To give you an idea of what healthcare analytics encompasses, here are some examples of healthcare analytics use cases that demonstrate the breadth and depth of modern healthcare analytics.

Analytics is often divided into three subcomponents–descriptive analytics, predictive analytics, and prescriptive analytics. Descriptive analytics encompasses using the analytic techniques previously discussed to better describe or summarize the process under study. Understanding how care is delivered is one process that stands to benefit from descriptive analytics.

How can we use descriptive analytics to better understand healthcare delivery? The following is one example of a visualization of a toddler's emergency department (ED) care record when they presented complaining of an asthma exacerbation (Basole et al., 2015). It uses structured clinical data commonly found in EMR systems to summarize the temporal relationships of the care events they experienced in the ED. The visualization consists of four types of activities–administrative (yellow), diagnostic (green), medications (blue), and lab tests (red). These are encoded by color and by y-position. Along the x-axis is time. The black bar on top is divided by vertical tick marks into hour-long blocks. This patient's visit lasted a little over two hours. Information about the patient is displayed before the black time bar.

While descriptive analytical studies such as these may not directly impact costs or medical care recommendations, they serve as a starting point for exploring and understanding the patient care and often pave the way for more specific and actionable analytical methods to be launched:

A central problem in medicine is identifying patients who are at risk of developing a certain disease. By identifying high-risk patients, steps can be taken to hinder or delay the onset of the disease or prevent it altogether. This is an example of predictive analytics at work–using information from previous events to make predictions about the future. There are certain diseases that are particularly popular for prediction research: congestive heart failure, myocardial infarction, pneumonia, and chronic obstructive pulmonary disease are just a few examples of high-mortality, high-cost diseases that benefit from early identification of high-risk patients.

Not only do we care about what diseases will occur in the future, we are also interested in identifying patients who are at risk of requiring high-cost treatments, such as hospital readmissions and doctor visits. By identifying these patients, we can take money-saving steps proactively to reduce the risk of these high-risk treatments, and we can also reward healthcare organizations that do a good job.

This is a broad example with several unknowns to consider. First: what specific event (or disease) are we interested in predicting? Second: what data will we use to make our predictions? Structured clinical data (data organized as tables) drawn from electronic medical records is currently the most popular data source; other possibilities include unstructured data (medical text), medical or x-ray images, biosignals (EEG, EKG), data recorded from devices, or even data from social media. Third: what machine learning algorithm will we use?

While making nice visualizations or predictions represent the sexier aspects of healthcare analytics, there are other types of analytics that are also important. Sometimes, it boils down to good, old number crunching. Monitoring the performance of physicians and healthcare organizations using healthcare measures is a good example of this type of analytical technique. Healthcare measures provide a mechanism by which individuals can measure and compare the compliance of participating providers on evidence-based medical recommendations. For example, it is a widely accepted recommendation that patients with diabetes receive foot exams to detect diabetic foot ulcers every three months by a physician.

A state-sponsored healthcare measure might specify guidelines for calculating the number of diabetic patients receiving care at an institution, and then determine the percentage of those patients that received appropriate foot care. Similar measures would exist for the common heart, lung, and joint diseases, among many others. This provides a way to identify the providers that provide the highest quality care, and these recommendations can be downloaded for public consumption. We will discuss specific healthcare measures in Chapter 6, Measuring Healthcare Quality.

In rare cases, healthcare analytics comprise medical technologies that are used to actually treat diseases, not just perform research on them. An example of this is neuroprosthetics. Neuroprosthetics can be defined as the enhancement of nervous system function using man-made devices. Neuroprosthetics research has enabled patients with disabilities such as blindness or paraplegia to recover some of their lost function. For example, a paralyzed patient may be able to move a computer cursor on a screen not with their hand, but by using their brain signals! In this specific application, recordings of the electrical activity of specific neurons are obtained, and a machine learning model is used to determine in which direction the cursor should move given the firing of the neurons. Similar analytics can be used for visual impairments, or for visualizing what a human is seeing. A second example includes implanting devices in the body that detect seizures before they occur and proactively administer preventive medication. Certainly, the sky is the limit for analytic-driven treatments.