NURS 8210 Week 10 Informatics Evaluation Plan Example

The Right Milk for the Right Baby: A RE-AIM Evaluation Plan for Barcode Scanning of Human Milk in Three Neonatal Intensive Care Units

Student Name

College of Nursing, Walden University

NURS 8210: Transforming Nursing and Healthcare Through Technology

Instructor Name

Month Day, Year

The Right Milk for the Right Baby: A RE-AIM Evaluation Plan for Barcode Scanning of Human Milk in Three Neonatal Intensive Care Units

The Problem as the Steering Group Sees It

Evaluation plans usually fail by measuring what is easy. This plan was written to measure what the rollout was bought to change. A composite health system with three neonatal intensive care units, one level IV and two level III, will introduce barcode scanning for expressed human milk. Each container will receive a printed label linking it to the mother and infant, and nurses will scan the infant's band and the container before every feeding. The scanner will warn if the milk belongs to another infant, has passed its expiration time or does not match an ordered fortifier.

The steering group's concern is feeding an infant another mother's milk. Such errors expose the infant to potential infection risk, lead to testing and disclosure conversations with two families, and damage the trust of mothers who work hard to provide milk. Before the change, the units rely on a two-person visual check of handwritten or printed labels, and errors are known only through incident reports. The system's leaders have also heard that one children's hospital reported administration errors falling to zero after it introduced centralized milk handling and barcode scanning, while the scanner caught dozens of attempts to feed the wrong milk in its first six months (Steele & Bixby, 2014). The second finding is the more important for evaluation: scanning reveals errors that visual checks never reported.

Framework: RE-AIM

The plan uses RE-AIM, which Glasgow et al. (1999) proposed to evaluate interventions on five dimensions rather than on effectiveness alone: reach, effectiveness, adoption, implementation and maintenance. The framework suits this rollout because a scanning system can be effective in principle and still fail in practice if nurses bypass it, if one unit adopts it and another does not, or if use fades once the go-live team leaves. All five dimensions are used, each with a defined measure.

Definitions of the Five Dimensions for This Rollout

Reach is the share of feedings of expressed human milk in which both the infant's band and the milk container were scanned before the feeding. Feedings of formula and direct breastfeeding are excluded, since they do not involve a stored container.

Effectiveness is the rate of misadministration events: an infant receiving milk from another mother, or expired milk, per 1,000 human milk feedings. Because scanning will reveal events previously missed, effectiveness also includes near misses intercepted by the scanner, reported separately.

Adoption is the share of units and of nursing staff, including float and agency nurses, who have used the scanner for milk feedings at least once in each month.

Implementation is fidelity: the share of scans that follow the intended sequence, with overrides and workarounds counted. A workaround might be scanning a spare label kept at the station instead of the container.

Maintenance is whether reach and effectiveness hold six and twelve months after the go-live team leaves, and whether the policy has been incorporated into orientation and annual competencies.

Measurement Table

Table 1 gives each measure its definition, data source, frequency and owner.

Table 1

Measures for the Human Milk Scanning Evaluation

Baseline data will come from incident reports and milk room records for the twelve months before go-live, with the caveat that baseline events were likely underreported.

The Baseline Problem

A plain before-and-after comparison of reported errors would mislead. Before scanning, errors were found only when someone noticed and reported them. After scanning, the system itself will record every attempt to feed the wrong milk. The reported rate of near misses will therefore almost certainly rise after go-live, not because care got worse but because the system can finally see. The plan addresses this by keeping reported misadministration events and scanner-intercepted near misses as separate measures, and by interpreting the first months of near-miss data as a new baseline rather than as a change.

Methods

The plan uses run charts rather than a trial design. The rollout will occur on all three units within a few months, and randomizing units or nurses is neither feasible nor ethical once the safety case is accepted. Run charts show data over time with the median marked, and simple rules can detect nonrandom signals such as shifts and trends (Perla et al., 2011). Go-live dates for each unit will be annotated, so a change that appears on each unit at its own go-live, and not before, is more credibly linked to scanning. Staggered go-lives across the three units strengthen that inference at no extra cost.

Observation audits will supplement the logs. Each month, a clinical nurse specialist will observe a sample of feedings on each unit to check whether scans match the intended sequence and to identify workarounds the log cannot show.

Declared Failure Threshold

Before go-live, the steering group has agreed on what would count as failure. If reach remains below 90% on any unit three months after its go-live, or if any wrong-infant feeding occurs after go-live in which the scanner was bypassed, the steering group will treat the rollout as failing on that unit and review workflow, hardware and training before continuing. Declaring these thresholds in advance prevents the evaluation from redefining success after the results are known.

Limitations

The plan cannot establish causation with the certainty of a trial. Other changes, such as a new milk room process or staffing changes, may coincide with go-live, and the steering group will log them. Incident reporting before and after scanning differs, which limits comparison of misadministration rates. And reach measured from the scanning log depends on complete feeding documentation; if nurses skip documentation as well as scanning, reach will be overstated. Observation audits partly address this.

Other hospitals have already reported that this kind of scanning can prevent wrong-milk feedings by catching attempts that would otherwise have gone unnoticed (Dougherty & Nash, 2009; Steele & Bixby, 2014). The evaluation's purpose is to show whether that happens in these three units, reliably and over time.

References

Dougherty, D., & Nash, A. (2009). Bar coding from breast to baby: A comprehensive breast milk management system for the NICU. Neonatal Network, 28(5), 321-328. https://doi.org/10.1891/0730-0832.28.5.321

Glasgow, R. E., Vogt, T. M., & Boles, S. M. (1999). Evaluating the public health impact of health promotion interventions: The RE-AIM framework. American Journal of Public Health, 89(9), 1322-1327. https://doi.org/10.2105/AJPH.89.9.1322

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

Steele, C., & Bixby, C. (2014). Centralized breastmilk handling and bar code scanning improve safety and reduce breastmilk administration errors. Breastfeeding Medicine, 9(9), 426-429. https://doi.org/10.1089/bfm.2014.0077