HI560 Unit 4 Data Quality Assessment Example

Course	HI560 Health Care Data Analytics
Unit	Unit 4
Paper type	Data quality assessment
Length	About 880 words, 6 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 4

Fit for One Use: A Unit 4 Data Quality Assessment of an MRI Completion Field That Erred in Opposite Directions at Two Sites

[Student Name]

Purdue University Global

HI560: Health Care Data Analytics

Unit 4 Data Quality Assessment

[Instructor Name]

[Date]

Thistlewood Health, its sites, the samples and every count are composite; the sources cited are real.

What this part is doingJudging quality against one named use, rather than in general, lets every finding be stated as a direction of bias the comparison will inherit.

The Use and the Fields That Decide It

The file will be used to compare the proportion of adult outpatient MRI exams that end incomplete at the two sites, adjusted for exam group and age. Five fields decide who enters the comparison and how it comes out: completion status, patient class, age, sedation ordered and exam group. Quality is judged against that use, following the view that data are of good quality when they suit the purpose at hand (Wang & Strong, 1996).

Framework

The checks are organized by the shared vocabulary that Kahn et al. (2016) proposed for secondary use of record data: conformance, meaning a value matches the format and coding the system promises; completeness, meaning values are present; and plausibility, meaning values are believable against other data or what is known. Automated checks ran on all 21,906 rows; a validation sample then compared the analytic file with the clinical record.

Automated Checks

1. Conformance: 112 accession numbers appeared twice, created when a technologist changed the protocol mid-exam; 230 rows lacked a scanner identifier.

2. Completeness: technologist reason blank on 22 percent of hospital and 9 percent of Fennimore incomplete exams; weight missing on 9 and 17 percent of rows.

3. Plausibility: 41 exams had a scan end time before the start time, all from one scanner's clock after a daylight saving change; 18 exams showed table times over three hours, all confirmed as real long cardiac or multi-region studies.

The Completion Field

The central finding concerns completion status. The radiology system lets a technologist mark an exam complete or incomplete, but the image archive shows which series were actually acquired, and the protocol lists which series are required. Comparing the three revealed two different errors.

At the hospital, 41 exams were coded complete although fewer series were acquired than the protocol requires and the technologist's note recorded that the patient asked to stop. In each case the radiologist had accepted a limited study, and the system has no limited status, so complete was used. Thirteen exams went the other way, coded incomplete with every required series present.

At Fennimore, 49 exams were coded incomplete although the patient finished the protocol later the same day in a second slot under a new accession number. The first accession was never updated. Twelve exams were coded complete with required series missing.

The hospital's error hides real incomplete exams and Fennimore's error invents them, so both push the apparent gap wider than it is.

What this part is doingStating the direction of each error, not just its count, is the move that tells the reader what the uncorrected file would have got wrong.

Validation Sample

One hundred exams per site were drawn at random from the analytic file and checked against the image archive and the technologist's note, applying one rule: an exam is complete if every required series was acquired during the visit, on the same day, across any accessions. The rebuilt status matched chart review on 99 of 100 hospital exams and all 100 Fennimore exams. The one disagreement was an exam in which the radiologist judged one required series unnecessary after viewing the others; the rule counts it as incomplete, which is conservative.

Reason Codes

The technologist's reason is not fit for analysis. At Fennimore, technologists choose claustrophobia whenever a patient asks to stop; at the hospital, many of the same events are coded motion, because the patient moved before asking. Reasons will be reported descriptively only.

Timeliness

The extract was pulled on January 15. At that point 1.2 percent of December exams still showed a status of in progress, because the technologist had not closed them. These were resolved from the image archive before analysis.

Effect of the Repairs

Before repair, the analytic file showed 331 incomplete hospital exams (3.36 percent) and 412 at Fennimore (5.16 percent), a gap of 1.80 points. After rebuilding completion from the image archive, collapsing duplicates and resolving open December exams, the counts are 359 (3.64 percent) and 375 (4.69 percent), a gap of 1.05 points. The repairs remove about four tenths of the apparent difference.

1. Hospital: 331 coded incomplete, plus 41 limited studies, minus 13 coded incomplete in error, equals 359.

2. Fennimore: 412 coded incomplete, minus 49 finished the same day, plus 12 coded complete in error, equals 375.

Verdict

The file is fit for the comparison after three repairs: completion rebuilt from acquired series against the protocol, limited studies counted as incomplete, and same-day completions credited as complete. Exam group, age, patient class and sedation passed every check. Weight should not be used for adjustment. Reviews of record quality work note that most assessments stop at completeness, while the more damaging problems, such as a field whose meaning differs by site, are found only by comparing against a second source (Weiskopf & Weng, 2013), which is what the image archive provided here. Two system changes would prevent the problems recurring: a limited study status in the radiology system, and a rule that links a same-day second slot to the first accession.

What this part is doingA verdict that names the conditions under which the data become fit gives the analysis a clear starting line instead of a vague caution.

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., 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

Wang, R. Y., & Strong, D. M. (1996). Beyond accuracy: What data quality means to data consumers. Journal of Management Information Systems, 12(4), 5-33. https://doi.org/10.1080/07421222.1996.11518099

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

The Purdue Global HI560 Unit 4 instructions

Unit 4 commonly asks you to assess the quality of your data set for the analysis you plan. Expect to name a framework, run checks, validate key fields against an independent source, explain the likely effect of each problem and give a verdict. Sections often ask for around six pages built around one table of findings. Judge quality against your specific use, starting with the fields that decide who enters the analysis and how groups compare. For each problem, state the direction it would push your result, not only its size. Describe any repair and show its effect on the numbers while keeping the original file. Many instructors expect recommendations that would prevent the problem in the source system.

Inside the HI560 Unit 4 example

The assessment names its use and the five fields that decide it, then introduces the Kahn framework. Automated checks are listed under conformance, completeness and plausibility. The central section compares the completion field with the image archive and protocol, showing 41 hospital limited studies coded complete and 49 Fennimore exams left incomplete after same-day completion, and the highlighted sentence explains that both errors widen the gap. A validation sample of 100 exams per site confirms the rebuilt status. Short sections judge the reason field unfit and resolve open December exams. An effects section shows the arithmetic for each site, and the verdict sets three repairs and two system changes, citing a review of record quality methods.

Reading the HI560 Unit 4 grading rubric

Rubrics for this assessment generally spread points over the framework, how rigorous the checks are, how effects are interpreted, the verdict and the writing. Interpretation often carries the most points, rewarded when the paper explains how each problem would bias the planned analysis. Rigor is judged on whether automated checks are paired with validation against an independent source. The framework earns credit when it organizes the work rather than appearing as a definition. A clear verdict with conditions is expected. Repairs should be documented with before and after figures. Tables, organization and APA citations complete the rubric. Assessments that list errors but never connect them to the planned comparison tend to lose points, and so do repairs reported without before and after counts.

HI560 Unit 4 help from the desk

The most common shortfall is reporting problems without their direction of bias. Another is relying on automated checks alone, which find impossible values but not wrong ones. Some students fix data silently, so later readers cannot tell what changed. Frameworks are often defined in full and then not used. A vague verdict, such as the data are mostly good, leaves the analysis without a starting line. If your file has different weak points, perhaps diagnosis codes or timestamps, tell us which fields decide your analysis and which second source you can reach, and we will draft an assessment around that pairing. Mention the planned comparison too, since every finding is judged against it.

Get an HI560 Unit 4 example written to your instructions

Share the data set the Unit 4 prompt supplies, and any second source it offers for comparison, such as a registry or abstract. Add the rubric. The first custom sample, free and returned within 24-48h, judges that file against the question your course poses and says which way each error would push the result. 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 4 questions, answered

What is a validation sample, and do I need one?

A set of records checked against an independent source, such as the chart, a registry or a certificate worksheet, to see whether plausible values are also correct. Automated checks find impossible values; only comparison finds wrong ones. If the prompt supplies a second source, use it. If not, say what source would be needed and why it matters for the question.

Which data quality framework fits health data analytics?

Kahn and colleagues' 2016 harmonized framework, built for secondary use of electronic record data, sorts checks into conformance, completeness and plausibility, which suits analytics well. AHIMA's data quality characteristics are common in HIM courses too. Whichever the course names, the framework should organize checks on the fields that matter for the question rather than generate a definition for every dimension.

Where can I find a free HI560 Unit 4 sample paper?

This page contains the complete Unit 4 assessment, including the opposite errors in the completion field, the validation sample and the verdict.

Why does the direction of a data error matter?

Because two errors that both push a gap wider can create a difference that is not real. Knowing the direction tells you how the uncorrected file would mislead.

What is a validation sample in a data quality assessment?

A set of records checked against an independent source, such as images, charts or registry entries, to see whether values that look plausible are actually correct.