HIM 550 Module 7 Final Project Milestone Three Example

Final Project Milestone Three: Fixing the Data Where It Is Made. Recommendations for Discharge Data Management

[Student Name]

Southern New Hampshire University

HIM 550: Data Management and Data Quality

Final Project Milestone Three

[Instructor Name]

[Date]

The organization, setting and figures below are a composite written as a model document. No real employer, client, colleague or patient is described.

Final Project Milestone Three: Fixing the Data Where It Is Made. Recommendations for Discharge Data Management

The first two milestones of this project showed that Cypress Hollow Medical Center's discharge data could answer leaders' questions only after substantial correction, and that the uncorrected data would have led them to the wrong conclusions about discharge timing. Correcting an extract once is not enough. The same problems will return every night unless the hospital changes how the data are created, loaded, owned and monitored. This milestone recommends six changes, each linked to a specific finding, and closes with an implementation table and the risks to watch.

Principles Behind the Recommendations

Three principles guide the choices. First, problems should be fixed where they begin, in registration screens, nursing workflows and load scripts, rather than by cleaning reports. Second, every data element that leaders use should have a named owner who decides what it means. Third, quality should be measured continuously, because a one-time assessment shows only a single moment. Rosenbaum (2010) argued that data stewardship depends on defined authority, transparent rules and accountability, and each recommendation below assigns all three.

Recommendation 1: Build the Measure on Departure Time

The discharge-before-noon measure should be rebuilt on the nursing departure time, which records when the patient leaves, rather than the registration time, which records when a clerk closes the encounter. Nursing leadership should own the departure time definition, and the business definition should appear in the data dictionary with its source field. The dashboard should show the departure-based rate from the next reporting month, with a note explaining why the number changed from about 29% to about 37%. Because 9% of departure times were missing, the nursing documentation should make the field required before a discharge can be completed.

Recommendation 2: Require Admission Source

Admission source should become a required field at registration, with a short list of values and brief help text. Because the blanks clustered on nights and weekends, the change should be paired with a review of registrar staffing and screen design during those hours rather than blaming staff. In the warehouse, blanks already present should be loaded as Missing, not Other, so reports show the gap honestly. The target is fewer than 2% missing within three months.

Recommendation 3: Keep Unit History

The unit dimension should be rebuilt so that each unit keeps a dated history, allowing discharges to remain linked to the unit as it existed at the time. The crosswalk created for the quality assessment supplies the starting history. Future reorganizations should trigger a documented update approved by the data governance committee before the change takes effect in any system. The same dated design should later extend to service lines and physician groups, which also change over time. Without it, every year-over-year comparison by unit carries a hidden error that analysts must explain away.

Recommendation 4: Audit Every Load

The nightly load should record rows extracted and rows loaded for each table and flag any difference or any count outside the expected range. Alerts should go to the analytics team before 6 a.m., and reports should display a notice when the previous night's load is incomplete. This prevents partial days from reaching leaders unnoticed, as happened in April. The audit should also keep a simple log of each night's counts, run time and any failures so analysts can check a date before trusting it, and a monthly summary should go to the governance committee so it can see whether loads are becoming less reliable.

Recommendation 5: Name Stewards for Discharge Data

The discharge data domain should have a data owner, the vice president for patient flow, who is accountable for its definitions, and data stewards in patient access, nursing and health information who each maintain the fields they create. Stewards should meet monthly with analytics to review quality measures and approve definition changes. The hospital's data governance committee, already chartered, should confirm these roles and resolve disagreements between departments.

Recommendation 6: Monitor Quality with Control Charts

Kahn et al. (2016) describe data quality checks for conformance, completeness and plausibility that can run routinely rather than once. Three checks should run every week: missing admission source, missing departure time and departure times earlier than admission. Plotting each on a control chart, as described by Mohammed et al. (2008), will let stewards tell ordinary variation from real deterioration and respond only when a signal appears. Each chart should list the steward who investigates a signal and the steps to take, so that a spike in missing admission sources on a holiday weekend leads to a conversation with registrars rather than a quiet correction in the warehouse.

Implementation Plan

Table 1 summarizes each recommendation with its owner, timeline and measure of success.

Table 1. Recommendations, Owners, Timelines and Measures

Note. Prepared by the author for the final data management plan.

Risks and Resistance

Two risks stand out. Leaders may resist a dashboard that jumps from 29% to 37% because it looks like a reporting trick; explaining the change in person, with both numbers side by side for a month, should build trust. Registrars may see a new required field as another burden during busy nights, so patient access should involve them in designing the screen. Neither risk is technical, which is why ownership and communication are part of every recommendation.

Conclusion

Each recommendation answers a finding from the earlier milestones, and together they move the hospital from correcting data after the fact to creating reliable data in the first place. The final project will combine these recommendations with the quality assessment and analysis into a complete data management plan.

References

Kahn, M. G., Callahan, T. J., Barnard, J., Bauck, A. E., Brown, J., Davidson, B. N., Estiri, H., Goerg, C., Holve, E., Johnson, S. G., Liaw, S.-T., Hamilton-Lopez, M., Meeker, D., Ong, T. C., Ryan, P., Shang, N., Weiskopf, N. G., Weng, C., Zozus, M. N., & Schilling, L. (2016). A harmonized data quality assessment terminology and framework for the secondary use of electronic health record data. eGEMs, 4(1), Article 1244. https://doi.org/10.13063/2327-9214.1244

Mohammed, M. A., Worthington, P., & Woodall, W. H. (2008). Plotting basic control charts: Tutorial notes for healthcare practitioners. Quality and Safety in Health Care, 17(2), 137-145. https://doi.org/10.1136/qshc.2004.012047

Rosenbaum, S. (2010). Data governance and stewardship: Designing data stewardship entities and advancing data access. Health Services Research, 45(5, Pt. 2), 1442-1455. https://doi.org/10.1111/j.1475-6773.2010.01140.x