LIST OF FIGURES

FIGURE
NUMBER CONTENT PAGE.NO
5.1 OUT PATIENT DATA 8

5.2 IN PATIENT DATA 8

5.3 9
ALL DATA

5.4 MAPPING SECURITY 9

5.5 GROUPING 10

5.6 11
CARDS
5.7 11
TOGGLE BUTTON
5.8 12
SUMMARY DSHBOARD

5.9 DETAIL PAGE 13

 ABSTRACT

This project examines healthcare waitlist data to uncover trends and insights crucial for
optimizing hospital management and improving patient care. Analyzing data from January
2018 to March 2021, the goal was to understand variations in waitlist durations across medical
specialties, patient types, and age groups.

The analysis began with data preparation, including cleaning, standardizing, and merging
multiple CSV files representing Inpatient and Outpatient data. This process involved renaming
fields for consistency, addressing formatting issues, and creating a unified dataset.

Using Power BI, we developed two key analytical pages: a Summary page and a Detailed page.
The Summary page features dynamic metrics and visualizations, such as comparisons of
current versus previous month waitlists, and toggle buttons for average versus median wait
times. It includes visualizations like donut charts for case type distribution, stacked column
charts for time band and age profile relationships, and line charts showing trends over time.
The Detailed page provides a matrix view of waitlists segmented by case type, age profile, time
band, specialty, and archive date.

Key findings revealed a significant increase in waitlist numbers, particularly among

outpatients, with the highest wait times observed in specialties like Accident & Emergency,
Dermatology, Clinical Genetics, Cardiology, and Pain Relief. Seasonal patterns and variations
in wait times were also identified.

The project demonstrates the effectiveness of data analytics in understanding healthcare

waitlist dynamics, offering actionable insights for better resource allocation and improved
patient satisfaction. By leveraging these insights, hospital management can enhance
operational efficiency and make informed decisions to address high wait times and optimize
patient care.
 CHAPTER - 1
INTRODUCTION

This project investigates patient waitlist data from 2018 to 2021, focusing on three patient
categories: Outpatient, Inpatient, and Day Case. Outpatient patients receive care and return
home the same day, Inpatient patients stay in the hospital for one or more nights for serious
treatment, and Day Case patients require more extensive care than Outpatients but are
discharged on the same day.

The project aims to track the current status of the patient waitlist, analyze historical trends by
patient category, and perform detailed analyses of specialties and age profiles. The analysis
begins with a Summary dashboard, designed to provide an overview of key metrics such as
month-to-month waitlist comparisons and trends. For example, the Summary dashboard
highlights a significant 26% increase in Outpatient waitlists over the past three years, while
Inpatient and Day Case waitlists have remained relatively stable. Additionally, it identifies the
Top 10 specialties with the longest wait times.

To facilitate deeper exploration, a Detailed View page was created. This interactive page allows
users to filter data by case type, specialty, age, and time band, providing a granular analysis
that supports more informed decision-making. This comprehensive approach aims to enhance
understanding of waitlist dynamics and improve hospital resource management and patient
care.
 CHAPTER - 2
PROBLEM STATEMENT

The aim of this project is to enhance the management of patient waitlists by analyzing data
from 2018 to 2021. Specifically, the project seeks to track current waitlist figures and compare
them to historical data to identify changes and trends. It also aims to analyze historical monthly
trends across Outpatient, Inpatient, and Day Case categories to uncover patterns in wait times.
Additionally, the project will perform a detailed analysis of wait times by medical specialty
and age group to highlight areas with significant delays and high-demand specialties. The goal
is to provide actionable insights that will support better resource allocation, improve patient
management, and optimize overall hospital efficiency.
 CHAPTER - 3
TECHNOLOGY ADOPTED
3.1. Power BI
Power BI is utilized in this project as the primary tool for analyzing patient waitlist data,
focusing on the Outpatient, Inpatient, and Day Case categories. The tool provides the ability to
create interactive dashboards and reports that allow for comprehensive monitoring of the
current waitlist status. It also facilitates in-depth analysis of historical trends, enabling a better
understanding of how wait times and patient demographics have evolved over time.
Additionally, Power BI supports detailed specialty and age profile analysis, allowing for
targeted insights into specific patient groups.

3.2. Data Visualization

Data visualization in Power BI plays a critical role in transforming raw patient waitlist data
into actionable insights. Through the use of bar charts, line graphs, and pie charts, complex
data is presented in a visually intuitive manner. These visualizations are essential for
identifying trends, distributions, and patterns across the dataset, which helps in highlighting
key areas of focus, such as peak wait times or specialty-specific bottlenecks. This visualization
process supports data-driven decision-making by providing clear and concise representations
of the underlying data.

3.3. Data Integration

Power BI’s data integration capabilities are leveraged to import and consolidate data from
various sources, including SQL databases and Excel files. This integration ensures that the
analysis is based on accurate, up-to-date information. During the data integration process, the
data is cleaned and transformed to align with the project’s analytical requirements. By
integrating multiple data sources, Power BI enables a comprehensive analysis of patient waitlist
information, ensuring that all relevant data is considered in the final reports.

3.4. Model Building

3.4.1. Data Preparation:
Data preparation is a crucial step in this project. The patient waitlist dataset is thoroughly
cleaned and transformed to address issues such as missing values and inconsistencies. The data
is then segmented into relevant categories—Outpatient, Inpatient, and Day Case—to allow for
focused analysis. Calculated fields are created within Power BI to enable the analysis of key
metrics, such as average wait times and demographic distributions, which are vital for
understanding the dynamics of the waitlist data.

3.4.2 Building Visual Models:

Visual models are constructed in Power BI by establishing relationships between data tables
and creating measures that facilitate in-depth analysis. For example, calculated columns are
used to derive essential metrics, while relationships between tables help in analyzing specialty-
specific waitlists. These models provide a foundation for the comprehensive analysis of patient
waitlist data, allowing for detailed reporting on various aspects of the data, such as wait times,
demographics, and specialty distribution.

3.5. Analysis Techniques

3.5.1. Trend Analysis:
Trend analysis is conducted using Power BI’s time series analysis features. Line charts and
trend lines are employed to visualize historical changes in waitlist data, enabling the
identification of peak periods and patterns in patient wait times. This analysis helps in
understanding how waitlist dynamics have shifted over time, providing valuable insights into
the factors influencing wait times.

3.5.2. Demographic Analysis:

Demographic analysis is performed using Power BI’s visualizations to analyze the age profiles
and specialty distributions of patients on the waitlist. Pie charts and bar charts are utilized to
represent the demographic distribution, helping to identify variations in wait times and patient
flow across different age groups and medical specialties.

3.6. Model Evaluation

Model evaluation is carried out through interactive dashboards in Power BI. Key performance
indicators (KPIs) and metrics are used to assess the accuracy and completeness of the data.
These dashboards provide real-time insights into the current patient waitlist status, allowing
stakeholders to make informed decisions based on the most recent data. Comparative analysis
features in Power BI enable the evaluation of different strategies.
 CHAPTER - 4
DETAILS OF TOOLS USED

4.1 Power BI

Power BI is a business analytics tool used for interactive data visualization and business
intelligence. For this project, Power BI was essential for loading, transforming, and visualizing
patient waitlist data. The tool's DAX functions were crucial for calculating dynamic metrics
and enabling toggle features for comparing average and median waitlist times. Its user-friendly
interface and strong data modeling capabilities made it the primary tool for building both
summary and detailed dashboards.

4.2 Python

Python was employed for initial data exploration and pre-processing tasks. Libraries like
Pandas and NumPy facilitated efficient manipulation and transformation of the CSV files
containing patient waitlist data. Python’s flexibility in handling large datasets and integrating
with various data sources was key to preparing the data for analysis in Power BI.

4.3 Excel

Excel was used for quick data checks and calculations. It also served as the storage medium
for the Mapping Specialty file, which was later imported into Power BI to create relationships
between specialties and their respective groups. Excel’s simplicity and familiarity made it a
convenient tool for managing reference data and performing initial data validations.

4.4 DAX (Data Analysis Expressions)

DAX is a formula language in Power BI used to create custom calculations. In this project,
DAX was used to create dynamic measures like Last Month Waitlist and Same Month Last
Year Waitlist. It also enabled the development of interactive features, such as toggle buttons
for switching between average and median waitlist views, enhancing the dashboard’s
functionality.
4.5 Web Scraping

Web scraping was used to collect relevant datasets from online sources. Tools and libraries
such as Beautiful Soup and Scrapy were utilized to extract data from websites and online
databases. The collected data was integrated with existing datasets, providing a comprehensive
view of patient waitlists and supporting more accurate analysis.
 CHAPTER - 5
SOLUTION OF THE PROBLEM

5.1 Dataset
• Outpatient: A patient attending for a short consultation or appointment.
• Inpatient: A patient admitted to stay in the hospital for one or more nights.
• Day Case: A patient undergoing a procedure that allows them to leave the same day.
• Archive_Date: The date of record.
• Specialty_HIPE: The specialty code.
• Specialty: The medical specialty being waited for.
• Case_Type: The classification of the patient (Inpatient/Day Case/Outpatient).
• Age_Profile: The age group of the person on the waitlist.
• Time_Bands: The waitlist time intervals.

5.2 Data Preprocessing

1. Data Unification:
o Inpatient Data: Four CSV files containing data for Inpatients and Day Cases.
o Outpatient Data: Four CSV files containing data for Outpatients.
o All CSV files within each patient type category (Inpatient and Outpatient) have
the same structure, allowing them to be appended directly.
o Minor differences exist between Inpatient and Outpatient data structures,
necessitating alignment before combining.

2. Data Cleaning and Transformation

o Field Renaming: The field "Specialty" in the Outpatient data was renamed to
"Specialty_Name" to match the Inpatient data.
o New Column Addition: A new column, "Case_Type," was added to the
Outpatient data with all rows populated as "Outpatient."
o Date Formatting: The Archive_Date field was converted from text to a date
format.
o Standardizing Time Bands and Age Profiles: Removed extra spaces and
inconsistencies in the Time_Bands and Age_Profile fields.
 3. Data Transformation
o Appending Data: Once the data was cleaned and standardized, the files were
appended together using Power BI to create a unified dataset named All_data.
o Ensuring Consistency: The transformation process ensured that all data across
the years and patient types were consistent and ready for analysis.

Fig 5.1 (Outpatient Data)

Fig 5.2(Inpatient Data)

5.3 Final Dataset Structure

• Patient Type: Differentiating between Inpatient, Day Case, and Outpatient.

• Specialty Information: Using the Specialty_Name field.
• Case_Type: A newly added column to Outpatient data, matching the structure of
Inpatient data.
• Archive Date: Properly formatted as a date.
• Time Bands and Age Profiles: Standardized and cleaned for accurate analyse.
 Fig 5.3 (All Data)

Fig 5.4 (Mapping Security)

We created a relationship between our All_Data table and our Mapping_Specialty data so that
we can refer to the specialty groupings when we do our analysis.

Fig 5.5 (Grouping)

5.4 Analysing the Data

1. Summary Page: This page will include dynamic cards for current vs. previous month
waitlists, Average vs. Median toggle buttons, and visualizations such as a donut chart
for case types, a stacked column chart for time bands vs. age profiles, and line charts
for Inpatients and Outpatients. A multi-row card will also display the top 5 specialties
based on waitlist metrics.
2. Detailed Page: This page will feature a matrix visualization to provide a detailed
breakdown of the waitlist data, including dimensions like Case_Type, Age_Profile,
Time_Band, Specialty, and Archive_Date, offering a comprehensive view of the data.
 5.5 Data Visualization

1. Total Current Month Waitlist vs. Total Previous Month Waitlist:

We created dynamic DAX measures to compare the current month's waitlist with the
previous month and the same month from the previous year. These measures allow for real-
time updates on the Power BI dashboard, ensuring accurate comparisons in the visualizations.

Fig 5.6 (Cards)

2. Average vs. Median Toggle button:

To address data outliers, we created toggle buttons that allow users to switch between
viewing the average and median waitlist values in the charts. We added slicers for "Average"
and "Median" and used DAX measures to calculate both. A final measure combines the two,
enabling dynamic chart updates based on the selected toggle option. This ensures flexible and
accurate data visualization.

Fig 5.7 (Toggle Button)

DASHBOARD PRESENTATION:

Fig 5.8 (Summary Page)

The waitlist has grown from 640K last year to 709K this year, mostly affecting outpatients.
The longest waits are 0-3 months and 18+ months. Inpatient and Day Case waitlists are stable,
while outpatient numbers continue to rise. The top 5 specialties with the highest wait times are
Accident & Emergency, Dermatology, Clinical Genetics, Cardiology, and Pain Relief.
 Fig 5.9 (Detail Page)

The Detailed page will include a matrix that shows the total patient waitlist, organized by
Case_Type, Age_Profile, Time_Band, Specialty, and Archive_Date. This setup will allow us
to analyse trends and patterns in the data across different categories, providing a clear and
detailed view of the waitlist.
 CHAPTER - 6

CONCLUSION

This project provided a detailed analysis of patient waitlist data from 2018 to
2021, utilizing Power BI and Python to convert raw data into actionable insights.
Through comprehensive data preprocessing, including feature scaling, date
handling, and feature-target splitting, the project ensured high data integrity. The
visualizations in Power BI offered interactive dashboards with features like
average vs. median toggles and specialty-specific charts, enabling in-depth
analysis of waitlist trends across Outpatient, Inpatient, and Day Case categories.
Additionally, predictive and classification models allowed for future trend
forecasting, aiding healthcare management in resource allocation and strategic
planning. This work highlights the potential of data-driven approaches to
optimize patient care and enhance healthcare system efficiency.
 CHAPTER - 7

FUTURE ENHANCEMENTS

To elevate this project, real-time data integration could be introduced for

continuous monitoring and swift adjustments to patient waitlists. Advanced
machine learning techniques, such as time series forecasting and clustering, could
enhance predictive accuracy and identify potential bottlenecks. Expanding the
dataset to include metrics like patient outcomes and resource utilization would
offer a more comprehensive analysis of healthcare efficiency. Enhancing Power
BI visualizations with interactive features, and integrating the system with
Electronic Health Records (EHR) and Hospital Information Systems (HIS),
would streamline data flow and improve decision-making. Additionally, applying
Natural Language Processing (NLP) to unstructured data, such as patient
feedback, could provide deeper insights and enrich the overall analysis.
 REFERENCE

1. Microsoft Corporation. (n.d.). Power BI Documentation. Retrieved from

https://docs.microsoft.com/en-us/power-bi/
2. LinkedIn Corporation. (n.d.). LinkedIn. Retrieved from https://www.linkedin.com
3. Microsoft Corporation. (n.d.). Microsoft Excel Documentation. Retrieved from
https://support.microsoft.com/en-us/excel
4. Giller, A., & Tredinnick, M. (2018). Power BI: An Introduction. Apress. ISBN: 978-
1484232348.
5. Power BI Community. (n.d.). Creating Dashboards in Power BI. Retrieved from
https://community.powerbi.com/t5/Dashboard-Design/bd-p/DashboardDesign
6. Python Software Foundation. (n.d.). Python Programming Language. Retrieved from
https://www.python.org
 APPENDIX

We have to use DAX and create new measures:

• DAX measure for last month's waitlist:

Last Month Waitlist = CALCULATE(SUM(All_Data[Total]),
All_Data[Archive_Date]=MAX(All_Data[Archive_Date])) + 0

• DAX measure for same month, last year's wait list:

PY Latest Month Waitlist = CALCULATE(SUM(All_Data[Total]),
All_Data[Archive_Date]=EDATE(MAX(All_Data[Archive_Date]), -12)) + 0

Average Waitlist = AVERAGE(All_Data[Total])

Median Waitlist = Median(All_Data[Total])

Now create one more measure that will allow us to interact with the slicer buttons:

Avg/Med Waitlist = SWITCH(VALUES('Calculation Method'[Calc Method]), "Average",

[Average Waitlist], "Median", [Median Waitlist])
 SAMPLE DATA:

INPATIENT DATASET

Patient_id Admission_date Discharge Specialty Waiting Patient type Age Gender Severity_
_date days level

0001 2021-01-10 2021-01- Cardiology 10 Emergency 55 M High

20

0002 2021-02-25 2021-03- Routine 45 F Medium

03 Orthopaedics 10
0003 2021-03-05 2021-03- Emergency 10 F High
15 Neurology 9
0004 2021-04-01 2021-04- General 9 Routine 35 M Low
10
0005 2021-05-20 2021-05- Cardiology 10 50 F High
Emergency
30
0006 2021-06-10 2021-06- Orthopedics 12 Routine 47 M Medium
20

0007 2021-07-15 2021-07- 14 Emergency 53 F High

Neurology
25

OUTPATIENT DATASET

Patient_id Appointment Specialty Waiting Patient type Age Gender Severity_

date days level

1001
2021-01-15 Dermatology 15 Routine 40 M Medium
1002
2021-02-10 Paediatrics 20 Routine 30 F Low
1003
2021-03-20 Orthopaedics 12 Emergency 55 M High
1004
2021-04-10 Neurology 25 Routine 60 F Medium
1005
2021-05-25 General 10 Emergency 50 M Low
1006
2021-06-15 Dermatology 18 Routine 35 F High

1007
2021-07-05 Pediatrics 22 Emergency 28 M Medium
 DAY CASE DATA

Patient_id Procedure date Specialty Waiting Patient type Age Gender Severity_
days level

2001 2021-01-20 Cardiology 8 Routine 45 F Medium

2002
2021-02-05 orthopaedics 10 emergency 50 M High

2003 2021-03-12 Neurology 5 Routine 55 F Low

2004
2021-02-05 Orthopaedics 10 Emergency 50 M Medium

2005 2021-05-30 Dermatology 12 Emergency 35 F High

2006 2021-06-10 Cardiology 9 Routine 48 M Medium

2007 2021-07-15 Orthopedics 11 Emergency 53 F High