OPTN Committee Data Request: Financial Analyses of Authors: Mark Schnitzler, PhD 1 Proximity Points Sommer Gentry, PhD 2 Presenter: Joshua Pyke, PhD 3 Scenarios 1. Saint Louis University; SRTR Senior Staff 2. United States Naval Academy; SRTR Senior Staff 3. Scientific Registry of Transplant Recipients 1
Committee Request • Committee requested that SRTR prepare analysis of the financial impact of each of the 28 models previously described. • Analysis should incorporate transplant center expenditures. • Analysis should include assessment of: Costs related to the transplant episode One year post-transplant costs Transportation costs 2
Cost Overview • Approach: combine cost-of-care models with simulated transplants to estimate system costs for 28 scenarios • Four cost models Pretransplant cost (Medicare data) Transplant episode + 1 year follow up cost (UHC data) Posttransplant cost (Medicare data) Transport cost (transport model) • Focused on underlying system costs, not charges or accounting 3
Cost Data: Medicare • Combined payments for all services from Medicare Parts A and B (inpatient, outpatient, home health, and hospice) • Payments for all services were summed and aggregated based on # of months each candidate spent on the waiting list and number of years (first 3 years only) posttransplant • Pretransplant costs included daily estimates of spending adjusted for lab MELD/PELD score and other patient characteristics for the duration of listing • Posttransplant cost per patient was estimated for 2 time periods: Early: 3 days before transplant to 1 year Late: year 1 to year 3 posttransplant • Posttransplant costs censored at time of retransplant to capture cost associated with first transplant • Medicare payments were minimally adjusted for wage/price differences by region, conforming with standard diagnosis related groups and the evaluation/management code-based fee schedule • Did not include organ acquisition cost under Medicare claims, since these are paid via institutional cost report 4
Cost Data: UHC • Dataset created by merging University HealthSystem Consortium (UHC) hospital cost accounting data with OPTN data for liver transplants performed between 2002 and 2013. • Transplant records were linked using date of transplant, age, and gender (no unique identifiers available for this dataset). • UHC data include patient-level cost data from administrative billing claims submissions, adjusted to costs using the transplant hospital’s Medicare cost-to-charge ratio and adjusted for geographic differential in wages. 5
Analytic Approach • Multivariate linear regression was used to estimate monthly (pretransplant) and total person-level (transplant episode and posttransplant) spending. Models were adjusted for recipient and donor factors relevant to • waitlist and transplant analysis. Recipient adjustment factors: Age, race, gender, blood group, diagnosis category, HCC exception status, diabetes, cerebrovascular disease, working for income, daily biologic MELD or biologic MELD at transplant Donor adjustment factors: Age, race, gender, blood group, cause of death, donation after circulatory death Costs were adjusted for inflation based on consumer price index • inflation reported by the Bureau of Labor Statistics. 6
Proximity Points Results: Cost Metrics
Overall cost (patient care and transport) 8
Pretransplant cost per patient-month 9
Pretransplant cost per year 10
Transplant cost per patient 11
Transplant cost per year 12
E xpected Cost per Transplant and 1 Y ear Follow Up by ME LD/ PE LD (The cost model is multivariate; this is the MELD/PELD portion assuming average values for other candidate and donor characteristics.) 13
Distribution of Lab ME LD/ PE LD at Transplant 14
Transplant Cost and UHC Data Summary • UHC data allows for better modeling of the relationship between cost and lab MELD/PELD at transplant • Transplant costs per patient increased slightly among 4-district in-district scenarios, remained largely stable across scenarios • Transplant center costs increase as lab MELD/PELD increases. However, modeled scenarios show little variation in distribution of lab MELD/PELD at transplant. For average patient with MELD 20-25, model predicts center cost increase of about $10,000 per MELD/PELD point Largest projected increase in national median MELD/PELD at transplant: 4 points (21→25) for 4 districts w/o proximity For 8 districts w/ proximity, projected national median MELD/PELD at transplant: 2 points higher (21→23) Less than 25% of transplants fall in the MELD/PELD 20-25 range where cost increases are steepest, so overall center transplant cost impact is only in the $10,000-$15,000 range 15
Posttransplant cost per patient-month 16
Posttransplant cost per year 17
Overall cost (patient care) 18
Transport cost per transplant 19
Transport cost per year 20
Overall cost (patient care and transport) 21
Proximity Points Financial Analysis Summary • Analysis modeled underlying system costs, not charges. • Overall projected differences in cost for transport and patient care among the 28 scenarios were small, with only about a 2% difference between the most and least expensive scenarios. • Pretransplant costs per year were somewhat lower for broader sharing scenarios due to fewer patient-months at high lab MELD/PELD. • UHC data allowed for modeling variability in transplant costs, which were somewhat higher per year for broader sharing scenarios due to increased MELD/PELD at transplant. • Posttransplant costs per year showed almost no variation between scenarios. While transport costs varied with higher costs estimated in scenarios that • involve more flying, they were a small proportion of total transplant costs, at about 3% of the total estimated cost of transplant care. • Decreased pretransplant costs balanced increased transplant and transport costs in most scenarios. Most redistricting options are projected to be cost- neutral or cost-saving. 22
Q & A
Supplemental Slides: Proximity Points Financial Analysis 24
Cost Data: Medicare • Data Linkage: Linked clinical and demographic information from OPTN data with Medicare billing claims for liver transplant candidates and recipients between 2002 and 2008 Medicare billing claims provided payment information for patients with Medicare fee-for-service primary or secondary insurance Beneficiary ID from Medicare files were linked to OPTN records using Social Security number, gender, and date of birth 25
Analytic Approach: Transportation Costs • For each transplant, transport mode was predicted using a transport model that assigns cost and mode of transportation based on estimated travel times (Gentry, Liver Transplantation , 20;1237-1243, 2014) • Mode was: By ground if driving time was less than 2 hours (or 1.5 hours in OPOs that use helicopters); By helicopter in OPOs that use helicopters where driving would take 1.5 hours or more and the distance was 100 miles or less; By air from the nearest airport for longer distances • Round-trip cost was estimated to be $1108 per team by ground transportation, $4742 per team by helicopter, and distance-dependent for flights. (Flight cost estimates based on transports for liver transplant in the Living Legacy Foundation OPO in 2013. Lynch, American Journal of Transplantation , 9:10;2416-2423, 2009) • Round-trip costs for flights were estimated by distance, using a formula of $7,767+($8.40 x round trip miles). This cost includes the cost for aircraft charter, fuel, aircraft crew, and airport fees. 26
Analytic Approach: Pretransplant Cost in LSAM • Over each 5-year simulation, candidates began accruing pretransplant (pre-tx) care costs from the beginning of simulation or date listed. The pre-tx period ended when the candidate underwent transplant or was removed from the list. Cost of pre-tx care was estimated based on the number of days the • candidate was on the list with a lab MELD/PELD score of 6-40. • We calculated average pre-tx care cost per patient month by: Summing total cost of pre-tx care was over each 5-year simulation Calculating patient pre-tx costs per month by dividing total pre- tx cost by cumulative months waiting. Averaging over number of candidates to find the reported pretransplant cost per patient month • Medicare cost data was used to estimate pretransplant cost 27
Analytic Approach: Transplant and 1-Y ear Follow- Up Cost in LSAM • Within each 5-year simulation, candidates who underwent transplant accrued the total cost of transplant procedure plus 1 year of follow-up care. • Transplant and 1 year cost per patient was calculated by dividing the total transplant and 1 year cost over the 5-year simulation by the number of patients who underwent transplant. • UHC cost data was used to estimate transplant and 1-year follow up cost. 28
Recommend
More recommend
