Five-Year Mortality Outcomes for Eating As Treatment (EAT), a Health Behavior Change Intervention to Improve Nutrition in Patients With Head and Neck Cancer: A Stepped-Wedge, Randomized Controlled Trial

Abstract

BackgroundMalnutrition affects up to 80% of patients with head and neck cancer (HNC), and is associated with higher burden of disease, poorer treatment outcomes, and greater mortality. The Eating As Treatment (EAT) intervention is a behavioural intervention, previously demonstrated to be effective for improved nutritional status, depression and quality of life, in patients with HNC. This paper examines the impacts of the EAT intervention on five-year mortality among participants. MethodsA multi-centre stepped-wedge randomised controlled trial was conducted in five Australian hospitals. Dietitians were trained to deliver EAT, a combination of motivational interviewing and cognitive behaviour therapy strategies, to patients with HNC receiving radiation therapy. This paper describes secondary analyses of survival benefit on an intention to treat basis. Differences in proportions of 5-year all-cause mortality between control and EAT intervention arms were analysed using multivariable logistic regression and 5-year survival rates were analysed using Cox proportional hazard regression. Analyses controlled for temporal effects (study time), hospital site (clustering), and baseline nutritional status differences. ResultsOverall, there were 64 deaths in the 5 years following enrolment; 36 (24%) among those assigned to the control condition and 28 (18%) among those assigned to EAT. Logistic regression showed statistically significant reduced odds in favour of EAT (OR= 0.33, 95%CI 0.11-0.96), with an absolute risk reduction of 17%(0.01 to 0.33%), a relative risk reduction of 55% (0.22-0.92), resulting in the number needed to treat of 6 (4 -13). Survival analysis found that risk of death was significantly reduced by the EAT intervention (Hazard Ratio 0.39, 0.16-0.96). ConclusionsEAT showed a statistical and clinically meaningful survival benefit, probably via improved nutrition during radiotherapy. This survival benefit strengthens the finding of the main trial showing that a behavioural intervention focussed on nutrition could improve HNC outcomes. Replication studies using stepped-wedge designs for implementation into clinical practice may be warranted.