The Grammar Behind Patient Experience Improvement


09.01.2017 • Zach Childers, PhD

iStock-484295463-1.pngThe richest indicator of a patient's good or bad experience is not stars or scores, but sentences.  

For that reason, a little bit of linguistic insight into how meaning is created at the sentence level can be useful for patient experience leaders to understand what their patients communicate, and thereby prioritize areas of opportunity.

The core of every sentence is a subject and a predicate. This fact of language was observed at least as far back as Aristotle.

In a simple sentence like "The doctors were kind," the subject "The doctors" and the predicate "kind" are joined by a linking verb (or "copula" in jargon). If all sentences were as simple as that, then automatically processing language for meaning would be an almost trivial task (almost, but not quite, due in part to the additional problem of word-sense disambiguation).

Sentences display much more complexity than the simple subject-predicate structure, however; the transformation of meaning through sentence structure can range from the subtle to the extreme.

A subtle and familiar transformation of sentence structure is that of the passive voice.

If you’re tracking the quality of physician communication in your hospital, you would be interested in sentences like "The doctor explained the procedure well." But, you would also be interested in its passive counterpart "The procedure was explained well by the doctor."

This is an example of a subtle transformation of meaning through different sentence structures. A comment analysis platform for patient experience built on “AI done right” (see previous article) would identify both of these sentences (as well as other structures, such as nominalizations: “The doctor’s explanation of the procedure”) as referring to the same topic and therefore belonging to the same patient experience category (doctor communication) despite their structural differences.

Let’s take a look at the difference between the two sentences below:

1. The patient told the doctor about potential prescription complications.

2. The patient was told by the doctor about potential prescription complications.

While these two sentences are made up of pretty much the same words, their meanings are quite different because what matters more than the words themselves is their relationship to each other as parts of a sentence.

AI-done-right would categorize sentence 1 under doctor listening and sentence 2 under doctor explanation as distinct areas of patient experience. The analytic capacity to categorize comments as belonging to these distinct patient experience categories is critical for patient experience leaders to glean actionable insights from patients’ comments and to prioritize their improvement initiatives.

Here’s another example of a significant meaning difference between two sentences, despite the fact that they are made up of the same words:

3. The nurses were kind I would go back

4. Were the nurses kind I would go back

Punctuation, especially in social media, cannot be relied on to disambiguate these kinds of sentences.

The answer to the question this raises, “Then how do I actually understand what my patients are saying?” is, of course, “through AI-done-right.”

At NarrativeDx, we’ve built our AI platform for patient experience with the linguistic knowledge on how putting the same words in different sequences can either transform or preserve meaning. We’re proud to contribute to the transformation of patient experience by bringing to the industry the analytical capability to amplify patients’ voices - regardless of how they put their words together.

Read our latest white paper and discover three important reasons why healthcare organizations benefit from using the right technology to hear patients accurately for actionable insights.

      READ WHITE PAPER

Zach Childers, PhD

Zach Childers, PhD

Linguistics Engineer, NarrativeDx

Get ”Insights into Patient Experience“ delivered to your inbox