This paper reviews the use of sleep-promoting medications in nursing home residents with reference to risks versus benefits. Up to two-thirds of elderly people living in institutions experience sleep disturbance. The aetiology of sleep disturbance includes poor sleep hygiene, medical and psychiatric disorders, sleep apnoea, periodic limb movements and restless leg syndrome. One key factor in the development of sleep disturbance in the nursing home is the environment, particularly with respect to high levels of night-time noise and light, low levels of daytime light, and care routines that do not promote sleep. Clinical assessment should include a comprehensive medical, psychiatric and sleep history including a review of prescribed medications. Nonpharmacological interventions for insomnia are underutilised in many clinical settings despite evidence that they are often highly effective. International studies suggest that 50-80% of nursing home residents have at least one prescription for psychotropic medication. Utilisation rates vary dramatically from country to country and from institution to institution. The most commonly prescribed medications for sleep are benzodiazepines and nonbenzodiazepine hypnotics (Z-drugs). The vast majority of studies of these medications are short-term, i.e. < or =2 weeks, although some longer extension trials have recently been carried out. Clinicians are advised to avoid long-acting benzodiazepines and to use hypnotics for as brief a period as possible, in most cases not exceeding 2-3 weeks of treatment. Patients receiving benzodiazepines are at increased risk of daytime sedation, falls, and cognitive and psychomotor impairment. Zaleplon, zolpidem, zopiclone and eszopiclone may have some advantages over the benzodiazepines, particularly with respect to the development of tolerance and dependence. Ramelteon, a novel agent with high selectivity for melatonin receptors, has recently been approved in the US. Use of the antidepressant trazodone for sleep in nondepressed patients is somewhat controversial. Atypical antipsychotics should not be used to treat insomnia unless there is also evidence of severe behavioural symptoms or psychosis.