HIM 400 Module 4 Project One Example

Project One: Rising for Eight Quarters, Trends and Patterns in Diabetes Control at Cold Brook Health

[Student Name]

Southern New Hampshire University

HIM 400: Communication and Technologies II

Project One

[Instructor Name]

[Date]

The organization, setting and figures below are a composite written as a model document. No real employer, client, colleague or patient is described.

Project One: Rising for Eight Quarters, Trends and Patterns in Diabetes Control at Cold Brook Health

Cold Brook Health's quality committee noticed that its poor diabetes control rate had crept upward and asked whether the change was real, where it was happening and why. This project answers those questions using registry data from January 2024 through December 2025. It defines the measure, tests the trend with a control chart, breaks the rise into its parts, compares patient groups and practices and ends with limitations and recommendations.

Data and Measure

The analysis uses the relational registry and the query specification from Modules Two and Three. Each quarter's denominator is adults aged 18 to 75 in the registry who had a primary care visit in the prior twelve months, excluding patients in hospice or palliative care; it ranged from 6,612 to 6,794. The numerator is patients whose latest HbA1c in the twelve-month window was above 9.0% or who had no test. Because the window rolls forward each quarter, consecutive quarters share nine months of data, which smooths the series and means small quarter-to-quarter changes should not be overread.

National data put the local figures in context. Ali et al. (2013) found that American adults with diabetes made real progress toward glucose, blood pressure and cholesterol goals between 1999 and 2010, yet a substantial share still fell short of recommended targets. Cold Brook's rate is therefore not unusual, but its direction is.

Is the Trend Real?

Table 1 shows the quarterly rates. A rise from 24.1% to 28.0% could still be random noise, so the analyst plotted the rates on a p-chart, the control chart used for proportions. Benneyan et al. (2003) describe how a control chart separates the ordinary scatter that any stable process produces, called common cause variation, from special cause variation, the kind that points to a change in the process itself. Mohammed et al. (2008) recommend deriving limits from the data and applying a small set of agreed rules for detecting signals.

Using the eight-quarter mean of 26.1% and an average denominator of about 6,700 patients, the three-sigma limits are roughly 24.5% and 27.7%. Three signals appear. The first quarter sits below the lower limit, the last quarter lies above the upper limit and the rate climbed at every one of the seven steps from quarter to quarter, a run longer than the six or so consecutive increases usually accepted as evidence of a trend. The rise is real.

Table 1. Quarterly Rate of Poor Control or No Test

Note. Rolling twelve-month window ending each quarter; composite registry data.

Splitting the Rise

The measure combines two different problems: patients who were tested and had high values, and patients who were not tested at all. Over the period, the no-test share rose 2.3 points, from 3.9% to 6.2%, while the tested-and-high share rose 1.6 points, from 20.2% to 21.8%. Almost 60% of the increase therefore comes from missing tests rather than worsening results. The no-test share began rising in the third quarter of 2024, shortly after Cold Brook closed its lab draw station in Harlow Falls, which had served two of the most rural practices.

Patterns Across Patient Groups

Table 2 compares subgroups in the final quarter. Poor control or no test was most common among adults aged 18 to 44, patients covered by Medicaid and patients living more than 30 miles from their practice. The distance pattern is the sharpest, and it is driven mostly by missing tests, which fits the draw station closure.

Table 2. Poor Control or No Test by Subgroup, Q4 2025

Note. Travel distance calculated from home ZIP code centroid to practice address.

Comparing Practices Fairly

Ranking the 14 practices by rate would be misleading, because small practices swing more by chance. Spiegelhalter (2005) proposed funnel plots, which place each unit's rate against its size with control limits that widen as size falls, so that only units outside the limits are treated as unusual. On a funnel plot around the system rate of 28.0%, twelve practices sit inside the 99.8% limits. Two sit above: Harlow Falls, with 310 patients and a rate of 39.4%, and Birch Hill, with 820 patients and a rate of 35.1%. Both lost access to the closed draw station. The practice with the lowest rate, 21.6%, is small enough that its result is still within the limits, so it should not be held up as a model on this evidence alone.

How Much Is Data, Not Care?

Some missing tests may not be missing. The Module Three validation found that 11 of 50 flagged patients had results in scanned outside reports or notes that the query could not read. Parsons et al. (2012) found that electronic quality measures often understated performance for the same reason. If the same share held across the no-test group, roughly a fifth of it may reflect results that exist but were never entered as structured data. This does not erase the trend, since the tested-and-high share also rose, but it means the outside lab interface discussed in Module One is part of the fix.

Limitations

The rolling window links neighboring quarters, so the run of increases partly reflects overlap, although the final quarter's position above the control limit does not depend on that. Distance was estimated from ZIP code centroids. The analysis describes associations and timing and cannot prove that the draw station closure caused the rise, and patient-level factors such as insulin access were not available.

Recommendations

Three actions follow from the findings. First, restore testing access near Harlow Falls and Birch Hill through a mobile draw schedule or point-of-care HbA1c testing during visits. Second, finish the outside lab interface and enter scanned results so the registry reflects care already delivered. Third, direct outreach toward adults under 45 and Medicaid patients, whose tested results are also worse. The committee should keep the p-chart monthly and use the funnel plot each quarter to see whether the two outlying practices return inside the limits.

Conclusion

Cold Brook's rise in poor diabetes control is a real trend, not noise. Most of it comes from patients who went untested, concentrated among those living far from care after a draw station closed, while a smaller part reflects worsening results among younger and Medicaid patients. Part of the gap is a data problem. Project Two will propose a platform to monitor these patterns continuously.

References

Ali, M. K., Bullard, K. M., Saaddine, J. B., Cowie, C. C., Imperatore, G., & Gregg, E. W. (2013). Achievement of goals in U.S. diabetes care, 1999-2010. New England Journal of Medicine, 368(17), 1613-1624. https://doi.org/10.1056/NEJMsa1213829

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

Mohammed, M. A., Worthington, P., & Woodall, W. H. (2008). Plotting basic control charts: Tutorial notes for healthcare practitioners. Quality and Safety in Health Care, 17(2), 137-145. https://doi.org/10.1136/qshc.2004.012047

Parsons, A., McCullough, C., Wang, J., & Shih, S. (2012). Validity of electronic health record-derived quality measurement for performance monitoring. Journal of the American Medical Informatics Association, 19(4), 604-609. https://doi.org/10.1136/amiajnl-2011-000557

Spiegelhalter, D. J. (2005). Funnel plots for comparing institutional performance. Statistics in Medicine, 24(8), 1185-1202. https://doi.org/10.1002/sim.1970