NR 542 Week 8 Data Management Paper Example

Data Management Paper: One Figure, 3.0% of Medicine Admissions Moving to the ICU Within a Day of Reaching the Floor, Built by Rules Written in Advance and Meaning Exactly What Its Author Says It Means, Nothing Beyond

Student Name

College of Nursing, Chamberlain University

NR 542: Managing Data and Information

Instructor Name

Month Day, Year

Data Management Paper: One Figure, 3.0% of Medicine Admissions Moving to the ICU Within a Day of Reaching the Floor, Built by Rules Written in Advance and Meaning Exactly What Its Author Says It Means, Nothing Beyond

The Claim

The claim of this paper is narrow on purpose. One figure, built by rules written down in advance, means what its author says it means and nothing beyond that. The figure is the rate of first-day moves from ward to intensive care for emergency admissions to the hospital medicine service in our 400-bed hospital: 3.0%, give or take about half a percentage point, for the 12 months from July to June. For more than a year before this course, a different figure, 8%, was quoted at quality meetings to argue that the emergency department was placing patients on floors when they should have gone to intensive care. That figure had no written definition. This paper describes how the new figure was built, what it means and what it must not be used to claim.

From a Quoted Number to a Question

The first step was to take the quoted figure apart. It left undefined who was counted, when the 24-hour clock started, which transfers counted and over what period. The second step was to translate the hospital medicine director's worry, whether patients were being placed on floors when they should have gone to intensive care, into questions the record could answer. Two could be answered: how often early transfers happen, and how that compares with the past and with published rates. One could not: whether each admission decision was reasonable at the time. Keeping the unanswerable question visible was essential. Hersh et al. (2013) warn that data produced for care and billing invite exactly this sort of overreach when used to judge clinical decisions.

Rules Before Results

The definition was written before any data were pulled: adult emergency admissions under the medicine service when they first reached a ward, with time zero at floor arrival rather than the admission order, excluding comfort care and planned transfers. The director confirmed each choice before the query ran. The reporting plan turned the definition into a numbered build sequence, each step stored as a saved query with its rule written in words, and added reconciliation checks, a version number and a change log. The rule that a rerun must reproduce the published figure exactly is the practical test of whether the rules really live in the document.

The confirmation step changed one choice. The director first wanted time zero at the admission order, because that is when the decision is made. After seeing that boarding times varied from one to fourteen hours, the director agreed that floor arrival better matched the question, which was about what happens once a patient is on the ward. Had the data been pulled first, that choice might have been made by whichever result looked more persuasive.

Structure and Quality

Two departments had reached different figures, 8% and 5%, because they read different tables: quality used the encounter table, with the admission order as time zero and the discharging attending as the team, while hospital medicine used a morning census that missed overnight transfers. Neither made an arithmetic error. The database structure write-up mapped each element of the definition to a table and column so that this could not recur silently.

The data quality audit then checked 200 encounters against the chart. It found a preselected admission source left unchanged in 5.5% of encounters, a unit coded as medical-surgical fourteen months after it became step-down, lags between real and recorded bed movement, duplicate encounters from before a system upgrade and missing team assignments. Correcting them moved the figure from 3.4% to 3.0%. Research on data quality in electronic health records describes dimensions such as completeness, correctness and currency (Weiskopf & Weng, 2013), and every problem found fell under one of them.

What the Figure Means

The figure means that roughly three of every hundred adults sent from the emergency department to a medicine ward bed end up in intensive care during their first day on the ward, for reasons other than a scheduled procedure. It is well below the figure that was quoted, and in the same range as the published experience of a large integrated system (Delgado et al., 2013). It is higher for night arrivals, and close to four in ten of the patients who moved had shown raised warning scores before they left the emergency department.

What the Figure Does Not Mean

The figure does not mean that 97% of admission decisions were right or that 3% were wrong. Some transferred patients deteriorated unpredictably; some patients who stayed on the floor might have been better placed in intensive care but did well anyway. It does not include step-down transfers, which would raise the figure to 4.9% if counted. It does not show whether the difference between night and day reflects patients, staffing or chance. And it does not include the 21 patients whose floor arrival could not be established. Each of these limits is stated in the report itself, because a figure separated from its limits is quickly used to support claims it cannot bear.

What Comes Next

Two developments will extend what the figure can show. First, three structured questions on the intensive care transfer order will capture the reason for transfer, whether the condition was new and whether intensive care was discussed before floor admission, with no preselected answers and a pilot before use. Kahn et al. (2016) advise defining new data elements, their allowed values and their expected completeness before collection , and those definitions have been written. Second, clinicians will review the 45 transferred patients with elevated warning scores, because only case review can address the question of judgment.

The broader lesson of the course is modest. Most bad figures in health care are not the result of bad arithmetic but of undefined terms, unread tables and untested fields. A figure earns trust not by being impressive but by being reproducible, bounded and honestly described. The new figure is less dramatic than the old one. It is also the first that anyone can check.

The new figure is not the end of the work. It will be reported monthly with its limits attached, and it will be retired or redefined if the hospital's units or record system change in ways the rules cannot absorb.

References

Delgado, M. K., Liu, V., Pines, J. M., Kipnis, P., Gardner, M. N., & Escobar, G. J. (2013). Risk factors for unplanned transfer to intensive care within 24 hours of admission from the emergency department in an integrated healthcare system. Journal of Hospital Medicine, 8(1), 13-19. https://doi.org/10.1002/jhm.1979

Hersh, W. R., Weiner, M. G., Embi, P. J., Logan, J. R., Payne, P. R., Bernstam, E. V., Lehmann, H. P., Hripcsak, G., Hartzog, T. H., Cimino, J. J., & Saltz, J. H. (2013). Caveats for the use of operational electronic health record data in comparative effectiveness research. Medical Care, 51(8, Suppl. 3), S30-S37. https://doi.org/10.1097/MLR.0b013e31829b1dbd

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 18. https://doi.org/10.13063/2327-9214.1244

Weiskopf, N. G., & Weng, C. (2013). Methods and dimensions of electronic health record data quality assessment: Enabling reuse for clinical research. Journal of the American Medical Informatics Association, 20(1), 144-151. https://doi.org/10.1136/amiajnl-2011-000681