In the dynamic healthcare landscape, data analysis is vital for optimizing patient care and hospital performance. This project utilizes Excel spreadsheets, "Quality Data" and the "Readmission Registry," to extract insights from diverse healthcare data.
I employed Python for data analysis, Tableau for visualizations, metrics, and KPIs, and SQL for data validation and querying.
Python for Data Understanding and Analysis:
I utilized Python to conduct comprehensive exploratory data analysis (EDA). This involved bivariate and multivariate analyses, as well as leveraging tools like Pandas Profiling and data prep for an in-depth understanding of the data. During the process, I identified missing values and addressed outliers.
Key components of the exploratory analysis included:
- Utilizing built-in libraries for data analysis, data preparation, and Pandas Profiling.
- Analyzing ED visits data based on gender and acuity.
- Examining patient data by gender and reasons for ED visits.
- Investigating the relationship between Length of Stay (LOS) and Expected LOS.
- Conducting demographic analysis, which involved scrutinizing data based on gender and organizational factors.
- Exploring factors such as age groups, language spoken, and months with the highest activity.
This analysis provided valuable insights into the dataset, facilitating a deeper understanding of the data's characteristics and relationships.
Link for python code:
Here are the key points from the provided data:
1. Most patients received treatment through referrals or organizational affiliations.
2. "Group, Inc" and "LLC" organizations have the highest number of patients in the dataset.
3. The majority of patients were born between 1975 and 1980.
4. The number of female patients is higher than male patients.
5. The number of Black/African American patients is approximately half that of White patients.
6. Most patients are English speakers.
7. Patient visits spiked in April 2019 and decreased in May 2019.
8. The number of new patients is half the number of follow-up patients.
9. Blood pressure Systolic and Pulse readings are evenly distributed.
10. Blood pressure Diastolic readings follow a normal distribution.
11. Blood Pressure Systolic, Blood Pressure Diastolic, and Pulse have the highest number of missing values.
12. Primary care visits are the most common reason for patient visits.
Days to readmission has a significant number of outliers. The average days to readmission id 5 for all the patients. All the patients who have been readmitted after 8 days are considered outliers.
MY Analysis:
The Data has more English speakers than Spanish speakers and a higher number of Black African/ American race patients than White race patients.
only one patient with cough was admitted to ED and the patient was a Male. Most of the patients have visited for Pneumonia and Fever.
Most patients were admitted to ED with Acuity of 1. All the age groups admitted to ED are of the same volume. Age does not make much difference in determining who gets admitted most.
Most male patients have been admitted to the ED than female patients. There are very less number of patients in Acuity 3.
The Expected LOS and LOS have very small difference for the patients who are discharged to home. Whereas, patients who have expired or have been transferred have a lot of differences between Expected LOS and the actual LOS.
highest readmission rate based on the service General Medicine 263
Tableau Analysis:
I used tableau to show key metrics and KPIS.
1. Length of Stay Analysis: We explore the Length of Stay (LOS), a key metric for assessing care efficiency, associated risks, and its correlation with various factors. A longer LOS can lead to increased risks and lower quality of care.
2. Readmission Analysis: We delve into the analysis of hospital readmission rates, identifying trends and factors contributing to patient re-entry within specific time frames after discharge.
3. ED Throughput Analysis: This assignment focuses on understanding the efficiency of the Emergency Department (ED) by examining the time from patient arrival to discharge and considering patient acuity and disposition.
4. Mortality Analysis: We assess hospital performance using the Observed-to-Expected (O:E) mortality ratio, examine mortality trends, and analyze the impact of primary diagnoses and demographic factors.
5. Ambulatory Visits Analysis: Ambulatory care, a significant contributor to healthcare costs, is analyzed, including completed visits, new visit types, no-show rates, appointment scheduling, and trends in visit types and no-show rates.
6. BP Control Analysis: Critical for cardiovascular health, we evaluate blood pressure control, control rate trends, and the influence of demographics and visit types on achieving controlled blood pressure.
Each analysis offers valuable insights for improving patient care, resource allocation, and overall hospital performance, aiding healthcare professionals and administrators in their decision-making.
In summary, these analyses have equipped me with essential skills to make data-driven decisions that improve patient care and hospital efficiency in the healthcare industry. The practical applications of calculated fields and data visualization techniques have deepened my understanding of healthcare performance metrics, making me a valuable asset in contributing to enhanced healthcare services and outcomes.
Tableau dashboard link: https://public.tableau.com/app /profile/sita4814
SQL Analysis:
Objectives worked on the given dataset:
Worked in the domain of Patient's Mortality, Vitals, Readmission,
Service and cause of death.
Triggers are created to
Raise the exception about wrong input.
Set the warning about the abnormality in vitals.
Set message on the Readmission registry flag.
Analysis of the expired patients based on different conditions like
Car Accident
Heart Failure
Analysis on the patient's visits to the hospital.
Classification of the patients by birth date.
SQL code link: sita47/HospitalData_SQLAnalysis (github.com)
Analysis:
our analysis of ambulatory visits data reveals that out of 945 patients, a significant proportion of 825 patients exhibited abnormal vital signs during their visits. Furthermore, when considering the type of admissions following general medicine, it is evident that the ICU (Intensive Care Unit) had the highest number of admissions. Lastly, our findings indicate that individuals born between the years 1971-1980 experienced the highest incidence of heart failure, followed closely by those born between 1961-1970.
Comments