An hfm Web Extra
Realizing the full benefits of business intelligence (BI) systems requires ensuring clean, reliable data from primary sources, creating dashboards that save time by focusing the decision making efforts of managers, and understanding that BI provides multidimensional analysis traditionally not available from use with reports.
BI users can correlate previously separate data elements to establish cause and effect of business issues in conjunction with trending to determine whether variances are isolated incidences or significant indicators that operations are improving or faltering. The BI tool allows the user to drill down to establish actual issues as opposed to addressing symptoms.
Take for example an OR manager who wishes to undertake an analysis to verify that the hospital is meeting a key quality standard by adhering to the practice of administering antibiotics (ABs) one hour before cut time and to ascertain whether case late starts are in any way associated with maintaining appropriate AB administration. The OR manager uses a dashboard with monthly trends of time between AB administration to cut time and case volumes that started more than five minutes after the scheduled start time.
While reviewing information, the manager discovers speculative high variance times indicating poor data entry, which must be corrected to ensure reliable data trending (see the exhibit below).
Having ensured the accuracy of the data, the manager is able to determine that antibiotic administration to cut time does not correlate to cases starting late. However, the manager also is able to determine that there was an isolated high level of inefficiency in December (see the exhibit below) .
The manager then focuses the analysis on the reason why inefficiency spiked in December. A drill-down analysis discloses a contributing factor was short staffing from increased paid time off (PTO) was a contributing factor. Based on this finding, administration is prompted to reevaluate the staff PTO policy, which caused the high level of staff absences in December.
Administration also wishes to optimize patient admissions and readmissions. So BI is used to trend patient visits with a length of stay (LOS) of only 1.0 day to isolate issues regarding lack of medical necessity (patient care that could be performed as outpatient), or discharging too early requiring readmission of patients (see the exhibit below).
The analysis discloses that one physician had five of the six patients with the highest number of readmissions (see the exhibit below).
With tangible information in hand, administrators are able to work with case management and physicians through July, August, and September to establishing appropriate medical necessity and standardize care to avoid readmissions. Continued trending following these efforts shows a reduction in 1.0 day LOS, from more than 200 a month to less than 100 per month, with an average decrease in average cost per visit and number of readmits. (see the exhibit below).
For more information,see Rose Rohloff's "Healthcare BI: A Tool for Meaningful Analysis," hfm, May 2011
Publication Date: Monday, May 02, 2011