Using topic coding to understand the nature of change language in a motivational intervention to reduce alcohol and sex risk behaviors in emergency department patients

https://doi.org/10.1016/j.pec.2016.05.003Get rights and content

Highlights

  • Greater discussion of drawbacks of a behavior does not predict more change talk.

  • Discussions of benefits of a behavior change often elicit support for change.

  • Patient talk supporting behavior change has similar focus for alcohol vs. sex risk.

  • Talk supporting sustaining a behavior differs in its focus for alcohol vs. sex risk.

Abstract

Objective

To elucidate patient language that supports changing a health behavior (change talk) or sustaining the behavior (sustain talk).

Methods

We developed a novel coding system to characterize topics of patient speech in a motivational intervention targeting alcohol and HIV/sexual risk in 90 Emergency Department patients. We further coded patient language as change or sustain talk.

Results

For both alcohol and sex, discussions focusing on benefits of behavior change or change planning were most likely to involve change talk, and these topics comprised a large portion of all change talk. Greater discussion of barriers and facilitators of change also was associated with more change talk. For alcohol use, benefits of drinking behavior was the most common topic of sustain talk. For sex risk, benefits of sexual behavior were rarely discussed, and sustain talk centered more on patterns and contexts, negations of drawbacks, and drawbacks of sexual risk behavior change.

Conclusions

Topic coding provided unique insights into the content of patient change and sustain talk.

Practice implications

Patients are most likely to voice change talk when conversation focuses on behavior change rather than ongoing behavior. Interventions addressing multiple health behaviors should address the unique motivations for maintaining specific risky behaviors.

Introduction

Coding patient language during health behavior counseling is an emerging method for understanding mechanisms responsible for behavior change, a necessary step for improving health behavior counseling [1], [2], [3]. The most widely employed system for coding patient language in health behavior counseling is the Motivational Interviewing Skill Code [MISC; 4]. The MISC codes patient language as reflecting either support for changing a target behavior (change talk) or for sustaining it (sustain talk) according to the tenets of Motivational Interviewing [MI; [5], [6], [7], [8], [9]]. The MISC has been applied primarily to alcohol-focused interventions [6], [10], [11], [12], but also to gambling [13], diet and nutrition [14], and sexual risk reduction [15] interventions. A meta-analysis indicated that the number of patient utterances coded as sustain talk and a composite measure of change and sustain talk (e.g., proportion change talk) are significant predictors of behavior change outcomes [16].

Typically, behavior change counseling covers a range of topics including a patient’s pattern of behavior, consequences of that behavior and of behavior change, barriers and facilitators of change, and change plans. These topics are not captured in any depth by the MISC or related coding systems. For example, the MISC can code that a patient utterance is a reason to change but does not code the nature of that reason. Reasons to change could include drawbacks of ongoing behavior (e.g., hangovers) or benefits of potential behavior change (e.g., having more energy). They also could involve general appraisals of a behavior as negative or outside of peer norms. Although behavior change interventions can focus on a range of topics, it is unknown the extent to which conversations that focus on certain topics are more likely to involve change vs. sustain talk.

The extent to which certain topics arise during behavior change counseling and are likely to involve change vs sustain talk may differ based on the target behavior under discussion. Interventions that focus on multiple health behaviors are becoming increasingly common [17], [18]. If topics that predominate in change or sustain talk depend on the target behavior, then counseling content can be altered to anticipate the differential importance of specific topics for each behavior. Pairing MISC coding with topic coding may provide important insights into the nature of motivational content across different targeted behaviors.

This study examined how topics of discussion in a health behavior intervention relate to patient change language across two target behaviors, alcohol use and sex risk. We employed an extension of the Generalized Medical Interaction Analysis System (GMIAS), a conversation coding system that permits detailed investigation of patient-provider interaction across various topics addressed in medical encounters [19], [20], [21], [22]. For this project, we expanded the GMIAS topic coding structure to include a new category and subcategories of topics that capture content typically covered in a health behavior intervention; we refer to this adapted system as the Generalized Behavioral Intervention Analysis System (GBIAS). We applied the GBIAS and the MISC to audio recordings of a brief motivational intervention to reduce risky alcohol use and sexual behaviors with Emergency Department (ED) patients who reported concurrent heavy drinking and sexual behavior that increases risk for HIV infection.

Section snippets

Method

Data were drawn from a randomized clinical trial of an MI intervention targeting both alcohol and sex-risk behaviors, which demonstrated efficacy for reducing both behaviors relative to a brief advice control [23]. The project was approved by hospital and university institutional review boards. Patients aged 18–65 were recruited from two community hospital EDs. Eligibility criteria included: 1) total score of ≥8 (males) or ≥6 (females) on the Alcohol Use Disorders Identification Test [24] or

Topic code frequencies

The second column of Table 2 shows the median number of utterances that received specific topic codes within 2.0 broken down by the target behaviors of alcohol use and sex risk. Overall, a substantially higher number of utterances concerned alcohol use compared to sex. Discussions of sexual consequences of drinking (e.g., increasing sexual pleasure or contributing to regretted experiences) were coded with alcohol use as the target behavior and comprised 2.4% of utterances on alcohol. For both

Discussion

This study demonstrated application of the newly developed Generalized Behavioral Intervention Analysis System (GBIAS) to the coding of sessions of an efficacious MI intervention that reduced both alcohol use and sexual risk behaviors [23]. Using the GBIAS, we reliably coded categories and subcategories of topics on which alcohol- and sex-related conversations focused. As described below, topic codes provided novel information about the content of the MI intervention that cannot be provided

Conflict of interest

The authors have no relevant conflicts of interest to disclose.

Role of the funding source

Support for this work was provided by the National Institute on Alcohol Abuse and Alcoholism (NIAAA), grants U24AA022003, 5R01AA009892-20, P01AA019072, and K23AA018126. Dr. Bryant from NIAAA was involved in conceptualizing and writing this manuscript. NIAAA did not have any other involvement in the project.

References (30)

  • W. Miller et al.

    Motivational Interviewing Helping People Change

    (2012)
  • T.B. Moyers et al.

    Client language as a mediator of motivational interviewing efficacy: where is the evidence

    Alcohol. Clin. Exp. Res.

    (2007)
  • T.B. Moyers et al.

    From in-session behaviors to drinking outcomes: a causal chain for motivational interviewing

    J. Consult. Clin. Psychol.

    (2009)
  • J.S. Baer et al.

    Adolescent change language within a brief motivational intervention and substance use outcomes

    Psychol. Addict. Behav.

    (2008)
  • D.C. Hodgins et al.

    Strength of commitment language in motivational interviewing and gambling outcomes

    Psychol. Addict. Behav.

    (2009)
  • Cited by (0)

    View full text