Two reviewers (Yang and Ho) independently reviewed the articles to determine whether they met the predetermined eligibility criteria. Their results were re-checked by another reviewer (Chien) and all three reviewers resolved any disagreement through discussion. The inclusion criteria are presented in Box 1. Trials were excluded if any participants had systemic disorders or if the control group was instructed to engage in stretching or low-intensity exercise. If multiple published reports

from the same trial were available, only the report that contained the most detailed and quantified information regarding both intervention and outcomes was analysed. Design • Randomised trial Participants • Middle-aged and older adults (> 40 yr) Intervention • Exercise learn more training program (aerobic or resistance exercise)

Outcome measures • Self-reported sleep quality (eg, PSQI questionnaire) Control • No training or health education Quality: The methodological quality of the selected trials was independently assessed by two reviewers (Yang and Ho) using the Physiotherapy Evidence Database (PEDro) scale ( Maher et al 2003, de Morton 2009). Any disagreement with regard to methodological quality were resolved by discussion. Participants: BLZ945 nmr Age, gender, and types of sleep problems were recorded to characterise the trials and to determine the similarity of participants between groups and between trials. Intervention: The target intensity, duration, and frequency of the exercise next training program, and the nature of the control intervention were recorded. Outcome measures: The objectively measured outcomes we considered were sleep onset latency, sleep duration, sleep disturbance, habitual sleep efficiency, daytime dysfunction and use of sleep medication. We also considered subjective measures of sleep quality using standardised instruments or scales, eg, the Pittsburgh Sleep Quality Index ( Buysse et al 1989). The Pittsburgh Sleep Quality Index is a widely used, self-rated sleep questionnaire for

measuring sleep quality. A total of 19 questions generate seven components, each with a score ranging from 0 (no difficulty) to 3 (severe difficulty). The components are subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleep medications, and daytime dysfunction. The seven component scores are also summed to generate a global Pittsburgh Sleep Quality Index score (ranging from 0 to 21), with a score of more than 5 indicating clinical sleep impairment. The analyses were performed using RevMan 5 softwarea. The standardised mean difference (SMD) and a 95% confidence interval (CI) of the post-intervention score or change in scores were calculated. An SMD of 0.5 indicates that the mean of the exercise group is half a standard deviation larger than the mean of the control group. An SMD of 0.8 is considered large, an SMD of 0.5 moderate, and an SMD of 0.2 small.