ID-5184 Wonca Abstracts supplement A-K 13-10-23 - Flipbook - Page 323
WONCA 2023 Supplement 1: WONCA 2023 abstracts (A–K)
A
B
C
Predicting mid- and late-life dementia risk in the primary
care setting: A nationwide screening population study
Dr Wonyoung Jung5, Sang Hyun Park2, Kyungdo Han3, Dong Wook Shin1,4
D
E
F
G
1
Department of Family Medicine and Supportive Care Center, Samsung Medical Center, 2Department
of Biostatistics, College of Medicine, The Catholic University of Korea, 3Department of Statistics and
Actuarial Science, Soongsil University, 4Department of Clinical Research Design and Evaluation,
Samsung Advanced Institute for Health Science and Technology (SAIHST), Sungkyunkwan University,
5
Kangdong Sacred Heart Hospital, Hallym University
Purpose
We aimed to develop a risk assessment model for predicting mid- and late-life dementia risk in a
primary care setting using a nationwide population-based screening cohort.
H
I
J
K
Methods
Individuals aged ≥40 years who participated in a general health examination in 2009 were used for
model development and validation. Age, body mass index (BMI), lifestyle behaviours, family history,
laboratory measurements and comorbidities were used to develop a five-year dementia risk model
for mid-life (40–59 years) and late-life (≥60 years) populations. The best-fit risk prediction models
were constructed by stepwise backward selection using the Cox proportional hazards model. The
risk model was converted into nomograms of risk score. Model performance was evaluated by
discrimination and calibration.
Results
During mean follow-up of 8.95 years, and with 59,947,524.07 person-years, we identified 349,292
cases of incident dementia (31,502 for mid-life, 317,790 for late life). The final risk model included 10
variables: age, sex, BMI, smoking habits, alcohol consumption, physical activity, type 2 diabetes,
hypertension, dyslipidaemia and estimated glomerular filtration rate. The concordance index of the risk
model was 0.764 (95% CI: 0.760, 0.769) for mid-life, and 0.743 (95% CI: 0.741, 0.744) for late life. Our
model correlated well in late life but overestimated the dementia risk in mid-life with high-risk estimates.
Conclusions
We developed a mid- and late-life dementia risk model using predictors readily available in primary
care settings with good performance. This can have clinical implications for risk-based cognitive
screening in primary care.
321