Abstract
Objective: To evaluate the clinical and economic burden of COPD patients to Medicaid. Study Design: Retrospective, observational matched cohort design. Methods: We calculated the incremental costs incurred and medical resources used by COPD patients relative to those without COPD. Data were obtained from 8 Medicaid states during 2003–2007. COPD patients were defined as Medicaid beneficiaries ≥40 years with a COPD diagnosis (ICD-9 CM: 491.xx, 492.xx, 496.xx) and treated with maintenance drugs for COPD. Patients were matched (1:3) to Medicaid beneficiaries without a COPD diagnosis on age, gender, race, index year, Medicare/Medicaid dual eligibility, and use of long-term care. Results were stratified by Medicare/Medicaid dual eligibility status and race. Results: A total of 10,221 COPD and 30,663 non-COPD patients were included. Cohorts were on average 65 years of age, 80% White, and 64.8% having Medicare/Medicaid dual eligibility. Inpatient hospitalizations and home healthcare visits/durable medical equipment were primary drivers of incremental medical costs. COPD patients were more than twice as likely to have a hospitalization (odds ratio [95% confidence interval] = 2.32 [2.19, 2.45]) or home healthcare visit/durable medical equipment (2.95 [2.82, 3.08]) compared to non-COPD patients. Medicaid incurred $2118/year in incremental costs due to COPD. On average, incremental costs were 7 times greater for non-dual-eligible patients ($4917) compared to dual-eligible patients ($667), and were more than double for Blacks compared to Whites ($4141 vs $1593). Conclusion: COPD imposes a substantial economic and clinical burden on the Medicaid program; this burden differs by dual eligibility status and race.
Introduction
Chronic obstructive pulmonary disease (COPD) is a preventable and treatable disease characterized by airflow limitation and gradual loss of lung function that is not fully reversible (Citation1). The disease poses a substantial clinical and economic burden as evidenced from several studies evaluating COPD costs from the perspective of national health authorities or specific payers (Citation2–6). The Medicaid program currently covers 48 million low-income and disabled people, and as a payer of healthcare services in the United States (US), accounts for over 15% of all national healthcare expenditures (Citation7). Yet to date, only one study has evaluated the economic burden of COPD in a Medicaid population (Citation3).
Evaluation of the burden of COPD in a Medicaid population is important and relevant for several reasons. Smoking –the most important risk factor for COPD –is reported to have at least a 50% higher prevalence in Medicaid beneficiaries compared to the general population (Citation8). Furthermore, low socioeconomic status has been found to be related to worse COPD outcomes (Citation9), which suggests the costs of COPD may be higher in a Medicaid population. Additionally, Medicaid enrollment is projected to increase to almost 70 million as a result of mandatory requirements of the 2010 Affordable Care Act (Citation10,11). Therefore, it may become increasingly important to benchmark costs incurred by Medicaid due to chronic conditions like COPD.
The study objective was to compare the clinical and economic burden between patients with and without COPD enrolled in the US Medicaid program, and to estimate the incremental costs of COPD. In the only study published in the Medicaid population on COPD costs, dual-eligible patients were excluded. Dual-eligible patients are eligible for both Medicaid and Medicare (Citation12).
For patients enrolled in both programs, Medicaid usually covers the costs of medical services related to patient responsibility (i.e., coinsurance, copays, deductibles, and premiums), while Medicare pays the bulk of the costs of the service that are not patient responsibility (Citation12). Dual enrolled patients are usually excluded from analysis of Medicaid costs because claims data are lacking for the amounts covered by Medicare precluding a comprehensive evaluation of costs to Medicaid. Also, prior research has shown possible racial disparities in the Medicaid population, with lower costs and use of healthcare services for African-Americans compared to Whites, contrary to expectations (Citation13). Therefore, the current study was focused on the incremental costs of COPD by Medicare/Medicaid dual eligibility status and race.
Methods
The study used data from the MarketScan Medicaid database, which contains the medical, surgical, and prescription drug experience of nearly 7 million Medicaid beneficiaries from 8 states. The data include records of inpatient services, inpatient admissions, outpatient services, prescription drug claims, information on long-term care, and demographic information for each patient. Data collection and analysis were in compliance with the Health Insurance Portability and Accountability Act (HIPAA). Furthermore, data on geographic region is not provided to maintain confidentiality of the states. Data from January 1, 2003 to December 31, 2007 was used (study period).
Study design and sample selection
A retrospective, matched cohort study design was used. Patients with COPD were identified as Medicaid beneficiaries diagnosed with COPD (International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9-CM] codes 491.xx, 492.xx, and 496.xx) and newly initiated on a maintenance medication (including short- and long-acting anticholinergics, inhaled corticosteroids, long-acting beta-agonists, and combination products of inhaled corticosteroids and long-acting beta-agonists during 01/01/2004 to 12/31/2006 (identification period). A cohort of non-COPD patients were identified as Medicaid beneficiaries without a COPD diagnosis but having at least 1 medical or pharmacy claim during each year of the identification period.
The first chronologically occurring date of maintenance medication claim during the identification period was termed as the index date for COPD patients. The index date for non-COPD patients was the date of their first medical or pharmacy claim during each calendar year of the identification period. Thus, there was a possibility of 3 index dates (2004, 2005, and 2006) for a non-COPD patient before the match process to enable matching on index year. A 1-year period before the index date (pre-index period) was used to characterize the study population at baseline. The 1-year period after the index date (follow-up period) was used to assess the outcomes of costs and resource use.
The final study population met the following criteria: 1) ≥40 years of age at index date, 2) continuously eligible to receive healthcare services in the pre-index and follow-up period, 3) enrolled in fee-for-service plans, and 4) absence of exclusionary co-morbid conditions; namely, cystic fibrosis, bronchiectasis, respiratory cancer, pulmonary fibrosis, pneumoconiosis, sarcoidosis, and pulmonary tuberculosis (including fibrosis due to tuberculosis).
A matched cohort analytical approach was adopted, where the COPD cohort was matched to the non-COPD cohort on a 1:3 basis on age (exact), gender, race, index year, Medicare/Medicaid dual eligibility, and use of long-term care during the pre-index period. The match process was run (without replacement) for each calendar year of the identification period since a patient in the non-COPD cohort could have more than one index date, and one of the variables for the match process was the index year (Appendix A1). The final study population included all matched patients from each year.
Outcome measures
The primary study outcome was an estimate of the incremental costs incurred due to COPD computed by subtracting total all-cause healthcare costs of non-COPD from COPD patients. The incremental cost approach has been used in the literature as an alternative to estimating disease-specific costs usually defined as costs on medical claims with a primary diagnosis of the disease (Citation2,3,Citation14,15). The incremental cost approach allowed capture of costs due to multiple co-morbidities that could be related to or exacerbated by COPD (e.g., congestive heart failure) but may not necessarily be coded with a primary diagnosis of COPD. The components of incremental costs were also computed and included medical, pharmacy, and long-term care costs. Costs were estimated as costs paid after applicable discounts but inclusive of any deductibles, coinsurance, or copayment, and were standardized to the last year of available data (2007) using the medical component of the Consumer Price Index (Citation16). The proportion of patients with all-cause medical resource use was also calculated, as were the number of the visits of each type of medical resource use, including inpatient hospitalizations, emergency department (ED) visits, physician visits, outpatient visits, home healthcare visits/durable medical equipment, and long-term care visits.
Analytical approach
The incremental costs due to COPD were computed by subtracting all-cause healthcare costs of the non-COPD cohort from that of the COPD cohort for the total and each of the components (pharmacy, medical, and long-term care). Although all-cause costs were captured (i.e., any diagnosis), they can overestimate costs due to COPD by including costs of conditions not usually theorized to be exacerbated by COPD. Therefore, in the current study, the conditions obtained from the Charlson co-morbidity index (CCI) (namely, malignant cancer with or without metastasis, diabetes with and without complications, liver disease, connective tissue disease, peptic ulcer disease, hemiplegia or paraplegia, renal disease, and acquired immunodeficiency syndrome) were adjusted in the multivariate analysis to eliminate any excess costs due to unrelated conditions. The adjusted costs differences were assessed using generalized linear models with a log-link using the appropriate distribution (usually gamma or Poisson), and accounting for matched pairs.
The method of recycled predictions was then utilized to obtain predicted costs for each cohort using coefficient values from the predicted equation (Citation17). The recycled predictions method calculates the mean predicted marginal outcome by the main predictor (i.e. COPD/non-COPD), thus allowing us to compare the predicted marginal outcome on average, while holding constant all other model covariates except the main predictor of interest. Conditional logistic regression was used to compare the odds of having all-cause resource utilization, and negative binomial regression was used to compare the difference in the number of all-cause visits between the cohorts. Stratified analyses were also conducted by Medicare/Medicaid dual eligibility status and race (White, Black, other). All statistical tests assessed a 2-sided hypothesis of no difference between the study cohorts at a significance level of 0.05, and were conducted using SAS statistical software version 9.2.
Results
Study sample
A total of 10,363 and 136,809 patients in the COPD and non-COPD cohorts, respectively, met all study criteria. The 1:3 matching process retained 10,221 COPD and 30,663 non-COPD patients (). The matched cohorts were on average 65 years of age with 37% male, 80% White, and 65% patients with Medicare/Medicaid dual eligibility. COPD patients had a significantly higher overall disease burden at baseline, with the CCI being twice that of the non-COPD cohort (2.0 vs 1.1, p < 0.001; ICD-9-CM codes for COPD were excluded from computation of the CCI). The proportion of patients with COPD-related co-morbidities (i.e., asthma, upper respiratory tract infection, lower respiratory tract infection, depression, and cardiovascular disease) and co-morbidities included in CCI was also higher for the COPD cohort.
Table 1. Baseline description of study sample: after matching
Outcomes
During follow-up, COPD patients incurred approximately $2000 higher costs/year compared to non-COPD patients after covariate adjustment (). The medical cost component contributed to a majority of the incremental costs ($1640) followed by the pharmacy cost component ($1176), whereas the long-term care cost component was found to be lower for COPD compared to non-COPD patients. Hospitalization and home healthcare visits/durable medical equipment contributed the majority of the incremental costs due to COPD.
Table 2. Average annual, adjusted incremental cost for Medicaid beneficiaries attributable to COPD
The stratified analyses revealed similar trends ( and ). However, important differences were noted in the magnitude of incremental costs when stratified by Medicare/Medicaid dual eligibility status or race. Incremental costs were almost 7 times higher in non-dual-eligible patients compared to dual-eligible patients (: $4917 vs $667). The proportionate increase in costs also differed, as these incremental costs represented a 27.5% and 4.6% increase for COPD compared to non-COPD patients, in non-dual-eligible and dual-eligible patients, respectively. The incremental costs for both pharmacy and medical cost components were higher in non-dual-eligible compared to dual-eligible patients. However, incremental pharmacy costs were 1.6 times higher ($1570 vs $977), and incremental medical costs were 7 times higher ($3746 vs $494) in the non-dual-eligible compared to the dual-eligible sample.
Figure 1. (a) Stratified analyses: Incremental all-cause annual costs per Medicaid recipient for COPD –Medicare/Medicaid dual eligibility status. (b) Stratified analyses: Incremental all-cause annual costs per Medicaid recipient for COPD –race.
Distinct differences were noted by race, too, with incremental costs being highest for Blacks (). There was more than a 2-fold difference in incremental costs between Whites and Blacks ($1593 vs $4141, respectively), with medical costs accounting for most of the difference.
Significant differences were noted between COPD and non-COPD patients in all types of medical resource utilization (). The differences were greatest in the use of inpatient hospitalizations (odds ratio, OR [95% confidence interval (CI)] = 2.32 [2.19, 2.45]), ED visits (OR [95% CI] = 2.22 [2.12, 2.32]), and home healthcare visits/durable medical equipment (OR [95% CI] = 2.95 [2.82, 3.08]), with COPD patients being more than twice as likely to have these types of resource use compared to non-COPD patients. Stratified analyses showed similar trends between COPD and non-COPD patients (). Notably, the rates of inpatient hospitalizations in COPD patients were higher in non-dual-eligible compared to dual-eligible patients, and in Blacks compared to Whites. The differences in the likelihood of use were also reflected in the amount of use as seen by the number of all-cause visits, and stratified analyses for this outcome showed similar trends (Appendix-A2).
Figure 2. All-cause adjusted healthcare resource use. COPD –chronic obstructive pulmonary disease; DME –durable medical equipment; ED –emergency department; IP –inpatient hospitalization; HH –home health; LTC –long-term care; OP –outpatient visit; Phy –physician visit. aPredicted proportions obtained from conditional logistic regression models controlling for malignant cancer with or without metastasis, diabetes with and without complications, liver disease (mild, moderate, severe), connective tissue disease, peptic ulcer disease, hemiplegia or paraplegia, renal disease, and AIDS. *Differences between COPD and non-COPD patients are statistically significant at p < 0.05.
Table 3. All-cause adjusted healthcare resource use for Medicaid beneficiaries: Stratified by Medicare/Medicaid dual eligibility status and race
Discussion
This retrospective database analysis found the US Medicaid program to incur a significant incremental cost/year due to COPD, representing a 14% increase in costs for COPD compared to non-COPD patients. Furthermore, Medicaid incurs incremental costs almost 7-fold higher for non-dual-eligible compared to dual-eligible patients. The incremental cost difference between these 2 patient populations were primarily due to medical costs; Medicaid only covers medical costs related to patient responsibility for dual-eligible patients, but covers the majority of the medical costs for non-dual-eligible patients (Citation12). Pharmacy costs, however, seems to be equally borne by Medicaid for both dual-eligible and non-dual-eligible patients. However, implementation of the Medicare Part D Low-Income Subsidy (LIS) program is likely to shift the burden of the pharmacy costs for dual-eligible patients further to Medicare (Citation18).
A review by Foster et al found that the studies conducted since 2000 showed a range of excess (or incremental) costs from about $6100 to $6600 (or $6369 to $6892 adjusted to 2007 US dollars [USD]) (Citation5). Mapel et al estimated COPD costs for a US managed care organization at $6039 (1997 USD) or $9037 (adjusted to 2007 USD) (Citation2). Grasso et al reported for the Medicare program $4971 (1992 USD) or $9180 (adjusted to 2007 USD) (Citation19). Menzin et al reported the COPD cost in Medicare managed plans to be $6300 (2004 USD) or $7132 (adjusted to 2007 USD) (Citation4). Miller et al derived an estimate of incremental costs of $4932 (2000 USD) or $6639 (adjusted to 2007 USD) using data from the Medical Expenditure Panel Survey (MEPS), which represents a sample of the US population (Citation20). Marton et al reported costs of COPD in 2 state Medicaid programs, which ranged from $5200 to $6500 (2001 USD) or $6692 to $8365 (adjusted to 2007 USD) (Citation3). The majority of patients in our study were dual-eligible, whereas Marton et al excluded all dual-eligible patients. The inclusion of dual-eligible patients may be the reason for the lower estimates ($2,118) in our study, but the estimate for non-dual-eligible patients ($4917) is closer to the previously reported estimates for the Medicaid population.
There has been limited literature in COPD on the impact of race on outcomes in COPD (Citation9,Citation13,Citation21,22). Recent analysis of the Maryland Medicaid population found some evidence of racial disparities, as Black adults were significantly less likely to use any type of medical service and had lower costs than Whites (Citation13). However, our study found similar use of services between Blacks and Whites with higher use for hospitalizations. Blacks had higher rates than Whites in our study, and the finding is similar to results from a nationwide sample comparing African-Americans to non-Hispanic Whites (Citation22). A recent prospective study found that the higher risk in Blacks compared to Whites in the risk of an acute exacerbation requiring an ED visit or hospitalization was explained entirely by socioeconomic status (Citation9). However, our sample includes patients from the same socioeconomic strata and yet a difference was noted. Regardless of the reason, the analysis showed that the magnitude difference in costs between Black COPD and non-COPD patients was far greater than between White COPD patients and non-White COPD patients, suggesting that Medicaid incurs a greater burden for Blacks than Whites due to COPD.
The findings of this study must be considered within the limitations of the data and study design. Administrative claims provide a relatively inexpensive and timely resource for understanding real-world patterns of care unlike the controlled environment of clinical trial data. However, all observational research using claims data are subject to limitations inherent to the limitations of claims data. Claims data are collected for the purpose of payment and not research, and as such are subject to possible coding errors, e.g. coding for the purpose of rule-out rather than actual disease and undercoding. While these limitations do not significantly reduce the strength of the study and its results, they must be considered during interpretation.
There are limitations to the generalizability of the results of this study. Our study included a sample of patients from 8 state Medicaid programs but cannot be generalized to the total Medicaid population as the geographical representation of the included states is unknown for data confidentiality reasons. Nevertheless, the similarity of our estimates to the study by Marton et al that used data from 2 states representing 25% of the Medicaid population provides some level of confidence in estimating the costs of COPD to Medicaid.
Other limitations pertain to study design specifically adjusting for only unrelated conditions in the incremental cost approach. Unrelated conditions were obtained from the CCI rather than a comprehensive listing of conditions unrelated to COPD. However, the CCI has been shown in several studies to be predictive of increased costs, and thus our objective of accounting for excess costs from unrelated conditions may be sufficiently addressed by only incorporating the conditions included in the CCI.
The study limitations however, do not mitigate the increased economic and clinical burden imposed due to COPD on the Medicaid population. From a respiratory disease perspective, asthma has primarily been assumed to be a key target for disease management efforts in Medicaid due to the predominant patient population demographic in Medicaid –women and children (Citation23). However, the current study highlights for clinicians and policymakers alike that COPD is an equally important respiratory condition for the Medicaid population to which disease management efforts must be targeted. There are numerous studies documenting the burden of COPD in the US in different populations (2, 4, 5, 19, 20) but relatively few in the Medicaid population (Citation3, Citation13). The results of the study provide an important benchmark for policymakers on the costs of COPD in the Medicaid population before the advent of Affordable Care Act.
Conclusions
In conclusion, this analysis has shown that the economic burden imposed by COPD on Medicaid is substantial. Non-dual-eligible patients in Medicaid incur 7-fold higher costs due to COPD compared to dual-eligible patients. On average, incremental costs were more than double for Blacks compared to Whites. Key reforms in healthcare, particularly the Affordable Care Act, might alter the Medicaid cost relationships identified in this study (Citation11). The latter are most likely to be non-dual-eligible patients, thus increasing the total costs for Medicaid.
Declaration of Interest Statement
Funding for this study and resultant publication was provided by GlaxoSmithKline (GSK ADC113900).
Author Anand A. Dalal is an employee of GSK and owns company stock. Author Anna O. D'Souza is an employee of Xcenda, LLC, a consulting company that received funding from GSK to conduct this research. Author Manan Shah was also employed by Xcenda at the time of this study and manuscript preparation. He is now employed by Bristol-Myers Squibb. Author Amol D. Dhamane was also employed by Xcenda at the time of this study and manuscript preparation. He is now employed by Boehringer-Ingelheim.
The authors alone are responsible for the content and writing of the paper.
Acknowledgments
The authors thank Gagan Jain, PhD, formerly employed at Xcenda, for medical writing services, which were compensated by GSK.
References
- Global Initiative for Chronic Obstructive Lung Disease. Pocket guide to COPD diagnosis, management, and prevention: updated 2010. Available from: http://www.goldcopd.org. Accessed Sep 23, 2011.
- Mapel DW, Hurley JS, Frost FJ, Petersen HV, Picchi M, Coultas DB. Health care utilization in chronic obstructive pulmonary disease. A case-control study in a health maintenance organization. Arch Intern Med 2000; 160:2653–2658.
- Marton JP, Boulanger L, Friedman M, Dixon D, Wilson J, Menzin J. Assessing the costs of chronic obstructive pulmonary disease: the state Medicaid perspective. Res Med 2006; 100:996–1005.
- Menzin J, Boulanger L, Marton JP, The economic burden of chronic obstructive pulmonary disease (COPD) in a U.S. Medicare population. Res Med 2008; 102:1248–1256.
- Foster T, Miller JD, Marton JP, Caloyeras JP, Russell MW, Menzin J. Assessment of the economic burden of COPD in the U.S.: a review and synthesis of the literature. COPD 2006; 3(4): 211e8.
- Simoni-Wastila L, Blanchette CM, Qian J, Burden of chronic obstructive pulmonary disease in Medicare beneficiaries residing in long-term care facilities. Am J Geriatr Pharmacother 2009 Oct; 7(5):262–270.
- California Health Care Almanac. Health Care Costs 101. Available from: http://www.chcf.org/∼/media/MEDIA%20LIBRARY%20Files/PDF/H/PDF%20HealthCareCosts11.pdf. Accessed Sep 23, 2011.
- Carr RM, Christiansen B, Jehn L, Matitz D. Meeting the challenge of tobacco use within the Medicaid population. Available from: http://www.ncbi.nlm.nih.gov/pubmed?term = %22WMJ+%3A+official+publication+of+the+State+Medical+Society+of+Wisconsin%22[Jour]+AND+2001[pdat]+AND+Carr[author]&cmd = detailssearch. WMJ 2001; 100(3):59–66.
- Eisner MD, Blanc PD, Omachi TA, Socioeconomic status, race and COPD health outcomes. J Epidemiol Community Health 2011 Jan; 65(1):26–34.
- The Urban Institute and Kaiser Commission on Medicaid and the Uninsured estimates based on data from Medicaid Statistical Information System (MSIS) reports from the Centers for Medicare and Medicaid Services (CMS). 2010. Available from: http://www.statehealthfacts.org/comparetable.jsp?ind = 198&cat = 4. Accessed Sep 23, 2011.
- Hislop RM. Health reform and Medicaid expansion. July 2010. Available from: http://www.healthcarereformmagazine.com/article/health-reform-and-medicaid-expansion.html. Accessed Jan 12, 2012.
- Centers for Medicare and Medicaid Services. Medicaid Coverage of Medicare Beneficiaries (Dual-eligibles) at a Glance. Available from: http://www.cms.gov/Outreach-and-Education/Medicare-Learning-Network-MLN/MLNProducts/downloads/Medicare_Beneficiaries_Dual_Eligibles_At_a_Glance.pdf. Accessed Jan 23, 2012.
- Shaya FT, Maneval MS, Gbarayor CM, Burden of COPD, asthma, and concomitant COPD and asthma among adults: racial disparities in a Medicaid population. Chest 2009 Aug 13; 6(2):405–411.
- Barnett A, Birnbaum H, Cremieux P, Fendrick MA, Slavin M. The costs of cancer to a major employer in the United States: a case-control analysis. Am J Manag Care 2000; 6:1243–1251.
- Tunceli O, Wade R, Gu T, Bouchard JR, Aagren M, Luo W. Cost of diabetes: comparison of disease-attributable and matched cohort cost estimation methods. Curr Med Res Opin 2010 Aug; 26(8):1827–1834.
- Bureau of Labor Statistics. Measuring price change for medical care in the CPI. U.S. Department of Labor. Available from: http://www.bls.gov/cpi/cpifact4.htm. Accessed Sep 23, 2011.
- Li Z, Mahendra G. Using “recycled predictions” for computing marginal effects. Paper 272–2010. SAS Global Forum 2010. Available from: http://support.sas.com/resources/papers/proceedings10/272–2010.pdf. Accessed May 1, 2012.
- Summer L, Hoadley J, Hargrave E. The Medicare Part D Low-Income Subsidy Program: Experience to Date and Policy Issues for Consideration. The Henry J. Kaiser Family Foundation. Publication # 8094. September 2010. Available from: http://www.kff.org/medicare/upload/8094.pdf. Accessed Jan 9, 2012.
- Grasso ME, Weller WE, Shaffer TJ, Diette GB, Anderson GF. Capitation, managed care, and chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1998; 158(1):133–138.
- Miller JD, Foster T, Boulanger L, Costs of chronic obstructive pulmonary disease (COPD) in the US: an analysis of Medical Expenditure Panel Survey (MEPS) Data. American Thoracic Society International Conference; May 21–26, 2004; Orlando, FL.
- Dransfield MT, Bailey WC. COPD: racial disparities in susceptibility, treatment, and outcomes. Clin Chest Med 2006; 27:463–471.
- Laditka JN, Laditka SB. Race, ethnicity and hospitalization for six chronic ambulatory care sensitive conditions in the USA. Ethn Health 2006 Aug; 11(3):247–263.
- Gillespie JL, Rossiter LF. Medicaid disease management programs findings from three leading US state programs. Dis Manage Health Outcomes 2003; 11(6):345–361.
APPENDIX
A1. Match Process for COPD and Non-COPD Cohort
Table
A2. All-cause, adjusted healthcare resource use per year for Medicaid beneficiaries