| Course | HI560 Health Care Data Analytics |
|---|---|
| Unit | Unit 10 |
| Paper type | Analytics project report |
| Length | About 1,120 words, 7 pages double-spaced |
| Format | APA 7 student paper |
| School | Purdue University Global |
| Program | MS in Health Informatics |
| Updated | October 2026 |
Free sample paper for HI560 Unit 10
From Question to Routing Rule: A Unit 10 Analytics Project Report on Incomplete MRI Exams at Two Thistlewood Sites
[Student Name]
Purdue University Global
HI560: Health Care Data Analytics
Unit 10 Analytics Project Report
[Instructor Name]
[Date]
Thistlewood Health, its sites, data and every estimate are composite; the sources cited are real.
Executive Summary
When patients who flagged claustrophobia at booking are set aside, Thistlewood's two MRI sites show no meaningful difference in completion. What separates them is a group of claustrophobic patients booked at Fennimore without access to an anxiolytic, which the hospital can give and Fennimore cannot. The report recommends routing those patients to the hospital or offering a nurse-supported medication option at Fennimore, two record system repairs and a rerun in six months, and it recommends against a completion target for Fennimore staff.
The Question
The analysis answers a rewritten version of the imaging director's question. In its final form it asks whether, over two years of adult outpatient scheduling, Fennimore's share of MRIs stopped short of their protocol exceeds the hospital's once exam group and age are made comparable. Whatever technologists try at the scanner when someone panics is not recorded anywhere, so that part of the original question is left to observation.
Data and Its Quality
The analysis rests on a repaired extract, and the repairs changed the answer. The extract held 21,906 exam rows from the radiology information system joined to the image archive's series log. Keeping only adult outpatient first attempts booked without sedation left 17,840 exams across the two sites. A quality assessment using the conformance, completeness and plausibility framework of Kahn et al. (2016) found that the completion field erred in opposite directions at the two sites: the hospital coded limited studies as complete, and Fennimore left exams marked incomplete after patients finished them the same day. Completion was rebuilt from acquired series against each protocol, which cut the apparent gap from 1.80 to 1.05 points. The rebuilt status matched chart review on 199 of 200 sampled exams.
Methods
Three methods were used, each chosen for the question at its stage. The crude difference was tested with a two-proportion z test and reported as a risk difference with a 95 percent interval; with these counts, the interval methods compared by Newcombe (1998) give the same answer. Differences in case mix were handled by direct standardization to the combined population, first across nine cells of exam group and age, then across eighteen cells with scheduled slot length added (Rothman et al., 2008). Last, after the seminar correction described below, the standardized comparison was repeated among patients who did not report claustrophobia at booking, and flagged patients were described separately by site and by whether an anxiolytic was given.
Results
The results table is the spine of the report. Each line gives an estimate of Fennimore's excess in incomplete exams per 100, with its interval, and the choice that produced it.
1. Coded data, before repair: 1.80 points.
2. Repaired data, crude: 1.05 points, interval 0.46 to 1.64.
3. After direct standardization by exam group and age band: 0.70 points, 0.10 to 1.29.
4. Also standardized for slot length: 0.38 points, minus 0.24 to 1.00.
5. Exam group and age, patients not flagged at booking: 0.12 points, minus 0.47 to 0.71.
The excess falls from 1.80 to 0.12 points as each choice is made visible, and the largest single drop comes from asking where anxious patients were booked.
Flagged patients tell the rest of the story. Of 1,140 adult outpatients who reported claustrophobia at booking, 702 were booked at the hospital, where 610 received an anxiolytic and ended incomplete about 5 percent of the time. Fennimore received 438, none medicated, and 21 percent ended incomplete, close to the 20 percent among the 92 hospital patients who declined medication. The analytic file's sedation exclusion had removed the hospital's medicated patients and kept all of Fennimore's flagged patients, which made a difference in access look like a difference in performance.
Rival Explanations
Three other explanations were considered. First, technologist skill: the near-identical rates among unflagged patients make a skill difference unlikely to be large, though a small one cannot be excluded. Second, slot length: Fennimore's shorter slots account for about a third of a point in the standardized comparison, but the interval is wide and the effect may overlap with the flagged patients, who are often booked in short slots. Third, scanner design: the hospital's scanner has a wider bore, which some patients find easier. The file cannot separate the scanner from the site, since each site has one magnet, and claustrophobic reactions are known to vary with exam setup (Dewey et al., 2007). This explanation remains open.
Recommendations
The recommendations are sized to the evidence.
1. From March 1, schedulers offer the hospital slot first to anyone who answers yes to the booking question. This addresses the main source of the gap at almost no cost.
2. Evaluate a nurse-supported anxiolytic option at Fennimore, with a decision by June 30, for patients who prefer the closer site.
3. Ask the booking question at every referral clinic, since it was missing for 14 percent of hospital exams.
4. Add a limited-study status and same-day accession linking to the radiology system, so the completion field can be trusted without rebuilding.
5. Rerun every step on the first half-year of data collected after the routing change, due October 31.
The report recommends against setting a completion target for Fennimore technologists. The evidence does not show that they perform worse, and a target would invite pressure on patients to continue scans they cannot tolerate.
Limitations
Five limitations remain. Weight could not be used for adjustment. Scanner design cannot be separated from site. The booking question was asked unevenly, so some anxious hospital patients were never flagged. The reason field is unreliable, so the analysis cannot say how many incomplete exams were due to claustrophobia rather than pain or motion. And 24 months of data cannot show whether the routing change will work; only the rerun can.
Reproducibility Appendix
The appendix lets another analyst reproduce every figure. It gives the extract query and date, the analytic file version, the third, saved on January 30, 2026, the rule for rebuilding completion from series counts, the narrowing steps with counts, the analysis script with its software versions, and the data dictionary from Unit 2.
How the Earlier Units Were Revised
Earlier work was revised rather than pasted. The Unit 6 limitation about sedated exams now points to the correction it should have prompted. The Unit 3 note on the booking question, recorded there without interpretation, is now central. Unit 9's summary is reused with one sentence tightened.
References
Dewey, M., Schink, T., & Dewey, C. F. (2007). Claustrophobia during magnetic resonance imaging: Cohort study in over 55,000 patients. Journal of Magnetic Resonance Imaging, 26(5), 1322-1327. https://doi.org/10.1002/jmri.21147
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., 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), 1244. https://doi.org/10.13063/2327-9214.1244
Newcombe, R. G. (1998). Interval estimation for the difference between independent proportions: Comparison of eleven methods. Statistics in Medicine, 17(8), 873-890. https://doi.org/10.1002/(SICI)1097-0258(19980430)17:8<873::AID-SIM779>3.0.CO;2-I
Rothman, K. J., Greenland, S., & Lash, T. L. (2008). Modern epidemiology (3rd ed.). Lippincott Williams & Wilkins.
The Purdue Global HI560 Unit 10 instructions
HI560 ends by asking for one report that pulls the course's separate assignments into a single line of argument. Expect an executive summary, the question, data and quality, methods, results, alternative explanations, recommendations, limitations and an appendix that supports reproducibility. Twenty pages with appendices is a typical scope. Begin each section with its conclusion so a reviewer can follow the chain of reasoning. Revise earlier units rather than copying them, and make sure every number agrees across sections. A results table listing each estimate and the choice that produced it is a powerful way to show how the answer evolved. Recommendations should be no stronger than the evidence.
How the HI 560 Unit 10 example is put together
The report begins with an executive summary and the final form of the question. A data section explains how the repaired completion field changed the gap and cites the Kahn framework. Methods describe the two-proportion test, direct standardization and the booking-flag comparison. The results table lists five estimates from 1.80 to 0.12 points, and the highlighted sentence identifies the largest single drop. Flagged patients are described by site and medication. Rival explanations cover technologist skill, slot length and scanner bore. Five recommendations follow, with a reasoned refusal to set a staff target. Limitations, a reproducibility appendix with a suppression rule and a note on how earlier units were revised close the report.
Reading the HI560 Unit 10 grading rubric
Final analytics reports are usually graded on integration, analytic rigor, clarity of argument, quality of recommendations and reproducibility. Integration and rigor often carry the most points: graders check that earlier work is revised, consistent and correctly applied. Clarity is rewarded when each section leads with its conclusion and the results table shows how choices affected the answer. Recommendations earn credit when they match the strength of the evidence and name owners or dates. Rival explanations and limitations must be specific. A reproducibility appendix is increasingly expected at the graduate level. APA format and professional presentation complete the rubric. Reports that paste earlier units unchanged lose points, and so do reports whose executive summary promises more than the results table shows.
HI560 Unit 10 help from the desk
A frequent reason final reports are returned is that numbers differ between sections because earlier units were pasted rather than reconciled. Another is a results section that shows only the final estimate, hiding how analytic choices shaped it. Some students recommend more than the evidence supports. Rival explanations are often omitted or dismissed in a sentence. Reproducibility appendices may lack the file version or inclusion rules. If your project analyzed a different question, such as readmissions or wait times, gather your nine earlier submissions, and we will reconcile their numbers into a single report built around your own results table.
Get an HI560 Unit 10 example written to your instructions
Gather the earlier units' results, or short summaries of them, with the final prompt and rubric; note any figures that changed along the way. The first sample is free and back within 24-48h, reconciled from question to recommendation, its results table showing how each analytic choice moved the estimate. The paper above is an original model document written by our desk, not a submitted student paper and not an official Purdue University Global document.
More HI560 papers and MS in Health Informatics samples
- HI560 Unit 1: Discussion Board Post
- HI560 Unit 2: Data Set Description
- HI560 Unit 3: Descriptive Statistics Report
- HI560 Unit 4: Data Quality Assessment
- HI560 Unit 5: Hypothesis Test Writeup
- HI560 Unit 6: Population Comparison
- HI560 Unit 7: Visualization Set
- HI560 Unit 8: Seminar Reflection
- HI560 Unit 9: Executive Summary
- HI555 Unit 10: Cycle Improvement Proposal
- HI550 Unit 10: Change Management Plan
- HI545 Unit 10: Program Effectiveness Review
- HI540 Unit 10: Project Closeout Report
HI560 Unit 10 questions, answered
How much of the earlier units' work should the final report reuse?
Most of the substance, revised. Condense each unit to what the argument needs, correct anything later units changed, and check that every number agrees across sections. Graders often read the final report beside earlier submissions, and a figure that silently differs, or a limitation that was resolved but still listed, suggests the pieces were never reconciled.
What belongs in a reproducibility appendix?
The data source and version, extraction date, the exact inclusion rules, any repairs made to the file, the script or step-by-step procedure, and software versions. A reader with the same file should reach the same numbers. A suppression rule for small cells also belongs here when results may circulate beyond the analytics team.
Can the final report conclude that the question has no clear answer?
Yes, if the evidence supports that, and it should still recommend something: further data, a targeted review, a repeat analysis. Concluding that the evidence cannot yet support action is itself a finding. What graders penalize is a conclusion that avoids committing to anything, or one that claims more certainty than the intervals allow.
What results table does the HI560 final report use?
One line per estimate with its interval and the choice that produced it, so readers watch the excess fall from 1.80 points in coded data to 0.12 among unflagged patients.
Where can I find a free HI560 Unit 10 sample paper?
This page holds the full Unit 10 report, from the question and repaired data to a five-step results table, recommendations and a reproducibility appendix.