The Effects of Communication Skills Training on Patients

Participation During Medical Interviews 

Recent models of physician-patient communication emphasize information exchange in promoting partnership. Although considerable attention has been given to physicians' information exchange, little research has examined patients' communication contributions. The purpose of this research was to test the effectiveness of a training booklet designed to enhance patients' communication skills in information exchange.

A nested design was used, such that 25 physicians each saw six patients, two patients in each of three communication skills interventions (i.e., trained, informed, control). The dependent variables included several discourse categories designed to assess patients' information seeking, provision, and verifying.

Results indicate that trained patients engaged in more effective and efficient information seeking, provided physicians with more detailed information about their medical condition, and used more summarizing utterances to verify information they received from physicians. Additionally, dyads consisting of trained patients demonstrated a more patient-controlled style of communication than did dyads consisting of informed or untrained patients.

Considerable research has been conducted on physician-patient communication over the last 30 years. [1-3] Most applications of this literature have been directed to improvingphysicians' communication, [4-6] while little attention has been given to patients' communication contributions to the medical interview. [7-10] On the other hand, studies indicatethat many patients could benefit from communication skills training. For example, patients typically do not ask physicians questions, even though virtually all patients claim theywant as much information as possible. [11-15] Other research shows that when patients do seek information they often do so indirectly. [16, 17] The purpose of this study is toassess the effects of a communication skills training intervention on patients' participation during a medical interview. The specific discourse outcomes of interest are addressed in the following sections.

1.1 Patient Participation

Over the last 40 years views of the physician-patient relationship have evolved from a paternalistic model to more recent views where medical decision making is emphasized as a joint negotiation and partnership between physician and patient. [18, 19] Of particular concern in more recent models is the extent and quality of information exchange. Information exchange is the fundamental basis by which patients participate in the medical interview and engage in shared decision making. [20] At the same time, information exchange is recognized as the primary basis by which physicians are able to make accurate diagnoses and effective treatment recommendations. [20, 21]

Despite the centrality of information exchange to the medical interview, methods for defining and measuring the communication of information have been lacking in several ways. [22-24] This is especially so for research into patients' communication, as relatively little attention has been given to patients' communicative contributions to medical discourse.

1.2 Patient Communication Skills Training

Anderson and Sharpe [7] identified only eight published reports devoted to patients’ communication skills training. Most of this effort has focused on patients' information seeking. The results of these studies have been inconsistent, with some research showing that trained patients ask more questions than untrained patients [25-27] and other studies reporting no significant difference in question asking between trained and untrained patients. [28, 29]

To our knowledge no attention has been given to patients' information provision in the communication skills training literature. [7] Given the importance of patients' information provision to physicians' diagnostic decision making, it is surprising that so little effort has been made to instruct patients in effective ways to provide physicians with information about such topics as medical history, symptoms, and treatment. Additionally, the recent emphasis on partnership, shared decision making, and patient participation further underscores the significant role patients' information provision skills play in the medical interview.

Little attention has been given to training patients in information-verifying skills. [7, 30] Such skills are important for patients to employ in checking their understanding of information that is given to them. [26, 30] For example, verifying strategies include requests for repeats of information, questions seeking clarification, and summaries of information that has been received.

Our previous research also supports the utility of training patients in these information exchange skills (i.e., information seeking, provision, and verifying). We have found that both physicians and patients associate these information-exchange skills with perceptions of communication competence during medical interviews, [31] and that such skills promote recall of treatment information [32] and compliance with recommended treatments. [33]

1.3 Components and Assessment of Information Exchange

The research into patient communication skills training has not only been limited with respect to what skills have been examined, it is also limited by the ways in which skills have been assessed.

Information seeking. Researchers typically assess patients' information seeking by counting the frequency of patients' questions. However, little effort has been made to distinguish among different functions of questions or to identify the content of questions. For example, the intent and function of some questions is to solicit unknown information, while other questions serve to clarify information or repair conversational mishaps in some way. [34] Similarly, there is an important difference between patients seeking information about medical topics (such as treatment risks and options) and patients asking questions about administrative matters (e.g., insurance coverage) or social topics (e.g., how was the physician's vacation).

In this study, three types of patient questions are defined, direct, assertive, and embedded. Direct questions consist of closed, moderately closed, or open questions that appear in standard form (i.e., raised inflection toward the end of the utterance and inverted subject-verb order) and function (i.e., are intended to solicit information on some specified topic). Assertive questions are declarative utterances but are marked in some way as serving an information-seeking function. For example, a physician might say, "I want you to tell me about this stomach pain you are having." Or a patient may say, "Another thing I want to ask. I'm having pain in my right shoulder and ..." (followed by a description of symptoms). The term "assertive" is used in reference to these information-seeking utterances to emphasize that they appear as assertions, rather than in standard interrogative form.

The third type of question, embedded, is the most indirect. These utterances are also phrased in declarative form, but they are not marked as having an information-seeking intent. For example, a patient might say, "I tried Pepcid, but Maalox worked better. I don't know why, it just did," or "I've been having trouble sleeping and I think it's due to the medication I'm taking." Within the context of the medical interview, these utterances suggest that the patient is actually asking for information--Why does Maalox work better? Could my medication be affecting my sleep?---even though the utterances are not in interrogative form or are not otherwise marked as information-seeking in intent. Embedded questions are indirect because they place the burden on the conversation participant to determine that information is being sought. Common to all three question types is that they are intended to seek new information.

Information provision. In addition to the lack of attention to patients' information provision, the research into physicians' information provision is often problematic. For example, it is not uncommon for physicians' information provision discourse to be coded using just one category. [22] Tuckett and his colleagues [23, 24] have admonished researchers for their lack of sophistication in assessing information provision, but relatively little improvement is evident in current literature. In this study, information provision was defined in terms of the amount of detail given either in response to a direct question or volunteered (i.e., not prompted by a conversation partner's information-seeking utterance). Three categories were used to assess information provision: solicited replies, elaborations, and unsolicited utterances.

A solicited reply is a direct answer to a question. Solicited replies stand on their own as a minimally adequate response to an information-seeking utterance, but they do not provide any additional information. Elaborations follow solicited replies, in that they add information beyond what is directly asked by the prompting question. For example, Physician: Do you drink alcohol? Patient: "Yes, I have two glasses of wine with my dinner each day." The patient's "yes" is a solicited reply, but the rest of the utterance is an elaboration because the patient provides information about what type of alcohol is consumed, how much is consumed, and how often, even though none of this information was requested in the physician's question. Unsolicited information consists of utterances that provide information without prompting from the conversational partner. They often appear as stand-alone topic changes or topic changes within utterances that address direct questions.

Information verifying. Relatively little attention has been given to patients' information verifying, even though some research indicates that it has clear benefits. [26, 30, 35] In this study, information verifying consists of utterances that seek to clarify information (e.g., Did you say take the medicine three times per day?), requests for repetition of information (e.g., Please say the name of that test again), formulations (i.e., summaries of information that was given), and forecasts of information that will be given or sought.

1.4 Content of Discourse

In general, the discourse categories used to examine information exchange in most studies are largely devoid of content, while emphasizing form and/or function. [24] In this study, all information-exchange categories were also coded for content. The content themes are listed in Table 1. The results reported in this study include participants' utterances about substantive medical topics only (i.e., the first six themes listed in Table 1, plus the "medical condition" theme).

Table 1

Content Themes of the Coding System.

Communication Skills Intervention

Anderson and Sharpe [7] report that a range of intervention strategies has been used by researchers to instruct patients in communication skills, including 10 to 20 minute, face to face didactic sessions, videotape, and printed materials. To our knowledge, no studies have systematically compared the effectiveness of various instructional modes or formats for teaching patients communication skills. However, Anderson and Sharpe report that the largest effect sizes were observed for modes of instruction that provided modeling and/or practice. This is consistent with results of related research into the effects of patient education on compliance with life style modifications [36-38] and considerable research into the effects of self efficacy and health. [39-41]

The method of instruction used in this study (i.e., a work booklet) provided patients with examples (similar to what may be presented via modeling) and a form of practice (achieved by prompting patients to write questions on a variety of topics and to list the details of their symptoms and related medical history). We compared patients who received this work booklet 2-3 days prior to their scheduled appointment with patients who received a brief summary of information contained in the booklet immediately prior to their appointment, and to patients who did not receive any form of communication skills intervention. We expected the work booklet patients to demonstrate more communication skills than patients who received only a brief summary of information because the work booklet provided information that was supplemented by examples and the opportunity to practice, or at least the opportunity to actively participate by writing concerns, goals, and the details of one's medical condition. In addition to providing a form of modeling and practice, the work booklet patients also were prompted to think about their medical condition, the reason(s) for their appointment, and what objectives they wanted to achieve well before their scheduled appointment. We suspect that most patients do not take the time or exert the effort necessary to think about these matters before their scheduled appointment, let alone 2-3 days prior to the appointment. Finally, we believe that for some patients, perhaps many, the work booklet legitimates active involvement and participation in the medical interview. Several researchers have observed that patients often do not actively participate in the medical interview. [13-15, 42, 43] The hypothesized reasons for low participation vary from factors such as age and culture to patients' concerns about appearing ignorant, foolish, or taking too much of the physician's time. While we do not advocate that the present work booklet addresses all, or even a large part, of the potential factors accounting for patients' low participation, we believe that the information in the booklet and its prompts to ask questions and otherwise actively participate in the medical interview may provide legitimization for such participation at least for some patients.

The purpose of this study was to assess the effectiveness of a communication skills intervention in promoting competent information exchange among patients and physicians. Accordingly, the following hypotheses were tested.

1.5 Hypotheses

H1: When compared to informed and untrained patients, trained patients will:

1a. demonstrate greater information seeking about medical themes;

1b. obtain more information about medical themes;

1c. provide more detailed information about their medical condition;

1d. demonstrate greater use of information-verifying utterances about medical themes.

H2: Dyads consisting of trained patients and physicians will demonstrate more patient-controlled communication about medical themes than dyads consisting of informed or untrained patients and their physicians.

2.1 Design

A nested design was used, such that patients were nested within physicians. Each physician was audio taped with six different patients, two patients in each of following three intervention conditions. The untrained group (N = 51) did not receive any intervention prior to their scheduled appointment. The trained group (N = 50) received an instructional booklet in the U. S. mail 2-3 days prior to their scheduled appointment, while an informed group (N = 49) received a brief written summary of the major points contained in the training booklet in the waiting room prior to seeing the physician.

2.2 Participants

Participants for this study were 25 family practice physicians and 150 patients. The physicians and patients were recruited at nine different locations in and around a large metropolitan area in central Ohio. Nine of the physicians practiced in a large clinic that is part of a university hospital complex, while the remaining 16 physicians practiced in private offices with two to four physicians per site. Table 2 contains demographic information relevant to the patient sample.

Table 2

Frequency Demographics and Mean Illness Severity for Patients in Three Intervention Groups.

^b These patients are new to the physician, but not to the clinic.

^c During the last 4 weeks, to what extent have your physical health or emotional problems interfered with your normal social activities with family, friends, neighbors or others (1 = not at all, 5 = extremely).

^d How much bodily pain have you had during the past 4 weeks (1= none, 6 = extremely severe).

^e In my judgment, my medical condition for this visit is (1 = not severe, 4 = very severe).

^f Medical decision-making for this visit was (1 = not complex, 4 = very complex).

^g This patient's medical condition (for this visit) is (1 = not severe, 4 = very severe)

Among physicians, 17 were males, 8 were females. Twenty two physicians were White, 3 were African American. On average, physicians were 11 years past residency (range: 1 month to 36 years).

2.3 Procedures

Data collection was completed at one location before moving to another site. The data were collected from July 15, 1997 to November 7, 1997.

Patient selection and assignment. Each patient listed on the appointment records for a given day was assigned a number. Patients were randomly selected and randomly assigned to an intervention, then telephoned to determine their interest in participating in the study. Patients were told that their physician had agreed to participate in a study of physician-patient communication and that they were being contacted to determine if they had interest in participating in the study as a patient. Overall, 84% of the patients contacted agreed to participate in the study.

Physician selection. Physicians agreed to participate in the study prior to data collection. They were told they would be audio taped with six different patients. They knew that a portion of the patients would receive an educational intervention of some kind, but they did not know any of the specific content or objectives of the intervention. To further mask intervention conditions, untrained patients were given a copy of the consent form with a cover exactly like the cover of the training booklet and brief summary given to trained and informed patients. While we did not have visual access to the participants, and thus could not observe possible nonverbal displays of these materials during the interview, we were able to monitor what was said. Only two patients (one trained and one informed) made explicit reference to the materials during the interview. Additionally, physicians typically did not know which interviews were being taped because they had no way of knowing if the microphone in the examination room was operational or not, and taping usually was done over the course of several hours whereby physicians saw a mixture of patients who were and were not part of the study.

Interventions. The 14-page training booklet was based on previous work in physician-patient communication [16, 31, 32, 35, 44, 45] and results of an assessment of earlier versions of the training materials. [45] The booklet was designed to instruct patients in information provision, seeking, and verifying. It was formatted like a workbook with examples and space for notes.

Regarding information provision, patients were first instructed to list the topics they wanted to discuss with the physician. Additionally, they were encouraged to consider any psychosocial issues relevant to their medical condition, such as worries, stress, or feelings of depression. Next, they were instructed to list items of personal and family history relevant to the topics they wanted to discuss (e.g., Had the patient seen a physician about the problem before? How was the problem treated?). Then, patients were instructed to respond to a series of questions regarding symptoms (e.g., What symptoms were experienced? How long had they experienced them? How often did they occur?). In addition, patients were asked to specify anything that helped to alleviate symptoms and what they expected the physician to do about their medical condition.

The next section of the booklet addressed information seeking. The following topics were covered, each with several sample questions: diagnosis, recommended medication(s), behavioral treatment recommendations (e.g., exercise, diet), and prognosis. This section ended with space for patients to write any additional questions they wanted to ask that were not already covered by the topics and sample questions.

The last section of the booklet was designed to instruct patients in information verifying. They were reminded that, when necessary, they could check on their understanding of information they received from the physician by asking questions of clarification, repeating what the physician had just said, or summarizing their understanding of what was said. Each of these strategies was illustrated by examples.

The booklet was analyzed for readability using the Flesch Reading Ease and Flesch-Kincaid Grade Level indices. The reading ease score was 68.96, which falls within the range for standard reading difficulty. The Flesch-Kincaid score was at the fifth grade level. An evaluation form was mailed with the booklet, which trained patients were asked to complete after using the booklet to prepare for their appointment. The evaluation form was developed for and used in previous research which pilot tested an earlier version of the training booklet. [45] Overall, the booklet was evaluated as useful and informative.

The informed group received a brief summary of the major points covered in the training booklet. Although they received information on major points, informed patients were not given sample questions or other examples to illustrate the ideas presented. However, they were encouraged to engage in such behavior as organizing their thoughts, writing down important items, expressing their concerns, asking questions, and using information-verifying strategies to make sure they understood information that was given to them. In virtually all instances, informed patients had considerable time to read and think about the recommendations provided before seeing the physician.

Administration of materials. All patients were met in the waiting room by one of the researchers or an assistant. They were given a pre-interview questionnaire to complete (not relevant to this report) and were asked to sign a consent form.

Trained patients were then asked if they experienced any problems using the training booklet, and the booklet was briefly examined for evidence of usage (e.g., written notes, underlining). Informed patients were given a brief summary of key points covered in the training booklet and encouraged to read the summary before seeing the physician. Untrained patients were simply told that they would soon be taken to an examination room to await the physician.

Two examination rooms at each site were equipped with wireless microphones. The recording equipment and base of operation at each site was not visible to the microphoned examination rooms. When patients were taken to the exam room, the recording equipment was turned on. When the physician arrived, recording began and a stop watch was started to record the length of the interview.

The entire interview was monitored as it was recorded. As soon as the interview ended, patients were taken to the waiting room and given two post-interview questionnaires to complete (not relevant to this report). When they completed the questionnaires they were paid ($30 for trained patients, who were asked for a larger time commitment, and $20 for untrained and informed patients). All patients were given a copy of the training booklet. They were encouraged to use it for their next physician appointment or consider sharing it with family or friends.

Telephone survey. Approximately two weeks after the taped appointment patients were telephoned and engaged in an interview designed to assess delayed recall of treatment information and compliance with treatment recommendations made during the taped interview. Recall and compliance data are not examined in this paper, but they are reported in Cegala et al. [33]

2.4 Discourse Coding

As indicated earlier, several categories of information exchange were coded. Three types of information-seeking questions were coded (direct, assertive, embedded); three categories of information provision were coded (solicited replies, elaborations, unsolicited statements), and four categories of information verifying were coded (clarifying questions, requests for repetitions, summaries, and forecasts of information that will be given or sought). Each of these information exchange categories was also coded for themes, as described earlier.

Coding was done using typed transcripts of the 150 medical interviews. All of the patients' utterances were coded but for physicians only information-provision utterances in response to patients' questions were coded. The senior author coded all 150 transcripts blind to the treatment condition of patients. Two research assistants coded a randomly selected 30% of the transcripts after being trained in unitizing and coding procedures. The research assistants also were blind to patients' treatment group.

The participants' discourse was segmented into T-units following rules provided by Hunt. [46, 47] Each T-unit contained one idea, or thought unit. The three coders independently unitized the transcripts prior to coding them. The obtained Kappa reliability among coders was .81.

After unitizing the transcripts, the three coders independently applied the coding scheme to each unit. Each unit was identified with a speaker code, a chronology code (indicating the discourse unit was about medical history or the current medical problem), a themes code, and an information-exchange code. The obtained reliabilities for chronology, themes, and information exchange codes using Holsti's [48] method were: .90, .73, and .82, respectively. The unitizing and content reliabilities obtained on 30% of the transcripts were considered acceptable and sufficient to assume that coding of the remaining 70% of the transcripts was also reliable.

Dependent variables. The dependent variable for hypothesis 1a was computed by summing the frequencies of patients' direct, assertive, and clarifying (i.e., an information-verifying category) questions about medical themes separately for utterances about history and the current medical problem. The frequency of patients' embedded questions was treated as a separate dependent variable. The frequency of physicians' information-provision units about medical themes was summed separately for utterances about history and the current medical problem to create the dependent variable for hypothesis 1b. In addition, the amount of information given by physicians per-patient-question-asked was computed by dividing the physicians' information-provision units by the frequency of patients' questions. The frequency of patients' information-provision units was summed (across solicited answers, elaborations, and unsolicited utterances) to create the dependent variable for hypothesis 1c. Analyses for this hypothesis were also computed separately on the frequencies of the three information-provision categories. The dependent variable for hypothesis 1d was computed by summing the frequencies of the three remaining categories of information verifying (i.e., repetitions, formulations, and forecasts).

Following Roter, Stewart, Putnam, Lipkin, Stiles, and Inui, [49] a measure of patient control was constructed by the following formula:

Patient Questions + Doctors' Information Units

_______________________________________

Doctors' Questions + Patients' Information Units

This measure is similar to one used by Greenfield, Kaplan, and Ware [50] and reflects a description of a patient-controlled/physician-controlled continuum described by Stewart. [51] The larger the score, the more patient-controlled the interview is in comparison to the physician. This score was used as the dependent variable to test hypothesis 2.

2.5 Data Analysis

MANOVAs or ANOVAs were used to analyze the data. Since patients were nested within physicians, the physician-by-treatment mean square was used as the error term. Planned paired contrasts were computed for each dependent variable, such that intervention groups were compared to each other.

Hypothesis 1b predicted that trained patients would engage in more information seeking about themes relevant to their medical condition than informed or untrained patients. Because there were too few questions about history (N = 39) and embedded questions (N = 37) to warrant including them as dependent variables, they were dropped from further analysis. An ANOVA was computed on the frequency of patients' medically related questions about their current medical problem (N = 548). The results of the ANOVA revealed a significant main effect for intervention (F [2,48] = 3.84, p = .03, eta² = .14). No other effects were significant. Paired contrasts on the means reported in Table 3 indicated that trained patients asked significantly more medically related questions than their informed (t = 2.08, p = .02, one-tailed) or untrained (t = 2.62, p = .005, one-tailed) counterparts. The contrast between informed and untrained patients was nonsignificant (t = 0.53, p = .62, two-tailed). These results support hypothesis 1a.

Table 3

Means and Standard Deviations^a for Patients' Questions About the Current Medical Problem.

Hypothesis 1b predicted that trained patients would elicit more information from physicians about medical themes than informed or untrained patients. Due to the low frequency of physicians' information-provision utterances about history (N = 55), only physicians' utterances about the current medical problem were analyzed (N = 1,862). The ANOVA on physicians' information-provision utterances resulted in a significant main effect for intervention (F [2,48] = 17.09, p < .0001, eta² = .42). No other effect was significant. Paired contrasts on the means reported in Table 4 indicated that trained patients obtained significantly more information than either informed (t = 4.94, p < .0001, one-tailed) or untrained (t = 5.17, p < .0001, one-tailed) patients. The contrast between informed and untrained patients was nonsignificant (t = 0.16, p = .87, two-tailed).

Table 4

Means and Standard Deviations^a for Information Units Elicited From Doctors About the Current Medical Problem.

Although these results show the relative amount of information obtained by patients, they do not reflect the efficiency of patients' information seeking. Accordingly, hypothesis 1b was also tested by examining how much information patients obtained per-question-asked of physicians. The results of this analysis also were significant for intervention (F [2,48] = 13.49, p < .0001, eta² = .36). No other effects were significant. Paired contrasts indicated that trained patients obtained significantly more information per-question-asked than either informed (t = 4.79, p < .0001, one-tailed) or untrained (t = 4.13, p < .0001, one-tailed) patients. The contrast between informed and untrained patients was nonsignificant (t = 0.72, p = .48, two-tailed). The means relevant to these tests are reported in Table 4. Overall, these results support hypothesis 1b.

Hypothesis 1c predicted that trained patients would provide more information about their medical condition than informed or untrained patients. A MANOVA on patient information provision units about history (N = 3,158) and about the current medical problem (N = 5,150) was computed to test this hypothesis. The multivariate result for intervention was borderline significant (Lambda = .8254, df = 4/94, p = .06, Wilks' effect size ß= .09). Examination of the univariate F tests revealed a nonsignificant effect for patients' information provision about history (F [2,48] = 0.63, p = .54, eta² = .03), and a significant effect for utterances about the current medical problem (F [2,48] = 3.86, p = .03, eta² = .14). Accordingly, paired contrasts were computed on the means for utterances about the current medical problem (see Table 5). These results indicated that trained patients provided significantly more information to physicians than either informed (t = 2.49, p = .008, one-tailed) or untrained (t = 1.88, p = .03, one-tailed) patients. The contrast between informed and untrained patients was nonsignificant (t = 0.73, p = .07, two-tailed). These results support hypothesis 1c.

Table 5

Means and Standard Deviations^a for Patients' Information Provision.

As a follow up analysis to hypothesis 1c, a MANOVA was computed using the separate categories of patients' information provision (i.e., solicited responses, elaborations, unsolicited statements) about the current medical problem as dependent variables. The result was significant for intervention (Wilks' lambda = .7658, df = 6/92, p = .05, Wilks' effect size = .12). Univariate tests indicated a significant result for elaborations (F [2,48] = 4.39, p = .02, eta² = .15) and nonsignificant results for solicited replies and unsolicited utterances (F [2,48] = 1.24, p = .30, eta² = .05; F [2,48] = 2.44, p = .10, eta² = .09, respectively). Table 5 contains the means and standard deviations associated with these tests. Paired contrasts on the elaboration means revealed that trained patients provided significantly more elaborations than either informed (t = 2.50, p = .007, one-tailed) or untrained (t = 1.80, p = .03, one-tailed) patients. The contrast between informed and untrained patients was nonsignificant (t = 0.73, p = .47, two-tailed).

Hypothesis 1d predicted that trained patients would use more information-verifying utterances than informed or untrained patients. Due to the low frequency of patients' information-verifying utterances, history and current problem utterances were combined (N = 170). The ANOVA on patients' verifying utterances was nonsignificant (F [2,48] = 2.24, p = .12, eta² = .09), although the means reported in Table 6 are in the predicted direction. As a follow up analysis, verifying categories that involved requests for repetition were combined, categories that involved summarizing information were combined, and the forecasting categories were combined.

Table 6

Means and Standard Deviations^a for Patients' Information Verifying.

These composite frequencies were used as dependent variables in a MANOVA. The multivariate results were nonsignificant (Lambda = .7807, df = 6/92, p = .07, Wilks' effect size = .12). However, the univariate F test for summarizing utterances was significant (F [2,48] = 3.57, p = .04, eta² = .13). Paired contrasts on these means (see Table 6) revealed that trained patients used significantly more summarizing utterances than either informed (t = 2.46, p = .008, one-tailed) or untrained (t = 1.79, p = .04, one-tailed) patients. The contrast between informed and untrained patients was nonsignificant (t = 0.69, p = .49, two-tailed).

Hypothesis 2 predicted that dyads consisting of trained patients and their physicians would demonstrate more patient-controlled communication about medical themes than dyads consisting of informed or untrained patients and their physicians. The ANOVA results were significant (F [2,48] = 4.21, p = .02, eta² = .15), such that dyads containing trained patients were more patient-controlled (mean = .38) than dyads consisting of either informed (mean = .26, t = 2.18, p = .02, one-tailed) or untrained (mean = .23, t = 2.74, p = .004, one-tailed) patients. The contrast between informed and untrained patients was nonsignificant (t = 0.52, p = .60, two-tailed). These results support hypothesis 2.

Models of physician-patient communication that emphasize joint participation and decision making between physicians and patients stress the importance of information exchange in promoting such a relationship. The attention to information exchange in these models reinforces the notion that patients must first and foremost be informed of their medical condition and options for treatment before they can participate in joint decision making. [19, 52] Accordingly, to the extent that patient communication skills training promotes competent information exchange between physicians and patients, it facilitates a partnership-type relationship whereby physicians and patients are more satisfied with health care and patients experience better health outcomes. [49, 53, 54]

Patients' information exchange discourse was examined with respect to information seeking, provision, and verifying. The results regarding patients' information seeking indicated that trained patients asked significantly more direct, assertive, and clarifying questions about their current medical problem than either informed or untrained patients. This result suggests that training enhanced more effective types of information seeking (i.e., question forms that are clearly recognized as information-seeking attempts by conversation participants, in this case, physicians). Moreover, the results of a post hoc analysis suggest that trained patients' questions were directed to significant, or complex, issues relevant to their medical condition. In particular, when physicians answered trained patients' questions they (the physicians) asked more questions and used more information-verifying utterances in the process of addressing the patients' concerns than when they addressed informed or untrained patients' questions (F [2, 48] = 3.32, p = .04, eta² = .12; F [2, 48] = 3.48, p = .04, eta² = .13 for questions and verifications, respectively). This suggests that trained patients' questions were sufficiently complex to require physicians to seek additional information and/or verification before they could address the patients' concerns. In addition to training's impact on the effectiveness of information seeking, results also indicated that trained patients elicited more information from physicians per-question-asked than either informed or untrained patients. Thus, communication skills training enhanced both the effectiveness and efficiency of patients' information seeking.

Regarding information provision, trained patients provided significantly more information about their current medical problem than either informed or untrained patients. Follow-up analyses on the individual categories of information provision indicated that this difference was primarily due to the frequency of trained patients' elaborations. Elaborations are utterances that provide detail beyond the minimum required to answer a direct question, thus trained patients provided physicians with more detailed information about their current medical condition than their informed or untrained counterparts. A post hoc analysis revealed that trained patients provided more elaborations per-question-asked by doctors (F [2, 48] = 3.32, p = .04, eta² = .12), thus further suggesting that trained patients addressed physicians' concerns in greater detail than other patients. Examination of the frequencies of the sub-categories of the symptom and treatment themes indicates that trained patients were especially informative about descriptions of their symptoms (e.g., where they occurred, how often they occurred, how long they lasted, explanation of co-occurrences). Trained patients' information provision about treatments focused especially on description of the treatment, explanation of how the treatment is done, and assessment of its effectiveness.

The results relevant to information verifying were less supportive of the communication skills intervention. However, one result concerning information verifying was significant both statistically and practically. Trained patients engaged in more summarizing of physicians' information provision utterances than did informed or untrained patients. While summarizing is only one form of information verifying, it is especially important for at least two reasons. First, summarizing what the physician has just said (e.g., instructions for taking medication) is an especially effective way to check one's understanding of information because it requires that one re-state the physician's utterances in terms that make sense to self. Such active participation in the consultation is likely to enhance patients' understanding and recall of information. [31, 33] Second, unlike asking clarifying questions or requests for re-statement, summarizing the physician's utterances indicates that patients have already processed enough of the physicians' talk to be able to synthesize essential information. This suggests that trained patients were more cognitively and communicatively engaged in the consultations then informed or untrained patients.

Overall, the results concerning the discourse variables indicate that trained patients were more communicatively active during their consultations than informed or untrained patients. This was further supported by the results concerning the patient-control measure. Dyads consisting of trained patients and their physicians were characterized as more patient-controlled than informed or untrained dyads. However, the actual degree of patient control was not excessive, thus allowing for, if not promoting, patient-physician partnership.

4. 1 Implications for Future Research

The consistent and often substantial differences observed between trained and informed patients suggest that communication skills training requires more than mere reference to desired behaviors. Consistent with Anderson and Sharpe's [7] review, modeling and practice appear to be important components of effective communication skills training. Along these lines, the work booklet in conjunction with other interventions may result in even more effective patient education and performance. [36-38] For example, we employed a face-to-face training procedure in previous research with considerable success. [32] Trained patients asked an average of 6.9 questions per consultation (compared to 4.5 here), accounting for 39% of the variance in intervention conditions (i.e., trained versus untrained). Similarly, trained patients used an average of 3.7 information-verifying utterances (compared to 1.5 here), accounting for 32% of the variance in intervention conditions. In part, the success of the face-to-face training intervention was probably due to our ability to tailor instruction specifically to patients' needs. Research is currently underway assessing the effects of a multi-component communication skills intervention with elderly patients. Additional work is needed to determine the optimal media and delivery for communication skills instruction, particularly with respect to matters of cost-effectiveness and dissemination. We believe a combination of the booklet with a brief follow-up videotape (e.g., played continuously in the clinic waiting room) may provide a cost-effective means of patient communication skills training.

4.2 Practice Implications

The additional, though moderate, effort to employ patient communication skills instruction is likely to enhance patients' participation in decisions about their health. As noted by Kaplan et al. [54] and others, such participation results in measurably better health outcomes. Consistent with this, results not reported here indicate that the trained patients in this sample were more compliant with treatment recommendations, especially behavioral treatments (e.g., diet and exercise) and follow up appointments, than their informed and untrained counterparts. [33] Trained patients also had greater delayed recall of treatment information. [55] In the age of managed care, it is perhaps especially important that these benefits of communication skills training can be obtained without increasing the length of appointments. As is the case for virtually all of the research on patient communication skills training, the interviews involving trained patients in this study were not significantly longer in duration than those interviews involving informed or untrained patients. [55]

Given the benefits of enhanced patient participation, a cost-effective means of communication skills training has important practical implications for clinical practice. The booklet tested in this study has the potential for such cost-effectiveness, as well as ease in dissemination. However, it is not clear to what extent, if any, the conditions of this study prompted trained patients to be more highly motivated, or perhaps obligated, to use the booklet than what may be so under routine circumstances (e.g., unannounced mailing from a clinic). After receiving an executive summary of the results, several physicians in the current sample asked for copies of the booklet, indicating that they planned to give them to particular patients who could benefit most from their use. We do not know if physicians used the booklets in this way, or if so, to what effect. Additional research is needed to determine the potential effectiveness of the booklet under more routine clinical circumstances.

This research was made possible by grant R03 HS90110-01T from The Agency for Health Care Policy and Research. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the Agency for Health Care Policy and Research.

1. Ong LML, DeHaes JCJM, Hoos AM, Lammes FB. Doctor-patient communication: A review of the literature. Soc Sci Med 1995; 40:903-918.

2. Roter DL, Hall JA. Studies of doctor-patient interaction. Annual Review of Public Health 1989; 10:163-180.

3.Thompson TL. Interpersonal communication and health care. In: Knapp ML, Miller GR, eds. Handbook of interpersonal communication. Newbury Park, CA: Sage, 1994:696-779.

4. Branch WT, Arky, R. A., Woo, B., Stoeckle, J. D., Levy, D. B., & Taylor, W. C. Teaching medicine as a human experience: A patient-doctor relationship course for faculty and first-year medical students. Ann Intern Med 1991; 114:482-489.

5. Kern DE, Grayson M, Barker LR, et al. Residency training in interviewing skills and the psychosocial domain of medical practice. J Gen Intern Med 1989; 4:421-431.

6. Smith RC, Lyles JS, Mettler J, et al. Effectiveness of intensive training for residents in interviewing skills. Ann Intern Med 1998; 128:118-126.

7. Anderson LA, Sharpe PA. Improving patient and provider communication: A synthesis and review of communication interventions. Patient Educ and Couns 1991; 17:99-134.

8. Sharf BF. Teaching patients to speak up: Past and future trends. Patient Educ Couns 1988; 11:95-108.

9. Simons-Morton DG, Mullen PD, Mains DA, Tabak ER, Green L. Characteristics of controlled studies of patient education and counseling for preventive health behaviors. Patient Educ Couns 1992; 19:175-204.

10. Webber GC. Patient education: A review of the issues. Med Care 1990; 28:1089-1103.

11. Beisecker AE, Beisecker TD. Patient information-seeking behaviors when communicating with doctors. Med Care 1990; 28:19-28.

12. Beisecker AE. Patient power in doctor-patient communication: What do we know? Health Communication 1990; 2:105-122.

13. Ley PL. Communicating with patients: Improving communication, satisfaction and compliance. New York: Croon Helm, 1988.

14. Parrott R. Exploring family practitioners' and patients' information exchange about prescribed medications: Implications for practitioners' interviewing and patients' understanding. Health Communication 1994; 6:267-280.

15. Street RL, Jr. Information-giving in medical consultations: The influence of parents' communicative styles and personal characteristics. Soc Sci Med 1991; 32:541-548.

16. Cegala DJ. A study of doctors' and patients' patterns of information exchange and relational communication during a primary care consultation: Implications for communication skills training. J Health Communication 1997; 2:169-194.

17. Frankel R. Talking in interviews: A dispreference for patient-initiated questions in physician-patient encounters. In: Psathas G, ed. Interaction competence. Washington, DC: International Institute for Ethnomethodology and Conversation Analysis & University Press of America, 1990:231-262.

18. Emanuel EJ, Emanuel LL. Four models of the physician-patient relationship. JAMA 1992; 267:2221-2226.

19. Charles C, Gafni, A., & Whelan, T. Shared decision-making in the medical encounter: What does it mean? (or it takes at least two to tango). Soc Sci Med 1997; 44:681-692.

20. Frederikson LG. Exploring information-exchange in consultation: The patients' view of performance and outcomes. Patient Educ Couns 1995; 25:237-246.

21. Sandler G. The importance of the history in the medical clinic and the cost of unnecessary tests. Am Heart J 1980; 100:928-931.

22. Roter DL, Hall JA, Katz NR. Patient-physician communication: A descriptive summary of the literature. Patient Educ Couns 1988; 12:99-119.

23. Tuckett D, Williams A. Approaches to the measurement of explanation and information-giving in medical consultations: A review of empirical studies. Soc Sci Med 1984; 18:571-580.

24. Tuckett DA, Boulton M, Olson C. A new approach to the measurement of patients' understanding of what they are told in medical consultations. J Health Soc Behav 1985; 26: 27-25.

25. Anderson LA, DeVellis MB, DeVellis RF. Effects of modeling on patient communication, satisfaction, and knowledge. Med Care 1987; 25:1044-1056.

26. Robinson EJ, Whitfield MJ. Improving the efficiency of patients' comprehension monitoring: A way of increasing patients' participation in general practice consultation. Soc Sci Med 1985; 21:915-919.

27. Roter DL. Patient question-asking in physician-patient interaction. Health Psychol 1984; 3:395-409.

28. Tabak ER. Encouraging patient question-asking: A clinical trail. Patient Educ Couns 1988; 12:37-49.

29. Thompson SC, Nanni C, Schwankovsky L. Patient-oriented interventions to improve communication in a medical office visit. Health Psychol 1990; 9:390-404.

30. Bertakis KD. The communication of information from physician to patient: A method for increasing patient retention and satisfaction. J Fam Pract 1977; 5:217-222.

31. Cegala DJ, Socha McGee D, McNeilis KS. Components of patients' and doctors' perceptions of communication competence during a primary care medical interview. Health Communication 1996; 8:1-28.

32. Socha McGee D, Cegala DJ. Patient communication skills training for improved communication competence in the primary care medical consultation. J Applied Communication Research 1998; 26:412-430.

33. Cegala DJ, Marinelli T, Post DM. The effect of patient communication skills training on treatment compliance in primary care. Arch Fam Med in press.

34. West C. Medical misfires: Mishearings, misgivings, and misunderstandings in physican-patient dialogues. Discourse Processes 1984; 7:107-134.

35. Cegala DJ, Coleman MT, Warisse J. The development and partial test of the Medical Communication Competence Scale (MCCS). Health Communication 1998; 10:261-288.

36. Burke LE, Dunbar-Jacob J. Adherence to medication, diet, and activity recommendations: From assessment to maintenance. J Cardiovasc Nurs 1995; 9:62-79.

37. Dunbar-Jacob J, Burke LE, Pyczynski S. Clinical assessment and management of adherence to medical regimens. In: Nicassio PM, Smith TW, eds. Managing chronic illness: A biopsychosocial perspective. Washington, D.C.: American Psychological Association, 1995: 313-349.

38. Rudd P. Clinicians and patients with hypertension: Unsettled issues about compliance. Am Heart J 1995; 130:572-579.

39. Bandura A. Self-efficacy: The exercise of control. New York: W. H. Freeman and Company, 1997.

40. McAuley E. Self-efficacy and the maintenance of exercise participation in older adults. Journal of Behavioral Medicine 1993; 16:103-113.

41. Schwarzer R. Self-efficacy in the adoption and maintenance of health behaviors: Theoretical approaches and a new model. In: Schwarzer R, ed. Self-efficacy: Thought control of action. Washington, D. C.: Hemisphere, 1992:217-243.

42. Beisecker AE. Older persons' medical encounters and their outcomes. Research on Aging 1996; 18:9-31.

43. Cassileth BR, Zupkis RV, Sutton-Smith K, March V. Information and participation preferences among cancer patients. Ann Intern Med 1980; 92:832-836.

44. Cegala DJ, McNeilis KS, Socha McGee D, Jonas AP. A study of doctors' and patients' perceptions of information processing and communication competence during the medical interview. Health Communication 1995; 7:179-203.

45. Cegala DJ, Drummond D, McCartney W, Marinelli T. The development and partial assessment of a patient communication skills training procedure, National Communication Association, Chicago, IL, November, 1997.

46. Hunt KW. Differences in grammatical structures written at three grade levels, the structures to be analyzed by transformational methods. Tallahassee, FL: U. S. Department of Health, Education, and Welfare, 1964.

47. Hunt KW. Grammatical structures written at three grade levels. Champaign, IL: National Council of Teachers of English, 1965.

48. Holsti OR. Content analysis for the social sciences and humanities. Reading, MA: Addison-Wesley, 1969.

49. Roter DL, Stewart M, Putnam SM, Lipkin M, Stiles W, Inui TS. Communication patterns of primary care physicians. JAMA 1997; 277:350-356.

50. Greenfield S, Kaplan S, Ware JE. Expanding patient involvement in care: Effects on patients' outcomes. Ann Intern Med 1985; 102:520-528.

51. Stewart MA. What is a successful doctor-patient interview? A study of interactions and outcomes. Soc Sci Med 1984; 19:167-175.

52. Quill TE, Brody H. Physician recommendations and patient autonomy: Finding a balance between physician power and patient choice. Ann Intern Med 1996; 125:763-769.

53. Hall JA, Roter DL, Katz NR. Meta-analysis of correlates of provider behavior in medical encounters. Med Care 1988; 26: 657-675.

54. Kaplan SH, Greenfield S, Gandek B, Rogers WH, Ware JE. Characteristics of physicians with participatory decision-making styles. Ann Intern Med 1996; 124:497-504.

55. Cegala DJ. Communication skills training for primary care patients. Rockville, MD: Agency for Health Care Policy and Research, 1998.

Top

Back to the Publication List