Table 1

Key features of included BZD-withdrawal trials

Study characteristicsBashir et al, 199416Cormack et al, 199417Heather et al, 200418
Numberof participants109222284
Setting; patientsUK general practice; patients with long-term BZD useUK general practice; patients with long-term BZD useUK general practice; patients with long-term BZD use
Follow-up6 months6 months6 months
InterventionsOne consultation from GP+ self-help bookleta to reduce gradually/stop drugGroup 1: single letter from GPb; group 2: letterb + 4-monthly information sheetsa on reducing medication and coping with withdrawalGroup 1: letter from GP to reduce gradually/stop drugb; group 2: consultation and self-help bookleta
ComparisonAllowed to continue usual doseAllowed to continue usual doseAllowed to continue usual dose
Outcome measures
BZD usagePrescription recordsPrescription recordsPrescription records
Health questionnaireGHQGHQ and SF-36
Methodological quality of studies
Adequate randomisation?Yes (used date of birth with baseline equality)YesYes
Baseline similarity?NoUnclearYes
Groups treated equally?YesUnclearUnclear
Loss to follow-up3.7% (BZD reduction), 15% (GHQ)5.8%3.9% (BZD reduction), 30% (GHQ)
Outcomes measures blind or objective?Unblinded but objective prescription recordsUnblinded but objective prescription recordsUnblinded but objective prescription records
  • a Sample not provided

  • b sample provided.

  • BZD = benzodiaze epines. GHQ = general health questionnaire.

  • SF = short-form.