Abstracting, Normalizing, and Reconciling Data
QUESTION
How do you evaluate outliers? How can outliers be used to determine root cause of data integrity issues? Research and briefly explain how CMS uses Winsorization when calculating cost performance.
Abstracting, Normalizing, and Reconciling Data
ANSWER
Abstracting, Normalizing, and Reconciling Data
Student name
Professor’s Name
Institutional Affiliation
Course name
Institution affiliation
Date
Abstracting, Normalizing, and Reconciling Data
Outliers are data points that appear not to belong to a provided set of data owing to measurement faults, unknown data structure, or faulty distribution assumption. Outliers are thus evaluated by considering these terminologies. To identify the fences for a certain data set, one may look for the interquartile range product by 1.5, add the result to q3, and subtract it from q1. The point that falls outside the outer fences is categorized as a major, while that that falls outside the data sets inner fences is called a minor (Tarazi, 2020).
Data integrity is one of the leading concerns in handling PHI information by nurses and healthcare facilities. Outliers may be employed to determine the root cause of integrity concerns by identifying observation through visual or graphical inspection. When cause variation is identified when the process goes out of control, creating outlier data points. The data points are then removed to give a realistic depiction of the process and give useful control limits. Besides, it ensures that actions are taken only when necessary and mitigates the added statistical control limits (Ventur, 2020).
Winsorization on the other hand is transformation of statistical data through the limitation of the extreme values in the data to minimize the effects of bogus outliers. The Center for Medicare and Medicaid Services employs winsorization in determining cost performance to identify equitable measures of performance on all primary care providers in the following way.
References
Tarazi, W. W. (2020). Associations between Medicaid expansion and nurse staffing ratios and hospital readmissions. Health Services Research, 55(3), 375-382.
Ventur, F. (2020). Deficient testing databases: a reliability-driven evaluation of privacy models providing a trade-off between data integrity and re-identification risk (Bachelor’s thesis, University of Twente).