CHDA Preparation: Healthcare Data Decisions and Practice Study Guide

Apply it

A hospital wants to study whether laboratory results available before discharge predict later utilization. The analyst needs timestamped clinical results and a defensible follow-up source, rather than assuming billing records contain the required laboratory values.

Apply it

A service adds appointment slots, improves documented follow-up completion, and tracks subsequent emergency visits. These represent structure, process, and outcome respectively; the analyst reports each separately rather than treating increased capacity as proof of improved health.

Apply it

Two systems both display glucose, but one records a blood measurement and another a urine measurement. The analyst checks observation attributes before mapping them, while using RxNorm separately to reconcile medication names across systems.

Apply it

Among 800 initially disease-free people followed completely for one year, 40 develop the condition. The one-year incidence proportion is 40 divided by 800, or 5%; existing cases would belong in a prevalence calculation instead.

Apply it

A manager reports rising denials using every submitted transaction as the denominator. The analyst separates original claims from resubmissions and distinguishes rejected submissions from adjudicated denials before assessing whether the underlying denial rate actually increased.

Apply it

For a hospital using Joint Commission accreditation to support deemed status, the analyst links applicable standards to evidence such as policies, clinical records, survey findings, and corrective actions. CMS remains the governmental authority; the accreditor evaluates compliance through its approved program. For Hospital IQR, the analyst instead builds the required measure submissions using the applicable reporting-year specifications, validates eligibility and calculation rules, and retains submission acknowledgments and completeness checks. CMS's [Hospital IQR measure guidance](https://qualitynet.cms.gov/inpatient/iqr/measures) identifies program-specific measures and reporting requirements. An accreditation decision and an accepted quality-data submission therefore answer different compliance questions and require separate evidence.

Apply it

A team measures turnaround from an order to a result but accidentally uses the interface receipt timestamp. The analyst checks field lineage and workflow timing, then selects timestamps that represent the intended start and completion events.

Apply it

A dashboard counts missed appointments, a model estimates future missed visits, and a scheduling optimization compares reminder allocations under staffing limits. The analyst distinguishes description, prediction, and prescription before choosing evaluation criteria for each proposed tool.

Apply it

Overall follow-up completion rises from 70% to 76%, but a language subgroup remains at 52%. The analyst checks subgroup denominators and missing language data, then presents the persistent gap alongside the overall improvement to leadership.

Apply it

An executive requests a utilization dashboard. Discussion establishes that the actual decision is which clinics need scheduling support next quarter, so the analyst focuses on appointment access and unused capacity rather than producing an unrestricted utilization report.

Apply it

For a discharge report, executives need service-level trends, unit managers need actionable local detail, and data stewards must verify field definitions. The analyst records these roles before agreeing on report views, validation responsibilities, and release timing.

Apply it

Two teams report different follow-up rates because one counts scheduled visits and the other completed visits. They agree on completed visits within seven days, define eligible discharges, and document how multiple visits and missing dates are handled.

Apply it

A thirty-day revisit report extracted on April 10 includes discharges through March 31. The analyst recognizes incomplete follow-up for late-March discharges and either delays reporting or restricts the cohort to records with a complete observation window.

Apply it

A request asks what proportion of patients had any emergency visit. One patient has six visits and another has none. The analyst creates patient-level indicators, producing one of two patients, rather than treating visits as the denominator.

Apply it

Before examining a service comparison, an analyst specifies eligible encounters, the outcome window, adjustment variables, missing-data checks, and the main comparison. A later subgroup exploration is labeled separately so readers can distinguish planned and exploratory findings.

Apply it

Midway through a clinic access analysis, the sponsor requests medication outcomes from another system. The analyst documents the additional linkage and validation work, then presents a revised timeline or a separately prioritized second phase for sponsor selection.

Apply it

A daily report requires yesterday’s finalized claims, but the source is updated monthly. The analyst identifies the timing conflict during feasibility review and agrees on a monthly claims report or a daily operational measure from another source.

Apply it

A dashboard’s encounter count changes after an interface update. The analyst follows lineage from the displayed metric through the reporting table and transformation rule to the source encounter-status field, identifying a newly included status as the cause.

Apply it

A team wants to understand appointment barriers that are absent from scheduling records. It adds a structured patient survey, tests question clarity, and compares respondents with the eligible population before treating reported barriers as representative of all patients.

Apply it

A January encounter is entered in February after the January extract closed. The analyst uses the source update timestamp to capture the late record, then assigns it to January by encounter date when rebuilding the monthly measure.

Apply it

Two hospitals both use patient number 1042 for different people. The analyst combines source-system identity with the local number, checks the enterprise linkage method, and reviews unmatched records before building a cross-hospital patient dataset for analysis.

Apply it

Only 4% of records lack discharge disposition overall, but one facility is missing 45%. The analyst investigates that facility’s interface and reports the localized problem before using complete records to compare performance across facilities or months.

Apply it

An encounter ends before its recorded start. The analyst checks source timestamps and timezone handling, corrects a verified conversion error, and flags unresolved cases rather than reversing every negative interval or silently replacing it with zero.

Apply it

A patient has two legitimate visits on the same day, and one visit appears twice after reprocessing. The analyst uses encounter identifiers and source version information to retain both visits while removing only the repeated encounter version.

Apply it

A local medication label omits dose form, producing several possible normalized matches. The analyst preserves the source label, requests additional attributes, and flags unresolved mappings instead of assigning the most common drug form to every ambiguous record.

Apply it

Clinic One completes 90 of 100 follow-ups and Clinic Two completes 1 of 10. The combined completion proportion is 91 divided by 110, or 82.7%, rather than the unweighted average of 90% and 10%.

Apply it

A conversion preserves every row but multiplies lengths of stay by twenty-four twice. Record-level examples and distribution checks reveal the unit error despite perfect count reconciliation, so the analyst corrects the transformation before releasing the monthly report.

Apply it

An encounter with a charge of 500 has three diagnosis rows. Summing the charge after a diagnosis join returns 1,500. The analyst preserves encounter-level charges and separately summarizes diagnoses before combining the results for the final report.

Apply it

A report must include every patient and any completed follow-up. The analyst places the completed-status condition in the left join’s ON clause so patients without qualifying follow-up remain present with missing follow-up fields in the output.

Apply it

A report lists clinics with at least fifty eligible visits during June. The analyst filters June visits in WHERE, groups by clinic, and applies HAVING COUNT(*) >= 50 to retain qualifying clinic summaries in the final result.

Apply it

Five encounter rows contain patient identifiers 11, 11, 12, NULL, and 13. COUNT(*) returns five, COUNT(patient_id) returns four, and COUNT(DISTINCT patient_id) returns three; the analyst selects the count that matches the stated reporting question.

Apply it

Five stays last 1, 2, 2, 3, and 22 days. The median is 2 days and the mean is 6 days. The analyst uses the median for typical experience while retaining the mean for resource planning.

Apply it

A histogram of turnaround times shows two clusters corresponding to routine and specialized tests. The analyst examines those services separately before selecting summaries, rather than interpreting the pooled mean as the typical performance of either workflow.

Apply it

The same twelve clinics report waiting times before and after a scheduling change. The analyst calculates each clinic’s difference and checks those differences before considering a paired t test, rather than treating the twenty-four measurements as independent observations.

Apply it

Two service estimates have similar means, but one confidence interval is much wider because it uses fewer observations. The analyst describes the second estimate as less precise and checks study assumptions before recommending a firm performance ranking.

Apply it

Events occur in 20 of 100 exposed people and 10 of 100 unexposed people. The risk ratio is 2, while the odds ratio is (20/80) divided by (10/90), or 2.25, reflecting different comparison measures.

Apply it

Hospital One has higher overall mortality but treats many more high-risk patients. The analyst compares hospitals within baseline-risk categories and finds similar outcomes, showing why the crude hospital comparison alone was inadequate for interpreting relative performance.

Apply it

A clinic compares its follow-up rate with a national figure but discovers that the reference uses fourteen days while its local measure uses seven. The analyst aligns definitions or presents the figures separately with the mismatch explained.

Apply it

A hospital comparison uses baseline age and documented preexisting conditions to account for patient mix. The analyst questions adding a complication that developed during the admission, because that event may reflect the care whose quality is being assessed.

Apply it

A stay of 180 days is confirmed as accurate. The analyst retains it for total utilization estimates, reports robust summaries for typical stays, and compares conclusions with and without the observation to show its influence transparently.

Apply it

An admission model performs exceptionally after using final discharge disposition. The analyst removes that future information, splits data appropriately, and fits imputation only within training data before reevaluating performance on untouched records available at the intended prediction time.

Apply it

A completion rate increases from 40% to 50%. The analyst reports a ten-percentage-point increase and a 25% relative increase, along with eligible counts, so leaders can understand both the scale and the proportional improvement.

Apply it

A large dataset identifies a statistically significant reduction of two seconds in a ninety-minute workflow. The analyst compares the effect and confidence interval with the service’s meaningful improvement threshold before describing the change as operationally important.

Apply it

One clinic records one event among five encounters and another records twenty among one hundred. Both rates are 20%, but the analyst displays counts and uncertainty rather than implying equal evidential strength or assigning a precise performance ranking.

Apply it

A satisfaction report includes only portal survey respondents, who are younger than the full clinic population. The analyst describes respondent characteristics and limits the conclusion to the observed sample instead of claiming the same satisfaction level for all patients.

Apply it

An outcome comparison excludes records without severity information. The analyst reports that exclusions are concentrated in one hospital and repeats the comparison under alternative handling assumptions, showing leadership whether the apparent hospital difference depends on those exclusions.

Apply it

Waiting times fall after a scheduling change, but staffing also rises during the same period. The analyst reports the observed improvement and competing explanation, then proposes a stronger comparison rather than attributing the entire change to scheduling.

Apply it

A service reviews weekly turnaround times across a year. The analyst plots weeks in order and annotates a system change, revealing a sustained level shift that a histogram of all weekly values would conceal from the viewer.

Apply it

Two hospitals compare lengths of stay using identical day intervals. Because their volumes differ greatly, the analyst uses clearly labeled relative frequencies and shared axes, allowing distributional comparison without mistaking the larger hospital’s higher counts for longer stays.

Apply it

A plot of clinic volume against waiting time shows two distinct clusters corresponding to service types. The analyst examines each service type separately rather than fitting one line and concluding that higher volume causes longer waiting throughout the organization.

Apply it

An executive receives a concise comparison with the main limitation and proposed action. Clinic managers receive the same validated definitions plus local detail, while technical reviewers receive calculation specifications and reconciliation evidence supporting the shared conclusions and report.

Apply it

A capacity dashboard shows unused slots by clinic, its last refresh time, and the active date filter. A blank source feed is labeled unavailable rather than displayed as zero capacity, preventing managers from acting on a misleading apparent shortage.

Apply it

A hospital’s adjusted outcome score improves, but completion of follow-up remains lower in one population group. The analyst presents both results, explaining that improved overall comparability does not resolve the subgroup gap requiring operational investigation and action.

Apply it

One clinic’s pilot suggests better follow-up, but the sample is small and staffing differs elsewhere. The analyst recommends a monitored expansion in comparable clinics with defined evaluation measures, rather than declaring that the approach will improve every service.

Apply it

A department adopts role-based access rules, assigns an owner to approve requests, and periodically compares active permissions with current responsibilities. The review finds transferred staff retaining old access, prompting removal and revision of the transfer procedure.

Apply it

An encounter table uses a unique encounter identifier and references a patient table through a foreign key. The analyst checks those constraints and relationship cardinality before joining a diagnosis dimension to a fact table for an encounter-level measure.

Apply it

An internal operations analyst needs encounter dates, service categories, and costs for a defined project. The covered entity evaluates that purpose and supplies authorized fields without unrelated narrative notes or direct identifiers that the analysis does not require.

Apply it

A provider requests records for treating a patient, while an analytics team separately requests records for an operations study. The covered entity evaluates the treatment exception for the provider request and applies the appropriate analytical access rules to the study.

Apply it

A release removes names but retains an admission date of March 14, 2026. For Safe Harbor, remove the month and day; admission year 2026 may remain unless it reveals an age over 89. Replace an exact age of 94 with age 90 or older, and remove or aggregate any date elements, including birth year, that reveal that age. Apply the same date rules to other individual-related dates, remove all other specified identifiers, and confirm there is no actual knowledge that the remaining information could identify an individual.

Apply it

A submitted quality total is challenged. The team identifies the approved output version, source snapshot, transformation logic, reviewer sign-off, and related access activity, allowing it to explain both how the total was produced and who handled the information.