STAT 451 Project Proposal

Group 19

Analyzing Healthcare Service Utilization by Age, Sex, and Payer Type

Objective:

The goal of this project is to investigate patterns in healthcare service usage by different age groups and genders across multiple payer types (Medicaid, Medicare, Out-of-Pocket, Private Health Insurance, and Other Payers). By analyzing trends over time, this project aims to identify key demographic patterns in service utilization and assess the impact of payer types on access to services such as Dental Services, Durable Medical Equipment, Home Health Care, Hospital Care, and Physician Services.

Research Questions:

How does healthcare service utilization vary across age groups and between genders?

Variables:

Payer: Types of healthcare payers (Medicaid, Medicare, Out-of-Pocket, Private Health Insurance, and Others)

Service: Type of healthcare service utilized (e.g., Dental Services, Home Health Care, Hospital Care)

Age Group: Categories by age, such as 0-18, 19-44, 45-64, 65-84, and 85+

Sex: Gender category (Males, Females, or Total)

Years: Time period from 2002 to 2020, showing service utilization counts per year

Method:

Step 1: use python to generate a bar charts to compare healthcare service usage across different age groups and genders, also adding a separate bar for each gender within each age group to better distinguish usage differences

Step 2: Use Chi-square test to determine if there are significant associations between gender, age group, and healthcare utilization.

Step 3 (Possible option): use regression for in-depth analysis to deeper view how demographic factors affect utilization quantitatively, using a regression model

(Possible option): Use logistic regression to analyze and quantify the impact of age groups and gender on the likelihood of using specific healthcare services. For example, using logistic regression to examine whether males aged 65 and above are more likely than females to use home health care services, or to estimate the probability of service usage for specific genders and age groups in the coming year or a designated future period.

(Possible option)：Use clustering analysis (such as K-Means) to identify similarities in healthcare service utilization patterns across different age groups, genders, and service types. By using variables such as "age group," "gender," and "annual utilization" as input features for clustering, we want to find groups with similar usage patterns without being limited to predefined age or gender categories.

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

# Load the dataset
data = pd.read_csv('age and Sex.csv')

# Reshape the dataset from wide to long format
data_long = pd.melt(data, id_vars=['Payer', 'Service', 'Age Group', 'Sex'], 
                    var_name='Year', value_name='Utilization')

# Convert the 'Year' column to numeric for plotting
data_long['Year'] = data_long['Year'].astype(int)

plt.figure(figsize=(12, 6))
sns.lineplot(x='Year', y='Utilization', hue='Age Group', data=data_long)
plt.title("Healthcare Service Utilization Over Time by Age Group")
plt.xlabel("Year")
plt.ylabel("Service Utilization")
plt.legend(title="Age Group")

# Set x-axis ticks to be integers
plt.xticks(data_long['Year'].unique())

plt.show()