This study will test whether a personalized mobile application that provides temperature-related health risk warnings and protective recommendations can reduce acute exacerbations in people with chronic obstructive pulmonary disease (COPD). Researchers will compare participants who receive the mobile application in addition to usual care with those who receive usual care alone. The main question is whether the mobile application reduces the number of COPD acute exacerbation events over 12 months. The study will also evaluate secondary outcomes including quality of life, functional capacity, symptoms, psychological status, physiological and biochemical indicators, environmental temperature risk perception, and participants' use of the mobile application.

This study is a proof-of-concept study using a cluster randomized controlled design, with hospitals serving as the unit of randomization. Four eligible hospitals in Beijing will be selected as study sites, and each hospital together with all eligible participants enrolled at that site will be treated as one cluster. Clusters will be randomly assigned in a 1:1 ratio to the intervention group or the control group. In addition to usual COPD care, participants in the intervention group will receive temperature-related health risk warning messages and corresponding protective recommendations generated from real-time meteorological data and individual characteristics through the personalized mobile application. Participants in the control group will receive usual COPD care only without access to the mobile application. The study includes a baseline assessment and a 12-month follow-up period. Study visits will occur at baseline and at 3, 6, 9, and 12 months after enrollment. Baseline assessments will include demographic information, air-conditioner use, smoking and alcohol consumption, health status, acute exacerbation events, quality of life, functional capacity, symptoms, psychological status, environmental temperature perception, and physiological and biochemical indicators. Follow-up assessments will collect acute exacerbation events, quality of life, functional capacity, symptoms, psychological status, physiological and biochemical indicators, environmental temperature perception and participants' use of the mobile application. Due to the cluster design and the nature of the mobile application intervention, study participants and research staff who collect visit data will be unmasked. However, data analysts and members of the endpoint adjudication committee will remain masked to the randomization assignments.

Hospital inclusion criteria:

1. Tertiary hospitals;
2. Hospitals with respiratory wards that admit patients with COPD;
3. Hospitals with a well-established electronic health record system and follow-up capability;
4. Hospitals able to assign a study coordinator.

Patient inclusion criteria:

1. Age >=40 years and <=80 years, with no restriction on sex;
2. Patients with stable chronic obstructive pulmonary disease;
3. At least 1 severe acute exacerbation requiring hospitalization, or at least 2 moderate acute exacerbations requiring outpatient or emergency care, within the previous year;
4. Local permanent residents with a duration of residence >=6 months;
5. Patients who voluntarily agree to participate, are expected to have good adherence, are able to sign informed consent;
6. Patients who own a smartphone or other smart device.

Hospital exclusion criteria:

1. Primary hospitals or private hospitals;
2. Hospitals unable to cooperate with the study protocol.

Patient exclusion criteria:

1. Patients with asthma;
2. Women who are known to be pregnant or breastfeeding, or who have a positive pregnancy test before cluster randomization;
3. Patients with other diseases that may affect participation in the trial, such as refractory hypertension or severe aphasia;
4. Patients currently receiving psychiatric or psychological treatment that may contaminate study results;
5. Expected survival time <1 year, such as in those with malignant tumors or severe cardiopulmonary disease;
6. Participation in other interventional clinical studies that may affect outcome assessment;
7. Any other circumstances in which the investigator considers the patient unsuitable for participation in this study or at significant risk, such as cognitive impairment preventing understanding of and/or compliance with study procedures and/or follow-up.

Due to the cluster design and the nature of the mobile application intervention, study participants and research staff who collect visit data will be unmasked. However, data analysts and members of the endpoint adjudication committee will remain masked to the randomization assignments.