The Four Habits Coding Scheme: Validation of an instrument to assess clinicians’ communication behavior

https://doi.org/10.1016/j.pec.2005.04.015Get rights and content

Abstract

Objective

To present preliminary evidence for the reliability and validity of the Four Habits Coding Scheme (4HCS), an instrument based on a teaching model used widely throughout Kaiser Permanente to improve clinicians’ communication skills.

Methods

One hundred videotaped primary care visits were coded using the 4HCS, and the data were assessed against a previously available data set for these visits, including the Roter Interaction Analysis System (RIAS), back channel responses, measures of nonverbal behavior, length of visit, and patients’ post-visit assessments.

Results

Levels of inter-rater reliability were acceptable, and the distribution of ratings across items indicated that physicians’ modal responses varied widely. Correlations between 4HCS ratings, RIAS, back channel responses, and non-verbal measures provided evidence of the instrument's construct validity.

Conclusions

The Four Habits Coding Scheme, an instrument that combines both evaluative and descriptive elements of physician communication behavior and is derived from a conceptually based teaching model, has the potential to be of utility to researchers and evaluators as well as educators and clinicians.

Practice Implications

The Four Habits Coding Scheme provides a template for both guiding and measuring physician communication behaviors.

Introduction

The Four Habits Model, used extensively to teach communication skills to thousands of clinicians in Kaiser Permanente, describes clusters of physician behaviors and skills associated with effective clinical practice and positive health outcomes [1], [2], [3]. The Four Habits (Invest in the Beginning; Elicit the Patient's Perspective; Demonstrate Empathy; and Invest in the End) lay out the basic tasks or functions of the medical interview, and also conceptualize how the elements of the interview relate to one another within and across medical visits. This model is consistent with the patient- and relationship-centered approaches to health care [4], [5], [6], and also derives from the three-function model of medical interviewing [7]. This paper presents the initial test of the Four Habits Coding Scheme (4HCS), an instrument for describing and evaluating clinician behavior based on this model.

Methods of describing and quantifying clinician communication behaviors have varied considerably, depending on the goals of the instrument developers. Educators, primarily interested in identifying and improving specific behaviors of medical students and residents, have used a wide variety of instruments. The Maastricht History Taking and Advice Checklist (MAAS) [8], [9] and the Calgary-Cambridge Guides [10], [11] are prime examples of comprehensive models and coding instruments that have been used widely and successfully for teaching and training. Although several rating formats have been used, checklists which code behavior as present or absent are the most common [12], [13].

Researchers interested in communication patterns between patients and clinicians have developed a number of more elaborate coding schemes to categorize clinician and/or patient behavior. The Roter Interaction Analysis Scheme (RIAS) [14], [15], the coding system most widely cited in the medical literature, assigns all physician and patient utterances into a set of 28 mutually exclusive categories, counting the frequency of each observed behavior. The Verona Medical Interview Classification System [16], [17] divides communication behaviors into 22 categories according to the three function model, and has now been factor analyzed to distinguish different strategies of exchanging information, building a relationship, and negotiating a treatment plan. The 4HCS, derived from a conceptual model of practice and tied closely to a teaching framework, sits at the intersection of the research and training worlds, and can be used successfully in either context.

The 4HCS consists of 23 items derived from the core skills referred to in the Four Habits Model (see Appendix A for a complete list of all items). It differs from existing instruments in several ways. First, the behaviors observed and coded are more broadly defined than those in a standard checklist. For instance, to determine whether the “clinician shows great interest in exploring the patient's understanding of the problem” the 4HCS coder has to consider and combine several discrete behaviors such as the number and type of questions asked, and the nonverbal and verbal signals that encourage the patient to tell his/her story. Second, rather than focusing on frequency counts of behavior, the 4HCS asks coders to distinguish among five levels of performance for each coded behavior category. Third, while the coding categories describe clinicians’ behavior, the underlying conceptual model implies distinctions that are actually evaluative. The data presented here represent an initial attempt to operationalize and code the behaviors subsumed by the Four Habits Model, and to establish the reliability and validity of the 4HCS.

Section snippets

Four Habits rating items

The 4HCS identifies a set of 23 behaviors, each associated with one of the Four Habits. Habit 1, Invest in the Beginning, contains six items that focus on creating rapport quickly and planning the visit (e.g., demonstrating familiarity with the patient and greeting the patient warmly). Habit 2, Elicit the Patient's Perspective, contains three items (eliciting the patient's understanding of the problem, understanding the patient's goals for the visit, determining the impact of the problem on the

Descriptive characteristics

The extent to which the clinicians engaged in each of the behaviors subsumed under the Four Habits varied considerably. As indicated in Table 1, the mean scores were particularly high on one component of Habit 1, demonstrating familiarity with the patient; and on three components of Habit 4, using the patient's concern to frame diagnostic information, offering information with little jargon, and making plans for follow-up. Mean scores were particularly low on two items: clinicians typically

Discussion

Our results indicate that the Four Habits Coding Scheme is a reliable and valid instrument that shows promise of practical utility. Concerning its reliability, the inter-rater coefficients, which ranged between .69 and .80, are at generally satisfactory levels, but they are not quite as high as some other instruments. Compared to coding schemes that use frequency counts or ratings of present versus absent, inter-rater reliability is harder to achieve with the 4HCS, first, because the families

Conclusions and practice implications

In this initial test, we have demonstrated that the 4HCS has many of the psychometric characteristics of a useful instrument, most notably construct validity in relation to the RIAS, back channel responses, and several indicators of non-verbal behavior. Because this instrument clusters clinicians’ behaviors and measures them in a way that is different from other current instruments, it stands independently as a useful addition to the field of research in physician communication.

The unique

Acknowledgements

The authors would like to thank Judith Hall for her generous assistance in making the videotapes and previously coded data available for this project, as well as for her advice and insights during the course of this work. We would also like to acknowledge Alithia Broderick and Chris Hummel for their efforts in refining the coding scheme and coding the tapes.

References (21)

There are more references available in the full text version of this article.

Cited by (189)

View all citing articles on Scopus
View full text