Linguistic Analysis of In-Office Dialogue Among Cardiologists, Primary Care Physicians, and Patients With Mixed Dyslipidemia

Linguistic Analysis of In-Office Dialogue Among Cardiologists, Primary Care Physicians, and Patients With Mixed Dyslipidemia Alan S. Brown, MDa, Lynn Cofer-Chase, MSN, RN, CLSa, and Corey A. Eagan, MPH (Cand)b,* An in-office linguistic study was conducted to assess physician–patient discussions of mixed dyslipidemia. Naturally occurring interactions among 12 cardiologists, 12 primary care physicians, and 45 of their patients diagnosed with low levels of high-density lipoprotein cholesterol and being treated with prescription niacin extended-release were recorded. The participants were interviewed separately after the visit. The transcripts were analyzed using sociolinguistic techniques. Determined from the time at talk and the number of questions asked, the patients were moderately engaged in the visit conversations; however, most communication was physician-driven. Only 6% of the average visit was dedicated to disease education. Conversations about dyslipidemia were characterized by numerous laboratory values but rarely contained clear benchmarking or goal setting. In the postvisit interviews, the patients demonstrated a lack of understanding about their lipid levels and the next steps they should take. Both “HDL” [high-density lipoprotein] and “good cholesterol” were the most frequently mentioned aspects of dyslipidemia in these conversations; however, most physicians did not contextualize these components such that the patients were able to understand and retain the information after the visit. Although the conversations about treatment with niacin extended-release contained detailed information about how to manage the side effect of flushing, they lacked a clear description of this side effect. Also, missing from the dialogue was a balanced discussion of risks and benefits. Communication gaps were observed in the discussions regarding mixed dyslipidemia and its treatment with niacin extended-release. In conclusion, additional research is warranted to assess the efficacy of communication strategies to educate both physicians and patients about this condition and its treatment. © 2010 Elsevier Inc. All rights reserved. (Am J Cardiol 2010;106:51–55) Cardiovascular disease (CVD) is the leading cause of mortality worldwide;1 however, patients’ understanding regarding CVD prevention has been shown to be insufficient.2 In addition, patients’ knowledge of individual risk status has been shown to be poor.3 Studies have also shown an inverse relation between high-density lipoprotein (HDL) cholesterol and the risk of developing CVD; research on the exact role of HDL cholesterol continues to progress.4 Niacin extended-release (NER) has been shown to be effective in increasing low HDL cholesterol levels,5–7 and strategies exist to manage its side effects.8 However, NER might be underused in community practice because of a perception of side effects that are more severe and less manageable than clinical data would suggest. To evaluate the discussions of HDL cholesterol and its treatment, an observational study was conducted using sociolinguistic methods.9 –11 The present study was designed to capture naturally occurring conversations among cardiologists, primary care physicians (PCPs), and patients with dyslipidemia. Of particular interest was patient understanding of the risks and benefits associated with treatment of low HDL cholesterol level with NER.

a

Midwest Heart Specialists, Naperville, Illinois; and bMBS/Vox, Parsippany, New Jersey. Manuscript received December 2, 2009; manuscript received and accepted February 7, 2010. This study was supported by funding from Abbott Laboratories, Abbott Park, Illinois. *Corresponding author: Tel: (973) 352-2067; fax: (973) 352-2567. E-mail address: [email protected] (C.A. Eagan). 0002-9149/10/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2010.02.014

Methods The Independent Investigational Review Board, Inc., approved the observational study in November 2007, and the study began in December 2007-January 2008. Invitations were mailed to 1,600 community-based cardiologists and PCPs from a list of high prescribers of NER. Of these physicians, 91 responded to the mailing. Of these 91, the first 30 who met the eligibility criteria (Table 1), were available during the appropriate period, and agreed to participate were enrolled in the present study. On previously agreed-on days, field researchers were sent to physicians’ offices. The office staff approached all patients who had a regularly scheduled appointment and met the eligibility criteria (Table 2) and invited them to take part in a study of physician–patient communication. All research participants were compensated for their participation; however, none was aware of the study sponsor. Informed written consent was obtained from all physicians and patients, after which the office visit was recorded on videotape and audiotape, without the researcher present in the examination room. Immediately after the visit, the patients participated in interviews with the researcher that were recorded on videoand audiotape. At the end of the day, the physicians were similarly interviewed regarding all patients who had participated that day. Medical records were brought into the interview to aid physicians with recall, but the records were not shared with the researcher. These postvisit interviews revealed the relative match or mismatch of the participants’ www.ajconline.org

52

The American Journal of Cardiology (www.ajconline.org)

Table 1 Physician eligibility criteria

Table 5 Patient demographics

Cardiologist or primary care physician Board certified In practice 2–30 years Spent ⱖ80% of time in direct patient care (vs administrative, research) Has not participated in market research regarding dyslipidemia in past month Not affiliated with Advertising/marketing agency Marketing research agency Food and Drug Association Pharmaceutical company Biotechnology company

Variable

Value

Patients (n) Men (%) Average age (years) Average co-morbidities reported Heart disease (n) Hypertension (n) Seasonal allergies (n) Diabetes (n) Previous heart attack, per physician report (n) Taking average of 1 prescription medication for non-cholesterol cardiovascular condition (n) Health insurance (%) Prescription drug coverage (%)

45 76% 55 3.5 15/45 14/45 13/45 8/45 8/45 45

Table 2 Patient eligibility criteria English-speaking adults who had received a diagnosis of low highdensity lipoprotein cholesterol Patients being treated with Niaspan (brand name) ⱕ6 months Patients being treated with niacin (over-the-counter product) were not eligible to participate Patients who were discussing initiation of Niaspan were eligible to participate

Table 3 Physician demographics Variable

Value

Total physicians (n) Primary care physicians Cardiologists Average practice time (years) Primary care physicians Cardiologists Average age (years) Men (%)

24 12 12 18 17 19 56.9 83%

Table 4 Visits recorded but excluded from final sample Reason for Exclusion Patient was prerecruited by physician Patient taking extended-release niacin ⬎6 months No high-density lipoprotein or extended-release niacin discussion Patient appeared mentally altered/incapacitated Total

98% 92%

remaining 52 patients, whose visits included relatively robust discussions of HDL cholesterol and/or treatment with NER, were enrolled until the desired sample size of 45 had been obtained (Table 5). All physician–patient interactions and postvisit interviews were transcribed using audio recordings. Videotapes provided a method for transcript quality control and facilitated the addition of nonverbal cues that facilitated the flow of dialogue (e.g., nodding, shrugging of shoulders). Body language and the tone of voice were not evaluated in the present study. The transcripts were analyzed using techniques from the field of interactional sociolinguistics.15,16 Specific linguistic analyses included, but were not limited to, quantification of the topics discussed and the time spent on each; quantification and a description of the questions asked and answered; analyses of the key lexical items (i.e. vocabulary or word choices); and “open door/close door” of topics put forth or blocked in conversation. Descriptive statistics were calculated. Results

Patients (n) 5 4 2 1 12

perceptions regarding the issues discussed at the visit, including diagnoses, medication regimens, side effects, and level of treatment satisfaction. The interview questions were designed to uncover what providers meant to convey and what the patients had understood. A sample size of 45 patients was chosen as a robust sample for in-depth analysis, supported by published work in the area of physician–patient communications.12–14 Of the 30 practices, 24 yielded ⱖ1 patient for the final sample (Table 3). None contributed ⬎3 patients, and the average per practice yield was 1.8 patients. A total of 64 interactions were recorded. A total of 12 patients were excluded from the final study sample for the reasons listed in Table 4. The

First, despite a moderate level of patient engagement, most dialogue was physician-driven. A minority of the visit time was devoted to disease education. The typical visit consisted of nearly 10 minutes of talk time. Visits with cardiologists were longer than those with PCPs (11 minutes, 27 seconds vs 9 minutes, 23 seconds). Patients demonstrated a moderate level of engagement by speaking 24% of the words in the dialogue. In addition, 36 (80%) of 45 patients asked questions regarding their lipids, an average of 4.4 questions per visit. More than ½ of the visit time was spent discussing lipid-related topics, including medications, side effects, and test results. Medication-related conversation comprised 18% of the average dialogue (Figure 1). Disease education consisted of only 6% of the visits. Second, the current dyslipidemia dialogue contained many “numbers” without clear benchmarking or goal setting, resulting in apparent patient confusion and a lack of an action plan. Although physicians provided many lipid values in these visits, HDL cholesterol was discussed most often. The HDL cholesterol numbers were also most often recalled by the

Preventive Cardiology/Physician-Patient Discussions of Dyslipidemia

53

Figure 1. Topics discussed and percentage of visits they comprised.

patients after their visit. Importantly, 45% of patients were already taking a statin to control their low-density lipoprotein (LDL) cholesterol. When sharing the test results, the physicians often did not focus on context. They rarely benchmarked the patients’ current test results against previous laboratory values, and they did not explicitly benchmark the patients’ results against the National Cholesterol Education Program guidelines or discuss clear goals. Most patients could not recall their specific laboratory values when asked after the visit; instead, they said things such as “too high” or “a little low.” During the postvisit interviews, the patients rated the importance of knowing their lipid values an average of “8” on a 1 to 10 scale, with 10 “very important.” However, some patients indicated that they relied on their physicians to indicate when they needed to be concerned or to take action about their lipids or cardiovascular health. Patients were unaware of the discrepancy between the importance they placed on their laboratory values and that they could not recall them. Third, although “HDL” and “good cholesterol” were the most frequently mentioned aspects of dyslipidemia during the office visits, most physicians did not successfully communicate its role to their patients. Both HDL and LDL cholesterol were discussed in detail during these 45 visits. In several instances, physicians described the link between the mechanism of action of HDL cholesterol on LDL cholesterol using metaphors focused on trucks. Three physicians in 4 visits likened HDL cholesterol to a truck carrying LDL cholesterol away. Two characterized HDL cholesterol as a “garbage” or “trash” truck, such as in the following example between a 53-year-old male patient and his PCP. Doctor: and finally we did get an improvement in the good cholesterol—the HDL cholesterol. For men we like to be over 40 —in the 40s. The last time you were 32 and this time 36. So, we’re still on the low side with the good cholesterol. Patient: Mm-hmm.

Doctor: It’s good and protective, because it goes into the arteries, it pulls the fat out of the arteries. So, anything that’s deposited by bad cholesterol, the LDL, is reversed by the good cholesterol. It goes in and acts like a trash truck. It removes the trash out of the arteries. Okay? In other cases, physicians intimated a relation in measured blood levels, such as in the following example in which a cardiologist, in a discussion with a 59-year-old female patient, described HDL cholesterol and LDL cholesterol as 2 sides of a “teeter-totter.” Doctor: The good cholesterol, you’re at the cutoff. The goal is over 40 and remember the issue is this—is that good cholesterol and bad cholesterol work kind of like—a teeter-totter is the way I always tell people. As the good goes up, the bad comes down. And the good cholesterol—the reason they call it that is it binds to the bad cholesterol and takes it to your liver and your liver gets rid of it. So, the stuff that plugs up your arteries, you can help get rid of, not just by getting—taking more medicine, but by elevating your good cholesterol. Triglyceride levels of ⬎200 mg/dl affected 17 patients (38%); however, only 7 of these patients’ visits contained a discussion of triglyceride levels. Triglycerides were discussed in visits with 11 of 27 patients with normal triglyceride levels. In addition, although current National Cholesterol Education Program guidelines stress the importance of non-HDL cholesterol in lipid management,17 only 2 visits contained mentions of “non-HDL cholesterol.” These both occurred appropriately in visits with patients whose triglyceride levels were ⬎200 mg/dl. Fourth, communication about NER was often limited, lacking a balanced risk/benefit discussion. Overall, the visits did not contain discussions of specific health-related benefits associated with NER, such as a reduced risk of CVD or the regression of plaque in the

54

The American Journal of Cardiology (www.ajconline.org)

Table 6 Physicians reported benefits of niacin extended-release (NER) were not discussed in detail–increasing good cholesterol mentioned most often Benefits Increases/regulates high-density lipoprotein/ good cholesterol Lowers triglycerides Lowers low-density lipoprotein/bad cholesterol Other patients have done well taking it Lowers risk of heart attack Is protective for your heart Protects heart from blockages/plaque in arteries Reduces risks of problems in arteries Helps more parameters than statins Switches small particles into big ones Total

Mentions (n) Visits (n) 26

26

8 5 3 1 1 1

8 5 3 1 1 1

1 1 1 48

1 1 1 31

Data compiled from 45 physician interviews after patient visits: “What did you discuss today with regard to treatment of dyslipidemia? Did you specifically discuss low high-density lipoprotein? Atherosclerosis? Risk factors?” (multiple mentions per visit possible).

arteries. Only 1 cardiologist stated in 1 visit that NER “. . . will improve your HDL and therefore protect you from having any heart blockages or development of plaque in the arteries.” After the visit, however, physicians reported having described numerous cardiovascular-related benefits of NER (Table 6). Despite the existence of clinical treatment guidelines, visits revealed minimal urgency to address low HDL cholesterol levels. Specifically, in 9 visits, the patients with low HDL cholesterol discussed, but did not initiate, treatment with NER. In 6 of these visits, physicians decided to wait for more blood work before initiating treatment with medication. Nearly all visits (89%) contained discussions about the side effects of medication to treat dyslipidemia. Physicians almost always initiated these discussions, and they rarely asked open-ended questions (n ⫽ 3) during these dialogues. Flushing was the side effect most often mentioned (87% of visits). However, few physicians set patient expectations or provided a description of what flushing actually is. Overall, the conversations did not include information regarding how often flushing episodes occur, how long they might last, and what sensations the patient would experience. Discussion The visits contained little focus on disease education overall. The long-term consequences of dyslipidemia (e.g., coronary artery disease, heart attack, or stroke) were infrequently explained to patients in detail. Although the office visits contained many discussions of the lipid levels, these were rarely benchmarked against previous results or used as a baseline for future goal setting. Instead, brief reviews of numbers and an absence of goal setting often resulted in patient confusion and an overall lack of urgency to treat. It was not clear whether this style of information delivery was representative of the physicians’ lack of understanding of, or low comfort level with, current—and often changing— target lipid guidelines, particularly surrounding HDL cholesterol.

In the present analysis, conversations about treatment with NER appeared to lack several components to ensure successful, long-term adherence to therapy. Discussions focused on side effects (e.g., flushing) but did not describe the benefits of therapy in a way that was perceived as meaningful to patients. In addition, when discussing flushing, physicians offered management tips but did not adequately set patients’ expectations with a description of the side effect itself (e.g., the exact sensations or the frequency and severity with which they might occur). This could negatively influence patient compliance, because, without appropriate information, patients might not be able to understand, recognize, and manage side effects such as flushing and might not understand the potential benefits of the therapy. Despite patients’ desire for understandable information about their medical conditions and related treatments,18 the present study, as has the other previous research,19 has illustrated the complexities in language used in discussions of dyslipidemia. To improve in-office communication around the complex topic of lipid management, physicians can implement several simple, patient-friendly strategies. These include the following: Offering clear, contextualized lipid targets rather than simply focusing on undifferentiated numbers Using an easy-to-understand metaphor, such as a garbage truck, to explain the role of HDL cholesterol in overall cardiovascular health Explaining the role of NER in the treatment of abnormal lipid values—and describing the specific benefits associated with that treatment (e.g., reduced risk of cardiovascular complications such as recurrent nonfatal myocardial infarction) Setting expectations regarding the side effects associated with NER by Defining expectations regarding flushing What it is (i.e., warmth, redness, itching, and/or tingling of the skin) How often it might occur How long an episode might last Providing information on how to manage flushing (e.g., take at bedtime with a low-fat snack, take with aspirin, avoid alcohol, spicy food and hot liquids) Offering patient educational materials and raising awareness regarding support programs Previous research has demonstrated that using similar patient-centered communication techniques can provide valuable clinical information and increase patient and provider satisfaction without a statistically significant increase in visit length.20 The present study had several limitations. The sample size, although appropriate for social science research, was qualitative and not projectable. Additional research with a larger sample could examine differences in conversational patterns between PCPs and specialists and could evaluate the roles of patient and physician age, gender, and ethnicity. The patients in this sample were limited to those being treated with prescription NER or considering initiation of such therapy. Future research could also compare differ-

Preventive Cardiology/Physician-Patient Discussions of Dyslipidemia

ences in communication between physicians and patients being treated with a range of therapeutic agents. Finally, the participants’ awareness of the recording of the office visit and postvisit interviews could be considered a limitation of the research. However, other studies have been conducted on this topic, and a number of them have concluded that the recording and interviews have little to no effect on research subjects’ behavior.21,22 In addition, to minimize the effect of the “observer’s paradox,”23 the participants were informed only that they were participating in a study about physician–patient communication; they were not informed about the specific focus of dyslipidemia or treatment with NER, and they were not made aware of the study sponsor. In conclusion, additional research is warranted to assess the efficacy of strategies to educate physicians and patients about the communication gaps observed in the discussions of HDL cholesterol and treatment with NER. 1. Hobbs FD, Erhardt LR, Rycroft C; From the Heart Study Investigators. The From the Heart study: a global survey of patient understanding of cholesterol management and cardiovascular risk, and physician–patient communication. Curr Med Res Opin 2008;24:1267–1278. 2. van Steenkiste B, van der Weijden T, Timmermans D, Vaes J, Stoffers J, Grol R. Patients’ ideas, fears and expectations of their coronary risk: barriers for primary prevention. Patient Educ Couns 2004;55:301– 307. 3. Whiteside C, Robbins JA. Cholesterol knowledge and practices among patients compared with physician management in a university primary care setting. Prev Med 1989;18:526 –531. 4. Kapur NK, Ashen D, Blumenthal RS. High density lipoprotein cholesterol: an evolving target of therapy in the management of cardiovascular disease. Vasc Health Risk Manag 2008;4:39 –57. 5. Capuzzi DM, Guyton JR, Morgan JM, Goldberg AC, Kreisberg RA, Brusco OA, Brody J. Efficacy and safety of an extended-release niacin (Niaspan): a long-term study. Am J Cardiol 1998;82:74U– 81U, 85U, 86U. 6. Goldberg AC. Clinical trial experience with extended-release niacin (Niaspan): dose-escalation study. Am J Cardiol 1998;82:35U– 41U. 7. Morgan JM, Capuzzi DM, Guyton JR. A new extended-release niacin (Niaspan): efficacy, tolerability, and safety in hypercholesterolemic patients. Am J Cardiol 1998;82:29U–34U, 39U– 41U. 8. Davidson MH. Niacin use and cutaneous flushing: mechanisms and strategies for prevention. Am J Cardiol 2008;101:14B–19B.

55

9. Ainsworth-Vaughn N. The discourse of medical encounters. In: Schiffrin D, Tannen D, Hamilton HE, eds. The Handbook of Discourse Analysis. Malden, MA: Blackwell, 2001:453– 469. 10. Heritage J. Conversation analysis and institutional talk: analyzing data. In: Silverman D, ed. Qualitative Research: Theory, Method and Practice. London: Sage, 1997:161–182. 11. Johnstone B. Qualitative Methods in Sociolinguistics, vol 123. New York: Oxford University Press, 2000. 12. Aranguri C, Davidson B, Ramirez R. Patterns of communication through interpreters: a detailed sociolinguistic analysis. J Gen Intern Med 2006;21:623– 629. 13. Davidson B, Blum D, Cella D, Hamilton H, Nail L, Waltzman R. Communicating about chemotherapy-induced anemia. J Support Oncol 2007;5:36 – 40, 46. 14. Hamilton HE, Nelson M, Martin P, Cotler SJ. Provider-patient inoffice discussions of response to hepatitis C antiviral therapy and impact on patient comprehension. Clin Gastroenterol Hepatol 2006; 4:507–513. 15. Gumperz JJ. On interactional sociolinguistic method. In: Sarangi S, Roberts C, eds. Talk, Work and Institutional Order Discourse in Medical, Mediation and Management Settings. Berlin: Mouton De Gruyter, 1999:453– 471. 16. Hamilton HE. Symptoms and signs in particular: the influence of the medical concern no the shape of physician–patient talk. Commun Med 2004;1:59 –70. 17. Blaha MJ, Blumenthal RS, Brinton EA, Jacobson TA; National Lipid Association Taskforce on Non-HDL Cholesterol. The importance of non-HDL cholesterol reporting in lipid management. J Clin Lipidol 2008;2:267–273. 18. Knapp P, Raynor DK, Berry DC. Comparison of two methods of presenting risk information to patients about the side effects of medicines. Qual Saf Health Care 2004;13:176 –180. 19. Durack-Bown I, Giral P, d’Ivernois JF, Bazin C, Chadarevian R, Benkritly A, Bruckert E. Patients’ and physicians’ perceptions and experience of hypercholesterolaemia: a qualitative study. Br J Gen Pract 2003;53:851– 857. 20. Hahn SR, Lipton RB, Sheftell FD, Cady RK, Eagan CA, Simons SE, Nelson MR. Healthcare provider-patient communication and migraine assessment: results of the American Migraine Communication study, phase II. Curr Med Res Opin 2008;24:1711–1718. 21. Pringle M, Stewart-Evans C. Does awareness of being video-recorded affect doctors’ consultation behavior? Br J Gen Pract 1990;40:455– 458. 22. Speer S, Hutchby I. From ethics to analytics: aspects of participants’ orientations to the presence and relevance of recording devices. Sociology 2003;37:315–337. 23. Labov W. Sociolinguistic Patterns. Philadelphia, PA: University of Pennsylvania Press, 1972:61.