Healthcare Analytics Made Simple

Techniques in healthcare computing using machine learning and Python

Vikas (Vik) Kumar

• Publisher: Packt Publishing
• Year: 2018
• Print ISBN: 9781787286702
• eBook ISBN: 9781787283220
• Edition: 1
• Topic: Computer Science
• Subtopic: Data Mining
• Index: Computer Science

Add a touch of data analytics to your healthcare systems and get insightful outcomes

Key Features

• Perform healthcare analytics with Python and SQL
• Build predictive models on real healthcare data with pandas and scikit-learn
• Use analytics to improve healthcare performance

Book Description

In recent years, machine learning technologies and analytics have been widely utilized across the healthcare sector. Healthcare Analytics Made Simple bridges the gap between practising doctors and data scientists. It equips the data scientists' work with healthcare data and allows them to gain better insight from this data in order to improve healthcare outcomes.

This book is a complete overview of machine learning for healthcare analytics, briefly describing the current healthcare landscape, machine learning algorithms, and Python and SQL programming languages. The step-by-step instructions teach you how to obtain real healthcare data and perform descriptive, predictive, and prescriptive analytics using popular Python packages such as pandas and scikit-learn. The latest research results in disease detection and healthcare image analysis are reviewed.

By the end of this book, you will understand how to use Python for healthcare data analysis, how to import, collect, clean, and refine data from electronic health record (EHR) surveys, and how to make predictive models with this data through real-world algorithms and code examples.

What you will learn

• Gain valuable insight into healthcare incentives, finances, and legislation
• Discover the connection between machine learning and healthcare processes
• Use SQL and Python to analyze data
• Measure healthcare quality and provider performance
• Identify features and attributes to build successful healthcare models
• Build predictive models using real-world healthcare data
• Become an expert in predictive modeling with structured clinical data
• See what lies ahead for healthcare analytics

Who this book is for

Healthcare Analytics Made Simple is for you if you are a developer who has a working knowledge of Python or a related programming language, although you are new to healthcare or predictive modeling with healthcare data. Clinicians interested in analytics and healthcare computing will also benefit from this book. This book can also serve as a textbook for students enrolled in an introductory course on machine learning for healthcare.

Making Predictive Models in Healthcare

This chapter is intended for all audiences and is an integral part of this book. We will demonstrate how to build predictive models for healthcare using example data and an example machine learning problem. We will preprocess the data one feature at a time. By the end of this chapter, you will understand how to prepare and fit a machine learning model to a clinical dataset.

Introduction to predictive analytics in healthcare

In Chapter 1, Introduction to Healthcare Analytics, we discussed the three subcomponents of analytics: descriptive analytics, predictive analytics, and prescriptive analytics. Predictive and prescriptive analytics form the heart of healthcare's mission to improve care, cost, and outcomes. That is because if we can predict that an adverse event is likely in the future, we can divert our scarce resources toward preventing the adverse event from occurring.

What are some of the adverse events we can predict (and then prevent) in healthcare?

• Deaths: Obviously, any death that is preventable or foreseeable should be avoided. Once a death is predicted to occur, preventative actions may include directing more nurses toward that patient, hiring more consultants for the case, or speaking to the family about options earlier rather than later.
• Adverse clinical events: These are events that are not synonymous with deaths, but highly increase the chances of morbidity and mortality. Morbidity refers to complications, while mortality refers to death. Examples of adverse clinical events include heart attacks, heart failure exacerbations, COPD exacerbations, pneumonia, and falls. Patients in which adverse events are likely could be candidates for more nursing care or for prophylactic therapies.
• Readmissions: Readmissions don't present an obvious danger to patients; however, they are costly, so preventable readmissions should, therefore, be avoided. Furthermore, readmission reduction is highly incentivized by the Centers for Medicare and Medicaid Services, as we saw in Chapter 6, Measuring Healthcare Quality. Preventative actions include assigning social workers and case managers to high-risk patients to assure that they are following up with outpatient providers and buying needed prescriptions.
• High utilization: Predicting patients who are likely to incur high amounts of medical spending again could potentially reduce costs by assigning more care members to their team and ensuring frequent outpatient check-ins and follow-ups.

Now that we've answered the "What?" question, the next question is, "How?" In other words, how do we make predictions about which care providers can act?

• First, we need data: The provider should send you their historical patient data. The data can be claims data, clinical transcripts, a dump of EHR records, or some combination of these. Whatever the type of data, it should eventually be able to be molded into a tabular format, in which each row represents a patient/encounter and each column represents a particular feature of that patient/encounter.
• Using some of the data, we train a predictive model: In Chapter 3, Machine Learning Foundations, we learned about what exactly we are doing when we train predictive models, and how the general modeling pipeline works.
• Using some of the data, we test our model's performance: Assessing the performance of our model is important for setting the expectations of the provider as to how accurate the model is.
• We then deploy the model into a production environment and provide live predictions for patients on a routine basis: At this stage, there should be a periodic flow of data from the provider to the analytics firm. The firm then responds with regularly scheduled predictions on those patients.

In the remainder of the chapter, we will go through the "How?" of building a predictive model for healthcare. First, we will describe our mock modeling task. Then, we will describe and obtain the publicly available dataset. After that, we will preprocess the dataset and train predictive models using different machine learning algorithms. Finally, we will assess the performance of our model. While we will not be using our models to make actual predictions on live data, we will describe the steps necessary for doing so.

Our modeling task – predicting discharge statuses for ED patients

Every year, millions of patients use emergency department facilities across the nation. The resources of these facilities have to be managed properly—if there is a large influx of patients at any given time, the staff and the available rooms should be increased accordingly. The mismatch between resources and patient influx could lead to wasted money and suboptimal care.

In this context, we introduce our example modeling task —predicting discharge statuses for patients presenting to the emergency room. The discharge status refers to whether patients are admitted to the hospital or sent home. Usually, the more serious cases are admitted to the hospital. Therefore, we are attempting to predict the outcome of the ED visit early on in the patient stay.

With such a model, the workflow of the hospital could be greatly improved, as well as resource flow. Many previous academic studies have looked at this problem (for an example, see Cameron et al., 2015).

You may be wondering why we didn't pick a different modeling task, such as readmission modeling or predicting CHF exacerbations. For one thing, the publicly available clinical data is very limited. The dataset that we chose is an ED dataset; there are no publicly available inpatient datasets available that are free to download without registering. Nevertheless, the task that we have chosen will serve our purposes in demonstrating how predictive healthcare models can be built.

Obtaining the dataset

In this section, we will provide step-by-step instructions for obtaining the data and its associated documentation.

The NHAMCS dataset at a glance

The dataset we have chosen for this book is part of the National Hospital Ambulatory Medical Care Survey (NHAMCS) public use data. It is survey data published and maintained by the US Center for Disease Control and Prevention (CDC). The home page for this data set is https://www.cdc.gov/nchs/ahcd/ahcd_questionnaires.htm.

• The NHAMCS data is survey-based data; it is populated by surveys sent to patients and healthcare providers that were seen in the hospital for encounters.
• The data files are in fixed-width format. In other words, they are text files in which each row is on a distinct line, and columns are each a set number of characters long. Information about the character length of each feature is available in the corresponding NHAMCS documentation.
• There are different sets of files depending on whether the data is from outpatient encounters or emergency department visits. We...

Table of contents

• Title Page
• Copyright and Credits
• Dedication
• Packt Upsell
• Foreword
• Contributors
• Preface
• Introduction to Healthcare Analytics
• Healthcare Foundations
• Machine Learning Foundations
• Computing Foundations – Databases
• Computing Foundations – Introduction to Python
• Measuring Healthcare Quality
• Making Predictive Models in Healthcare
• Healthcare Predictive Models – A Review
• The Future – Healthcare and Emerging Technologies
• Other Books You May Enjoy