Methodology for data set development

The use of computers for multidisciplinary patient documentation has increased rapidly over the last 15 years. Each computerized patient care documentation system includes a predetermined set of data that is collected and stored based on the individual patient's clinical course and needs. To make the best use of the data collected, it is desirable for computerized documentation systems to use free text as minimally as possible. Coded data are easier to search and use for future analyses. The purpose of this study was to develop and formatively evaluate a methodological approach for use when developing data sets for computerized patient care documentation systems. The study identified knowledge engineering techniques that could be useful during data set development. Four knowledge acquisition techniques were identified as particularly applicable to data set development. These included literature review, expert interview, data analysis and focus groups. These identified techniques were formalized into a methodology for data set development. The methodology was presented in document form including text, diagrams and examples. Evaluation of the development methodology document was done by expert review and solicitation of expert opinion of the methodology via a questionnaire. The research questions addressed the clarity and usefulness of the newly proposed methodology. Eleven experts in data set development and/or data set use participated in the evaluation. The study sample represented a broad range of knowledge and experience in data set development and use. The results of the questionnaire and the comments received from the reviewers indicated the Data Set Development Methodology (DSDM) was fairly well defined and is useful for data set development. The mean scores for all questions asking about the clarity and the usefulness of the document were above 5 (range = 1-7, with 1 = strongly disagree, 7 = strongly agree). These reviewer scores support the formalization of this DSDM for data set development. The DSDM provides clinical data set developers of various skill levels and resources, a documented, consistent, structured and economical tool for the creation of their needed and /or desired data set.