Evaluating Gaps in Care of Malnourished Patients

To assess the prevalence of gaps in care among malnourished adult medicine patients, we conducted a retrospective study of all adult patients (>18 years old) admitted to a medicine unit at NewYork-Presbyterian- Weill Cornell Medical Center from September 1- November 30, 2014 and diagnosed with malnutrition during that admission. Two independent researchers reviewed the medical charts of 242 patients identified through Humedica with either severe or non-severe malnutrition ICD9 codes (262 or 263.0). Each admission/readmission was counted as a separate episode of care. Readmission was defined as those patients readmitted within 3 months of initial admission. Note, we considered defining readmission to be within 30 days of the initial admit date, but the number of participants meeting this criterion in our sample was too small. Patients were excluded if they were admitted to an intensive care unit (ICU) at any point during their stay. Two RDNs reviewed the first 43 charts and inter-reliability was calculated as well as the individual Cohen's kappa for all questions. Reviewers discussed questions with lower agreement in order to avoid future discrepancies.

A gap in communication with medical team was defined as a delay in the approval or implementation of an appropriate nutritional recommendation made by the RDN (i.e. recommendations or pending orders were not verified/ordered within 48 hours). Procedures/testing gap was defined as when a feeding was held in preparation for or related to procedures or testing (i.e. operating room, imaging, swallowing evaluation, etc.). Discharge gap was defined as when the diet order in the discharge instructions given to patients was not consistent with the RDN's recommendations (i.e. RDN recommends an ONS to continue post discharge but discharge orders do not include it).

Fisher's exact and chi-square tests were used to evaluate the following categorical variables: gender, severe versus non-severe malnutrition and malnutrition context. All statistical analysis was performed with Stata/MP 14.1. This study had Institutional Review Board approval.