Requirements

1. Spotfire DXP version 1.0 or later.

Instructions

1. Right-click on the following link and save the sample data file to your computer: ClinicalLabs.csv

   Here is a snippet of this file:
   
   Notice the format of the data: Each row is a labs result for a single subject on a specific visit.

2. Open Spotfire DXP. Drag and drop the ClinicalLabs.csv file onto the DXP window. Spotfire shows you the Import Settings dialog box. Click OK.

   Spotfire reads in the data, automatically recognizes all the data fields, and creates an initial scatter plot of PatientID vs. Value. Spotfire also creates filters along the right for each data field.

3. Close the graph by right-clicking on it and selecting Close.

4. Click the bar chart icon.

   Spotfire displays a bar chart, initially configured with a bar for each VisitName, and the height of the bars reflecting the sum of the PatientIDs for the collection of patients tested at each visit.

5. We would first like to see the number of lab results for each lab test.
   
   On the Y axis, click on the arrow that's part of the "Sum(Patient ID)" label, move the mouse over Aggregation at the top of the selection box, and select Count.

   Now the bar heights reflect the number of tests performed at each visit. We see that we have very few test results for Day28.

   On the X axis click on the arrow that's part of the "Visitname" label, and select LabTest.

   Now the bars represent each lab test, and the bar heights show how many results we have for each test. Hover over any of the bars to see the name of the lab test and the number of results

6. We would also like to see the number of Out-of-Range lab values for each lab test.
   
   Find the Out of Range filter in the filter panel. Drag it on top of the bar chart, and drop it on the color palette icon. .

   Now we can see that there are a lot of Low lab results for the BUN test. We also see a large number of High lab results for TCHOL and THT4.

7. Click on any bar segment to see details of those lab results in the Details-on-Demand panel in the lower-right.

8. We'll adjust the chart to be more informative.
   
   
   • Right click on the bar chart and select Properties. Select the Appearance tab (on the left hand side). Check the box for All bars should have maximum height.
   • In the Properties window, select the X-axis tab (left hand side). In the Scale labels section (middle of window) click on the Vertically button and change the value of Max number of labels to 50. Close the Bar Chart Properties window.
   • Drag the bottom line of the chart up a bit to make room for the test names along the bottom.

   We now see the relative number of High,Low,Normal, and Empty lab results instead of the absolute number.

9. Let's compare our treatment groups. If the relative number of Out-of-Range results (High or Low) is greater in the Active group than in the Placebo there may be a safety issue with the protocol.
   
   Find the Group filter in the filter panel. Drag it on top of the bar chart, and drop it on the icon of 2 bars charts which are side-by-side .

   We now see the relative number of Out-of-Range labs for our Active and Placebo treatment groups.

10. Earlier we saw that the BUN and TCHOL lab tests showed high numbers of Out-of-Range values. Let's confirm if there is a safety issue by comparing Active vs Placebo for these specific tests.
    
    Find the LabTest filter in the filter panel. Move the filter slide so that BUN is selected.

    All the other lab tests are filtered away in our visual. We see that there are actually relatively more Out-of-Range (Low) values in the Placebo group.

    Move the filter slide so that TCHOL is selected.

    We see that over 70% of the results for the Active group are High, compared with 41% for the Placebo group. This may indeed be a safety issue.

    We can repeat this for any of the lab tests in our dataset. Check out the Lymphocytes or Chlor lab tests.

    You can continue to investigate this data by interacting with other filters. For example, instead of looking at Active vs. Placebo, you can examine Out-of-Range labs broken down by Site, Race, Sex, etc., by dragging any one of these filters onto the side-by-side icon .