Developing Item Banks for Patient-Reported Health Outcomes
Over the past 30 years, health outcome assessments have been increasingly used in clinical trials and clinical practice settings. Health outcomes assessment is now commonly incorporated into clinical practice and patient health care. In clinical trials and comparative effectiveness research studies, health outcomes assessments are used to demonstrate treatment effectiveness (Ahmed et al., 2012). Health outcomes are also playing a larger role in health system performance assessments by providing critical information on health care quality from the consumer perspective (National Quality Forum, 2013). However, for years, the development of health outcomes assessments was not systematic or consistent, and multiple assessments for different purposes and of varying quality have been developed. Many instruments were developed to meet the immediate needs of a specific clinical study or clinical development program, and often multiple instruments covering the same health domains were simultaneously developed by competing health outcomes researchers. The development of item banks using item response theory (IRT) methods for different health domains provides a way to standardize the health outcomes assessment by drawing from a pool of items that are systematically and consistently developed and that have scores on a standardized metric (Cella et al., 2007; Cella et al., 2010). In addition, a tailored health outcomes assessment can be developed using a subset of items from an item bank so that the assessment is more efficient and less burdensome to patients yet retains adequate precision.