Hospital malnutrition has been widely recognised as a problem for some decades. It has been shown that poor nutritional status of patients is associated with worsening of physical functions, increased probability of complications, longer recovery and, because of that, higher treatment costs. Regrettably there is no common (and widely accepted) policy on how to detect malnourished patients or those at risk. However, many screening tools have been developed with this purpose. In the UK, MUST scoring is recommended, and was adopted more than 2 years ago by the University College London Hospitals (UCLH) as a mandatory screening tool. Although screening tests are meant to be simple and uncomplicated it has appeared that they are being performed less often and less well than hoped and expected. Concerning the usefulness and accuracy of the MUST scoring, we now address the following questions: “What are the real differences in the conditions of patients with widely diverse screening results?” and “Does the professional background of a person performing the screening have any influence on the results?”

The MUST scores obtained by nursing staff of 53 patients admitted to the Gastroenterology/Gastrointestinal surgery ward of UCLH were recorded. Cases lacking data were noted. Subsequently a new MUST score for each patient was calculated by a dietician and an additional set of data was obtained to satisfy the NRS and SNAQ screening tools and to complete the ESPEN Nutrition Day questionnaire. Comparison of data from medical/nursing notes and from the trained dietician was made.

Hospital staff had screened only 36 of a group of 53 patients (67.9%) at the time of review. In this subgroup of 36 patients, the observations of the regular clinical staff were compared with those of the dietician (reference). Pre-assessment and ward staff identified 7 patients with a score of 2 or higher (threshold to refer to a dietician), while the dietician recognised each of these and a further 10 patients (all in patients screened in the pre-assessment context). No overrated scores were found; that is, the dietician did not downgrade any results greater than or equal to 2 obtained by the clinical team although there were minor discrepancies which would not have affected the decision to refer for dietetic assistance (e.g. 6 versus 4, or 3 versus 5).

Concerning comparison of NSR, MUST (by specialist dietician) and SNAQ screening tests, the following results were obtained. All tests were concordant on recommending referral to the dietician in 13 out of 53 cases (24.5%), which is in line with estimates found in current literature. Two or more tests agreed in 26 cases (49.1%), while each alone would have prompted referral in 32, 22 and 22 cases, respectively. This suggests that there is still a place to improve their accuracy or at least decrease their tolerance.

There is a strong need to improve the rate of patients being screened by hospital staff and the accuracy of the results. Further research regarding the means to achieve these goals is needed, whether by further education of the staff or by modification of the screening tools, none of which appears fully able to provide all that is required as currently implemented.