Alzheimer's disease (AD) and other dementias pose a major health and economic challenge to U.S. society. At least one in eight Baby Boomers will develop AD before death (
Plassman et al., 2007- Plassman B.L.
- Langa K.M.
- Fisher G.G.
- Heeringa S.G.
- Weir D.R.
- Ofstedal M.B.
- Wallace R.B.
- et al.
Prevalence of dementia in the United States: The Aging, Demographics, and Memory Study.
) and the average annual monetary cost of care for each AD patient is estimated to be between $41,689 and $56,290 (
Hurd et al., 2013- Hurd M.D.
- Martorell P.
- Delavande A.
- Mellen K.J.
- Langa K.M.
Monetary costs of dementia in the United States.
). However, a gap exists in literature investigating the differences in the health and economic burden of AD by demographic and socioeconomic features to guide more “targeted” interventions for the prevention, treatment, and care of AD patients.
One of the most established differences in AD risk across demographic categories is gender based, although the financial implications are unknown. Women face greater risks of either being diagnosed with AD or serving as informal caregivers for their family members with AD. Using nationally representative data,
Hurd et al., 2013- Hurd M.D.
- Martorell P.
- Delavande A.
- Mellen K.J.
- Langa K.M.
Monetary costs of dementia in the United States.
found that surviving elderly women at age 65 or older have 38% higher (12.1% vs. 8.8%) risk to develop AD than surviving men. The National Alliance for Caregiving reported that approximately 60% to 70% of the informal care givers for AD patients are females, and females often take the most physically and mentally challenging roles in caregiving, including bathing, feeding, toileting, and so on, whereas men usually take the less stressful roles of reading, transportation, and spiritual support (
). Therefore, this study aimed to quantify the health and economic burdens of AD by gender.
Methods
Unlike the majority of existing literature that investigates the economic cost of AD at annual levels, this study takes an innovative lifetime perspective to investigate the burden of AD by gender, over the course of the disease. Because AD is an extremely challenging and complicated chronic disease that could last more than a decade and requires a large variety of clinical and nursing care, the net cost of AD per capita throughout the lifetime is more informative for policy discussions than the annual cost. The lifetime cost is determined by three factors: 1) the probability to develop the disease before death, 2) the duration of the disease, and 3) the component and intensity of the formal or informal care that patients received over the course of illness. This study captures the gender differences in the lifetime cost of AD, to compare the costs of both formal and informal care reimbursed by a third party, paid out of pocket, or that are uncompensated, throughout the lifetime. These costs include Medicare cost for formal clinical and nursing care, Medicaid cost for clinical and long-term care (LTC), out-of-pocket cost for assisted living (ASL) facility care, out-of-pocket cost for home health (HH) care, and informal care provided by family members. This study used secondary data without identifying information that could link to the human subject. The data were analyzed on an authorized and protected computer on the Emory University campus. The Institutional Review Board of Emory University waived the review of human subject protection.
The data used in this study is from the Cost and Use Files of the Medicare Current Beneficiary Survey (MCBS) from 2000 to 2010. The MCBS is collected by Centers for Medicare & Medicaid Services (CMS) as a nationally representative sample of Medicare beneficiaries. It merges the survey information of the respondents' demographic, socioeconomic, and general health with their Medicare and Medicaid claims data. The details of MCBS are introduced by
Respondents were identified as having AD based on any one of these three criteria: 1) a yes response for a survey question regarding whether the respondent has been diagnosed with Alzheimer's disease or dementia; 2) ICD-9 codes of inpatient and outpatient claims that indicate dementia (ICD-9-CM 290 or 331.0), and 3) outpatient prescription drugs claims showing utilization of drugs prescribed for dementia, including donepezil (Aricept), revastigmine (Exelon), galantamine (Reminyl or Razadyne), and memantine (Namenda;
Taylor et al., 2009- Taylor D.H.
- Østbye T.
- Langa K.M.
- Weir D.
- Plassman B.L.
The accuracy of Medicare claims as an epidemiological tool: The case of dementia revisited.
,
Taylor et al., 2001- Taylor D.H.
- Schenkman M.
- Zhou J.
- Sloan F.A.
The effect of disability and comorbidity on the cost of Alzheimer's disease and related dementias.
,
Yang et al., 2012- Yang Z.
- Zhang K.
- Lin P.
- Clevenger C.
- Atherly A.D.
A longitudinal analysis of the lifetime cost of dementia.
) We used these three criteria to include all possible AD cases, to avoid underestimates of the prevalence; 93% of the AD sample were identified by survey and/or the claims data, and the remaining 7% were identified by prescription drug utilization.
To focus on the gender differences, this study conducts a multistage analysis from the lifetime perspective among men and women separately. Such method is an expansion of the lifetime estimation method introduced in earlier publications (
Lin et al., 2014- Lin P.J.
- Yang Z.
- Howard F.
- Cohen J.
- Neumann P.J.
Unintended benefits: The potential economic impact of addressing risk factors to prevent Alzheimer's disease.
,
Yang et al., 2012- Yang Z.
- Zhang K.
- Lin P.
- Clevenger C.
- Atherly A.D.
A longitudinal analysis of the lifetime cost of dementia.
) First, we used regression analysis to estimate the risk of experiencing AD before death, the probability of death conditional on AD status, and the annual costs conditional on AD status and time to death. Then, a cohort-based, counterfactual simulation was conducted to estimate the duration of the disease after onset, the duration of time that AD patients will spend in the community versus in LTC facilities, and the lifetime incremental costs of AD on Medicare and Medicaid. The cohort-based demographic features were incorporated in the simulation, and all the costs were inflated to 2013 dollar value, based on Consumer Price Index of Medical Goods published by Bureau of Labor Statistics (
), and all the cost estimates from the simulation are discounted to current 2013 dollar value for the current cohort at age 65. The details of the model can be found in the previous publication of
Yang et al., 2012- Yang Z.
- Zhang K.
- Lin P.
- Clevenger C.
- Atherly A.D.
A longitudinal analysis of the lifetime cost of dementia.
, and results of the regression models by gender will be provided by the authors upon request.
We calculated four types of cost related to AD: 1) Medicare cost for the reimbursement of clinical care, 2) Medicaid cost for the reimbursement of LTC, 3) out-of-pocket cost for ASL and HH care, and 4) the cost of uncompensated, informal care. The first two types of costs are borne by publicly financed entitlement programs, and the last two types of cost are usually borne by the family of AD patients and their caregivers.
The costs of Medicare and Medicaid are obtained from MCBS directly, because the claims data of MCBS provide the amount of reimbursement from Medicare or Medicaid to clinical or LTC services. The lifetime incremental Medicare and Medicaid costs of AD are then calculated by counterfactual simulation. For other types of cost, we imputed cost data from other sources into the simulated course of illness to obtain the lifetime estimates.
For ASL and HH cost, to simplify the estimation, we assume spouses are the primary caregivers for each other. The cohort covered by the MCBS survey and used in the simulation were born in the 1920s and 1930s, and entered into marriage age in 1940s and 1950s; according to the data from the National Vital Statistics System, the marriage rate among this cohort was close to 83% among both men and women in the 1940s and 1950s, and we used this marriage rate data in our analysis (
,
Perrin et al., 1973Perrin E.B., Fisher G.F., Simmons W.R., Hanlon J.J., Kelly J.E., Minty E.E.., … Hetzel AM. (1973). “100 years of marriage and divorce statistics United States, 1867–1967: Data from National Vital Statistics System, Series 21(24). Washington, DC: U.S. Department of Health Education, and Welfare.
,
). We are aware that the marriage rate of this cohort after age 65 is likely lower because part of the population could have been widowed or divorced. However, that level of technical details is beyond the scope of this particular study because identification of marriage and divorce rates by age, gender, cohort, and AD status is quite challenging, and accounting these factors will not change the conclusions significantly. Therefore, we use the 83% marriage rate for the analysis, but acknowledge it as a limitation of this particular study.
Although ASL facilities provide nursing care, we count the AD patients in ASL as living in the community instead of LTC facilities, because MCBS data only specified nursing homes as LTC facilities. According to the National Association of Professional Geriatric Care Managers, currently there are around 735,000 residents in ASL nationwide, according to data provided by managers of ASL facilities, and 40% of those patients have AD (
). This accounts for about 5% of the 5.8 million AD/dementia patients. Therefore, we assume that there is a 5% probability that AD patients who stay in community will use ASL services, and we use this probability to calculate the out-of-pocket cost of ASL care for AD patients by gender in this study.
Besides the cost of ASL, this study assumes that the rest of community-dwelling AD patients use some form of HH care. We calculate the cost of HH care for AD patients in the community to be based on the probability that they will require HH care (1 minus 5% to use ASL at 95%) and impute the average annual cost of HH care as estimated by
Hurd et al., 2013- Hurd M.D.
- Martorell P.
- Delavande A.
- Mellen K.J.
- Langa K.M.
Monetary costs of dementia in the United States.
at $5,678 per year.
Finally, the majority — 70% — of informal caregivers for AD/dementia patients are women. In this study, the informal care cost per AD calculation is based on the duration of time the AD patient spends living in the community, the average marriage rate of 83%, the probability of being a caregiver by gender, and the average annual cost of informal care based on replacement costs of $27,789 as estimated by
Hurd et al., 2013- Hurd M.D.
- Martorell P.
- Delavande A.
- Mellen K.J.
- Langa K.M.
Monetary costs of dementia in the United States.
. We are aware that no “gold standard” exists for estimating the cost of informal care, and we choose to use the replacement cost from
Hurd et al., 2013- Hurd M.D.
- Martorell P.
- Delavande A.
- Mellen K.J.
- Langa K.M.
Monetary costs of dementia in the United States.
for two reasons. First, this is the most recent publication using nationally representative data. Second, we believe this particular estimate is conservative, to match with the economic theory of opportunity cost—low-income women are more likely to be caregivers themselves instead of hiring caregivers because of the lower opportunity cost of lost income.
Article info
Publication history
Accepted:
June 4,
2015
Received in revised form:
May 29,
2015
Received:
June 20,
2014
Footnotes
Funding source: This study is not supported by external funding, and there is no conflict of interest for either of the authors.
Copyright
© 2015 Jacobs Institute of Women's Health. Published by Elsevier Inc. All rights reserved.