Monday, October 19, 2015: 10:00 AM - 11:30 AM
Grand Ballroom FG (Hyatt Regency St. Louis at the Arch)

10:00 AM

Natalia Olchanski, MS, Joshua T. Cohen, Ph.D., Peter J. Neumann, Sc.D., John B. Wong, MD and David M. Kent, MD, MSc, Tufts Medical Center, Boston, MA


Purpose:   Because risk often varies across patients enrolled in randomized trials, average population trial and cost-effectiveness analysis (CEA) results do not apply to many patients.  Risk prediction models make it possible to incorporate individual risk and clinical effectiveness information to compute individualized CEA and identify patients for whom therapy is most appropriate.  The expected value of individualized care (EVIC) refers to the incremental monetized value of customizing care in this manner, compared to a uniform recommendation (treat all or none) based on the population average incremental cost-effectiveness ratio (ICER).  We explore factors influencing EVIC.

Methods:  We developed a general framework to calculate individualized ICERs as a function of individual outcome risk.  For a case study (tPA vs. streptokinase to treat possible myocardial infarction), we used simulation to explore how EVIC is influenced by risk model discriminatory power (c-statistic), population outcome prevalence, model calibration and willingness-to-pay (WTP) thresholds.  We characterized individual risk using beta distributions, the parameters for which we estimated from population prevalence and the risk model c-statistic.  We derived this relationship empirically from a database of 25 large randomized clinical trials.  We assumed that treatment effect (relative risk reduction) and life expectancy do not vary according to patient risk.


Results: For unbiased models (figure), EVIC is always non-negative and improvements in discrimination better targets treatment, increasing EVIC.   In our example, lower population outcome prevalence also increases EVIC.  EVIC also depends on the WTP threshold; when the WTP and population average ICER are close, EVIC increases because individualized information more often changes treatment decisions.  In contrast to unbiased prediction models, miscalibrated models can introduce mistakes and hence have a negative EVIC.  When the population average ICER is near the WTP threshold, decision making is more sensitive to bias and these “mistakes” are more common (making EVIC more negative).  In simulated examples, EVIC decreased for model c-statistic values of 0.6-0.7 and increased for c-statistic values of 0.8-0.9.


Conclusions:  In general, higher predictive model c-statistic values produce higher EVIC values.  This benefit is greatest when the population ICER is near the WTP threshold.  When models are miscalibrated, greater discriminating power can paradoxically reduce the EVIC under some circumstances.

Figure:  EVIC as a function of c-statistic at several levels of outcome prevalence and WTP thresholds

10:15 AM

Kejing Jiang, Stanford University Department of Management Science and Engineering, Stanford, CA, Jason Andrews, Stanford University School of Medicine, Stanford, CA and Jeremy D. Goldhaber-Fiebert, PhD, Stanford Health Policy, Centers for Health Policy and Primary Care and Outcomes Research, Department of Medicine, Stanford University, Stanford, CA
Purpose: In the previous 4 decades, there have been 19 major outbreaks of Ebola virus disease, but none have approached the magnitude of the 2014 outbreak in Liberia nor have they occurred in such large, urban settings. For the 2014 outbreak, outcry erupted due to delays in implementing safe burial practices and social distancing interventions. While experts expect future urban outbreaks, empirical evaluations of alternative population control strategies are infeasible, necessitating simulation modeling approaches to aid preparedness. 

Method: We developed a 5-compartment dynamic transmission model of Ebola for the 2014 Liberian outbreak and performed literature review to characterize model inputs and their uncertainty. We matched 2-week moving averages of new Ebola cases reported by the World Health Organization both before widespread burial and social distancing interventions began in Liberia (prior to September 2014) as well as afterwards (through mid-May 2015) by performing 10,000 Neldor-Mead search calibrations from random starting sets of inputs. By simultaneously calibrating to both periods, we recovered natural history parameters and intervention effectiveness. The objective function of the calibration was a weighted sum-of-squares where weights were the inverse of the standard error of the observed estimates under the binomial distribution. For analyses of alternative timings of interventions, we sampled 1,000 calibrated parameter sets with replacement from the 10,000, weighting the sampling by an approximation of the likelihood function so that better-fitting sets were more likely to be sampled.

Result: Compared to the observed 10,604 cumulative Ebola cases from the current outbreak in Liberia, our model predicts 10,519 cases [95%CrI: 9,755-10,992]. If interventions had been implemented earlier by 1 month, total cases are predicted at 1,904 [95%CrI: 1,359-2,951]. At 2 months earlier, these figures are 485 cases [95%CrI: 273-1,041]. 

Conclusion: Initiating safe burial and social distancing interventions earlier via better surveillance and epidemic preparedness has the potential to substantially decrease the impact of future Ebola epidemics in urban African settings.

10:30 AM

Haomiao Jin, MS1, Shinyi Wu, PhD1 and Paul Di Capua, MD, MBA2, (1)University of Southern California, Los Angeles, CA, (2)University of California, Los Angeles, Los Angeles, CA
Purpose: Approximately 30% of diabetes patients are suffering from depression, but nearly half of them are undiagnosed. Although universal depression screening improves diagnosis rates, it is a labor-intensive intervention. This study developed a clinical forecasting model for automatically detecting comorbid depression among patients with diabetes and applied the model to derive a screening policy to improve efficiency of depression screening.

Method: Machine learning methods were used to develop the model to forecast occurrence of major depression, measured by Patient Health Questionnaire 9-item score≥10. Predictors were selected using a correlation-based subset evaluation method from 20 risk factors of depression. Two linear models, Ridge logistic regression and multilayer perceptron, and two nonlinear models, support vector machine and random forest, were trained and validated on data pooled from two safety-net clinical trials of diabetes and depression (N=1793). The model with the best overall predictive ability, measured by area under receiver-operating curve (AUROC), was chosen as the ultimate model. Depression identification rate and measures relevant to provider resource and time were compared between a model-based policy that screens only patients predicted as being depression and alternative policies. These policies include universal screening and partial screening based on certain risk factors of depression such as depression history, diabetes severity, or either criteria.

Result: Seven predictors were selected to develop the prediction model: 1) gender, 2) Tolbert diabetes self-care 3) number of diabetes complications, 4) previous diagnosis of major depression, 5) number of ICD-9 diagnoses in past 6 months, 6) chronic pain, and 7) self-rated health status. Ridge logistic regression with the above seven predictors had the best overall predictive ability (AUROC=0.81) and was chosen as the ultimate model. Compared to universal screening, the model-based policy can save about 50-60% of provider resources and time but will miss identification of about 30% of depression cases. Partial-screening policy based on depression history alone yielded a very low rate of depression identification. Two other partial screening policies have depression identification rates similar to model-based policy but cost more in resources and time.

Conclusion: The depression prediction model developed in this study has compelling predictive ability. By adopting the model-based depression screening policy, healthcare providers can better prioritize the use of their resources and time while increasing efficiency in managing their patient population with depression.

10:45 AM

Mina Kabiri, MS1, Walid Gellad, MD, MPH1, Jagpreet Chhatwal, PhD2, Michael Dunn, MD, FACP3, Julie Donohue, PhD1 and Mark S. Roberts, MD, MPH1, (1)Department of Health Policy and Management, University of Pittsburgh, Pittsburgh, PA, (2)The University of Texas MD Anderson Cancer Center, Houston, TX, (3)Division of Gastroenterology, Hepatology and Nutrition, University of Pittsburgh School of Medicine, Pittsburgh, PA
Purpose: New highly effective therapies have changed the treatment paradigm for hepatitis C virus (HCV), with very high cure rates. However, the unknown true prevalence of HCV-infected individuals and their distribution of disease stages, along with high cost of treatment, present challenges for healthcare payers like state Medicaid programs with limited budgets for HCV treatment. Our objective was to estimate the prevalence of HCV in the Pennsylvania (PA) Medicaid.

Method: We used PA Medicaid claims data from 2007­–2012 to identify individuals diagnosed with HCV, individuals who received HCV therapies, and those who developed advanced liver disease due to HCV infection. To estimate the current HCV prevalence, we used an innovative approach of combining the results of claims data with a validated microsimulation model that accurately predicted national HCV prevalence in the United States. In this process, we accounted for trends in Medicaid enrollment, and adjusted the rates of treatment contraindications, such as substance abuse, that are often higher in Medicaid populations. We calibrated our model such that it simulated the observed number of patients diagnosed with HCV, and “hard outcomes” (liver transplants, hepatocellular carcinoma, decompensated cirrhosis) from 2007­–2012 claims data. Our model included historic as well as current HCV screening and treatment recommendations. From the calibrated PA model, we estimated the number of patients who will need treatment in 2015 and beyond by disease stage (represented by fibrosis scores F0–F4) and by HCV genotype.

Result: Our calibrated model matched the number of individuals with HCV diagnoses based on PA Medicaid claims data at 26,400 in 2012. The model estimated 22 liver transplants in 2012, closely matching the true incidence found in claims data. Our model estimated that 46,400 beneficiaries were infected with HCV in 2015, of whom 65% were aware of their disease, and 72% were treatment naïve. In the following 10 years, 8,500 new patients would be added to PA Medicaid either because of HCV screening or new enrollments.

Conclusion: We provide a novel approach to estimate the prevalence of HCV by using a combination of claims data and simulation modeling. Our results can assist state Medicaid programs in effective allocation of their resources to manage HCV patients in a rapidly changing clinical and policy environment.

11:00 AM

Mary-Ellen Hogan, BScPhm, PharmD, MSc1, Nicholas Mitsakakis, MSc PhD2, Vibhuti Shah, MD MSc3, Joel Katz, BA, MA, PhD4, Anna Taddio, BScPhm, MSc, PhD5 and Murray D Krahn, MD, MSc, FRCPC2, (1)Leslie Dan Faculty of Pharmacy, University of Toronto, Toronto, ON, Canada, (2)Toronto Health Economics and Technology Assessment (THETA) Collaborative, University of Toronto, Toronto, ON, Canada, (3)Institute of Health Policy, Management and Evaluation, University of Toronto, Department of Paediatrics, Mount Sinai Hospital, Toronto, ON, Canada, (4)Department of Psychology, York University, Toronto General Research Institute and Department of Anesthesia and Pain Management, University Health Network, Toronto, ON, Canada, (5)Leslie Dan Faculty of Pharmacy, University of Toronto; Department of Child Health Evaluative Sciences, Pharmacy, Hospital for Sick Children, Toronto, ON, Canada
Purpose:      Almost 1 in 5 adults has chronic pain.  New interventions are being developed to manage this widespread condition and cost-utility analyses of these technologies require robust data.  We aimed to estimate utilities using a population based sample of adults with chronic pain and examine the contribution of several factors.   Methods:    Health Utilities Index Mark 3 (HUI3) values, self-reported race/ethnicity and presence of arthritis, back problems, migraines, heart disease, stroke, diabetes and cancer were obtained from the Ontario survey responses of the Canadian Community Health Survey (CCHS) 2009-10.  The CCHS questions for presence of pain, severity and disability from pain were used to identify and stratify patients with chronic pain.  Income, aggregated diagnosis groups (ADGs, Johns Hopkins ACG system, a measure of comorbidity), age and sex were obtained from linked administrative data.  Ordinary least squares regression was used to investigate the impact of variables on utility.    Results:    A total of 15,901 responses for adults 18 – 64 years of age were available for analysis and 4,116 reported chronic pain.  In the pain cohort, mean age was 48 years (SD 12); 59% were female.  The average number of ADGs was 4.2 (SD 2.9).  People with chronic pain had a mean utility of 0.60, 0.22 points below the overall sample mean (see table).  Increasing income quintile was associated with an increase in utility (p<0.001) as was black race (p < 0.05) (versus Caucasian).  Aboriginal ethnicity was associated with a decrease in utility (p<0.001).   Presence of the following conditions was associated with a decrease in utility:  migraine, back problems, arthritis, suffering from the effects of a stroke, heart disease, diabetes and an additional ADG (all p<0.001).  Age, sex, and having cancer were not significantly associated with utility change.     Conclusions:    Utilities in people with chronic pain were very low and decreased with greater pain and more activity limitations.  A decrement of 0.22 is larger than seen with heart disease, diabetes, COPD, asthma and epilepsy.1  To our knowledge, this study is the first to estimate utilities in patients with chronic pain at the population level.  This data will be useful to inform future cost-utility analyses.   1Mittmann N, Trakas K, Risebrough N, Liu B.  Utility scores for chronic conditions in a community-dwelling population.  Pharmacoeconomics 1999; 15(4):369-376.

11:15 AM

Melissa Ross, MA1, John F.P. Bridges, PhD2, Xinyi Ng, BSc (Pharm)3, Emily J. Frosch, M.D.4, Gloria M. Reeves, M.D.5 and Susan dosReis, PhD1, (1)University of Maryland School of Pharmacy, Baltimore, MD, (2)Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, (3)University of Maryland Baltimore, Baltimore, MD, (4)Johns Hopkins School of Medicine, Baltimore, MD, (5)University of Maryland School of Medicine, Baltimore, MD
   Purpose: To understand the relationship between treatment-initiation concerns and outcomes of an evidence-based treatment among caregivers of a child with attention-deficit/hyperactivity disorder (ADHD).

   Methods: Caregivers of a 4-14 year-old child diagnosed with ADHD were recruited from pediatric primary and mental health clinics and family support groups across Maryland.  A case 1, balanced incomplete block design (BIBD), best-worst scaling (BWS) experiment assessed caregivers' most important concerns when initiating ADHD treatment.  Participants completed 16 choice tasks, each showing 6 of the 16 concerns from which one most and one least important concern was selected.  Demographic characteristics, caregiver-reported improvements resulting from medication, and additional desired ADHD changes also were reported. Preference utilities were estimated using conditional logit, effects coding, and assuming sequential best-worst responses. Scores were ranked to assess relative importance. Latent class analysis (LCA) was conducted to determine if there were distinct segments that prioritized concerns differently.  Reported outcomes were estimated based on the observed impact of treatment and additional desired changes.    Results: The 184 participants (m=42 years) were primarily the biological mother and Caucasian (68%).  Children were mostly male (71%) and using medication (81%).  The top-ranked utility scores influencing whether to engage in treatment were the child becoming a successful adult (1.71, p<.0001), school behavior improvements (1.55, p<.0001), and the doctor addressing their concerns (1.39, p<.0001).  Least important to treatment initiation were school pressures to medicate (-2.01, p<.0001) followed by issues related to stigma.  LCA yielded a three segment solution: short-term impact (36%), long-term impact (40%), and side-effects/safety (24%).  When considering caregiver-reported outcomes behavioral (48%), executive functioning (57%), and mood (10%) improvements were noted.  Of those reporting improvement in behavior, executive functioning, and school, 60%, 60%, and 45%, respectively, desired additional improvement in their child's ADHD (p<.05).       Conclusions: While 76% of caregivers' priorities when considering engaging in ADHD treatment were focused on outcome improvement, the majority had not realized full improvement, despite using an evidence-based medication treatment.  Those whose priorities and expectations are not met over the course of treatment are at high risk for disengagement from care. Additional research is needed to implement a shared-decision making process over the course of care to ensure that stated preferences for treatment outcomes correspond with observed patient-reported outcomes.