Does receiving referral and transfer patients make hospitals expensive?

In 1984-1985 Medicare data, hospital cost per case was strongly correlated both with the percentage of patients admitted as transfers (r = 0.314, P less than 0.0001) and the hospital's referral index (r = 0.512; P less than 0.0001). These relationships remained highly significant after correcting for the hospital wage level, case mix index, teaching activity, urban or rural location, and number of beds. When individual patients in the same hospital and the same diagnosis-related group (DRG) were compared, charges were 37.5 +/- 0.9% higher for transfer admissions, which accounted for only 31.3% of the higher costs of hospitals that received transfers, and charges were not significantly related to referral status (P greater than 0.10). At least for Medicare patients, therefore, the higher costs of transferred patients accounted for less than one third of the higher costs of hospitals that receive such patients, and the costs associated with referred patients did not account for the higher costs of hospitals receiving referrals. Those hospitals may be more expensive because of the features that attract transfers and referrals.