Table 1.
Characteristics and methodological quality of randomised trials in review
| Katz, 199925 | De Deyn, 199924 | Chan, 200123 | Brodaty, 200322 | Street, 200026 | |
|---|---|---|---|---|---|
| Interventions | Risperidone v placebo | Risperidone v haloperidol v placebo | Risperidone v haloperidol | Risperidone v placebo | Olanzapine v placebo |
| No of patients | 625 (462 risperidone, 163 placebo) | 344 (115 risperidone, 115 haloperidol, 114 placebo) | 58 (29 risperidone, 29 haloperidol) | 337 (167 risperidone, 170 placebo) | 206 (159 olanzapine, 47 placebo) |
| Drug dose (per day) | Fixed: 0.5, 1, or 2 mg | Flexible: mean 1.1 mg risperidone, 1.2 mg haloperidol | Flexible: mean 0.85 mg risperidone, 0.9 mg haloperidol | Flexible: mean 0.95 mg | Fixed: 5, 10, or 15 mg |
| Duration (weeks) | 12 | 12 | 12 | 12 | 6 |
| Mean age (years) | 82.6 | 81.3 | 80.5 | 83.0 | 82.8 |
| Diagnosis | 73% AD, 15% VaD, 11% mixed | 67% AD, 26% VaD, 7% mixed | 79% AD, 21% VaD | 58% AD, 29% VaD, 13% mixed | 100% AD |
| Mean MMSE (out of 30)27 | 6.6 | 8.4 | 8.0 | 5.5 | 6.7 |
| Assessment of quality: | |||||
| Randomisation* | 1 | 2 | 1 | 1 | 2 |
| Blinding† | 2 | 2 | 1 | 2 | 2 |
| Withdrawals/dropouts‡ | 1 | 1 | 1 | 1 | 1 |
| Jadad quality score§ | 4 | 5 | 3 | 4 | 5 |
| Concealment¶ | 0 | 1 | 0 | 0 | 1 |
| Follow up** | 2 | 2 | 2 | 2 | 2 |
AD=Alzheimer's disease; MMSE=Folstein mini mental state examination; VaD=vascular dementia.
0=not randomised or pseudo-randomised; 1=randomised, method not described; 2=randomised, appropriate method reported.
0=no blinding or inappropriate method; 1=double blind, blinding not reported; 2=double blind, blinding appropriate.
0=not described for each group; 1=described by group.
Score out of 5, based on scores for randomisation, blinding, and withdrawals/dropouts; does not consider concealment of allocation or adequacy of follow up.17
0=no concealment process described; 1=concealment process described.
0=no or inadequate reporting of follow up; 1=reported, data not analysed according to intention to treat; 2=reported, data analysed according to intention to treat.