Abstract
BACKGROUND: Malnutrition is prevalent among patients within certain cancer types. There is lack of universal standard of care for nutrition screening and a lack of agreement on an operational definition and on validity of malnutrition indicators. OBJECTIVE: In a secondary data analysis, we investigated prevalence of malnutrition diagnosis with 3 classification methods using data from medical records of a National Cancer Institute-designated comprehensive cancer center. METHODS: Records of 227 patients hospitalized during 1998 with head and neck, gastrointestinal, or lung cancer were reviewed for malnutrition based on 3 methods: (1) physician-diagnosed malnutrition-related International Classification of Diseases, Ninth Revision codes; (2) in-hospital nutritional assessment summaries conducted by registered dietitians; and (3) body mass indexes (BMIs). For patients with multiple admissions, only data from the first hospitalization were included. RESULTS: Prevalence of malnutrition diagnosis ranged from 8.8% based on BMI to approximately 26% of all cases based on dietitian assessment. κ coefficients between any methods indicated a weak (κ = 0.23, BMI and dietitians; and κ = 0.28, dietitians and physicians)-to-fair strength of agreement (κ = 0.38, BMI and physicians). CONCLUSIONS: Available methods to identify patients with malnutrition in a National Cancer Institute-designated comprehensive cancer center resulted in varied prevalence of malnutrition diagnosis. A universal standard of care for nutrition screening that uses validated tools is needed. IMPLICATIONS FOR PRACTICE: The Joint Commission on the Accreditation of Healthcare Organizations requires nutritional screening of patients within 24 hours of admission. For this purpose, implementation of a validated tool that can be used by various healthcare practitioners, including nurses, needs to be considered.