Changes in Antipsychotic
Pharmacotherapy and Healthcare Costs
Following a New Diagnosis of Diabetes
among Patients with Schizophrenia
Douglas L. Leslie, PhD
Robert Rosenheck, MD
New England Mental Illness Research, Education and
Clinical Center (MIRECC);
VA Northeast Program Evaluation Center (NEPEC);
Yale Medical School
Background
First-generation antipsychotics were introduced
in the 1950’s
Second-generation antipsychotics
–
–
–
–
clozapine, olanzapine, risperidone, quetiapine
At least as effective as first-generation drugs
Have fewer side effects
Associated with weight gain and diabetes
FDA warning
Objectives
To determine risk of new-onset diabetes
associated with antipsychotic medications.
To identify the costs associated with
treatment of new-onset diabetes.
To determine whether patients switch
antipsychotic medications after diagnosis
of diabetes.
Sample
Patients with schizophrenia treated in the VA
Healthcare System.
No pre-existing diabetes.
“Stable” on an antipsychotic medication.
Followed for up to 2 years between June 1999
and September 2001.
56,849 patients included in the study.
Methods
Patients with a new diagnosis of diabetes
were identified.
Cox proportional hazards model.
A matched control group of patients who did
not develop diabetes was identified.
Medication changes and costs were
compared.
Study timeline
June 1999
September 2001
D
Diabetes patient
Stable period
(3 months)
Diabetes
onset date
D
Matched nondiabetes patient
Stable period
(3 months)
Pseudo diabetes
onset date
Results
Patients were followed for an average of
15.7 months
4,132 patients had a new diagnosis of
diabetes – annual incidence rate of 4.36%
Unadjusted diabetes risk was highest for
clozapine and olanzapine
Total healthcare costs were $3,104 higher
for patients with diabetes
Fitted survival functions from the Cox proportional
hazards model predicting time to diabetes onset
1.00
Survival
0.98
0.96
0.94
First-generation: ref
risperidone: HR=1.01
olanzapine: HR=1.15*
quetiapine: HR=1.20
clozapine: HR=1.57*
0.92
0.90
0.88
0
5
10
Month
15
20
25
Percentage of patients who switch/discontinue
antipsychotic medication after diabetes onset
No
diabetes
9.8%
With
diabetes
4.6%
p
0.092
olanzapine
25.2%
23.1%
0.179
quetiapine
40.1%
31.9%
0.053
risperidone
28.2%
24.2%
0.023
conventional
24.3%
22.0%
0.097
Last medication
before “diabetes date”
clozapine
Diabetes risk and costs attributable
to second-generation drugs
Average cost
per day of
Attributable
therapy
cost per day*
$0.134
$8.34
Medication
clozapine
Attributable
risk
2.03%
olanzapine
0.63%
$0.042
$6.86
quetiapine
0.80%
$0.053
$3.72
risperidone
0.05%
$0.003
$3.65
* Based on average cost of $198/month per case of diabetes
Conclusions
Very high rate of diabetes onset in this
population (4-5% per year)
Results do not support a “class effect”
Onset of diabetes does not affect
antipsychotic pharmacotherapy
Diabetes risk and treatment costs
attributable to second-generation drugs
are small