Leveraging Latent Dirichlet Allocation in processing free-text personal goals among patients undergoing bladder cancer surgery

Yuelin Li, Bruce D. Rapkin, Thomas M. Atkinson, Elizabeth Schofield, Bernard H. Bochner

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Purpose: As we begin to leverage Big Data in health care settings and particularly in assessing patient-reported outcomes, there is a need for novel analytics to address unique challenges. One such challenge is in coding transcribed interview data, typically free-text entries of statements made during a face-to-face interview. Latent Dirichlet Allocation (LDA) offers statistical rigor and consistency in automating the interpretation of patients’ expressed concerns and coping strategies. Methods: LDA was applied to interview data collected as part of a prospective, longitudinal study of QOL in N = 211 patients undergoing radical cystectomy and urinary diversion for bladder cancer. LDA analyzed personal goal statements to extract the latent topics and themes, stratified by time, and on things patients wanted to accomplish and prevent. Model comparison metrics determined the number of topics to extract. Results: LDA extracted seven latent topics. Prior to surgery, patients’ priorities were primarily in cancer surgery and recovery. Six months after the surgery, they were replaced by goals on regaining a sense of normalcy, to resume work, to enjoy life more fully, and to appreciate friends and family more. LDA model parameters showed changing priorities, e.g., immediate concerns on surgery and resuming employment decreased post-surgery and were replaced by concerns over cancer recurrence and a desire to remain healthy and strong. Conclusions: Novel Big Data analytics such as LDA offer the possibility of summarizing personal goals without the need for conventional fixed-length measures and resource-intensive qualitative data coding.

Original languageEnglish (US)
JournalQuality of Life Research
DOIs
StatePublished - Jan 1 2019

Fingerprint

Urinary Bladder Neoplasms
Interviews
Urinary Diversion
Cystectomy
Longitudinal Studies
Neoplasms
Prospective Studies
Delivery of Health Care
Recurrence

Keywords

  • Big Data analysis
  • Bladder cancer
  • Latent Dirichlet Allocation
  • Qualitative data
  • Text analysis

ASJC Scopus subject areas

  • Public Health, Environmental and Occupational Health

Cite this

Leveraging Latent Dirichlet Allocation in processing free-text personal goals among patients undergoing bladder cancer surgery. / Li, Yuelin; Rapkin, Bruce D.; Atkinson, Thomas M.; Schofield, Elizabeth; Bochner, Bernard H.

In: Quality of Life Research, 01.01.2019.

Research output: Contribution to journalArticle

@article{c4aa51370a9743f08fad84d14ace8b63,
title = "Leveraging Latent Dirichlet Allocation in processing free-text personal goals among patients undergoing bladder cancer surgery",
abstract = "Purpose: As we begin to leverage Big Data in health care settings and particularly in assessing patient-reported outcomes, there is a need for novel analytics to address unique challenges. One such challenge is in coding transcribed interview data, typically free-text entries of statements made during a face-to-face interview. Latent Dirichlet Allocation (LDA) offers statistical rigor and consistency in automating the interpretation of patients’ expressed concerns and coping strategies. Methods: LDA was applied to interview data collected as part of a prospective, longitudinal study of QOL in N = 211 patients undergoing radical cystectomy and urinary diversion for bladder cancer. LDA analyzed personal goal statements to extract the latent topics and themes, stratified by time, and on things patients wanted to accomplish and prevent. Model comparison metrics determined the number of topics to extract. Results: LDA extracted seven latent topics. Prior to surgery, patients’ priorities were primarily in cancer surgery and recovery. Six months after the surgery, they were replaced by goals on regaining a sense of normalcy, to resume work, to enjoy life more fully, and to appreciate friends and family more. LDA model parameters showed changing priorities, e.g., immediate concerns on surgery and resuming employment decreased post-surgery and were replaced by concerns over cancer recurrence and a desire to remain healthy and strong. Conclusions: Novel Big Data analytics such as LDA offer the possibility of summarizing personal goals without the need for conventional fixed-length measures and resource-intensive qualitative data coding.",
keywords = "Big Data analysis, Bladder cancer, Latent Dirichlet Allocation, Qualitative data, Text analysis",
author = "Yuelin Li and Rapkin, {Bruce D.} and Atkinson, {Thomas M.} and Elizabeth Schofield and Bochner, {Bernard H.}",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/s11136-019-02132-w",
language = "English (US)",
journal = "Quality of Life Research",
issn = "0962-9343",
publisher = "Springer Netherlands",

}

TY - JOUR

T1 - Leveraging Latent Dirichlet Allocation in processing free-text personal goals among patients undergoing bladder cancer surgery

AU - Li, Yuelin

AU - Rapkin, Bruce D.

AU - Atkinson, Thomas M.

AU - Schofield, Elizabeth

AU - Bochner, Bernard H.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Purpose: As we begin to leverage Big Data in health care settings and particularly in assessing patient-reported outcomes, there is a need for novel analytics to address unique challenges. One such challenge is in coding transcribed interview data, typically free-text entries of statements made during a face-to-face interview. Latent Dirichlet Allocation (LDA) offers statistical rigor and consistency in automating the interpretation of patients’ expressed concerns and coping strategies. Methods: LDA was applied to interview data collected as part of a prospective, longitudinal study of QOL in N = 211 patients undergoing radical cystectomy and urinary diversion for bladder cancer. LDA analyzed personal goal statements to extract the latent topics and themes, stratified by time, and on things patients wanted to accomplish and prevent. Model comparison metrics determined the number of topics to extract. Results: LDA extracted seven latent topics. Prior to surgery, patients’ priorities were primarily in cancer surgery and recovery. Six months after the surgery, they were replaced by goals on regaining a sense of normalcy, to resume work, to enjoy life more fully, and to appreciate friends and family more. LDA model parameters showed changing priorities, e.g., immediate concerns on surgery and resuming employment decreased post-surgery and were replaced by concerns over cancer recurrence and a desire to remain healthy and strong. Conclusions: Novel Big Data analytics such as LDA offer the possibility of summarizing personal goals without the need for conventional fixed-length measures and resource-intensive qualitative data coding.

AB - Purpose: As we begin to leverage Big Data in health care settings and particularly in assessing patient-reported outcomes, there is a need for novel analytics to address unique challenges. One such challenge is in coding transcribed interview data, typically free-text entries of statements made during a face-to-face interview. Latent Dirichlet Allocation (LDA) offers statistical rigor and consistency in automating the interpretation of patients’ expressed concerns and coping strategies. Methods: LDA was applied to interview data collected as part of a prospective, longitudinal study of QOL in N = 211 patients undergoing radical cystectomy and urinary diversion for bladder cancer. LDA analyzed personal goal statements to extract the latent topics and themes, stratified by time, and on things patients wanted to accomplish and prevent. Model comparison metrics determined the number of topics to extract. Results: LDA extracted seven latent topics. Prior to surgery, patients’ priorities were primarily in cancer surgery and recovery. Six months after the surgery, they were replaced by goals on regaining a sense of normalcy, to resume work, to enjoy life more fully, and to appreciate friends and family more. LDA model parameters showed changing priorities, e.g., immediate concerns on surgery and resuming employment decreased post-surgery and were replaced by concerns over cancer recurrence and a desire to remain healthy and strong. Conclusions: Novel Big Data analytics such as LDA offer the possibility of summarizing personal goals without the need for conventional fixed-length measures and resource-intensive qualitative data coding.

KW - Big Data analysis

KW - Bladder cancer

KW - Latent Dirichlet Allocation

KW - Qualitative data

KW - Text analysis

UR - http://www.scopus.com/inward/record.url?scp=85062023040&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062023040&partnerID=8YFLogxK

U2 - 10.1007/s11136-019-02132-w

DO - 10.1007/s11136-019-02132-w

M3 - Article

JO - Quality of Life Research

JF - Quality of Life Research

SN - 0962-9343

ER -