Now that we've defined and introduced healthcare analytics, it's important to give some background on the knowledge from which it draws. Healthcare analytics can be viewed as the intersection of three fields: healthcare (Healthcare Analytics), mathematics (Math), and computer science (CS), as seen in the following diagram. Let's explore each of these three areas in turn:

Healthcare is the domain-knowledge pillar of healthcare analytics. Here are some of the significant healthcare areas of knowledge that comprise healthcare analytics:

• Healthcare delivery and policy: An understanding of how the healthcare industry is structured, who the major players in healthcare are, and where the financial incentives lie can only help us in improving healthcare analytics endeavors.
• Healthcare data: Healthcare data is rich and complex, whether it is structured or unstructured. However, healthcare data collection often follows a specific template. Knowing the details of the typical history and physical examination (H&P) and how data is organized in a medical chart goes a long way in helping us turn that data into knowledge.
• Clinical science: A familiarity with medical terminology and diseases helps in knowing what's important in the vast ocean of medical information. Clinical science is commonly divided into two areas: physiology, or how the human body functions normally, and pathology, or how the human body functions with a disease. Some basic knowledge of both can be very helpful in doing effective healthcare analytics.

An introduction to healthcare for healthcare analytics will be provided in Chapter 2, Healthcare Foundations.

The second pillar of our healthcare analytics triumvirate is mathematics. We are not trying to scare you off with this list; a detailed knowledge of all of the following areas is not a prerequisite for doing effective healthcare analytics. A basic knowledge of high school math, however, may be essential. The other areas are most helpful while understanding the machine learning models that allow us to predict diseases. That being said, here are some of the significant mathematical domains that comprise healthcare analytics:

• High school mathematics: Subjects such as algebra, linear equations, and precalculus are essential foundations for the more advanced math topics seen in healthcare analytics.
• Probability and statistics: Believe it or not, every medical student takes a class in biostatistics during their training. Yes, effective medical diagnosis and treatment rely heavily on probability and statistics, including concepts such as sensitivity, specificity, and likelihood ratios.
• Linear algebra: Commonly, the operations done on healthcare data while making machine learning models are vector and matrix operations. You'll effectively perform plenty of these operations as you work with NumPy and scikit-learn to make machine learning models in Python.
• Calculus and optimization: These last two topics particularly apply to neural networks and deep learning, a specific type of machine learning that consists of layers of both linear and nonlinear transformations of data. Calculus and optimization are important for understanding for how these models are trained.

An introduction to mathematics and machine learning for healthcare analytics will be provided in Chapter 3, Machine Learning Foundations.

Here are some of the significant computer science domains that comprise healthcare analytics:

• Artificial intelligence: At the center of healthcare analytics is artificial intelligence or the study of systems that interact with their environment. Machine learning is a subarea within artificial intelligence, in which predictions are made about future events using information from previous events. The models that we will study in the later parts of this book are machine learning models.
• Databases and information management: Healthcare data is often accessed using relational databases, which can often be dumped by electronic medical record (EMR) systems on demand, or which are located in the cloud. SQL (short for Structured Query Language) can be used to select the specific data in which we are interested and to make transformations on that data.
• Programming languages: A programming language provides an interface between the human programmer and the ones and zeros inside of a computer. A programming language allows a programmer to provide instructions to the computer to make calculations on data that humans cannot practically do. In this book, we will use Python, a popular and emerging programming language that is open source, comprehensive, and features plenty of machine learning libraries.
• Software engineering: Many of you are presumably learning about healthcare analytics because you are interested in deploying production-grade healthcare applications in your workplace. Software engineering is the study of the effective and efficient building of software systems that satisfy user and customer requirements.
• Human-computer interaction: The end users of healthcare analytics applications usually don't use programming to obtain their results, but instead rely on visual interfaces. Human-computer interaction is the study of how humans interact with computers and how such interfaces can be designed. A current hot topic in medicine is how EMR applications can be made more intuitive and palatable to physicians, rather than increasing the number of mouse clicks they must make per patient while writing notes.

Computer science is so pervasive in healthcare analytics that almost every chapter in this book deals with it.