【摘要】 目的：综合医院护士是识别和处理躯体-精神共病患者的核心力量，但现行教育培训体系下其精神卫生专业储备常面临瓶颈。本研究旨在通过潜在剖面分析(latent profile analysis,LPA)，精准识别重庆市三级综合医院护士心理健康素养的潜在类别，并探讨不同群体剖面的特征及其社会人口学、职业环境相关因素，从而为制订靶向性、分层精准干预策略提供科学依据。方法：本研究采用横断面调查设计和多阶段分层整群抽样方法，于2023年4月至5月，基于重庆市38个区县的国内生产总值进行高、中、低经济发展层级分层后，在各层级中随机抽取共9家三级综合医院，然后以科室为单位，对符合纳入与排除标准的在岗护士进行整群调查。研究工具包括自行设计的一般资料问卷及中文版心理健康素养量表(Chinese version of the Mental Health Literacy Scale,MHLS-C)；其中，对MHLS-C的分析包括心理障碍知识(以下简称“知识”)、寻求信息和帮助的能力(以下简称“能力”)、心理障碍的识别(以下简称“识别”)及对精神疾病患者的接纳(以下简称“接纳”)4个维度。使用LPA进行模型拟合，评价指标包括信息熵及各类拟合准则；由于存在医院与研究对象的嵌套结构，采用多变量广义估计方程(generalized estimating equations,GEE)分析相关因素，并以科室作为聚类变量。综合考量贝叶斯信息准则(Bayesian information criterion,BIC)、样本量校正的BIC(sample-size adjusted BIC,aBIC)、赤池信息准则(Akaike information criterion,AIC)及临床解释意义的分析结果，以识别潜在剖面。结果：最终共纳入1 492名研究参与者，MHLS-C总分为92.38±10.04。结合LPA结果，综合考量BIC、aBIC、AIC及临床解释意义的分析结果，最终识别出3个具有显著异质性的潜在类别：剖面1在“知识”“能力”“识别”维度的得分极低，但“接纳”的维度中等，命名为“低认知-中接纳型”(n=90,6.03%)；剖面2的各项认知指标得分中等，但“接纳”维度的得分为全样本最低，命名为“中认知-低接纳型”(n=886,59.38%)；剖面3的认知水平较高，但“接纳”维度的得分中等，命名为“高认知-中接纳型”(n=516,34.59%)。以“高认知-中接纳型”为参照组的多变量GEE分析结果显示，曾接受专业心理干预(OR=3.742,95%CI 2.124～6.592)、未接受心理知识相关培训(OR=2.136,95%CI 1.705～2.677)及年收入≤10万元(OR=2.682,95%CI 1.284～5.605)的护士更倾向于归入“低认知-中接纳型”(均P<0.05)；而男性(OR=2.104,95%CI 1.309～3.382)、年龄≥30岁(OR=1.476,95%CI 1.013～2.150)、已婚(OR=1.358, 95%CI 1.069～1.725)、在急危重症医学科(OR=2.286,95%CI 1.543～3.387)或外科(OR=1.499,95%CI 1.141～1.969)工作、未接受心理知识培训(OR=1.573,95%CI 1.326～1.866)、在区县医院工作(OR=1.353,95%CI 1.043～1.754)及年收入≤10万元(OR=1.558,95%CI 1.206～2.014)的护士，归入“中认知-低接纳型”的可能性显著更高(均P<0.05)。结论：重庆市三级综合医院护士的心理健康素养存在显著的非线性群体异质性。个人心理求助经历、专业培训缺失及高压执业环境是决定剖面归属的关键因素。未来建议摒弃“一刀切”模式，实施分层精准干预：针对“低认知-中接纳型”护士，重在知识补盲；针对“中认知-低接纳型”护士，需推行抗污名化与认知减负策略；针对“高认知-中接纳型”护士，则需强化人文反思与心理赋能，最终构建全人护理生态。更多还原

【Abstract】 Objective: Nurses in general hospitals play a critical role in identifying and managing patients with comorbid physical and mental health conditions. However, under current educational and training systems, their professional competence in mental health often faces limitations. This study aims to identify latent profiles of mental health literacy among nurses in tertiary general hospitals in Chongqing, China, using latent profile analysis(LPA), and to explore associated sociodemographic and occupational factors, thereby providing a scientific basis for targeted and stratified intervention strategies.Methods: A cross-sectional study was conducted using a multistage stratified cluster sampling method from April to May 2023. Based on the gross domestic product levels of 38 districts and counties in Chongqing, regions were stratified into high-, medium-, and low-economic levels. A total of nine tertiary general hospitals were randomly selected from each stratum. Nurses meeting the inclusion criteria were recruited using department-based cluster sampling. Data were collected using a self-designed demographic questionnaire and the Chinese version of the Mental Health Literacy Scale(MHLS-C). The MHLS-C comprised four dimensions: knowledge of mental disorders(“knowledge”), ability to seek information and help(“ability”), recognition of mental disorders(“recognition”), and acceptance of individuals with mental illness(“acceptance”). LPA was performed to identify latent profiles, with model fit evaluated using entropy and multiple information criteria. Considering the nested structure of participants within hospitals, multivariable generalized estimating equations(GEE) were used to analyze associated factors, with departments treated as clustering units. The latent profiles are identified based on LPA, and considering the Bayesian information criterion(BIC), sample-size adjusted BIC(aBIC), Akaike information criterion(AIC), and clinical interpretability. Results: A total of 1 492 valid participants were included, with a mean MHLS-C score of 92.38±10.04. Based on LPA, and considering the BIC, aBIC, AIC, and clinical interpretability, 3 distinct latent profiles were identified: Profile 1(“low cognitionmoderate acceptance”, n=90, 6.03%): very low scores in knowledge, ability, and recognition, but moderate acceptance; Profile 2(“moderate cognition-low acceptance”, n= 886, 59.38%): moderate cognitive scores but the lowest acceptance; Profile 3(“high cognitive-moderate acceptance”, n=516, 34.59%): high cognition levels with moderate acceptance. Using Profile 3 as the reference, multivariable GEE analysis showed that nurses who had received professional psychological interventions(OR=3.742, 95% CI 2.124 to 6.592), had not received mental health training(OR=2.136, 95% CI 1.705 to 2.677), or had an annual income ≤100 000 CNY(OR=2.682, 95% CI 1.284 to 5.605) were more likely to be classified into Profile 1(all P<0.05). Male sex(OR=2.104, 95% CI 1.309 to 3.382), aged≥30 years(OR=1.476, 95% CI 1.013 to 2.150), being married(OR=1.358, 95% CI 1.069 to 1.725), working in intensive care units(OR=2.286, 95% CI 1.543 to 3.387) or surgical departments(OR=1.499, 95% CI 1.141 to 1.969), lack of mental health training(OR=1.573, 95% CI 1.326 to 1.866), employment in county-level hospitals(OR=1.353, 95% CI 1.043 to 1.754), and annual income ≤100 000 CNY(OR=1.558, 95% CI 1.206 to 2.014) were significantly associated with Profile 2(all P<0.05).Conclusion: Significant heterogeneity exists in mental health literacy among nurses in tertiary general hospitals in Chongqing. Personal help-seeking experience, lack of professional training, and high-pressure work environments are key determinants of profile membership. Future interventions should move beyond a “one-size-fits-all” approach and adopt stratified strategies: for the “low cognition-moderate acceptance” group, emphasis should be placed on improving knowledge; for the “moderate cognition-low acceptance” group, anti-stigma interventions and cognitive burden reduction are needed; for the “high cognition-moderate acceptance” group, efforts should focus on enhancing humanistic reflection and psychological empowerment. Such tailored approaches may ultimately contribute to the development of a holistic nursing care system.更多还原