Patient Count Analysis of Sepsis Data set using Tableau

The first step of data analysis is to have a basic segregation of the dataset.

Segregation helps us to understand which part of the dataset needs to be focused and drilled down to draw patterns or insights which will further help the experts in decision making.

Understanding the Domain of Data:

• The given data set is clinical data-patient records to be precise.

• Size of the sample: 1.5M

• The data has 43 features which can broadly be classified into

• Demographics

• Vital Signs

• Laboratory values

• Total no. of patients: 40336

  
  

What is sepsis?

SEPSIS is a life threatening condition that arises when the body's response to an infection will damage it's own organs. The infection will weaken the immune system causing one/more body organs to dysfunction and if severe it could ultimately result to death. Not all the infections can cause SEPSIS but any infection has potential to lead to SEPSIS.

Basic segregation of Sepsis dataset using Tableau

When a patient is admitted to a hospital

• Sepsis label is 0 (No sepsis on admission)

• Sepsis label is 0 and 1(No sepsis on admission and then converted to sepsis)

• Sepsis label 1 (Sepsis on Admission)

Using the above data, a meaningful insight of the dataset is shown below:

Fig 1. Segregation of Patients based on Sepsis label

In the dataset, the total count of patients is 40,336. Out of which we have 426 sepsis patients and 2,506 onset-sepsis patients, i.e patients who did not have sepsis when they were admitted in the hospital but developed later during their stay in hospital.

Fig 2. Representation of Age wise Break up of Sepsis and Non Sepsis patient count with a segregation of Female and Male count.

• Age bin to the left side of the chart spans across age groups of 10-100.

• The red bars of the butterfly chart represents Sepsis patient count and the green bars represent Non Sepsis patient count.

• One chart each for Female and Male.

• Higher numbers on the age group 60-70.

• So the next set of analysis will be focusing more on the age group 60-70.

  
  

Blood pressure variations is one of the ways of understanding the patients' health condition. Representing Blood pressure details in a chart to decide which part of the dataset needs to be drilled down deeper.

Below is a chart with Blood Pressure details of the patients:

Fig 3. BP Analysis of all the Patients in the data set with sepsis condition as filter.

Fig 3.1 Shows the same chart with filtered data- Shows only Onset Sepsis and Sepsis patients BP analysis

Like this the sepsis data set can be drilled down more with a comparison analysis of more vital signs to derive more meaningful insights.