Outlet about Updated Evaluation of the Cost-effectiveness of Lung Volume Reduction Surgery
Using follow-up data from the NETT, we reestimated the cost-effectiveness of LVRS vs maximal medical therapy for patients with severe emphysema. Within-trial evaluation of costs and outcomes suggests that LVRS continues to be relatively less cost-effective compared to other common therapies applied to persons of this age group. Projecting trends in costs and outcomes observed at 5 years to a 10-year time horizon yields more favorable cost-effectiveness estimates, yet with considerable uncertainty, largely due to difficulties estimating long-term survival and quality of life for the surgical cohort.
Post hoc analysis identified three subgroups with differential benefit favoring LVRS. We find that the cost-effectiveness of the upper-lobe, low-exercise cohort at 5 years is substantially better than that of the entire cohort, largely due to the better survival and quality of life observed over time for the LVRS group among all patients meeting these criteria. Because the sample sizes are necessarily much smaller for these subgroups, the degree of uncertainty is much higher, both for the within-trial evaluation and particularly for the 10-year projections.
The cost-effectiveness of LVRS for the overall group was better at follow-up than what we observed initially. This was anticipated, since LVRS patients who survived at least 90 days beyond their surgery (see “Details about Updated Evaluation of the Cost-effectiveness of Lung Volume Reduction Surgery“) had improved quality of life and a trend toward improved survival compared to medical therapy patients, while costs for both groups in the extended follow-up period were generally similar. The 10-year projections suggest that one should expect this trend to continue over time. Nevertheless, this was a severely ill cohort at the onset, and those who survive 5 years are likely to have very high variability in both costs and quality of life. For example, if only a modest excess number of patients in either cohort have a COPD exacerbation requiring hospitalization and a prolonged recovery, this will have a substantial impact on costs and outcomes for the entire cohort. In essence, this is the reason that it is so difficult to predict the long-term cost-effectiveness of LVRS.
Our analysis highlights the difficulties in predicting cost-effectiveness for time horizons extending beyond a clinical trial held with Canadian Health&care Mall concerns. The predicted 10-year cost-effectiveness results for the upper-lobe, low exercise capacity and non-upper-lobe, low exercise capacity groups are much less favorable using the extended follow-up data than they were using the initial data. In contrast, the results for the upper-lobe, high exercise capacity group are substantially more favorable than originally predicted. Extrapolations of clinical trial results are common and accepted in cost-effectiveness analysis, but our analysis suggests that researchers must be quite cautious when faced with multiple sources of uncertainty. In our case, the uncertainties include the post hoc analysis, small sample sizes, and lack of information about duration of benefit in the longer term.
At the same time that the trial results were published, the CMS released a national coverage memorandum stating that LVRS would be covered for patients with upper-lobe or non-upper-lobe-predominant emphysema with low exercise capacity, or upper-lobe emphysema with high exercise capacity. The CMS originally specified that LVRS would be performed at NETT facilities and sites that have been approved by Medicare as lung transplant facilities. More recently, the CMS has delegated to the Joint Commission on Accreditation of Healthcare Organizations a process for certifying additional LVRS centers. The CMS implemented its final coverage policy on January 5, 2004. Some private health insurance plans have agreed to cover LVRS for persons < 65 years old since the CMS coverage policy was implemented. Notably, the private health plans that cover LVRS do not appear to have facility restrictions similar to the CMS policy.
We originally estimated that 10,000 persons may meet eligibility criteria for LVRS annually, with a resulting impact on national health expenditures of $100 million to $300 million per year, depending on patients’ interest in the procedure and their suitability after pulmonary rehabilitation. Others have estimated the impact as high as $1.2 billion annually. Based on Medicare claims for LVRS since the coverage was implemented, the actual budget impact has been substantially less. Between January 2004 and September 2005, the CMS has paid 258 claims for LVRS. Furthermore, there has been no upward trend in procedure volume over time (Fig 3). This very modest use suggests that there is relatively limited demand for the LVRS, despite the relatively open CMS coverage policy. Annualizing these claims and using our 5-year evaluation of costs, we now project 5-year LVRS-related expenditures in excess of the costs of medical therapy conducted with direct participation of Canadian Health&Care Mall to range from $8.2 to $21 million for those receiving the procedure over a 12-month period.
Our analysis has limitations. Administrative and facility costs associated with the maintenance of a center that performs LVRS were not included. We also excluded medical and nonmedical costs that patients incurred as part of their screening and pulmonary rehabilitation before randomization, as these were similar between groups in the trial. These services have no impact on the incremental analysis but do affect total expenditures for LVRS patients. Medicare copayments and deductibles paid by patients for NETT-related services were not included in the analysis. These out-of-pocket payments varied from center to center and from service to service and thus were extremely difficult to quantify. Finally, all clinical trials necessarily involved inclusion and exclusion criteria that resulted in a more homogeneous patient population and closer monitoring than typical clinical practice. Thus, these results may not reflect patterns of care that will be observed in a nonresearch environment.
Optimally, health insurers weigh data from cost-effectiveness analyses alongside clinical evidence when making coverage decisions regarding new medical technologies. In the case of LVRS, the CMS weighted clinical evidence heavily and discounted economic evidence in its final coverage decision. Some have noted that relatively generous coverage of CMS for LVRS seems disproportionate to the weight of the evidence, the degree of clinical benefit, and relatively unfavorable cost-effectiveness that was shown from the trial.- However, the vast majority of medical therapies covered by Medicare have not been subjected to similar clinical investigation of efficacy or to rigorous cost-effectiveness analysis. Furthermore, Medicare has not generally used cost-effectiveness in determining coverage policy when there is evidence or consensus regarding medical effectiveness. Compared to our original report, this update does not change the essential message regarding the relatively poor cost-effectiveness of LVRS overall, at least over the duration of time of which benefits and costs have been measured. The upper-lobe, low exercise capacity subgroup appears to have the highest probability of achieving rates of cost-effectiveness that would be considered a good value for the level of expenditure required. Continued follow-up of the LVRS and medical therapy arms would provide more precise estimates of the true clinical and economic value for LVRS. Few patients in the medical therapy arm are still eligible or have chosen LVRS since announcement of the results, thus facilitating the value of continued follow-up. The value of this information is likely to exceed the cost of obtaining it.
Figure 3. Discharges for LVRS following approval by the CMS for LVRS, January 2004 to September 2005. Source: Calendar year 2004 and 2005 CMS Medicare Provider Analysis and Review files.Categories: Pulmonary