G APPLICATION OF COST EFFECTIVENESS ANALYSES: FROM PREVENTION TO TREATMENT

Tuesday, October 22, 2013: 10:30 AM - 12:00 PM
Key Ballroom 5-6 (Hilton Baltimore)
Category Reference for Presentations
AHEApplied Health Economics DECDecision Psychology and Shared Decision Making
HSPHealth Services, and Policy Research METQuantitative Methods and Theoretical Developments

* Candidate for the Lee B. Lusted Student Prize Competition

Session Chairs:
Anirban Basu, PhD and Ankur Pandya, PhD
10:30 AM
G-1
(AHE)
Constantinos I. Michaelidis, BA1, Richard K. Zimmerman, MD, MPH1, Mary Patricia Nowalk, PhD1, Michael J. Fine, MD, MSc1 and Kenneth J. Smith, MD, MS2, (1)University of Pittsburgh School of Medicine, Pittsburgh, PA, (2)University of Pittsburgh, Pittsburgh, PA
Purpose: To evaluate the cost-effectiveness of procalcitonin-guided antibiotic therapy in outpatient management of ARTIs in adults.

Method: We developed a cost-effectiveness model based on data from two randomized controlled trials and modeled the different cohorts in two separate analyses. Cohort one comprised all adults with ARTIs presenting in the outpatient setting. Cohort two comprised all adults with ARTIs presenting in the outpatient setting and judged by their providers to require antibiotics after clinical evaluation. Both analyses assumed a societal perspective and a clinic visit time horizon. We compared two strategies: usual care and procalcitonin-guided antibiotic therapy. The primary differences between strategies were the procalcitonin testing costs and fewer antibiotic prescriptions in the procalcitonin strategy. The primary outcome was cost per antibiotic avoided. Because there is not an accepted societal willingness-to-pay threshold for this outcome, we developed an estimate for this threshold based on the societal cost of antibiotic resistance per outpatient antibiotic prescription. For each cohort, we performed base case cost-effectiveness analyses and examined model robustness to parameter variation in sensitivity analyses. 

Result: We estimated the willingness-to-pay threshold as $43 (range $0–$333) per antibiotic avoided, reflecting the estimated societal cost of antibiotic resistance per outpatient antibiotic prescription. In the cohort including all adults with ARTIs, the procalcitonin strategy cost more ($49 vs. $15 per patient) and reduced the number of antibiotic prescriptions (0.14 vs. 0.37 antibiotic prescriptions per patient) compared to usual care, resulting in an incremental cost-effectiveness ratio (ICER) of $149 per antibiotic prescription avoided. In a probabilistic sensitivity analysis, the likelihood of procalcitonin being favored over usual care was 2.8%. In the cohort including adult ARTIs judged by their providers to require antibiotics, the procalcitonin strategy cost more ($51 vs. $29 per patient) and reduced the number of antibiotic prescriptions (0.25 vs. 0.97 antibiotic prescriptions per patient) compared to usual care, resulting in an ICER of $31 per antibiotic prescription avoided. In a probabilistic sensitivity analysis, the likelihood of procalcitonin being favored over usual care was 58.4%. 

Conclusion: Procalcitonin-guided antibiotic therapy for outpatient management of ARTIs in adults is cost-effective compared to usual care when the societal costs of antibiotic resistance are considered and procalcitonin testing is limited to adults with ARTIs judged by their providers to require antibiotics after clinical evaluation.

10:45 AM
G-2
(AHE)
Hassan M.K. Ghomrawi, PhD1, Ashley A. Eggman, MS1, Sophia Paul2 and Andrew Pearle, MD3, (1)Weill Cornell Medical College, New York, NY, (2)Hospital for Special Surgery, New York, NY, (3)Hopsital for Special Surgery, New York, NY
Purpose: Unicompartmental knee arthroplasty (UKA) is an alternative to total knee arthroplasty (TKA) in patients with unicompartmental knee osteoarthritis. UKA has higher revision rates 5-10 years postoperatively, yet requires little rehabilitation, has fewer complications, and may offer patients higher function. Two prior cost-effectiveness (CE) analyses limited to older patients (age 65+) found little benefit to UKA. With rising demand expected among patients <65 who would favor better function and quicker return to work (50% of 3.4 million in 2030), we evaluated CE of UKA vs. TKA in younger and older patients.

Method: We developed a Markov, state-transition model to determine the CE of UKA vs. TKA in hypothetical younger (age 45) and older (age 65) patients. Patients transition to either a full- or limited-function state postoperatively based on their Western Ontario and McMaster Universities Osteoarthritis Index score. The limited-function state is associated with lower quality-of-life (QOL) and increased costs compared to full-function. Patients in either state may experience an implant failure. We assumed that limited-function state patients fail at a higher rate than full-function state and UKA patients only have home health rehabilitation while TKA patients could also have either inpatient rehabilitation or skilled nursing facilities options. Failure rates were calculated from 20-year follow-up data in the Swedish national registry. Procedure and complication costs were based on DRG and ICD-9 codes. Inpatient rehabilitation, outpatient healthcare utilization and return to work costs were derived from the literature. QOL data and transition probability to limited-function state were also from the literature.  We report costs, quality-adjusted life years (QALYs), and ICERs over the lifetime of the patient from a societal perspective in US dollars (discounted at 3% annually).

Result: UKA dominated TKA in both younger and older patients in the base case  In the younger patients only, TKA has an ICER <$100,000/QALY when the annual revision rate for UKA increased from 1.0% in the base case to 2.45% beginning at age 66 .  Results were not sensitive to lowering TKA rehab costs. Increasing UKA’s transition probability to the limited-function state to equal that of TKA produced an ICER >$100,000/QALY.

Conclusion: Unlike previous studies, we found UKA had substantial economic value in older and younger patients. Our results indicate that UKA should be recommended for eligible patients regardless of age.

11:00 AM
G-3
(AHE)
Sze-chuan Suen, MS1, Kimberly M. Brayton, MD, JD1, Vishal G. Patel, MD2, Douglas K. Owens, MD, MS3 and Jeremy D. Goldhaber-Fiebert, PhD1, (1)Stanford University, Stanford, CA, (2)University of Texas Southwestern Medical Center, Dallas, TX, (3)Veterans Affairs Palo Alto Health Care System and Stanford University, Stanford, CA

Purpose: For patients undergoing elective percutaneous coronary intervention (PCI), shortening hospitalization has important cost implications and safety considerations. In studies of same-day discharge (SDD) compared to routine overnight observation in highly selected PCI populations, the overall safety profile is uncertain as some outcome were better with SDD and others were not. We assessed the cost-effectiveness of SDD for all eligible PCI patients.

Method: We developed a decision-analytic Markov model of patients post PCI to compare same-day discharge to routine overnight observation. Patients were followed from their index PCI over their remaining lifetime, including the possibility of repeat coronary procedures. We derived inputs from registries, randomized trials, and meta-analyses. The probabilities for SDD adverse outcomes were death (0.6%), major bleed (MB) (0.9%), myocardial infarction (MI) (2.6%), and target vessel revascularization (1.1%); overnight observation outcomes differed only for MI, at 0.6%. We adjusted mortality rates for previous adverse procedure outcomes. We estimated costs using micro-costing methods (2012 USD) and measured outcomes in QALYs, both considered from a societal perspective and discounted at 3% annually.

Result: Compared to current practice (routine overnight observation), same-day discharge costs $1,639 less but delivers 0.0028 fewer QALYs ($585,000 saved per QALY lost). This finding is sensitive to adverse outcome probabilities. SDD costs less and provides more QALYs for values of safety measures within the 95% confidence intervals when the probabilities of adverse outcomes for SDD patients are lower (<0.57% for death, <0.60% for MB, or <2.32% for MI) or when the probabilities of adverse outcomes for overnight observation are higher (>0.63% for death, >0.90% for major bleeding events, or >2.89% for MI). In contrast, SDD costs more and is less effective if risks of SDD adverse events are substantially higher (>6.46% for MI) but still within the 95% confidence interval. Probabilistic sensitivity analyses reveal that SDD is frequently cost-effective even at higher willingness-to-pay thresholds (see figure).

Conclusion: While same-day discharge can likely reduce costs in many health systems, its health benefits relative to overnight observation appear close and uncertain. The attractiveness of same-day discharge may depend on the ability of particular systems to monitor and quickly respond to adverse events out of the hospital.  Research in this area should prioritize establishing the safety profile of same-day discharge in representative patient populations.

11:15 AM
G-5
(AHE)
Carolina Barbosa, PhD, MSc, PharmD1, Emily A. Smith2, Thomas J. Hoerger3, Nancy Fenlon2, Sarah F. Schillie2, Christina Bradley3 and Trudy V. Murphy2, (1)RTI International, Chicago, IL, (2)Centers for Disease Control and Prevention, Atlanta, GA, (3)RTI International, Research Triangle Park, NC
Purpose: To analyze the cost-effectiveness of the national perinatal hepatitis B prevention program (PHBPP) over the lifetime of the 2009 U.S. birth cohort and compare the costs and outcomes of the strategy represented by the program to 2 alternatives: Strategy 1: no prevention and Strategy 2: current routine vaccination schedule. The PHBPP is targeted at the identification and management of infants born to hepatitis B surface antigen (HBsAg)-positive mothers.

Method: A decision analytic tree and a long-term Markov model represented the risk of perinatal and early childhood (“postnatal” ages 1 to 5 years-old) infections under different vaccination alternatives, and the long-term health and economic consequences of hepatitis B infection. Outcome measures were the number of perinatal infections and postnatal infections from infants born to HBsAg-positive women, quality-adjusted life years (QALYs), lifetime costs, and incremental cost per QALY gained. The health outcomes and total costs of each strategy were compared in the form of incremental cost effectiveness ratios (ICERs). Costs were evaluated from the health care system perspective and expressed in U.S. dollars at a 2010 price base. Probabilistic sensitivity analysis was conducted to incorporate uncertainty in model parameters. Uncertainty in the model was described using cost-effectiveness acceptability curves. One-way sensitivity analyses were conducted to estimate the impact of changing each key parameter individually. A series of multi-way sensitivity analyses changed several variables simultaneously to create hypothetical scenarios.

Result: In all analyses, the PHBPP increased QALYs and led to higher reductions in the number of perinatal and postnatal infections than the alternative strategies. Compared with Strategy 2, the PHBPP was associated with 2,351 fewer total infections (1,485 perinatal and 866 postnatal), 2,304 less QALYs lost, and an ICER of $2,602 per QALY. When the PHBPP was compared with a hypothetical scenario of no prevention, it was associated with 9,159 fewer total infections (5,902 perinatal and 3,257 postnatal), 8,772 fewer QALYs lost, and an ICER of $1,785 per QALY. In sensitivity analyses, the cost-effectiveness ratios were robust to variations in model inputs, including instances where the program was both more effective and cost saving.

Conclusion: This study indicated that the current PHBPP represents a cost-effective use of resources. Ensuring the program reaches all pregnant women could prevent further hepatitis B-related morbidity and mortality.

11:30 AM
G-6
(AHE)
Syed Mohiuddin1, Katherine Payne1, Iain Bruce2, Peter Callery1, Nicola Harman1, Bill Shaw1, Tri Tat1, Stephanie Tierney1, Paula Williamson3 and Kevin O'Brien1, (1)University of Manchester, Manchester, United Kingdom, (2)Central Manchester University Hospital, Manchester, United Kingdom, (3)University of Liverpool, Liverpool, United Kingdom
Purpose: There is a paucity of evidence to guide the treatment of Otitis Media with Effusion (OME), which is a common problem causing hearing impairment in children with Cleft Palate (CP). The surgical insertion of grommets is being used as a means of correcting hearing impairment and preventing complications of untreated OME, but there remains active discussion in the medical community about whether the benefits of grommets outweigh the risks. This study aimed to use value of information analysis within a mixed-methods feasibility study to inform future research requirements and priorities for the treatment of children with CP and OME.

   Method: A model-based early economic analysis compared the surgical insertion of grommets with two non-surgical alternatives (hearing-aids; do-nothing) in the management of persistent bilateral OME in children with CP. The model assumed a 2-year time horizon and UK NHS perspective. Outcomes were valued using Quality-Adjusted Life-Years (QALYs) estimated by linking published utility values with hearing gain (measured in decibels). Multiple data sources were used including: systematic reviews of the effectiveness, resource use and utility literature and supplemented with published expert opinion. Probabilistic sensitivity analysis was used to quantify parameter uncertainty. The results of an expected value of perfect information analysis was combined with the findings from the qualitative views of children and parents about willingness to take part in a potential trial and important outcomes of OME management.

   Result: The probabilistic incremental cost-effectiveness ratio for grommets insertion versus hearing-aids was £14,333 per QALY (95% CI: £8,311 to £20,355) and versus do-nothing was £18,036 per QALY (-£3,790 to £39,862). A key source of uncertainty was the assumed link with changes in utility values and hearing gain. The expected value of perfect information analysis indicated that the value of future research work in this area is worthwhile. Children and parents had clear ideas about outcomes that were important to them, which extended beyond hearing gain. However, more than half of the parents (63%) were negative or unsure about their child taking part in a future trial.

 

   Conclusion: This study has shown that value of information analysis has an important role in informing the need for future research but should be part of a wider programme of work to understand the feasibility and design of future trials.