ID-5184 Wonca Abstracts supplement A-K 13-10-23 - Flipbook - Page 285
WONCA 2023 Supplement 1: WONCA 2023 abstracts (A–K)
A
B
C
D
E
F
G
H
I
J
K
Construction and validation of nomogram model for
risk prediction of anxiety in elderly patients with chronic
diseases in China
Mr Yadong Hu, Rongying Wang
河北医科大学第二医院
Objective
To investigate the influencing factors of anxiety in elderly patients with chronic diseases, construct a
risk prediction nomogram model and verify it.
Methods
Using the data of the 2018 China Follow-up Survey of Health Influencing Factors of the Elderly
(CLHLS), a total of 9187 elderly people aged 65 years and above from 23 provincial-level administrative
units in China with chronic diseases were included, and 9187 patients with chronic diseases were
divided into 7349 cases in the model group and 1838 cases in the verification group by random
number table method. In the model group, univariate logistic regression analysis calculated the
influencing factors of anxiety in elderly patients with chronic diseases, and then multivariate logistic
regression analysis was used to construct a nomogram model for risk prediction of anxiety in elderly
patients with chronic diseases. The model was evaluated by area under the curve (AUC) and internal
calibration. The validation group was used as an externally validated dataset to further test the
accuracy of the model through the C-index and decision curve.
Results
Among the 9187 elderly patients with chronic diseases, anxiety occurred in 1176, including 955 in
the model group and 221 in the validation group. The four variables of gender, economic situation,
sleep status and life satisfaction were statistically significant in the multivariate logistic model and were
selected to build the nomogram model. The AUC value was 0.729 (95% confidence interval [CI]: 0.711,
0.746) for the model group and 0.756 (95% CI: 0.721, 0.791) for the validation group.
Conclusion
We developed and verified a nomogram model for risk prediction of anxiety in elderly patients with
chronic diseases, which has good discrimination and calibration. The model has more reliable
predictive performance and has a good net benefit value in predicting the occurrence of anxiety.
283