Learning Adaptive Treatment Strategies in Mental Health

描述（申请人提供）：严重抑郁症和精神分裂症等慢性精神疾病给个人和社会带来沉重负担。高质量的临床护理可以减轻负担，但这样的护理需要连续的临床决策有关的治疗。该项目的长期目标是改善连续的临床决策。具体目标涉及帮助临床医生解决以下问题的方法：应该等待多长时间才能确定患者没有从治疗中获得足够的益处？如果确定患者没有从治疗中充分获益，那么接下来应该提供哪种治疗？治疗顺序是否应根据患者特征和结局（如疾病特征、反应、副作用负担和依从性）而有所不同？在这个项目中，从计算机科学，工程学和统计学中提取的方法被推广并适用于临床试验数据，以解决这些问题。这些方法将使用NIMH资助的两项大型试验的数据进行开发和验证。该项目将由一个合作小组进行，其中包括一名计算机科学家，两名精神病学家和一名统计学家。

DESCRIPTION (provided by applicant): Chronic mental illnesses such as major depression and schizophrenia impose a heavy burden both on the individual and society. High quality clinical care can reduce the burden but such care demands sequential clinical decision making concerning treatment. The long term goal of this project is to improve sequential clinical decision making. The specific aims concern methodological approaches for helping clinicians address questions such as: How long should one wait to decide that a patient is not deriving sufficient benefit from treatment? If it is established that a patient is not sufficiently benefiting from treatment then which treatment should be provided next? Should the sequence of treatments vary according to patient characteristics and outcomes such as disease features, response, side effect burden, and adherence? In this project methods taken from computer science, engineering and statistics are generalized and adapted for use with clinical trial data so as to address these kinds of questions. The methods will be developed and validated using data from two large NIMH funded trials. This project will be conducted by a collaborative team involving a computer scientist, two psychiatrists and a statistician.