Hospital Readmission Analysis and Plan

#webdev #excel #ai #healthcare

Healthcare costs are significantly increased by hospital readmissions and longer lengths of stay. My goal was to analyze healthcare data in order to uncover trends, patterns and relationships between patient data during their hospital stay as a way to predict risk of readmission or prolonged hospital length of stay, then use my findings to create an actionable plan to reduce length of stay and readmission risk.
I found a publicly available data set from Kaggle involving 25,000 hospital patients. Used Microsoft Excel to organize the data into subsections and then analyze it. Utilized pivot tables and graphs to visualize the results. Prompted chat GPT to create an organized report of key findings and actionable plan to increase safe discharges and reduce readmission rates and lengths of stay.
Below is my report, along with links for Excel tables and graphs, as well as discharge checklist:

Report
Analyzing Readmission Risk and Recommendations for Safer Discharges

Purpose and Executive Summary
Hospital readmissions and longer lengths of stay significantly increase healthcare costs. This report analyzes readmission risk based on a dataset from Kaggle of 25,000 patients, with an overall readmission rate of 47%. Key factors such as length of stay (LOS), age, primary diagnosis, admitting physician specialty, and medication count were identified as predictors of readmission. The report provides actionable recommendations, including enhanced discharge planning, tailored interventions for high-risk groups, and improved medication management, to reduce readmissions and improve patient outcomes.
Introduction
Background
Hospital readmissions are a significant indicator of healthcare quality and efficiency. This analysis aims to understand the factors contributing to readmission risk and propose strategies to mitigate these risks. The dataset includes 25,000 patients, with a mean LOS of 4.45 days and a 47% overall readmission rate.
Objectives
Identify individual and combined factors influencing hospital readmissions.
Use insights to guide the development of a discharge planning committee and strategies to reduce readmissions.
Key Insights
3.1 Length of Stay (LOS)
Findings: Readmission risk is lowest on day 1, increases steadily, peaking at 7-10 days (60-65% risk), and slightly decreases for LOS of 13-14 days (45% risk).
Implications: Prolonged stays signal potential complications, while slightly longer stays may allow for improved recovery and discharge readiness.
Recommendation: Focus on discharge readiness reviews and targeted interventions for patients with LOS between 7-10 days.
3.2 Age Group
Findings: Patients aged 70-89 exhibit the highest readmission rates due to challenges such as comorbidities, limited resilience, and insufficient post-discharge support.
Implications: Elderly patients are a high-risk group requiring enhanced follow-up care and tailored discharge protocols.
Recommendation: Introduce age-specific discharge plans, including caregiver involvement and telehealth follow-ups.
3.3 Primary Diagnosis
Findings: Patients with diabetes (54% readmission rate) and respiratory conditions (49%) are at the highest risk.
Implications: These chronic conditions require comprehensive management and clear post-discharge instructions.
Recommendation: Provide disease-specific education and assign specialized care teams to manage high-risk diagnoses.
3.4 Admitting Physician Specialty
Findings: Emergency/Trauma and Family/General Practice specialties coincide with higher readmission risks.
Implications: These specialties often manage acute or complex cases, increasing the likelihood of readmission.
Recommendation: Foster collaboration between admitting physicians and multidisciplinary teams for improved discharge planning.
3.5 Medication Management
Findings: Patients with 1-12 medications have low risk, but risk increases sharply for 13-27 medications (50-55%) and decreases for 28+ medications.
Implications: High medication counts may indicate treatment complexity or adherence challenges.
Recommendation: Conduct detailed medication reviews, simplify regimens, and provide patient education tools.
Combined Analysis of Factors
4.1 LOS and Age Group
Finding: Older patients (70-89) with prolonged stays face disproportionately higher readmission risks.
Recommendation: Combine age- and LOS-specific interventions, such as extended inpatient care and tailored post-discharge support.
4.2 Admitting Physician and Primary Diagnosis
Finding: Diabetes and respiratory conditions are common among patients admitted via Emergency/Trauma.
Recommendation: Develop diagnosis-specific care pathways, ensuring close follow-ups for high-risk conditions.
4.3 Medications and Diagnosis
Finding: High medication counts (13-27) are often associated with chronic conditions like diabetes.
Recommendation: Provide additional medication counseling and simplify regimens for these patients.
4.4 LOS and Medications
Finding: Longer stays correlate with higher medication counts, compounding readmission risk.
Recommendation: Streamline medication plans during the hospital stay and align them with discharge goals.
Recommendations
5.1 Safe Discharge Planning
Implement a structured discharge checklist addressing high-risk factors like LOS, age, and diagnosis.
Assign care coordinators for patients flagged as high-risk.
5.2 Enhanced Follow-Up Programs
Conduct follow-up calls within 48 hours of discharge for high-risk patients.
Use telehealth services to address post-discharge concerns.
5.3 Multidisciplinary Collaboration
Form a discharge planning committee with physicians, nurses, pharmacists, and social workers.
Tailor interventions based on specialty-specific and diagnosis-specific risk factors.
5.4 Medication Management
Simplify complex regimens and ensure patients understand their medication plans.
Use digital tools for reminders and adherence tracking.
Conclusion
This analysis highlights the multifactorial nature of hospital readmissions and the need for targeted interventions to address high-risk factors. By focusing on LOS, age, diagnosis, and medication management, hospitals can significantly reduce readmission rates and improve patient outcomes. Implementing structured discharge planning and follow-up processes, supported by multidisciplinary teams, is essential for achieving these goals. In order to implement our findings we have also designed a checklist to be used by hospital case managers for following and discharging patients
Appendices
Discharge checklist
Excel visualizations
Excel data tables summarizing key metrics
Original dataset