In Patients Admitted to a Home Rehabilitation Service, Is Remote Completion of the Patient-Generated Subjective Global Assessment Physical Examination Using Still Images Captured by Allied Health Assistants a Valid Alternative to an In-Person Physical Examination?

Abstract

Background: There is increasing provision of telehealth services, including nutrition services. However, remote nutrition assessments are challenging due to difficulties in conducting physical assessments remotely, a crucial component of assessing nutritional status.

Objective: The aim of this study was to evaluate whether remote completion of the Patient-Generated Subjective Global Assessment physical examination using still images captured by allied health assistants (AHAs) is a valid alternative to an in-person physical examination in patients admitted to a home rehabilitation service.

Design: This study was cross-sectional in design.

Participants/setting: This study involved 104 adults admitted to the home rehabilitation service at Southern Adelaide Local Health Network, Adelaide, Australia, over 2 sampling periods in 2019 and 2020 who were receiving home visits by an AHA and were engaged in rehabilitation activities.

Main outcome measures: Validity of the still image-based physical assessment was determined using still images collected by an AHA and an in-person physical assessment completed by a dietitian from each participant. A dietitian blinded to the in-person results later assessed the de-identified still images to determine the presence and extent of deficit at each anatomical site and overall physical examination component of the Patient-Generated Subjective Global Assessment.

Statistical analyses performed: Percentage agreement, weighted κ, sensitivity, and specificity between the still image based and in-person physical examinations were determined to assess agreement between the 2 methods of assessment.

Results: The still image based physical examination achieved a percentage agreement of 75% against the in-person examination, with a weighted κ of 0.662 (95% confidence interval 0.516-0.808) and a sensitivity-specificity pair of 76.6% and 89.1%.

Conclusions: Physical examination using still images collected by AHAs achieved percentage agreement, κ, and sensitivity and specificity compared with an in-person physical examination that is consistent with or superior to commonly adopted nutrition screening and assessment tools. There is potential for implementation of this method to facilitate remote nutritional assessments by dietitians; however, further work is needed to ensure dietitians are able to assess still images reliably.

Keywords: Nutrition assessment; Patient-Generated Subjective Global Assessment; Physical examination; Remote.