Comparison of cholinesterase inhibitor utilization patterns and associated health care costs in Alzheimer's disease.

BACKGROUND: Sustained treatment with a cholinesterase inhibitor (ChEI) is used in the management of the symptoms of Alzheimer's disease (AD). However, the characteristic declines in learning and memory seen in AD may erode the patient's ability to adhere to medication regimens with or without caregiver support.

OBJECTIVES: To examine differences by type of ChEI in (1) monthly prevalence of use, (2) nonpersistence, (3) switching from the index drug to another ChEI, (4) number of days on therapy, (5) medication possession ratio (MPR), and (6) an estimate of the relationship of these characteristics to total annual health care expenditures.

METHODS: Data were from the MarketScan Medicare Supplemental and Coordination of Benefits 2001-2003 database, which comprised 1.47 million Medicare beneficiaries during this 3-year time period. Inclusion criteria were: (1) aged 65 years or older; (2) at least 1 claim with an International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code 331.0 for AD in any of 15 diagnosis fields on outpatient claims or any of 2 diagnosis fields on inpatient claims at any time during 18 months of observation; (3) at least 1 pharmacy claim for donepezil, galantamine, or rivastigmine preceded by a 6-month period without a ChEI claim; and (4) at least 12 months of follow-up data, for a minimum 18 months continuous enrollment. Multivariate analyses, including logistic regression and exponential conditional mean models, tested for cohort differences in ChEI utilization, controlling for demographics, region of the country, type of insurer, and the Charlson Comorbidity Index (comorbid diagnoses). Using exponential conditional mean models, we also examined the relationship between utilization characteristics and all-cause (i.e., not specific to AD) health care expenditures for a 12-month period, including inpatient and outpatient (physician) care, laboratory and radiology services, emergency room (ER) use, prescription drugs, and long-term care services (e.g., nursing home care, home health visits) paid by Medicare or private insurance, but excluding long-term care services paid by Medicaid. Expenditure was defined as allowed charge (i.e., the total payment received by the service provider including plan and patient paid amounts.)

RESULTS: More than 70% of the patients who received ChEI therapy and who otherwise met the inclusion criteria were excluded from this study due to the absence of at least 1 claim with a diagnosis for AD. Of the 3,177 patients included in the study, the index ChEI was donepezil for 62.8% of the patients (n=1,994); 17.2% received galantamine (n=546) and 20.1% received rivastigmine (n=637). The total number of days of index therapy dispensed was greater for those starting on donepezil (mean [median, SD] days=226 [263, 115]) compared with rivastigmine (206 [233, 120], P<0.001), but was not significantly different compared with galantamine (216 [250, 119], P=0.085). Monthly prevalence of use was similar for the 3 drugs until month 5 when a smaller proportion of rivastigmine patients had index medication on hand (65.9%) compared with 72.1% of donepezil patients (P=0.003) and 72.7% of galantamine patients (P=0.012). At 12 months, the likelihood of receiving the index ChEI was higher for donepezil (61.1%) than for either rivastigmine (50.1%, P<0.001) or galantamine (56.4%, P=0.048) and was higher for galantamine than for rivastigmine (P=0.030). The rate of switching for donepezil patients was significantly lower (14.5%) than the switch rate for rivastigmine patients (21.5%, P<0.001) and was similar to the switch rate for galantamine patients (15.0%, P=0.781 for donepezil vs. galantamine; P=0.004 for galantamine vs. rivastigmine). Rates of nonpersistence, measured as having at least 1 gap in therapy of 30 days or more during the 1-year follow-up, were 63.5% for donepezil, 63.7% for galantamine (P=0.933 for donepezil vs. galantamine), and 68.0% for rivastigmine (P=0.042 for donepezil vs. rivastigmine). MPRs and total days supply of any ChEI did not significantly differ among the 3 drugs. Results of multivariate models showed that, controlling for index ChEI drug, each additional month of ChEI treatment was associated with a reduction of 1% in total all-cause health care costs. The mean (SD) total all-cause 1-year health care costs for patients initiated on the 3 ChEIs were not significantly different: $12,112 ($16,437) for donepezil, $12,137 ($19,154) for galantamine (P=0.978), and $12,853 ($14,543) for rivastigmine (P=0.278).

CONCLUSIONS: During the first year following initiation of ChEI therapy, patients initiated on donepezil had a greater days supply of the index medication than did patients initiated on rivastigmine. At 12 months following treatment initiation, the proportion of patients in therapy was higher for donepezil than for either rivastigmine or galantamine and was higher for galantamine than for rivastigmine. Patients treated with either donepezil or galantamine were less likely to switch from the index drug to another ChEI than were patients treated with rivastigmine. All-cause 1-year health care costs for patients initiated on the 3 ChEIs were not significantly different.