Health Care Utilization and Costs of Alzheimer's

0 downloads 0 Views 51KB Size Report
Home Health Visits. Costs. $548. $251. $296.
528

July-August 2002

Family Medicine

Health Care Utilization and Costs of Alzheimer’s Disease: The Role of Co-morbid Conditions, Disease Stage, and Pharmacotherapy Howard Fillit, MD; Jerrold W. Hill, PhD; Robert Futterman, PhD Background and Objectives: Studies on the relationship between Alzheimer’s disease (AD) and health care costs have yielded conflicting results. This study analyzed the relationship between co-morbid conditions and health care utilization and costs for patients with AD and estimated costs by stage of disease and receipt of pharmacotherapy. Methods: We conducted a retrospective analysis of administrative data for 1,366 patients with AD and 13,660 age-gender matched controls enrolled in a large Medicare managed care organization (MCO). Co-morbid conditions were based on the diagnostic classifications from the Charlson co-morbidity index. Health care costs and utilization for MCO-covered services for cases were compared to controls. We used presence of complications of AD associated with later-stage disease to classify patients as having earlier- or later-stage AD. Results: After controlling for co-morbid conditions, age, and gender, annual costs were $3,805 higher for AD patients, resulting in excess costs of $5 million to the MCO. For seven of the 10 most prevalent co-morbidities for AD patients, adjusted costs were higher for AD patients compared with controls with the same condition. Higher costs were attributable to higher inpatient and skilled nursing facility costs. Costs for patients classified as earlier-stage AD were 44% higher than controls and significantly higher for eight of 10 co-morbid conditions when compared with controls with the same conditions. Costs for AD patients receiving treatment by a cholinesterase inhibitor were $2,408 lower than AD patients not receiving therapy. Conclusions: Utilization and costs for patients with AD were higher compared to controls and were substantially higher for patients with both AD and co-morbid diseases commonly targeted for disease management. Earlier-stage AD and receipt of pharmacotherapy were associated with lower costs. These findings indicate that better treatment and care management of AD could reduce the costs of co-morbid illnesses commonly suffered by AD patients. (Fam Med 2002;34(7):528-35.)

While there are many published studies on the relationship between Alzheimer’s disease (AD) and medical costs, the results of those studies have been equivocal. Several studies1,2 report that costs for managed care patients with AD and related dementias (ADRD) were higher than in controls without ADRD, and several similar studies report higher Medicare- and Medicaidcovered and caregiver costs for community-dwelling patients with AD.3-7 In contrast, a study of patients identified with AD or other dementias (ADOD) on the Medi-

From the Institute for the Study of Aging (Drs Fillit and Hill); the Department of Geriatrics and Adult Development, Mount Sinai Medical Center (Dr Fillit); and Medical and Quality Informatics, HIP of Greater NewYork (Dr Futterman); New York.

care Current Beneficiary Survey reports no difference in direct medical costs compared with beneficiaries without ADOD.8 Also, a study of patients with AD and other dementias in the last 3 years of life reports lower costs for AD patients compared with controls and no difference in costs for patients with other dementias, compared with controls.9 The differences in cost estimates are most likely attributable to criteria for sample selection and statistical methods used to control for differences between dementia patients and controls. Specifically, some studies include only AD patients 3,5,6,10-12 while others include patients with AD or other dementias,1,2,7 some include community-dwelling and nursing home patients in the same sample9 while others segregate these groups,7,8 and some measure both medical and caregiver costs 3,7,10

Award-winning Papers From the AAFP 2001 Annual Scientific Assembly while others focus on medical costs exclusively.1,2,6,11-13 Further, some studies identify patients from medical claims1,6 while others identify patients from medical records or other clinical criteria.5,7,9-11 More-recent cost estimates may also be influenced by whether AD patients are receiving one of four prescription cholinestera se-inhibiting drugs (tacr ine, 1 4 donepe zil,1 5 ,16 rivastigmine, 17,18 and galantamine19,20 ) for the treatment of mild to moderate AD. Whether or not the types of dementias studied (AD only versus AD and other dementias) influence cost estimates is a question of practical importance. Cost estimates specific to AD patients will better inform decisions on therapies and interventions specific to AD. For example, evidence that AD increases health care costs and available therapies reduce costs will encourage coverage of these therapies in prescription drug plans and encourage physicians to prescribe them. To assess the economic effect of cholinesterase inhibitors, it is necessary to determine cost estimates by stage of AD because these drugs are indicated only for patients with mild to moderate AD. Cost estimates by disease stage are also required for evaluating the costeffectiveness of screening for AD and case management of AD patients in the community, since AD patients with mild to moderate disease are more likely to receive these interventions. Separate cost estimates for those receiving therapy and those not receiving therapy are essential, however, for assessing the financial effects of formulary decisions on coverage and clinical decisions on prescribing. To address these issues, this study examines medical costs in a Medicare managed care plan for patients with the specific diagnosis of AD. Costs are estimated separately for all AD patients and AD patients classified as earlier-stage AD. For those with earlier-stage disease, we estimate costs for those receiving either of the two prescription drugs available for the treatment of AD during the study period (tacrine or donepezil). Finally, we examine the effect of these factors on costs for patients with co-morbid conditions. Methods Medical costs and health services utilization for AD patients were compared to an age-gender matched set of controls in a large mixed-model Medicare managed care plan for the time period of October 1, 1997, through September 30, 1999. Estimates of costs, utilization, and diagnoses were derived from administrative data on enrollment, medical claims, pharmacy claims, and encounter records of services. Case-control differences in costs and utilization were analyzed by co-morbid condition from the Charlson co-morbidity index.21,22

Vol. 34, No. 7

529

Subjects The subjects for this study were a subset of the sample previously analyzed.2 The original sample included 3,934 patients with an International Classification of Diseases, Ninth Revision (ICD-9) diagnosis of AD (331.0), senile dementia (all 290 codes), senility (797), drug-induced dementia (292.82), alcoholic dementia (291.2), dementia in condition classified elsewhere (294.1), dialysis dementia (294.8), Pick’s disease (331.1), and senile degeneration (331.2). Of the 3,934 dementia patients previously studied, 1,366 patients with the specific ICD-9 code for AD were selected. Ten controls were selected for each AD patient, matched by age and gender (Table 1). As in our previous study,2 control observations were weighted so that the length of enrollment for AD patients and controls was equivalent. The managed care organization (MCO) from which the subjects received services maintains exclusive relationships with several group model independent practice associations (IPAs) providing capitated primary and specialist care to most of its Medicare members. However, some members receive health care services from providers working under a fee-for-service arrangement with the MCO. The IPAs do not bill the MCO but, instead, submit a record for each service with a procedure code. The procedure code for each service was matched to the Medicare physician fee schedule to impute costs to the MCO for services provided by the IPAs. Identifying Co-morbidity The co-morbidity classifications comprising the Charlson co-morbidity index22 were used to identify

Table 1 Demographic Characteristics of Samples

Age 65–74

Alzheimer’s Disease Women Men # (%) # (%) 236 (28.64) 177 (32.66)

Controls Women Men # (%) # (%) 2,360 (28.64) 1,770 (32.66)

75–84

368 (44.66)

253 (46.68)

3,680 (44.66) 2,530 (46.68)

85–94

208 (25.24)

107 (19.74)

2,080 (25.24) 1,070 (19.74)

>94 Total

12

(1.46)

824 (100.00)

Women

59.6%

Average age (years)

79.8

5

(.92)

542 (100.00)

120

(1.46)

50

(.92)

8,240 (100.00) 5,420 (100.00) 59.5%

77.7

79.5

77.5

530

July-August 2002

members’ co-existing medical conditions other than AD. Members were classified as having the co-morbid condition if they had a qualifying diagnosis appearing on one or more claims or encounter records. Costs and Utilization Tota l costs were computed as amounts paid (or imputed amounts paid) for medical services and prescription drugs. The procedure code for each service was matched to the Medicare physician fee schedule to impute costs to the MCO for services provided by the IPAs. Measures of utilization included hospita l admissions, hospital days, skilled nursing facility (SNF) days, e me rge nc y room visits, home health visits, and physician office visits. Average costs for the study period were converted to annual costs by dividing average cost by average member years of enrollment.

Family Medicine

Table 2 Mean Annual Costs and Utilization, by Place of Service Adjusted for Patient Characteristics, AD Patients Versus Controls AD Patients $9,737

Controls $5,932

Difference, AD Patients Versus Controls $3,805

P Value