Development of a screening tool for assessing risk of undernutrition in patients in the community

Background: This study was undertaken to develop a simple tool for assessing risk of undernutrition for use by the primary health care team. Methods: We devised a list of 25 questions that might be predictive of nutritional risk. Community nurses administered the questions, which required only yes/no answers, to 507 patients in their care. Within 4 days a dietitian carried out a full nutritional assessment. Discriminant analysis was used to determine which questions were predictive of nutritional risk as assessed by a dietitian. Multiple regression analysis was used to derive a simple equation to assign weightings to those questions. Results: Five hundred and seven patients completed the study. Nine questions were found to be associated with nutritional risk as assessed by a dietitian. The final tool, which has three categories of scores: 0-6 = not at risk; 7-16 = possible/probable risk; greater than or equal to 17 = malnourished, has a positive predictive value of 94.6% and a negative predictive value of 81.1%. Conclusion: This simple tool can be used to assess risk of undernutrition in patients in primary care, to increase awareness of the importance of nutrition in patients in the community and to ensure that patients are given appropriate dietary advice and/or referred to a dietitian.