Social Science & Medicine 54 (2002) 10111023

Communicative competence in the delivery of bad news

Cathy Gillottia, Teresa Thompsonb,*, Kelly McNeilisc
b a Department of Communication, Purdue University Calumet, USA Department of Communication, University of Dayton, Dayton, OH 45460-1410, USA c Department of Communication, Southwest Missouri State University, USA

Abstract Grounded in the Cegala and Waldron (Communication Studies 43 (1992) 105) model of communicative competence, the present study applied the McNeilis (Health Communication 13 (2001) 5) provider-patient coding scheme to video tapes of 3rd year medical students delivering bad news to a standardized patient. The goal of the study was to understand the specic communicative moves that are associated with perceptions of competence during bad news delivery. The coding scheme assesses Content, Acknowledgment Tokens, Interruptions, Alignment, and Function of the message. Na.ve observers also evaluated the tapes on several items, assessing empathy and communicative eectiveness. Nonmedical talk was the most common type of content, followed by discussion of the current health problem. Neither acknowledgment tokens nor interruptions were frequent. The most common function of a message was a closed question, followed by explanations, assertions, and open questions. Summing across the functions indicated that information giving was the most common behavior. The perceivers data showed fairly neutral assessments of the medical students}they were generally not evaluated very positively, although they were not disliked. Regression analyses indicated numerous specic communicative behaviors that were associated with judgments of competence. Statements falling into the Nonspecic Content category were associated with more positive perceptions, while relational statements, moderately closed questions, solicited answers, expansions, restatements, assertions, explanations, open questions, bracketing, and small talk as well as information verifying, seeking, and giving (summed functions) led to more negative perceptions. The results indicate that the delivery of bad news requires communicative moves that dier from other kinds of medical communication. Depending on the results of future analyses of this topic, health care providers may be well advised to focus little of their communication on information seeking, giving, or verifying during the initial bad news delivery consultation, but rather to save most communication of information for a follow-up scheduled shortly afterwards. # 2002 Elsevier Science Ltd. All rights reserved.
Keywords: USA; Doctorpatient communication; Competence

Communicative competence in the delivery of bad news Research on the interaction between patients and their health care providers has grown tremendously over the last 20 years (see reviews by Thompson, 1994, 1998, 2000). One theme that pervades this literature is the need for an understanding of the communicative behaviors that are more vs. less competent within such interac*Corresponding author. Tel.: +1-937-229-2379. E-mail address: thompson@udayton.edu (T. Thompson).

tions. The research on the outcomes of health care interaction indicates that how patients and care providers interact matters}it impacts such outcomes as patient compliance with treatment regimens (DiMatteo, Reiter, & Gambone, 1994), the ling of malpractice suits (Vincent, Young, & Phillips, 1994), patient recovery (Anderson, 1987), numerous physiological and medical outcomes (Kaplan, Greeneld, & Ware, 1989; Van Veldhuizen-Scott, Widmer, Stacey, & Popovich, 1995), medical costs (Lieberman, 1992), pain (King, 1991), mortality (Lieberman, 1992), and patient understanding

0277-9536/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 2 7 7 - 9 5 3 6 ( 0 1 ) 0 0 0 7 3 - 9

1012

C. Gillotti et al. / Social Science & Medicine 54 (2002) 10111023

of the diagnosis, prognosis, and treatment plan (Thompson, 2000). How competently providerpatient interaction occurs, then, is worthy of study. One framework that has been suggested for the study of providerpatient interaction is the concept of communication competence (Kasch, 1984; Kasch, Kasch, & Lisnek, 1998; Kreps & Query, 1990; Morse & Piland, 1981). Research on communicative competence argues that some communicative behaviors are more eective in terms of meeting goals than are other behaviors, and that context determines the behaviors that are most appropriate and eective. A context that is particularly troubling for both patients and health care providers is the delivery of bad news (Gillotti & Applegate, 2000). This is, of course, a context with which care providers are frequently faced, but at which they are not always successful. Care providers experience great discomfort in such situations (Bowers, 1999; Maynard, 1991) and do not typically deliver the bad news very eectively (Franks, 1997; Girgis, Sanson-Fisher, & Schoeld, 1999; Parathian & Taylor, 1993; Roth & Nelson, 1997; Salander, Bergenheim, Bergstroem, & Henriksson, 1998; Vetto, 1999). This, then, is a context in which the notion of communicative competence has particular applicability. The goal of the present study is to apply the communicative competence framework to the bad news delivery context in an attempt to ascertain those communicative behaviors that are associated with more vs. less competent bad news delivery. Communicative competence The notion of communicative competence requires a dyadic perspective. Competence in communication is determined by both conversational partners, involves knowledge of how to communicate, references actual communicative behavior, and reects the communicators success at achieving goals (Parks, 1994). The particular competence framework selected for the present study was the model developed by Cegala and Waldron (1992). This model was selected because it privileges participants language-in-use and is context bound. These characteristics of the model make it especially relevant to the providerpatient context. Although, as mentioned above, much research has been conducted on the providerpatient relationship, little of this research has been theoretically grounded. This has limited the applicability and usefulness of the ndings. The Cegala and Waldron (1992) competence model is based upon four assumptions, each of which reects a synthesis of current theorizing on competence. First, competence is best dened by how interactants align their utterances as they coordinate goals (Delia & OKeefe, 1982; Pearce & Cronen, 1980). Next, denitions of competence must be context-specic and

acknowledge that competence evaluations are situational (Fisher, 1982; Watzlwawick, Beavin, & Jackson, 1967). Third, competence is dyadic and must be measured by looking at patterns of interaction (Watzlawick et al., 1967). Finally, individual dierences in competence are partially determined by cognitive/ aective processes involving the interpretation and production of messages during interactions (Applegate, 1990; Burleson, 1984; Delia & OKeefe, 1982). Thus, competence requires participants to align their goals and necessitates an ability to grasp the meaning and intent of the other. These assumptions are also consistent with a constructivist, person-centered perspective on competence (Applegate, 1990; Burleson, 1984; Delia & OKeefe, 1982), which will serve as the basis for the measurement of competence perceptions in the present study. The CegalaWaldron model, while building upon the same assumptions as a constructivist perspective, allows for more precise measurement of speech variables and the sequential nature of language. The application of this model to medical interaction requires an understanding of the objectives of talk in this context. This necessitates a task analysis that reveals participants likely goals in the situation and the communicative moves that will likely allow the accomplishment of goals (McFall, 1982). Meta-analytic reviews of research on providerpatient interaction (Hall, Roter, & Katz, 1988; Roter, Hall, & Katz, 1988) and research by Cegala and colleagues (Cegala, McNeilis, McGee, & Jonas, 1995; Cegala, McGee, & McNeilis, 1996) reveal that the primary communicative tasks in the medical context are information exchange and relational development. The importance of information exchange in the health care context has been discussed by numerous researchers (e.g., Beisecker, 1990; Beisecker & Beisecker, 1990; Guttman, 1993; Katz, Gurevitch, Peled, & Danet, 1969; Pendleton & Bochner, 1980; Roter, 1989; Roter & Frankel, 1992; Street, 1991a; Waitzkin, 1984, 1985), although little research has actually focused on the specics of how such information exchange takes place. Additionally, physicians and patients also communicate with each other to build a relationship (Ben-Sira, 1980; Smith & Hoppe, 1991), which then impacts trust, respect, loyalty, and satisfaction with health care (Cegala et al., 1996). We know little about the relationship between the information exchange and relationship development functions of medical communication (for exceptions see Buller & Street, 1991; Cegala et al., 1996; Roter, 1989; Street, 1991b). The coding scheme used in the present study}the Communication and Competence System (CACS, McNeilis, 2001)}addresses both of these components of medical exchange and overcomes many of the limitations of other commonly used methods. Past methods have focused upon rating scales, coding systems that are not theoretically grounded, and

C. Gillotti et al. / Social Science & Medicine 54 (2002) 10111023

1013

conversation analysis, which, while interesting, are not useful for quantitative assessments relating communication patterns to outcome variables. Grounded in Cegala and Waldrons competence model, the CACS coding scheme can be used to look at communicative competence within the medical context more microscopically than past coding schemes while still allowing the quantitative comparisons necessary to study outcomes. For a more complete description of the development of this coding scheme, see McNeilis (2001). The coding scheme is a multidimensional system that tracks participants responsiveness to previous utterances on global, topical levels, and to the intent and meaning of the interactional partner. The scheme assesses communicative functions such as information giving, information seeking, information verifying, and socio-emotional communication using 31 dierent categories; it also addresses communicative content issues (9 categories) and conversational alignment (9 categories, including acknowledgement tokens and interruptions). No other coding scheme oers such a detailed assessment of competence in communication, while still being theoretically grounded (Thompson, 2001). The particular medical context to which the coding scheme was applied was the delivery of bad news. Bad news delivery The study of bad news delivery may seem insignicant among the many communicative tasks clinicians must undertake. However, while the interactions are routine in some respects, they are particularly challenging in others (Maynard, 1991). Bad news is typically dened as information that negatively alters (Buckman, 1992, p. 15) the patients perspective of his or her future. The interaction still involves information giving and seeking, as do most medical consultations, but the emotional component and subsequent patient retention are dierent than other medical interactions (Buckman, 1992). Not much is known about the actual training of health care professionals in this area (Sharp, Strauss, & Lorch, 1992). Some clinicians have argued that what training does take place is inadequate (Davis, 1991; Falloweld, 1993; McLauchlan, 1990; Miranda & Brody, 1992; Quill & Townsend, 1991; Speck, 1991). It has also been argued that even though appropriate theoretical frameworks exist which could assist health care providers to accomplish their interactional goals more successfully in this situation, professionals generally do not frame their presentation of bad news according to any theoretical underpinning (Gillotti & Applegate, 2000). Additionally, these pieces are predominantly authored by clinicians and while they may serve as sound advice for interpersonal communication, little empirical research has been completed that would

provide systematic insight on improving communication competencies of clinicians who unfortunately must engage in the task of delivering bad news. What is suggested are practical strategies and behaviors focusing on when, where, with whom and how the news should be delivered (Brewin, 1991; Charlton, 1992; Davis, 1991; Falloweld, 1993; Graham, 1991; McLauchlan, 1990; Miranda & Brody, 1992; Ptacek & Eberhardt, 1996; Quill & Townsend, 1991; Speck, 1991; Statham & Dimavicius, 1992). This literature is summarized by Ptacek and Eberhardt (1996) and Ellis and Tattersall (1999). Many factors aect care providers abilities to deliver bad news successfully and competently. First, care providers must face the issue of medical disclosure (Waitzkin, 1985), which includes attention to patient autonomy and patient rights (Sell, Devlin, & Bourke, 1993). Other concerns include the uncertainty of medical prognosis (Miranda & Brody, 1992); feelings of failure and attention to face needs in the interaction (Goman, 1959; Miranda & Brody, 1992); expressions of emotion (DiMatteo, 1979; Falloweld, 1993; Kaiser, 1993; Krahn, Hallum, & Kime, 1993; Maynard, 1989; Maynard, 1991; Ptacek & Eberhardt, 1996; Sharp et al., 1992; Swanson, 1993; Wesley, 1996); and lack of training and socialization to a detached style of interaction (Flynn & Hekelman, 1993; Haerty, 1991; McWhinney, 1989; Mizrahi, 1991; Novack, Volk, Drossman, & Lipkin, 1993; Rappaport & Witzke, 1993). Physician expressions of compassion, however, do reduce patient anxiety (Fogarty, Curbow, Wingard, McDonnell, & Somereld, 1999). An examination of the research suggests that much is left to be uncovered in the study of communication competence in this context. A more complete review of literature on bad news delivery can be found in Gillotti and Applegate (2000). Research question The delivery of bad news is a problematic task for health care providers. The present study attempted to examine the bad news delivery process within a competence framework. Applying the theoretical model of communication competence should illuminate particular features of bad news delivery and health care provider competence in terms of specic communicative utterances. The goal of the present study is not to look at the eectiveness of bad news delivery in the target population per se, but to look at the relationships between the very specic communicative categories identied in the coding scheme as they relate to perceptions of competence. We examine not the question of how eectively the medical students in the present study delivered bad news, but what communicative behaviors were associated with perceptions of competence. Thus, the following research question was

1014

C. Gillotti et al. / Social Science & Medicine 54 (2002) 10111023

addressed: What communicative moves are associated with perceptions of competence during a bad news delivery context?

Method There were three phases to the procedure. In the rst phase, 3rd year medical students delivering bad news to a simulated patient were videotaped. The tapes were then coded using the McNeilis (2001) providerpatient interaction coding scheme}the Coordination and Competence System (CACS). The competence of the medical students was then rated by outside observers in an attempt to determine those communicative behaviors that best predict perceptions of competence. Thus, data collection and coding involved several phases. Phase I}taping medical students bad news delivery Participants for Phase I were recruited from a class of third year medical students at a large Southern university. Fifty-four out of 90 medical students agreed to participate in the study. Thirty-nine of the 54 were male, and 15 were female. The participants ranged in age from 23 to 51. The training received by the medical students had already included discussion of bad news delivery. The medical students were videotaped in an 11 min roleplay interaction during the students clinical performance examination for their internal medicine clerkship. This 6-h examination tests students clinical and interpersonal skills. The medical students move from station to station completing dierent tasks. Many of the stations require the students to interact with actors playing patients. The actors were hired by the medical school to play the various standardized patients. The actors for this role play were trained by the senior investigator and the clinicians in charge of the clerkship. There were a total of three actresses who participated in this bad news role play over the course of a 14-month data collection period. Videotaping took place on three separate occasions, as the medical students do the rotations for their clerkships in groups of 30. Each medical student was given 11 min to inform the standardized patient that she was HIV positive. The clinicians in charge of the course and the senior investigator created the character to be played by the actress. The patient, Mrs. Murphy, was 36 years old with two children, ages eight and ten. After 15 years of marriage she discovered that her husband had been repeatedly unfaithful throughout the course of their marriage. After being separated from her husband for six months, she decided to have an HIV test. She has returned to the physicians oce to nd out her results.

This role play was one of many in which the medical students participated during the course of their 6-h examination. All 90 students were required to participate in the role play even though only 54 consented to have their videotaped interactions used in the analyses. The senior investigator was located outside the room designated for the bad news role play station. During the videotaped sessions, the medical students approached the consultation room and were given 2 min to read the instructions for that testing station posted on the door. The bad news station instructions informed the student to tell the patient she was HIV positive and that the results had been conrmed by a Western-blot test. The medical students were also informed that the probable contraction point for the HIV virus was Mrs. Murphys husband. Neither the students nor the actress were given a script. However, the actresses were instructed by the investigator to play the interaction with two dominant emotions, distress and anger, in order to create a challenging context within which the medical students were to deliver the bad news. According to Platt and Keller (1994), one of the most dicult situations faced by physicians involves strong emotional displays from patients. Great consistency was noted in the behavior of the actresses for all medical students by the raters of the tapes (described below), the investigators, and the physician in charge of the testing situation. All three actresses went through similar training sessions with the physician in charge of the test and the senior investigator. The actresses for the 2nd and 3rd rounds were given a copy of the tape of the 1st actress so that they could model their behavior on her behavior. For the purposes of the present study, the context of bad news delivery was dened as encompassing the entire interaction, not just the sentence or two in which the diagnosis of HIV was actually stated. The rationale for this was based upon the observation that past research on bad news has typically focused on attempts at providing additional information or comforting patients about the diagnosis. Those aspects of bad news delivery have been deemed important in past research, and appeared relevant in the present videotapes. The actual statement of prognosis}frequently requiring only a sentence or two}is more straightforward and of less apparent research interest. Phase II}coding communicative behaviors Each medical students tape was coded using the CACS providerpatient coding scheme developed by McNeilis (2001). Based upon the Cegala and Waldron (1992) context-bound model of communicative competence, the CACS focuses upon alignment of utterances as communicators coordinate goals on a turn by turn

C. Gillotti et al. / Social Science & Medicine 54 (2002) 10111023

1015

basis. Alignment implies the interpretation and production of messages that facilitate individual and mutual goals and entails participants attention to the intent and meaning of each others messages. The coding system centers on message content, alignment (two levels) and function of messages. Global message content is assessed with nine categories that range from medical concerns such as diagnosis and treatment to nonmedical concerns. Judgments of goal-matching and contextual appropriateness are inherent in the categorizations. While competent communication cannot occur if interactants are talking about two dierent topics, eective interaction management (Wiemann, 1977) also requires local topic synchronization or uptake. Acknowledgment tokens and interruptions are coded as part of the message content. Matters such as topic change and issue versus event extensions are also coded, allowing for assessment of the extent to which participants address central, peripheral, or dierent substantive matters relative to the previous utterance. This is assessed in the alignment portion of the coding scheme. Finally, the coding scheme assesses how utterances mesh functionally, for example, looking at whether an answer follows a question. Functional meshing contributes to conversational coherence, and the attainment of information and relational goals. Content categories include history taking, current problems, diagnosis, treatment, procedure, prognosis, nonmedical, behavioral, and nonspecic. Alignment has two levels. One involves simple acknowledgement tokens and overlapping talk (interruptions). The second involves more topic assessments of alignment, including issue extensions, event extensions, continuers, pop extensions, topic changes, rst utterances, and immediate pops. Pop extensions and immediate pops refer to statements that return to a previously discussed topic, but vary in how long ago that topic was mentioned. Finally, the 31 functional categories involve tasks relating to information giving, seeking, and verifying, and socioemotional statements. The complete list of categories will be presented in Table 1 in Results section. Each utterance is given ve dierent codes: Content (9 categories), acknowledgement token (absent or present), interruption (absent or present), alignment (7 categories), and function (31 categories). The unit of analysis was the utterance, dened as a word or series of words spoken by an individual constituting a thought or partial thought that may or may not be interrupted by or overlap with other talk by the partner. If there is overlap, the rst unit ends there and the next unit begins with the overlapped talk by the partner. Multiple utterances can occur within one speaking turn. During development of the coding scheme, McNeilis (1996) reported unitizing reliability of 0.96 using Holstis (1969) formula.

Table 1 Frequencies of language categories Lang. category Content History Current Problem Treatment Procedure Prognosis Diagnosis Nonmedical Behavioral Nonspecic Acknow. Token Interruption Alignment: Continuer Event Extension Issue Extension Pop Extension Topic Change First Utterance Immediate Pop Function Information-seeking Closed question Mod. closed question Open question Embedded question Information-giving Solicited answer Elaboration Unsolicited info. Expansion Assertion Agreement Disagreement Correction Justication Explanation Bracketing Verifying Cond. relevant ques. Formulation Restatement Socio-emotional Legitimizing Aect Naming Apology Relational Reinforcement Small Talk Humor Miscellaneousa Polite Directive Directive Qualied Directive Compliance Freq. 45 382 110 125 143 90 1046 146 186 240 105 M 0.83 7.07 2.04 2.31 2.65 1.67 19.37 2.76 3.51 4.53 1.98 SD 1.22 7.78 2.85 2.42 3.00 1.44 10.47 4.51 3.87 4.85 2.90

147 578 785 96 430 73 65

2.77 10.91 14.81 1.81 8.11 1.38 1.23

3.45 6.63 11.09 1.85 5.66 2.11 1.64

321 120 234 35 61 74 64 87 300 96 3 2 36 318 7 15 17 66 86 12 32 80 18 14 1 27 53 5 8

6.06 2.26 4.23 0.66 1.15 1.40 1.21 1.64 5.66 1.81 0.06 0.04 0.68 6.00 0.13 0.28 0.32 1.25 1.63 0.23 0.61 1.51 0.35 0.26 0.02 0.51 1.00 0.09 0.15

5.13 2.56 3.20 0.96 1.55 2.49 1.62 2.02 9.28 2.88 0.23 0.19 1.21 5.91 0.34 0.57 0.78 1.83 1.73 0.51 0.88 1.80 0.79 0.56 0.14 1.03 1.86 0.41 0.41

1016 Table 1 (continued) Lang. category Hedging Incomplete Composite function scores Information-seeking Information-giving Verifying Socio-emotional Total
a

C. Gillotti et al. / Social Science & Medicine 54 (2002) 10111023

Freq. 10 76

M 0.19 1.43

SD 0.59 2.71

700 940 105 571 2353

13.21 17.74 1.98 10.77

6.65 11.77 2.41 7.44

Miscellaneous categories were not included within the composites.

McNeilis also reported reliability with two dierent coders on the various components of the coding scheme: Content=0.83 and 0.80; Acknowledgement Tokens=0.68 and 0.72; Interruptions=1.00 and 1.00; Alignment=0.80 and 0.76, Function=0.70 and 0.70. In the present study, 18 undergraduate coders were trained by one primary coder who had past experience in the use of the coding scheme. Each coder worked on some practice data to establish reliability with the primary coder. The coders were allowed to code the data tapes when practice reliability exceeded 0.80 on unitizing and each individual category. Each coder then coded four of the medical student interactions. Thus, some, but not all, interactions were coded by more than one rater. In such cases, the data from the coder with the highest reliability score were used for analysis. Although both the communication of the medical student and the simulated patient were coded, only the data from the medical students were analyzed for this paper since it was only the competence of the medical students that was under investigation. The sequential nature of the data is still taken into account by the coding scheme, however, in that judgements of each statement made by a medical student were assessed for competence in the context of the utterances preceding and following it. Phase III}assessment of communicative competence The nal phase of the project involved seeking the opinions of lay evaluators of the medical students performances in terms of communication competence. The sample of evaluators consisted of 527 (246 male and 281 female) undergraduate students enrolled in 20 lower division communication courses at the same university. The undergraduate student sample ranged in age from 20 to 53. The lay evaluators represented all academic majors, as the courses in which the data were evaluated fullled a university-wide requirement.

Each undergraduate lay evaluator viewed three videotaped interactions and rated the three medical students on a global communication competence scale. The interviews shown to each class of undergraduates were determined by the order in which the medical students went through the role play. Therefore, the rst three videotaped interactions were shown to the rst class of undergraduate students and so on until all the medical students were evaluated. The undergraduate students were oered an alternative exercise if they did not wish to participate in the evaluation of the medical students. Instrumentation. The general communication competence scale used by the undergraduate, lay evaluators was a 14-item Likert scale modied from Woods (1993). The original items of the scale were derived from competence scales developed by Hosman (1987) and Planalp and Tracy (1980). In Woods (1993) research the items reected three constructs, including general communication competence, social attractiveness, and persuasiveness. The persuasion items were removed and replaced with dimensions that reected person-centered communication behavior, such as autonomy granting and empathic behaviors. This perspective, derived from a constructivist view, is consistent with the assumptions of the CegalaWaldron model. Items ranged from 1 to 5, with a higher score indicating more of the variable. A factor analysis using an oblique rotation was conducted to illuminate the factors of the modied scale. The oblique rotation was chosen due to the probable relationship between the items of the scale (Ferguson & Takane, 1989). For instance, social attractiveness was correlated to communication competence, and empathy to social attractiveness. All the items of the scale were positively correlated with one another to some degree. The factor analysis yielded two factors. The rst factor contained 11 items and addressed those items previously regarded as empathy, social attractiveness, global likability, and four new items related to person-centered communication. Thus, the empathy factor, included questions one, three, four, ve, six, eight, and 1014 (Cronbachs alpha=0.94). Three items cleanly loaded on factor two, the general communication skills factor. Questions two, seven and nine reected speech skills, such as being understandable, easy to follow, and communicating eectively (alpha=0.82). A copy of the questionnaire is available from the senior investigator. Table 2 in Results section presents these items grouped within factors. Data analysis Linear multiple regressions were used to examine communicative behaviors that best predicted ratings of competence. The frequencies of the various coded competence scores were used to predict competence

C. Gillotti et al. / Social Science & Medicine 54 (2002) 10111023 Table 2 Means and standard deviations of perceived competence Variable Empathy Caring Sensitivity Treated patient as an individual Nice Polite Attentive Friendly Pleasant Courteous I liked the medical student I. Overall empathy (averaged) Eectiveness Understandable Easy to follow II. Overall skills (averaged) Mean 2.47 2.57 2.49 2.68 2.88 2.97 2.88 2.69 2.55 2.80 2.40 2.67 2.51 2.74 2.62 2.62 SD 1.34 0.58 0.63 0.47 0.45 0.45 0.39 0.48 0.50 0.51 0.59 0.49 0.57 0.43 0.48 0.47

1017

ratings. The competence ratings were averaged across evaluators of each medical student to yield a mean score for each item per student. The mean competence ratings were then summed to create scores for empathic behavior (alpha on this averaged scale=0.99) and skills (alpha=0.93), consistent with the factor structure mentioned above. In addition to the composites, individual competence ratings were used as dependent measures in follow-up analyses in order to examine the data in more detail. The CACS coding data were also used both as individual categories and as composites. These variables were converted to ratios based upon the total number of utterances per student, as there was great variability amongst the medical students in the number of utterances.

Results Frequencies of communicative behaviors The frequencies with which each of the CACS coded behaviors were noted are reported in Table 1. As the table indicates, the most commonly occurring type of Content was Nonmedical talk. This includes small talk and other topics not directly relevant to the health problem. This topic far exceeded any of the other categories. Within the coded interactions, the medical student and the simulated patient frequently discussed the patients personal situation}her husbands indelity, their current separation, and her children. Because of the nature of the context, very little time was spent on

history-taking. The next largest category was the discussion of the current problem. Neither acknowledgement tokens, which would indicate explicit recognition of the partners previous statement, nor interruptions were very frequent in the present data. Acknowledgement tokens, however, occurred more frequently than interruptions did. Alignment behaviors also showed some variability, with issue extensions being most frequent, followed by event extensions (a continuation of the previous topic, but with some shifting) and topic change. The most frequent function of a message was as a closed question. Second most common was an explanation. The medical students frequently explained the nature of HIV, the type of testing, and so forth, to the patient. The number of explanations was closely followed, however, by assertions. It is likely that the assertions accompanied the explanations. Open questions were also relatively common. All of the remaining functions occurred fewer than 100 times in the 54 interactions. For each medical student, the frequencies of various functions were then summed to represent four basic types of communicative behaviors. Information seeking included closed questions, moderately closed questions, open questions, and embedded questions. Information giving was composed of solicited answers, elaborations, unsolicited information, expansions, assertions, agreement, disagreement, corrections, justications, explanations, and bracketing (mentioning that a topic will be discussed later). Socio-emotional statements were legitimizing aect, naming (commenting on an emotion), apologies, relational statements, reinforcement, small talk, humor and three types of directives. Information verifying was represented by conditionally relevant questions, formulations summarizing the gist of what someone has said, and restatements. These groupings were determined a priori and were based on the theoretical model and past literature. The most common of these function types was information giving, followed by information seeking, and socio-emotional statements. Information verifying statements were much less frequent, accounting for only 105 out of the 2253 coded statements. Note that some components may not sum to 2253 because of missing data. Assessment of communicative competence The means and standard deviations of the competence ratings are reported in Table 2. A higher score indicates more of the variable being measured. The highest score was for politeness. The medical students were generally seen as polite. Attentiveness and niceness scored slightly less than politeness. The lowest score was for liking}the respondents were generally only neutral in their feelings

1018

C. Gillotti et al. / Social Science & Medicine 54 (2002) 10111023 Table 3 Beta weights and probabilities. Analyses with composite behavior ratios as predictors Behavioral variable Verifying Verifying Evaluative variable Beta Probability Partial r 0.35 0.32 0.32 0.31 0.27 0.29 0.39 0.31 0.31

toward the medical students. Liking was followed closely by empathy}the students were not seen as very empathic. It is interesting to note that this item, however, also demonstrated much greater variability than did the other items. Fairly low scores (relative to the other items) were also noted on eectiveness, caring, and pleasantness. On the summed Empathic/Person Centeredness (items 1, 36, 8, and 1014) and Skills (items 2, 7, 9) measures, means also indicated fairly neutral scores. Relations between communication and competence A series of multiple regression analyses was conducted to determine the relationships between the communication frequency data from the CACS and the assessments of competence. Ratios were created for each medical student because of the great variability in the numbers of utterances coded for individual students. The frequencies were divided by the total number of coded utterances. Regressions were rst conducted on the summed Function scores (information giving, seeking, and verifying and socio-emotional statements) as predictors of the Empathic and Skills scores, then as predictors of the individual competence ratings. To examine the impact of the specic communicative behaviors more directly, the next set of analyses used the individual CACS categories (as ratios) as they predicted overall Empathy and Skills. While this approach resulted in a large number of regression tests, with an increased possibility of Type 1 error, the exploratory nature of the research dictated a need for in-depth examination. Predictor variables were not highly correlated within each regression analysis, so multicollinearity was not a problem. Signicant standardized beta weights and probability estimates are reported in Tables 3 and 4. The regressions indicated that increased utilization of the information-verifying ratio was associated with less perceived Skill (composite), less empathy, eectiveness, understandability, politeness, ease in following, and niceness. The information-seeking and informationgiving ratios were both associated with lower levels of friendliness (see Table 3). The second set of regression analyses looked at how ratios of the individual coding categories predicted the overall Empathy and Skills scores. More Empathy was associated with fewer moderately closed questions, solicited answers, expansions, restatements, assertions, relational statements, explanations, and small talk. Higher levels of perceived Skills were associated with fewer open questions, solicited answers, expansions, formulations, assertions, relational statements, bracketing, and nonspecic content (see Table 4). To explore the data further, all of the variables were examined for gender dierences. This included looking

Composite skills Individual empathy item Verifying Eectiveness Verifying Niceness Verifying Understandability Verifying Politeness Verifying Easy to follow Info-seeking Friendliness Info-giving Friendliness

0.36 0.02 0.33 0.02 0.33 0.32 0.29 0.30 0.41 0.43 0.47 0.03 0.03 0.05 0.04 0.005 0.03 0.03

Table 4 Beta weights and probabilities. Analyses with individual ratios as predictors Behavioral predictor variable Nonspecic content Restatement Moderately closed question Solicited answer Expansion Assertion Relational Explanation Small talk Open question Solicited answer Expansion Formulation Assertion Relational Bracketing Composite criterion variable Skills Empathic Empathic Empathic Empathic Empathic Empathic Empathic Empathic Skills Skills Skills Skills Skills Skills Skills Beta Probability

0.31 0.40 0.72 0.42 0.75 0.79 0.54 0.73 0.38 0.50 0.43 0.70 0.47 0.75 0.53 0.68

0.05 0.05 0.003 0.05 0.008 0.03 0.007 0.03 0.04 0.03 0.05 0.02 0.03 0.05 0.01 0.01

for gender dierences (1) in the behaviors of the medical students, (2) based upon the gender of the coder, and (3) based upon the gender of the 3rd party evaluators. Although a few small dierences were noted, the number was small enough that it could have been due to experiment-wise error.

Discussion These results indicate some potentially interesting patterns for our examination of communicative competence within the bad news delivery context. As we interpret these results, it is important to note that the

C. Gillotti et al. / Social Science & Medicine 54 (2002) 10111023

1019

communicative tasks involved in the delivery of bad news may well dier from the requirements of other medical communication. The competence ratings indicate that the medical students were generally not very well liked, although they were not disliked. Most respondents were rather neutral in their feelings toward the medical students. Perhaps this was because the students were not seen as very empathic. In a bad news delivery context, empathy would likely be seen as more important than it might otherwise be. The scores also indicated relatively little amounts of caring, eectiveness, and pleasantness. Pleasantness, of course, might not be a very appropriate behavior for a bad news delivery context, but caring and eectiveness would be appropriate. The students, however, did not do a bad job of delivering the bad news. Perceived competence of communicative behaviors The analyses of the composite communicative functions, as ratios, indicated that information verifying was perhaps the most interesting behavior. More information verifying statements led to perceptions of less skill, eectiveness, understandability, and ease in following. It may be that the information-verifying statements made the interaction appear more awkward. Other analyses indicated that use of verifying statements also had less than positive relational consequences, in that it was associated with less empathy, politeness, and niceness. When a patient has just been told that she is HIV+, raters may well assume that she should not be forced to verify information at that point. Too much information seeking and giving were associated with less friendliness and negative outcomes. It is likely that raters were aware that the bad news delivery context is dierent from other medical contexts, and that care providers should not focus upon information giving or seeking when delivering bad news. When patients are receiving bad news, they may need time to absorb the diagnosis before being overwhelmed with additional information. While it appears that socio-emotional behaviors would be seen as appropriate in bad news delivery, the category of socio-emotional functions included such inappropriate behaviors as small talk, directives, and humor as well as more appropriate ones that might comfort or legitimate aect. Perhaps this is why the composite socio-emotional function was not a signicant predictor of perceived competence, though individual items, such as small talk, were signicant although negatively related. The overall Empathy and Skills behaviors were also associated with several variables. Communicators were seen as more empathic when they relied on fewer moderately closed questions, solicited answers, expansions, restatements, assertions, relational statements,

explanations, and small talk. All of these except relational statements and small talk are communicative moves that could be seen as overkill or information overload in the bad news delivery process. As noted above, small talk, too, could be seen as inappropriate when a patient is being informed that she is HIV+. Relational statements are intended to establish a relationship of some sort. While patients who are receiving bad news may want to be comforted, they may not be quite ready to establish much of a relationship with the care provider. Certainly a patient who is being told that she is HIV+ would have other things on her mind beyond building a relationship with the care provider, although relationship building might well be a concern at a later point. Within the context of an established providerpatient relationship, relationship building may be more relevant than in the present context. The lack of an established relationship is one of the limitations of the present study. Similarly, communicators were seen as more skilled when they relied on fewer open questions, solicited answers, expansions, formulations, assertions, relational statements, bracketing, and nonspecic statements. Bracketing statements indicate that a topic will be discussed later in a visit. While putting o a patients concerns might be acceptable in some contexts, it is likely not seen as acceptable in the bad news delivery process. Expansions, formulations, and assertions are all information exchange types of statements, which would also be seen as inappropriate when the patient is preoccupied with the bad news she has received. Implications It is interesting that alignment codes, conceptually an important portion of the competence model, did not predict communicative skills in the medical students. Alignment has not, however, been a very predictive category in other research applying this model to medical communication (McNeilis, 1995, 1996, 2001). However, when alignment codes are combined with function categories, they tend to become more predictive. The code system is multidimensional and sequential and may require that categories be combined for better comparisons. Such future analyses are in process. It is important to note that the results of this study do provide some evidence that the criteria for communicative competence in the bad news delivery context dier from the traditional medical interview (McNeilis, 2001). In the traditional medical interview, information exchange utterances are more likely to relate to perceptions of skill and relational competence. This is especially true of information verifying and following up on patient initiated topics. Some qualitative research has shown that highly competent care providers com-

1020

C. Gillotti et al. / Social Science & Medicine 54 (2002) 10111023

municate more information about diagnosis, prognosis, and treatment options (McNeilis, 1995). Additional research has also determined that consultation interviews may be characterized by more relationally oriented statements, humor, and legitimizing aect than was found in the bad news context (see also Cegala, 1997). Such communication occurs primarily during nonmedical discussion. Research into providerpatient interaction in other contexts typically indicates that patients desire more information than they usually receive from physicians, especially about prognoses (Thompson, 2000). However, with regard to the delivery of bad news, the question is one of timing. A patient receiving bad news may not be able to absorb much additional information at that point. An astute care provider may plan to keep the initial bad news consultation shorter and schedule a follow-up visit at which additional information is provided fairly shortly thereafter. Limitations The care providers within the present study were 3rd year medical students, not practicing physicians. This limits the generalizability of the study. Although it might be assumed that in a testing situation such as this, students would attempt to communicate as eectively as possible, an examination of the tapes did not indicate very skillful approaches. The neutral scores they received are consistent with this assessment. The low level of skills is cause for concern if the students were indeed trying to be as eective as possible. It is, of course, just as important to provide evidence for poor skills to be eliminated as for good skills to be learned. Those who have more experience in medical communication in general and in bad news delivery, in particular, would most likely have developed their own variations on how to deliver bad news based upon their experiences, thus leading to more variability in the data and perhaps more interesting and realistic results. The relationship between perceived competence and communicative behaviors may be dierent in a more experienced sample. A second limitation was the use of non-patient judges of communicative competence whose perceptions may dier from patients. The third-party observers in the present study were likely focusing on the giver of bad news rather than on their own internal anxieties and needs, as would be the case with real patients. The judges in this study may also have been more demanding because they were in an academic setting and may be better educated than the general population. Within the present study the actresses were asked to rate the competence of the medical students, but their ratings showed very little variance and were of little use for data

analysis. Real patients may show more variance. Using actual bad news delivery would, of course, also greatly increase the validity of the ndings in many other ways, too. It should be kept in mind, however, that the goal of the study was not to determine how care providers deliver bad news, but to determine some rather specic relationships between certain communicative behaviors and perceived competence. The large number of coders used in the present study made maintaining reliability more dicult than it might otherwise have been. Although all of the coders established reliability with one primary coder at the outset, and then only coded four interactions, there was variability amongst coders. The results generated by some coders included more utterances than those generated by others. There were also dierences amongst the coders in terms of unitizing, with some coders recording longer utterances than others. This variability was the rationale for dividing the variables by the number of utterances from each interaction for some of the analyses, but was also a limitation to the present study and should be overcome in future research. Future research should also investigate alternative methods of assessing competence. The measure used in the present study was derived from a constructivist, person-centered communication perspective. While this perspective is consistent with the assumptions of the CegalaWaldron model, it is not the only one that could be used to assess competence (see Cegala, Coleman, & Turner, 1998, for a discussion of this topic).

Conclusion The present ndings oer some tentative conclusions regarding the relations between specic communicative behaviors and perceptions of competence during the process of delivering bad news. It is important for care providers who will be delivering bad news to know that the requirements of delivering bad news dier from those of other medical interactions. These ndings indicate that when delivering bad news, it may be less important to communicate large amounts of information or attempt to verify information than would normally be the case in health care interaction. The validity of such a conclusion, however, depends on the replication of the present ndings in other bad news delivery contexts, varied by gender, health problems, and life circumstances.

References
Anderson, E. A. (1987). Preoperative preparation for cardiac surgery facilitates recovery, reduces psychological distress and reduces the incidence of acute postoperative hyperten-

C. Gillotti et al. / Social Science & Medicine 54 (2002) 10111023 sion. Journal of Consulting and Clinical Psychology, 55, 513520. Applegate, J. L. (1990). Constructs and communication: A pragmatic integration. In G. Neimeyer, & R. Neimeyer (Eds.), Advances in personal construct psychology (pp. 203 230). Greenwich, CT: JAL. Beisecker, A. E. (1990). Patient power in doctor-patient communication: What do we know? Health Communication, 2, 105122. Beisecker, A. E., & Beisecker, T. D. (1990). Patient information-seeking behaviors when communicating with doctors. Medical Care, 28, 1928. Ben-Sira, Z. (1980). Aective and instrumental components of the physicianpatient relationship: An additional dimension of interaction theory. Journal of Health and Social Behavior, 28, 170180. Bowers, L. J. (1999). Back to basics. . . Ive got some bad news. . .. Topics in Clinical Chiropractic, 6(18), 6971. Brewin, T. B. (1991). Three ways of giving bad news. The Lancet, 337, 12071210. Buckman, R. (1992). How to break bad news. Baltimore, MD: The Johns Hopkins Univ. Press. Buller, D., & Street Jr., R. (1991). The role of perceived aect and information in patients, evaluations of health care and compliance decisions. Southern Communication Journal, 56, 230237. Burleson, B. R. (1984). Comforting communication. In H. E. Sypher, & J. L. Applegate (Eds.), Communication by children and adults: Social cognitive and strategic processes (pp. 63104). Beverly Hills, CA: Sage. Cegala, D. J. (1997). A study of doctors and patients patterns of informative exchange and relational communication during a primary medical consultation. Journal of Health Communication, 2, 169194. Cegala, D. J., Coleman, M. T., & Turner, J. W. (1998). The development and partial assessment of the medical communication competence scale. Health Communication, 10, 261288. Cegala, D. J., McGee, D. S., & McNeilis, K. S. (1996). Components of patients and doctors perceptions of communication competence during a primary care medical interview. Health Communication, 8, 128. Cegala, D. J., McNeilis, K. S., McGee, D. S., & Jonas, P. (1995). A study of doctors and patients perceptions of information processing and communication competence during the medical interview. Health Communication, 7, 179203. Cegala, D. J., & Waldron, V. R. (1992). A study of the relationship between communicative performance and conversation participants thoughts. Communication Studies, 43, 105123. Charlton, R. C. (1992). Breaking bad news. The Medical Journal of Australia, 157, 615621. Davis, H. (1991). Breaking bad news. The Practitioner, 235, 522526. Delia, J. G., & OKeefe, B. J. (1982). Impression formation and message production. In M. E. Rolo, & C. R. Berger (Eds.), Social cognition and communication (pp. 3372). Beverly Hills, CA: Sage.

1021

DiMatteo, M. R. (1979). A social-psychological analysis of physician-patient rapport: Toward a science of the art of medicine. Journal of Social Issues, 35, 1233. DiMatteo, M. R., Reiter, R. C., & Gambone, J. C. (1994). Enhancing medication adherence through communication and informed collaborative choice. Health Communication, 6, 253266. Ellis, P. M., & Tattersall, M. H. (1999). How should doctors communicate the diagnosis of cancer to patients? Annals of Medicine, 31, 336341. Falloweld, L. (1993). Giving sad and bad news. The Lancet, 341, 476478. Ferguson, G. A., & Takane, Y. (1989). Statistical analysis in psychology and education. Hilldsale, NJ: McGraw-Hill Book Company. Fisher, B. A. (1982). The pragmatic perspective of human communication: A view from systems theory. In F. E. X. Dance (Ed.), Human communication theory (pp. 192219). New York: Harper and Row. Flynn, S. P., & Hekelman, F. P. (1993). Reality shock: A case study in the socialization of new residents. Family Medicine, 25, 633636. Fogarty, L. A., Curbow, B. A., Wingard, J. R., McDonnell, K., & Somereld, M. R. (1999). Can 40 seconds of compassion reduce patient anxiety? Journal of Clinical Oncology, 17, 371379. Franks, A. (1997). Breaking bad news and the challenge of communication. European Journal of Palliative Care, 4(2), 6165. Gillotti, C. M., & Applegate, J. L. (2000). Explaining illness as bad news: Individual dierences in explaining illness-related information. In B. Whaley (Ed.), Explaining illness (pp. 101120). Mahwah, NJ: Lawrence Erlbaum Associates. Girgis, A., Sanson-Fisher, R. W., & Schoeld, M. J. (1999). Is there consensus between breast cancer patients and providers on guidelines for breaking bad news? Behavioral Medicine, 25(2), 6977. Goman, E. (1959). The presentation of self in everyday life. Garden City, NY: Doubleday. Graham, J. R. (1991). Touching and imparting of bad news. The Lancet, 337, 16081609. Guttman, N. (1993). Information exchange in medical encounters: Problems and problems. In B. D. Ruben, & N. Guttman (Eds.), Caregiver patient communications (pp. 151 168). Dubuque, IA: Kendall Hunt. Haerty, F. W. (1991). Into the valley: Death and the socialization of medical students. New Haven, CT: Yale University Press. Hall, J. A., Roter, D. L., & Katz, N. R. (1988). Meta-analysis of correlates of provider behavior in medical encounters. Medical Care, 26, 657675. Holsti, O. R. (1969). Content analysis for the social sciences and humanities. Reading, MA: Addison-Wesley. Hosman, L. A. (1987). The evaluational consequences of topic reciprocity and self-disclosure reciprocity. Communication Monographs, 54, 420435. Kaiser, R. M. (1993). The challenge of breaking bad news. Hospital Practice, 28(10A), 814. Kaplan, S. H., Greeneld, S., & Ware, J. E. (1989). Impact of the doctor-patient relationship on the outcomes of chronic disease. In M. Stewart, & D. Roter (Eds.), Communicating

1022

C. Gillotti et al. / Social Science & Medicine 54 (2002) 10111023 Novack, D. H., Volk, G., Drossman, D. A., & Lipkin, M. (1993). Medical interviewing and interpersonal skills teaching in US medical schools: Progress, problems, and promise. Journal of the American Medical Association, 269, 21012105. Parathian, A. R., & Taylor, F. (1993). Can we insulate trainee nurses from exposure to bad practice? A study of role play in communicating bad news to patients. Journal of Advanced Nursing, 18, 801807. Parks, M. R. (1994). Communicative competence and interpersonal control. In M. Knapp, & G. R. Miller (Eds.), Handbook of interpersonal communication (pp. 589619). Thousand Oaks, CA: Sage. Pearce, W. B., & Cronen, V. E. (1980). Communication, action, and meaning: The creation of social realities. New York: Praeger. Pendleton, D., & Bochner, S. (1980). The communication of medical information in general practice consultations as a function of patients social class. Social Science and Medicine, 14A, 669673. Planalp, S., & Tracy, K. (1980). Not to change the topic but. . .: A cognitive approach to the management of conversation. In D. Nimmo (Ed.), Communication Yearbook 4 (pp. 237 258). New Brunswick, NJ: Transaction. Platt, F. W., & Keller, V. F. (1994). Empathic communication: A teachable and learnable skills. Journal of General Internal Medicine, 9, 222225. Ptacek, J. T., & Eberhardt, T. L. (1996). Breaking bad news: A review of the literature. Journal of American Medical Association, 276, 496502. Quill, T. E., & Townsend, P. (1991). Bad news: Delivery, dialogue, and dilemmas. Archives of Internal Medicine, 151, 463468. Rappaport, W., & Witzke, D. (1993). Education about death and dying during the clinical years of medical school. Surgery, 113, 163165. Roter, D. (1989). Which facets of communication have a strong eect on outcome}A meta-analysis. In M. Stewart, & D. Roter (Eds.), Communicating with medical patients (pp. 183 196). Newbury Park, CA: Sage. Roter, D. L., Hall, J. A., & Katz, N. R. (1988). Patient physician communication: A descriptive summary of the literature. Patient Education and Counseling, 12, 99119. Roter, D., & Frankel, R. (1992). Quantitative and qualitative approaches to the evaluation of the medical dialogue. Social Science & Medicine, 34, 10971103. Roth, N. L., & Nelson, M. S. (1997,). HIV diagnosis rituals and identity narratives. AIDS Care, 9, 161179. Salander, P., Bergenheim, A. T., Bergstroem, P., & Henriksson, R. (1998). How to tell cancer patients: A contribution to a theory of communicating the diagnosis. Journal of Psychosocial Oncology, 16(2), 7993. Sell, L., Devlin, B., & Bourke, S. (1993). Communicating the diagnosis of lung cancer. Respiratory Medicine, 87, 6163. Sharp, M. C., Strauss, R. P., & Lorch, S. C. (1992). Communicating medical bad news: Parents experiences and preferences. Journal of Pediatrics, 121, 539546. Smith, R. C., & Hoppe, R. B. (1991). The patients story: Integrating the patient- and physician-centered approaches to interviewing. Annals of Internal Medicine, 115, 470477. Speck, P. (1991). Breaking bad news. Nursing Times, 87, 2526. Statham, H., & Dimavicius, J. (1992). How do you give the bad news to parents? Birth, 19, 103104.

with medical patients (pp. 228245). Newbury Park, CA: Sage. Kasch, C. R. (1984). Interpersonal competence and communication in the delivery of nursing care. Advances in Nursing Science, X, 7188. Kasch, C. R., Kasch, J. B., & Lisnek, P. (1998). Womens talk and nurse-client encounters: Developing criteria for assessing interpersonal skill. Scholarly Inquiry for Nursing Practice, 12, 269287. Katz, E., Gurevitch, M., Peled, T., & Danet, B. (1969). Doctor patient exchanges: A diagnostic approach to organizations and professions. Human Relations, 22, 309324. King, P. E. (1991). Communication, anxiety, and the management of postoperative pain. Health Communication, 3, 127138. Krahn, G. L., Hallum, A., & Kime, C. (1993). Are there good ways to give bad news? Pediatrics, 91, 578582. Kreps, G. L., & Query, J. L. (1990). Health communication and interpersonal competence. In G. M. Phillips, & J. T. Woods (Eds.), Speech communication essays to commemorate the 75th anniversary of the Speech Communication Association (pp. 293323). Carbondale, IL: Southern Illinois University Press. Lieberman, D. (1992). The computers potential role in health education. Health Communication, 4, 211226. Maynard, D. W. (1989). Notes on the delivery and reception of diagnostic news regarding mental disabilities. In D. T. Helm, W. T. Anders, A. J. Meehan, & A. W. Rawls (Eds.), The interactional order: New directions in the study of social order (pp. 5467). New York: Irvington. Maynard, D. W. (1991). Bearing bad news in clinical settings. In B. Dervin, & M. J. Voigt, Progress in communication sciences, Vol. 10 (pp. 143172). Norwood, NJ: Ablex. McFall, R. M. (1982). A review and reformulation of the concept of social skills. Behavioral Assessment, 4, 133. McLauchlan, C. A. J. (1990). Handling distressed relatives and breaking bad news. British Medical Journal, 301, 11451149. McNeilis, K.L. (1995). A preliminary investigation of a coding scheme to assess communication competence in the primary care medical interview. Unpublished doctoral dissertation, The Ohio State Univ., Columbus, OH. McNeilis, K.L. (1996). A preliminary investigation of a coding scheme to assess communication competence in the primary care medical interview. Paper presented at the Annual Convention of the Speech Communication Association, San Diego, Nov. McNeilis, K. L. (2001). Analyzing communication competence in medical consultation. Health Communication, 13, 518. McWinney, I. (1989). The need for a transformed clinical method. In M. Stewart, & D. Roter (Eds.), Communicating with medical patients (pp. 2540). Beverly Hills, CA: Sage. Miranda, J., & Brody, R. V. (1992). Communicating bad news. Western Journal of Medicine, 156, 8385. Mizrahi, T. (1991). Getting rid of patients; contradictions in the socialization of physicians. Brunswick, NJ: Rutgers University Press. Morse, B. W., & Piland, R. N. (1981). An assessment of communication competencies needed by intermediatelevel health care providers: A study of nursepatient, nursedoctor, and nursenurse communication relationships. Journal of Applied Communication Research, 9, 3041.

C. Gillotti et al. / Social Science & Medicine 54 (2002) 10111023 Street, R. L. (1991a). Information-giving in medical consultations: The inuence of patients communicative styles and personal characteristics. Social Science & Medicine, 32(5), 541548. Street, R. L. (1991b). Accommodation in medical consultations. In H. Giles, J. Coupland, & N. Coupland (Eds.), Contexts of accommodation (pp. 131156). Cambridge: Cambridge University Press. Swanson, R. W. (1993). Psychological issues in CPR. Annuals of Emergency Medicine, 22, 350. Thompson, T. L. (1994). Interpersonal communication and health care. In M. L. Knapp, & G. R. Miller (Eds.), Handbook of interpersonal communication (pp. 696725). Thousand Oaks, CA: Sage. Thompson, T. L. (1998). The patient/health professional relationship. In L. D. Jackson, & B. Duy (Eds.), Health communication research: A guide to developments and directions (pp. 3756). Westport, CT: Greenwood Press. Thompson, T. L. (2000). The nature and language of explanations. In B. Whaley (Ed.), Explaining illness (pp. 339). Mahwah, NJ: Lawrence Erlbaum Associates. Thompson, T. L. (2001,). Introduction to special issue on coding providerpatient interaction. Health Communication, 13, 14. Van Veldhuizen-Scott, M. K., Widmer, L. B., Stacey, S. A., & Popovich, N. G. (1995). Developing and implementing a

1023

pharmaceutical care model in an ambulatory care setting for patients with diabetes. Diabetes Educator, 21, 117123. Vetto, J. T. (1999). Teaching medical students to give bad news: Does formal instruction help? Journal of Cancer Education, 14, 1317. Vincent, C., Young, M., & Phillips, A. (1994). Why do people sue doctors? A study of patients and families taking legal action. Lancet, 343, 16091613. Waitzkin, H. (1984). Doctorpatient communication: Clinical implications of social scientic research. Journal of the American Medical Association, 252, 24412446. Waitzkin, H. (1985). Information giving in medical care. Journal of Health and Social Behavior, 26, 81101. Watzlawick, P., Beavin, J., & Jackson, D. (1967). Pragmatics of human communication. New York, NY: Norton. Wesley, G. (1996). Aective orientation: Its role in compliancegaining in physicianpatient and nursepatient interactions. Unpublished doctoral dissertation, University of Kentucky, Lexington. Wiemann, J. M. (1977). A model of communicative competence. Human Communication Research, 3, 195213. Woods, E. (1993). An integration of the assessment of verbal and nonverbal components of interpersonal communication competence from a constructivist perspective. Unpublished doctoral dissertation, Dept. of Comm., University of Kentucky, Lexington.