Preparing a large data set for analysis: using the Minimum Data Set to study perineal dermatitis
Article first published online: 24 OCT 2005
Journal of Advanced Nursing
Volume 52, Issue 4, pages 399–409, November 2005
How to Cite
Savik, K., Fan, Q., Bliss, D. and Harms, S. (2005), Preparing a large data set for analysis: using the Minimum Data Set to study perineal dermatitis. Journal of Advanced Nursing, 52: 399–409. doi: 10.1111/j.1365-2648.2005.03604.x
- Issue published online: 24 OCT 2005
- Article first published online: 24 OCT 2005
- Accepted for publication 30 June 2005
- Brown's model;
- large data sets;
- nursing homes;
- perineal dermatitis;
Aim. The aim of this paper is to present a practical example of preparing a large set of Minimum Data Set records for analysis, operationalizing Minimum Data Set items that defined risk factors for perineal dermatitis, our outcome variable.
Background. Research with nursing home elders remains a vital need as ‘baby boomers’ age. Conducting research in nursing homes is a daunting task. The Minimum Data Set is a standardized instrument used to assess many aspects of a nursing home resident's functional capability. United States Federal Regulations require a Minimum Data Set assessment of all nursing home residents. These large data would be a useful resource for research studies, but need to be extensively refined for use in most statistical analyses. Although fairly comprehensive, the Minimum Data Set does not provide direct measures of all clinical outcomes and variables of interest.
Method. Perineal dermatitis is not directly measured in the Minimum Data Set. Additional information from prescribers’ (physician and nurse) orders was used to identify cases of perineal dermatitis. The following steps were followed to produce Minimum Data Set records appropriate for analysis: (1) identification of a subset of Minimum Data Set records specific to the research, (2) identification of perineal dermatitis cases from the prescribers’ orders, (3) merging of the perineal dermatitis cases with the Minimum Data Set data set, (4) identification of Minimum Data Set items used to operationalize the variables in our model of perineal dermatitis, (5) determination of the appropriate way to aggregate individual Minimum Data Set items into composite measures of the variables, (6) refinement of these composites using item analysis and (7) assessment of the distribution of the composite variables and need for transformations to use in statistical analysis.
Results. Cases of perineal dermatitis were successfully identified and composites were created that operationalized a model of perineal dermatitis.
Conclusion. Following these steps resulted in a data set where data analysis could be pursued with confidence. Incorporating other sources of data, such as prescribers’ orders, extends the usefulness of the Minimum Data Set for research use.