NURS 6231 Week 9 Outcomes Dashboard Critique Example

All Green, Nothing Seen: Would a Worsening Unit Show Up on the Nursing Quality Dashboard?

Student Name

College of Nursing, Walden University

NURS 6231: Healthcare Systems and Quality Outcomes

Instructor Name

Month Day, Year

All Green, Nothing Seen: Would a Worsening Unit Show Up on the Nursing Quality Dashboard?

Each month, the nursing division of a composite 350-bed hospital reviews a one-page quality dashboard. For the past year every cell has been green. This critique asks whether green means the units are doing well or that the dashboard cannot tell. It takes each design choice in turn and tests it against one question: if a unit were getting worse, would the dashboard look any different? For most of the choices, the answer is no.

The Dashboard as Its Readers See It

The dashboard has four panels. Each shows a single number for the division for the current month: injurious falls per 1,000 patient days, catheter infections per 1,000 catheter days, stage 2 and deeper pressure injuries acquired in the hospital per 1,000 patient days, and the HCAHPS top-box percentage, the share of patients choosing the most favorable answer, for nurse communication. Each number sits in a cell colored red, yellow or green. Green means the figure is at or better than a target, and each target is a national mean from a benchmarking vendor. Yellow means within 10% of the target. Beside each cell is last month's figure. The division's 14 inpatient units do not appear, and neither do the numbers of falls, catheter days or completed surveys behind the rates.

Aggregation to the Division

The division figure pools 14 units. Suppose falls with injury on one 24-bed unit doubled over six months while two larger units improved modestly. The pooled rate could stay flat, and the cell would stay green. A worsening unit would not look different, because it does not appear. The same is true of every panel. A division total answers the question of how the division is doing, but the people who can act, unit managers and their staff, need to know how their unit is doing, and the dashboard withholds that.

One Month Beside the Last

Each cell shows this month and last month. Two points cannot distinguish a real change from ordinary variation. Deming (1986) warned that reacting to each rise and fall in a stable process as if it had a special cause makes performance worse, since each adjustment adds its own variation. The dashboard invites exactly that reaction. When a cell turns yellow, the division asks for an explanation; when it returns to green, the question is dropped. Neither response is informed by whether anything changed.

A run chart solves this. Plotting each month's value in time order with a median line, and applying simple rules, for instance treating six straight months on the same side of the median as a signal, lets a reader see a shift that no single month would reveal (Perla et al., 2011). A unit that slowly deteriorated would show a run above its median months before any single month crossed a target. On the current dashboard, the same unit would look no different until one month happened to cross the line.

Missing Denominators and Sample Sizes

The dashboard shows rates without the counts behind them. For the infection panel, the rate depends on catheter days, and a unit that sharply reduced catheter use could see its rate rise while infections fell. For the survey panel, the problem is larger. Hospital-level HCAHPS results for public reporting pool four quarters of surveys, and the program expects larger hospitals to collect at least 300 completed surveys a year (Centers for Medicare & Medicaid Services [CMS], n.d.). A unit's monthly score may rest on a dozen returned surveys. At that size, two unhappy patients can move a unit's top-box percentage by more than 15 points. The dashboard gives no sign of this, so readers cannot tell a real decline from a small sample. A worsening unit and an unlucky one look the same, and so do an improving unit and a lucky one.

Targets Set at the National Mean

Green means average or better. For falls with injury, a unit at the national mean still injures patients at a rate the organization would not choose if it were stated as a count. Setting the target at the mean is a design decision, and its effect is that average performance is shown as success. It also means that a unit that was well above average and slid back to the mean stays green throughout its decline. A target tied to each unit's own trajectory, such as its median over the prior year, would show that slide.

The critique does not propose abandoning external comparison. For surveillance measures such as the infection rate, the national reporting system compares facilities through a standardized infection ratio that adjusts for type and mix (Centers for Disease Control and Prevention [CDC], 2025). That ratio belongs on a hospital-level report. On a unit's dashboard, the more useful comparison is with the unit's own past.

Outcomes Only

Every measure on the dashboard counts harm after it has happened. A unit whose hourly rounding had quietly stopped, or whose catheters were staying in longer, would show nothing until falls or infections rose, and with low monthly counts that could take many months. A worsening process would not look different on this dashboard until patients had been hurt.

Redesign

Each proposal answers one of the flaws above. First, show each unit, with the division total as the last row, so that a declining unit is visible. Second, replace the two-month comparison with a 24-month run chart for each unit and measure, with the median marked and runs flagged. Third, print the count and denominator beside every rate, and the number of completed surveys beside every HCAHPS score, graying out any score based on fewer than 30 surveys. Fourth, set each unit's reference line at its own prior-year median, keeping national benchmarks for the hospital-level report. Fifth, pair each outcome with one process measure that moves before harm: observed hourly rounding with falls, catheter device utilization with infections, and completion of turning schedules with pressure injuries.

The redesign does not add every measure available. Four outcomes and four paired processes fit on one page per unit, and a dashboard's value depends as much on what it leaves off as on what it shows.

Conclusion

Tested against a single question, the dashboard fails most of its design choices. Aggregation hides units, two-point comparisons hide trends, missing denominators hide sample size, mean-based targets hide decline, and outcome-only panels hide the processes that fail first. A redesigned dashboard that shows units, time, counts, the unit's own history and leading measures would let a worsening unit look different, which is the purpose of reviewing it.

References

Centers for Disease Control and Prevention. (2025). National Healthcare Safety Network (NHSN) patient safety component manual. https://www.cdc.gov/nhsn/pdfs/pscmanual/pcsmanual_current.pdf

Centers for Medicare & Medicaid Services. (n.d.). HCAHPS: Patients' perspectives of care survey. https://hcahpsonline.org/

Deming, W. E. (1986). Out of the crisis. MIT Center for Advanced Engineering Study.

Perla, R. J., Provost, L. P., & Murray, S. K. (2011). The run chart: A simple analytical tool for learning from variation in healthcare processes. BMJ Quality & Safety, 20(1), 46-51. https://doi.org/10.1136/bmjqs.2009.037895