A linguistic study of patient-centered interviewing: Emergent interactional effects Ashley M. Hesson a, *, Issidoros Sarinopoulos a , Richard M. Frankel b , Robert C. Smith a a Michigan State University College of Human Medicine, East Lansing, USA b Indiana University School of Medicine, Indianapolis, USA 1. Introduction Patient-centered interviewing (PCI) has been extensively studied, with recent work specifically defining PCI behaviorally and providing evidence that it is effective in improving patient outcomes [1–5]. This paper describes the emergent linguistic effects that distinguish standard clinician-centered interviewing (CCI) from a well-established, evidence-based method that integrates PCI and CCI, summarized in Table 1. RCTs based on behaviorally defined interviewing methods have focused on indirect measures of the method’s effect including: patient satisfaction [6–10], use of health services [9,11], health status outcomes [12–17], and quality of life metrics [16,18]. While these measures demonstrate validity, they do not represent the direct effects of a given method on the provider–patient interaction (PPI). The field needs direct measures to provide explanatory data about the mechanism(s) of successful interview- ing methods. For valid methods like the PCI method studied here, the next step is to achieve understanding of the intervening factors between PCI behaviors, patient perceptions, and outcomes. Recent research highlights the ways in which purely outcome or interaction-based research can be misleading: providing incomplete or incorrect characterizations of the relationship between an intervention and its effects [19,20]. These findings point to a need a mechanistic understanding of patient-centeredness. For example, using the Roter Interactional Assessment System (RIAS [21]), Roter and Cooper [13,14] studied links between PCI and health outcome/ satisfaction measures, triangulating the method, the intermediary mechanism, and patient outcomes. However, RIAS codes operation- ally define communication behaviors that can overlap with PCI methods, complicating the process of defining the predictor and the predicted. Though we do not claim to have solved this problem, we focus on stylistic (i.e., phonological, structural, and organizational) features that are not part of the PCI method. Patient/physicians’ use of these linguistic features is largely subconscious and independent from the behavioral steps of the PCI. Instead of applying the RIAS or other communication-behavior coding system, we turned to the discourse analytic approach of interactional sociolinguistics [22–24] to identify direct measures of patient-centeredness, stylistic rather than behavioral features of the PPI. Stylistic markers are also relatively unexplored in medical Patient Education and Counseling 88 (2012) 373–380 A R T I C L E I N F O Article history: Received 31 January 2012 Received in revised form 2 June 2012 Accepted 8 June 2012 Keywords: Provider–patient relationship Patient-centered interviewing linguistics Sociolinguistics Discourse analysis Communication A B S T R A C T Objectives: To evaluate interactional effects of patient-centered interviewing (PCI) compared to isolated clinician-centered interviewing (CCI). Methods: We conducted a pilot study comparing PCI (N = 4) to CCI (N = 4) for simulated new-patient visits. We rated interviews independently and measured patient satisfaction with the interaction via a validated questionnaire. We conducted interactional sociolinguistic analysis on the interviews and compared across three levels of analysis: turn, topic, and interaction. Results: We found significant differences between PCI and CCI in physician responses to patients’ psychosocial cues and concerns. The number and type of physician questions also differed significantly across PCI and CCI sets. Qualitatively, we noted several indicators of physician–patient attunement in the PCI interviews that were not present in the CCI interviews. They spanned diverse aspects of physician and patient speech, suggesting interactional accommodation on the part of both participants. Conclusions: This small pilot study highlights a variety of interactional variables that may underlie the effects associated with patient-centered interviewing (e.g., positive relationships, health outcomes). Question form, phonological accommodation processes, and use of stylistic markers are relatively unexplored in controlled studies of physician–patient interaction. Practice implications: This study characterizes several interactional variables for larger scale studies and contributes to models of patient-centeredness in practice. ß 2012 Elsevier Ireland Ltd. All rights reserved. * Corresponding author at: B312 Clinical Center, East Lansing, MI 48824, USA. Tel.: +1 517 353 3730; fax: +1 517 432 1326. E-mail address: bartell6@msu.edu (A.M. Hesson). Contents lists available at SciVerse ScienceDirect Patient Education and Counseling jo ur n al h o mep ag e: w ww .elsevier .co m /loc ate/p ated u co u 0738-3991/$ – see front matter ß 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.pec.2012.06.005