Walid F Gellad

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Are you Walid F Gellad?

Claim your profile

Publications (62)446.28 Total impact

  • Walid F Gellad
    Journal of General Internal Medicine 08/2015; DOI:10.1007/s11606-015-3492-2 · 3.42 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Quality improvement efforts are frequently tied to patients achieving ≥80% medication adherence. However, there is little empirical evidence that this threshold optimally predicts important health outcomes. To apply machine learning to examine how adherence to oral hypoglycemic medications is associated with avoidance of hospitalizations, and to identify adherence thresholds for optimal discrimination of hospitalization risk. A retrospective cohort study of 33,130 non-dual-eligible Medicaid enrollees with type 2 diabetes. We randomly selected 90% of the cohort (training sample) to develop the prediction algorithm and used the remaining (testing sample) for validation. We applied random survival forests to identify predictors for hospitalization and fit survival trees to empirically derive adherence thresholds that best discriminate hospitalization risk, using the proportion of days covered (PDC). Time to first all-cause and diabetes-related hospitalization. The training and testing samples had similar characteristics (mean age, 48 y; 67% female; mean PDC=0.65). We identified 8 important predictors of all-cause hospitalizations (rank in order): prior hospitalizations/emergency department visit, number of prescriptions, diabetes complications, insulin use, PDC, number of prescribers, Elixhauser index, and eligibility category. The adherence thresholds most discriminating for risk of all-cause hospitalization varied from 46% to 94% according to patient health and medication complexity. PDC was not predictive of hospitalizations in the healthiest or most complex patient subgroups. Adherence thresholds most discriminating of hospitalization risk were not uniformly 80%. Machine-learning approaches may be valuable to identify appropriate patient-specific adherence thresholds for measuring quality of care and targeting nonadherent patients for intervention.
    Medical care 06/2015; 53(8). DOI:10.1097/MLR.0000000000000394 · 2.94 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Little is known about the health status of the 7.3 million Americans who enrolled in insurance plans through the Marketplaces established by the Affordable Care Act in 2014. Medication use may provide an early indicator of the health needs and access to care among Marketplace enrollees. We used data from January-September 2014 on more than one million Marketplace enrollees from Express Scripts, the largest pharmacy benefit management company in the United States. We compared the characteristics and medication use between early and late Marketplace enrollees and between all Marketplace enrollees and enrollees with employer-sponsored insurance. Among Marketplace enrollees, we found that those who enrolled earlier (October 2013-February 2014) were older and used more medication than later enrollees. Marketplace enrollees, as a whole, had lower average drug spending and were less likely to use most medication classes than the employer-sponsored comparison group. However, Marketplace enrollees were more likely to use medicines for hepatitis C and particularly for HIV. Project HOPE—The People-to-People Health Foundation, Inc.
    Health Affairs 05/2015; 34(6):1-8. DOI:10.1377/hlthaff.2015.0016 · 4.32 Impact Factor
  • Value in Health 05/2015; 18(3):A127-A128. DOI:10.1016/j.jval.2015.03.743 · 2.89 Impact Factor
  • Y. Tang · C.H. Chang · H. Huskamp · W.F. Gellad · J.M. Donohue
    Value in Health 05/2015; 18(3):A128. DOI:10.1016/j.jval.2015.03.747 · 2.89 Impact Factor
  • Value in Health 05/2015; 18(3):A129. DOI:10.1016/j.jval.2015.03.751 · 2.89 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background: Health insurance claims data may play an important role for healthcare systems and payers in monitoring the non-medical use of prescription opioids (NMPO) among patients. However, these systems require valid methods for identifying NMPO if they are to target individuals for intervention. Limited efforts have been made to define NMPO using administrative data available to health systems and payers. We conducted a systematic review of publications that defined and measured NMPO within health insurance claims databases in order to describe definitions of NMPO and identify areas for improvement. Methods: We searched eight electronic databases for articles that included terms related to NMPO and health insurance claims. A total of 2,613 articles were identified in our search. Titles, abstracts, and article full texts were assessed according to predetermined inclusion/exclusion criteria. Following article selection, we extracted general information, conceptual and operational definitions of NMPO, methods used to validate operational definitions of NMPO, and rates of NMPO. Results: A total of seven studies met all inclusion criteria. A range of conceptual NMPO definitions emerged, from concrete concepts of abuse to qualified definitions of probable misuse. Operational definitions also varied, ranging from variables that rely on diagnostic codes to those that rely on opioid dosage and/or filling patterns. Quantitative validation of NMPO definitions was reported in three studies (e.g., receiver operating curves or logistic regression), with each study indicating adequate validity. Three studies reported qualitative validation, using face and content validity. One study reported no validation efforts. Rates of NMPO among the studies' populations ranged from 0.75-10.32%. Conclusions: Disparate definitions of NMPO emerged from the literature, with little uniformity in conceptualization and operationalization. Validation approaches were also limited, and rates of NMPO varied across studies. Future research should prospectively test and validate a construct of NMPO to disseminate to payers and health officials.
    Substance Abuse 02/2015; 36(2). DOI:10.1080/08897077.2014.993491 · 1.62 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Older adults with diabetes and dementia are at increased risk for hypoglycemia and other adverse events associated with tight glycemic control and are unlikely to experience long-term benefits. We examined risk factors for tight glycemic control in this population and use of medications associated with a high risk of hypoglycemia in the subset with tight control. This retrospective cohort study of national Veterans Affairs (VA) administrative/clinical data and Medicare claims for fiscal years (FYs) 2008-2009 included 15,880 veterans aged ≥65 with type 2 diabetes and dementia and prescribed antidiabetic medication. Multivariable regression analyses were used to identify sociodemographic and clinical predictors of hemoglobin A1c (HbA1c) control (tight, moderate, poor, or not monitored) and in patients with tight control, subsequent use of medication associated with a high risk of hypoglycemia (sulfonylureas, insulin). Fifty-two percent of patients had tight glycemic control (HbA1c <7% [53 mmol/mol]). Specific comorbidities, older age, and recent weight loss were associated with greater odds of tight versus moderate control, whereas Hispanic ethnicity and obesity were associated with lower odds of tight control. Among tightly controlled patients, 75% used sulfonylureas and/or insulin, with higher odds in patients who were male, black, or aged ≥75; had a hospital or nursing home stay in FY2008; or had congestive heart failure, renal failure, or peripheral vascular disease. Many older veterans with diabetes and dementia are at high risk for hypoglycemia associated with intense diabetes treatment and may be candidates for deintensification or alteration of diabetes medications. © 2015 by the American Diabetes Association. Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered.
    Diabetes Care 01/2015; 38(4). DOI:10.2337/dc14-0599 · 8.57 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Effective medications are a cornerstone of prevention and disease treatment, yet only about half of patients take their medications as prescribed, resulting in a common and costly public health challenge for the US health care system. Since poor medication adherence is a complex problem with many contributing causes, there is no one universal solution. This paper describes interventions that were not only effective in improving medication adherence among patients with diabetes, but were also potentially scalable (ie, easy to implement to a large population). We identify key characteristics that make these interventions effective and scalable. This information is intended to inform health care systems seeking proven, low resource, cost-effective solutions to improve medication adherence.
    Patient Preference and Adherence 01/2015; 9:139-49. DOI:10.2147/PPA.S69651 · 1.49 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Long-acting insulin analogues (eg, insulin glargine and insulin detemir) are an alternative to neutral protamine Hagedorn (NPH) insulin for maintaining glycemic control in patients with diabetes. Clinical trials comparing analogue insulin and NPH have neither been adequately powered nor had sufficient follow-up to examine long-term health outcomes. To compare the effects of NPH and long-acting insulin analogues on long-term outcomes. This retrospective observational study relied on administrative data from the Veterans Health Administration and Medicare from 2000 to 2010. Local variations in analogue insulin prescribing rates were used in instrumental variable models to control for confounding. Outcomes were assessed using survival models. The study population included US veterans dually enrolled in Medicare who received at least 1 prescription for oral diabetes medication and then initiated long-acting insulin between 2001 and 2009. Outcomes included ambulatory care-sensitive condition (ACSC) hospitalizations and mortality. There was no significant relationship between type of insulin and ACSC hospitalization or mortality. The hazard ratio for mortality of individuals starting a long-acting analogue insulin was 0.97 (95% CI, 0.85-1.11), and was 1.05 (95% CI, 0.95-1.16) for ACSC hospitalization. Differences in risk remained insignificant when predicting diabetes-specific ACSC hospitalizations, but starting on long-acting analogue insulin significantly increased the risk of a cardiovascular-specific ACSC hospitalization. We found no consistent difference in long-term health outcomes when comparing use of long-acting insulin analogues and NPH insulin. The higher cost of analogue insulin without demonstrable clinical benefit raises questions of its cost-effectiveness in the treatment of patients with diabetes.
    The American journal of managed care 01/2015; 21(3):e235-43. · 2.17 Impact Factor
  • Drug and Alcohol Dependence 01/2015; 146:e126. DOI:10.1016/j.drugalcdep.2014.09.260 · 3.28 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: When incretin mimetic (IM) medications were introduced in 2005, their effectiveness compared with other less-expensive second-line diabetes therapies was unknown, especially for older adults. Physicians likely had some uncertainty about the role of IMs in the diabetes treatment armamentarium. Regional variation in uptake of IMs may be a marker of such uncertainty. To investigate the extent of regional variation in the use of IMs among beneficiaries and estimate the cost implications for Medicare. This was a cross-sectional analysis of 2009-2010 claims data from a nationally representative sample of 238 499 Medicare Part D beneficiaries aged ≥65 years, who were continuously enrolled in fee-for-service Medicare and Part D and filled ≥1 antidiabetic prescription. Beneficiaries were assigned to 1 of 306 hospital-referral regions (HRRs) using ZIP codes. The main outcome was adjusted proportion of antidiabetic users in an HRR receiving an IM. Overall, 29 933 beneficiaries (12.6%) filled an IM prescription, including 26 939 (11.3%) for sitagliptin or saxagliptin and 3718 (1.6%) for exenatide or liraglutide. The adjusted proportion of beneficiaries using IMs varied more than 3-fold across HRRs, from 5th and 95th percentiles of 5.2% to 17.0%. Compared with non-IM users, IM users faced a 155% higher annual Part D plan ($1067 vs $418) and 144% higher patient ($369 vs $151) costs for antidiabetic prescriptions. Among older Part D beneficiaries using antidiabetic drugs, substantial regional variation exists in the use of IMs, not accounted for by sociodemographics and health status. IM use was associated with substantially greater costs for Part D plans and beneficiaries. © The Author(s) 2014.
    Annals of Pharmacotherapy 12/2014; 49(3). DOI:10.1177/1060028014563951 · 2.92 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Clinical trials often compare hypoglycemic medications on the basis of glycemic control but do not examine long-term outcomes (e.g., mortality). This study demonstrates an alternative approach to lengthening clinical trials to assess these long-term outcomes. To use observational quasi-experimental methods using instrumental variables (IVs) to compare the effect of two hypoglycemic medications, sulfonylureas (SUs) and thiazolidinediones (TZDs), on long-term outcomes. This study used administrative data from the Veterans Health Administration and Medicare from 2000 to 2010. The study population included US veterans dually enrolled in Medicare who received a prescription for metformin and then initiated SUs or TZDs. Patients could either continue on or discontinue metformin after the initiation of the second agent. Treatment was defined as starting either a SU or a TZD. Local variations in SU prescribing rates were used as instruments in IV models to control for selection bias. Survival models predicted all-cause mortality, ambulatory care sensitive condition hospitalizations, and stroke or heart attack (acute myocardial infarction). Starting on SUs compared to TZDs significantly increased the likelihood of experiencing mortality and ACSC hospitalization. The estimated hazard ratio for the effect of starting on SUs compared to TZDs was 1.50 (95% confidence interval [CI] 1.09-2.09) for all-cause mortality, 1.68 (95% CI 1.31-2.15) for ambulatory care sensitive condition hospitalization, and 1.15 (95% CI 0.80-1.66) for acute myocardial infarction or stroke. Our findings suggest increased risk of major adverse events associated with SUs as a second-line agent. Quasi-experimental IV methods may be an important alternative to lengthening clinical trials to assess long-term outcomes. Copyright © 2014 International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. All rights reserved.
    Value in Health 12/2014; 17(8):854-62. DOI:10.1016/j.jval.2014.08.2674 · 2.89 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Self-monitoring of blood glucose is a costly component of care for diabetes mellitus, with unclear benefits for patients not taking insulin. Veterans with dual Department of Veterans Affairs (VA) and Medicare benefits have access to test strips through both systems, raising the potential for overuse.
    JAMA Internal Medicine 11/2014; 175(1). DOI:10.1001/jamainternmed.2014.5405 · 13.25 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Background: Nursing home patients with dementia may be more likely to suffer adverse drug events from suboptimal prescribing. Previous studies have not used national samples, nor have they examined multiple types of suboptimal prescribing by dementia severity. Objective: To examine the prevalence of and factors associated with potentially suboptimal prescribing in older veteran nursing home patients with dementia. Methods: This is a retrospective descriptive study of 1303 veterans 65 years or older admitted between January 1, 2004, and June 30, 2005, with dementia for long stays (90+ days) to 133 Veterans Affairs Community Living Centers. Dementia severity was determined by the Cognitive Performance Scale and functional status dependences. Results: Overall, 70.2% with mild-moderate dementia (n = 1076) had underuse because they did not receive an acetylcholinesterase inhibitor (AChEI), and 27.2% had evidence of inappropriate use because of a drug-disease or drug-drug-disease interaction. Of the 227 with severe dementia, 36.1% had overuse by receiving an AChEI or lipid-lowering or other agents, and 25.1% had evidence of inappropriate use as a result of a drug-disease or drug-drug interaction. Multinomial logistic regression analyses among those with mild to moderate dementia identified that living in the South versus other regions was the single factor associated with all 3 types of suboptimal prescribing. In those with severe dementia, antipsychotic use was associated with all 3 suboptimal prescribing types. Conclusions: Potentially suboptimal prescribing was common in older veteran nursing home patients with dementia. Clinicians should develop a heightened awareness of these problems. Future studies should examine associations between potentially suboptimal prescribing and health outcomes in patients with dementia.
    Annals of Pharmacotherapy 11/2014; 49(1). DOI:10.1177/1060028014558484 · 2.92 Impact Factor
  • JAMA Pediatrics 09/2014; 168(11). DOI:10.1001/jamapediatrics.2014.1647 · 4.25 Impact Factor
  • Timothy S Anderson · Chester B Good · Walid F Gellad
    JAMA The Journal of the American Medical Association 08/2014; 312(5):558-9. DOI:10.1001/jama.2014.7578 · 30.39 Impact Factor
  • Yan Tang · Walid F Gellad · Aiju Men · Julie M Donohue
    [Show abstract] [Hide abstract]
    ABSTRACT: Little is known about how Medicare Part D plan features influence choice of generic versus brand drugs. To examine the association between Part D plan features and generic medication use. Data from a 2009 random sample of 1.6 million fee-for-service, Part D enrollees aged 65 years and above, who were not dually eligible or receiving low-income subsidies, were used to examine the association between plan features (generic cost-sharing, difference in brand and generic copay, prior authorization, step therapy) and choice of generic antidepressants, antidiabetics, and statins. Logistic regression models accounting for plan-level clustering were adjusted for sociodemographic and health status. Generic cost-sharing ranged from $0 to $9 for antidepressants and statins, and from $0 to $8 for antidiabetics (across 5th-95th percentiles). Brand-generic cost-sharing differences were smallest for statins (5th-95th percentiles: $16-$37) and largest for antidepressants ($16-$64) across plans. Beneficiaries with higher generic cost-sharing had lower generic use [adjusted odds ratio (OR)=0.97, 95% confidence interval (CI), 0.95-0.98 for antidepressants; OR=0.97, 95% CI, 0.96-0.98 for antidiabetics; OR=0.94, 95% CI, 0.92-0.95 for statins]. Larger brand-generic cost-sharing differences and prior authorization were significantly associated with greater generic use in all categories. Plans could increase generic use by 5-12 percentage points by reducing generic cost-sharing from the 75th ($7) to 25th percentiles ($4-$5), increasing brand-generic cost-sharing differences from the 25th ($25-$26) to 75th ($32-$33) percentiles, and using prior authorization and step therapy. Cost-sharing features and utilization management tools were significantly associated with generic use in 3 commonly used medication categories.
    Medical care 06/2014; 52(6):541-8. DOI:10.1097/MLR.0000000000000142 · 2.94 Impact Factor
  • Source
    Value in Health 05/2014; 17(3):A253. DOI:10.1016/j.jval.2014.03.1474 · 2.89 Impact Factor
  • JAMA The Journal of the American Medical Association 04/2014; 311(13):1353-5. DOI:10.1001/jama.2013.284925 · 30.39 Impact Factor

Publication Stats

423 Citations
446.28 Total Impact Points


  • 2008–2015
    • University of Pittsburgh
      • • Department of Medicine
      • • Division of General Internal Medicine
      Pittsburgh, Pennsylvania, United States
  • 2009–2012
    • RAND Corporation
      Santa Monica, California, United States
    • Brooks Rand
      Seattle, Washington, United States
  • 2006–2008
    • Brigham and Women's Hospital
      • • Division of General Internal Medicine and Primary Care
      • • Department of Medicine
      Boston, MA, United States
  • 2007
    • Harvard University
      Cambridge, Massachusetts, United States
    • Harvard Medical School
      • Department of Medicine
      Boston, Massachusetts, United States