Tue May 10
2:00 PM - 3:00 PM
On-Demand Podium Session
View anytime on this date, after 2:00pm, until June 17. Applications of Discrete Choice Experiments for Patient Preference Elicitation
On-demand
Moderator
Ellen Janssen, PhD
The Janssen Pharmaceutical Companies of J&J, Baltimore, MD, USA
P53: Patient Preferences for Multiple Myeloma Treatment: A Stated Preference Survey Using Discrete Choice Experiment and Swing Weighting
2:15PM - 2:30PM
Janssens R 1 , Lang T2 , Vallejo A2 , Galinsky J2 , Morgan K2 , Plate A2 , Verschueren M3 , Schoefs E1 , Vanhellemont A4 , Delforge M4 , Hellem Schjesvold F5 , Cabezudo E6 , Vandebroek M3 , Stevens H7 , Simoens S1 , Huys I1 1 KU Leuven, Leuven, VBR, Belgium, 2 Myeloma Patients Europe, Brussels, Belgium, 3 KU Leuven, Leuven, Belgium, 4 University Hospitals Leuven, Leuven, Belgium, 5 Oslo University Hospital, Oslo, Norway, 6 H. Moises Broggi / ICO-Hospitalet, Barcelona, Spain, 7 Université libre de Bruxelles, Brussels, Belgium
OBJECTIVES: Patient preferences have become an important focus for inclusion in drug development and evaluation. Understanding patient preferences is especially valuable in Multiple Myeloma (MM), where the rapid development of treatments with diverging benefit-risk profiles raises uncertainty about what matters most to patients. This study aimed to investigate which treatment attributes (side-effects, symptoms, efficacy outcomes, uncertainties) are most important to MM patients.
METHODS: A preference survey incorporating Discrete Choice Experiment (DCE) and Swing Weighting (SW) was widely disseminated through the European MM patient population. The survey was developed using a qualitative study during which MM patients (n=24) reached consensus on the attributes and levels included in the survey. MM patients and patient organizations provided extensive feedback during the survey development and piloting.
RESULTS: 393 MM patients across 21 countries participated. Patients were heterogeneous regarding years since diagnosis (M: 6) and prior therapies (M: 3). While life expectancy was most important to patients with the most and least prior therapies, quality of life-related attributes such as pain and mobility problems were most important to participants with intermediary treatment experience. Significant preference heterogeneity was revealed depending on participants’ side-effect and symptom experience. Participants highlighted the difficulty of trading-off between life expectancy and quality of life, and between physical and mental health. Patients demanded psychological support to cope with their symptoms, side-effects and uncertainties.
CONCLUSIONS: Preferences elicited from patients reveal the need for the systematic inclusion and prioritization of quality of life-related treatment outcomes by drug developers, regulators, Health Technology Assessment bodies, and healthcare providers in MM drug development, evidence generation, evaluation, and clinical practice. In order for patients to make informed choices in preference studies, researchers should involve patients and patient organizations during the selection of the attributes, levels, and explanations, how the preference questions are asked, the visuals and survey technology.
P56: Patient Preferences in Multiple Myeloma: A Discrete Choice Experiment
2:00PM - 2:15PM
Tervonen T1 , Duenas A2 , Collacott H2 , Lam A3 , Gries KS3 , Carson R3 , Trevor N4 , Krucien N2 , He J 3 1 Evidera, London, UK; University Medical Center Groningen, London, UK, 2 Evidera, London, LON, UK, 3 Janssen Global Services, LLC, Raritan, NJ, USA, 4 Janssen-Cilag Ltd, High Wycombe, BKM, UK
OBJECTIVES
: This study assessed the preferences of patients with multiple myeloma (MM) for treatment attributes and evaluated the impact of health-related quality of life (HRQoL) on those preferences.
METHODS
: Patients in the UK, France, and Germany with physician-confirmed transplant eligible (TE) or transplant ineligible (TIE) newly diagnosed MM (NDMM) or relapsed/refractory MM (RRMM) completed a discrete choice experiment (DCE). HRQoL was assessed using the EuroQoL Five Dimension (EQ-5D) questionnaire. Preferences for four benefit attributes (increased life expectancy, increased time to relapse, reduced pain, and reduced fatigue), three administration attributes (method of administration, frequency of administration, and monitoring), and one risk attribute (risk of severe infection) were analyzed with a multinominal logit model. Differences between subgroups were also analyzed.
RESULTS
: 300 patients completed the DCE (TE NDMM, n=108; TIE NDMM, n=105; RRMM, n=87). Median age was 68 years; most common symptoms were fatigue (69%), bone pain (63%), and sleepiness/tiredness (55%). Median EQ-5D score was 0.8 (IQR 0.7-0.9), with no differences by disease stage. Patients most valued reduced pain and fatigue and increased life expectancy. In general, patients favored shorter injection/monitoring time and less frequent administration. Patients were willing to make trade-offs, such as accepting a reduction of ≥ 2.7 years of life expectancy for no pain vs extreme pain and ≥ 2.0 years for no fatigue vs constant fatigue. Patients with lower EQ-5D scores were willing to trade more years of life expectancy than those with higher scores for reductions in pain or fatigue. Disease stage (NDMM vs RRMM), country, and age did not affect preferences.
CONCLUSIONS
: Patients with MM valued treatments that reduced pain and fatigue and were willing to trade life expectancy for better quality of life. HRQoL influenced patient preferences, suggesting that this is an important attribute that should be considered as part of the treatment decision-making process.
P55: A Discrete Choice Experiment of Patient-Informed Preferences for Major Depressive Disorder Treatment
2:30PM - 2:45PM
Dosreis S 1 , Amill-Rosario A1 , Ali C2 , Elonge E3 , Xie R4 , Chapman R5 , Slejko JF1 1 University of Maryland School of Pharmacy, Baltimore, MD, USA, 2 University of Maryland School of Pharmacy, Sandy Spring, MD, USA, 3 University of Maryland Baltimore, Baltimore, MD, USA, 4 The Innovation and Value Initiative, Newton, MA, USA, 5 The Innovation and Value Initiative, Alexandria, VA, USA
OBJECTIVE : Major depressive disorder (MDD) affects 10% of US adults, of which 30% experience moderate/severe impairment despite treatment. While value assessment often relies on clinically derived measures, how MDD treatment impacts individuals’ daily lives and what matters most to them is also relevant for value assessment. The study aimed to assess preferences for patient-informed attributes of MDD treatment.
METHODS
: A discrete choice experiment (DCE), developed with meaningful stakeholder engagement, was administered to 150 community-dwelling adults aged 18 and older diagnosed with MDD. We excluded post-partum depression and comorbid bipolar, psychosis, and cognitive disorders. A diverse sample from across the US was recruited via ResearchMatch.org. The DCE had six attributes, each with three levels: treatment modality, time to effect, days of hopefulness, improvement in productivity, impact on relationships, and out-of-pocket costs. An orthogonal array design with 100 D-efficiency yielded six choice tasks. A conditional logit model generated preference weights for each attribute level. The relative attribute importance is the proportion each attribute contributes to the total variance in preference weights.
RESULTS
: Participants were 40 years-old on average, 62% male, 28% female, and 10% non-binary/transgender. Race/ethnicity demographics reflected 45% White, 15% Black, 19% Latino, 13% Asian, and 8% mixed race. The relative attribute importance, in rank order, is out-of-pocket costs (30%), social relationships (29%), hopeful days (18%), treatment modality (9%), productivity (7%), and time to treatment effect (7%). Participants preferred treatment modalities that included medication, psychotherapy, and other services (e.g., brain stimulation); improved social relations; hopefulness six-days per week, and
< $90 USD monthly out-of-pocket costs.
CONCLUSION : Among individuals with MDD, social and life impact outcomes are preferred over time to treatment effect or productivity, which are more commonly used in economic evaluation. To reflect patient value more completely, these attributes should be considered in value assessments as well.
P54: Patient Preferences for Treatment of BCG-Unresponsive Non-Muscle Invasive Bladder Cancer: A Discrete Choice Experiment
2:45PM - 3:00PM
Collacott H 1 , Rentz A2 , Krucien N1 , Heidenreich S1 , Ghatnekar O3 1 Evidera, London, LON, UK, 2 Evidera, Bethesda, MD, USA, 3 Ferring International PharmaScience Center, Copenhagen, Denmark
OBJECTIVES:
First-line treatments for high-risk non-muscle invasive bladder cancer (NMIBC) include transurethral resection of the bladder tumour and Bacillus Calmette-Guerin (BCG) intravesical therapy. If BCG therapy fails, patients are offered bladder removal surgery (radical cystectomy (RC)), a major surgical intervention with impact on quality of life. This study quantifies the trade-offs patients are willing to make when choosing between RC and an alternative medical treatment. METHODS:
An online discrete choice experiment was completed by adults in the UK, France, Germany and Canada who were, or had been, treated with BCG therapy, or have undergone RC for NMIBC. Participants repeatedly chose between immediate RC and two hypothetical medical treatments that varied in: time-to-RC; risk of progressing to MIBC (muscle invasive bladder cancer) while on treatment; risk of serious side effects (SAEs); and administration. Preferences were analysed with an error-component logit model. Relative attribute importance (RAI) scores and minimum acceptable time to RC were obtained to explore patients’ treatment priorities.
RESULTS:
The study included 107 patients (64% male, average age 63 years, 93% diagnosed with NMIBC 1-5 years before; 82% Eastern Cooperative Oncology Group performance status ≤2). Delay in time to RC had the largest influence on treatment preferences (RAI = 55.0%), followed by reducing the risk of progressing to MIBC while on treatment (RAI = 25.1%), and less frequent administrations (RAI 11.6%). Risk of SAEs (RAI = 8.3%) was the least important attribute. Patients were willing to accept a 20% increase in the risk of progressing to MIBC for an additional 2.3-years until RC, and a 10% increase in the risk of SAEs for an additional 0.8-years until RC.
CONCLUSIONS:
Patients treated with BCG for NMIBC placed a high importance on delaying RC and were willing to accept increased risks of both SAEs and progression to MIBC for prolonged bladder preservation.
Real-World Evidence for Comparative Effectiveness, Safety and Adherence Evaluations
On-demand
Moderator
Laura Bozzi, MS, PhD
Janssen of J&J, Titusville, NJ, USA
Laura Bozzi, MS, PhD, is Manager of Benefit-Risk Assessment in the Global R&D Epidemiology Department at Janssen Pharmaceuticals of Johnson & Johnson in New Jersey. Laura started at Janssen in July 2021 and her role focuses on structured benefit-risk assessment to support regulatory submissions across therapeutic areas. Additionally, Laura leads patient preference studies to inform benefit-risk assessment and support internal decision-making. Prior to joining Janssen, Laura completed her doctorate at the University of Maryland, Baltimore (UMB) in the Pharmaceutical Health Services Research program where she received a R36 Health Services Research Dissertation Grant from the Agency for Health Research and Quality. As a graduate research assistant at UMB, she was apart of the Patient-Driven Values in Healthcare Evaluation (PAVE) Center where her research consisted of the design, implementation, and analysis of discrete choice experiments to understand patient preferences as it pertains to treatment decisions in the areas of COPD and Major Depressive Disorder. Additionally, she has served project coordinator for PAVE, developing high level strategy for engaging internal and external stakeholders (e.g. non-profit organizations, patient advocated, and industry partners) on PCORI, AHRQ, and PhRMA funding and research opportunities. She obtained her Bachelor of Science degree in Biological Sciences from the University of Connecticut and her Master of Science degree in Epidemiology & Genetics from UMB.
P74: Real-World Comparative Effectiveness of Pegfilgrastim Biosimilars Versus Originator
2:15PM - 2:30PM
Wang CY 1 , Park H1 , Heldermon CD2 , Vouri SM1 , Brown JD1 1 Center for Drug Evaluation & Safety, Department of Pharmaceutical Outcomes and Policy, Gainesville, FL, USA, 2 University of Florida, College of Medicine, Gainesville, FL, USA
OBJECTIVES:
Real-world evidence on the clinical effectiveness of pegfilgrastim biosimilars are limited. This study compared the incidence of febrile neutropenia (FN) between pegfilgrastim biosimilars (pegfilgrastim-jmdb, pegfilgrastim-cbqv) and originator users.
METHODS:
A retrospective cohort study using 2019 IBM MarketScan Commercial and Medicare Supplemental database was conducted in adult patients with cancer initiating myelosuppressive chemotherapy courses. At least 2 cancer diagnoses (at least 7 days apart within +/- 30 days of the initiation) were required. The following exclusion criteria were applied: (1) ≥2 solid cancers; (2) acute myeloid leukemia; (3) autologous peripheral blood progenitor cell collection; (4) bone marrow transplantation; (5) using pegfilgrastim on-body-injector/with unknown route; (6) using more than 1 granulocyte-colony stimulating factors product. Cumulative incidence of FN associated hospitalization was measured by ICD-10 diagnosis codes (neutropenia, fever, or infection) in the first cycle. After 1:1 propensity score (PS) matching, we compare FN risk between biosimilars and originator users using equivalence (with a margin of 6%) and superiority hypothesis tests.
RESULTS:
A total of 2,045 patients were included (445 used pegfilgrastim-jmdb, 636 used pegfilgrastim-cbqv, and 964 used pegfilgrastim originator). After PS matching, 13 out of 445 originator users and 17 out of 445 pegfilgrastim-jmdb users developed FN (risk difference was 0.9%; p-value was 0.4575 for superiority test indicating no difference ;p-value was <0.0001 for equivalence test indicating statistical equivalence). After PS matching, 14 out of 633 originator users and 16 out of 633 pegfilgrastim-cbqv users developed FN (risk difference was 0.32%; p-value was 0.7117 for superiority test indicating no difference; p-value was <0.0001 for equivalence test indicating statistical equivalence).
CONCLUSIONS:
In this real-world study of patients with cancer receiving myelosuppressive chemotherapy, there was no difference in FN risk between patients receiving pegfilgrastim originator and biosimilars in the first cycle.
P76: Comparative Adherence Trajectories across Oral Disease-Modifying Agents in Multiple Sclerosis
2:30PM - 2:45PM
Earla JR 1 , Hutton GJ2 , Aparasu RR3 1 University of Houston, College of Pharmacy, Fords, NJ, USA, 2 Baylor College of Medicine Medical Center, McNair Campus, Houston, TX, USA, 3 University of Houston, College of Pharmacy, Houston, TX, USA
OBJECTIVES:
The oral Disease-Modifying Agents (DMAs) such as fingolimod (FIN), teriflunomide (TER), and dimethyl fumarate (DMF) are convenient alternatives to injectable DMAs for Multiple Sclerosis (MS). However, there is limited evidence regarding the comparative adherence patterns across different oral DMAs. This study compared the adherence trajectories between FIN, TER, and DMF users with MS.
METHODS:
A retrospective longitudinal study was conducted involving adults (≥18 years) identified with MS (ICD-9/10-CM:340/G35 and a DMA prescription) from the 2015–2019 IBM MarketScan Commercial Claims Database. Patients were classified as incident FIN-, TER- or DMF-users based on the index DMA with one year of washout period. The DMA adherence trajectories based on Proportion Days Covered (PDC) were examined using Group-Based Trajectory Modeling (GBTM) during the one year after the treatment initiation. Generalized boosting models (GBM)-based Inverse Probability Treatment Weights (IPTW) were incorporated in multinomial logistic regression to assess the comparative adherence trajectories across oral DMAs with FIN group as reference category.
RESULTS:
The study cohort consisted of 1,913 MS patients who were initiated with FIN (24.2%,n=462), TER (23.9%,n=468), and DMF (51.9%,n=993) during 2016–2018. The adherence rate (PDC≥0.8) among FIN, TER, and DMF users was found to be 70.8%(n=327), 59.6%(n=273), and 61.0%(n=606), respectively. The GBTM grouped study subjects into three adherence trajectories – Complete Adherers-59.1%, Slow Decliners-22.6% and Rapid Discontinuers-18.3%. The multinomial logistic regression model involving GBM-based IPTW revealed that TER (adjusted odds ratio [aOR]-2.32, 95% CI:1.57-3.42) and DMF (aOR-2.50, 95% CI:1.62-3.88) users had higher odds to be rapid discontinuers relative to FIN users. In addition, TER users are more likely (aOR-1.50, 95% CI:1.06-2.13) to be slow decliners compared to FIN.
CONCLUSIONS:
Teriflunomide and dimethyl fumarate were associated with poorer adherence trajectories than fingolimod. More research is needed to evaluate the clinical implications of these adherence trajectories of oral DMAs for the management of MS.
P73: Real-World Effectiveness and Safety of Non-Vitamin K Antagonist Oral Anticoagulants (NOAC) in Patients with Non-Valvular Atrial Fibrillation (NVAF): A Retrospective Cohort Study in Singapore
2:00PM - 2:15PM
Foo W1 , Hui T 2 , Ong SKB1 , Ng KH1 1 Agency for Care Effectiveness, Ministry of Health, Singapore, Singapore, 2 Agency for Care Effectiveness, Ministry of Health, Singapore, 01, Singapore
OBJECTIVES:
Non-vitamin K antagonist oral anticoagulants (NOACs) are increasingly being used over warfarin for stroke prevention in non-valvular atrial fibrillation (NVAF) patients. However, there are limited studies on clinical outcomes of NOAC in a multi-ethnic Asian country such as Singapore. This study aims to compare the real-world effectiveness and safety between NOAC and warfarin in NVAF patients and assess the potential long-term impact on the healthcare system in Singapore.
METHODS:
Using government’s linked national health record databases, patients aged 18 years and above initiated on NOAC or warfarin for NVAF from 2015 to 2018 were included in this retrospective cohort study. Patient’s demographics, comorbidities, and comedications were balanced using propensity score matching. Primary efficacy and safety endpoints were stroke or systematic embolism (SE) and intracranial haemorrhage (ICH) or gastrointestinal (GI) bleeding, respectively. Effect size was estimated using hazard ratios (HRs) with 95% confidence intervals (CIs) from Cox regression. A Markov model was used to extrapolate the number of deaths and hospitalisations avoided with NOAC initiation.
RESULTS:
After propensity score matching, 3,315 comparable patients in each cohort remained. Over a median follow-up period of 2.0 years, NOAC users were less likely to develop a stroke or SE than warfarin users (HR 0.83; 95% CI 0.70-0.99). NOAC initiation was associated with a lower incidence rate of any ICH or GI bleed (HR 0.77; 95% CI 0.66-0.91) and all-cause mortality (HR 0.81; 95% CI 0.70-0.94). NOAC initiation was estimated to avoid 1,270 deaths, 622 stroke/SE hospitalisations and 1,311 ICH/GI bleed hospitalisations, which could contribute to $28 million saved over 10 years.
CONCLUSION : Consistent with clinical trials and other real-world findings, this large cohort study confirmed that NOAC reduced risk of stroke or SE, ICH/GI bleed and avoid all-cause mortality, and has potential to bring about cost savings to the healthcare system.
P75: Impact of High-Dose Vs Standard-Dose Influenza Vaccine on Respiratory Hospitalizations Among Adults
2:45PM - 3:00PM
Bianchini M 1 , Wright GC2 , Anderson HD2 , Perraillon MC2 , Lindrooth RC2 1 University of Colorado Anschutz, Denver, CO, USA, 2 University of Colorado Anschutz, Aurora, CO, USA
OBJECTIVES:
The high-dose (HD) influenza vaccine is approved for use in adults ages ≥65 with evidence for reducing influenza infections and hospitalizations. The objective of this study was to determine the impact of the HD vs standard-dose (SD) influenza vaccine on hospitalizations among adults ages 50-80.
METHODS:
A fuzzy regression discontinuity design was used to estimate the causal effect of the HD vaccine on respiratory hospitalizations. This design takes advantage of the discontinuity in likelihood of receiving the HD vaccine at age 65 to compare the outcomes of adults immediately above and below age 65. Data was extracted from IQVIA claims database for adults with an insurance claim for a HD or SD vaccine during the 2012-2018 influenza seasons (September-April). Outcomes and covariates were identified using International Classification of Diseases codes. The primary outcome was respiratory-related hospitalization. Covariates included demographics, comorbidities, and history of receiving the HD vaccine.
RESULTS:
The study included 384,180 individuals. The HD vaccine was used in 0.3% of adults 50-64 and 52% of adults 65 and older. Receipt of the HD vaccine decreased the probability of respiratory-related hospitalization by 0.5% points (95% CI -0.8%, -0.2%) compared to the SD vaccine (p=0.002) for adults who received HD because it was approved for their age group. Results were robust to various model specifications and sensitivity tests. CONCLUSIONS:
The HD vaccine reduced the rate of respiratory-related hospitalization compared to the SD vaccine among adults who received HD vaccine because of their age. The results suggest that extending approval to adults ages 50-64 would reduce respiratory-related hospitalizations among those who become newly eligible.
HEOR Studies in Medication Management
On-demand
P70: Evaluation of Cost and Health Utilization Outcomes of Medicare Beneficiaries Enrolled in a Medication Synchronization Program
2:00PM - 2:15PM
Waghmare P
OBJECTIVES:
Medication synchronization (med-sync) aligns patients’ chronic medications to a predetermined routine pickup date at a community pharmacy. An appointment-based model (ABM) med-sync service includes a comprehensive medication review or other clinical appointment at the pharmacy. We compared the cost and healthcare utilization outcomes of Medicare beneficiaries enrolled in an ABM med-sync program to beneficiaries not enrolled in such a program.
METHODS:
This retrospective cohort study included Medicare beneficiaries obtaining medications from pharmacies providing ABM med-sync. Medicare inpatient, outpatient, emergency, and pharmacy claims data from 2014 to 2016 were used to create med-sync (n=13,193) and non-med-sync (n=156,987) cohorts. All patients were followed longitudinally for 12 months before and after a 2015 index/enrollment date. Baseline characteristics including age, gender, race, geographical region, income-based enrollment status, copayment, and urban residence were utilized to create a logistic regression model for propensity score matching. A 1:1 greedy nearest neighbor matching algorithm was adapted for sequentially matching both cohorts. Difference-in-differences (DID) was used to compare mean changes in costs and utilization outcomes between cohorts.
RESULTS:
After matching, 13,193 beneficiaries in each cohort were used for analysis. Mean outpatient, emergency, pharmacy, and total costs increased before and after enrollment for both cohorts. No significant DID in costs were observed between cohorts. Healthcare utilization mean DID were significantly greater in the non-med-sync cohort compared to the med-sync cohort for outpatient visits (DID: 1.17, p<0.0001), emergency department visits (DID: 0.03, p=0.0372) and pharmacy fills (DID: 1.93, p<0.0001). There was no significant DID for inpatient visits between cohorts.
CONCLUSION: Outpatient, emergency, and pharmacy utilizations changes were significantly higher in the non-med-sync cohort compared to the med-sync cohort in the 12-months after enrollment. Lower pharmacy utilization could be due to optimization of therapy during medication reviews of ABM med-sync.
P69: Assessing Pharmacist-Administered Influenza Vaccinations and the Potential Impact of State Policies on Vaccination Rates: A Claims Data Analysis
2:15PM - 2:30PM
Tran J
Objective: Pharmacists have been permitted to immunize in all US states for over a decade, historically under enabling laws and varying requirements such as collaborative practice agreements (CPAs). This study aimed to characterize trends in influenza vaccinations over time and evaluate whether changes in policies related to CPAs impacted annual influenza vaccination rates. Methods: Real-world influenza vaccination trends were observed using IBM MarketScan administrative claims data for adults with employer-sponsored commercial plans for 2010 to 2019. We identified states with relevant policy changes using the National Association of Boards of Pharmacy (NABP) Surveys of Pharmacy Laws. Difference-in-differences analyses evaluated the effect of adding (Nebraska in 2017) or removing (Illinois in 2011, Montana in 2014) a CPA requirement on state annual influenza vaccination rates in the overall outpatient setting (i.e., pharmacy, physician offices, outpatient clinics) and in pharmacy only. These states were compared to similar states without CPA-related changes during this time period. Results : We identified 263,817 pharmacist-administered influenza vaccines in 2010, accounting for 9.1% of outpatient influenza vaccinations identified, and this reached 1.27 million by 2019, or 37.0% of outpatient influenza vaccinations. No significant effects of CPA-related policy changes were observed with vaccinations administered at pharmacies. However, the odds of outpatient influenza vaccination were 4% lower after Nebraska added CPA requirements (p=0.005) and 10% higher after Montana removed CPA requirements (p<0.001). No effects were seen in Illinois. Conclusion : The number and proportion of influenza vaccinations at pharmacies increased between 2010 to 2019 in the MarketScan commercial population. Although we found CPA-related changes impacted outpatient influenza vaccinations in two states, they did not affect vaccinations specifically administered in the pharmacy. Our study suggests CPAs may no longer have an enabling effect as vaccinations at pharmacies have become more common. Further research is needed to better understand their impact on outpatient vaccinations.
P71: Inpatient and Outpatient Medical Costs, Hospitalizations, and Length of Stay Associated with Adherence to Non-cycled Oral Antineoplastics Among Oncology Patients
2:30PM - 2:45PM
Staskon F
OBJECTIVES:
Increasing utilization of non-cycled antineoplastic therapies for a variety of cancers has generated interest in applying a proportion of days covered (PDC) adherence metric, and examining costs and health outcomes associated with adherent PDC (≥80%).
METHODS:
This retrospective study used MarketScan Commercial Claims and Encounters databases from 2017--2019. PDC was calculated for 2018 or 2019 utilizing non-cycled medications (monthly dosages for 31 products) from 15 therapeutic categories. New therapy/diagnosis indications were inferred from the 2017 files. Continuous enrollment was required as were ICD oncology diagnosis codes for those 18—65 years of age. Sample exclusion criteria were deaths, hospice care, inpatient transplant services, and females on fertility therapy. Models examined adherence level (PDC≥80% or not) and 9 covariates, with interaction terms, for significant associations on medical costs (inpatient and outpatient), hospitalizations, and length of days stay (LOS). Finally, for patients meeting 2019 PDC criteria, combined yearly results were modeled for associations.
RESULTS:
Of the 5,694 patients meeting 2018 PDC criteria, 2,034 retained a 2019 PDC, with adherence at 73.2% in 2018 and 67.9% in both years. In 2018, 34.8% were new to therapy/diagnosis and a mean age was 52.4 years. In 2018, 16.2% were hospitalized with LOS=9.5, and 24.1% were hospitalized in both years with LOS=8.8. Models for 2018 found significant effects favoring the adherent cohort with reduced medical costs (-$9,600, p<.0001), odds of hospitalization (0.72, p<.0007), and oncology-related LOS (-1, p<.02). Results for combined years indicated adherence effects for reduced medical costs (-$9,356, p<.0001), odds for hospitalization (0.6, p<.0002), and oncology-related LOS (-2.9, p<.0007). Significant covariate interactions are discussed in the poster.
CONCLUSION: Remaining adherent to oral antineoplastic therapy was associated with lower medical costs, fewer hospitalizations, and a shorter LOS, even across multiple years. These reductions were dependent on comorbidities, new to therapy/diagnosis, or metropolitan area.
P72: Designing a Value-Based Formulary for Kaiser Permanente Washington: A Hypothetical Case Study of Diabetes Mellitus Medications
2:45PM - 3:00PM
Chen Y 1 , Loucks AR2 , Sullivan SD1 , Pearson SD3 , Martin P2 , Yeung K2 1 Comparative Health Outcomes, Policy, and Economics (CHOICE) Institute, University of Washington, Seattle, WA, USA, 2 Kaiser Permanente Washington, Seattle, WA, USA, 3 Institute for Clinical and Economic Review, Boston, MA, USA
OBJECTIVES:
Value-based formulary (VBF) design sets lower copayments for pharmaceuticals with a high assessed value and higher copayments for drugs with a low assessed value. Previous VBFs focused on reducing cost-sharing without matching each drug's copayment to its assessed value. Our aim is to outline a novel VBF design for Kaiser Permanente Washington (KPWA) beneficiaries and estimate possible spending changes.
METHODS:
Formulary design exercise involved consultation with multiple stakeholders including the P&T chair, formulary design manager, clinical, and health economics experts. We designed a 4-tier VBF with exclusions based on incremental cost-effectiveness ratios (ICERs): Tier 1: ICER ≤$50,000, with zero cost-sharing. Tier 2: ICER ≤$100,000, with cost-sharing of $30. Tier 3: ICER ≤$150,000, with cost-sharing of 10% and no ability to apply coupons. Tier 4: ICER ≤$200,000, with cost-sharing of 30% and no coupons. Excluded drugs: ICER >$200,000 or dominated. We assessed value of each drug primarily using cost-effectiveness analysis from available Institute for Clinical and Economic Review evidence reports. We aligned drug prices with KPWA prices for 23 diabetes mellitus (DM) medicines to derive plan-specific incremental cost per QALY gained estimates. Using pharmacy claims (2019-2020), we identified a cohort of 40,150 beneficiaries who were on the included DM medicines. We forecasted future health plan spending and out-of-pocket (OOP) costs with VBF, using published own-price elasticity estimates.
RESULTS:
The average age of the cohort was 55, and 51% of them were female. When compared to the traditional formulary, the VBF is estimated to reduce total annual health plan spending by 10%, saving $87 in annual spending per member and $91 in annual OOP spending per member. Sensitivity analyses using various price elasticity values showed declines in all spending outcomes.
CONCLUSIONS:
Designing a VBF in a U.S. employer-based health plan has the potential to save money for both the plan’s and the patient's OOP expenses.
Cost Effectiveness Analysis in Oncology Studies
On-demand
Moderator
Koen Degeling, PhD, MSc, BSc
Lumen Value & Access – a Healthcare Consultancy Group Company, New York, NY, USA
Dr Koen Degeling is a Research Scientist, Health Economic Modelling & Advanced Analytics at Lumen Value & Access, a Healthcare Consultancy Group company. He was trained as an Industrial Engineer specializing in Healthcare Technology and Management and holds a PhD in Advanced Health Economic Modelling from the University of Twente in the Netherlands. Prior to joining Lumen Value & Access, Koen worked on real-world data-driven health economic and health services research projects at the Cancer Health Services Research department of the University of Melbourne in Australia, where he continues to be involved as an honorary fellow. He is an active ISPOR member and currently serves on the Editorial Advisory Board for Values & Outcomes Spotlight and ISPOR New Professionals Steering Committee, is involved in several short courses and workshops, and has served as global chair, committee co-chair, and chapter president within the ISPOR Student Network.
P64: Cost-Utility Analysis Comparing Mastectomy and Lumpectomy in Early Stage Breast Cancer Surgeries in Sweden - a Life-Time Approach Using Markov Modelling
2:00PM - 2:15PM
Pham PD
Background: Breast cancer is the most common cancer type among Swedish women. Although the treatments of breast cancer has developed dramatically over years, there lacks health economics assessments on these treatments, especially cost-utility analysis. This study performed a cost-utility analysis comparing three available surgical procedures, including mastectomy, lumpectomy without irradiation and lumpectomy with irradiation in Swedish context. Methods: A 6-state Markov model with 20-year time horizon was used to investigate the cost-utility of three alternatives. Transition probabilities parameters were obtained based on the best available evidence in Sweden and similar contexts. Both healthcare and societal perspectives were considered in cost estimation. An individual dataset from Federation of Swedish county councils and municipalities (SKL) was used to calculate costs. Quality-adjusted life years (QALYs) were used to calculate utility, based on a study in a similar context of Norway due to the lack of utility evidence in Sweden. Deterministic univariate and multivariate sensitivity analysis were performed to handle uncertainties. Results: Lumpectomy with irradiation dominated two other options and was the most cost-effective treatment option for early stage breast cancer in Sweden. Moving from mastectomy to lumpectomy without irradiation, the incremental cost-effective ratio (ICER) was 153,350 SEK in healthcare perspective and 32,017 SEK in societal perspective, per QALY gained. This ICER should still be considered as cost-effective in Swedish context. Conclusions: This study revealed the cost and utilities in life-time approach of three surgical procedures in early stage breast cancer treatments in Sweden, namely mastectomy, lumpectomy without irradiation and lumpectomy with irradiation. The results showed that lumpectomy with irradiation appeared to be the most cost-effective option during 20-year follow-up in both healthcare and societal perspectives. Future studies should be expanded to estimate the more reliable parameters in Sweden to build up the consistency of our findings.
P61: Cost Utility Analysis of Circulating Tumour DNA Guided Adjuvant Chemotherapy in Stage II Colon Cancer
2:15PM - 2:30PM
To YH 1 , Degeling K2 , Kosmider S3 , Wong R1 , Lee M1 , Dunn C1 , Gard G1 , Jalali A1 , Wong V1 , IJzerman M2 , Gibbs P1 , Tie J1 1 Walter and Eliza Hall Institute of Medical Research, Parkville, VIC, Australia, 2 The University of Melbourne, Melbourne, VIC, Australia, 3 Western Health, Footscray, VIC, Australia
Background: There is currently potential overuse of adjuvant chemotherapy (AC) in patients with stage II colon cancer (CC) given the uncertain survival benefit in unselected patients. A circulating tumour DNA (ctDNA) approach has the ability to improve patient selection for AC, defining patients who may benefit from treatment (ctDNA positive) and those who will not (ctDNA negative). This study aimed to estimate the health and economic impact of ctDNA-guided prescription of AC for stage II CC. Methods: A cost-utility analysis was performed comparing ctDNA-guided AC prescription for stage II CC to standard of care (SOC), where 22.6% of SOC patients received AC, all ctDNA-positive patients (8.7%) received AC and all ctDNA-negative patients (91.3%) did not. A third preference-sensitive ctDNA strategy was included where 6.8% of ctDNA-negative patients would receive AC to reflect potential non-compliance. A state-transition model was populated utilising a landmark cohort study investigating the prognostic value of ctDNA and clinical registries. The analysis employed an Australian payer perspective and lifetime horizon. Extensive scenario and probabilistic analyses quantified model uncertainty. Results: Compared to SOC, the ctDNA and preference-sensitive ctDNA strategies increased quality-adjusted life years (QALYs) by 0.20 (95% confidence interval -0.40 to 0.81) and 0.19 (-0.40 to 0.78), and resulted in incremental costs of AUD - 4,215 (-17,651 to 9,216) and AUD - 2,450 (-15,472 to 10,570), respectively. ctDNA remained cost-effective at a willingness-to-pay of AUD 20,000 per QALY gained throughout most scenario analyses in which the proportion of ctDNA-positive patients cured by AC and compliance to a ctDNA negative test results were decreased. Conclusions : ctDNA-guided AC is a potentially cost-effective strategy to improve patient selection for adjuvant chemotherapy in resected stage II colon cancer. Expanding the analysis with results from ongoing randomised clinical studies will be important to reduce uncertainty in the model.
P62: Cost-Effectiveness of Atezolizumab Plus Cobimetinib and Vemurafenib in the Treatment of BRAF V600 Mutation-Positive Metastatic Melanoma
2:45PM - 3:00PM
Cai C 1 , Yunusa I2 , Tarhini A3 1 University of South Carolina, Columbia , SC, USA, 2 University of South Carolina, Columbia, SC, USA, 3 University of South Florida Morsani College of Medicine, Tampa, FL, USA
OBJECTIVES:
To evaluate the cost-effectiveness of atezolizumab and vemurafenib plus cobimetinib vs. vemurafenib plus cobimetinib alone in patients with newly diagnosed unresectable BRAF V600 mutated metastatic melanoma from the US healthcare perspective.
METHODS:
This economic evaluation study used a three-state partitioned survival model to assess the cost-effectiveness of the triplet combination of PD-L1 inhibitor atezolizumab plus BRAF inhibitor vemurafenib plus MEK inhibitor cobimetinib. The observed overall survival curves and progression free survival curves were digitized from the IMspire150 trial (January 2017 – April 2018) and the long-term survivals (over a life-time horizon) beyond the end of the study were extrapolated using seven different survival models. Life-years (LYs) gained and quality-adjusted life-years (QALYs), costs, and incremental cost-effectiveness ratio (ICER) expressed as cost per LYs and per QALYs saved were estimated. The cost and health preference data were collected from a literature review.
RESULTS:
Adding atezolizumab to vemurafenib and cobimetinib provided an additional 3.267 QALYs compared to doublet regimen of vemurafenib plus cobimetinib, at an ICER of $271,669/QALY, which is not considered cost-effective at the willingness-to-pay (WTP) threshold of $150,000/QALY. However, the scenario analyses found that atezolizumab combined with vemurafenib plus cobimetinib could be cost-effective at a 20-year (ICER, $121,432/QALY) and 30-years ($98,092/QALY) time horizons when both strategies were stopped after two years of treatments, and over a lifetime horizon ($122,220/QALY) when immunother
