Methods for quality-of-life studies.
Methods for quality-of-life studies.
复制标题
生活质量研究方法。
DOI：
10.1146/annurev.pu.15.050194.002535
复制
发表时间：
1994
影响因子：
20.8
通讯作者：
Nackley,JF
中科院分区：
文献类型：
--
作者：
Testa,MA;Nackley,JF
在健康的观察性和干预性研究中使用患者的生活质量结果的方法来自于一个大的和多样化的研究方法领域。生活质量被概念化的多维方式将影响衡量它的方式和衡量的复杂性。在研究的最早阶段，人们必须依靠测试和测量、调查研究、心理计量学和社会计量学等领域常见的方法来测量无法直接观察到的结构。衡量绩效的指标既可以集中在量表在非干预性横断面研究中的执行能力，也可以集中在干预性纵向研究中。稳定性指数、内部一致性指数、对生活质量真实变化的反应指数和对治疗效果的敏感性指数可被用来评估量表的充分性，作为感兴趣的因变量。受访者的差异可能是由于不同的记者(患者、配偶、医生)、管理的方式和形式(长形式与短形式；自我管理与采访)以及评估环境(诊所、家庭)等因素造成的。最后，由于生活质量研究通常涉及推论统计和假设检验，因此应遵循好的研究设计的统计学和流行病学原则。此外，在设计科学假设时，应该考虑到量表的可靠性、响应性和敏感性，并应该通过基于干预的验证来具体说明生活质量效应大小的含义。设计考虑必须解决以下统计问题：权力、通过外部标准验证来确定效果大小、纵向数据、退出和提前终止的影响、上限和下限效应、以及个体反应和敏感性的异质性。在这个医疗改革和财政紧缩的时代，估计用于药物经济学模型的生活质量总结参数的问题正受到越来越多的关注。虽然医疗决策理论自20世纪70年代初以来一直使用成本-效果模型和质量调整的寿命年，但对人口参数的估计以区分不同的医疗干预措施相对较新。从风险、收益和成本方面评估与医疗干预相关的患者结果显然将是卫生保健改革的主要重点。生活质量研究的新方法的发展应该建立在临床研究、流行病学、生物统计学、经济学和行为科学领域已经建立的坚实基础之上。
Methodologies involving the use of quality-of-life patient outcomes in observational and interventional studies of health are drawn from a large and diverse field of research methods. The multidimensional way in which quality of life is conceptualized will affect the way it is measured and the complexity of the measurement. At the earliest stages of research, one must rely on methods common to the fields of tests and measurement, survey research, psychometrics and sociometrics to measure constructs that are not directly observable. Indices measuring performance can either focus on the scale's ability to perform in noninterventional, cross-sectional studies or interventional, longitudinal studies. Indices of stability, internal consistency, responsiveness with respect to true changes in quality of life, and sensitivity to treatment effects can be used to assess the scale's adequacy as a dependent variable of interest. Respondent variability can occur due to factors such as different reporters (patient, spouse, physician), the manner and form of administration (long form vs short form; self-administration vs interview) and the assessment environment (clinic, home). Finally, since quality-of-life research often involves inferential statistics and hypothesis testing, the statistical and epidemiologic principles of good study design should be followed. In addition, one should account for the reliability, responsiveness, and the sensitivity of the scale when designing the scientific hypotheses, and should specifically address the meaning of quality-of-life effect sizes by interventional-based validation. Design considerations must address the statistical issues of power, the determination of effect sizes through validation by external criteria, longitudinal data, effects of withdrawal and early termination, ceiling and floor effects, and heterogeneity of responsiveness and sensitivity among individuals. The problem of estimating quality-of-life summary parameters for use in pharmacoeconomic models is receiving increasing attention in this era of health-care reform and fiscal restraint. While medical decision theory has used cost-effectiveness models and quality-adjusted life years since the early 1970s, estimation of population parameters to differentiate among different medical interventions is relatively new. The assessment of the patient outcomes associated with medical interventions in terms of the risks, benefits and costs will clearly be a major focus of health-care reform. Development of new methodologies in quality-of-life research should build upon the strong foundation already established in the areas of clinical research, epidemiology, biostatistics, economics and behavioral science.(ABSTRACT TRUNCATED AT 400 WORDS)