MN504 Unit 6 seminar reflection example

What Else Changed That Year? A Seminar Reflection on a Before-and-After Report I Had Championed

[Student Name]

Purdue University Global

MN504: Evidence-Based Practice

Unit 6 Seminar Reflection

[Instructor Name]

[Date]

The report and seminar participants are composites. Classmates and the instructor are identified by role only.

Reflection

Going into the Unit 6 seminar, I was proud of one row in my evidence table. It described an internal quality report from a sister hematology-oncology unit in our hospital system, where central line infections fell from 2.34 to 1.08 per 1,000 line days after daily chlorhexidine bathing began. I had placed it first for fit, because its patients were the closest to ours, and I planned to lean on it in my synthesis as the local proof that bathing works on an oncology unit.

In the seminar, the instructor asked each of us to list anything else that changed on our lead study's unit during its after period. I had not asked that question. A classmate who had read the same report pointed to one sentence in its methods: chlorhexidine-impregnated dressings for central lines were introduced on that unit during the same months as the bathing. I had read that sentence and passed over it. Chlorhexidine dressings have their own evidence for reducing catheter-related infection in critically ill adults (Timsit et al., 2009). The drop in infections could have come from the dressings, from the bathing, from both or from neither, and the report's design had no way to tell them apart.

Rereading the report afterward, I could name the threats. A before-and-after comparison without a concurrent control group cannot rule out co-interventions, other changes introduced during the same period, or secular trends, changes that were already under way across the hospital regardless of the intervention (Harris et al., 2006). Infection rates were falling across the whole system that year after a new central line insertion checklist. With only 9 infections before and 4 after, the rate ratio of 0.46 had a confidence interval from 0.14 to 1.49, which crosses 1. The apparent halving was never as solid as I had treated it, even before the seminar.

The seminar also made me look at why I had ranked the report first. I had followed my own rule, fit before level, and that rule was sound: evidence from patients like ours matters. But I had let high fit stand in for high confidence. The Johns Hopkins model places an internal quality report at Level V for a reason, since such reports are designed to track a unit's own performance rather than to test a cause (Dang et al., 2022). I had rated it correctly in one column and then argued from it as if the other columns did not exist. Fit tells me whether a result would apply to our unit if it were true; level and quality tell me how likely it is to be true. I need both before I lean on anything.

I went back through my table after the seminar and asked the same question of every row. The cluster-randomized trials were protected from co-interventions by design, because units in both arms experienced the same hospital-wide changes during the same months. The hospital-wide implementation study in row 3 was not; it described a year in which several infection-prevention changes were introduced together, and I have lowered my confidence in its outcome figures for the same reason.

What the seminar undermined was the claim that bathing caused the drop. What survives is different and still useful. The report describes patients very like ours, receiving chemotherapy through ports, and it shows that nursing assistants on a busy oncology unit could deliver daily cloth baths to more than 90 percent of eligible patients for a year. That is evidence about feasibility and fit, not about effect, and it is worth having. I have also learned a design habit I will keep: when I read any pre-post comparison, I will look for everything else that changed in the same window before I believe the result.

In my synthesis in Unit 7, this report will move from the lead to a supporting role. I will cite it for feasibility and for the similarity of its population, quote its interval whenever I mention its result, and name the dressing change as a co-intervention. The lead evidence on effect will come from the randomized trials, with the caution that most of them studied intensive care patients rather than patients like ours. I will also add a short paragraph to the synthesis on co-interventions, so that anyone reading my proposal understands why a promising local result is not carrying the argument, and so that our own pilot, if it happens, records every other change made on the unit during the same months.

References

Dang, D., Dearholt, S. L., Bissett, K., Ascenzi, J., & Whalen, M. (2022). Johns Hopkins evidence-based practice for nurses and healthcare professionals: Model and guidelines (4th ed.). Sigma Theta Tau International.

Harris, A. D., McGregor, J. C., Perencevich, E. N., Furuno, J. P., Zhu, J., Peterson, D. E., & Finkelstein, J. (2006). The use and interpretation of quasi-experimental studies in medical informatics. Journal of the American Medical Informatics Association, 13(1), 16-23. https://doi.org/10.1197/jamia.M1749

Timsit, J.-F., Schwebel, C., Bouadma, L., Geffroy, A., Garrouste-Orgeas, M., Pease, S., Herault, M.-C., Haouache, H., Calvino-Gunther, S., Gestin, B., Armand-Lefevre, L., Leflon, V., Chaplain, C., Benali, A., Francais, A., Adrie, C., Zahar, J.-R., Thuong, M., Arrault, X., ... Lucet, J.-C. (2009). Chlorhexidine-impregnated sponges and less frequent dressing changes for prevention of catheter-related infections in critically ill adults: A randomized controlled trial. JAMA, 301(12), 1231-1241. https://doi.org/10.1001/jama.2009.376