Limited Agreement on Etiologies and Signs/Symptoms among Registered Dietitian Nutritionists in Clinical Practice

Clinical studies that establish the reliability and validity of nutrition diagnoses are absent from the literature.

The purpose of this study was to assess the reliability and validity of the diagnoses, etiologies, and signs/symptoms within the clinical practice of registered dietitian nutritionists (RDNs).

Nine pairs of RDNs randomly selected adult patients to diagnose. The two RDNs in each pair independently assessed nutritional data from the same patient. Each RDN chose one nutrition diagnosis and rated the presence or absence of the etiologies and signs/symptoms. Clinical reliability was determined with percentage of agreement for the diagnoses, and percentage agreement, kappa coefficient, and the proportions of positive agreement (p(pos)) and negative agreement (p(neg)) for the etiologies and signs/symptoms. Clinical validity was calculated according to a clinical diagnostic validity score for etiologies and signs/symptoms.

These RDNs practiced either in an acute-care facility (n=10; 58.8%), ambulatory/outpatient facility (n=3; 17.6%), or both (n=4; 23.5%). Nutrition diagnoses were selected for 316 patients. Two raters selected the same diagnosis for 121 patients (38% agreement). Agreement was moderate (κ=0.54) for etiologies and at the lower end of substantial (κ=0.63) for signs/symptoms. For etiologies p(pos)=0.71 and p(neg)=0.78 and for signs/symptoms p(pos)=0.82 and p(neg)=0.79, indicating that the raters agreed on the presence of each as well as the absence. The overall clinical diagnostic validity score for etiologies was 0.33 and for signs/symptoms was 0.44, signifying they were not characteristic of the diagnoses.

Although RDNs knew a nutrition problem was present, they were in poor agreement as to the most important diagnosis, etiologies, and signs/symptoms. The lack of agreement indicated inconsistencies in determining a primary diagnosis among these RDNs.