Natural Language Processing to Extract Meaningful Information from Patient Experience Feedback

Abstract

Background: Due to reimbursement tied in part to patients' perception of their care, hospitals continue to stress obtaining patient feedback and understanding it to plan interventions to improve patients' experience. We demonstrate the use of natural language processing (NLP) to extract meaningful information from patient feedback obtained through Press Ganey surveys.

Methods: The first step was to standardize textual data programmatically using NLP libraries. This included correcting spelling mistakes, converting text to lowercase, and removing words that most likely did not carry useful information. Next, we converted numeric data pertaining to each category based on sentiment and care aspect into charts. We selected care aspect categories where there were more negative comments for more in-depth study. Using NLP, we made tables of most frequently appearing words, adjectives, and bigrams. Comments with frequent words/combinations underwent further study manually to understand factors contributing to negative patient feedback. We then used the positive and negative comments as the training dataset for a neural network to perform sentiment analysis on sentences obtained by splitting mixed reviews.

Results: We found that most of the comments were about doctors and nurses, confirming the important role patients ascribed to these two in patient care. "Room," "discharge" and "tests and treatments" were the three categories that had more negative than positive comments. We then tabulated commonly appearing words, adjectives, and two-word combinations. We found that climate control, housekeeping and noise levels in the room, time delays in discharge paperwork, conflicting information about discharge plan, frequent blood draws, and needle sticks were major contributors to negative patient feedback. None of this information was available from numeric data alone.

Conclusion: NLP is an effective tool to gain insight from raw textual patient feedback to extract meaningful information, making it a powerful tool in processing large amounts of patient feedback efficiently.

Fig. 1

Preprocessing of the raw text. All text converted to lowercase. Numbers, special characters, and extra spaces removed. Certain words were similar in meaning, for example doc, dr, and doctors, all three were replaced with physician. Words were also reduced to base form, for example explained, explaining, and explain mean the same thing, therefore reduced to one word.

Fig. 2

Natural language processing methods: ( A ) part of speech tagging, using representation of words and their tags; ( B ) output of the parts of speech tagger. Each word is coupled with its tag; ( C ) summary of neural network based model; and ( D ) evaluation metrics of the model.

Fig. 3

( A , B ) Graphical representation of the training process of the model. Accuracy did not improve significantly after 15 epochs, but loss continued to decrease and started going up after 35 epochs.

Fig. 4

Classification of mixed sentiment comments. All mixed comments were split into sentences. The deep learning model trained with the preclassified positive and negative comments was then used to classify the sentences into positive, negative, or neutral.

Fig. 5

Word Cloud representations. The larger the word, the more frequent it appeared in the comments. ( A ) “Nurse” and “doctor” are the most frequent words in both positive and negative comments; ( B ) “room” also appears more frequently in negative comments.

Fig. 6

Comments in each category: ( A ) based on sentiment and ( B ) based on care aspect.

Fig. 7

Comments in each category based on care aspect: ( A ) positive comments and ( B ) negative comments.

Fig. 8

Positive and negative comments in each category based on care aspect. Most care aspect categories had more positive comments than negative except for comments about room, tests and treatments, and discharge.

Fig. 9

Classification of mixed sentiment feedback by deep learning model. A mixed comment may have many sentences; this figure shows how different sentences in one mixed feedback may be classified based on sentiment by the classification model.