You are here
Burden of multiple sclerosis on direct, indirect costs and quality of life: National US estimates
Multiple Sclerosis and Related Disorders, 2, 3, pages 227 - 236
MS imposes a significant burden on patients, caregivers, employers, and the healthcare system.
To comprehensively evaluate the US MS burden using nationally representative data from the Medical Expenditure Panel Survey.
We identified non-institutionalized patients aged ≥18 with MS (ICD-9 code 340) from 1998 to 2009 and compared them to individuals without an MS diagnosis (non-MS) during the interview year. The cohorts were compared using multivariate regression on direct costs, indirect costs (measured in terms of employment status, annual wages, and workdays missed), and health-related quality of life (HRQoL; measured using Short Form 12, SF-6 Dimensions, and quality-adjusted life years [QALYs]).
MS prevalence was 572,312 (95% CI: 397,004, 747,619). Annual direct costs were $24,327 higher for the MS population (n=526) vs. the non-MS population (n=270,345) (95% CI: $22,320, $26,333). MS patients had an adjusted 3.3-fold (95% CI: 2.4, 4.5) increase in the odds of not being employed vs. non-MS individuals and a 4.4-fold higher adjusted number of days in bed (95% CI 2.97, 6.45). On average, MS patients lost 10.04 QALYs vs. non-MS cohort.
MS was associated with higher healthcare costs across all components, reduced productivity due to unemployment and days spent in bed, and lower HRQoL.
- US multiple sclerosis (MS) prevalence was 572,312 (95% CI: 397,004, 747,619).
- Annual direct costs were $24,327 higher for the MS population (n=526).
- MS patients lost on average 10.04 quality-adjusted life years vs. non-MS cohort.
- MS was associated with higher direct and indirect costs and lower HRQoL.
Keywords: Multiple sclerosis, Burden of illness, Direct costs, Quality of life, Reduced productivity, MEPS.
MS, a chronic and debilitating inflammatory autoimmune disorder of the central nervous system, afflicts approximately 400,000 people in the United States (Zwibel and Smrtka, 2011 and Goldberg et al, 2009) and 2.1 million people worldwide ( Zwibel and Smrtka, 2011 ). MS is among the most common contributors to neurologic disability in young and middle-aged adults ( Phillips, 2004 ) with detrimental effects on patients' productivity and health-related quality of life (HRQoL). Peak age of onset is approximately 30 years ( Olofsson et al., 2011 ). Because patients with MS typically have a long life span and increased healthcare utilization ( Prescott et al., 2007 ), MS imposes a significant burden on patients and their families/caregivers, employers, and the healthcare system.
Average total (direct and indirect) costs in a study of 1909 patients identified from the North American Committee on Multiple Sclerosis Patient Registry (NARCOMS) were estimated at $47,215 per patient per year, 63% for direct medical and nonmedical costs ($29,634) and 37% for total indirect costs including early retirement and informal care ($17,581) ( Kobelt et al., 2006b ). Ivanova et al. (2012) reported annual average total (direct and indirect) costs, excluding costs of disease modifying treatments (DMTs), ranging from $14,095 for DMT-adherent patients to $16,638 for those with poor adherence to DMTs. In a retrospective claims-based analysis of privately insured US employees with MS, total indirect costs (2006 US dollars) were more than 4 times higher for MS patients than for matched controls (P<0.0001) ( Ivanova et al., 2009 ).
A less widely reported component of the costs of MS is the additional burden on HRQoL ( Orlewska, 2006 ), commonly referred to as “intangible costs”. Many studies assessing MS burden do not include HRQoL in their estimates due in part to challenges in monetizing HRQoL measures ( Orlewska, 2006 ). One HRQoL-informed measure that has been monetized is quality-adjusted life years (QALYs), a common metric used in cost-effectiveness research ( Neumann and Greenberg, 2009 ). QALYs include two components: an HRQoL-informed utility weight (quality-adjusted component) and survival (life years component). Utility values are anchored at 0 for death and 1 for perfect health. For each time interval of survival, the utility weight is multiplied by the survival time and summed overall survival times thus producing QALYs. The QALY therefore allows for the translation of disparate risks and benefits into one common outcome measure that can be compared within and across diseases.
As the incidence of MS is projected to grow 5% by 2013 ( National Multiple Sclerosis Society, 2011 ), and as new DMTs are expected to be introduced to the market that could influence treatment choice and cost, further examination of the economic and intangible burden of MS would add to the platform for evidence-based decisions.
The primary objective of the present study was to conduct a comprehensive evaluation of the current burden of MS in the United States using nationally representative data from the Medical Expenditure Panel Survey (MEPS), a public-access, large-scale database that links direct cost information with information on productivity and HRQoL. MEPS also makes adjustments for socioeconomic and demographic characteristics. This is the first time this unique dataset has been utilized to evaluate the comprehensive burden of MS including direct costs, indirect costs, and HRQoL.
2. Materials and methods
2.1. Data source
MEPS tracks individual and household demographic, socioeconomic, and health-related characteristics ( Agency for Healthcare Research and Quality, 2013 ), providing a nationally representative sample of the US civilian non-institutionalized population (the study population of inference). The Household Component of the Medical Expenditure Panel Survey (MEPS-HC) is a nationally representative survey of the US civilian noninstitutionalized population. The sampling frame is drawn from respondents to the National Health Interview Survey, which is conducted by the National Center for Health Statistics ( Cohen et al., 2009 ). The MEPS-HC collects data from a nationally representative sample of households through an overlapping panel design. A new panel of sample households is selected each year, and data for each panel are collected for two calendar years. The two years of data for each panel are collected in five rounds of interviews that take place over a 2.5 year period. This provides continuous and current estimates of health care expenditures at both the person and household level for two panels for each calendar year. To provide estimates that are representative of a national US population, the MEPS-HC panels have oversampled subgroups of individuals such as Hispanics, African-Americans, Asians, low-income households, and those likely to incur high medical expenditures. The most recent MEPS-HC panels have data on approximately 15,000 families and 35,000 individuals ( Agency for Healthcare Research and Quality, 2013 )
At the time of the analysis, the most recent available file was for 2009. We combined the 1998–2009 MEPS full-year consolidated, medical, and pharmacy utilization data files to generate an analytical cohort with robust sample size. The Colorado Multiple Institutional Review Board considers this research to be exempt from requiring board approval since MEPS is a publicly available data source.
2.2. Independent variables
The MS study population was defined as having a medically coded diagnosis with ICD-9 code 340 during the interview year. Variables were created from the MEPS Household Component, which comprises questionnaires fielded to household members and their medical providers. We limited the study populations to patients aged ≥18 years. This constituted a prevalent cohort of adults with MS (n=526) between the years 1998 and 2009. The comparison population comprised all individuals who did not have a MS diagnosis (non-MS,n=270,345 for non-weighted sample) during the interview year. Therefore, the comparison population included non-institutionalized individuals from the full disease states excluding MS.
The Chronic Conditions Indicator (2012) was used to determine total comorbidity burden for MS and non-MS study populations.
We created indicator variables for each mutually exclusive demographic category: gender; age in years (18–34, 35–49, 50–64, 65–79, ≥80); race (White, Black, American Indian, other); ethnicity (Hispanic, non-Hispanic); marital status (married, widowed, divorced/separated, never married); educational status (no degree, high school/GED, bachelors, masters, or Ph.D.); family income category (using United States Census Bureau (2012) definitions for poor, near poor, low income, middle income, high income); and health insurance status (public, private, uninsured).
2.3. Outcome variables
All-cause and MS-specific healthcare utilization were determined for annual numbers of outpatient visits, emergency room visits, inpatient days, and annual number of fills for prescription medications excluding refills for non-DMTs. Prescription medications were annual counts of all prescribed medications purchased during the interview year including refills. Due to limitations of MEPS in terms of how injectable medications are captured (i.e., J codes for injectable medications are not captured, and data on DMTs may be missing for covered medications), we did not include DMTs in the prescription medication utilization category. We used estimates from the literature to calculate the average annual DMT cost per MS patient as follows: 85% of MS patients have relapse-remitting MS (RRMS), ( National MS Society, 2013 ) approximately 67% of RRMS patients are treated with DMTs ( Morrow, 2007 ) with an average adherence of 75% ( Steinberg et al., 2010 ). With an average 2011 annual cost of DMTs of approximately $40,000 ( Yamamoto and Campbell, 2012 ), the DMT annual cost estimate per MS patient=0.85×0.67×0.75×$40,000=$17,085.
Total healthcare expenditures included direct payments to providers for all healthcare utilization (as defined above) during the year, including out-of-pocket payments and payments by private insurance, Medicaid, Medicare, and other sources, adjusted to 2011 US dollars using the medical care component of the Consumer Price Index ( US Department of Labor, 2012 ). All costs were converted to 2011 US dollars ( US Department of Labor, 2012 ) in order to interpret (a) the results in the most recent calendar year the data were analyzed (which was 2011), (b) to represent aggregate costs across different years and, (c) to compare costs across the study cohorts. Total all-cause expenditures included all non-DMT prescriptions and medical utilization from MEPS as well as our literature-based estimate of DMT annual cost.
Measures of productivity (employment, wages, and missed days) were created to assess indirect costs. We assessed the annual number of days spent in bed due to illness or injury for the entire study population. Respondents who indicated zero wages during each of three questionnaire rounds in a given year were considered to be “not employed.” “Not employed” is different from a standard “unemployed” definition in that some “not employed” people are not seeking employment and therefore have zero wages. For the entire study population and for the employed (wage>$0) subpopulation, we defined total annual wage income as annual wages earned in 2011 US dollars for individual respondents.
HRQoL measures included the Short Form 12 (SF-12) Health Status Measure, Short Form 6 Dimensions (SF-6D), and quality-adjusted life years (QALYs). The SF-12 has two validated subscales for physical and mental health status ( Ware Jr. et al., 1996 ), which are normalized to general community populations and internationally validated, with mean value of 50 and standard deviation of 10; group means as well as patient scores below 47 are interpreted as clinically lower than average health status ( Ware Jr. et al., 1996 ). The SF-6D provides a means for using the SF-12 in economic evaluation by estimating a preference-based single index measure for health (preference score or utility) from these data using general population values ( Brazier and Roberts, 2004 ). Preference scores (utility) derived from the SF-6D are anchored at 1.0 for perfect health and 0.0 for death. The quality-adjusted component or utility decrement was estimated from the difference in MS vs. non-MS SF-6D preference scores. Life years were estimated from average survival of MS patients compared with the general population ( Kingwell et al., 2011 ) QALY estimates for MS and non-MS cohorts were calculated by multiplying SF-6D preference score (utility) by average survival.
2.4. Statistical analyses
Descriptive statistics for clinical and demographic characteristics were calculated using Chi-square tests of proportions (survey design-basedF-test) for variable distributions across MS and non-MS study populations. The US proportion totals were estimated using the most recently available 2008 and 2009 MEPS data.
Adjusted analyses for count-based outcomes measures, including healthcare utilization and days spent in bed, were modeled using negative binomial regression (incident rate ratios for MS vs. non-MS). Continuous measures including healthcare cost categories, wages, SF-12 mental and physical subscores, and the SF-6D preference score were modeled using ordinary least squares regression (marginal effect for MS vs. non-MS). The binary unemployed variable was modeled using logistic regression (odds ratio for MS vs. non-MS). All adjusted analyses, except those evaluating employment, included the following covariates in the regression models: gender, race, ethnicity, region, family income, age, insurance status, education, number of chronic conditions other than MS, and survey year. Income was excluded from the not employed equation.
As a sensitivity analysis, we ran regression models without adjusting for chronic conditions other than MS to provide a less conservative estimate of burden, since many comorbid conditions for the MS population are a result of MS and not vice versa (e.g., anxiety, depression, fatigue, and sleep disturbances) (Mills and Young, 2011, Wood et al, 2013, Li et al, 2012, and Heesen et al, 2006).
We used STATA version 11 (StataCorp LP, College Station, Texas, USA) for all statistical analyses.
To adjust for the complex sample design, MEPS sampling weights were applied in all analyses to obtain nationally representative estimates and robust standard errors.
There were 526 respondents (0.21% survey sample weight-adjusted) with a verified MS diagnosis in the 12-year cohort, (1998–2009), with an estimated annual prevalence of 572,312 (95% confidence interval [CI] 397,004, 747,619) in the United States. MS and non-MS cohorts were similar on the demographic dimensions of race and education ( Table 1 ). Compared with the non-MS population, the MS population was more likely to be female (77% vs. 52%), older by about 4 years on average, non-Hispanic (97% vs. 88%), divorced or separated (26% vs. 13%), have lower income status, have 1.2 more chronic conditions on average, and was more likely to have public insurance (33% vs. 15%).
|Average adult US population (2008–2009) b|
|Multiple sclerosis||No multiple sclerosis||P-value a||Multiple sclerosis||No multiple sclerosi|
|Age (mean, SE)||49.97||0.93||45.82||0.13||<0.01|
|Highest degree completed||0.30|
|High school, GED||295||53.44||131,333||50.42||308,759||114,000,000|
|Family income category||0.01|
|Number of chronic conditions||<0.01|
|Chronic conditions indicator (mean, SE)||2.68||0.16||1.51||0.01||<0.01||–||–|
a Pearson survey weighted F-test for the difference between multiple sclerosis and no multiple sclerosis.
b Average adult MEPS weight adjusted US population (2008–2009).
All-cause healthcare utilization was significantly higher for the MS cohort vs. the non-MS cohort ( Table A.1 ). These utilization trends were statistically significantly higher for inpatient and outpatient visits, emergency room visits, and medications.
By summing over all-cause expenditures and adjusting for measured characteristics, annual direct medical expenditures were 5.1 times higher, or $24,327 greater (95% CI: $22,320, $26,333), for the MS population than for the population without MS after controlling for all chronic conditions ( Table 2 ). The cost components with the greatest difference between the groups were DMT costs, outpatient visits, non-DMT prescription medications, and inpatient hospitalizations. For example, adjusted outpatient visit annual costs were $3443 (95% CI $2442, $4445) higher for the MS population.
|Unadjusted a||Adjusted b|
|Multiple sclerosis (N=526)||No multiple sclerosis (N=270,345)||Multiple sclerosis vs. no multiple sclerosis|
|Mean||SE||Mean||SE||Marginal effect c||95% CI|
|Outpatient visits d||$1938||$283||–||–||–||–|
|Emergency room visits||$70||$30||–||–||–||–|
|Prescription medications including refills (excluding DMT's) e||$1110||$152|
|Total MS expenditures excluding prescriptions||$3220||$465||–||–||–||–|
|Total MS health care expenditures (excluding DMT's)||$8150||$672||–||–||–||–|
|All cause expenditures|
|Outpatient visits f||$6204||$542||$1667||$22||$3443||($2442, $4445)|
|Emergency room visits||$290||$49||$154||$3||$64||(−$27, $155)|
|Inpatient days||$4018||$768||$1568||$31||$1546||($39, $3054)|
|Other expenditures||$357||$92||$45||$2||$290||($109, $470)|
|Literature-based DMT cost estimate g||$17,085||–||–||–||$17,085||–|
|Prescription medications including refills (excluding DMT's) e||$3599||$346||$941||$10||$1834||($1242, $2427)|
|Total expenditures excluding prescription medications||$11,367||$1003||$3791||$43||$5407||($3523, $7292)|
|Total all cause expenditures||$32,051||$1125||$4732||$48||$24,327||($22,320, $26,333)|
a All expenditure categories except dental, vision, and total home health non-agency are significantly different at P<0.05 based on Wald F-test for the difference between multiple sclerosis and no multiple sclerosis.
b Adjusted for gender, race, ethnicity, region, family income, age, insurance status, education, number of chronic conditions other than multiple sclerosis, and survey year.
c Marginal effect from OLS regression models.
d MS-specific outpatient visits include medical provider visits and hospital outpatient visits.
e Prescription medications including refills excludes MS-related disease-modifying therapies (DMT's).
f All-cause outpatient visits include medical provider visits, hospital outpatient visits, total home health agency, and total home health non-agency.
g DMT cost estimate from literature. No uncertainty was built into this cost estimate.
The average unadjusted proportion of individuals not employed during a year was 60% for MS and 27% for non-MS, and the MS population had an adjusted 3.3-fold (95% CI: 2.4, 4.5) increase in the odds of not being employed during the year vs. the non-MS population ( Table 3 ). The adjusted number of days spent in bed (IRR of 4.4; 95% CI 2.97, 6.45) was significantly higher for the MS population. Total annual adjusted income was lower for the MS population by an average of $6767 (95% CI: $11,422, $2111), but the annual adjusted income for the employed subpopulation was not statistically different for the MS vs. non-MS populations.
|Unadjusted a||Adjusted b|
|Multiple sclerosis||No multiple sclerosis||Multiple sclerosis vs. no multiple sclerosis|
|Mean||SE||Mean||SE||Adjusted coefficient||95% CI|
|Adult population (18 and over)|
|Not employed (percentage) c||60%||–||27%||–||3.30||(2.44, 4.47)|
|Total number of days spent in bed d||29.21||3.73||5.32||0.09||4.38||(2.97, 6.45)|
|Total annual wage income e||$19,526||$2933||$28,699||$258||−$6767||(−$11,422, −$2111)|
|All employed adults (18 and over)|
|Total annual wage income e , f||$47,859||$5857||$39,448||$273||$5080||(−$4354, $14,513)|
a All estimates were significant at P<0.05 based on the Wald F-test for the difference between multiple sclerosis and no multiple sclerosis.
b Adjusted for gender, race, ethnicity, region, family income, age, insurance status, education, number of chronic conditions other than multiple sclerosis, and survey year.
c Odds ratio from one logistic regression model.
d Incidence Rate Ratio (IRR) from negative binomial regression models.
e Marginal effect from OLS regression model; wage income expressed in 2011 US dollars.
f All adults 18 and over with any paid employment through the year (i.e., any wage greater than zero).
Generic HRQoL, preference scores, and QALYs were all significantly lower for the MS population ( Table 4 ). The adjusted SF-12 physical summary score was 11.89 lower (95% CI: 13.44, 10.35), on average, for the MS population, whereas the mental summary score was 2.94 lower (95% CI: 4.28, 1.61). The average expected QALYs for those with MS was 10.04 (95% CI: 10.52, 9.44) quality-adjusted years fewer than for those without MS.
|Multiple sclerosis||No multiple sclerosis||Adjusted analysis|
|Mean (95% CI)||Mean (95% CI)||Marginal eff.a and b (95% CI)|
|SF-12 physical summary score c||34.10 (32.37, 35.83)||49.57 (49.45, 49.68)||−11.89 (−13.44, −10.35)|
|SF-12 mental summary score c||45.74 (44.19, 47.30)||51.00 (50.90, 51.09)||−2.94 (−4.28, −1.61)|
|Health-related quality of life preference score (SF-6D) c||0.632 (0.613, 0.651)||0.794 (0.792, 0.795)||−0.111 (−0.127, −0.094)|
|Expected survival in years from median age of 31 (age of MS onset)||47.5 (45.3, 49.7)||53.5 (51.3, 55.7)||−|
|Average expected QALYs d||30.02 (28.63, 31.41)||42.50 (40.73, 44.22)||−10.04 (−10.52, −9.44)|
a Adjusted for age, gender, race, marital status, region, education, family income, insurance, survey year, and chronic conditions other than multiple sclerosis.
b All adjusted estimates significant at P<0.05 when compared to the reference group (no multiple sclerosis).
c Marginal effects from linear regression models.
d Average QALYs based on median age of 31 years at MS onset with life expectancy of 47.5 (45.3, 49.7) for MS and 53.5 (51.3, 55.7) for no MS ( Kingwell et al., 2011 ). Marginal effect estimated from adjusted preference score as well as expected survival.
3.1. Sensitivity analysis results
When other chronic conditions were not adjusted for, average total all-cause expenditures were $25,934 higher (compared to $24,327 higher when adjusting for other chronic conditions) for the MS vs. the non-MS population. Similarly, the SF-12 physical summary score was 13.57 lower (compared to 11.89 lower), the SF-12 mental summary score was 3.88 lower (compared to 2.94 lower), and the HRQoL preference score (SF-6D) was 0.132 lower (compared to 0.111 lower) for the MS population than for the non-MS population.
We observed a higher annual prevalence ofN=572,312 (0.21%) of MS in our MEPS-identified study population than the commonly cited national estimate of 400,000 persons (0.13% of an estimated US population of 300 million) ( Zwibel and Smrtka, 2011 ; Goldberg et al., 2009 ; National Multiple Sclerosis Society, 2011 ; Noonan et al., 2002 ; Loma and Heyman, 2011 ; Asche et al., 2010 ; Vlahiotis et al., 2010) . Past primary data studies on the prevalence for MS vary from 58 to 177 per 100,000 population (0.06–0.18%) in the United States (Noonan et al, 2002, Baum and Rothschild, 1981, Anderson et al, 1992, Williamson and Henry, 2004, Mayr et al, 2003, and Noonan et al, 2010), but were mainly regional cohorts or of specific subpopulations that are not necessarily representative of the US. A 2010 review of the global impact of MS estimated the prevalence within the US to be 135/100,000 (0.14%) ( Trisolini et al., 2010 ). While anecdotal reports suggest that the prevalence of MS within the US may be higher, our study is the only current empirical evidence of the most recent estimate (2008–2009) of the prevalence of MS within the US.
Female MS patients outnumbered male MS patients by 3:1 in our study, which is consistent with a recent estimate by the National MS Society stating that MS is 2–3 times more common in women ( National Multiple Sclerosis Society, 2011 ) and with current research reporting that the female to male gender ratio of MS patients has been increasing over the past few decades (Sellner et al, 2011 and Orton et al, 2006) from the commonly cited 2:1 ratio reported in past studies (Olofsson et al, 2011 and Casetta et al, 2009).
Comparisons of our MS and non-MS populations revealed higher healthcare resource use across all types of utilization and higher healthcare costs across all cost components, after adjusting for relevant covariates. We also observed reduced productivity in the MS population due to days spent in bed and unemployment. HRQoL preference scores and average expected QALYs were significantly lower for the MS population. In summary our study was able to provide both overall and component specific costs for both the direct and indirect costs for MS patients. In doing so our findings provided updated estimates on the economic burden of this disease adjusted for the most recent year (2011) in which the analysis was conducted. However, there are some limitations on how our findings may impact future costs, primarily as they relate to DMT costs. We only used a literature based estimate to capture DMT costs which does not account for the newer oral drugs that have been approved since 2010 (namely fingolimod, teriflunomide and dimethyl fumarate). Thus future direct costs are likely to be higher estimates than our findings due to the inclusion of these drugs and their associated prescription costs which are between $50,000–60,000 per patient per year.
This study's total MS-attributable direct costs (annual average per patient of $24,327) is consistent with past literature. Rajagopalan et al. (2011) reported indirect costs related to short-term disability ranging from $1627 to $3164 prior to DMT initiation, and from $1131 to $2603 after DMT initiation. Productivity losses due to reduced work time or early retirement translated into a cost of $18,436 in the study of Kobelt et al., 2006b . We did not calculate dollar values for MS patients' missed workdays or monetize short-term disability costs, and therefore cannot provide comparisons for these findings. We did, however, assess total annual wages and found that, on average, MS patients had $6767 less income than non-MS individuals, which primarily appears to impact those who were not employed.
We estimated a loss of 10.04 QALYs for the MS population vs. the non-MS population as well as a utility decrement of 0.111 per person year. Our utility decrement may be compared with the annual utility decrement of 0.26 per person in a US study ( Kobelt et al., 2006b ) and a mean decrement of 0.27 per patient in a European study, ( Kobelt et al., 2006a ) both conducted by Kobelt et al. The mean utility score in the MS sample from their US study was 0.698, whereas our mean utility score for the MEPS MS sample was 0.632 (SF-6D preference score). The estimates in utility score for the MS population are comparable, but we found our average population without MS utility scores to be 0.794, whereas Kobelt et al. assumed a utility estimate for their control group to be much closer to 1.0 (perfect health). Therefore, the average utility decrement per patient year attributed to MS was much larger in their study compared to our MEPS findings.
The projected average lifetime QALYs lost due to MS at 10.04 has a meaningful impact and is an important component of overall burden since willingness-to-pay values for one QALY are typically thought to be in the $50,000–100,000 range (Wundes et al, 2010 and Lee et al, 2012). Willingness-to-pay is defined as the maximum amount that a purchaser or payer would pay for a good, service, or reduction in the risk of death and illness. Willingness-to-pay is a method of measuring the value a purchaser or payer places on a good, service, or reduction in the risk of death or illness by estimating the maximum dollar amount a purchaser or payer would pay in order to obtain the good, service, or reduction in risk ( Gold et al., 1996 ). When used in terms of a QALY, willingness-to-pay is referring to how much a purchaser or payer would be willing to pay for one perfect year of life.
The management of MS is complicated by the fact that costs related to common comorbid conditions, such as depression, anxiety, fatigue, and sleep disturbances, (Mills and Young, 2011, Wood et al, 2013, Li et al, 2012, and Heesen et al, 2006) must be considered in addition to the costs of MS itself. Thus, without controlling for any chronic conditions that may result from the disease itself, we may in fact obtain a more accurate estimate of direct costs for MS. Our sensitivity analysis confirmed this estimate to be $25,934, or 6.6% higher than the estimate of $24,327 after controlling for all chronic conditions. The HRQoL scores showed a similar trend. Although this is a more “real” estimate of direct costs, it may still be conservative as only chronic conditions were excluded in the sensitivity analysis and symptoms such as fatigue and bowel and bladder dysfunction, which also have costs associated with their treatment, were not excluded in the estimations.
To provide an additional comparative context for the interpretation of our findings, we compared our MS population to patient populations with other debilitating chronic conditions, including rheumatoid arthritis, diabetes, inflammatory bowel disease, depression, acute coronary syndrome, and asthma. MS was associated with higher annual lost wages per person than any of the compared conditions ( Fig. A.1 a). Similarly, the adjusted decrement in the SF-12 physical summary score was highest for MS even when compared to rheumatoid arthritis ( Fig. A.1 b). The adjusted decrement in the SF-12 mental summary score for MS was second only to depression ( Fig. A.1 c). The adjusted utility decrement was once again highest for MS, followed by depression and rheumatoid arthritis ( Fig. A.1 d).
Finally, the findings of our studies are limited to a non-institutionalized population with a mean age of 50 years. The MEPS database also only captures non-institutionalized patients so we were not able to include patients in nursing homes or hospices in our study as well. It is estimated that approximately 20–25% of patients with MS will need long-term care during the course of their disease, and an estimated 5% will reside in a in a long term nursing facility ( Stopl-Smith et al., 1998 ). The mean age of MS patients in a 2002 analysis of nursing home residents was older than in our analysis (72 years vs. 50 years) ( Buchanan et al., 2006 ). Making the assumption that nursing home patients are likely to have higher EDSS scores than noninstitutionalized patients, we can approximate the impact on direct and indirect costs on these patients. Patwardhan et al. (2005) found that both direct and indirect costs rise continuously with increasing EDSS category. The rise in indirect costs appears at lower EDSS scores in other words, indirect costs rose at lower EDSS scores compared to direct costs. So we can speculate that in a non-institutionalized population, we will expect to see higher direct costs but lower indirect costs since the majority of indirect costs are due to the loss in productivity during the ambulatory years of a patient's life.
With regard to the impact of our results on children with MS, we would not have been able to include any significant amount of children in our study as the sample sizes in the MEPS database were less than 20 patients. Accordingly it is uncertain what the impact of including children in our analysis would be on direct medical costs given these very limited sample sizes. Productivity loss (indirect costs) will not be applicable to children directly but could influence their parents and caregivers substantially. For the reasons previously mentioned we were not able to include children and subsequently their caregiver burden. A review of the economic burden of undertaken by families caring for a child with disabilities by the Canadian Office of Health Technology Assessment found that indirect costs between 5% and 12% of a family's income can be lost in the provision of this care ( Anderson et al., 2007 ). Thus we can speculate the indirect costs in the form of caregiver productivity losses, could pose a substantial burden for children with MS.
Certain limitations must be considered when interpreting these study results, such as the fact that the MEPS dataset does not include information on disease severity, which could affect the estimates. As DMT utilization was rarely captured in the MEPS dataset, we excluded the costs of DMTs from our prescription expenditures but included a literature-based estimate of annual DMT costs for completeness. This study's estimates of MS burden are likely conservative for at least two reasons: (1) our analysis is limited to the ambulatory outpatient MS population, as the MEPS dataset does not encompass institutionalized populations (an estimated 5–10% of MS patients require extended or permanent nursing home care ( Buchanan et al., 2010 ; Northrop and Frankel, 2009 )); and (2) we adjusted analyses for any other chronic conditions, whereas some chronic conditions can be considered to have been caused by MS (e.g., depression, anxiety) and therefore could be built into burden estimates. We acknowledge that there are many possible reasons for why the prevalence as well as the gender ratio may differ across studies. We consider our study to be a rigorous observational study and therefore place it in the context of other epidemiologic studies by providing additional references for such studies.
Finally, MS-specific Expanded Disability Status Scale or patient-determined disease step scores are not available in the MEPS dataset, so progression of disability and associated costs could not be measured.
Based on our findings, MS prevalence is likely higher than previous conservative US estimates, especially among women. MS patients have higher healthcare costs and utilization than their non-MS counterparts, and are at higher risk for not being employed and having lower HRQoL. Further, MS direct costs, indirect costs, and intangible burden in the MS population are higher across most domains as compared with other debilitating chronic conditions.
This project was supported by an unrestricted grant from Biogen Idec. The content is solely the responsibility of the authors.
Study concept and design—all authors. Acquisition of the data—Dr. Ghushchyan. Statistical analysis and data interpretation—primarily conducted by Dr. Ghushchyan with support from Dr. Campbell, Mr. McQueen, and Dr. Nair. Critical revision of the manuscript for important intellectual content—all authors. Study supervision—Dr. Campbell and Dr. Nair.
Conflict of interest
Dr. Campbell is a consultant for Amgen Inc., AstraZeneca Inc., Biogen Idec, and for VeriTech Inc., and has received grant support from the Centers for Disease Control and Prevention and the Agency for Health Care Research and Quality.
Dr. Ghushchyan is a consultant for Janssen Services and Biogen Idec. and has received grant support from Amgen Inc, Janssen Services and Biogen Idec.
Mr. McQueen is a consultant for Biogen Idec and VeriTech Inc. and has received grant support from PhRMA Foundation.
Dr. Cahoon-Metzger is an employee and stock holder of Biogen Idec.
Dr. Livingston is an employee and stock holder of Biogen Idec.
Dr. Vollmer is a consultant for PRIME Education Inc., Projects in Knowledge Inc., Guidepoint Global, Esai Pharmaceuticals, Biogen-Idec, Sanofi-Aventis Group, Teva, Eli Lilly & Co, Xenoport, Novartis, Schering-Plough Biopharma, Ono, Elan Pharmaceuticals, and Acorda Pharma, and has received grant support from Acorda Therapeutics, Barrow Neurological Foundation, Biogen Idec, BioMS Medical, Biosite, Daiichi Sankyo, Elan Pharmaceuticals, Eli Lilly and Company, EMD Serono, Genentech Inc., Genzyme Corp., Merck Serono SA, Novartis, Ono Pharmaceuticals, PDL BioPharma, Pfizer, Inc., Sanofi-aventis, Teva Pharmaceuticals, The National Institutes of Health, The National Multiple Sclerosis Society, The Rocky Mountain Multiple Sclerosis Society, and Translational Genomics Research Institute.
Dr. Corboy is a consultant for Celgene and Teva Neurosciences, has received grant support from JDRF and Webb–Waring Institute, NIH, Teva, Pfizer, Novartis, BioMS, Genentech, Eli Lilly & Co, and Celgene, and is an investigator on trials funded by Novartis Pharmaceuticals, Eli Lily & Co, Celgene, and NIH.
Dr. Miravalle is a consultant for Bayer Healthcare Pharmaceuticals Inc, Biogen Idec, and Questcor Pharmacueticals, Inc; is on the speakers bureau for EMD Serono and Teva Neuroscience; and has received consulting fees from Healthcare Pharmaceuticals Inc., Biogen Idec, Questcor Pharmaceuticals Inc., Projects in Knowledge, and Teva Neuroscience.
Dr. Schreiner has no conflicts of interest to report.
Ms. Porter has no conflicts of interest to report.
Dr. Nair is a consultant for Janssen Services and Biogen Idec. and has received grant support from Amgen Inc., Janssen Services, Takeda Pharmaceuticals, Daiichi Sankyo, Biogen Idec, Novartis, Ethicon, Eli Lilly and Pfizer.
- et al., Agency for Healthcare Research and Quality. Medical Expenditure Panel Survey. 〈http://meps.ahrq.gov/mepsweb/〉 [accessed 08.02.13].
- Anderson et al., 2007 D. Anderson, S. Dumont, P. Jacobs, L. Azzaria. The personal costs of caring for a child with a disability: a review of the literature. Public Health Reports. 2007;122:3-16
- Anderson et al., 1992 D.W. Anderson, J.H. Ellenberg, C.M. Leventhal, S.C. Reingold, M. Rodriguez, D.H. Silberberg. Revised estimate of the prevalence of multiple sclerosis in the United States. Annals of Neurology. 1992;31(3):333-336
- Asche et al., 2010 C.V. Asche, M.E. Singer, M. Jhaveri, H. Chung, A. Miller. All-cause health care utilization and costs associated with newly diagnosed multiple sclerosis in the United States. Journal of Managed Care Pharmacy. 2010;16:703-712
- Baum and Rothschild, 1981 H.M. Baum, B.B. Rothschild. The incidence and prevalence of reported multiple sclerosis. Annals of Neurology. 1981;10(5):420-428
- Brazier and Roberts, 2004 J. Brazier, J. Roberts. The estimation of a preference-based measure of health from the SD-12. Medical Care. 2004;42:851-859
- Buchanan et al., 2006 R.J. Buchanan, R.A. Martin, W. Suojin, MyungSuk Kim. Racial analysis of longer stay nursing home residents with multiple sclerosis. Ethnicity & Disease. 2006;16:159-165
- Buchanan et al., 2010 R.J. Buchanan, D. Radin, C. Huang, et al. Caregiver perceptions associated with risk of nursing home admission for people with multiple sclerosis. Disability and Health Journal. 2010;3:117-124
- Casetta et al., 2009 I. Casetta, T. Riise, M. Wamme Nortvedt, et al. Gender differences in health-related quality of life in multiple sclerosis. Multiple Sclerosis. 2009;15:1339-1346
- et al., Chronic Conditions Indicator for ICD-9-CM. 〈http://www.hcup-us.ahrq.gov/toolssoftware/chronic/chronic.jsp〉 [accessed 26.08.12].
- Cohen et al., 2009 J.W. Cohen, S.B. Cohen, J.S. Banthin. The Medical Expenditure Panel Survey: a national information resource to support healthcare cost research and inform policy and practice. Medical Care. 2009;47:S44-S50
- Gold et al., 1996 M. Gold, J.E. Siegel, L.B. Russell, M.C. Weinstein. Cost-effectiveness in health and medicine. (Oxford University Press, New York, 1996)
- Goldberg et al., 2009 L.D. Goldberg, N.C. Edwards, C. Fincher, et al. Comparing the cost-effectiveness of disease-modifying drugs for the first-line treatment of relapsing-remitting multiple sclerosis. Journal of Managed Care Pharmacy. 2009;15:543-555
- Heesen et al., 2006 C. Heesen, L. Nawrath, C. Reich, et al. Fatigue in multiple sclerosis: an example of cytokine mediated sickness behaviour?. Journal of Neurology, Neurosurgery, and Psychiatry. 2006;77:34-39
- Ivanova et al., 2012 J.I. Ivanova, R.E. Bergman, H.G. Birnbaum, et al. Impact of medication adherence to disease-modifying drugs on severe relapse, and direct and indirect costs among employees with multiple sclerosis in the US. Journal of Medical Economics. 2012;15:601-609
- Ivanova et al., 2009 J.I. Ivanova, H.G. Birnbaum, S. Samuels, et al. The cost of disability and medically related absenteeism among employees with multiple sclerosis in the US. Pharmacoeconomics. 2009;27:681-691
- Kingwell et al., 2011 E. Kingwell, M. van der Kop, Y. Zhao, et al. Relative mortality and survival in multiple sclerosis: findings from British Columbia, Canada. Journal of Neurology, Neurosurgery, and Psychiatry. 2011;83:61-66
- Kobelt et al., 2006a G. Kobelt, J. Berg, D. Atherly, et al. Costs and quality of life in multiple sclerosis: a cross-sectional study in the United States. Neurology. 2006;66:1696-1702
- Kobelt et al., 2006b G. Kobelt, J. Berg, P. Lindgren, et al. Costs and quality of life of patients with multiple sclerosis in Europe. Journal of Neurology, Neurosurgery, and Psychiatry. 2006;77:918-926
- Lee et al., 2012 S. Lee, D.C. Baxter, B. Limone, et al. Cost-effectiveness of fingolimod versus interferon beta-1a for relapsing remitting multiple sclerosis in the United States. Journal of Medical Economics. 2012;15:1088-1096
- Li et al., 2012 Y. Li, K.L. Munger, S. Batool-Anwar, et al. Association of multiple sclerosis with restless legs syndrome and other sleep disorders in women. Neurology. 2012;78:1500-1506
- Loma and Heyman, 2011 I. Loma, R. Heyman. Multiple sclerosis: pathogenesis and treatment. Current Neuropharmacology. 2011;9:409-416
- Mayr et al., 2003 W.T. Mayr, S.J. Pittock, R.L. McClelland, N.W. Jorgensen, J.H. Noseworthy, M. Rodriguez. Incidence and prevalence of multiple sclerosis in Olmsted County, Minnesota, 1985–2000. Neurology. 2003;61(10):1373-1377
- Mills and Young, 2011 R.J. Mills, C.A. Young. The relationship between fatigue and other clinical features of multiple sclerosis. Multiple Sclerosis. 2011;17:604-612
- Morrow, 2007 T.J. Morrow. The costs and consequences of multiple sclerosis relapses: a managed care perspective [published online ahead of print March 23, 2007]. Journal of the Neurological Sciences. 2007;256(Suppl. 1):S39-S44
- et al., National MS Society. 〈http://www.nationalmssociety.org/about-multiple-sclerosis/relapsing-ms/relapsing-remitting-ms-rrms/how-rrms-differs-from-progressive-courses-of-ms/index.aspx〉 [accessed 08.02.13].
- et al., National Multiple Sclerosis Society. Who gets MS? 〈http://www.nationalmssociety.org/about-multiple-sclerosis/what-we-know-about-ms/who-gets-ms/index.aspx〉 [accessed 20.09.11].
- Neumann and Greenberg, 2009 P.J. Neumann, D. Greenberg. Is the United States ready for QALYs?. Health Affairs (Millwood). 2009;28:1366-1371
- Noonan et al., 2002 C.W. Noonan, S.J. Kathman, M.C. White. Prevalence estimates for MS in the US and evidence of an increasing trend in women. Neurology. 2002;58:136-138
- Noonan et al., 2010 C.W. Noonan, D.M. Williamson, J.P. Henry, R. Indian, S.G. Lynch, J.S. Neuberger, et al. The prevalence of multiple sclerosis in 3 US communities. Preventing Chronic Disease. 2010;7:1
- Northrop and Frankel, 2009 D.E. Northrop, D. Frankel (Eds.) Nursing home care of individuals with multiple sclerosis: guidelines and recommendations for quality care. Monograph (National Multiple Sclerosis Society Professional Resource Center, New York, NY, 2009)
- Olofsson et al., 2011 S. Olofsson, A. Wickstrom, A. Hager Glenngard, et al. Effect of treatment with natalizumab on ability to work in people with multiple sclerosis: productivity gain based on direct measurement of work capacity before and after 1 years of treatment. BioDrugs. 2011;25:299-306
- Orlewska, 2006 E. Orlewska. Economic burden of multiple sclerosis: what can we learn from cost-of-illness studies?. Expert Review of Pharmacoeconomics and Outcomes Research. 2006;6:145-154
- Orton et al., 2006 S.M. Orton, B.M. Herrera, I.M. Yee, et al. Canadian Collaborative Study Group. Sex ratio of multiple sclerosis in Canada: a longitudinal study. Lancet Neurology. 2006;5:932-936
- Patwardhan et al., 2005 M.B. Patwardhan, D.B. Matchar, G.P. Samsa, D.C. McCrory, R.G. Williams, T.T. Li. Cost of multiple sclerosis by level of disability: a review of literature. Multiple Sclerosis. 2005;11:232-239
- Phillips, 2004 C.J. Phillips. The cost of multiple sclerosis and the cost effectiveness of disease-modifying agents in its treatment. CNS Drugs. 2004;18:561-574
- Prescott et al., 2007 J.D. Prescott, S. Factor, M. Pill, et al. Descriptive analysis of the direct medical costs of multiple sclerosis in 2004 using administrative claims in a large nationwide database. Journal of Managed Care Pharmacy. 2007;13:44-52
- Rajagopalan et al., 2011 K. Rajagopalan, R.A. Brook, I.A. Beren, et al. Comparing costs and absences for multiple sclerosis among US employees: pre- and post-treatment initiation. Current Medical Research and Opinion. 2011;27:179-188
- Sellner et al., 2011 J. Sellner, J. Kraus, A. Awad, et al. The increasing incidence and prevalence of female multiple sclerosis—a critical analysis of potential environmental factors. Autoimmunity Reviews. 2011;10:495-502
- Steinberg et al., 2010 S.C. Steinberg, R.J. Faris, C.F. Chang, et al. Impact of adherence to interferons in the treatment of multiple sclerosis. Clinical Drug Investigation. 2010;30:89-100
- Stopl-Smith et al., 1998 K.A. Stopl-Smith, E.J. Atkinson, M.E. Campion, P.C. O'Brien, M. Rodriguez. Health care utilization in multiple sclerosis: a population-based study in Olmsted County, MN. Neurology. 1998;50:1594-1600
- Trisolini et al., 2010 M. Trisolini, A. Honeycutt, A. Wiener, S. Lesesne. Global economic impact of multiple sclerosis. ((Multiple Sclerosis International Federation) Prepared by RTI, Research Triangle Park, 2010)
- et al., United States Census Bureau. Income data. Available from: 〈http://www.census.gov/hhes/www/income/data/〉 [accessed 26.08.12].
- et al., US Department of Labor. Bureau of Labor Statistics. Consumer Price Index. 〈http://www.bls.gov/cpi/〉 [accessed 26.08.12].
- Vlahiotis et al., 2010 A. Vlahiotis, R. Sedjo, E.R. Cox, et al. Gender differences in self-reported symptom awareness and perceived ability to manage therapy with disease-modifying medication among commercially insured multiple sclerosis patients. Journal of Managed Care Pharmacy. 2010;16:206-216
- Ware et al., 1996 J. Ware Jr., M. Kosinski, S.D.A. Keller. 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Medical Care. 1996;34:220-233
- Williamson and Henry, 2004 D.M. Williamson, J.P. Henry. Challenges in addressing community concerns regarding clusters of multiple sclerosis and potential environmental exposures. Neuroepidemiology. 2004;23(5):211-216
- Wood et al., 2013 B. Wood, I. van der Mei, A.L. Ponsonby, et al. Prevalence and concurrence of anxiety, depression and fatigue over time in multiple sclerosis. Multiple Sclerosis. 2013;19:217-224
- Wundes et al., 2010 A. Wundes, T. Brown, E.J. Bienen, et al. Contribution of intangible costs to the economic burden of multiple sclerosis. Journal of Medical Economics. 2010;13:626-632
- Yamamoto and Campbell, 2012 D. Yamamoto, J.D. Campbell. Cost-effectiveness of multiple sclerosis disease-modifying therapies: a systematic review of the literature. Autoimmune Diseases. 2012;:784364 10.1155/2012/784364 (Epub 2012 December 6)
- Zwibel and Smrtka, 2011 H.L. Zwibel, J. Smrtka. Improving quality of life in multiple sclerosis: an unmet need. American Journal of Managed Care. 2011;17(Suppl. 5):S139-S145
a University of Colorado, Skaggs School of Pharmacy and Pharmaceutical Sciences, Aurora, CO, USA
b Biogen Idec, Medical and Outcomes Science, Weston, MA, USA
c University of Colorado School of Medicine, Department of Neurology, Aurora, CO, USA
d Mastic Beach, NY, USA
Correspondence to: Kavita Nair, Center for Pharmaceutical Outcomes Research (CePOR), Skaggs School of Pharmacy and Pharmaceutical Sciences, University of Colorado, Anschutz Medical Campus, 12850 East Montview Boulevard, Aurora 80045, CO, USA. Tel.: +1 303 724 2635, cell: +1 720 277 5855; fax: +1 303 724 0979.
© 2013 Elsevier B.V., All rights reserved.