BMC Psychiatry (Apr 2023)

Construction and validation of cognitive frailty risk prediction model for elderly patients with multimorbidity in Chinese community based on non-traditional factors

  • Shuzhi Peng,
  • Juan Zhou,
  • Shuzhen Xiong,
  • Xingyue Liu,
  • Mengyun Pei,
  • Ying Wang,
  • Xiaodong Wang,
  • Peng Zhang

DOI
https://doi.org/10.1186/s12888-023-04736-6
Journal volume & issue
Vol. 23, no. 1
pp. 1 – 10

Abstract

Read online

Abstract Background and objectives Early identification of risk factors and timely intervention can reduce the occurrence of cognitive frailty in elderly patients with multimorbidity and improve their quality of life. To explore the risk factors, a risk prediction model is established to provide a reference for early screening and intervention of cognitive frailty in elderly patients with multimorbidity. Methods Nine communities were selected based on multi-stage stratified random sampling from May–June 2022. A self-designed questionnaire and three cognitive frailty rating tools [Frailty Phenotype (FP), Montreal Cognitive Assessment (MoCA), and Clinical Qualitative Rating (CDR)] were used to collect data for elderly patients with multimorbidity in the community. The nomogram prediction model for the risk of cognitive frailty was established using Stata15.0. Results A total of 1200 questionnaires were distributed in this survey, and 1182 valid questionnaires were collected, 26 non-traditional risk factors were included. According to the characteristics of community health services and patient access and the logistic regression results, 9 non-traditional risk factors were screened out. Among them, age OR = 4.499 (95%CI:3.26–6.208), marital status OR = 3.709 (95%CI:2.748–5.005), living alone OR = 4.008 (95%CI:2.873–5.005), and sleep quality OR = 3.71(95%CI:2.730–5.042). The AUC values for the modeling and validation sets in the model were 0. 9908 and 0.9897. Hosmer and Lemeshow test values for the modeling set were χ2 = 3.857, p = 0.870 and for the validation set were χ2 = 2.875, p = 0.942. Conclusion The prediction model could help the community health service personnel and elderly patients with multimorbidity and their families in making early judgments and interventions on the risk of cognitive frailty.

Keywords