The effect of patient and site characteristics on predictors of care experiences among PCQ-H sample population

1. After consideration of persons with missing variables, the final model consists of 355 informative study participants. Overall R-squared for the model shown was 0.36, F = 8.95, p ≤0.0001. For variables composed of multiple categories, P values reflect type 3 tests of fixed effects.2. General Health is construed as a 5-step variable (Poor, Fair, Good, Very Good, and Excellent).3. Depression/Anxiety and Severe Psychiatric Symptoms scores are computed from relevant subscales on the Colorado Symptom Index (see Methods).4. Drug Severity and Alcohol Severity derived from World Health Organization Alcohol, Smoking and Substance Involvement Screening Test Global Continuum Illicit Drug Use Score and Specific Current Alcohol Involvement Score, respectively.5. Perceived Extent of Choice derived from 4-point Likert response to the item “I can switch primary care providers if I ask”.6. Social Support derived from 3 items of the “Strong Ties” scale related to companionship and friendship (see Methods).1.

1. 在剔除变量存在缺失的受试者后，最终模型纳入355名信息完整的研究参与者。所呈现模型的总体决定系数（R-squared）为0.36，F值为8.95，p≤0.0001。对于包含多个类别的变量，其P值对应固定效应的第三类检验（type 3 tests of fixed effects）。
2. 一般健康状况被构建为5级变量，分级依次为差、一般、良好、优秀及极佳。
3. 抑郁/焦虑与严重精神症状得分通过科罗拉多症状指数（Colorado Symptom Index）的相关子量表计算得出（详见方法部分）。
4. 药物严重程度与酒精严重程度分别源自世界卫生组织酒精、吸烟与物质使用筛查测试的全球非法药物使用连续体得分与当前特定酒精涉入得分。
5. 感知选择程度基于对"若提出申请，可更换全科医疗服务提供者"这一问题的4级李克特（Likert）量表应答结果计算得到。
6. 社会支持源自"强纽带（Strong Ties）"量表中3项与陪伴及友谊相关的条目（详见方法部分）。