| Category Reference | |||
|---|---|---|---|
| BEC | Behavioral Economics | ESP | Applied Health Economics, Services, and Policy Research |
| DEC | Decision Psychology and Shared Decision Making | MET | Quantitative Methods and Theoretical Developments |
* Candidate for the Lee B. Lusted Student Prize Competition
Purpose: This analysis assessed reduction of cardiovascular (CV) hospitalizations in the first 12 months of the ATHENA trial and the associated cost savings in the US.
Method: The ATHENA trial randomized atrial fibrillation/flutter (AF/AFL) patients (mean age 71.6 years) with ≥1 other CV risk factor to dronedarone (n=2,301) or placebo (n=2,327), plus standard care. In this cost analysis, hospitalization costs, derived from claims data for a US cohort of ‘ATHENA-like’ AF/AFL patients with Medicare supplemental insurance (n=10,200), were applied to hospitalization events occurring during the first 12 months of the ATHENA trial. Cost inputs (2008 values) were (i) weighted mean CV hospitalization costs, categorised according to admission cause, and (ii) Diagnosis Related Groups costs of hospitalizations for adverse events (AEs) in ATHENA. Cost variations were assessed using Monte Carlo sensitivity analysis.
Result: During the first 12 months of ATHENA, overall CV hospitalizations fell by 29% with dronedarone (33.36 vs. 47.19 events/100 patients, dronedarone vs. placebo). Based on the observed hospitalizations and derived costs, the overall cost savings with dronedarone were estimated at (mean ± SD) $1,328 ± 176 per patient (Table). The estimated savings in CV hospitalization costs (mean $1,341 per patient) heavily outweighed the added estimated AE hospitalization costs (mean $12 per patient). Sensitivity analysis showed the cost offset ranged between $594−$2,124 over 10,000 cycles of Monte Carlo simulation.
Conclusion: Dronedarone offers early cost benefits in AF/AFL, producing estimated mean hospital-related cost savings of $1328 per patient within the first 12 months of treatment in the ATHENA population.
| Hospitalization cause | Hospitalizations/100 patients1 | Default cost/hospitalization | Hospitalization cost saving/patient | |
Placebo + standard care | Dronedarone + standard care | |||
| Myocardial infarction/unstable angina | 3.14 | 1.69 | $17,360 | $250 |
| Cardiac arrhythmia & conduction disorders | 26.69 | 15.38 | $8,601 | $972 |
| Cardiovascular surgery | 2.58 | 2.52 | $21,233 | $12 |
| Worsening heart failure, pulmonary edema/cardiac dyspnea | 5.54 | 4.26 | $9,945 | $128 |
| Implantation of cardiac device | 2.15 | 1.69 | $18,272 | $83 |
| Transient ischemic attack/stroke | 1.59 | 1.30 | $9,006 | $26 |
| Other cardiovascular2 | 5.50 | 6.52 | $12,807 | −$130 |
| Total cardiovascular hospitalizations | 47.19 | 33.36 | − | $1,341 |
| Adverse events3 | 0.21 | 0.48 | $4,681 | −$12 |
| All hospitalization events | 47.40 | 33.84 | − | $1,328 |
| 1. Numbers rounded from 3 decimal places; 2. Cardiac transplantation, cardiovascular infection, pulmonary embolism/deep vein thrombosis, non-fatal cardiac arrest, major bleeding, atherosclerosis-related, syncope, blood pressure-related, stable angina pectoris or atypical chest pain; 3. Non-cardiovascular and treatment-related. | ||||
Purpose: To evaluate the relative cost-effectiveness of six non-invasive cardiac imaging tests in stable outpatients with suspected coronary artery disease (CAD) including: 1) stress echocardiography (Echo), 2) stress Echo with contrast agent (Echo+Contrast), 3) stress Echo with contrast agent used only if initial results are not interpretable (Echo>Contrast), 4) 64-slice computer tomography angiography (CTA), 5) cardiac magnetic resonance imaging (CMRI), and 6) stress single-photon emission computed tomography (SPECT).
Method: A decision-analytic Markov model was constructed to simulate the costs and consequences of diagnostic testing in a hypothetical cohort of patients presenting with chest pain in an ambulatory setting, with an intermediate risk of CAD after clinical evaluation. Resource use and costs were derived from Ontario data sources, including the Ontario Health Insurance Plan Schedule of Benefits and Ontario Case Costing Initiative. Estimates of diagnostic test characteristics (sensitivity, specificity) were identified by systematic review and statistically pooled using a bivariate regression approach. Data sources for other model parameters were published data identified by systematic review. The analysis took the perspective the Ontario public health care system and was conducted over a lifetime time horizon. Costs were expressed in 2008-2009 Canadian prices. The primary outcome was quality-adjusted life years (QALYs). Costs and QALYs were discounted at an annual rate of 5%. Cost-effectiveness was evaluated using two conventional willingness-to-pay thresholds: $50,000 per QALY and $100,000 per QALY. Uncertainty around the results was explored with probabilistic sensitivity analysis with 10,000 simulations.
Result: Echo>Contrast was the least expensive test (expected lifetime costs of $21,536) with expected lifetime QALYs of 10.02. CTA was more slightly more expensive ($21,618) and effective (expected lifetime QALYs of 10.05) than Echo>Contrast; thus CTA was cost-effective relative to Echo>Contrast with an incremental cost-effectiveness ratio of $2,958 per QALY. CTA dominated Echo, Echo+Contrast, CMRI, and SPECT, and extendedly dominated Echo. The probability that CTA was cost-effective at a willingness-to-pay of $50,000 per QALY and $100,000 per QALY was 0.929 and 0.934, respectively. Varying individual parameter values across plausible ranges in a series of 1-way sensitivity analyses did not change the finding that CTA was cost-effectiveness.
Conclusion: Sixty-four slice computer tomography angiography appears to be a cost-effective non-invasive cardiac imaging option for intermediate risk patients with suspected CAD in an ambulatory setting.
Purpose: To evaluate the cost-effectiveness of using a web-based self-assessment tool to detect cases of Hypercholesterolaemia and subsequently treat with statin therapy.
Methods: Data was collected from 25,364 users of the HeartAware risk factor self-assessment tool administered through a nationwide network of 127 hospitals and clinics. The web-based tool asked participants to report several risks factors for heart disease including: low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, systolic and diastolic blood pressure, diabetes and smoking status, medical history, and family history of disease. Responses enabled the calculation of heart disease risk. Participants identified as high risk were eligible for selection by hospitals or clinics for no-cost clinical screening of the same risk factors. Participants with no history of heart disease and those with both self-reported and clinically measured risk factors were included in the analytic sample. A decision-tree determined if sample members would qualify for statin therapy based on inter-rater agreement of self-reported and clinical measures, prior usage of cholesterol lowering medications, and clinical guidelines for statin therapy established by the Adult Treatment Panel III. HeartAware screening costs were measured along with costs of follow-up testing and treatment for those identified for statin therapy. Cost avoidance associated with reduced risk of heart disease was calculated. Life years gained (LYG) as a result of statin therapy was used to calculate cost per LYG. Sensitivity analysis was also performed for scenarios of low statin adherence and enhanced screening methodologies. Findings were compared to prior studies of cost-effectiveness for opportunistic and universal familial hypercholesterolaemia screening as well as accepted thresholds for cost per LYG.
Results: The analytic sample contained 5,884 participants, with 225 eligible for statin therapy. HeartAware resulted in a cost per LYG of $16,665. Sensitivity analysis for 50% statin adherence resulted in a cost per LYG of $16,428, while enhanced screening methodologies indicated a cost per LYG between $7,620 and $14,607.
Conclusion: Prior studies of opportunistic and universal screening indicate a cost per LYG of $20,313 and $23,413 respectively. Accepted thresholds for cost per LYG are between $35,000 and $60,000. HeartAware is more cost effective than these established screening methodologies, and also favorable relative to accepted willingness to pay thresholds. As such, it should be considered a viable alternative screening method for heart disease.
Purpose: The American Diabetes Association (ADA) developed a questionnaire-based scoring system to screen for undiagnosed diabetes, wherein persons with a score of ≥10 are considered at high risk and recommended for further screening. We assessed the cost-effectiveness of the recommended cutoff score of 10 and other alternative cut-points of the scoring system.
Method: We used a validated simulation model to estimate the lifetime cost-effectiveness associated with a 1 point increment in risk score from 5 to 15. We used data from the National Health and Nutritional Examination Survey (2007) to estimate the prevalence and characteristics of the undiagnosed diabetes population, and sensitivities and specificities of each alternative cutoff. Persons who screened positive were assumed to receive a follow-up diagnostic test and intensive glycemic management if confirmed to have diabetes. Outcomes were measured by expected life-years, quality-adjusted life-years (QALYs), and medical costs. Incremental Cost-Effectiveness Ratio (ICER) of one cutpoint was measured by the incremental cost per QALY gained comparing with its next higher cutpoint. The analysis was conducted from a societal perspective.
Results: The proportion of undiagnosed diabetes detected, health benefit, cost and ICER by alterative cutoff score are presented in the table. A lower cutpoint resulted in a larger proportion of the undiagnosed diabetes detected and greater health benefits, but also in higher medical costs and higher ICER. The cutpoints in the range of 11 to 15 have ICERs lower than $50,000 and the cutpoint of 10 was associated with an ICER of $55,000/QALY.
Cutpoint | Undiagnosed Diabetes Cases Detected,% | Life-year Gained† | QALY Gained† | Incremental Cost | Cost per QALY |
5 | 77.8 | 0.009 | 0.008 | 1,420 | 176,000 |
6 | 70.5 | 0.010 | 0.012 | 1,088 | 94,000 |
7 | 64.6 | 0.012 | 0.013 | 911 | 72,000 |
8 | 64.3 | 0.015 | 0.014 | 809 | 60,000 |
9 | 57.1 | 0.016 | 0.013 | 794 | 59,000 |
10 | 31.4 | 0.017 | 0.014 | 787 | 55,000 |
11 | 21.1 | 0.018 | 0.016 | 760 | 48,000 |
12 | 29.7 | 0.018 | 0.016 | 718 | 44,000 |
13 | 18.8 | 0.021 | 0.016 | 677 | 43,000 |
14 | 7.3 | 0.029 | 0.015 | 620 | 42,000 |
15 | 1.1 | - * | - | - |
|
Conclusions: There was a tradeoff between the total health benefit and economic efficiency by lowering the cutoff score. If $50,000/QALY were used as the acceptable willingness-to-pay threshold, a cutoff score of ≥11 should be selected.
Purpose: For the initial assessment of patients with stable chest pain syndrome, coronary CT angiography (CTA) has evolved as an alternative to functional testing (FT) for the detection of obstructive coronary artery disease (CAD). However, uncertainty remains about its overall diagnostic value including the identification of non-obstructive CAD given the current absence of treatment. Our objective was to evaluate clinical outcomes, costs, and cost-effectiveness of different anatomic and functional test modalities in the light of potential treatments for non-obstructive CAD.
Methods: Design: Cost-effectiveness analysis using a microsimulation model to simulate incidence and progression of CAD (non-obstructive and obstructive) as a function of patient age, gender and cardiac risk profile. Mortality risk depended on patient's demographics, CVD and treatment status. Potential treatment effect on non-obstructive CAD was based on decreasing the Framingham risk score (hypothetical life-style modifications) and secondary prevention studies. Target population: Patients with chronic chest pain syndrome. Time horizon: Diagnostic phase, lifetime. Discount rate: 3%. Perspective: Societal. Interventions: (1) preventive treatment (SOC) to (2) CTA (CTA), (3) stress-EKG/stress-echo/SPECT (in 20%, 50%, and 30%) (FT), (4) FT followed by CTA if FT positive or indeterminate (FT-CTA), (5) CT followed by FT if CTA positive or indeterminate (CTA-FT). Outcomes: Diagnostic results, discounted quality-adjusted life expectancy (QALE) and lifetime costs, incremental cost-effectiveness ratio (ICER).
Results: In our base case population (males, 50 years, low risk for CAD) the prevalence of CAD was estimated at 53% (13% obstructive). FT correctly identified 13% (10%) at $469/patient; CTA 44% (12%), CTA-FT 49% (9%), FT-CTA 17% (9%) at $599, $663, and $605 per patient, respectively. The model predicted an average remaining life expectancy of 22.01 quality-adjusted life years (QALY) for SOC and 22.27, 22.33, 22.20 and 22.21 QALYs for FT, CTA, CTA-FT, and FT-CTA, respectively. This resulted in an ICER of $13,800/QALY for FT compared to SOC, and of $20,000/QALY for CTA vs. FT; CTA-FT and FT-CTA were both dominated. When applying potential treatment benefit to patients with non-obstructive CAD, CTA dominated most other strategies across a broad range of CAD prevalences (figure).
Conclusion: Preliminary analyses indicate that CTA is cost-effective compared to functional testing as an initial evaluation method for patients with chronic chest pain. These results are independent of treatment effect on non-obstructive CAD.
Purpose: Delivery of preventive services falls short of guideline recommendations. We evaluate the multilevel factors associated with missed opportunities to deliver evidence-based preventive services during periodic health examinations (PHE).
Method: Physician subjects (N=64) were general internal medicine and family physicians practicing in 2007-2009 with an integrated delivery system in southeast Michigan. Patient subjects (N=484) were insured, aged 50-80 years, and due for colorectal cancer screening. Office visit audio-recordings were used to ascertain physician recommendation for/delivery of 19 services recommended by the US Preventive Services Task Force and Advisory Committee on Immunization Practices. A patient survey and claims data were used to determine patient service eligibility/due status. Alternating logistic regression with individual service delivery as the outcome evaluated patient, physician, visit and contextual factors associated with missed opportunities. Models nested services within patients and patients within physicians as well as controlled for service type.
Result: Among N=2662 services for which patients were due, 46% were not delivered. Services with highest rates of missed opportunities included aspirin counseling (82%), vision screening (81%) and influenza vaccination (80%). Those with lowest rates included colorectal cancer (7%) hypertension (8%) and breast cancer (10%) screening. Regression results indicated the likelihood of a missed opportunity increased with patient age (OR=1.03; 95% CI= 1.01-1.05) and each additional concern the patient raised (1.24; 1.09-1.40), decreased with increasing patient body mass index (0.98; 0.97-1.00) and each additional minute after scheduled appointment time the physician first presented (0.99; 0.98-1.00), and was greater if the physician used the electronic medical record (EMR) in the exam room (1.40; 1.06-1.86), was of a different gender than the patient (1.37; 1.05-1.79), and had seen the patient in the past 12 months (1.40; 1.06-1.86).
Conclusion: Almost half of recommended preventive services are not delivered to patients during PHEs. A combination of patient, physician, visit and contextual factors are associated with missed opportunities. While physicians appear not to skip delivery of due services when running late, delivery can be compromised when patients raise competing demands, and when the EMR is used in the exam room. The public health and economic impact of missed opportunities to deliver preventive services is profound and warrants additional studies to understand the complex interplay of factors that support and compromise preventive service delivery.