HIM 220 Module 2 Data Quality Short Paper Example

Measuring What We Complain About: A Data Quality Assessment at Kettle River Health

[Student Name]

Southern New Hampshire University

HIM 220: Healthcare Data Management

Module Two Short Paper

[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.

Measuring What We Complain About: A Data Quality Assessment at Kettle River Health

Everyone at Kettle River Health has a story about bad data: a patient listed twice, a report that does not match another, a lab value that seems impossible. Stories do not tell leaders which problems are largest or which to fix first. This paper uses recognized data quality frameworks to define what good data means, measures five known problems and ranks them for action.

Quality Depends on Use

Data quality is not absolute. The same data can be adequate for billing and inadequate for research or equity reporting. Kahn et al. (2016) built their harmonized framework around fitness for use, asking whether data are good enough for a specific purpose. A missing race field does not affect a claim, but it makes it impossible to tell whether readmissions differ by race. Every assessment below therefore names the use at stake.

Five Dimensions

Weiskopf and Weng (2013) reviewed studies assessing electronic health record data quality and identified five recurring dimensions. Completeness asks whether a value is present. Correctness asks whether it is true. Concordance asks whether different sources agree. Plausibility asks whether a value makes sense, such as a birth date that precedes an admission. Currency asks whether data are up to date. They also described assessment methods, from comparing records with a gold standard to checking distributions for implausible patterns.

Conformance, Completeness and Plausibility

Kahn et al. (2016) sorted quality checks into three families. The first, conformance, tests whether values obey the required formats, allowed values and relationships between tables. Completeness checks for gaps. Plausibility tests whether a value could be true, judged alone and against the values recorded before and after it. They distinguished verification, checking data against internal expectations such as allowed values, from validation, checking data against an external source such as chart review. Most routine checks are verification; validation is slower but reveals errors that look plausible.

Problem 1: Missing Race and Ethnicity

A query of all patients seen in the past year found race or ethnicity missing for 19% and recorded as declined for another 7%. Missingness varied by where patients registered: 31% in the emergency department compared with 12% in clinics. This is a completeness problem that threatens equity reporting. It is common. Polubriaginof et al. (2019) found race or ethnicity unknown for a quarter of patients in large national databases and for most patients at one New York health system, yet 86% of patients gave meaningful answers when they recorded the information themselves.

Problem 2: Suspected Duplicate Records

The master patient index software flagged 6.4% of records as possible duplicates, meaning two records that likely belong to one person. A manual review of 200 flagged pairs confirmed 71% as true duplicates. This is a uniqueness problem, part of plausibility in Kahn's framework, and it matters for safety because allergies or results in one record may be missing from the other. Most duplicates arose at registration, when staff created new records rather than finding existing ones under a misspelled name or changed address.

Problem 3: Impossible Timestamps

A simple verification check found 212 inpatient encounters last year whose discharge time preceded the admission time, and 38 with lengths of stay over 400 days. These plausibility failures came mostly from downtime periods, when paper records were entered later with estimated times. They distort length-of-stay statistics and readmission windows, which depend on accurate dates and times.

Problem 4: Outdated Problem Lists

For patients who saw a primary care clinician during the last twelve months, only 58% had a problem list reviewed or updated during that time. Many lists still showed resolved conditions, such as a treated pneumonia from years before. This is a currency and correctness problem that affects care coordination and any analysis that uses the problem list to identify chronic conditions, such as a registry of patients with diabetes.

Problem 5: Unmapped Laboratory Results

About 12% of results from the system's outside reference laboratory arrive with local test names but no standard LOINC code. McDonald et al. (2003) described LOINC as a universal code system for laboratory observations that allows results from different laboratories to be recognized as the same test. Without it, a hemoglobin A1c from the reference laboratory may not appear in the diabetes dashboard. This is a conformance problem.

Ranking the Problems

Table 1 summarizes the five problems with their dimension, measurement, affected uses and a priority based on risk and reach. Duplicate records rank highest because they create direct patient safety risks. Missing race and ethnicity rank next because they block equity analysis the board has requested.

Table 1. Data Quality Assessment Summary

Note. Measurements from the author's queries and reviews of system data.

Common Causes

Most problems begin at the point of capture. Registration workflows allow new records without an adequate search, ask about race and ethnicity inconsistently and lack standard wording. Downtime procedures let estimated times enter the record. Interfaces accept unmapped lab codes without alerting anyone. Fixing downstream reports cannot solve problems created upstream.

Who Should Own Each Problem

Measurement alone does not fix anything; each problem needs an owner. Duplicate records belong with the HIM department's master patient index team and the patient access manager, since registration creates most of them. Race and ethnicity collection belongs with patient access and the equity office. Laboratory mapping belongs with the laboratory information system analyst. Problem list currency belongs with primary care leadership, and downtime timestamps with the nursing informatics team. The governance paper later in the course will formalize these roles as data stewards.

Conclusion

Named dimensions and simple measurements turned vague complaints into five defined problems with sizes, causes and priorities. The assessment shows that data quality depends on use, that verification catches many problems cheaply and that validation is needed for others. The next module examines the databases and standards that shape how data are stored and exchanged.

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

McDonald, C. J., Huff, S. M., Suico, J. G., Hill, G., Leavelle, D., Aller, R., Forrey, A., Mercer, K., DeMoor, G., Hook, J., Williams, W., Case, J., & Maloney, P. (2003). LOINC, a universal standard for identifying laboratory observations: A 5-year update. Clinical Chemistry, 49(4), 624-633. https://doi.org/10.1373/49.4.624

Polubriaginof, F. C. G., Ryan, P., Salmasian, H., Shapiro, A. W., Perotte, A., Safford, M. M., Hripcsak, G., Smith, S., Tatonetti, N. P., & Vawdrey, D. K. (2019). Challenges with quality of race and ethnicity data in observational databases. Journal of the American Medical Informatics Association, 26(8-9), 730-736. https://doi.org/10.1093/jamia/ocz113

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