A comparison of nursing and medical diagnoses in predicting hospital outcomes.

UNLABELLED: The main premise of the Nursing Minimum Data Set (NMDS) is that nursing data should be included in the hospital discharge abstract. Yet to date, little empirical evidence has been published to measure the efficacy or usefulness of these nursing data elements. We report the results of a comparison between a daily collection of nursing assessments using nursing diagnoses (NDX) to the Diagnostic Related Group (DRG) and the All Payer Refined DRG (APR-DRG) in their ability to predict three common outcome variables: hospital days, ICU day, and total charges. A secondary data analysis was performed from a large existing data set of four years patient data from a Midwest University hospital. FINDINGS: NDX is significantly associated with hospital length of stay, ICU length of stay, and total charges. NDX also improves explanatory power when added to models with DRG or APR-DRG. This suggests that nursing data compliments existing data and is not redundant with the DRG or APR-DRG. The findings also suggest that NDX explains a different portion of the variance of the three outcome variables in this series. The results of this study support the argument that nursing data should be included in the hospital discharge abstract.

Development of a computerized database for a nursing quality management program.

There has been little study on how nurse managers collect, store, process and retrieve quality management data, yet nearly every nursing department has an existing program to provide these functions. We investigated our current paper-based system and found several deficiencies. Guided by structured interviews, task analysis, and focus groups we developed a networked computerized quality management database to provide more timely data reporting and consistency of analysis. This paper describes the development and implementation of the system with discussion on how it will be evaluated in the future.

Patient outcomes and nurses' classification data.

The patient classifications done by nurses for all adult patients (n = 15,500) discharged from an urban teaching hospital in one year were retrieved and analyzed by discharge status. Classification results were summarized by physical-functional, psychological-social, and dependence categories and were associated with discharge disposition; patients discharged home were less dependent than others discharged to nursing homes or those who died in the hospital. Diagnosis related group (DRG) payment weights were somewhat independent of the patient classification scores and were not associated with adverse outcomes.

Automating and improving the data quality of a nursing department quality management program at a university hospital.

BACKGROUND: The development and implementation of a relational database program for nursing quality management at a university hospital was stimulated by a lack of consistent data management and analysis tools in the existing noncomputerized program. PROGRAM DEVELOPMENT AND IMPLEMENTATION: An initial software prototype implemented in the critical care service included data collection instruments for five areas: medication errors, patient falls, returns to an intensive care unit within 48 hours, hospital-acquired skin breakdown, and unplanned extubations. Access to the database was limited and paper reports only were disseminated on a scheduled basis. In a second phase, the database is being deployed throughout the nursing department using a local area network. Nurse managers will enter and interact with the quality database online and have access to graphics, reports, and action plan development. POSSIBLE ERRORS: A wide range of potential errors influences decisions on how to collect, store, retrieve, and process quality management data. Each type of error affects the nurse manager's ability to identify significant patterns or trends that are amenable to intervention. There is no right way of constructing and implementing a quality improvement database; only an optimum balance between cost, complexity, and efficacy. SUMMARY AND CONCLUSIONS: Initial feedback from end uses has been positive. A three-year experience with a personal computer database suggests that the personal computer-based information technology is appropriate for small to medium applications and can support departmentwide CQI efforts. A case scenario using simulated data is included to illustrate the use of computerized reports in assessing and taking action on an increase in falls.