Hostname: page-component-54dcc4c588-54gsr Total loading time: 0 Render date: 2025-09-19T19:49:45.047Z Has data issue: false hasContentIssue false

Prediction of dementia following traumatic injury with risk score (the DEMTIS): a multivariable prediction model development study based on Hong Kong electronic health records

Published online by Cambridge University Press:  26 August 2025

C. S. Fung*
Affiliation:
Department of Psychiatry
H. F. Chung
Affiliation:
Department of Psychology
S. H. Cheung
Affiliation:
Department of Psychology
N. Dong
Affiliation:
Department of Psychology
I. T. K. Hui
Affiliation:
Department of Psychology
C. S. M. Wong
Affiliation:
School of Public Health, The University of Hong Kong, Hong Kong, Hong Kong
W. C. Chang
Affiliation:
Department of Psychiatry
P. W. Cheng
Affiliation:
Department of Psychiatry
*
*Corresponding author.

Abstract

Core share and HTML view are not available for this content. However, as you have access to this content, a full PDF is available via the ‘Save PDF’ action button.
Introduction

Patients following traumatic injury (TI) are at increased risk of developing dementias, yet an efficient, validated screening instrument is lacking.

Objectives

In the current study, we developed the Dementias following Traumatic Injury Screening (DEMTIS) score, a brief post-TI dementias screening tool.

Methods

We identified 270,843 electronic health records from Hong Kong patients admitted for TI between 2001 and 2021. The records were randomly split into training (80%; n=258,739) and testing cohort (20%; n=50,883). The DEMTIS was developed based on a backward stepwise multivariate Cox proportional hazard model predicting first-ever dementia diagnosis. Competing risk survival analyses were used to predict the risk of Alzheimer’s disease (AD) and vascular dementia (VD), while taking the risk of other dementias into account. Model discrimination of the three scores was evaluated using concordance statistics (c-statistic) calculated as the area under the receiver operating characteristic curve. Statistical significance was set at p<.01.

Results

The 5-year, 10-year, and 20-year risk of all-cause dementias following TI was 2.1% (95% CI 0.020-0.021), 3.8% (95% CI 0.037-0.039), and 6.5% (95% CI 0.063-0.066), respectively. The final model included sex, TI characteristics, physical covariates, history of mood and anxiety disorders, and cerebral degenerative disease (See Table 1). The population mean of DEMTIS was 59.45 (SD=21.29). The optimal threshold of DEMTIS predicting dementia was determined at 75 using the closest top left rule. Individuals at high risk (DEMTIS≥75) were associated with a 6.0% (95% CI 0.059-0.061) risk of dementia in 5 years, whereas those at low risk were associated with a 0.5% (95%CI 0.004-0.005) risk (see Figure 1; Figure 2). The model predicting 5-year dementia has an overall c-statistic of 0.835 (95% CI 0.832-0.839) in the testing data (see Figure 3). We further developed risk scores for 5-year AD and VD based on the findings from competing risk models; the c-statistics of model for AD and VD are 0.857 (95% CI 0.844-0.871) and 0.837 (95%CI 0.821-0.853) respectively.Table 1:

Calculation of the DEMTIS

VariablesDEMTISDEMTIS-ADDEMTIS-VD
Sex
Male000
Female211
Age of TI111
Position of TI
Head injury1-12
Torso121
Upper Limb222
Lower Limb000
Fracture typesNA
Opened-1-1
Closed00
Unspecified-1-1
Hypertension
Yes2-13
No000
Hyperlipidemia
Yes2-12
No000
CCI-1-12
Cerebral degenerative disease
Yes333
No000
Mood and anxiety disorderNA
Yes22
No00

Image 1:

Image 2:

Image 3:

Conclusions

As a novel, easily accessible screening instrument, DEMTIS can identify patients at elevated risk of dementia following TI. It assists clinicians in evaluating patients’ risk of dementia and providing personalized care.

Disclosure of Interest

None Declared

Information

Type
Abstract
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
© The Author(s), 2025. Published by Cambridge University Press on behalf of European Psychiatric Association
Submit a response

Comments

No Comments have been published for this article.