Table 2.
Impact of FDA Warning on Prevalent Second Generation Antipsychotic Users, by Diagnosis
| Bipolar Disorder | Schizophrenia | |||||
|---|---|---|---|---|---|---|
| Modeled Outcomea | Risk | Risk Ratio | 95% Confidence Interval |
Risk | Risk Ratio | 95% Confidence Interval |
| Second Generation Antipsychotic Use | ||||||
| Pre-Advisory Period (Jan 01 – Jan 03) | .66 | REF | REF | .70 | REF | REF |
| Early- Advisory (Feb 03 – Nov 03) | .64 | .96 | .93–.99 | .70 | 1.00 | 1.00–1.01 |
| Advisory Period (Dec 03 – Aug 04) | .62 | .94 | .91–.98 | .69 | .99 | .99–1.00 |
| Initial Post-Advisory (Sept 04 – Mar 05) | .61 | .93 | .89–.97 | .68 | .98 | .97–.99 |
| PDL / CATIE / Dementia Warning (April 05 – Sept 05) | .61 | .92 | .88–.96 | .67 | .96 | .95–.97 |
| Post-PDL/CATIE / Dementia Warning (Oct 05 – Dec 06) | .52 | .78 | .74–.82 | .60 | .87 | .85–.88 |
| Alternative Medication Useb | ||||||
| Pre-Advisory Period (Jan 01 – Jan 03) | .22 | 1.00 | REF | .07 | 1.00 | REF |
| Early- Advisory (Feb 03 – Nov 03) | .31 | 1.38 | 1.31–1.45 | .07 | 1.02 | .99–1.06 |
| Advisory Period (Dec 03 – Aug 04) | .31 | 1.38 | 1.29–1.47 | .07 | 1.02 | .98–1.07 |
| Initial Post-Advisory (Sept 04 – Mar 05) | .32 | 1.44 | 1.33–1.55 | .07 | 1.00 | .95–1.05 |
| PDL / CATIE / Dementia Warning (April 05 – Sept 05) | .31 | 1.39 | 1.28–1.51 | .07 | 1.00 | .95–1.06 |
| Post-PDL/CATIE / Dementia Warning (Oct 05 – Dec 06) | .30 | 1.33 | 1.22–1.46 | .07 | .96 | .91–1.02 |
| Second Generation Antipsychotic Discontinuationsc | ||||||
| Pre-Advisory Period (Jan 01 – Jan 03) | .03 | 1.00 | REF | .03 | 1.00 | REF |
| Early- Advisory (Feb 03 – Nov 03) | .02 | .75 | .60–.94 | .02 | .75 | .69–.82 |
| Advisory Period (Dec 03 – Aug 04) | .02 | .74 | .58–.95 | .02 | .79 | .71–.87 |
| Initial Post-Advisory (Sept 04 – Mar 05) | .02 | .73 | .55–.98 | .02 | .80 | .72–.90 |
| PDL / CATIE / Dementia Warning (April 05 – Sept 05) | .02 | .88 | .66–1.18 | .03 | .96 | .86–1.08 |
| Post-PDL/CATIE / Dementia Warning (Oct 05 – Dec 06) | .04 | 1.37 | 1.11–1.68 | .04 | 1.46 | 1.35–1.58 |
| Second Generation Antipsychotic Interruptionsd | ||||||
| Pre-Advisory Period (Jan 01 – Jan 03) | .09 | 1.00 | REF | .08 | 1.00 | REF |
| Early- Advisory (Feb 03 – Nov 03) | .08 | .81 | .74–.90 | .07 | .82 | .80–.85 |
| Advisory Period (Dec 03 – Aug 04) | .08 | .80 | .72–.89 | .07 | .82 | .79–.85 |
| Initial Post-Advisory (Sept 04 – Mar 05) | .08 | .82 | .72–.93 | .07 | .82 | .78–.85 |
| PDL / CATIE / Dementia Warning (April 05 – Sept 05) | .07 | .80 | .70–.91 | .08 | .90 | .86–.93 |
| Post-PDL/CATIE / Dementia Warning (Oct 05 – Dec 06) | .07 | .77 | .68–.86 | .07 | .82 | .80–.85 |
Generalized estimating equations were used to account for the repeated measures on each subject. We estimated each model using PROC GENMOD (SAS, 9.1) with a Poisson distribution, log link. The correlation structure was unspecified. Control variables included sex, age, race, Medicaid eligibility status, and the presence of the following during the 6 months prior to the index prescription date: receipt of care from a psychiatrist, inpatient mental health treatment, dementia, alcohol or substance abuse, epilepsy, complicating conditions, diabetes, hyperlipidemia, hypertension, heart disease, or obesity. PDL / CATIE / Dementia Warning = Period including the Medicaid Preferred-Drug List policy change, publication of the CATIE trial results, and the FDA’s black box warning for the risk of mortality among elderly patients with dementia.
Medication alternatives included first generation antipsychotics (for enrollees with schizophrenia or bipolar disorder), and lithium, lamotrigine, carbamazepine, or valproate (for enrollees with bipolar disorder).
Discontinuations were defined as 6 or more months of no second generation antipsychotic use following a month of second generation antipsychotic use.
Interruptions were defined as a gap in second generation antipsychotic fills of 1 to 5 months with second generationantipsychotic use before and after the gap.