DEVELOPING AN EFFECTIVE ADULT NUTRITION SCREENING TOOL FOR A COMMUNITY-HOSPITAL

Objective To develop a sensitive and specific nutrition screening tool that conforms to the requirements of the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) and is effective in any given hospital population. Design A two-step prospective evaluation was performed in two unrelated community hospitals to determine the effectiveness of a current nutrition screening tool, and to develop a new, more sensitive tool that permits efficient use of available personnel. Statistical analyses Standard calculations for sensitivity, specificity, positive predictive value, and negative predictive Value were used to compare the results of various screening methods with those of full nutrition assessments. Stepwise discriminant analysis was used to determine specific parameters that contributed to the diagnosis of malnutrition and to derive a predictive equation based on these parameters. Methods In part 1, 100 patients admitted consecutively were evaluated using a subjective screening tool; the results were compared with those from full nutrition assessment of the same patients. Stepwise discriminant analysis identified three parameters most closely related to malnutrition, and a multivariant equation relative to nutrition risk was derived. in part 2, 151 randomly selected adult medical and surgical patients were evaluated at admission with the same Screening tool used in part 1. Additionally, prealbumin level was determined on admission. The results of the screening, the discriminant analysis equation from part 1, and both methods incorporating the prealbumin level were compared with the results of a full nutrition assessment. Results In part 1, discriminant analysis identified total lymphocyte count, percentage weight loss, and serum albumin level as the three best indicators of malnutrition. A multivariant equation incorporating these objective parameters was derived. In part 2, the equation yielded better sensitivity and specificity results than other screening methods. Inclusion of prealbumin level did not improve screening methods. Conclusions The revised screening tool described can be used as the basis of an effective screening program that meets the proposed JCAHO nutrition care standards.