BACKGROUND: Few studies on care transitions following acute stroke have evaluated whether referral to community-based rehabilitation occurred as part of discharge planning. Our objectives were to describe the extent to which patients discharged home were referred to community-based rehabilitation and identify the patient, hospital, and community-level predictors of referral. METHODS: We examined data from 40 North Carolina hospitals that participated in the COMPASS (Comprehensive Post-Acute Stroke Services) cluster-randomized trial. Participants included adults discharged home following stroke or transient ischemic attack (N=10 702). In this observational analysis, COMPASS data were supplemented with hospital-level and county-level data from various sources. The primary outcome was referral to community-based rehabilitation (physical, occupational, or speech therapy) at discharge. Predictor variables included patient (demographic, stroke-related, medical history), hospital (structure, process), and community (therapist supply) measures. We used generalized linear mixed models with a hospital random effect and hierarchical backward model selection procedures to identify predictors of therapy referral. RESULTS: Approximately, one-third (36%) of stroke survivors (mean age, 66.8 [SD, 14.0] years; 49% female, 72% White race) were referred to community-based rehabilitation. Rates of referral to physical, occupational, and speech therapists were 31%, 18%, and 10%, respectively. Referral rates by hospital ranged from 3% to 78% with a median of 35%. Patient-level predictors included higher stroke severity, presence of medical comorbidities, and older age. Female sex (odds ratio, 1.24 [95% CI, 1.12–1.38]), non-White race (2.20 [2.01–2.44]), and having Medicare insurance (1.12 [1.02–1.23]) were also predictors of referral. Referral was higher for patients living in counties with greater physical therapist supply. Much of the variation in referral across hospitals remained unexplained. CONCLUSIONS: One-third of stroke survivors were referred to community-based rehabilitation. Patient-level factors predominated as predictors. Variation across hospitals was notable and presents an opportunity for further evaluation and possible targets for improved poststroke rehabilitative care. REGISTRATION: URL: https://www.clinicaltrials.gov ; Unique identifier: NCT02588664.

中文翻译：

---

急性中风后转介至社区康复：COMPASS 实用试验的结果

背景：很少有关于急性中风后护理过渡的研究评估转诊社区康复是否作为出院计划的一部分。我们的目标是描述出院回家的患者被转介到社区康复的程度，并确定患者、医院和社区层面的转诊预测因素。 方法：我们检查了参与 COMPASS（急性中风后综合服务）整群随机试验的 40 家北卡罗来纳州医院的数据。参与者包括中风或短暂性脑缺血发作后出院回家的成年人 (N=10 702)。在这次观察分析中，COMPASS 数据补充了来自不同来源的医院级和县级数据。主要结局是出院时转介至社区康复（身体、职业或言语治疗）。预测变量包括患者（人口统计、中风相关、病史）、医院（结构、流程）和社区（治疗师供应）措施。我们使用具有医院随机效应的广义线性混合模型和分层后向模型选择程序来确定治疗转诊的预测因素。 结果：大约三分之一 (36%) 的中风幸存者（平均年龄，66.8 [SD, 14.0] 岁；49% 女性，72% 白人）被转介至社区康复中心。转诊至物理治疗师、职业治疗师和言语治疗师的比例分别为 31%、18% 和 10%。医院转诊率从 3% 到 78% 不等，中位数为 35%。患者层面的预测因素包括中风严重程度较高、存在合并症和年龄较大。女性（比值比，1.24 [95% CI，1.12–1.38]）、非白人种族（2.20 [2.01–2.44]）和拥有医疗保险（1.12 [1.02–1.23]）也是转诊的预测因素。生活在物理治疗师供应较多的县的患者转诊率较高。各医院转诊的大部分差异仍无法解释。 结论：三分之一的中风幸存者被转介至社区康复中心。患者层面的因素作为预测因素占主导地位。各医院之间的差异显着，为进一步评估和改善中风后康复护理的可能目标提供了机会。 登记：网址：https://www.clinicaltrials.gov; 唯一标识符：NCT02588664。