Content validation for level of use of feature rich systems: a Delphi study of electronic medical records systems

Shaw, N. and Manwani, S. (2013) Content validation for level of use of feature rich systems: a Delphi study of electronic medical records systems. Information Research, 18 (1). 558. ISSN 1368-1613 (Shaw, N. & Manwani, S. (2013). Content validation for level of use of feature rich systems: a Delphi study of electronic medical records systems. Information Research, 18(1) paper 558. [Available at http://InformationR.net/ir/18-1/paper558.html])

Abstract/Summary

Introduction. Feature usage is a pre-requisite to realising the benefits of investments in feature rich systems. We propose that conceptualising the dependent variable 'system use' as 'level of use' and specifying it as a formative construct has greater value for measuring the post-adoption use of feature rich systems. We then validate the content of the construct as a first step in developing a research instrument to measure it. The context of our study is the post-adoption use of electronic medical records (EMR) by primary care physicians. Method. Initially, a literature review of the empirical context defines the scope based on prior studies. Having identified core features from the literature, they are further refined with the help of experts in a consensus seeking process that follows the Delphi technique. Results.The methodology was successfully applied to EMRs, which were selected as an example of feature rich systems. A review of EMR usage and regulatory standards provided the feature input for the first round of the Delphi process. A panel of experts then reached consensus after four rounds, identifying ten task-based features that would be indicators of level of use. Conclusions. To study why some users deploy more advanced features than others, theories of post-adoption require a rich formative dependent variable that measures level of use. We have demonstrated that a context sensitive literature review followed by refinement through a consensus seeking process is a suitable methodology to validate the content of this dependent variable. This is the first step of instrument development prior to statistical confirmation with a larger sample.