Sepsis Apache analysis

Sepsis is a serious medical condition in which a person’s body does not respond to an infection properly. Generally, the human immune system sends signals to the infected area to deal with the infection locally, but in sepsis the response becomes dysregulated and can cause widespread inflammation all across the body. This, if not treated immediately can lead to tissue damage, organ failure and in some cases even death.

Stages of Sepsis

       I.          SIRS: Systematic Inflammatory Response Syndrome is the initial stage of sepsis where the body’s defense system starts reacting to a harmful stressor and the situation begins to exacerbate. A harmful stressor can be either an infection, a surgery or an acute inflammation. SIRS is identified if any of the below four medical conditions develops –

• WBC>12000 or WBC<4000

• Temp>38.5°C or Temp<35°C

• Tachycardia which means HR>90 bpm

• Tachypnea which means RR>20 breaths per minute or PaCO2<32 mmHg

      II.          Sepsis: Sepsis occurs when the body’s defenses start working against the body rather than for it. This causes organ dysfunction like altered mental status, decreased urine output.

    III.          Severe Sepsis or Septic Shock: this is a lethal medical condition where the patient’s condition worsens beyond a point significantly decreasing blood pressure and this may also result in fatality.

To measure the risk of mortality of patients, Apache II and Sofa scores are widely used. Here, I will talk more about the Apache II score.

Apache II

Apache II score is basically the sum of Acute physiological score, age points and chronic health points. Acute physiological score is in turn a sum of 12 individual physiological variables (biomarkers). Apache score ranges from 0 to 71. Increasing score means increasing risk of hospital death.

We can define the predicted mortality rate based on Apache score as below:

•  Apache score 0 – 4: ~4% mortality

• Apache score 5 – 9: ~8% mortality

• Apache score 10 – 14: ~15% mortality

• Apache score 15 – 19: ~25% mortality

• Apache score 20 – 34: ~40% mortality

• Apache score 25 – 29: ~55% mortality

• Apache score 30 – 35: ~75% mortality

• Apache score above 35: ~85% mortality

I would like to demonstrate some analysis done on a group of sepsis patients using the Sepsis dataset.

Demographic Analysis

It is observed that among 40,336 patients, most patients fall under the 25% mortality rate category. It means although many people have contracted sepsis, not a lot of them are at a high risk of death. A total of 791 patients are extremely critical according to their apache scores.

Now that we know the distribution of patients across various death rates, we can analyze what percentage males and females are at the risk of dying. Maximum number of males and females have just 25% risk of mortality. There are a significant number of men greater than women who are at a higher risk of losing their lives due to sepsis. Especially, in the 75% and 85% death rate category there are more men than women.

Since apache score ranges from 0 to 71, it becomes quite difficult to analyze patients’ health with it. For this reason we can categorize the scores into low, medium and high. Let’s categorize these scores as below:

• Apache score <= 14: Low

• 15 < Apache score <= 29: Medium

• Apache score > 30: High

If we fan out the patients in all age groups according to the apache category, number of patients in medium risk is the highest and that too concentrated between 60-80 age groups. But surprisingly, people with high risk are spread across 50 to mid 70 years of age. People between 50-60 might have a better survival chance if they receive prompt treatment.

ICU Analysis

Let’s take a look at how many patients were admitted in the ICU ward and what was their apache score.

On day1, patient count is higher for apache score 17 and for day 3 there are most patients with score 19. It indicates that health of some patients is deteriorating as they have to stay longer in the ICU and their risk of mortality is also increasing.

Mortality in the ICU is higher on day 1 compared to day 3 for APACHE patients" suggests that patients who are admitted to the intensive care unit with higher APACHE scores, indicating greater severity of illness, are more likely to die earlier in their ICU stay, particularly within the first 24 hours (day 1), compared to later in their ICU stay (day 3).

Organ Dysfunction

We all know that prolonged suffering can lead to multiple organ failure in a patient.

In the above chart we see that as the apache score increases, severity of illness increases and the patients are more likely to develop dysfunction in specific organs such as heart and kidneys. More specifically kidney failure seems to be more prominent. We also see that there is one patient with score 43 who has developed a heart failure and might be at a life-threatening stage.

Like apache, SIRS score is another measure that indicates the level of gravity of the situation. In the below chart, I wanted to analyze how many patients transferred from SIRS stage to further stages.

Here, we have taken to into account SIRS Score 3 and 4 and catgorized patients on high mortality rates only. The white circle shows the total number of patients having SIRS score 3 and 4. The second circle has 1009 patients with SIRS score 3 and 60 with SIRS score 4. Further segregating these patients we have maximum number of patients in 55% mortality rate in both the sections and then 85%. The final circle shows what Apache score is responsible for their respective mortality rate in descending order.

Hence, we find that apache score is a very crucial tool in examining the health deterioration of patients with sepsis. In this dataset not many patients are in an acute danger but still they need immediate treatment.

Hope you all find my analysis agreeable!