NU800 Unit 7 Methods Section Draft Example

Before Lunch and Before Dinner, Triggered by the Order: A Methods Section Draft Specifying a Glucocorticoid Hyperglycemia Bundle on Two Composite Medical Units Closely Enough That Another Unit Could Run It From the Page

[Student Name]

Purdue University Global

NU800: DNP Project Proposal

Unit 7 Methods Section Draft

[Instructor Name]

[Date]

The units, volumes and dates are composite; figures are bracketed. The section is written in the future tense, as in a proposal.

Design

This will be a quality improvement project with a pre-post comparison, using weekly run charts to compare [16] baseline weeks, drawn retrospectively, with [12] implementation weeks that follow a [two]-week education period. Its report will follow SQUIRE 2.0, the consensus standard for writing up improvement work (Ogrinc et al., 2016). The design cannot establish causation, because other changes over the period could affect results, but it can show whether measures shift in a way unlikely to be chance and whether the bundle was delivered as planned.

Setting

The project will take place on 4 North and 4 South, two adjacent [30]-bed adult medical units at [Lakeside Medical Center], a [340]-bed community teaching hospital. The units share a hospitalist group, a pharmacist who covers both during the day, and a point-of-care glucometer system that uploads each reading with a timestamp and patient identifier. Across both units there are about [90] glucocorticoid patient-days a week. Lunch trays arrive between [1130] and [1215], and dinner trays between [1630] and [1715].

Sample

The sample will include every patient-day on either unit for adults receiving a systemic glucocorticoid at a daily dose of at least [20] mg in prednisone equivalents, during the baseline and implementation periods, an expected [1,440] and [1,080] patient-days respectively. Equivalence will be calculated from a table built into the order set, for example [4] mg of dexamethasone or [16] mg of methylprednisolone equaling [20] mg of prednisone. Patient-days will be excluded for patients on insulin infusions, in diabetic ketoacidosis or a hyperosmolar state, pregnant or receiving comfort care only.

The Intervention

The bundle has four parts, each described below at the level of who acts, when and in response to what. First, when a prescriber enters a qualifying glucocorticoid order, the order set will schedule two capillary checks a day, one ahead of the midday meal and one ahead of the evening meal, for the first [48] hours in all patients, and at bedtime for patients with known diabetes, continuing beyond [48] hours if any reading exceeds 180 mg/dL. Checks will be done by the nurse or patient care technician within [30] minutes before the tray arrives. Second, when two readings exceed 216 mg/dL within [24] hours, the nurse will page the prescriber using a scripted message that includes the readings and a pharmacy-approved suggestion of morning NPH insulin at [0.1] units per kilogram for every [10] mg of prednisone-equivalent, to a maximum of [0.4] units per kilogram, given with the steroid dose, consistent with guidance that intermediate-acting insulin can match a morning steroid's effect (Roberts et al., 2018). Third, when a steroid dose is reduced, the order set will prompt the prescriber to reduce NPH by the same proportion. Fourth, before discharge, any patient with a reading above 200 mg/dL during the stay and no diabetes diagnosis will receive a follow-up plan: an A1C if not done, and a primary care appointment within [two] weeks.

The page will carry both the readings and a suggested dose, because a prescriber who receives a number with no proposal tends to add correctional insulin and move on.

Implementation Procedures

During the two education weeks, every nurse and technician on both units will attend a [15]-minute huddle session on the steroid glucose curve and the bundle, with attendance recorded. Hospitalists will receive the same content at their weekly meeting. A break-coverage list naming pre-lunch checks will be posted, and electronic prompts will fire at [1100] and [1600]. The project lead will round on each unit twice a week during implementation to answer questions and observe checks.

Measures

Four measures will be collected. The primary outcome counts, for each week, how many monitored steroid days contained a glucose above 250 mg/dL, and divides by the number of monitored steroid days, expressed per 100. A steroid day counts as monitored when it has two or more glucose results and at least one of them falls after [1100]; this restriction, explained in the PICOT worksheet, stops improved detection from appearing as worse control. The main process measure is afternoon check completion, reported as two separate figures, one for the pre-lunch check and one for the pre-dinner check, each as a share of eligible steroid days. Further process measures cover how often qualifying readings lead to a page, and how often a page leads to a scheduled insulin order within [24] hours. The balancing measure counts steroid days with any glucose under 70 mg/dL, per 100 steroid days.

Fidelity

Delivery of the bundle will be checked separately from its outcomes. Each week the project lead will record the share of qualifying steroid orders that triggered the order set, the share of pre-lunch and pre-dinner checks completed within [30] minutes of the tray, the share of escalation criteria followed by a page within [two] hours, and the share of steroid reductions followed by an NPH adjustment. These fidelity measures will show whether a lack of change in outcomes reflects a bundle that did not work or a bundle that was not delivered, which is the question an evaluation of this kind most often cannot answer.

Data Collection

Glucose readings will be exported weekly from the glucometer system and linked to steroid orders from the pharmacy record. Escalations will be identified from paging logs and nursing notes. Discharge plans will be reviewed in the charts of patients newly hyperglycemic. Baseline data for the [16] weeks will be drawn the same way from stored records.

Analysis

Each measure will be charted by week, and the median of the baseline weeks will be extended into the implementation period. Run chart rules will be used to identify nonrandom change: a run of at least six consecutive weeks falling on a single side of the median, a trend of five or more points in one direction, an unusual number of runs, or an astronomical point (Perla et al., 2011). Because weekly denominators will be modest, results will also be summarized as pooled rates for each period, with a simple comparison of proportions reported descriptively.

Protection of Participants

The hospital's institutional review board will be asked to confirm that the project counts as quality improvement and does not require review as research. Data will be extracted by the project lead and stored without names, using a project number linked to medical record numbers in a separate, locked file. No patient will receive care outside the hospital's approved protocols, and the NPH suggestion will be an option for the prescriber, not a standing order.

References

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 & Safety, 25(12), 986-992. https://doi.org/10.1136/bmjqs-2015-004411

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

Roberts, A., James, J., Dhatariya, K., & Joint British Diabetes Societies for Inpatient Care. (2018). Management of hyperglycaemia and steroid (glucocorticoid) therapy: A guideline from the Joint British Diabetes Societies (JBDS) for Inpatient Care group. Diabetic Medicine, 35(8), 1011-1017. https://doi.org/10.1111/dme.13675