A comparison study of the ability of various ergonomically based methods for transferring patients between wheelchairs and toilets to prevent back problems was performed. Five manual methods were evaluated: lifting by two persons, gently rocking and pulling the patient with a gait belt (method-A) or walking belt (method-B) using two persons, and rocking and pulling the patient with a walking belt or sling using one person. The mechanically assisted methods utilized hoists with a fixed base (hoist-T), an adjustable base (hoist-H), or a semiadjustable base (hoist-A). The methods were tested during simulated patient lifting using six female nursing students, mean age 22.5 years. The lifts were videotaped and goniometric measurements were made of the subjects' postures. Pulling forces exerted on the gait and walking belts and sling were measured. The subjects were asked to rate their perceived levels of exertion during the simulated lifts. Mean trunk flexion moments, erector spinae forces, and L5/S1 compressive forces during the lifts varied from 93 to 200 newton meters, 1,861 to 4,100 newtons (N), and 1,974 to 4,800N, respectively. The highest forces occurred during the two person manual lift and the lowest when method-A and method-B were used. Manual lifting was judged to be the most stressful and the least comfortable and secure. Hoist-A was rated the least stressful and the most comfortable and secure. An intervention study using method-B and hoist-A was conducted in two units of a nursing home. Fifty seven nursing assistants and 140 patients were involved in the study. After being trained in using the lifting technique, back injury data of the nursing assistants were recorded for 4 months and compared to data for the 4 years preceding the intervention. The nurses were asked to rate their perceived level of exertion associated with using the device. Nurses who used method-B and hoist-A rated their work as very light compared to being hard or moderately hard before the intervention. The incidence and back injury severity rates were reduced from 83 to 43 and 634 to 0 per 100, respectively, after the intervention. The authors conclude that use of ergonomically based methods for transferring patients shows promise for reducing back injuries to nursing personnel.