Statistically Significant, Clinically Small and Reported in Full: Interpreting a Coaching Project's Results Against a Minimal Clinically Important Difference and Writing Them Up to SQUIRE 2.0
[Student Name]
University of Phoenix
DNP/701: Biostatistics and Epidemiology
Week 8 Assignment
[Instructor Name]
[Date]
The clinic and its figures are a composite written for a model paper.
My DNP project evaluated nurse telephone coaching for adults with diabetes in our rural clinic network. After this course, I have results from three analyses: a comparison of A1c change between coached and uncoached patients, a comparison of goal attainment and an interrupted time series of our poor-control rate. This paper interprets those results for decision makers and plans how I will report them.
The Results in Brief
The coached group's A1c fell 0.5 percentage points more than the comparison group's, with a 95% confidence interval of 0.05 to 0.95 and a p value of .03. Goal attainment was 15 points higher, with an interval of minus 2 to plus 32 and a p value of .09. The time series showed a significant change in trend, with poor control about 6 points lower than projected by month 36.
Statistical Versus Clinical Significance
A statistically significant result tells me that the data are unlikely under the null hypothesis; it does not tell me whether the difference matters to patients. Jaeschke et al. (1989) introduced the minimal clinically important difference, the smallest change in an outcome that patients perceive as beneficial and that would lead a clinician to change management, and developed an approach anchoring changes in questionnaire scores to patients' global ratings of change.
Applying the Concept to A1c
Before analysis, our project team agreed that an A1c difference of 0.5 percentage points would be clinically meaningful for our population, since it would change medication decisions and is associated with lower risk of complications. Our point estimate, 0.5, equals that threshold. But the confidence interval, from 0.05 to 0.95, includes values well below it. The result is statistically significant, but we cannot be confident that the true effect reaches the clinically important level.
The p value said coaching did something; the confidence interval said it might have done too little to matter or nearly twice what we hoped.
Reading Confidence Intervals
Du Prel et al. (2009) explain that p values indicate whether to reject a null hypothesis, while confidence intervals show the range in which the true value lies with a stated level of confidence, along with the direction and size of the effect, enabling conclusions about clinical relevance. They recommend reporting both. For goal attainment, the p value of .09 alone might suggest no effect, but the interval shows that the data are compatible with anything from a small harm to a large benefit, which means the pilot was underpowered, not that coaching failed.
Why the Threshold Was Set in Advance
Setting the clinically important difference after seeing the results would invite bias, since a team could choose a threshold the results happened to meet. Our team documented 0.5 points in the project plan before data collection, which protects the interpretation from that temptation.
Clinical Meaning of Goal Attainment
Turned into a count, the 15-point gap means roughly seven patients must be coached for one more to reach goal. For a program costing roughly $120 per patient over six months, that is about $840 per additional patient reaching goal. Framing results in these terms helps decision makers weigh them.
Integrating the Three Analyses
The two-group comparison and the time series point in the same direction. The time series, with 36 months of data and control for baseline trends, adds confidence that the improvement is real. The two-group comparison, limited by nonrandom groups and small size, provides an estimate of effect for individual patients. Together they suggest coaching is associated with improvement of about half a point, with uncertainty about its exact size.
What I Will Recommend
Continue coaching, since the evidence favors it, the cost is modest and harm is unlikely. Evaluate it further with a larger sample and randomized rollout across sites, to estimate the effect more precisely and address site differences.
Consistency Across Methods
Two analyses with different weaknesses, a nonrandom two-group comparison and a 36-month time series, arrived at similar conclusions. Agreement between methods that fail in different ways increases confidence that the effect is real rather than an artifact of one design.
Reporting With SQUIRE 2.0
Ogrinc et al. (2016) developed SQUIRE 2.0 through a consensus process as guidelines for reporting quality improvement work. The guidelines ask authors to describe the problem and available knowledge, the rationale and theory, the context, the intervention and how it was studied, the measures, the analysis, ethical considerations, the results including unintended consequences and the interpretation, limitations and conclusions.
Applying SQUIRE to My Project
My final paper will describe the local problem, including our 27% poor-control rate; the rationale for coaching; the clinic context, including staffing and the rural population; the coaching protocol in enough detail to replicate it; the study design with both analyses; the measures and their definitions; the statistical methods including autocorrelation checks; and results with confidence intervals. I will report the concurrent pharmacist addition as a contextual factor and limitation.
Unintended Consequences
SQUIRE asks for unintended consequences. I checked hypoglycemia-related visits, which did not increase, and nurse workload, which rose by about four hours per week per nurse. Both are reported.
Patient Perspective
The minimal clinically important difference was originally anchored to patients' own ratings of change (Jaeschke et al., 1989). A1c is a laboratory value, not something patients feel. In a future phase, I will add a patient-reported measure of diabetes distress, for which an established threshold for meaningful change would let us ask whether patients themselves noticed a difference.
Limitations
Limitations include nonrandom comparison groups, a small sample for goal attainment, a concurrent change in pharmacist staffing and reliance on health record data. I will state each and its likely effect on the findings.
What the Results Do Not Show
The analyses do not show that coaching works equally for all patients. Subgroup results by age and site were too small to interpret, and I will not report them as findings. They do not show long-term effects beyond 18 months, and they do not establish cause with the certainty a randomized trial would.
Sustainability
Improvement that fades when a project ends does little for patients. I will track the poor-control rate for a year after the project and report whether the trend continued, which SQUIRE encourages as part of interpreting results.
Communicating to Different Audiences
For the clinic board, a one-page summary with the time series graph and one sentence on each finding. For clinicians, the effect sizes and intervals. For a professional audience, a SQUIRE-structured manuscript and poster.
Conclusion
My project's main result is statistically significant but uncertain in clinical importance, because the confidence interval extends below the minimal clinically important difference set in advance. A second outcome was imprecise rather than null. Reading intervals, integrating analyses and reporting with SQUIRE 2.0 will let me present what the project found honestly: coaching likely helps, by an amount that further study should pin down.
References
du Prel, J.-B., Hommel, G., Röhrig, B., & Blettner, M. (2009). Confidence interval or p-value? Part 4 of a series on evaluation of scientific publications. Deutsches Ärzteblatt International, 106(19), 335-339. https://doi.org/10.3238/arztebl.2009.0335
Jaeschke, R., Singer, J., & Guyatt, G. H. (1989). Measurement of health status: Ascertaining the minimal clinically important difference. Controlled Clinical Trials, 10(4), 407-415. https://doi.org/10.1016/0197-2456(89)90005-6
Ogrinc, G., Davies, L., Goodman, D., Batalden, P., Davidoff, F., & Stevens, D. (2016). SQUIRE 2.0 (Standards for QUality Improvement Reporting Excellence): Revised publication guidelines from a detailed consensus process. BMJ Quality and Safety, 25(12), 986-992. https://doi.org/10.1136/bmjqs-2015-004411
How this DNP 701 Week 8 example is structured
The DNP/701 Week 8 work usually closes with interpreting statistics for a DNP project. This paper brings the course together by moving from numbers to meaning: whether a result matters to patients, how certain it is and how to report it honestly and completely. Students search this week as DNP 701 Week 8, DNP701 Wk 8 or DNP/701 Wk 8; all three are the same assignment.
DNP/701 Week 8 questions, answered
What does DNP/701 Week 8 usually ask for?
Many sections ask students to interpret statistical findings for a DNP project, including clinical versus statistical significance, and to plan how results will be reported.
What is a minimal clinically important difference?
The smallest change in an outcome that patients perceive as beneficial and that would lead to a change in management; it helps judge whether a statistically significant result matters in practice.
What is SQUIRE 2.0?
The Standards for Quality Improvement Reporting Excellence, a set of guidelines for reporting quality improvement work, covering the problem, rationale, intervention, study of the intervention, measures, analysis, results and interpretation.
Write yours, or have the desk draft it
This paper is an original model document written by our desk, not a submitted student paper and not an official University of Phoenix document. Read it for the moves, then write your own to the instructions in your classroom. If you want one built to your exact prompt and rubric, the first custom sample is free and arrives in 24 to 48 hours.