 diabetes and comorbid depression. Generalized estimating equations were used with a longitudinal national cohort of 740,197 veterans with type 2 diabetes. MNA was the main outcome defined by <80% medication possession ratio for diabetes medications. The primary independent variable was comorbid depression. Analyses were adjusted for the longitudinal nature of the data and covariates including age, sex, marital status, and rural/urban residence. In adjusted models, MNA was higher in non-Hispanic blacks (NHBs) (odds ratio [OR] 1.58 [95% confidence interval-CI: 1.57, 1.59]), Hispanics (OR 1.34 [95% CI: 1.32, 1.35]), and the other/missing racial/ethnic group (OR 1.37 [95% CI: 1.36, 1.38]) than in non-Hispanic whites (NHWs). In stratified analyses, the odds of MNA associated with depression were highest in NHWs (OR 1.14 [95% CI: 1.12, 1.15]) and were significantly associated in the other 3 minority racial/ethnic groups. MNA was lower in rural than urban NHWs (OR 0.91 [95% CI: 0.90, 0.92]), NHBs (OR 0.92 [95% CI: 0.91, 0.94]), and the other/unknown racial/ethnic group (OR 0.89 [95% CI: 0.88, 0.90]), but higher in rural Hispanic patients (OR 1.12 [95% CI: 1.09, 1.14]). Depression was associated with increased odds of MNA in NHWs, as well as in minority groups, although associations were weaker in minority groups, perhaps as a result of the high baseline levels of MNA in minority groups. There were also differences by race/ethnicity in MNA in rural versus urban subjects.

Journal of General Internal Medicine, Mar 23, 2017

BACKGROUND: Diabetes disproportionately affects African Americans and is associated with poorer o... more BACKGROUND: Diabetes disproportionately affects African Americans and is associated with poorer outcomes. Selfmanagement is important for glycemic control; however, evidence in African Americans is limited. OBJECTIVE: To assess the efficacy of a combined telephone-delivered education and behavioral skills intervention (TBSI) in reducing hemoglobin A1c (HbA1c) levels in African Americans with type 2 diabetes, using a factorial design. DESIGN: This is a four-year randomized clinical trial, using a 2 x 2 factorial design.: Participants: African American adults ≥18 years) with poorly controlled type 2 diabetes (HbA1c ≥9%) were randomly assigned to one of four groups: 1) knowledge only, 2) skills only, 3) combined knowledge and skills (TBSI), or 4) control group. Intervention: All participants received 12 telephone-delivered 30min intervention sessions specific to their assigned group. Participants were assessed at baseline and 3, 6, and 12 months. Main measure: The primary outcome was HbA1c at 12 months post-randomization in the intentto-treat (ITT) population. KEY RESULTS: Two hundred fifty-five participants were randomly assigned to the four groups. Based on the ITT population after multiple imputation, the analysis of covariance with baseline HbA1c as the covariate showed that HbA1c at 12 months for the intervention groups did not differ significantly from that of the control group (knowledge: 0.49, p = 0.123; skills: 0.23, p = 0.456; combined: 0.48, p = 0.105). Absolute change from baseline at 12 months for all treatment arms was 0.6. Longitudinal mixed effects analysis showed that, on average, there was a significant decline in HbA1c over time for all treatment groups (−0.07, p < 0.001). However, the rates of decline for the intervention groups were not significantly different from that of the control group (knowledge: 0.06, p = 0.052; skills: 0.02, p = 0.448; combined: 0.05, p = 0.062). Results from per-protocol populations were similar. CONCLUSIONS: For African Americans with poorly controlled type 2 diabetes, combined education and skills training did not achieve greater reductions in glycemic control (i.e., HbA1c levels) at 12 months compared to the control group, education alone, or skills training alone. This trial is registered with ClinicalTrials.gov, identifier no. NCT00929838.

Diabetes Care, Nov 14, 2012

OBJECTIVEdTo examine the longitudinal effects of medication nonadherence (MNA) on key costs and e... more OBJECTIVEdTo examine the longitudinal effects of medication nonadherence (MNA) on key costs and estimate potential savings from increased adherence using a novel methodology that accounts for shared correlation among cost categories. RESEARCH DESIGN AND METHODSdVeterans with type 2 diabetes (740,195) were followed from January 2002 until death, loss to follow-up, or December 2006. A novel multivariate, generalized, linear, mixed modeling approach was used to assess the differential effect of MNA, defined as medication possession ratio (MPR) $0.8 on healthcare costs. A sensitivity analysis was performed to assess potential cost savings at different MNA levels using the Consumer Price Index to adjust estimates to 2012 dollar value. RESULTSdMean MPR for the full sample over 5 years was 0.78, with a mean of 0.93 for the adherent group and 0.58 for the MNA group. In fully adjusted models, all annual cost categories increased ;3% per year (P = 0.001) during the 5-year study time period. MNA was associated with a 37% lower pharmacy cost, 7% lower outpatient cost, and 41% higher inpatient cost. Based on sensitivity analyses, improving adherence in the MNA group would result in annual estimated cost savings ranging from

arXiv (Cornell University), Nov 18, 2015

Positive continuous outcomes with a point mass at zero are prevalent in biomedical research. To m... more Positive continuous outcomes with a point mass at zero are prevalent in biomedical research. To model the point mass at zero and to provide marginalized covariate effect estimates, marginalized two part models (MTP) have been developed for outcomes with lognormal and log skew normal distributions. In this paper, we propose MTP models for outcomes from a generalized gamma (GG) family of distributions. In the proposed MTP-GG model, the conditional mean from a two-part model with a three-parameter GG distribution is parameterized to provide regression coefficients that have marginal interpretation. MTP-gamma and MTP-Weibull are developed as special cases of MTP-GG. We derive marginal covariate effect estimators from each model and through simulations assess their finite sample operating characteristics in terms of bias, standard errors, 95% coverage, and rate of convergence. We illustrate the models using data sets from The Medical Expenditure Survey (MEPS) and from a randomized trial of addictive disorders and we provide SAS code for implementation. The simulation results show that when the response distribution is unknown or mis-specified, which is usually the case in real data sets, the MTP-GG is preferable over other models.