Publications

2016

Lo-Ciganic, Wei-Hsuan, Walid F Gellad, Haiden A Huskamp, Niteesh K Choudhry, Chung-Chou H Chang, Ruoxin Zhang, Bobby L Jones, Hasan Guclu, Seth Richards-Shubik, and Julie M Donohue. (2016) 2016. “Who Were the Early Adopters of Dabigatran?: An Application of Group-Based Trajectory Models.”. Medical Care 54 (7): 725-32. https://doi.org/10.1097/MLR.0000000000000549.

BACKGROUND: Variation in physician adoption of new medications is poorly understood. Traditional approaches (eg, measuring time to first prescription) may mask substantial heterogeneity in technology adoption.

OBJECTIVE: Apply group-based trajectory models to examine the physician adoption of dabigratran, a novel anticoagulant.

METHODS: A retrospective cohort study using prescribing data from IMS Xponent™ on all Pennsylvania physicians regularly prescribing anticoagulants (n=3911) and data on their characteristics from the American Medical Association Masterfile. We examined time to first dabigatran prescription and group-based trajectory models to identify adoption trajectories in the first 15 months. Factors associated with rapid adoption were examined using multivariate logistic regressions.

OUTCOMES: Trajectories of monthly share of oral anticoagulant prescriptions for dabigatran.

RESULTS: We identified 5 distinct adoption trajectories: 3.7% rapidly and extensively adopted dabigatran (adopting in ≤3 mo with 45% of prescriptions) and 13.4% were rapid and moderate adopters (≤3 mo with 20% share). Two groups accounting for 21.6% and 16.1% of physicians, respectively, were slower to adopt (6-10 mo post-introduction) and dabigatran accounted for <10% share. Nearly half (45.2%) of anticoagulant prescribers did not adopt dabigatran. Cardiologists were much more likely than primary care physicians to rapidly adopt [odds ratio (OR)=12.2; 95% confidence interval (CI), 9.27-16.1] as were younger prescribers (age 36-45 y: OR=1.49, 95% CI, 1.13-1.95; age 46-55: OR=1.34, 95% CI, 1.07-1.69 vs. >55 y).

CONCLUSIONS: Trajectories of physician adoption of dabigatran were highly variable with significant differences across specialties. Heterogeneity in physician adoption has potential implications for the cost and effectiveness of treatment.

Donohue, Julie M, Sharon-Lise T Normand, Marcela Horvitz-Lennon, Aiju Men, Ernst R Berndt, and Haiden A Huskamp. (2016) 2016. “Regional Variation in Physician Adoption of Antipsychotics: Impact on US Medicare Expenditures.”. The Journal of Mental Health Policy and Economics 19 (2): 69-78.

BACKGROUND: Regional variation in US Medicare prescription drug spending is driven by higher prescribing of costly brand-name drugs in some regions. This variation likely arises from differences in the speed of diffusion of newly-approved medications. Second-generation antipsychotics were widely adopted for treatment of severe mental illness and for several off-label uses. Rapid diffusion of new psychiatric drugs likely increases drug spending but its relationship to non-drug spending is unclear. The impact of antipsychotic diffusion on drug and medical spending is of great interest to public payers like Medicare, which finance a majority of mental health spending in the US.

AIMS: We examine the association between physician adoption of new antipsychotics and antipsychotic spending and non-drug medical spending among disabled and elderly Medicare enrollees.

METHODS: We linked physician-level data on antipsychotic prescribing from an all-payer dataset (IMS Health's XponentTM) to patient-level data from Medicare. Our physician sample included 16,932 US. psychiatrists and primary care providers with > 10 antipsychotic prescriptions per year from 1997-2011. We constructed a measure of physician adoption of 3 antipsychotics introduced during this period (quetiapine, ziprasidone and aripiprazole) by estimating a shared frailty model of the time to first prescription for each drug. We then assigned physicians to one of 306 U.S. hospital referral regions (HRRs) and measured the average propensity to adopt per region. Using 2010 data for a random sample of 1.6 million Medicare beneficiaries, we identified 138,680 antipsychotic users. A generalized linear model with gamma distribution and log link was used to estimate the effect of region-level adoption propensity on beneficiary-level antipsychotic spending and non-drug medical spending adjusting for patient demographic and socioeconomic characteristics, health status, eligibility category, and whether the antipsychotic was for an on- vs. off-label use.

RESULTS: In our sample, mean patient age was 62 years, 42% were male, and 86% had low-income. Half of antipsychotic users in Medicare had an on-label indication. The weighted average propensity to adopt the three new antipsychotics varied four-fold across HRRs. For every one standard deviation increase in the propensity to adopt there was a 5% increase in antipsychotic spending after adjusting for covariates (adjusted ratio of spending 1.05, 95% CI 1.01-1.08, p = 0.005). Physician propensity to adopt new antipsychotics was not associated with non-drug medical spending (adjusted ratio 0.96, 95% CI 0.91-1.01, p < 0.117).

DISCUSSION: These findings suggest wide regional variation in physicians' propensity to adopt new antipsychotic medications. While physician adoption of new antipsychotics was positively associated with antipsychotic expenditures, it was not associated with non-drug spending. Our analysis is limited to Medicare and may not generalize to other payers. Also, claims data do not allow for the measurement of health outcomes, which would be important to evaluate when calculating the value of rapid vs. slow technology adoption.

Chidi, Alexis P, Cindy L Bryce, Julie M Donohue, Michael J Fine, Douglas P Landsittel, Larissa Myaskovsky, Shari S Rogal, Galen E Switzer, Allan Tsung, and Kenneth J Smith. (2016) 2016. “Economic and Public Health Impacts of Policies Restricting Access to Hepatitis C Treatment for Medicaid Patients.”. Value in Health : The Journal of the International Society for Pharmacoeconomics and Outcomes Research 19 (4): 326-34. https://doi.org/10.1016/j.jval.2016.01.010.

BACKGROUND: Interferon-free hepatitis C treatment regimens are effective but very costly. The cost-effectiveness, budget, and public health impacts of current Medicaid treatment policies restricting treatment to patients with advanced disease remain unknown.

OBJECTIVES: To evaluate the cost-effectiveness of current Medicaid policies restricting hepatitis C treatment to patients with advanced disease compared with a strategy providing unrestricted access to hepatitis C treatment, assess the budget and public health impact of each strategy, and estimate the feasibility and long-term effects of increased access to treatment for patients with hepatitis C.

METHODS: Using a Markov model, we compared two strategies for 45- to 55-year-old Medicaid beneficiaries: 1) Current Practice-only advanced disease is treated before Medicare eligibility and 2) Full Access-both early-stage and advanced disease are treated before Medicare eligibility. Patients could develop progressive fibrosis, cirrhosis, or hepatocellular carcinoma, undergo transplantation, or die each year. Morbidity was reduced after successful treatment. We calculated the incremental cost-effectiveness ratio and compared the costs and public health effects of each strategy from the perspective of Medicare alone as well as the Centers for Medicare & Medicaid Services perspective. We varied model inputs in one-way and probabilistic sensitivity analyses.

RESULTS: Full Access was less costly and more effective than Current Practice for all cohorts and perspectives, with differences in cost ranging from $5,369 to $11,960 and in effectiveness from 0.82 to 3.01 quality-adjusted life-years. In a probabilistic sensitivity analysis, Full Access was cost saving in 93% of model iterations. Compared with Current Practice, Full Access averted 5,994 hepatocellular carcinoma cases and 121 liver transplants per 100,000 patients.

CONCLUSIONS: Current Medicaid policies restricting hepatitis C treatment to patients with advanced disease are more costly and less effective than unrestricted, full-access strategies. Collaboration between state and federal payers may be needed to realize the full public health impact of recent innovations in hepatitis C treatment.

Chhatwal, Jagpreet, Xiaojie Wang, Turgay Ayer, Mina Kabiri, Raymond T Chung, Chin Hur, Julie M Donohue, Mark S Roberts, and Fasiha Kanwal. (2016) 2016. “Hepatitis C Disease Burden in the United States in the Era of Oral Direct-Acting Antivirals.”. Hepatology (Baltimore, Md.) 64 (5): 1442-50. https://doi.org/10.1002/hep.28571.

UNLABELLED: Oral direct-acting antivirals (DAAs) represent a major advance in hepatitis C virus (HCV) treatment. Along with recent updates in HCV screening policy and expansions in insurance coverage, treatment demand in the United States is changing rapidly. Our objective was to project the characteristics and number of people needing antiviral treatment and HCV-associated disease burden in the era of oral DAAs. We used a previously developed and validated Hepatitis C Disease Burden Simulation model (HEP-SIM). HEP-SIM simulated the actual clinical management of HCV from 2001 onward, which included antiviral treatment with pegylated interferon (Peg-IFN)-based therapies as well as the recent oral DAAs, risk-based and birth-cohort HCV screening, and the impact of the Affordable Care Act. We also simulated two hypothetical scenarios-no treatment and treatment with Peg-IFN-based therapies only. We estimated that in 2010, 2.5 (95% confidence interval [CI], 1.9-3.1) million noninstitutionalized people were viremic, which dropped to 1.9 (95% CI, 1.4-2.6) million in 2015, and projected to drop below 1 million by 2020. A total of 1.8 million HCV patients will receive HCV treatment from the launch of oral DAAs in 2014 until 2030. Based on current HCV management practices, it will take 4-6 years to treat the majority of patients aware of their disease. However, 560,000 patients would still remain unaware by 2020. Even in the oral DAA era, 320,000 patients will die, 157,000 will develop hepatocellular carcinoma, and 203,000 will develop decompensated cirrhosis in the next 35 years.

CONCLUSIONS: HCV-associated disease burden will still remain substantial in the era of oral DAAs. Increasing HCV screening and treatment capacity is essential to further decreasing HCV burden in the United States. (Hepatology 2016;64:1442-1450).

Huskamp, Haiden A, Shelly F Greenfield, Elizabeth A Stuart, Julie M Donohue, Kenneth Duckworth, Elena M Kouri, Zirui Song, Michael E Chernew, and Colleen L Barry. (2016) 2016. “Effects of Global Payment and Accountable Care on Tobacco Cessation Service Use: An Observational Study.”. Journal of General Internal Medicine 31 (10): 1134-40. https://doi.org/10.1007/s11606-016-3718-y.

BACKGROUND: Tobacco use is the leading cause of preventable death and disability. New payment and delivery system models including global payment and accountable care have the potential to increase use of cost-effective tobacco cessation services.

OBJECTIVE: To examine how the Alternative Quality Contract (AQC) established in 2009 by Blue Cross Blue Shield of Massachusetts (BCBSMA) has affected tobacco cessation service use.

DESIGN: We used 2006-2011 BCBSMA claims and enrollment data to compare adults 18-64 years in AQC provider organizations to adults in non-AQC provider organizations. We examined the AQC's effects on all enrollees; a subset at high risk of tobacco-related complications due to certain medical conditions; and behavioral health service users.

MAIN MEASURES: We examined use of: (1) any cessation treatment (pharmacotherapy or counseling); (2) varenicline or bupropion; (3) nicotine replacement therapies (NRTs); (4) cessation counseling; and (4) combination therapy (pharmacotherapy plus counseling). We also examined duration of pharmacotherapy use and number of counseling visits among users.

KEY RESULTS: Rates of tobacco cessation treatment use were higher following implementation of the AQC relative to the comparison group overall (2.02 vs. 1.87 %, p < 0.0001), among enrollees at risk for tobacco-related complications (4.97 vs. 4.66 %, p < 0.0001), and among behavioral health service users (3.67 vs. 3.25 %, p < 0.0001). Statistically significant increases were found for use of varenicline or bupropion alone, counseling alone, and combination therapy, but not for NRT use, pharmacotherapy duration, or number of counseling visits among users.

CONCLUSIONS: In its initial three years, the AQC was associated with increases in use of tobacco cessation services.

Lo-Ciganic, Wei-Hsuan, Julie M Donohue, Bobby L Jones, Subashan Perera, Joshua M Thorpe, Carolyn T Thorpe, Zachary A Marcum, and Walid F Gellad. (2016) 2016. “Trajectories of Diabetes Medication Adherence and Hospitalization Risk: A Retrospective Cohort Study in a Large State Medicaid Program.”. Journal of General Internal Medicine 31 (9): 1052-60. https://doi.org/10.1007/s11606-016-3747-6.

BACKGROUND: Numerous interventions are available to boost medication adherence, but the targeting of these interventions often relies on crude measures of poor adherence. Group-based trajectory models identify individuals with similar longitudinal prescription filling patterns. Identifying distinct adherence trajectories may be more useful for targeting interventions, although the association between adherence trajectories and clinical outcomes is unknown.

OBJECTIVE: To examine the association between adherence trajectories for oral hypoglycemics and subsequent hospitalizations among diabetes patients.

DESIGN: Retrospective cohort study.

PATIENTS: A total of 16,256 Pennsylvania Medicaid enrollees, non-dually eligible for Medicare, initiating oral hypoglycemics between 2007 and 2009.

MAIN MEASURES: We used group-based trajectory models to identify trajectories of oral hypoglycemics in the 12 months post-treatment initiation, using monthly proportion of days covered (PDC) as the adherence measure. Multivariable Cox proportional hazard models were used to examine the association between trajectories and time to first diabetes-related hospitalization/emergency department (ED) visits in the following year. We used the C-index to compare prediction performance between adherence trajectories and dichotomous cutpoints (annual PDC <80 vs. ≥80 %).

RESULTS: The mean annual PDC was 0.58 (SD 0.32). Seven trajectories were identified: perfect adherers (9 % of the cohort), nearly perfect adherers (31.4 %), moderate adherers (21.0 %), low adherers (11.0 %), late discontinuers (6.8 %), early discontinuers (9.7 %), and non-adherers with only one fill (11.1 %). Compared to perfect adherers, trajectories of moderate adherers (HR = 1.48, 95 % CI 1.25, 1.75), low adherers (HR = 1.51, 95 % CI 1.25, 1.83), and non-adherers with only one fill (HR = 1.35, 95 % CI 1.09, 1.67) had greater risk of diabetes-related hospitalizations/ED visits. Predictive accuracy was improved using trajectories compared to dichotomized cutpoints (C-index = 0.714 vs. 0.652).

CONCLUSIONS: Oral hypoglycemic treatment trajectories were highly variable in this large Medicaid cohort. Low and moderate adherers and those filling only one prescription had a modestly higher risk of hospitalizations/ED visits compared to perfect adherers. Trajectory models may be valuable in identifying specific non-adherence patterns for targeting interventions.

Donohue, J M, M Mastrovich, and K J Resch. (2016) 2016. “Spectrally Engineering Photonic Entanglement With a Time Lens.”. Physical Review Letters 117 (24): 243602. https://doi.org/10.1103/PhysRevLett.117.243602.

A time lens, which can be used to reshape the spectral and temporal properties of light, requires the ultrafast manipulation of optical signals and presents a significant challenge for single-photon application. In this work, we construct a time lens based on dispersion and sum-frequency generation to spectrally engineer single photons from an entangled pair. The strong frequency anticorrelations between photons produced from spontaneous parametric down-conversion are converted to positive correlations after the time lens, consistent with a negative-magnification system. The temporal imaging of single photons enables new techniques for time-frequency quantum state engineering.

Huskamp, Haiden A, Marcela Horvitz-Lennon, Ernst R Berndt, Sharon-Lise T Normand, and Julie M Donohue. (2016) 2016. “Patterns of Antipsychotic Prescribing by Physicians to Young Children.”. Psychiatric Services (Washington, D.C.) 67 (12): 1307-14.

OBJECTIVE: Antipsychotic use among young children has grown rapidly despite a lack of approval by the U.S. Food and Drug Administration (FDA) for broad use in this age group. Characteristics of physicians who prescribed antipsychotics to young children were identified, and prescribing patterns involving young children and adults were compared.

METHODS: Physician-level prescribing data from IMS Health's Xponent database were linked with American Medical Association Masterfile data and analyzed. The sample included all U.S. psychiatrists and a random sample of 5% of family medicine physicians who wrote at least ten antipsychotic prescriptions per year from 2008 to 2011 (N=31,713). Logistic and hierarchical binomial regression models were estimated to examine physician prescribing for children ages zero to nine, and the types and numbers of ingredients used for children versus adults ages 20 to 64 were compared.

RESULTS: Among antipsychotic prescribers, 42.2% had written at least one antipsychotic prescription for young children. Such prescribing was more likely among physicians age ≤39 versus ≥60 (odds ratio [OR]=1.70) and physicians in rural versus nonrural areas (OR=1.11) and was less likely among males (OR=.93) and graduates of a top-25 versus a lower-ranked U.S. medical school (OR=.87). Among physicians who prescribed antipsychotics to young children and adults, 75.0% of prescriptions for children and 35.7% of those for adults were for drugs with an FDA-approved indication for that age. Fewer antipsychotic agents were prescribed for young children (median=2) versus adults (median=7).

CONCLUSIONS: Prescribing antipsychotics for young children was relatively common, but prescribing patterns differed between young children and adults.

Lo-Ciganic, Wei-Hsuan, Walid F Gellad, Haiden A Huskamp, Niteesh K Choudhry, Chung-Chou H Chang, Ruoxin Zhang, Bobby L Jones, Hasan Guclu, Seth Richards-Shubik, and Julie M Donohue. (2016) 2016. “Who Were the Early Adopters of Dabigatran?: An Application of Group-Based Trajectory Models.”. Medical Care 54 (7): 725-32. https://doi.org/10.1097/MLR.0000000000000549.

BACKGROUND: Variation in physician adoption of new medications is poorly understood. Traditional approaches (eg, measuring time to first prescription) may mask substantial heterogeneity in technology adoption.

OBJECTIVE: Apply group-based trajectory models to examine the physician adoption of dabigratran, a novel anticoagulant.

METHODS: A retrospective cohort study using prescribing data from IMS Xponent™ on all Pennsylvania physicians regularly prescribing anticoagulants (n=3911) and data on their characteristics from the American Medical Association Masterfile. We examined time to first dabigatran prescription and group-based trajectory models to identify adoption trajectories in the first 15 months. Factors associated with rapid adoption were examined using multivariate logistic regressions.

OUTCOMES: Trajectories of monthly share of oral anticoagulant prescriptions for dabigatran.

RESULTS: We identified 5 distinct adoption trajectories: 3.7% rapidly and extensively adopted dabigatran (adopting in ≤3 mo with 45% of prescriptions) and 13.4% were rapid and moderate adopters (≤3 mo with 20% share). Two groups accounting for 21.6% and 16.1% of physicians, respectively, were slower to adopt (6-10 mo post-introduction) and dabigatran accounted for <10% share. Nearly half (45.2%) of anticoagulant prescribers did not adopt dabigatran. Cardiologists were much more likely than primary care physicians to rapidly adopt [odds ratio (OR)=12.2; 95% confidence interval (CI), 9.27-16.1] as were younger prescribers (age 36-45 y: OR=1.49, 95% CI, 1.13-1.95; age 46-55: OR=1.34, 95% CI, 1.07-1.69 vs. >55 y).

CONCLUSIONS: Trajectories of physician adoption of dabigatran were highly variable with significant differences across specialties. Heterogeneity in physician adoption has potential implications for the cost and effectiveness of treatment.