204480 Natural language processing to evaluate text fields in Omaha System data

Tuesday, November 10, 2009

Colleen M. Hart, MS, RN , School of Nursing, University of Minnesota, Minneapolis, MN
Madeleine J. Kerr, PhD, RN , School of Nursing, University of Minnesota, Minneapolis, MN
Genevieve B. Melton, MD, MA , School of Medicine & Institute of Health Informatics, University of Minnesota, Minneapolis, MN
Karen A. Monsen, PhD RN , School of Nursing, University of Minnesota, Minneapolis, MN
Bonnie Westra, PhD, RN , School of Nursing, University of Minnesota, Minneapolis, MN
Debra A. Solomon, MSN, RN, FNP , Fairview Lakes HomeCaring & Hospice, Chisago City, MN
Jill E. Timm, JD, RN , Department of Public Health and Environment, Washington County, Stillwater, MN
Standardized data sets from computerized clinical information systems offer rich content regarding client characteristics, interventions and outcomes. One standardized classification, the Omaha System, is commonly used for computerized documentation of home visiting practice. The Omaha System consists of three components: Problem Classification Scheme, Intervention Scheme, and Problem Rating Scale for Outcomes. Practitioner documentation consists of structured data from the three components and associated descriptive free text. To date, only structured data have been available for research and program evaluation. However, techniques such as Natural Language Processing (NLP) have the capacity to analyze free text documentation. Medical NLP systems allow computerized processing of clinical text into a structured format taking into account the complexities of language within the health domain. The purpose of this study is to identify system and user issues by using NLP to analyze free text fields utilized in documentation with the Omaha System in the context of associated structured terms and definitions.

The investigators will analyze de-identified data from a public health family home visiting program (6,680 visits, 1,079 clients) and a home care, palliative care, and hospice program (55,021 visits, 2,309 clients) that implement the Omaha system. An established medical natural language processor will be used in the analysis of text fields. Text will be converted to standard concepts, taking into account negation, timing, and modifiers to the text. These concepts will be organized and compiled into categories, generating standard summary statistics of this data. By comparing the categories of concepts contained within the Omaha System and free text data, the investigators will identify potential areas of improvement and modification for the Omaha system and prospective user issues with computerized data entry in this setting.

Learning Objectives:
1. Describe an application of natural language processing (NLP) with home visiting practice data. 2. Identify potential system issues uncovered by NLP. 3. Identify potential user issues discovered by NLP.

Keywords: Evaluation, Human Populations

Presenting author's disclosure statement:

Qualified on the content I am responsible for because: I am a co-investigator on the research project.
Any relevant financial relationships? No

I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines, and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed in my presentation.