Referral of patients from long-term to acute-care facilities.

Medical care available to residents of nursing homes and chronic-care hospitals was assessed by studying transfers of such persons to the emergency room of an acute-care hospital. One hundred patients transferred from nursing homes and 16 patients transferred from chronic-care hospitals were compared with 338 elderly patients from home (control group). Elevated temperature (greater than or equal to 102 degrees F) was found in 17.3 per cent of nursing home patients and 30.0 per cent of chronic-case hospital patients, compared with 1.8 per cent of controls (P less than .05); mental status abnormalities were found in 66.1 per cent of patients from nursing homes and in 90.9 per cent of those from chronic-care hospitals, compared with 36.2 per cent of controls (P less than .025). In addition, patients from chronic-care hospitals, but not those from nursing homes, often showed substantial abnormalities of blood pressure and pulse. Thus, the patients from nursing homes tended to be slightly sicker than controls, and those from chronic-care hospitals considerably sicker. The probability of requiring admission to the hospital was the same for residents of nursing homes and persons living at home (44.0 per cent and 43.2 per cent, respectively), but was higher for persons from chronic-care hospitals (81.3 per cent, P less than .005). It was also determined whether a disproportionately large percentage of transfers from extended-care facilities occurred at night or on weekends. Among nursing home patients, 12.6 per cent of emergency room visits occurred on weekends, compared with 20.0 per cent of chronic-care hospital patients and 24.3 per cent of controls. Thus, no evidence for "dumping" of patients was found. It was concluded that utilization of a hospital emergency room by nursing home patients is very similar to that by home residents, suggesting an inadequacy of on-site medical services, whereas that by chronic-care hospital patients is restricted to major illness, which is entirely appropriate.