DNP 810 Module 4 PDSA Design and Family of Measures Example

Predict, Test, Learn: A Three-Cycle PDSA Design and Family of Measures for On-Time Second Antibiotic Doses

Student Name

Doctor of Nursing Practice Program, Aspen University

DNP 810: Evidence-based Practice for Quality Improvement

Instructor Name

Month Day, Year

Predict, Test, Learn: A Three-Cycle PDSA Design and Family of Measures for On-Time Second Antibiotic Doses

The Plan-Do-Study-Act cycle is the most widely used method for testing changes in health care improvement, and one of the most misused. Teams often call a single large rollout a PDSA cycle, skip predictions and collect data only at the end. This paper designs a series of PDSA cycles for the redosing project for septic patients still in the emergency department after admission, following the core features of the method, and defines the outcome, process and balancing measures that will tell the team whether things got better.

What Good PDSA Practice Requires

A systematic review of 73 published articles reporting use of PDSA in health care assessed them against key features of the method: iterative cycles, prediction-based testing, small-scale initial tests and use of data over time. Many reports did not follow these features. Just 14 of the 73 reports laid out a full series of connected cycles, and 7 of the 47 examined closely relied on data gathered monthly or more often when choosing each next cycle (Taylor et al., 2014). The design below is built to avoid those failures: each cycle has a stated prediction, starts small and uses frequent data to decide the next step.

The Three Questions

The Model for Improvement asks three questions before any cycle begins. What are we trying to accomplish? To reduce the proportion of sepsis patients boarding in the emergency department whose second antibiotic dose has a major delay, meaning the dose arrives a quarter or more of the dosing interval late, from 38% to 15% within six months. What would show the team things are better? Through the family of measures described below, plotted over time. What change can we make that will result in improvement? A redosing process in which the emergency pharmacist places the follow-up dose order as soon as the first dose is charted, and the nurse receives a task reminder before it is due, adapted from a pharmacist-driven approach that reduced major delays from 48% to 13% at an academic emergency department (Payne-Cardona et al., 2021).

Each answer is written so that the whole team, including night staff who will not attend planning meetings, can repeat it in a sentence.

Planned Cycles

Cycle 1 tests the process on one day shift with one pharmacist and two nurses caring for sepsis boarders. The prediction is that the pharmacist will enter second-dose orders for all eligible patients within 30 minutes of the first dose, and that nurses will find the task reminder useful. The team will observe, time each step and interview the participants at the end of the shift. Cycle 2, planned after one week of daily single-shift tests and adjustments, extends the process to all day shifts for two weeks. The prediction is that at least 80% of eligible patients will have the second dose ordered within 30 minutes and that major delays will fall below 20% for those patients. Cycle 3 extends it to nights and weekends, when pharmacist coverage is thinner, for four weeks, with a prediction that on-time performance will be lower on nights and that a remote pharmacist verification pathway will be needed.

After each cycle, the team studies the data against the prediction, decides whether to adopt, adapt or abandon the change, and documents the reasoning. If the night shift cannot meet the target with the remote pathway, the team will test an alternative, such as allowing the emergency physician to enter the second-dose order with the first, before expanding further.

The Family of Measures

For the outcome, the team counts how many eligible patients have a major second-dose delay, as a percentage, calculated weekly during testing and monthly after. The numerator is patients whose second dose was given at or beyond 125% of the recommended interval after the first dose; the denominator is adults admitted with sepsis whose initial antibiotic was given in the emergency department and who remained there when the second dose was due. Time from first to second dose, in hours, is a secondary outcome.

Process measures tell the team whether staff are actually doing the new steps: the percentage of eligible patients with a second-dose order entered within 30 minutes of the first dose, and the percentage of task reminders acknowledged by the nurse. Balancing measures check for unintended consequences: duplicate antibiotic doses caused by both the emergency and inpatient teams ordering the second dose, pharmacist time spent per patient, and nurse-reported interruptions from the new reminders. Duplicate doses matter most, because a process that prevents late doses could create extra ones if admission orders are not reconciled.

Leisman and colleagues found that delays were most frequent with shorter dosing intervals, so results will also be stratified by six-hour, eight-hour and twelve-hour intervals, where improvement should be largest for the shortest (Leisman et al., 2017).

Data Collection and Display

Data come from the electronic medication administration record, using barcode scan times, which are more reliable than manually entered times, and from the pharmacy order log. The informatics analyst will build a weekly report listing eligible patients and their first and second dose times. Each measure will be plotted weekly against its baseline median, annotated with each cycle's start date, so that the team can see whether changes coincide with shifts in performance.

The team will also keep a simple cycle log: for each test, the date, scope, prediction, result, lesson and next step. The log turns the project's learning into a record that later chapters, and other teams considering the same change, can use.

Conclusion

A well-designed PDSA plan answers the Model for Improvement's three questions, tests small and predicts before testing, and uses frequent data to decide what comes next. For the redosing project, three expanding cycles, with a family of measures that includes duplicate doses as a balancing measure, provide a disciplined path from one shift's test to a department-wide change, and avoid the shortcuts that weaken so many published improvement reports.

References

Leisman, D., Huang, V., Zhou, Q., Gribben, J., Bianculli, A., Bernshteyn, M., Ward, M. F., & Schneider, S. M. (2017). Delayed second dose antibiotics for patients admitted from the emergency department with sepsis: Prevalence, risk factors, and outcomes. Critical Care Medicine, 45(6), 956-965. https://doi.org/10.1097/CCM.0000000000002377

Payne-Cardona, M., San Luis, V. A., Aazami, R., Dermendjieva, M., Erin, M., Kirkwood, J., Tong, C., Marks, G., Smith, E. A., Torbati, S. S., & Gilmore, J. F. (2021). Pharmacist driven antibiotic redosing in the emergency department. The American Journal of Emergency Medicine, 50, 160-166. https://doi.org/10.1016/j.ajem.2021.07.039

Taylor, M. J., McNicholas, C., Nicolay, C., Darzi, A., Bell, D., & Reed, J. E. (2014). Systematic review of the application of the plan-do-study-act method to improve quality in healthcare. BMJ Quality & Safety, 23(4), 290-298. https://doi.org/10.1136/bmjqs-2013-001862