Pilot investigation shows high sensitivity and specificity using an Abridged Subjective Global Assessment without physical examination in a tertiary hospital; pertinence for use amongst those without COVID-19 when physical distancing required during the pandemic

Summary

Background

Completing malnutrition assessments when physically distant has been an immediate challenge during the COVID-19 pandemic. Even during periods of physical distancing, continuing nutrition assessments amongst those without COVID-19 is vital given that high malnutrition prevalence exists in clinical settings. The investigation aim was to assess the reliability of utilising the validated Subjective Global Assessment (SGA) tool, without compared to with physical examination.

Methodology

Original paper-based SGA documentation from a hospital-wide audit was reassessed by a blinded experienced clinician using history alone without reviewing documented physical examination. Participants included adults admitted to a tertiary hospital with no maternity or obstetric services. Those terminally ill, undergoing end-of-life palliative care, with disordered eating or admitted to emergency or intensive care units were excluded. McNemar's test assessed paired categorical data. Cohen's kappa coefficient assessed inter- and intra-rater reliability. Sensitivity, specificity, positive and negative predictive values were completed.

Results

There was no significant difference in malnutrition identification (p < 0.454) with 97% (473/489) of assessments identical. High sensitivity (87.2%, 68/78), specificity (98.9%, 405/411), positive (91.9%, 68/74) and negative (97.6%, 405/415) predictive values were evident. High inter- and intra-rater reliability was confirmed (kappa values 0.875 and 0.987).

Conclusion

The Abridged-SGA utilising the four key factors of the SGA history identified many malnourished amongst those without COVID-19 who otherwise would not be identified when physical distancing is required due to the pandemic. It did not overestimate malnutrition. Until alternative means of assessing physical parameters remotely are validated, the pragmatic value of practitioners' judgement when utilising the Abridged-SGA was confirmed.