Getting On Pathway?
The rising cost of healthcare is a topic of constant discussion and equally constant consternation. A lot of money is spent on healthcare in America, far more than in any other Western country, and over time, there has been tremendous inflation in this spending, particularly when indexed against the GDP.1 Oncology seems a particularly prominent passenger on this runaway train. Although oncology spending in both the commercial and Medicare sectors represents less than 10% of the total healthcare spend,2 several big-ticket items bring constant attention to this specialty. And why not, with immunotherapy costing $100,000 per course and oral targeted agents costing $5000 to $10,000 per month? These mind-boggling price tags seem out of reach for the average patient.
A major reason for the concern over the cost of these agents is the perceived lack of value. The clinical trials that have led to the approval of these agents often employ end points, or outcomes, that address only part of the value assessment patients must make. High cost may interfere with living longer or better regardless of disease-free survival or disease stability.
The ability to rank “relative values” associated with our therapeutic choices should empower us to make wise decisions in healthcare. It allows us to maximize value. And to the extent that we can improve outcomes, it allows us to improve quality of care. Since the healthcare dollar has finite limits, objective, evidence-based treatment should help us do better with the dollars available.
Sadly, ample evidence suggests we have far to go. The best marker of how far short we are falling is the regional variability in healthcare spend. Seniors in Minneapolis cost Medicare $3341 per year whereas in Miami, they cost Medicare $8881 per year, 2½ times as much.3 These variations are even more exaggerated when care during the last 6 months of life is analyzed. The data from the Dartmouth Atlas are most compelling.4 Utilization of all healthcare services, including hospital days, ICU days, and doctors visits, goes up. But what is most alarming is the variability from state to state and from region to region. For example, hospital days in the last 6 months range from 7 in Utah and Oregon to almost 16 in DC and almost 12 in Florida. Office visits follow the same trends. Not unexpectedly, hospice enrollment follows an inverse relationship. Without question, socioeconomic factors explain some of the variability and there is no clear “quality standard” which allows definition of an appropriate benchmark. However, this much variability cannot be viewed as good.
This type of variability certainly occurs in oncology care, even on a much smaller scale. Our studies looking at charges for lung cancer therapy within a single large practice shows wide interphysician differences.5 Notably, spend was not correlated at all with better survival (and ). Patients cared for by the most expensive doctors did just as well (or just as poorly) as those cared for by the most thrifty doctors. In oncology, the spend is spread among several services and even several sites of service. Although there are some publications based on Surveillance, Epidemiology, and End Results data in the Medicare population,6 there was very little in the published literature to guide us in identifying where opportunities lie. We therefore entered into a partnership with Milliman to study costs of cancer care in the commercial population.7
The study analyzed data from Medstat 2007, a claims database of commercial insurers, including several BCBS plans, covering 28 million lives. We focused on the top 10 cancer diagnoses, which represented 70,000 individuals. Cancer was relatively uncommon in this cohort, only 0.7% of all covered individuals, but cancer claims made up almost 10% of the cost. The 10 cancer diagnoses represented about 65% of all cancer diagnoses in this cohort, with breast cancer being the most common diagnosis. Among the cancer patients, about 25% received chemotherapy in a given year. The average cost per cancer patient receiving chemotherapy was about $110,000 per year, 4 times as much as the average cancer patient not receiving chemotherapy. Further, the cost for cancer patients receiving chemotherapy was dramatically higher, more than 20-fold greater, than the cost for patients with other chronic diseases like diabetes.
So how is the money spent? The majority of the charges were for outpatient services (). In analyzing the outpatient costs, only 25% of costs were for chemotherapy. The remainder of the costs included imaging, outpatient surgery, and other ambulatory services. We also found there were about 2 hospitalizations per year and 2 emergency department (ED) visits per year in chemotherapy-treated patients, half of which could be attributed to chemotherapy toxicity. Unfortunately, we found relatively low hospice utilization rates. And more concerning, we found a large cohort of patients who died in the hospital and received chemotherapy within 14 days of death.
As with the published experience with the Dartmouth Atlas, we identified marked regional variations in care (). The range in chemotherapy costs was almost a third higher in the highutilizing regions, and not surprisingly, hospital admits and ED visits varied 2-fold and 4-fold respectively from lowto high-utilizing regions. This data set identifies opportunities for cost control, especially in chemotherapy costs, hospital and ED costs, and end-of-life costs.
The US Oncology Network has, for the last 6 years, focused on the ability to control costs by adopting an evidence- based treatment platform, Level I Pathways. These evidence-based treatment guidelines are embedded in the iKnowMed EHR and offered as a decision support tool, with reports that allow implementation of a process improvement methodology. We have reported on the impact of these evidence-based pathways in our network in patients with lung cancer and colorectal cancer.
Our first study, in non-small cell lung cancer patients, was a collaboration with Aetna.8 Over an 18-month period, we looked at cost of care in 1400 patients
at 8 sites in the US Oncology Network. Clinical data including stage of disease and line of therapy were documented. We studied chemotherapy costs as well as costs associated with acute care visits, E and M codes, laboratory, and supportive care drugs. Within our sites of service, 1100 of 1400 patients were treated according to guidelines, considered “on pathway,” and 300 of 1400 were treated “off pathway.” There was a substantial cost difference between the groups, with a 35% reduction in chemotherapy costs associated with being treated on pathway ($18,000 vs $27,700). Importantly, survival of those treated on pathway was identical to those treated off pathway ( and ).
We also analyzed the effect of pathways compliance in colon cancer.9 In this second study, we looked at costs associated with pathway status (on or off) in a national database with the assistance of Milliman. Since survival data were not available in this claims system, we analyzed survival with respect to pathways status within iKnowMed. We again found a substantial savings associated with being “on pathway,” about 35%. Interestingly, we found evidence that disease-free survival and overall survival with treatment on pathway were at least as good as in the offpathway cohort. A preliminary review of inpatient charges and hospitalization rates within the on pathway/off pathway cohorts suggests reduced costs and utilization in the on-pathway group, providing further evidence of cost savings associated with adherence to these evidence-based treatment guidelines.
This body of data makes a compelling case for value associated with a disciplined, cost-conscious, evidence-based approach to chemotherapy decision making. This value derives from clearly reduced costs but also improved outcomes. The value associated with this approach is physician developed and driven. And a similar physician developed and driven approach may well be feasible to attack costs associated with hospitalizations and ED visits and especially end-of-life care, where there is ample opportunity for improvement.
There are clearly challenges to the broad adoption of this approach. First, there are innumerable evidence-based treatment guidelines. Most of these are at best cost neutral and at worst cost insensitive. It would be a mistake to assume an equal benefit across all of these guidelines. Second, this approach requires a genuine attempt at process improvement. Reporting and accountability are mandatory. Without physician buy-in, the process is doomed to failure. Finally, there is widespread dissatisfaction with the current oncology reimbursement methodology. A transition to a methodology that financially rewards high-level, quantifiable performance will greatly enhance likelihood of general acceptance. There will need to be trust between payers and providers to succeed in this system. Ultimately, the reward will be improved value and enhanced quality.
Author Affiliations: From New York Oncology Hematology and Albany Cancer Center (MK), Albany NY; Texas Oncology (JRH), Austin, TX; and US Oncology (MK, JRH), Houston, TX.
Funding Source: None.
Author Disclosures: Drs Kolodziej and Hoverman report employment with US Oncology, Houston, TX.
Authorship Information: Concept and design (MK, JRH); acquisition of data (MK); analysis and interpretation of data (MK, JRH); drafting of the manuscript (MK, JRH); critical revision of the manuscript for important intellectual content (MK, JRH); provision of study materials or patients (MK); and supervision (MK).1. Borger C, Smith S, Truffer C, et al. Health spending projections through 2015: changes on the horizon. Health Aff (Millwood). 2006;25:w61- w73.
2. Tangka FK, Trogdon JG, Richardson LC, Howard D, Sabatino SA, Finkelstein EA. Cancer treatment cost in the United States: has the burden shifted over time? Cancer. 2010;116(14):3477-3484.
3. Congressional Budget Office. Geographic Variation in Healthcare Spending 2005. www.cbo.gov/ftpdocs/89xx/doc8972/2-15-GeogHealth.pdf.
4. The Trustees of Dartmouth College. The Dartmouth Atlas of Health Care. http://www.dartmouthatlas.org/.
5. Hoverman JR, Robertson SM. Lung cancer: a cost and outcome study based on physician practice patterns. Dis Manag. 2004;7:112-123.
6. Hassett MJ, Neville BA, Weeks JC. The relationship between cost, quality, and outcomes among women with breast cancer in SEER-Medicare. J Clin Oncol. 2011;29:384S(suppl; abstr 6001).
7. Kolodziej M, Hoverman JR, Garey JS, et al. Benchmarks for value in cancer care: an analysis of a large commercial population. J Oncol Pract. 2011;7(5):301-306.
8. Neubauer MA, Hoverman JR, Kolodziej M, et al. Cost-effectiveness of evidence-based treatment guidelines for the treatment of non-small-cell lung cancer in the community setting. J Oncol Pract. 2010;6(1):12-18.
9. Hoverman JR, Cartwright TH, Patt DA, et al. Pathways, outcomes, and costs in colon cancer: retrospective evaluations in two distinct databases. J Oncol Pract. 2011;7(3)(suppl 3):52s-59s.