BACKGROUND: Donepezil is a new drug recently approved in the United States
and Canada for the treatment of Alzheimer's disease (AD). We estimated the
cost-effectiveness of donepezil 5 mg daily as an adjunct to usual care in t
he management of persons with mild-to-moderate AD defined as a Mini-Mental
Health State Examination (MMSE) score in the range 10 to 26.
METHODS: Treatment effect data as MMSE change-over-baseline scores were obt
ained from a 30-week placebo-controlled trial of donepezil. MMSE scores bey
ond observed trial data were estimated using a Markov model with 10 cycles
of 24 weeks based on the placebo MMSE progression observed in the trial. Da
ta from AD subjects in the Canadian Study of Health and Aging were used to
estimate costs of nursing home care, community services, medications, and c
aregiver rime as a function of MMSE score. A clinic-based cohort study from
Alberta was used to estimate the distribution of AD patients by MMSE score
presenting for treatment. The effectiveness measure for the economic model
was expected time (over 5 years) spent with nonsevere AD (MMSE greater tha
n or equal to 10).
RESULTS: Over 5 years of treatment, donepezil is predicted to reduce health
care costs by CA$929 per patient but increase caregiver time costs by CA$4
8 per patient for an overall cost saving to society of CA$882 per patient.
Patients not receiving donepezil are predicted to spend 2.21 years of the 5
years in nonsevere AD compared with 2.41 years for created patients (a gai
n of just over 2 months). Sensitivity analysis reveals that cost savings pe
r patient increase if more AD patients are assumed to survive to 5 years; h
owever, if donepezil treatment continues when patients' MMSE score falls be
low 101 the incremental cost is higher for treatment at CA$1554 per patient
.
CONCLUSION: Based on the limited available data, our model predicts that th
e use of donepezil for mild-to-moderate AD in Canada is associated with low
er 5-year costs and less time spent with severe AD when compared with the a
lternative of usual care with no donepezil therapy. As more reliable long-t
erm data became available, these predictions should be confirmed and/or upd
ated.