Japanese Journal of Smoking Control Science
Online ISSN : 1883-3926
Volume vol.12, Issue 08
Displaying 1-1 of 1 articles from this issue
  • :A Latent Class Model Approach
    Kazuaki Uchikawa, Michiko Watanabe
    2018 Volume vol.12 Issue 08 Pages 1-9
    Published: 2018
    Released on J-STAGE: June 02, 2021
    Aim: The current research carried out patient classification based on the individual characteristics found in the diagnostic information of clinical records while also outlining continued consultation in an outpatient smoking cessation clinic. The purpose of this research was to uncover the characteristics of patients who discontinue attendance at outpatient smoking cessation clinics.
    Method: The research design was a retrospective cohort study in a single institution. There were 374 participants drawn from 468 patients seen at Hospital A between April 1st, 2008, and March 31st, 2014, after excluding 81 patients whose initial consultation data were lost and a further 13 patients who discontinued treatment with the consent of their physician. Patients were classified using latent class analysis, and the characteristics of each class was subsequently identified.
    Results: A total of four classes—one continuer class and three discontinuer classes—were identified for patients who attended at least five continued-consultation sessions. Moreover, classes categorized according to patient characteristics were named: achiever (70.3% class size), early-stage discontinuer (8.5% class size), middle-stage discontinuer (10.3% class size), and final-stage discontinuer class (10.9% class size). Furthermore, when patients were classified according to a 90% attribution probability, it was possible to classify 98.1% of participants into one of the classes.
    Conclusions: It was revealed that patients can be classified according to time of discontinuation using the latent class model. This suggests that it could be possible for medical practitioners to predict time of discontinuation from patient information gathered at the initial consultation. If discontinuation timing could be predicted at the initial consultation, it would be possible to offer more suitable smoking cessation support to individual patients. Thus, it is thought that this could lead to improved rates of continued attendance at outpatient smoking-cessation clinics and to consequently ameliorate clinic-based complete smoking cessation and healthy life expectancy.
    Download PDF (773K)