NURS 6231 Week 7 Variation Analysis Example

Same Bundle, Different Floors: A Variation Analysis of Injurious Falls on Two Medical Units

Student Name

College of Nursing, Walden University

NURS 6231: Healthcare Systems and Quality Outcomes

Instructor Name

Month Day, Year

Same Bundle, Different Floors: A Variation Analysis of Injurious Falls on Two Medical Units

A composite 300-bed hospital runs one fall prevention bundle on all adult medical floors. The bundle calls for hourly rounding, a bed alarm for patients at high fall risk and scheduled toileting for patients who need help to walk. Two of its medical floors, 4 West and 5 East, each have 30 beds and about 900 patient days a month. Over the last two years, 4 West recorded 26 falls with injury and 5 East recorded 9. This analysis first tests whether that gap reflects a real and stable difference or ordinary month-to-month noise, and then compares how the bundle is practiced on each floor, whom each floor serves and how each floor is built. The analysis finds two stable processes producing different results, and it attributes most of the gap to scheduled toileting, which happens on 5 East and not on 4 West, and the remainder to the layout of 4 West.

Is the Difference Real?

Every process produces some variation. The distinction Shewhart drew, and Deming later made central to management, separates common-cause variation, which is built into a stable process and produces a predictable range of results, from special-cause variation, which arises from something outside the usual process (Deming, 1986; Shewhart, 1931). A manager who reacts to common-cause variation as if each bad month had a special cause will chase noise. A difference between two units could also be noise, so it has to be tested before it is explained.

Table 1 shows the monthly counts of injurious falls on each unit. Because patient days were close to 900 each month, counts can be read directly, and rates are given for each year.

Table 1

Injurious Falls by Month, 4 West and 5 East, Two Years

Each unit's data was plotted on its own u-chart, a control chart for counts of events per unit of exposure, with limits calculated from that unit's two-year average (Benneyan et al., 2003). Table 2 summarizes the two charts.

Table 2

Control Chart Parameters, Injurious Falls per 1,000 Patient Days

Both charts show stable processes. No month falls beyond the upper limit on either unit, and neither chart shows a run or trend that the standard rules would flag. The worst months on 4 West, including a June with three injurious falls, are within that unit's ordinary range and do not need a special explanation.

The difference between the units is also real. If 4 West ran the same process as 5 East, a month with at least one injurious fall would be expected about 31% of the time. 4 West had at least one in 20 of its 24 months. Taken together, the two-year totals of 26 and 9 on equal exposure would occur by chance less than 1% of the time if the units had the same underlying rate. The gap is not a bad stretch on 4 West; it is the difference between two stable processes. The next question is what those processes are.

Rare events limit what u-charts can detect. With fewer than two events a month, a modest shift within a unit might take many months to show, and Benneyan et al. (2003) noted alternatives, such as charts of the time between events, for low rates. For the purpose here, comparing two units over two years, the u-charts and totals are sufficient.

The Bundle as Practiced

Both units document the bundle nearly completely. Hourly rounding appears in 96% of required hours on 4 West and 95% on 5 East. A policy audit would rate them equal. Direct observation tells a different story. The analysis used 20 hours of observation on each unit, spread across day, evening and night shifts, with an observer recording whether a staff member actually entered each room during each hour and what was offered.

Three differences emerged. First, observed rounding reached patients in 81% of hours on 5 East and 38% on 4 West, where rounding was often documented from the doorway or after the fact. Second, 5 East applies bed alarms by protocol to every patient with a high fall risk score, while 4 West leaves the decision to the nurse. Third, 5 East offers each patient who needs help to walk a trip to the toilet on a two-hour schedule during the day and at 10 p.m., 2 a.m. and 5 a.m., as part of each round. 4 West has no toileting schedule; patients call when they need to go.

Whom Each Floor Serves

Before any of these differences can be credited, case mix has to be checked. Over the two years, the median age was 71 on 4 West and 70 on 5 East. Positive delirium screens on the Confusion Assessment Method occurred in 14% and 13% of admissions. Sedative-hypnotic orders were active for 22% and 24% of patients. Both units admit from the same emergency department under the same bed placement rules. The patients are similar enough that case mix does not explain a threefold difference in injurious falls.

How Each Floor Is Built

The buildings differ. 5 East is a racetrack with the nurses' station at the center and no room more than about 20 meters away. 4 West is an L-shaped wing with the station at the corner, and its 10 rooms at the end of the long arm sit more than 30 meters from the station, out of sight of it.

Dividing the Gap

The incident reports allow the gap to be divided. Of 4 West's 26 injurious falls, 15 occurred while the patient was getting up to use the bathroom or commode without help; on 5 East, 3 of 9 did. That difference of 12 falls is about 70% of the 17-fall gap. It matches what is known about hospital falls: half of inpatient falls in one prospective study were related to elimination, and those falls carried more than twice the odds of injury (Hitcho et al., 2004). A patient offered the toilet on a schedule is less likely to get up alone in a hurry. Because scheduled toileting on 5 East is delivered during rounds that actually happen, the rounding and toileting differences act as one mechanism, and the analysis credits them together.

Of the remaining 5 falls in the gap, 4 West's non-toileting falls clustered in the far rooms of the long arm, where 7 of its 11 such falls occurred, although those rooms hold a third of its beds. Distance from the station plausibly delays the response to a patient getting up. The analysis assigns the remainder to layout.

Bed alarms are not assigned a share. In a cluster randomized trial on 16 units, a program that greatly increased bed alarm use produced no reduction in falls (Shorr et al., 2012). The difference in alarm practice between the floors is real, but the evidence does not support treating it as a cause.

Conclusion

Two floors following one bundle on paper run two different processes, and each is stable. The charts establish that the gap is real, case mix does not explain it, and observation shows where it comes from: patients on 5 East are taken to the toilet on a schedule during rounds that happen, and patients on 4 West get up alone, some of them far from the station. The analysis attributes about 70% of the gap to scheduled toileting within real rounding and the rest to layout. Closing it will require changing the process on 4 West, not exhorting its staff, since a stable process will keep producing the same results until it is redesigned.

References

Benneyan, J. C., Lloyd, R. C., & Plsek, P. E. (2003). Statistical process control as a tool for research and healthcare improvement. Quality and Safety in Health Care, 12(6), 458-464. https://doi.org/10.1136/qhc.12.6.458

Deming, W. E. (1986). Out of the crisis. MIT Center for Advanced Engineering Study.

Hitcho, E. B., Krauss, M. J., Birge, S., Dunagan, W. C., Fischer, I., Johnson, S., Nast, P. A., Costantinou, E., & Fraser, V. J. (2004). Characteristics and circumstances of falls in a hospital setting: A prospective analysis. Journal of General Internal Medicine, 19(7), 732-739. https://doi.org/10.1111/j.1525-1497.2004.30387.x

Shewhart, W. A. (1931). Economic control of quality of manufactured product. Van Nostrand.

Shorr, R. I., Chandler, A. M., Mion, L. C., Waters, T. M., Liu, M., Daniels, M. J., Kessler, L. A., & Miller, S. T. (2012). Effects of an intervention to increase bed alarm use to prevent falls in hospitalized patients: A cluster randomized trial. Annals of Internal Medicine, 157(10), 692-699. https://doi.org/10.7326/0003-4819-157-10-201211200-00005