Automated malnutrition screening for use with community dwelling older adults

Overall problem to be addressed in the subsequent external funding proposal: Malnutrition affects 33% (95%CI 27-39%) of people in the community setting.1 This condition leads to reduced quality of life, reduced immune function and healing, reduced muscle strength and function, and increased health service utilisation and costs.2 As the population of older Australians grows, there is a growing need to optimise care provision in the community setting, supporting individuals to live healthier lives in their own homes. Functional decline due to malnutrition is difficult to reverse, thus, early management is critical.3 However, most older adults are not routinely screened for malnutrition, leading to delayed management.4 The subsequent external funding proposal will aim to develop and implement a malnutrition early detection and management clinical pathway for use in the community setting.
Seeding grant proposal aim: To modify an automated malnutrition screening tool for use with community dwelling older adults.
Importance of seeding grant proposal to the larger project: Identification of malnutrition is a pre-requisite to implementing any malnutrition intervention. While there are several validated malnutrition screening instruments,2 these all require individuals to have a consultation with a trained assessor. This seeding grant proposal to develop automated malnutrition screening using routinely collected data supports the first step of early identification required for the larger project. Automated screening will enable malnutrition identification that is scalable with minimal additional cost or workforce requirements.
Consumer/partner engagement within seeding proposal: The direction of the larger project was guided by the Residential Opinions of Ageing Reference group (ROAR) which was formed as part of prior work on improving malnutrition in the residential aged care setting. The ROAR includes perspectives from consumers with lived experience, aged care providers, clinicians, researchers, and policy advocacy bodies.5 The ROAR identified concerns about individuals arriving from the community to residential aged care already malnourished, leading to the proposed study focused on pre-emptively addressing malnutrition in the community setting. The ROAR were involved in the development of an automated malnutrition screening in the residential setting and will also contribute to the modification of this approach to the community setting (see methods). This seeding proposal is also being conducted in partnership with Feros Care who were involved in the development of the research plan and will facilitate recruitment of participants.
Seeding grant methods:
Study design: A diagnostic prediction model modification study will be conducted. The developed model will be based on the automated malnutrition screening instrument that uses only routinely collected data developed for the residential aged care setting,6 and modified for the community setting. The reference standard used for diagnosing malnutrition will be the Subjective Global Assessment where malnutrition status is determined based on dietitian conducted medical history and physical examination.7
Participants: Participants will be recruited from partner organisation Feros Care, an in-home care aged care provider. Participants will be eligible if they are aged ≥65 years and have a home care package. A target of 60 participants is sought.
Data collection: Data will be collected from residents by dietitians and extracted from Feros Care records.
Data analysis: The automated malnutrition screening indicator will be derived using multivariate binary logistic regression.
Model interpretation: Consultation with the ROAR will be used to determine the appropriate threshold probability to use as a cut-off point for classifying malnutrition based on the prediction model. This requires consideration for the consequences of false negatives (i.e not treating malnourished individuals) versus false positives (i.e. treating normally-nourished individuals).8
Feasibility: The project will be conducted from June to December 2024. Timely commencement and progress of the study is supported by prior ROAR input to study concept, existing written agreement from partner organisation Feros Care, existing ethics approval (GUHREC #2024/191), and an existing workforce of research-trained dietitians for data collection. Recruitment and data collection is expected to occur over 4-months from June-September, followed by data-analysis, ROAR group engagement, and write-up completed by December 2024.