SHBC1422

J.C.N.CHU¹, A.Y.M.HUM¹, P.KAUR², Z.J.LOW¹, W.S.TAN², Y.Y.DING³, P.KANNAPIRAN² 

Tan Tock Seng Hospital¹, National Healthcare Group HQ², Yishun Community Hospital³ 

Dementia, particularly dementia type and advanced dementia (AD), is commonly under coded and complex to diagnose. Commonly used assessment tools to determine dementia severity (e.g. Abbreviated Mental Test and Mini-mental status examination) relies on verbal responses which could be hampered by aphasia or issues related to articulation in patients with AD. The absence of a formal AD diagnosis could affect the active management of dementia during hospitalizations and patients’ access to relevant support services. We aim to understand how AD can be determined from Electronic Health Records (EHRs) even when it is not formally documented or assessed by using the keywords described in Reisberg Functional Assessment Staging Tool (FAST).

EHRs of patients who were admitted to Tan Tock Seng Hospital (TTSH) with a primary diagnosis of either “delirium”, “dementia”, “pneumonia”, “sepsis”, or “enteral feeding” and a secondary diagnosis of “dementia” during July 2017 to October 2017 were retrospectively screened to determine AD.

1200 out of 1979 patients were screened to have dementia of advanced severity. Of the 1200 patients, 57 had MMSE scores indicative of AD (MMSE≤10), 3 were diagnosed by clinicians to be at FAST Stage 7 and 297 had an “AD” diagnosis by clinicians. The remaining 843 patients had keywords listed in the EHRs that were descriptive of the FAST stages 7A-7F, depicting AD.

AD is underdiagnosed and the FAST allows for identification of patients who display AD symptoms by identifying commonly used keywords in EHRs that were similar to the FAST 7 staging description of AD.