NU710 Unit 8 Glucose Meter Integration Evaluation Example

Ninety Seconds Between the Meter and the Chart: A Unit 8 Simulation Evaluation of Wireless Glucose Meter Integration Before Go-Live

[Student Name]

Purdue University Global

NU710: Advanced Health Information and Technologies

Unit 8 Device Integration Evaluation

[Instructor Name]

[Date]

The hospital, the meters, the simulation and its findings are composite and illustrative; the studies cited are real.

Why Integration Matters for Insulin

With the current meters, a nurse checks a glucose at the bedside, writes or remembers the number, docks the meter later and enters the correction dose from the scale. Results reach the chart minutes to hours after they are taken. The wireless meters are meant to close that gap: the reading appears in the record within seconds, and the medication record calculates the correction dose from it, reducing transcription and calculation errors.

Integration changes more than speed. It moves a safety check from the nurse's head into an interface between two systems built by different companies. Sittig and Singh (2010) warn that health IT risks arise from interactions among hardware, software, clinical content, the user interface, people and workflow, not from any one of them. A meter and a record that each work correctly can still produce an unsafe result together. Qian and Siau (2025), reviewing connected health devices, note that continuous data from smart sensors can support decision-making but bring challenges of their own in reliability, interoperability and security.

Method: A Simulation Before Go-Live

The evaluation followed the approach Manning et al. (2025) used for infusion pump integration at a pediatric hospital, where experts from six units used real-world scenarios in simulation to test integration before implementation. Their pumps met technical integration requirements, but the simulation revealed problems with workflow efficiency, the timing of communication between systems and the limits of automation.

Here, nine nurses, three each from a medical-surgical, a cardiac and an oncology unit, worked through eight scripted scenarios over two days in the simulation center, using the new meters connected to a copy of the production record. Scenarios included a routine pre-meal check, a check on a patient with two armbands from a transfer, a low reading below 70, a reading during a Wi-Fi dead spot, a patient with a similar name in the next bed and a quality control lockout. An informatics nurse timed each step, and a pharmacist reviewed every calculated dose. Nurses completed a short usability questionnaire after each day.

Findings

Technical performance. All 144 readings in the scenarios reached the record, and every calculated correction dose matched the scale. Patient identification by armband scan worked in every case where the scan was done.

Risk one: timing. Median time from reading to display in the record was 9 seconds, but in the scenario simulating a weak signal, results took up to 90 seconds, and the meter showed "sent" before the record received the value. In three of nine runs, nurses opened the medication record during that gap and saw the previous reading, from four hours earlier, with a correction dose calculated from it. Two of the three nurses began preparing that dose before the new reading appeared.

The meter said sent; the chart still showed the morning; and two nurses started drawing up the wrong dose.

Risk two: identification workarounds. In the two-armband scenario, the meter rejected the first scan. Four nurses entered the patient's record number by hand to continue, which bypasses the scan's protection. The meter allows manual entry by default.

Risk three: trust in automation. In usability comments, six of nine nurses said they would stop checking the scale themselves once the record calculated doses. That is the purpose of automation, but it means the calculated dose must be right every time, including when the reading is stale.

Usability. Nurses rated the meters easy to use, with a median of 6 on a 7-point scale, and liked not having to dock them.

What the Nurses Said

In the debrief, nurses described the stale-reading scenario as the one that worried them most, because nothing on the screen looked wrong. One cardiac nurse said that on a busy evening she would never have noticed the time stamp. Oncology nurses raised a different concern: their patients often have glucose checks at unusual times around chemotherapy and steroids, and they wanted to be sure the record would not calculate a correction from a reading taken for a different purpose. Their comments shaped two of the recommendations below and added a scenario to the repeat simulation.

What These Findings Mean

The meters work, and the integration does what it was designed to do. The risks lie at the boundary between devices, network and workflow, exactly where Manning et al. (2025) found theirs. If the hospital had gone live without simulation, the first discovery of the stale-reading problem would have been a patient receiving correction insulin for a glucose that had already changed, potentially a low.

Recommendations Before Go-Live

1. Show the time of the reading next to the dose. The medication record will display the time of the glucose used for any calculated dose, in bold, and will refuse to calculate a correction from a reading more than 30 minutes old, prompting the nurse to wait or recheck.

2. Confirm receipt on the meter. The vendor will change the meter's message from "sent" to "received by chart" only after the record acknowledges the result.

3. Restrict manual patient entry. Manual entry of a record number will be turned off, with an override available only to charge nurses, and every override reported.

4. Fix the dead spots. Facilities will survey Wi-Fi coverage on every inpatient unit before go-live, starting with the rooms identified in the simulation.

5. Train for automation limits. Training will include the stale-reading scenario, so nurses learn to check the time of the reading before giving a calculated dose.

6. Repeat the simulation after the changes, before the first unit goes live.

Measures After Go-Live

The integration will be monitored for the share of glucose readings reaching the record within 60 seconds, manual identification overrides per month, correction doses given from readings more than 30 minutes old, as found on audit, and hypoglycemia below 70 mg/dL within six hours of correction insulin, compared with the three months before go-live.

Conclusion

The wireless meters and their connection to the record met every technical requirement in simulation. The simulation also showed that a reading delayed by a weak signal could lead nurses to prepare a dose from an earlier value, that a default manual entry option invites identification workarounds and that nurses will trust calculated doses once they exist. Displaying the time of the reading, confirming receipt, restricting manual entry, fixing coverage and training for these limits should come before go-live, not after the first error.

References

Manning, A. P., Camacho, C. A., Rehm, D. M., & Skeens, M. A. (2025). Evaluating infusion pump integration in a pediatric setting using simulation: A usability and safety assessment. Clinical Simulation in Nursing, 109, 101855. https://doi.org/10.1016/j.ecns.2025.101855

Qian, Y., & Siau, K. L. (2025). Advances in IoT, AI, and sensor-based technologies for disease treatment, health promotion, successful ageing, and ageing well. Sensors, 25(19), 6207. https://doi.org/10.3390/s25196207

Sittig, D. F., & Singh, H. (2010). A new sociotechnical model for studying health information technology in complex adaptive healthcare systems. Quality and Safety in Health Care, 19(Suppl 3), i68-i74. https://doi.org/10.1136/qshc.2010.042085