Cecil G. Sheps Center for Health Services Research, University of North Carolina, 725 Martin Luther King, Jr. Blvd., Chapel Hill, NC 27599-7590, USA.
Psychiatr Serv. 2013 Apr 1;64(4):312-7. doi: 10.1176/appi.ps.201200096.
A previous study of recovery-oriented assertive community treatment (PACT) found large differences over three years in use of state psychiatric hospitals between PACT participants and consumers in a matched control group, especially for PACT participants with significant previous psychiatric hospitalization. This study extended these findings by examining the timing of PACT effects.
Generalized estimating equation models of monthly cost data for state, local, and crisis hospital use estimated the time-varying effects of participation in one of ten PACT teams in Washington State. Data from PACT participants (N=450) and propensity score-matched consumers (N=450) were included. Additional analyses determined whether effects differed by prior state hospital use.
Differences in costs between PACT and control participants were largest immediately after PACT enrollment and tapered off. During the first quarter after enrollment, monthly per-person costs for state hospital use were $3,458 lower for PACT enrollees than for control participants. A composite measure of psychiatric hospital costs (state and local hospitals and local crisis stabilization units) declined by $3,539 monthly during the first quarter after PACT enrollment (p<.01). Differences were noted up to 27 months after enrollment, when the difference in the composite costs measure became insignificant compared with the prior quarter (months 25-27) (p>.05). Differences were larger for PACT enrollees with greater baseline state hospital use.
The time-varying estimates may have implications for the length and intensity of ACT enrollment. However, the optimum time for receipt of ACT services needs to be considered in the context of outcomes other than hospitalization alone.
先前一项针对以康复为导向的坚定社区治疗（PACT）的研究发现，在三年内，PACT 参与者与匹配对照组中的消费者在使用州立精神病院方面存在很大差异，尤其是对于有过大量先前精神病住院经历的 PACT 参与者。本研究通过检查 PACT 效果的时间来扩展这些发现。
使用广义估计方程模型对州、地方和危机医院使用的月度成本数据进行建模，以估计参与华盛顿州十个 PACT 小组之一的参与者的时间变化效应。研究数据来自 PACT 参与者（N=450）和倾向得分匹配的消费者（N=450）。此外的分析确定了效果是否因先前的州立医院使用而异。
PACT 和对照组参与者之间的成本差异在 PACT 入组后立即最大，并逐渐减少。在入组后的第一个季度，PACT 参与者的州立医院每月个人费用比对照组参与者低 3458 美元。在 PACT 入组后的第一个季度，精神科医院费用的综合衡量指标（州和地方医院以及地方危机稳定单位）每月下降 3539 美元（p<.01）。在入组后 27 个月内仍能观察到差异，此时复合成本指标的差异与前一季度（第 25-27 个月）相比不再显著（p>.05）。对于基线州立医院使用量较大的 PACT 参与者，差异更大。
时变估计可能对 ACT 参与的时长和强度有影响。然而，需要在考虑住院以外的其他结果的情况下，确定接收 ACT 服务的最佳时间。
