The Hospital Safety Net: A Predictive Approach to Patient Readmissions-PART 2
A Descriptive Analysis of Hospital Readmissions
Hospital efficiency does not depends upon how many beds a hospital has or how quickly they clean the rooms, change the bed lines, arranging medical supplies, making the room ready for next patient or updating the records. The actual story begins only after a patient leave to their home. What generally happens—how soon someone recovers, any complications arises, and how often they return back because of illness—this leads us to decide how the hospital really performed.
By going through the registry of 720 hospital records on admissions, we analyse the post-discharge journeys. The data help to discover the clear patterns that helps to explain why some patients recover smoothly without any struggle but others do. These observations don’t just gives us individual results—they describe system-level factors that can help to improve recovery, care and long-term success.
1. General Descriptive Statistics: The Clinical Baseline
The data shows the picture of a high-complexity care environment, where no two recovery journeys look like same. Some patients recover too quickly, while others stays in the hospital for longer period of time and also with higher physiological risks. On average, patients spend nearly 8.8 days in the hospital—but that number hides different experiences among patients like some recover fast, some recover slow and some has very serious conditions.
A significant group of patients carries high clinical risk, which helps us to explain why recovery times changes so dramatically. These differences are very important, because they not only gives individual outcomes, but also helps to analyse how the hospitals plan care unit, allocate needed resources, and support patients even after they discharged.
Average Expected Mortality: 33.8%
Average Expected Length of Stay (LOS): 8.82 Days
Average Time to Readmission: 5.36 Days (for those who return)
The data gives us the information that for patients who do leave to their home, the "re-crash" happens very quickly— generally within the first week of post-discharge.
2. Categorical Distributions: Where Care should be Concentrated
Generally not all the patients across the hospital experience same level of Clinical risk. The facility in the hospital operates across 6 various services and also manages 17 unique primary diagnoses.
Service Volume: The hospital is anchored by Cardiology, the ICU and General Medicine. They were generally consider as high-volume workload running department of the hospital. Anyway we cant compare volume to the readmission risk. Because, for example Cardiology has high volume but its readmission rate is moderate and ICU has high volume and also higher risk.
Diagnosis: From the data there are 17 unique diagnoses but let us consider "Top 10" because they are responsible for major readmission. We see patients with heavy concentration of acute and chronic conditions, ranging from high risk acuity Strokes and Heart Failure to more common seasonal cases like the Flu. Strokes and Heart Failure are difficult to manage. It require proper lifestyle and medication at home and also multiple organ systems.
3. Readmission Analysis: Reality-The "Return"
The most critical measurements to be consider for any hospital is its readmission rate. From the analysis we find out that the significant portion of patients are struggling to maintain health post-discharge.
Global Readmission Rate: 26.94% (Approximately 1 in 4 patients).
Post-Discharge ED Visits: 47.36%. From this we came to know that half of the patients are coming for immediate visit within a week of discharge.
The Golden Window: If the patients stays at home for more than ten days from the day of discharge then recovery phase of the patient increases exponentially.
Interestingly, there is a perfect correlation in analysing this data between Emergency Department visits and readmissions, resulting the ED as the primary "early warning" zone for failing recovery of the patients.
4. Visualizing the Risks (Coding & Insights)
To analyse these insights, we use Python's pandas, matplotlib and seaborn libraries. Presented below are the coding logic and the resulting visualisations.
Plot 1: Top 10 Diagnoses by Readmission Rate
Some diagnosis has highest risk when compare to others. Stroke and Heart Failure has the highest readmission rate.
Plot 2: Readmission Rate based on Hospital Service
The ICU has the highest readmission rate at 41.1% and Neurology (29.2%). Conversely, Orthopaedics has efficient readmission rate of 4.4%.
Plot 3: Expected Mortality vs. Readmission (Risk Profiling)
At the time of initial admission, patients who are all readmitted consistently are at higher Expected Mortality rate which is analysed based on clinical risk vs outcomes.
Results for Summary
The data gives us clear picture of where the hospital's "safety net" needs to be strengthened.
Based on services provide by the hospital readmission rates are analysed
Conclusion on Descriptive Analysis
From the descriptive analysis proves that the readmissions are concentrated in high-acuity services and specific chronic diagnoses. By targeting on the ICU and Neurology departments, concentrate on the discharge process with patients admitted for Stroke and Heart Failure, the hospital has a efficient opportunity to lower its 26.9% readmission rate.
