DNP 750 Week 6 Methods, Measures and Data Collection Example

Counting What the Question Asks: Operational Measures, Data Collection Procedures and an Analysis Plan for a Diabetes Distress Screening Pilot

[Student Name]

University of Phoenix

DNP/750: DNP Applied Project I

Week 6 Assignment

[Instructor Name]

[Date]

The health center, staff and figures are composites written for a model paper.

The practice question for this project names three outcomes: the share of eligible diabetes visits that include a distress screen, the share of positive screens that receive follow-up within 14 days, and the change in distress among patients who screen positive. This paper sets out the methods that will produce those numbers. It defines who is included, describes the procedures in the order a patient would meet them, presents the instruments and their properties, writes each measure as a calculation, and explains how the data will be collected, protected and analyzed.

Eligibility

An eligible visit is any in-person or video appointment held during the 12 implementation weeks for adult patients (18 and over) with type 2 diabetes recorded on the problem list, when the visit is scheduled as a diabetes follow-up or a chronic care visit. Visits for acute problems only, such as a sore throat, are excluded because they are too short for screening and follow-up. Patients with type 1 diabetes, documented dementia or an active behavioral health crisis at the visit are excluded from the screening workflow but receive usual care. A patient may have more than one eligible visit, but the distress outcome uses only the first positive screen.

Procedures

At rooming, the medical assistant hands the patient a tablet showing the two-item screen in English or Spanish, with a one-sentence explanation that the questions are about the stress of living with diabetes and that answers go to the care team. The tablet scores the screen automatically. A mean score of 3 or higher triggers a prompt for the patient to complete the full 17-item scale on the same tablet before the clinician enters. The clinician sees both results in the visit note and may discuss them. Before the patient leaves, the care manager is messaged; she reviews the subscale scores that day and either meets the patient in clinic or calls within 14 days. At that contact, she and the patient choose a response: problem-solving support for regimen distress, a scheduled review with the clinician for physician-related distress, a family or peer resource for interpersonal distress, or a warm handoff to the behavioral health consultant for high emotional burden. A repeat full scale is given at the next diabetes visit or by phone at 12 weeks.

Instruments

The Diabetes Distress Scale screener (DDS2) asks how much the patient has been bothered by feeling overwhelmed by the demands of living with diabetes and by feeling that they are often failing with their diabetes routine, each rated from 1, not a problem, to 6, a very serious problem. Fisher et al. (2008) derived it from the full scale and found that a mean of 3 or more identified patients with at least moderate distress and that it flagged 95% of patients with at least moderate distress on the full scale while correctly clearing 85% of those without it.

The 17-item Diabetes Distress Scale (Polonsky et al., 2005) uses the same six-point response format and yields a total mean item score and four subscale scores covering the emotional load of the disease, strain in the relationship with the clinician, strain from the treatment routine and strain in relationships with family and friends. Internal consistency exceeded .87 for the total and subscales in the development samples, and scores correlated as expected with depressive symptoms, self-care and A1C. Because the scale reports its result as a mean item score from 1 to 6, the same cut points can be applied to the total and to each subscale, which lets the care manager see at a glance whether the distress is mainly about the regimen, the clinician or the family. Spanish versions of both instruments, available from the developers, will be used for patients who prefer Spanish.

Measures

Measure 1, screening rate: eligible visits with a completed DDS2 divided by all eligible visits, reported weekly. Measure 2, timely follow-up: the number of positive DDS2 results followed by a charted care manager call or visit no later than day 14, over the number of positive DDS2 results, reported weekly with a two-week lag. Measure 3, distress change: the mean difference between the first and the 12-week total scale score among patients with both scores, reported at the end, together with the number moving from the high level (3.0 or above) to a lower level. Measure 4, a balancing measure, is the median minutes added to rooming, timed by the medical assistant champion on 20 visits in each improvement cycle. An exploratory measure, A1C at baseline and at the first result after 12 weeks, will be reported descriptively only.

Baseline Data

For the 12 weeks before implementation, the quality coordinator will extract eligible visits and depression screens. Because no distress screening existed, the baseline for measures 1 and 2 is expected to be zero, and the baseline period mainly confirms the volume of eligible visits and the stability of the workflow.

Data Collection

The electronic health record analyst will build structured fields for both instruments and a report listing, for each eligible visit, the date, a study number, screen status and score, full scale scores and the date of the first care manager contact. The quality coordinator will run the report every Monday and send it to me as a file stored on the health center's secure drive. I will enter weekly counts into run charts and keep a separate log of improvement cycle notes and rooming times.

Analysis Plan

Weekly values for measures 1, 2 and 4 go onto run charts whose center line is the median of the first weeks. Perla et al. (2011) explain how to read such charts. Six points in a row above the median, or six in a row below it, count as a shift; five points in a row that each rise, or each fall, count as a trend; and an unusually small or large number of runs also points to real change rather than chance. These rules will be used to judge whether each improvement cycle changed the process. For measure 3, each patient's first and 12-week scores will be compared using a paired t test, switching to the Wilcoxon signed-rank test if the differences prove clearly skewed, with a two-sided alpha of .05; the mean change and its 95% confidence interval will be reported rather than the p value alone. Subgroups by language and age will be described but not tested, since the numbers will be small.

Sample Size

About 200 patients with type 2 diabetes are expected to attend an eligible visit in 12 weeks. If about a third screen positive and most complete a repeat scale, around 50 to 60 patients will have paired scores, which is enough to estimate a moderate mean change with reasonable precision in a quality improvement project.

Missing Data

A screen will count as completed only if both items are answered. Patients who screen positive but do not finish the full scale will be offered it by phone. Patients without a repeat score will be described, including their first scores, so that any difference from those who completed both can be seen.

Data Management and Protection

Reports will use study numbers instead of names, the key linking numbers to records will stay with the quality coordinator, and files will remain on the secure drive. No data will be kept on personal devices, and results will be reported only in aggregate, with counts under 11 suppressed.

Conclusion

Each outcome in the practice question now has a defined measure, a data source and a time point, and the analysis plan matches the type of data. The methods are specific enough for another site to repeat. The next paper will address feasibility, ethics and site approval.

References

Fisher, L., Glasgow, R. E., Mullan, J. T., Skaff, M. M., & Polonsky, W. H. (2008). Development of a brief diabetes distress screening instrument. Annals of Family Medicine, 6(3), 246-252. https://doi.org/10.1370/afm.842

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

Polonsky, W. H., Fisher, L., Earles, J., Dudl, R. J., Lees, J., Mullan, J., & Jackson, R. A. (2005). Assessing psychosocial distress in diabetes: Development of the Diabetes Distress Scale. Diabetes Care, 28(3), 626-631. https://doi.org/10.2337/diacare.28.3.626