Information Retrieval: A Biomedical and Health Perspective

Citation

Hersh, W. (2020). Information Retrieval: A Biomedical and Health Perspective (4th ed.). Springer.

Chapter Summary

The fourth edition is an extensively revised text of about 410 pages in Springer's Health Informatics series, organised as eight chapters. The publisher describes it as covering the theory, implementation and evaluation of information retrieval systems in the biomedical and health domain, with chapters on theory, practical applications, evaluation and research directions.

Chapter 1: Foundations

The opening chapter (about 40 pages) is the book's foundation: the seven chapters that follow are titled for the parts of a retrieval system and its use, and this chapter comes before all of them.

Chapter 2: Information

At roughly 100 pages this is the longest chapter, about a quarter of the book. Its title names the information that retrieval systems are built to serve, and its length makes it a main block of reading rather than a preamble.

Chapter 3: Content

This chapter (about 40 pages) turns from information in general to content. It sits between the chapter on information and the chapters on indexing and retrieval.

Chapter 4: Indexing

Indexing (about 40 pages) is the first chapter named for a step in the retrieval process itself and comes immediately before the chapter on retrieval.

Chapter 5: Retrieval

The retrieval chapter (about 35 pages) shares its name with the book's subject. It follows indexing directly, and the two are best read as a pair.

Chapter 6: Access

A shorter chapter (about 27 pages) on access. The publisher's description stresses use by clinicians, consumers, researchers and others, and this is the chapter whose title points to the user side of retrieval.

Chapter 7: Evaluation

Evaluation (about 47 pages) is the chapter given to judging retrieval systems. The publisher lists evaluation, together with theory, implementation and research directions, among the book's main concerns.

Chapter 8: Research

The final chapter (about 70 pages, the second longest) looks at where applications and research systems are heading. The publisher stresses this direction of travel across a range of areas, including use by clinicians, consumers and researchers. The fourth edition also carries a short published correction to the retrieval chapter.

Key Concepts

Chapters 1–3: Foundations; Information; Content

• Biomedical and health information retrieval: finding relevant material in the biomedical and health domain, the subject of the whole book.
• Information and content: the two chapters after the foundations treat information and content as separate topics.

Chapters 4–6: Indexing; Retrieval; Access

• Indexing: a step in the retrieval process, treated in its own chapter.
• Retrieval: the searching that the book is named for.
• Access: the user side of retrieval systems, covered in its own chapter.

Chapters 7–8: Evaluation; Research

• Evaluation: judging retrieval systems, named in the publisher's description alongside theory and implementation.
• Research directions: where current applications and research systems are heading, in the publisher's words.

Critical Analysis

Strengths

• A single, focused subject. The book covers one topic, searching for biomedical and health information, and its eight chapters follow a clear arc from foundations through content, indexing, retrieval and access to evaluation and research.
• Written for the field it serves. The publisher positions the book as a guide for both the trainee and the experienced biomedical informatician, and the author is a physician and professor and chair of a medical informatics and clinical epidemiology department, so the domain is treated from inside medicine.
• Evaluation and research get real space. Chapters 7 and 8 together take about 115 pages, over a quarter of the book, which suits readers who need to judge systems and follow current research and not only build them.
• Established title. The fourth edition follows earlier ones; the publisher presented the third as the first book to cover searching from a biomedical perspective, and describes the fourth as extensively revised.

Limitations

• Narrow scope. It is a specialist text on retrieval, not a general health informatics survey; topics such as electronic records, standards or decision support lie outside its chapters.
• Pitched at informaticians. The publisher presents it as a resource for the trainee and the experienced biomedical informatician, so readers with no informatics background will find it easier after a general course.
• Dated by publication. The edition appeared in 2020, so developments since then are not in it; the author's website for the book is where errata and updates would be posted.

Where It Fits in the Course

In the medical informatics course, this is the specialist book for one strand. Enrico Coiera's Guide to Health Informatics and the eighth edition of Health Informatics: Practical Guide, which Hersh edits, are broad introductory surveys and are better first reads. The Shortliffe edited volume is a wider textbook whose application chapters include information-retrieval systems. Principles of Health Interoperability deals with the standards that let systems exchange data, a different problem from finding information in them, and Robert Wachter's The Digital Doctor is a reported account of the digitisation of American medicine rather than a technical text.

Read Hersh after one of the broad surveys, when searching, indexing and evaluation of health information have become the topic of interest; it can also serve as the main text for a short module on retrieval inside a larger informatics course.

How to Use This Book

The book suits graduate students and practitioners in biomedical and health informatics, and others who need to understand how health information is searched for. A first course in health or biomedical informatics is a sensible preparation.

A practical plan:

• Core: read Chapters 1 to 5 in order. Chapter 2 is the longest, so allow it a separate sitting, and read Chapters 4 and 5 together because retrieval depends on indexing.
• Then: Chapter 6 on access and Chapter 7 on evaluation, which are the chapters most relevant to anyone who has to judge a system or its users.
• Optional or later: Chapter 8 on research is best read last, and again after a year or two, since it concerns where the field is heading.

Alongside the reading, try the search systems the book discusses on real questions, compare the results of different searches on the same question, and keep a short log of what was found and missed. The author's website for the edition is the place to look for errata and updates.