Actual Medical and Pharmacy Costs for Bariatric Surgery
Actual Medical and Pharmacy Costs for Bariatric Surgery
Unlike previous studies that have reported cost data for up to 2 years postsurgery, the present study provides data on medical and pharmacy expenditures for a 6-year postsurgical period and compares WLS patients to obese nonsurgical patients. In our sample of individuals insured through the state-managed health insurance agency, no significant differences in nonpharmaceutical medical care costs between WLS patients and obese control patients were observed in the first 3 years postsurgery; however, by the fourth year postsurgery, costs were significantly lower for individuals who had undergone WLS, and these lower medical care costs continued through 6 years postsurgery.
This study provides support for a sustained reduction in medical care costs following WLS, with a return on investment starting at year 4 postsurgery. Our results are inconsistent with those of Cremieux et al, however, who estimated a complete return on investment from laparoscopic surgery within 2 years. We can only speculate on reasons for the differences between our results and those of Cremieux et al. First, our study used a much smaller sample size, and thus may be less reliable. Second, Cremieux et al used cost estimates beginning 18 months postsurgery rather than actual expenditures. Our study used actual claims data for a 2-year presurgical and 6-year follow-up period. Third, Cremieux et al examined a nationwide sample of privately insured individuals, whereas our study focuses on Louisiana state-managed medically insured state employees and their dependents. It is possible that unique cultural or socioeconomic features of our sample, or the unique features of our managed care system, account for the disparate findings. Finally, the costs of WLS were lower for Cremieux et al compared with the present study ($17,000 vs $25,000). In addition, it is possible that baseline medical care costs differed for these two samples, potentially affecting the return on investment.
Our examination of specific subcategories of medical claims costs revealed no differences in office visit, emergency department, or laboratory/pathology postsurgical costs between the WLS group and the nonsurgical group. It is unclear why there are no reported differences between the non-WLS and WLS group; however, one may speculate that the costs of these claims were relatively low initially, thus making it difficult to produce a statistically significant finding. In the one distinct subcategory in which a sustained cost difference was found (sleep facilities), the cost differences between the two groups is small, highly variable from year to year, and accounts for only a small portion of total medical costs. Most of the differences in costs were for expenditures not categorized into the existing categories (ie, "remaining" costs). This suggests the need for more fine-grained analyses of post-WLS medical expenditures changes in future studies. Interestingly, costs for the nonsurgical group increased substantially during the first year postsurgery, primarily due to an increase in "remaining medical services." Unfortunately, the cause for this is unknown.
Our results are similar to past research showing a rapid decrease in total pharmacy expenditures following WLS; however, whereas most previous researchers have examined medication costs for a few years postsurgery, our study was able to report on medication cost changes for 6 years postsurgery. Unfortunately, the lower total pharmacy costs for WLS patients seen in the first few years after surgery were not maintained, despite the finding that the WLS group had significantly lower expenditures for antidiabetic agents, antihypertensive agents, and dyslipidemic agents through all 6 postsurgical years. Notably, the results of the present study are consistent with a report from the Swedish Obese Subjects study, which also reported the absence of a reduction in total medication costs for a 6-year post-WLS period.
Our study suggests that a long-term return on investment and future cost savings may not occur in all populations and is not likely to be obtained as rapidly as some studies have suggested. It should be noted that researchers have cautioned against the use of return on investment as a requirement for insurance agencies to reimburse for WLS, and instead argue that the ability of WLS to treat medical conditions with superior cost-effectiveness than other approaches is a more appropriate guide for reimbursement determinations. In addition, it is relevant that in this and other similar studies, the effects of WLS on changes in worker productivity, which may be more costly than the medical costs, are not considered.
There are limitations to our study. First, although we examined subcategories of cost, we were not able to examine all of the possible subcategories and thereby fully describe all medical and pharmacy cost changes. The results of the present study suggest that future studies may benefit from analyzing a wider spectrum of medical and pharmacy cost subcategories. An additional limitation is that we do not have full demographic data (eg, race, educational attainment) or comprehensive medical data on the non-WLS obese group, and thus cannot fully attest to the comparability of our control group. Nonetheless, the non-WLS group comprised individuals who were initially interested in WLS, and their healthcare costs were similar to that of the WLS group before surgery, suggesting that both groups were relatively similar with respect to the variables examined.
Discussion
Unlike previous studies that have reported cost data for up to 2 years postsurgery, the present study provides data on medical and pharmacy expenditures for a 6-year postsurgical period and compares WLS patients to obese nonsurgical patients. In our sample of individuals insured through the state-managed health insurance agency, no significant differences in nonpharmaceutical medical care costs between WLS patients and obese control patients were observed in the first 3 years postsurgery; however, by the fourth year postsurgery, costs were significantly lower for individuals who had undergone WLS, and these lower medical care costs continued through 6 years postsurgery.
This study provides support for a sustained reduction in medical care costs following WLS, with a return on investment starting at year 4 postsurgery. Our results are inconsistent with those of Cremieux et al, however, who estimated a complete return on investment from laparoscopic surgery within 2 years. We can only speculate on reasons for the differences between our results and those of Cremieux et al. First, our study used a much smaller sample size, and thus may be less reliable. Second, Cremieux et al used cost estimates beginning 18 months postsurgery rather than actual expenditures. Our study used actual claims data for a 2-year presurgical and 6-year follow-up period. Third, Cremieux et al examined a nationwide sample of privately insured individuals, whereas our study focuses on Louisiana state-managed medically insured state employees and their dependents. It is possible that unique cultural or socioeconomic features of our sample, or the unique features of our managed care system, account for the disparate findings. Finally, the costs of WLS were lower for Cremieux et al compared with the present study ($17,000 vs $25,000). In addition, it is possible that baseline medical care costs differed for these two samples, potentially affecting the return on investment.
Our examination of specific subcategories of medical claims costs revealed no differences in office visit, emergency department, or laboratory/pathology postsurgical costs between the WLS group and the nonsurgical group. It is unclear why there are no reported differences between the non-WLS and WLS group; however, one may speculate that the costs of these claims were relatively low initially, thus making it difficult to produce a statistically significant finding. In the one distinct subcategory in which a sustained cost difference was found (sleep facilities), the cost differences between the two groups is small, highly variable from year to year, and accounts for only a small portion of total medical costs. Most of the differences in costs were for expenditures not categorized into the existing categories (ie, "remaining" costs). This suggests the need for more fine-grained analyses of post-WLS medical expenditures changes in future studies. Interestingly, costs for the nonsurgical group increased substantially during the first year postsurgery, primarily due to an increase in "remaining medical services." Unfortunately, the cause for this is unknown.
Our results are similar to past research showing a rapid decrease in total pharmacy expenditures following WLS; however, whereas most previous researchers have examined medication costs for a few years postsurgery, our study was able to report on medication cost changes for 6 years postsurgery. Unfortunately, the lower total pharmacy costs for WLS patients seen in the first few years after surgery were not maintained, despite the finding that the WLS group had significantly lower expenditures for antidiabetic agents, antihypertensive agents, and dyslipidemic agents through all 6 postsurgical years. Notably, the results of the present study are consistent with a report from the Swedish Obese Subjects study, which also reported the absence of a reduction in total medication costs for a 6-year post-WLS period.
Our study suggests that a long-term return on investment and future cost savings may not occur in all populations and is not likely to be obtained as rapidly as some studies have suggested. It should be noted that researchers have cautioned against the use of return on investment as a requirement for insurance agencies to reimburse for WLS, and instead argue that the ability of WLS to treat medical conditions with superior cost-effectiveness than other approaches is a more appropriate guide for reimbursement determinations. In addition, it is relevant that in this and other similar studies, the effects of WLS on changes in worker productivity, which may be more costly than the medical costs, are not considered.
There are limitations to our study. First, although we examined subcategories of cost, we were not able to examine all of the possible subcategories and thereby fully describe all medical and pharmacy cost changes. The results of the present study suggest that future studies may benefit from analyzing a wider spectrum of medical and pharmacy cost subcategories. An additional limitation is that we do not have full demographic data (eg, race, educational attainment) or comprehensive medical data on the non-WLS obese group, and thus cannot fully attest to the comparability of our control group. Nonetheless, the non-WLS group comprised individuals who were initially interested in WLS, and their healthcare costs were similar to that of the WLS group before surgery, suggesting that both groups were relatively similar with respect to the variables examined.
Source...