0 Replies Latest reply on Aug 30, 2017 7:40 AM by Hope Stiles

    Feedback on sketched dashboards

    Hope Stiles

      Good Morning,

       

      I am working on a new project and have some sketches of what I think the dashboards are going to look like.  I'm hoping you all might have some insight and help me think of things I might have forgotten to consider or have some tips to make these even more effective.  Any feedback is appreciated.

       

      First, the project starts with a main Outcomes Dashboard.  This one can be filtered by market and facility, and will show Readmissions, HAI/Falls, Patient Experience, Mortality, and Financial information.  The user should be able to choose one of those categories and navigate to 1. a breakdown of the outcome or 2. a process measure dashboard for the outcome.  I haven't yet figured out the best way to allow that navigation choice without a lot of buttons and some of the users will be on an ipad so I'm hesitant to put links in the tooltips.  Here is what I have:

      (at least 6 data sources)

      I'm thinking I'll use control charts to the mortality and readmissions rates to spot where something seems to be off while putting the index values in a table.

      The patient experience portion will show like 18 rolling months.  That scribble is where there is a lightly shaded area for the 90th percentile.  The circles are something we refer to as the Patient Experience Statistical Significance Indicator which helps leaders tell is the movement they see if unlikely due to change/normal variation.  I think the rest can probably be understood without explanation.

       

      I have sketched out 4 of the Breakdown dashboards.  They will not have filters on them, as they will be for whatever filters were chosen on the main outcomes dashboard.  The user does need to be able to navigate back to the main outcome dashboard.

        First is Readmissions.

      There are sparklines for the Index and rate with a text value for the YTD value next to the end of the line.  I'm planning to have the axes for each metric start at the lowest value in the series and end at the highest value in the series.  After those two sections it will show by text the FYTD numerator/denominator.  Next is a strip plot for the days to readmission with the first 72 hours being shaded and then a text value for the number of 72 hour readmissions.  (2 data sources)

       

      Next is HAI/Falls:

      It is kind of a spin off of small multiples.  The top row are HAIs, second row SSIs, and the 3rd row are falls.  Each has a bar graph showing the number per month.

      We also want to show the number of days since the last event, which is what the # are above the charts.  One concern I have with this one is the top two rows come from 1 source and the bottom row another (2 data sources).  Then I think I'll have the days since # sheet floating right above the charts which means I'm looking at 6 worksheets.  From a performance standpoint I'm concerned this will be slower.  I've considered using stacked bars for Class 1 ortho and NHSN SSIs but think I'm leaning away from it and perhaps put that info in the tooltip or something.  And this doesn't include anything for YTD and I think that might be important but I'm not sure where to put it.

       

      Next is Patient Satisfaction:

      We'll show each of the domains here with a rolling 18 months using the same shading and Patient Experience Statistical Significance Indicator as on the Main Outcomes dashboard.  Then a bullet chart for the FYTD with a line for last year's performance.  I'm struggling with the shading in the background.  We usually color code our HCAHPS into 4 categories (red is <50th, Orange is >=50th and <75th, Yellow is >=75th and <90th, and green is >=90th).  Four seems like too many but three might confuse people.  I won't use the actual colors anyway, but I'm thinking of using shades of gray.  The last column would show the number of surveys that need to move to a top box answer to reach the 90th percentile.  For the domains with multiple questions, I calculated this for each question and am thinking I'll show the max value as the value for the domain.  It won't be precise but is the best I've come up with given the methodology to get the domain score.  I'm also debating if this should show how many need to move for the FYTD or the most recent month. Once you're half way into the year, the FYTD value might be too overwhelming where the most recent month should seem attainable. (Only 1 data source )

       

      And the last one I have sketched is Mortality:

      Just like for Readmissions, There are spark lines for the Index and rate with a text value for the YTD value next to the end of the line.  I'm planning to have the axes for each metric start at the lowest value in the series and end at the highest value in the series.  After those two sections it will show by text the FYTD numerator/denominator.  Then there will be a strip plot for the age of patients with a reference line at 66.  I'm wondering if I should use control charts on at least the rate, but there isn't a lot of room there and I don't want it to be too cluttered.  But we don't really have targets for the rates and what I want them to see if when something should cause them to be alerted to look deeper.  (2 data sources)

       

      The financial stuff is just going to be a place holder for now so I don't have anything sketched there.

       

      Any feedback is welcome and I'm working in 10.1.

       

      Thanks,

      Hope