Project Brief

The project brief in short was to evaluate whether the HCAHPS survey has been successful in accomplishing the goal to "create incentives for hospitals to improve their quality of care". The survey data encompass 9 years, as a data-analyst your report should answer questions regarding quality improvement scores over the last 9 years and areas most suitable to try to increase quality to improve the patient experience.

Philosophy and Approach

After cleaning, modelling and playing with the data (and getting a feeling for trends), my approach evolved into “mention historical trends, show the current status of measures surveyed and focus on weaknesses and opportunities”. Actionable insights, no interaction, the last survey results are the start for future improvement.

For the report I only used the top-box scores, in other words the percentage of people rating their experience in the most positve box. The why is in the section modelling.

Key Reporting Areas

Hospital experience in general

I looked at the trend for general appriciation of the hospital, the willingness to recommend the hospital and the willingness to answer a survey.

Surveyed measures in KPI Cards

For each surveyed measure I created a KPI card with the average top-box score in the 2021-2022 report. A bar visual shows the average top box score and also the minimal and maximal score, thus providing an inside in the value range 2021-2022 . At the bottom of the card the progression in quality improvement in the range 2013 – 2022 as experienced by patients is shown.

If the average top-box score is < 65%, the value is colored red. If the 2013-2022 overall progression is < -2%, this is also shown in red. Both 65 and -2 are my personal perception of dangerzone (or opportunity).

Also on the KPI card are top improvers and bottom performers. A top improver is a state which has made the most progress in quality improvement (last surveys value – first surveys value). A bottom perfomer is a state which had the lowest value for the measure in the latest report. If possible, if the top-box score was low, I tried to insert as much top improvers as possible. This part ended up a bit messy, see Technical design below.

Technical - Design

Modelling

I dropped the national results data for the analysis because this was a summary of the state results. A few things I wondered about while modelling:

• it looks like for part of the surveys (see questions) the middle box value has a more negative sentiment than the bottom box value. I switched those value although I never used them. I wondered why, I wondered if I should trust them, and I wondered how the questions were presented to the patients (order of answers).
• There were missing state values, Wikipedia helped. I created an updated States.csv.
• After loading this file (and playing around a bit more) I discovered some states (Virgin Islands & Puerto Rico) have submitted surveys which were not processed into state results. Why?

KPI Cards

The KPI Cards were created as SVG image URL for the new KPI Card visual in Power BI.

I got rid of the callout value (and label) by creating a trigger measure for the SVG image and hiding the callout value (and label). The result was a grid to show my SVG images.

The starting point for the SVG image was a SVG template published by Kerry Kolosko (thank you Kerry) :(https://kerrykolosko.com/portfolio/kpi-card/).

Things got messy when I wanted to insert the top improver and bottom performer. Text-wrap in SVG is not that easy (understatement). I had and still have a TOPN measure with an output string of unknown length. I moved the output from the SVG image to a separate Power BI Card visual to make it possible to manipulate the output a bit more.

Overall Design

With my goals in mind (summary of history regarding quality, current status and actionable insights) I decided not to add much “design” to the dashboard. When I googled the HCAHPS logo and inserted it into the report I took some colors from their website and used them in the resport and I am afraid that is just it.

"Thank you" to Gerard Duggan for writing such structured summaries for his projects. I copied one and have tried to describe as good as I can my process. I'm not going to use an online tool to improve the english of this text, I used such a tool for the report itself.