Abstract: Objective To identify latent profiles of dyspnea beliefs in elderly patients with severe bronchial asthma and analyze the influencing factors of different latent profiles, in order to provide a basis for implementing individualized psychological interventions. Methods A convenience sampling method was used to select elderly patients with severe bronchial asthma admitted to Shanghai Changzheng Hospital from June 2024 to June 2025 as the study subjects. Data were collected using a general information questionnaire, the Breathlessness Beliefs Questionnaire, the Self-Rating Anxiety Scale (SAS), and the modified British Medical Research Council Dyspnea Scale. Latent profile analysis was employed to identify potential categories of dyspnea beliefs, and multivariate logistic regression analysis was used to explore the influencing factors. Results A total of 306 patients were included, with an effective response rate of 98.37% (301/306). Latent profile analysis revealed that the dyspnea beliefs of elderly patients with severe bronchial asthma could be categorized into three distinct profiles: a low-beliefs group (18.9%), a moderate-beliefs-avoidance group (40.9%), and a high-beliefs-fear group (40.2%). Multivariate logistic regression analysis indicated that monthly household income per capita, dyspnea severity, regular physical exercise, combined with chronic obstructive pulmonary disease, regular follow-up, and SAS score were independent influencing factors for different dyspnea belief profiles. Conclusions Dyspnea beliefs among elderly patients with severe bronchial asthma exhibit significant heterogeneity and are influenced by a combination of physiological, psychological, and social factors. Healthcare providers should recognize the characteristic differences among patients with different profiles, with a particular focus on those with multiple risk factors, to develop targeted intervention strategies.

Key words: aged, severe bronchial asthma, dyspnea beliefs, potential profile analysis, influencing factor