The doctor-patient relationship is the focus of attention in the medical field. General practice clinic is an important place for providing continuous and comprehensive medical services, and the communication ability of general practitioners (GPs) is particularly important for improving the doctor-patient relationship. However, there are few studies on the current situation of doctor-patient communication in general practice clinic in China.

To understand the characteristics and current situation of doctor-patient conversation in general practice clinic, in order to provide empirical support for improving communication skills of GPs, optimizing doctor-patient communication strategies, and building a harmonious doctor-patient relationship.

Convenience sampling method was used to select eight GPs from one tertiary hospital and three community health service centers (stations). Natural observation method was used to collect 100 doctor-patient conversation corpus from May to August 2024. The Roter Interaction Analysis System (RIAS) was used to classify, code and quantify the doctor-patient conversation, describe the topic structure of the conversation, and analyze the discourse differences.

A total of 18 667 doctor-patient utterances were collected; 8 784 from doctors and 9 883 from patients.The verbal dominance ratio was 0.89, the patient-centeredness score was 1.27, and the psycho-biomedical ratio was 0.10 in the general practice clinic. The three most frequent topics of physician discourse were biomedical information giving (28.78%), active dialogue (25.18%) and biomedical information seeking (13.79%), while the three most frequent topics of patient discourse were biomedical information giving (34.68%), active dialogue (28.45%) and social language (14.82%). GPs were more likely to ask for biomedical information (P=0.001) and psycho-social information (P=0.003) than patients, but less likely to give biomedical information (P<0.05) and psycho-social information (P<0.001) than patients. Patients had more social language than doctors (P=0.003). The verbal dominance ratio, biomedical information seeking, biomedical information giving, biomedical advice and guidance, and step language of GPs in tertiary hospitals were higher than those in community GPs (P<0.05), while the psycho-biomedical ratio, psycho-social advice and guidance, and social language of GPs in tertiary hospitals were lower than those in community GPs (P<0.05). The frequency of biomedical information seeking, biomedical information giving and positive dialogue of GP-patient in the tertiary hospitals were higher than those in the community GP-patient (P<0.05), while the social language of GP-patient in the community was significantly higher than that in the tertiary hospitals (P<0.01) .

The patient occupies a dominant position in the discourse of the doctor-patient conversation in the general practice clinic. The theme of the doctor-patient conversation is mainly biomedical, and the degree of "patient-centered" is relatively good. Compared with patients, general practitioners had more "seeking discourses" and less "giving discourses". Compared with the GPs in the community, the GPs in the tertiary hospitals have a higher proportion of language dominance, more biomedical discourse, and less psycho-social discourse. GPs should pay more attention to patients' emotional needs, information acquisition needs and psycho-social factors of the disease, so as to improve doctor-patient communication skills and medical service quality.