HIM 690 Module 9 Final Project Example

Final Project: Present on Arrival, or Not? A Capstone Study of Admission Flag Accuracy for Four Hospital Complications at Cimarron Heights Medical Center

[Student Name]

Southern New Hampshire University

HIM 690: Health Information Management Capstone

Final Project

[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: Present on Arrival, or Not? A Capstone Study of Admission Flag Accuracy for Four Hospital Complications at Cimarron Heights Medical Center

Executive Summary

Admission flags on inpatient claims decide whether a wound, infection, clot or fall injury counts as harm that occurred in the hospital. This capstone re-read 317 charts at Cimarron Heights, a composite academic hospital in Tulsa, and compared each coded flag with a blinded second reading. Flags agreed with the chart 76% of the time overall, close to earlier research, but only 61% of the time for pressure wounds, where errors overwhelmingly labeled hospital-acquired wounds as pre-existing. A missing first-day skin and risk check was the strongest predictor of error, and the coders interviewed described settling on yes when they could neither see nursing findings nor ask nurses. Five recommendations follow, and the hospital's reported pressure wound rate should be expected to rise as the data improve.

Problem and Questions

An internal audit had found many coded pressure wounds flagged as present on admission with nothing in the first day's notes to support it. The study asked how often flags agreed with an independent reading for four complications, which way disagreements ran and whether a missing first-day check was associated with disagreement after accounting for other factors. Pine et al. (2007) showed why the flag matters for measurement: adding it to claims data substantially improved risk adjustment, but only if the flag is right.

Methods as Carried Out

The study used an explanatory sequential mixed methods design, a record review followed by interviews, approved by the hospital's review board with a HIPAA waiver. A stratified random sample of 318 fiscal 2026 discharges was drawn; one record lacking transfer documents was excluded, leaving 317, with 77 rather than 80 fall records because that stratum had no more eligible charts. Two credentialed coders abstracted every chart blind to the coded flag. Every tenth record was read twice, and whole-study kappa was 0.85, within the strong range McHugh (2012) describes. One decision rule, on wounds noted at arrival but staged later, was clarified mid-study and applied to all earlier records, changing three. The logistic model had about eleven events per predictor term, meeting the guidance of Peduzzi et al. (1996). Eight interviews were analyzed with thematic analysis (Braun & Clarke, 2006).

Results

Agreement was 242 of 317 records, 76.3%, or about 77% weighted to the hospital's year, compared with 74% in a large California audit (Goldman et al., 2011). By complication, agreement was 61.3% for pressure wounds, 83.8% for urinary infections linked to catheters, 82.5% for venous clots and 77.9% for fall injuries. Of 75 disagreements, 44 were over-reports and 31 under-reports, a split that could be chance overall but not for wounds, where 26 of 31 were over-reports. In the model, a missing first-day check carried adjusted odds of disagreement of 3.4, pressure wounds 2.6 compared with urinary infections and facility admission 1.9; coder experience showed no association. A separate payment review found seven claims that may have been overpaid, referred to compliance.

Integrated Findings

Interviews explained the numbers. Nurses described deferring skin inspection during busy night admissions. Coders described choosing yes for facility patients with any skin mention, because a no would imply the hospital caused harm, and they had no way to query nurses and seldom opened the nursing flowsheet where wound findings sit. Read together, as Fetters et al. (2013) recommend for mixed methods, the evidence supports two explanations for wound over-reporting, missing documentation and a coding default, and weakens a third, late-appearing deep tissue injury, which accounted for only four of 26 wound over-reports.

Discussion

Overall accuracy at Cimarron Heights resembles what earlier research found, but the direction of its wound errors does not. In the California audit, teaching hospitals leaned toward the opposite error, too few conditions marked present on arrival (Goldman et al., 2011); this teaching hospital marks too many pressure wounds that way. The interviews suggest a reason specific to wounds: of the four complications, wounds are the one coders feel least able to judge and the one most often documented by nurses rather than physicians. The practical result is that fewer wounds are counted as the hospital's than truly were, which matters because, as Bahl et al. (2008) showed, these flags change the safety rates hospitals report and are judged by.

Limitations

The study has clear limits. It covers one hospital and one year. Its gold standard comes from the same record as the coded flag, so a condition never documented anywhere cannot be detected, which likely understates under-reporting. The design is cross-sectional, so the link between missing checks and wrong flags is an association supported by interviews, not proof of cause. Eight interviews cannot represent every coder and nurse. The payment review covered only the sample and was not extrapolated.

Recommendations

Five recommendations follow from the findings: make the first-day skin and Braden check a required step before the nursing admission note can be signed; let coders send structured questions to nurses; replace the yes default with a policy that flags must reflect documentation; show first-day skin findings inside the coding workspace; and repeat the study's audit every quarter with the same rules and reliability checks. Each has an owner, a cost mostly in staff time and a measure, with a target of at least 85% wound flag agreement within a year. Leaders should be told in advance that the reported wound rate may rise as accuracy improves.

Competencies Demonstrated

The project shows data quality and governance through the abstraction tool, rules and reliability results; research and analytics through the sampling, model and integration; privacy and compliance through the review board approval, waiver, handling log and compliance referral; finance through the payment review; and leadership through the presentation to the quality committee and the recommendations now being taken up by nursing and coding.

Conclusion

A one-character field on a claim decides whether harm is counted. At Cimarron Heights it was wrong in two of five pressure wound records, mainly because no one looked at the skin on the first day and because coders, unable to see or ask, assumed the best. The fixes are modest and the measurement method to prove them is already built.

References

Bahl, V., Thompson, M. A., Kau, T.-Y., Hu, H. M., & Campbell, D. A. (2008). Do the AHRQ patient safety indicators flag conditions that are present at the time of hospital admission? Medical Care, 46(5), 516-522. https://doi.org/10.1097/MLR.0b013e31815f537f

Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77-101. https://doi.org/10.1191/1478088706qp063oa

Fetters, M. D., Curry, L. A., & Creswell, J. W. (2013). Achieving integration in mixed methods designs: Principles and practices. Health Services Research, 48(6pt2), 2134-2156. https://doi.org/10.1111/1475-6773.12117

Goldman, L. E., Chu, P. W., Osmond, D., & Bindman, A. (2011). The accuracy of present-on-admission reporting in administrative data. Health Services Research, 46(6pt1), 1946-1962. https://doi.org/10.1111/j.1475-6773.2011.01300.x

McHugh, M. L. (2012). Interrater reliability: The kappa statistic. Biochemia Medica, 22(3), 276-282. https://doi.org/10.11613/BM.2012.031

Peduzzi, P., Concato, J., Kemper, E., Holford, T. R., & Feinstein, A. R. (1996). A simulation study of the number of events per variable in logistic regression analysis. Journal of Clinical Epidemiology, 49(12), 1373-1379. https://doi.org/10.1016/S0895-4356(96)00236-3

Pine, M., Jordan, H. S., Elixhauser, A., Fry, D. E., Hoaglin, D. C., Jones, B., Meimban, R., Warner, D., & Gonzales, J. (2007). Enhancement of claims data to improve risk adjustment of hospital mortality. JAMA, 297(1), 71-76. https://doi.org/10.1001/jama.297.1.71