D910 is Data Analysis for Healthcare Decisions, carried at WGU as MHA 6911 and worth three competency units. It teaches principles and techniques of data analysis applied to healthcare decision making, which is a narrower promise than a statistics course makes. You are not being trained to be an analyst. You are being trained to be the executive who can tell whether an analysis supports the conclusion attached to it, and who can produce one that survives the same scrutiny from a board.
Start from the decision, not from the dataset
Most weak submissions in this course begin with data. The student opens a file, computes everything computable, and produces a document full of correct numbers that recommends nothing in particular. Aspects asking for interpretation and recommendation then have very little to score.
Reverse the order. Write the decision first: whether to add a second scheduling clerk, whether to extend clinic hours on Saturdays, whether to change the discharge follow up call script. The decision determines the measure, the measure determines the analysis, and the analysis determines which chart belongs in the document. That chain also protects you from the most common overreach in this subject, which is comparing two groups that were never comparable to begin with.
The second habit worth building early is naming the unit of analysis in one sentence. A row in your data might be a patient, a visit, a claim, a shift or a department. Rates computed across mixed units are wrong in ways that are hard to see and easy for an evaluator with an analytics background to spot.
Sizing an analytic deliverable from the aspects
Your scored aspects live in the Course of Study rather than the catalog. In analytic tasks they usually split into three families: preparing and describing the data, performing and explaining the analysis, and interpreting it into a recommendation. Count how many fall in each family before you begin, because the third family is where most of the writing goes and the least of the student effort usually does.
The word budget, worked. Take a 1,800 word report with nine scored aspects, exhibits excluded from the count. Reserve 120 words for a purpose statement naming the decision, leaving 1,680, or about 187 words per aspect flat. Analytic aspects are efficient in prose because a table carries the detail, so let the data preparation and calculation aspects run at 130 each. That releases roughly 340 words, which should go to interpretation and recommendation, taking those to about 300 each. Charts and tables do not count toward the argument, so never let an exhibit stand in for a sentence: every exhibit needs a line above it saying what to look at and a line below saying what it means.
Build the exhibits before you write the prose. Writing first tends to produce claims that the finished chart does not actually support, and rewriting around a chart is slower than choosing the right chart.
The shape of an analysis a board would accept
Follow any template your directions supply. Where they do not, this arrangement puts the material in the order a decision maker reads it and keeps each scored aspect visible.
| Section | What it contains | The check before you move on |
|---|---|---|
| Decision and question | The choice being made and the question the data must answer | Could someone disagree with the decision as stated |
| Data description | Source, period, number of records, unit of analysis, known gaps | Is the denominator stated |
| Preparation | Cleaning steps, exclusions and the reason for each | Would another analyst reproduce your file |
| Descriptive findings | Central tendency, spread and distribution, with the right chart | Does the chart type match the data type |
| Comparison or trend | The specific contrast, with the method named | Are the groups comparable, and if not, what did you adjust |
| Interpretation | What the result means for this organization | Is any word here stronger than the analysis supports |
| Limitations | What the data cannot show | Have you named at least one real limit, not a token one |
| Recommendation | The action, the owner, the cost and the measure to watch | Could this be implemented on Monday |
The preparation row is the one students skip and evaluators use. Exclusions change results, and a document that reports 412 records analyzed out of 480 received, with the reason for the difference, is far more credible than one that quietly reports 412.
Evidence craft when the evidence is a number
Numbers carry an authority they have not always earned, and graduate analytic writing is expected to manage that.
- Report the measure of spread alongside any average. A mean wait of 22 minutes hides a very different situation depending on whether the spread is four minutes or forty.
- Use the median for skewed operational data such as length of stay, cost per case and wait times, and say why you chose it.
- Label every axis, unit and period on every exhibit. An unlabeled chart is an unscoreable chart.
- Keep percentages honest by stating the base. A 50 percent increase in a category with four events is noise wearing a suit.
- Say whether a difference matters operationally, not only whether it is detectable. Statistical language belongs beside a practical judgment.
- Cite your data source in APA the same way you would cite a paper, including the version or extraction date.
One sentence separates good analytic writing from great: naming the alternative explanation. If readmissions fell in the same quarter that admissions criteria tightened, say so. Evaluators reward the analyst who looks for the confound rather than the one who reports the favorable number.
What separates Competent from a submission sent back
Returns in this course are usually about the join between number and judgment. The analysis is correct and the recommendation does not follow from it. Or the interpretation restates the figure in words, which is a translation rather than a finding. Or the report presents six exhibits and never says which one drove the decision.
Work that passes on the first read has a visible chain. Decision stated at the top. Data described with a denominator. Preparation steps listed. One or two exhibits that are actually referenced in the argument. An interpretation that says what the numbers mean for this organization, with a because clause. A limitation paragraph that names something real. And a recommendation with an owner, a cost order of magnitude and a follow up measure.
WGU performance assessment work can be revised and resubmitted with no grade penalty, so a return is time rather than damage, but analytic tasks are unusually annoying to rebuild because a change to an exclusion rule ripples through every exhibit. Build the file once, carefully, and document what you did while you are doing it.
If your plan pairs this course with a proctored objective assessment, our involvement is preparation only. We work through practice problems, drill the vocabulary and give an honest readiness call. We are not in the room for a proctored assessment, and we never hold portal credentials.
Six mistakes that cost time in D910
- Analysis before decision. Computing everything and choosing a conclusion afterwards produces a report with no argument in it.
- Averages with no spread. Operational data is rarely symmetrical, and the mean alone often describes a situation nobody experiences.
- Charts that decorate. Every exhibit should be referenced by a sentence that tells the reader what to see in it.
- Silent exclusions. Dropping records without saying so is the analytic equivalent of an uncited claim.
- Causal language on a comparison. Two groups differing is not one group causing the other, and the wrong verb costs an interpretation aspect.
- No cost on the recommendation. A decision memo without a resource line reads as an analyst's note rather than an administrator's.
How we work this course with you
Send the task directions, the rubric and any data file the task supplies, and you get back the decision statement written first, the measure that answers it, the exhibits that are worth building and the ones that are not, and a section plan with word counts that protects the interpretation and recommendation aspects. On review we read the numbers against the sentences to find any claim the analysis does not actually support, and we check that every exhibit is referenced by name somewhere in the argument. Most students find that the second pass is faster than the first, because by then the file is documented and the only work left is prose.
Questions D910 students ask
How much spreadsheet skill do I need for this course?
What if the supplied data has obvious errors in it?
Can I use my own workplace data?
Building the D910 analysis?
Send the MHA 6911 directions and any data file. You get a decision statement, an exhibit plan and a section budget back.
Where D910 sits in WGU's programs
The July 2026 catalog places this code in 1 current WGU program. Open a program page for the complete standard path and term positions. The live Degree Plan remains authoritative after transfer credit, substitutions, and mentor planning.
The assessments, one by one
The public catalog does not publish this course's PA/OA identity or task count. WGU Tutors publishes at most one PA manual per course and only from a WGU-controlled public rubric. Until that source exists, PA help begins from the student's real Course of Study and OA support remains preparation only.