Five RCTs (294 participants) were included in the analysis. Only one RCT was appropriately randomised. Two RCTs were described as double-blinded and blinded outcome assessors. No RCTs reported blinding of participants or investigators. Attrition was described in all five RCTs and was less than 10% to 15% in four of these. Attrition was properly analysed in four RCTs. Follow-up ranged from nine weeks to 24 months.
Sleep outcomes (objective measures): There was very low evidence reporting mixed effects in the short term (post treatment). One RCT found improvements in sleep latency and sleep efficiency for temazepam compared to CBT but two RCTs found no differences for temazepam or triazolam and CBT. There was moderate evidence suggesting that CBT for insomnia was more effective than zopiclone and zolpidem in the short term (two RCTs).
At long term follow-up (six to 24 months) there was moderate evidence suggesting that CBT for insomnia was more effective than temazepam and triazolam (three RCTs) and low grade evidence that CBT was more effective than zopiclone (one RCT). There was no long term follow-up for zolpidem.
Quality of life: No significant between-group differences were reported for quality of life outcomes. Improvements on the State-Trait Anxiety Index was reported for CBT groups compared to zopiclone (one RCT). Adverse events were rarely reported.
Other findings were reported in the review but they were difficult to follow.