HLTH 4393 Module 3 Quality Data Analysis Example

Stable, Predictable and Too Slow: A Run Chart, p-Chart and Pareto Analysis of Twenty-Six Weeks of Emergency Troponin Turnaround

Student Name

American College of Education

HLTH4393: Quality Management for Healthcare Administrators

Module 3 Assignment

Instructor Name

October 19, 2026

The Data

Earlier modules settled what to measure for emergency troponin testing and walked the process from order to result. This paper turns to the numbers themselves: 26 weeks of results, April through September, drawn from the timestamps in the laboratory information system. The emergency department ordered 4,290 high-sensitivity troponin tests in that period, about 165 a week, and 2,488 of them, or 58%, met the 60-minute standard.

Three questions guide the analysis. Is the process stable, or is its performance changing? If it is stable, is it capable of meeting the target of 90%? And which causes account for most of the late results? Each question calls for a different tool: a run chart and a control chart for the first two, a Pareto chart for the third.

Run Chart

In the account given by Perla et al. (2011), a run chart is simply a line of values kept in the order they occurred and centered on their median, and they give four probability-based signals to look for. A shift is at least six successive weeks sitting on one side of the median line. A trend is at least five successive values each climbing, or each falling, from the one before. The count of runs, the stretches between crossings of the median, can be too small or too large to be chance. The fourth signal, which they call astronomical, is a single week so out of line with the others that no rule is needed to spot it. Their point is that keeping time order teaches a team far more than a quarterly average does.

The weekly percentages ranged from 49% to 67%, apart from one week, with a median of 58%. No stretch of weeks stayed on the same side of the median for more than four, so there was no shift. The longest sequence of increasing or decreasing points was three, so there was no trend. The 26 points crossed the median in a way that produced 13 runs, within the expected range for that number of points. Week 14 stood out at 31%. The laboratory's records show that the main analyzer was out of service for nine hours that week and specimens were sent to a backup instrument with a longer cycle. Apart from that week, the run chart shows no evidence that performance is improving or getting worse.

Control Chart

A control chart adds statistically calculated limits to the display, which helps distinguish common cause variation, the ordinary noise of a stable process, from special cause variation arising from something outside the usual system (Benneyan et al., 2003). Because the measure is a proportion with roughly equal weekly samples, a p-chart is appropriate. With a center line of 0.58 and an average weekly sample of 165, the standard error is about 0.038, which places the upper control limit near 69.5% and the lower limit near 46.5%.

Every week except week 14 falls inside the limits, and week 14 falls well below the lower limit, confirming it as special cause variation. The conclusion is important for management. Nothing in this process is broken; it reliably delivers an on-time result a little more than half the time, because that is all its design allows. A stable process cannot be pushed to 90% by urging staff to work harder, because the variation that remains is built into the system. Reaching the target will require changing the process itself. Treating an ordinary bad week as a crisis, as happened after two weeks near 50% in the spring, only produces reactions to noise.

Stratification

A single weekly percentage can hide differences between groups, so the results were stratified in two ways suggested by the process map. By shift, 64% of day-shift results met the standard, compared with 58% in the evening and 47% at night. By location, 63% of results for patients in rooms met it, compared with 49% for patients in hallway beds. First troponins met the standard 66% of the time and second troponins only 49%. Each difference points back to a finding from the process map: one phlebotomist at night, a shared tube station, hallway beds far from the label printer, and a second draw that depends on memory. The night and hallway gaps also bear on the equity aim, since patients placed in hallway beds at night are consistently served worse than others.

Pareto Analysis

To rank the causes of late results, the team reviewed a random sample of 200 results that missed the standard and assigned each a primary cause from its timestamps and the emergency record. Late second draws accounted for 74 results, or 37%. Specimens held before sending, with more than 20 minutes from collection to laboratory receipt, accounted for 46, or 23%. Delayed first draws, usually for patients in hallway beds or still in the waiting area, accounted for 30, or 15%. Hemolysis redraws accounted for 22, or 11%. The laboratory's shift-change queue accounted for 14, or 7%, and other causes, including instrument maintenance and mislabeled specimens, for the remaining 14.

Arranged in descending order with a cumulative line, the chart shows that the first three causes account for 75% of late results, and all three occur before the specimen reaches the laboratory. The ranking matches the hypotheses the team formed during mapping, with one exception: delayed first draws were more common than expected, largely because of patients waiting in the triage area before a bed was assigned.

Hemolysis as a Separate Signal

Hemolysis redraws were only the fourth-largest category, but they cause the longest individual delays and add a second needle stick, so the team looked at them separately. Over the 26 weeks, 2.8% of troponin specimens were rejected as hemolyzed. When the rejections were split by how the blood was drawn, 3.4% of specimens taken through a newly placed intravenous catheter were hemolyzed, compared with 0.6% of those drawn by straight needle venipuncture. The direction matches a randomized crossover comparison run in another hospital's emergency department, where Lowe et al. (2008) found 5.6% of catheter-drawn samples hemolyzed against 0.3% by venipuncture, although this hospital's gap is narrower. The finding matters for the next module in two ways. It suggests a change that could remove most redraws, and it warns that any change that pushes nurses to draw faster through catheters could raise the balancing measure defined in the first module. Hemolysis will therefore be tracked on its own chart alongside the primary measure.

Conclusions for Improvement

Three conclusions follow. First, the process is stable, so improvement requires redesigning it, and progress should be judged against the run chart rules and control limits rather than against any single week. Second, the most productive targets are when the second draw happens and the holding of specimens before they are sent, which together explain 60% of late results; the laboratory's own queue is a small contributor. Third, night shift and hallway beds deserve particular attention because they concentrate failures. The next module will plan tests of change aimed at the top two causes, and the p-chart built here will serve as the baseline against which those tests are judged.

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

Lowe, G., Stike, R., Pollack, M., Bosley, J., O'Brien, P., Hake, A., Landis, G., Billings, N., Gordon, P., Manzella, S., & Stover, T. (2008). Nursing blood specimen collection techniques and hemolysis rates in an emergency department: Analysis of venipuncture versus intravenous catheter collection techniques. Journal of Emergency Nursing, 34(1), 26-32. https://doi.org/10.1016/j.jen.2007.02.006

Perla, R. J., Provost, L. P., & Murray, S. K. (2011). The run chart: A simple analytical tool for learning from variation in healthcare processes. BMJ Quality & Safety, 20(1), 46-51. https://doi.org/10.1136/bmjqs.2009.037895