Technical level: Low (includes some machine learning algorithm formulas as examples)
Who is this book intended for: Not data scientists. The technical content on machine learning & associated is high-level and at the primer level, although it does cover many different areas. Healthcare information and examples given are extremely high level, and include many of the boilerplate, speculative analytic applications: VR, image processing, simulation. Possibly individuals new to data science or completely new to healthcare would find the content valuable.
Chapter summaries:
1. What is AI?
To it's credit, this book does not excessively over-hype AI, and also includes some discussion on giving the author's definition of 'data science' and 'machine learning'. Approximately 5 pages out of 18 on healthcare.
2. Data
Primer level discussion on types of data and attributes of those data types. Also, some discussion on 'big data' as a specific type of data vs small data. Includes discussion on larger topics such as policies and data governance, ethics and stewarship, etc. Approximately 6 pages out of 52 specifically address healthcare use cases for data including: wait times/scheduling, readmissions, utilizing EHR data, value-based care, healthcare IoT, evidence-based medicine, and public health.
3. Machine Learning
General introduction to machine learning as a concept. 0 out of 41 pages on healthcare.
4. Machine Learning Algorithms
Very high-level introduction to a wide-variety of common algorithms. Typically 1-2 paragraph discussion of each algorithm with some diagrams and formulas to assist in explaining. Includes approximately 12 pages on NLP as a concept & specific methods. This chapter does include a health care example (focused on diabetes) for approximately 2 out of 67 pages.
5. Evaluating Learning for Intelligence
Additional information on fitting, tuning, and evaluating machine learning results. This chapter is brief (17 pages) but at least includes some important caveats for a non-technical or inexperienced audience (e.g. causation vs correlation).
6. Ethics of Intelligence
Approximately 46 pages of ethical implications of the use of data and the practice of data science. A fairly wide variety of topics are covered, none comprehensively, but again with the goal of introducing novels concept to an audience it's nice to see this topic taken seriously and given some discussion.
7. Future of Healthcare
The real dedicated chapter on healthcare in this book is 48 pages and covers approximately 43 topics. Don't expect to get up to speed on how to actually do healthcare data science here with each topic typically getting a few paragraphs. The sections that get the most attention are 'connected medicine' (including discussion of virtual assistants and remote monitoring) at 7 pages, blockchain (naturally) at 3 pages, and VR/AR at 3 pages.
8. Case Studies
This chapter highlights 5 different actual use cases of some type of 'machine learning' in healthcare. Use cases were solicited on twitter with these 5 chosen based on a set of criteria (1 case study does appear to be from a close member of the author's own digital diabetes group). Several of these case studies appear to be academic study results, even listed in scientific paper format; one case study appears to be from the published results of a RCT. Some of these vignettes may be interesting (particularly if the reader is interested in diabetes), but I'm never exactly sure of what the reader is intended to take-away from this type of high-level vignette, short of pure inspiration around a concept or high-level application.
