Table 12 Fisch et al., 2003
Domain | Description | Review author’s judgement |
|---|---|---|
Sequence generation | Randomization by means of a preprinted randomization table. | “Yes” |
Allocation concealment | The exact method is not described. | “Unclear” |
Blinding of participants, personnel, and outcome | “The issue is not addressed by the authors. | “Unclear” |
Incomplete outcome data | 163 patients were randomized and 159 allocated to receive medication. The patients were included in the analysis if they provided data at least two assessments (baseline and one of the next four). 64 patients were evaluable in the fluoxetine group and 65 in the placebo group. The reasons for dropouts are not fully presented. The authors used a modification of completers’ analysis. The missing data imputation method was the best change score, which is defined as the difference between the baseline score and the average of the best consecutive scores. According to the authors’ opinion this is a valid statistical method for longitudinal data. To our opinion the best change score belongs to the inappropriate imputation methods. | “No” |
Selective outcome reporting | No study protocol available. All pre-specified outcomes are reported in the pre-specified way. | “Yes” |
Other sources of bias | There were many loss of data especially at the fourth assessment. This could influence the intervention effect estimate. | “No” |