Cost Effectiveness of an Intervention in Hospitalized Patients With Disease-related Malnutrition

Objectives

1. -After a nutritional screening of all patients, evaluate the effectiveness of an intervention in patients admitted to a hospital in relationship to disease-related malnutrition (DRM) by comparing 3 cohorts: cohort-1: organized intervention to patients with DRM or at risk; group 2: intervention on demand by health personnel; Group 3: usual care without any explicit intervention. Effectiveness will be evaluated in relation to multiple outcomes such as: length of hospital stay of the index admission; and the following parameters during admission and until one year after admission, mortality, occurrence of complications (mainly infectious, or surgical in surgical patients), readmissions, use of health resources after discharge (primary care consultations, hospital and emergency consultations), changes in weight during admission and up to one year, changes in their health related quality of life (SF-36 and/or EQ-5d) and fragility (Barthel I.) from admission to one year after admission.
2. -Evaluate the cost and cost-effectiveness of those strategies.
3. - Evaluate in the three groups the evolution of patients in relation to the previous parameters of results according to their classification after screening at the time of admission to DRM, risk of DRM or no problem of DRM detected. Secondary Objectives
4. - Evaluate the three groups of patients classified according to their level of DRM according to their level of fragility and in relation to their evolution during admission and up to one year (based on all the outcome parameters indicated above)
5. - Identify risk factors of worse evolution during hospital admission and up to one year of follow up based on the previously described outcome parameters and study the role that the DRM level plays as predictor of evolution.

DESIGN: Prospective cohorts with three branches. SCOPE: Three public hospitals. The first center, where the study intervention will be carried out (branch 1), is the Galdakao - Usánsolo University Hospital that serves a population of about 300,000 urban inhabitants and semi-urban The second center, the University Hospital of Basurto (HUB), where the intervention will be carried out by demand (branch 2), serves the population of Bilbao, some 350,000 urban inhabitants. The third center, Donostia University Hospital (HUD) where there will be no specific organized intervention (branch 3), attends the population of San Sebastián-Donostia and surrounding areas covering a population of about 400,000 inhabitants of urban and semi-urban characteristics. All of them have similar human and technological resources being the population that attend similar on sociodemographic and clinical characteristics.

SUBJECTS OF THE STUDY. See detailed information in other sections. SAMPLE SIZE. It is estimated at least 10-15% of DRM and up to 30-35% of patients at risk of DRM (with a global of almost 40-50%) in the selected pathologies of patients.

Sample size calculation. We have made the estimates for an alpha error of 0.05 and a power of 80%. We have chosen the following outcome parameter in which we expect relevant differences between branches: we estimate that the average length of stay will decrease from an average of 12 days (DS: 5) in the DRM group or at risk without any explicit intervention (HUD) to an average of 10 days (DS: 5) in the DRM group or at risk where the intervention will be performed (HGU) for which we will need 100 patients by each comparisons group but we will try to recruit at least 100 patients from the previous three nutritional groups per center. We estimate a percentage of losses in the follow-up of 10-15% so, overall, 350 patients per center will be sufficient to meet the objectives and respond to the hypotheses of the study. Data from 2017 for each hospital and established patients selection criteria indicate that we should have more than those patients admitted by hospital in 12 months.

DESCRIPTION OF THE SCREENING: all admitted patients of the three centers that meet the selection criteria will be evaluated in the first 72 hours after admission and again upon discharge (if a patient remained admitted more than one week, this assessment will be repeated weekly). The nutritional risk assessment will be done by applying the Malnutrition Universal Screening Tool (MUST) and the Global Leadership Initiative on Malnutrition (GLIM) questionnaires.

VARIABLES TO COLLECT: in summary, A.-From the clinical history: Sociodemographic data will be collected (age, gender, educational level ..) and clinical data such as: comorbidities, based on the Charlson Comorbidity Index , weight, height, health habits, usual treatments (dose, number of drugs), diagnosis of admission (Major Diagnostic Categories ICD-10), severity of pathology (TNM if pathology tumor), source of admission (emergency vs scheduled), admission service. During admission: complications, treatments, procedures and referral to special units (ICU,...). Analytical data available (biochemistry and blood cells count, including urea, creatinine, glucose, electrolytes, albumin, pre-albumin, lymphocytes, total cholesterol, triglycerides, transferrin, and liver function tests). This analytics will be obtained in the first 48 hours of admission and in surgical patients before the intervention if scheduled. They will be repeated at discharge and weekly in prolonged income.

Main results of interest: see separate section. B.-Health-related quality of life questionnaires: 1.- The patient's functional will be calculated using the Barthel index. 2. - General health and utilities: they will be measured by the EuroQol-5d questionnaire, which will mainly be used for economic analysis. 3.-SF-36.

STATISTIC ANALYSIS. A descriptive analysis of the entire sample and univariate analysis to determine potential variables related to the outcomes of interest will be performed. In the multivariable analysis, different models will be used according to the dependent/outcome variable of interest. In the case of dichotomous dependent variables (mortality, complications, readmissions ...) logistic regression models will be used. For continuous dependent variables (changes in HRQoL) generalized linear models will be used; and for dependent variables that have other distributions, such as length of stay, Poisson regression models will be used. The main comparison will be between the intervention groups (HGU) versus intervention by demand (HUB) versus control group (HUD). In all multivariables analyzes, the use of a Propensity Score (PS) as an adjustment system for the differences between the characteristics of the patients of the three centers will be considered.

Economic analysis: A cost-effectiveness analysis of the intervention corresponding to the first branch will be carried out with respect to the second and third branches. Only medical costs associated with medical assistance will be included. The measure for the cost-effectiveness analysis will be the incremental cost for each year of life adjusted by Quality Adjusted Life Year (QALY) earned. The QALY will be calculated from the EuroQol-5d-5l questionnaire scores, in the baseline moment and one year after admission, and the mortality variable up to the year.

PROBLEMS AND LIMITATIONS OF THE STUDY. 1.-As in any study prospective study, main problem is the losses in the follow up. To reduce them, patients will be fully informed and a close contact will be established during follow-up to all patients. 2.-This project is proposed as an intervention study adapted to the current circumstances of the usual clinical care practice, which should guarantee a greater feasibility of a subsequent implementation of this patient care strategy if our results show its benefits. 3.- The comparability between the different cohorts will be guaranteed in the statistical analysis by the adjustment by the propensity scores. 4.-The strategy of the study poses an ethical problem since after screening in all the centers some of the patients will be identified as with DRM or at risk and therefore this will be communicated to the physicians in charge of each patient so that they proceed in the way they create timely by writing down their attitude therapy. We will collect information about any additional nutritional strategy given to any patient.

ETHICAL AND CONFIDENTIALITY ASPECTS. The project has been evaluated and approved by the research commissions of all the participant centers and approved by the accredited Clinical Research Ethics Committee (CEIC) of the Basque Country. EU and Spanish laws on personal data protection will be followed, ensuring that patient information data obtained cannot be associated with identifiable persons. Study data base will be encrypted and protected to assure confidentiality. Only research personal will have access to it with specific passwords.