Instrumental, Affective, and Patient-Centered Communication Between Cardiologists and Patients with Low Socioeconomic Status: An Observational Study

ABSTRACT

One of the essential elements in managing health is having adequate communication with health care providers. Unfortunately, patients with a low socioeconomic status (SES) often experience less adequate communication with their doctor. In the current study, we explore and compare the communication of both doctors and patients from lower and higher sociodemographic backgrounds on three factors: instrumental, affective, and patient-centered communication. In total, 45 cardiology consultations were observed, transcribed, and coded (16 low-SES, 16 middle-SES, 13 high-SES). Our analyses showed that, compared to higher-SES patients, low-SES patients voiced less of their concerns, answered questions of the doctor more often with one word, and expressed less utterances overall. Naturally, we found that doctors expressed more utterances overall toward low-SES patients. For doctors, no differences regarding instrumental, affective, or patient-centered communication were found. These findings suggest that low-SES patients are more passive communicators and communication differences based on SES exist predominantly for patients’ communication. The revealed communication differences may lead to a less adequate interaction and potentially worse patient outcomes, further increasing the socioeconomic health gap. Hence, doctors should become even more aware of socioeconomic patient communication differences so that they can appropriately encourage low-SES patients to become more active communicators.

The fundamental idea of equality in health care is that people in equal need should have equal treatment, equal access to treatment, and equal treatment outcomes. Despite many efforts, health inequalities due to differences in socioeconomic status (SES) are still commonplace (Cao et al., 2022; Shavers, 2007; Verlinde et al., 2012). SES is defined as an individual’s position within society in comparison to others and their access to resources (Allen, 2020; Wani, 2019). Low SES is typically associated with higher morbidity and mortality rates (e.g., Loef et al., 2021; Walker et al., 2014), and, compared to people with a high-SES, people with a low-SES generally live 15 years longer in poorer health, and live 6 years less (Centraal Bureau voor de Statistiek, 2017). These SES-related mortality and morbidity rates can be explained by personal conditions, such as level of education, living conditions, health literacy, and income, but also by social conditions, such as the quality of interpersonal relationships.

Adequate communication is an important prerequisite for maintaining high-quality interpersonal relationships, also when it comes to doctor-patient communication (Ha & Longnecker, 2010). Adequate doctor-patient communication is reached when both parties are satisfied, and when the patient recalls health information and is adherent to treatment (e.g., Chandra & Mohammadnezhad, 2021; Ong et al., 1995). Ideally, the interaction should meet the needs of the patient regarding three aspects: instrumental communication (e.g., information giving, question asking), affective communication (e.g., counseling, empathy), and patient-centeredness (e.g., shared decision making, partnership building; Ong et al., 1995; Willems et al., 2005)

Numerous communication differences exist between lower and higher SES patient consultations based on instrumental, affective, and patient-centered communication, with both the patient and doctor impacting doctor-patient communication and its quality. For example, low-SES patients often use a more passive communication style, eliciting a less empathic communication style from their doctor (Willems et al., 2005). Paradoxically, research has shown that low-SES patients specifically value empathy in their doctor (Jessup et al., 2017; Ong et al., 1995). Furthermore, low-SES patients generally receive less disease-related information because doctors assume these patients do not fully understand medical information (Aelbrecht et al., 2015). Access to and understanding of health information are, however, a prerequisite for participating in shared decision making (SDM; Murray et al., 2007). Leaving low-SES patients ill-informed may hamper them from actively participating in the consultation.

Both doctors and patients play a significant role in adequate doctor-patient communication (Schwartzberg et al., 2007; Street et al., 2007), which calls for simultaneously studying how both parties communicate in medical consultations. However, when it comes to communication research, a substantial number of studies have adopted a single-focus approach investigating either the patient’s or doctor’s perspective, with ample studies focusing on patient outcomes as a result of the doctor’s communication (Allen, 2020; Arora, 2003; Desjarlais-deKlerk & Wallace, 2013). Moreover, although research has shown that the numerous differences in doctor-patient can be attributed to SES (Allen et al., 2019; Verlinde et al., 2012; Willems et al., 2005), only few studies have explored these SES-related differences using observations of consultations. Observations, however, are particularly suitable to capture communication of doctors and patients simultaneously, allowing for better insights into their roles in doctor-patient communication.

In this study, we aim to fill these research gaps by systematically exploring how doctors and low-SES patients communicate during consultations, using in-depth observational data. Although previous research has established solid frameworks of doctor-patient communication (e.g., Verlinde et al., 2012; Willems et al., 2005), no studies to date have investigated the impact of SES on doctor-patient communication for the three theorized communication aspects (i.e., instrumental communication, affective communication, and patient-centeredness). Hence, a theory-driven and systematic exploration of these communication aspects from both patients’ and doctors’ perspectives will be conducted to gain novel insights for research and practice into communication with this vulnerable group of patients. In this explorative study, the following research question will be addressed: What is the impact of SES on doctor-patient communication in terms of instrumental, affective, and patient-centered communication during medical consultations?

Socioeconomic status

It is well established that SES is one of the most crucial factors affecting health (Chan et al., 2018). People with low-SES typically have a worse health status than their higher-SES counterparts (Dobransky & Hargittai, 2012). Low-SES often means dealing with poor (work and living) conditions and poor housing quality (Stormacq et al., 2019). These unfavorable conditions are associated with a greater risk for various health issues including communicable diseases, diabetes, depression, and cardiovascular diseases (e.g., Topel et al., 2019; Verma et al., 2017). Consequently, this group is more likely to use health care services, shown by their higher number of visits to general practitioners, more frequent use of medications, and higher hospital admissions (Loef et al., 2021). Since the low-SES population is already vulnerable, receiving adequate care is crucial to regain and maintain their health.

Unfortunately, receiving good health care is not self-evident for low-SES patients as they tend to experience barriers to seeking and using health resources (Nijman et al., 2014). They often have less access to health care, lower health literacy levels, and lower proficiency in using health sources (Malat et al., 2006; Stormacq et al., 2019). Research has shown that various SES indicators, such as education level, have been associated with several health literacy domains, including appraisal and understanding of health information, with disadvantaged socioeconomic groups displaying lower health literacy skills (Beauchamp et al., 2015). Moreover, research has shown that those with a low-SES have fewer positive feelings about the health-care system and the interpersonal aspects of the medical encounter (Malat et al., 2006). In particular, low-SES patients frequently feel less understood and supported by health care practitioners, partly due to experiencing a large distance between them and their doctors (Jessup et al., 2017). As low-SES patients are especially vulnerable to the power imbalance of the doctor-patient relationship, using adequate communication is key to appropriately involve low-SES patients in the interaction (Durand et al., 2014; Ousseine et al., 2019).

Despite increased academic attention for SES in doctor-patient communication, considerable weaknesses exist in the current literature concerning the measurement of SES. SES should typically be measured by education, income, and occupation, as these indicators best reflect someone’s socioeconomic risks and resources (Cambois et al., 2020), but many studies fail to do so. Often, many studies operationalize SES by only one socioeconomic indicator (Laaksonen et al., 2005), typically education level (Allen, 2020; Gagné & Ghenadenik, 2018; Walker et al., 2014). Although education level is one of the most important indicators, it fails to capture other SES-related aspects, including work-related conditions such as income and occupation (Cambois et al., 2020). Furthermore, there is much variation in how SES has been defined and measured, which hampers comparability of these studies (Smith, 2021). Consequently, many of the study findings so far might have drawn incorrect conclusions about the role of SES, which may ultimately lead to wrong health care interventions for low-SES populations.

Socioeconomic status and doctor-patient communication

SES also impacts various aspects of doctor-patient communication (Allen, 2020; Verlinde et al., 2012). To fully grasp the impact of SES, we will first look at what doctor-patient communication entails in general. Much like other forms of social interactions, both interaction partners – here: doctors and patients – affect communication (Arora, 2003; Desjarlais-deKlerk & Wallace, 2013; Schwartzberg et al., 2007). In the medical encounter, doctors and patients have to cooperate and coordinate their talk and, in doing so, they influence one another (Street et al., 2007). Understanding this interactional process is key to understanding the quality of care patients receive (Desjarlais-deKlerk & Wallace, 2013).

Doctor-patient communication almost always involves interactions between individuals in non-equal positions (Ong et al., 1995). As doctors are often a dominant communication partner in the consultation room (Siminoff et al., 2006), quality of care depends for a large part on doctors’ communication style and skills (Street et al., 2007). As such, it is crucial that doctors provide clear medical information, but also that they signal emotional support and show interest in the patient’s concerns, questions, and preferences (Desjarlais-deKlerk & Wallace, 2013; Street et al., 2007, 2015). At the same time, quality of care also depends on the patient’s communication style, as their communication can influence doctors’ communication behavior and beliefs (Hall, 2003; Street et al., 2007). For instance, when patients ask more questions, make requests, and offer their opinions, doctors tend to give more information (Street et al., 2007; Willems et al., 2005). Similarly, patients who express themselves more affectively receive more support, empathy, and reassurance from their doctor (Willems et al., 2005).

Instrumental communication and SES

For doctor-patient communication, the interplay of influences can occur for various types of communication, such as instrumental communication. Providing instrumental communication is one of the main purposes of the medical encounter, and refers to the problem-focused area of the consultation focusing on processes of information exchange and discussion of biomedical issues (Ong et al., 1995). Its primary goal is treating patients’ illness and health concern (Ong et al., 1995). Instrumental communication often includes talking about symptoms, explaining tests and medications, and inquiring information about the patient’s (health) status (Desjarlais-deKlerk & Wallace, 2013).

There are numerous differences regarding the level of instrumental communication for low- versus high-SES patients (Verlinde et al., 2012). First, from the doctor’s perspective, doctors generally give less disease-related information to patients with a lower SES (Aelbrecht et al., 2015, 2019; Willems et al., 2005). Doctors assume that patients of lower sociodemographic backgrounds are not interested in learning about their health or that these patients do not fully understand disease-related information (Aelbrecht et al., 2015; Street, 1991). This subsequently leads to enlarging rather than closing the gap between low and high-SES patients.

Second, from the patient’s perspective, patients with a low SES tend to disclose less information about their health problems (Willems et al., 2005), and ask less questions during consultations (Street, 1991). Thus, not only does the doctor provides less disease-related information, but the patient also impedes the interaction at the level of instrumental communication. As a result, doctors adapt their information giving and tend to provide more information to high-SES than low-SES patients (Blanquicett et al., 2007; Willems et al., 2005). As such, low-SES patients are at risk to end up in a downward spiral: low-SES patients’ passive communication style, the doctor’s expectations of their needs and abilities, and the doctor’s reaction to these patients’ passive communication style all contribute to low-SES patients being less well informed, and them being less likely to partake in important matters such as SDM.

Affective communication and SES

Receiving medical information can be a stressful event for patients. Affective communication has the potential to reduce these feelings by creating an atmosphere of positive affect and social support. This is accomplished through communicative behaviors, such as showing concern, engaging in social talk, paraphrasing, and reassurance (Willems et al., 2005). Hence, affective communication relates to the care part of the consultation and enlarges the potential of developing a trusting doctor-patient relationship (Ong et al., 1995). Having a trusting interpersonal relationship alleviates the stressful consultation and therefore has the potential to lead to a more accurate diagnosis and greater compliance (Haskard et al., 2009; Sep et al., 2014).

Research has shown that especially lower educated patients value the affective part of the consultation compared to higher educated patients (Aelbrecht et al., 2015; Jessup et al., 2017; Ong et al., 1995). According to Aelbrecht et al. (2015), one explanation for this could be that the affective part of the consultation is the field where lower educated patients feel more competent to make comments than they do for the instrumental part of the consultation. Nonetheless, low-SES patients tend to behave more passively in their affective behaviors during the consultation (Willems et al., 2005). For example, low-SES patients voice less of their concerns (Street & Haidet, 2011) and show less emotions (Willems et al., 2005). Higher-SES patients show more affective expressiveness which, in turn, positively influences affective expressiveness of the doctor (Willems et al., 2005). Likewise, doctors express more emotional utterances to higher-SES patients (Siminoff et al., 2006). Moreover, doctors tend to listen less attentively to low-SES patients and engage less in small talk with them (Allen, 2020). As small talk creates an interpersonal bond, low-SES patients may experience less adequate doctor-patient communication. For low-SES patients in particular, a good relationship can help offset the less favorable communication effects of SES (Walker et al., 2014).

Patient-centeredness and SES

Over the years, patient-centeredness has acquired a central role in health care (Chandra & Mohammadnezhad, 2021). Patient-centered care includes behaviors, such as partnership building, SDM, and seeing the patient as a unique person with individual needs (Carrard et al., 2018). In terms of doctor-patient communication, it means that the doctor is more responsive to the patient’s perspective (Chandra & Mohammadnezhad, 2021). A doctor can provide patient-centered communication by approaching the patient with open-ended questions, not interrupting the patient, making sure the patient understands them, understanding the patient’s perspective of the health problem, and engaging in active listening (Hashim, 2017).

Patient-centered care also requires active participation of patients (Halabi et al., 2020). Patient participation includes patient behaviors, such as actively sharing health information, being involved in SDM, and expressing opinions about treatment (Longtin et al., 2010). In patient-centered health care, patients are increasingly expected to be involved in their own care. However, patients’ SES influences the ability to be involved in active participation (Halabi et al., 2020). For low-SES patients specifically, active participation can be difficult as they have a harder time grasping health information than high-SES patients, which may be a result of their lower health literacy skills (Beauchamp et al., 2015.)

Research has shown that low-SES patients tend to participate less actively in doctor-patient interactions compared to higher SES patients (Allen, 2020). Low-SES patients talk less, show less socioemotional behavior, ask fewer questions, and express less concerns (Street & Haidet, 2011; Verlinde et al., 2012; Willems et al., 2005). This could explain why doctors behave less patient-centered with low-SES patients compared to those with a higher SES (Allen et al., 2019; Willems et al., 2005). These patient behaviors come at the expense of receiving more patient-centered care, as doctors may feel like they have to behave in a more directive way (Street & Haidet, 2011; Verlinde et al., 2012). Patient-centeredness and patient-participation thus seem to influence each other.

The present study

Systematic differences seem to exist in how instrumental, affective, and patient-centered communication prevail in doctor-patient communication across different SES levels. These differences can contribute to low-SES patients feeling less understood, less supported, and less qualified to actively participate (Allen, 2020; Halabi et al., 2020; Jessup et al., 2017). Research has already shown that communication differences exist, but recent studies have not yet compared the relative amount of instrumental, affective, and patient-centered communication in consultations, nor have they explored if and to what extent these communication aspects can be linked to socioeconomic differences between patients. More insights into these differences are needed as this will enrich our knowledge on low-SES patient communication. Therefore, in this explorative study the following research questions are formulated:

RQ1:

How do doctors and patients with a low-SES communicate during their consultation in terms of (a) instrumental communication, (b) affective communication, and (c) patient-centeredness? RQ2: How do these interactions differ from higher-SES patients regarding (a) instrumental communication, (b) affective communication, and (c) patient-centeredness?

Methods

Consultations and study populations

A total of 45 doctor-patient consultations were observed at the cardiology department at Elisabeth-Tweesteden Hospital (ETZ) in the Netherlands. Since low-SES has been associated with increased risk to develop cardiovascular diseases (Bucki et al., 2017; Topel et al., 2019; Verma et al., 2017), we investigated doctor-patient communication in the context of cardiology. In our study, the consultations were always preceded by a physical examination by the nurse, which included making an electrocardiogram (ECG), and, oftentimes, by a cycling test and/or an echo of their heart.

Both cardiologists and patients were part of the study population. All cardiologists of the hospital were eligible for the study. A total of 17 cardiologists were approached and 15 agreed to participate in the study. Patients were eligible for the study if they 1) spoke Dutch, and 2) were newly referred patients of the cardiologist. Patients’ participation was voluntary, as the study did not influence the consultation nor their treatment. In total, 48 participants were eligible to participate in the study but the data of three patients were not considered for data analysis, because of a missing survey (n = 1), and a missing part of a consultation (n = 2).

Setting and procedure

Data were collected between November 2020 and March 2021. The first author identified eligible patients at the hospital using the hospital’s appointment system. This system showed photos next to the patients’ name and highlighted newly referred patients. When eligible patients were identified, the first author approached them in the waiting room and asked if they were willing to participate. They were told that – if agreed upon – an audio recorder would be used to record their consultation. Patients who were willing to participate signed informed consent and completed a short questionnaire that assessed demographic variables, including those used to determine SES (i.e., income, education, household composition, occupation, region of residence). As soon as the patient was called into the consultation room of the cardiologist, the first author would place an audio recorder on the table, but did not physically attend the consultation. At the end of the consultation, the first author retrieved the audio recorder and thanked the patient for participating. Ethical approval was obtained from the local feasibility advisory committee of the hospital and the medical ethical committee [NW2020-23].

SES measurements

SES was measured by patients’ education level, neighborhood, income, household composition, and occupation. To determine SES, patients received separate scores (i.e., 1 = low SES, 2 = middle SES, 3 = high SES) for education, neighborhood, and a combined assessment of income, household composition, and occupation based on fixed thresholds. If a minimum of two out of three variables fell into the same level, the patient was classified as having that specific SES level. If all three variables fell into different levels, patients received a middle-SES score. As education is considered the most used SES indicator (Gagné & Ghenadenik, 2018; Walker et al., 2014), education level was the decisive factor in unclear cases, such as lacking information on income or if the three SES scores deviated much (e.g., two high-level scores, one low-level score). Table 1 displays examples of how SES was determined.

Table 1. Four examples of SES determination.

#	Education	Education score	Area-level	Area-level score	Combined assessment income, household composition, occupation	Combination score	Assigned SES-level
P1^a	Lower vocational education	1	−0.36	1	1001–1500; living together, no children; permanent disability, not mentioned	1	Low (three low-SES scores)
P2^b	Higher vocational education	3	0.16	2	1501–2000; living together, one child; fulltime employment, caregiver for the elderly	1	Middle (three different SES scores)
P3^c	Higher vocational education	3	−0.22	1	Not mentioned^d; single, no children; fulltime employment, manager	3	High (two high-, one low-SES score, but high education level → high-SES score)
P4^e	Secondary vocational education	2	0.07	2	1501–2000; living together, two children; fulltime employment, not mentioned	1	Middle (two middle-, one low-SES score)

Note: A score of 1 means low-level score, 2 middle-level score, and 3 high-level score. ^aP1 shows an example of a clear SES case in which all scores were the same. ^bP2 shows an example of a mix of all SES-levels, therefore, receiving a middle-SES score. ^cP3 shows an example of a patient with two of the same scores and one other level score, with this third score is deviating much from the other two scores. As one of the same scores includes education level (which is decisive in cases as such), the patient received a high-SES score. ^dFor this case, income was not mentioned. To determine the score for the combined assessment, we looked at household composition (single, no children), occupation situation (full time working), and profession (manager). Based on these indicators, a high-level score was assigned. ^eP4 shows an example of two of the same scores and one other level score, which does not deviate much for the other two scores. The patient therefore received a SES score corresponding with the other two scores; here middle-SES score.

Education

Patients’ education level was assessed by asking about their highest completed education level. Assigning a low, middle, or high education level score was based on the Dutch standard education classification. A low education level score was given to no education completed, primary school, pre-vocational education, and lower vocational education. A middle score was given to senior secondary vocational education, senior general secondary education, and pre-university education. A high score was given to higher vocational education, and university education.

Neighborhood

To determine patients’ area-level deprivation, a status score was derived from patients’ postal code. Status scores indicate the social status of a postal code area, compared to other postal code areas in the Netherlands. They are based on characteristics of people living in that area, the mean income in a neighborhood, the percentage of people with a low income, the percentage of people with a low education level, and the percentage of people without a job (Sociaal-Cultureel Planbureau, 2017). To assign SES-level scores, the following thresholds were used: lower than 0.01 indicating a low neighborhood-level, scores between 0.01 and 0.33 indicating middle neighborhood levels, and scores of 0.34, and higher indicating high neighborhood levels (Sociaal-Cultureel Planbureau, 2017).

Income, household composition, and occupation

A combination score was created based on patients’ personal income, household composition, and occupation. To determine a score, a decision tree was created. The first step was to look at someone’s income: if income was less than €1,500.-, it was assigned a low-level score. If income was more than €3,500.-, it was assigned a high-level score (CBS, 2017; Goderis et al., 2019). For average incomes (i.e., 1,500.- to €3,500.-), occupation situation and household composition were considered. See Table 1 for more details about and examples of how low-, middle-, or high-level scores were assigned.

Coding procedure and measurements

Development of codebook

Guided by the theoretical perspective of this study, we created a codebook based on the studies of Willems et al. (2005), Street (1991), and the RIAS (Roter & Larson, 2002). The main units of analysis were utterances, defined as expressions in which one linguistic idea arises (Street, 1991). Utterances were binary coded with 0 (utterance does not fit a category) and 1 (utterance fits a category).

Following our codebook, we first coded the speaker, so who pronounced the utterance (e.g., doctor or patient), whether the utterance was a closed or open question, and whether the utterance contained instrumental or affective communication. If the latter question was answered with “yes,” the utterance was further coded in terms of instrumental communication and affective communication. Both contained various subcategories (e.g., diagnosis, social talk, verbal attention). It was possible to have utterances coded for both instrumental communication and affective communication at the same time, but an utterance could only be coded once within an instrumental and once within an affective category. Table 2 presents an overview of the (sub)categories of instrumental and affective communication expressed by doctors and Table 3 presents this same overview for categories expressed by patients. Finally, all utterances were coded for patient-centeredness (e.g., the doctor asked the patient for their opinion, preferences, or wishes) and patient participation (e.g., the patient answered with one word).

Table 2. Codebook categories and examples for doctors’ utterances and Cohen’s kappa scores.

Category	Explanation	κ^a
Instrumental communication
Diagnosis	Medical condition of the patient and results from tests (e.g., ECG, scans)	.77
Generalinformation	Medical information nonspecific to the personal health of the patient, explaining biology	.47
Psychosocial talk	Overall mood and psychological problems	.44
Lifestyle talk	Food, diet, exercise, and smoking	.86
Procedural talk	Practical, logistical, and orientation parts of the consult	.62
Affective communication
Supportive talk	Sympathy and empathy, sensitivity, comforting the patient	.79
Negative talk	Critique and disapproval, indifference	.28^b
Social talk	Non-medical talk, chit-chat	.71
Verbal Attention	Paraphrasing, showing they are listening	.73
Patient-Centeredness
Understanding	Asking the patient if they understand them	1.00^b
Clarification	Asking for clarification	.88
Preference	Asking for the opinion, preferences, or wishes	1.00^b
Next steps	Asking what the patients wants with regard to treatment, testing, and next steps	.29^b

Note: ^aValues between .41 and .60 were considered moderate agreements; between .61 and .79 solid agreements; .80 and higher nearly perfect agreements (Burla et al., 2008). ^bSome values are below .41 and are therefore on the low end, some values are on the high end with 1.0. This is due to the very low number of utterances within those categories. All utterances within these categories were manually checked for acceptable levels of agreement by the first author.

Table 3. Codebook categories and examples for patients’ utterances and Cohen’s kappa scores.

Category	Explanation	κ^a
Instrumental communication
Biomedical talk	Medical condition of the patient and results from tests (e.g., ECG, scans)	.82
Psychosocial talk	Overall mood and psychological problems	.49
Lifestyle talk	Food, diet, exercise, and smoking	.88
Procedural talk	Practical, logistical, and orientation parts of the consult	.51
Affective communication
Concern expression	Expressing worries and fears	.75
Supportive talk	Expressing reassurance and gratitude	.62
Negative talk	Being critical, clearly disagreeing	.26^b
Social talk	Non-medical talk, chit-chat	.73
Verbal Attention	Paraphrasing, showing they are listening	.77
Patient-Centeredness
One-word answer	Answers a question with one word (yes/no)	.64
Clarification	Asking for clarification	.51

Note: ^aValues between .41 and .60 were considered moderate agreements; between .61 and .79 solid agreements; .80 and higher nearly perfect agreements (Burla et al., 2008). ^bThis low agreement is due to the very low number of utterances within the negative talk category. All utterances within that category were manually checked for acceptable levels of agreement by the first author.

Coding procedure

All audio recordings were transcribed verbatim by three student assistants [LD, MB, FS]. The first author checked the transcripts and divided the expressions of both parties into utterances. An example of an utterance is: “How do you feel today?.” All utterances were entered into an Excel file where the utterances could be coded. Utterances were arranged in chronological order of the interaction. Although each utterance was to be coded separately, the context of the utterance and/or the entire conversation were considered in assigning a code.

Two coders [MB, FS] were familiarized with the codebook and coding procedure in three coding training sessions. Both coders were instructed to each code roughly 25% (n = 3,308) of the same consultations to determine intercoder reliability. In case of insufficient reliability (Cohen’s kappa <.41; Burla et al., 2008), a new training session was performed to discuss these scores and make final adaptations to the codebook. Categories with such insufficient scores (i.e., supportive talk, expressed concerns, negative talk, and all patient-centeredness categories) were newly coded by a third coder [AN] and the first author. After newly coding, reliability scores were again computed. The final Cohen’s kappas of doctors’ and patients’ subcategories are found in respectively Tables 2 and 3. Intercoder reliabilities for the first part of the codebook, in which we coded the origin of the utterance, were: the speaker (κ = .98); question-asking (κ = .85); and the question being open or closed (κ = .80).

Analyses

In total, 11,536 utterances were expressed. All utterances not containing instrumental, affective, or patient-centered communication (e.g., knocking on the door, typing on the computer, “erm,” unclear utterances, etc.) and/or utterances not spoken by the doctor or patient (e.g., utterances spoken by a partner or an assistant) were deleted from the dataset. This resulted in a dataset of 10,444 utterances, which represented 90.5% of all utterances. We looked at the data in two ways. First, we explored quotes and absolute frequencies of utterances in the (subcategories of) instrumental, affective, and patient-centered communication to obtain a general impression of the communication between doctors and low-SES patients in the consultations. Second, we performed (multivariate) analyses of variance (ANOVAs and MANOVAs) to examine SES differences concerning the various subcategories. As the consultation length varied between 34 and 602 utterances, proportions scores of communication utterances were calculated to account for variety in consultation length. The proportion scores were used as the dependent variables for the (M)ANOVAs. Proportions were calculated by dividing the number of utterances within a category by the total number of utterances by the doctor or patient (depending on whether utterances within a category were expressed by the doctor or the patient). For post-hoc tests, Bonferroni corrections were used. The data were analyzed using SPSS version 28.

Results

Patient characteristics and consultation descriptives

Patients (N = 45) were on average 61 years old (M = 60.64, SD = 16.07, range = 20 to 90) and 25 patients (55.6%) were women. Sixteen patients were classified as low-SES (35.6%), 16 as middle-SES (35.6%), and 13 patients as high-SES (28.8%). In total, 10,444 utterances were analyzed, of which 5,799 utterances (55.5%) were expressed by doctors and 4,645 utterances (44.5%) by patients. See Table 4 for an overview of the total utterances expressed by doctors and patients divided by patients’ SES levels.

Table 4. Total utterances of doctors and patients.

	Low-SES (n = 16)	Middle-SES (n = 16)	High-SES (n = 13)	Total (N = 45)
All utterances	3,272 (31.3)	3,762 (36.0)	3,410 (32.7)	10,444 (100)
Doctor	1,981 (34.2)	2,048 (35.3)	1,770 (30.5)	5,799 (100)
Patient	1,291 (27.8)	1,714 (36.9)	1,640 (35.3)	4,645 (100)

Note: Descriptives represent absolute frequencies with percentages between parentheses.

When comparing the proportions of the total number of utterances, significant SES differences were found, F(2, 42) = 4.45, p = .018, η_p² = .18. Doctors expressed proportionally more utterances in consultations with low-SES patients (M_doctor = 0.38; SD = 0.12) compared to consultations with high-SES patients (M_doctor = 0.47; SD = 0.10, p = .030). This means that low-SES patients expressed less utterances (M_{low-SESpatient} = 0.62; SD = 0.12) than high-SES patients (M_{high-SESpatient} = 0.53; SD = 0.10, p = .030). No significant differences were found between the low-SES and middle-SES groups (M_doctor = 0.54; SD = 0.05; M_{middle-SESpatient} = 0.46, SD = 0.05, p = .063) or the middle-SES and high-SES groups (p = 1.000).

Instrumental communication

Instrumental communication of the doctor

Instrumental communication utterances of the doctor (n = 3,563) were identified and categorized into five subcategories (RQ1a): diagnosis, procedural talk, general medical information, lifestyle talk, and psychosocial talk. Most instrumental utterances of the doctor fell into the diagnosis category (n = 2,487; 69.8%). An example of a diagnosis utterance of a doctor was: “I don’t see any anemia or other things that could cause this.” The second most commonly expressed instrumental utterances were those regarding procedural talk (n = 749; 21.0%). An example utterance belonging to procedural talk was: “So what I will do is; I will order the test and then I call you.” The third subcategory was general information (n = 148; 4.2%), which can be illustrated by the following quote: “A normal pumping function of the heart is above 55%.” The next subcategory was called lifestyle talk (n = 121; 3.4%), which contained utterances such as: “And alcohol, do you use that?.” The lowest number of utterances were expressed in the psychosocial talk category (n = 61; 1.7%). An example of psychosocial talk was: “Ah and do you experience stress or anxiety?.” Table 5 displays an overview of these subcategories across patients’ SES levels.

Table 5. Total utterances of instrumental communication by the doctor.

	Low-SES (n = 16)	Middle-SES (n = 16)	High-SES (n = 13)	Total (n = 45)
Instrumental	1,304 (36.6)	1,243 (34.9)	1,016 (28.5)	3,563 (100)
Diagnosis	942 (37.9)	836 (33.6)	709 (28.5)	2,487 (100)
Procedural talk	263 (35.1)	302 (40.3)	184 (24.6)	749 (100)
General information	47 (31.8)	50 (33.8)	51 (34.4)	148 (100)
Lifestyle talk	27 (22.3)	37 (30.6)	57 (47.1)	121 (100)
Psychosocial talk	26 (42.6)	19 (31.2)	16 (26.2)	61 (100)

Note: Descriptives represent absolute frequencies with percentages between parentheses.

Across different levels of SES (RQ2a), there were no SES differences with regard to the proportion of instrumental communication of the doctor in total, F(2, 42) = 1.55, p = .069, η_p² = .07. There were also no significant differences for the instrumental communication subcategories: diagnosis: F(2, 42) = 1.60, p = .214, η_p² = .07; procedural talk: F(2, 42) = 0.76, p = .475, η_p² = 03; general information: F(2, 42) = 0.03, p = .968, η_p² = .00; lifestyle talk: F(2, 42) = 1.94, p = .157, η_p² = .08; and psychosocial talk: F(2, 42) = 0.05, p = .949, η_p² = .00). These results suggest that doctors, on average, did not vary their instrumental communication based on patients’ SES.

Instrumental communication of the patient

Instrumental communication utterances of the patients were divided into four different categories (RQ1a): biomedical talk, lifestyle talk, psychosocial talk, and procedural talk. The category with the highest number of utterances was biomedical talk (n = 1,978; 76.0%). An example of biomedical talk of patients is: “And then I got nauseous which stayed like that.” The second category, lifestyle talk (n = 233; 9.0%), is best illustrated with the following quote: “Look, I have been eating normally, yes, normal.” An example of psychosocial talk (n = 207; 8.0%) of the patient is: “Well, it could also be because I can’t really relax.” The final category, procedural talk, had the lowest number of utterances (n = 181; 7.0%) An example of procedural talk is: “So I just need to go the counter to make an appointment?.” An overview of these subcategories across the various SES levels is shown in Table 6.

Table 6. Total utterances of instrumental communication by the patient.

	Low-SES (n = 16)	Middle-SES (n = 16)	High-SES (n = 13)	Total (n = 45)
Instrumental	748 (28.8)	959 (36.9)	892 (34.3)	2,599 (100)
Biomedical talk	593 (30.0)	737 (37.2)	648 (32.8)	1,978 (100)
Lifestyle talk	45 (19.3)	91 (39.1)	97 (41.6)	233 (100)
Psychosocial talk	61 (29.5)	54 (26.1)	92 (44.4)	207 (100)
Procedural talk	49 (27.1)	77 (42.5)	55 (30.4)	181 (100)

Note: Descriptives represent absolute frequencies with percentages between parentheses.

Regarding SES differences in instrumental communication (RQ2a), no significant differences were found in the three SES-groups regarding the proportion of instrumental communication in total, F(2, 42) = 0.76, p = .476, η_p² = .04 Furthermore, no significant differences were revealed for the instrumental communication of the patient subcategories: biomedical talk: F(2, 42) = 0.56, p = .574, η_p² = .03; lifestyle talk: F(2, 42) = 2.04, p = .143, η_p² = .09; psychosocial talk: F(2, 42) = 0.97, p = .389, η_p² = .04; procedural talk: F(2, 42) = 0.48, p = .622, η_p² = .02. Thus, on average, patients did not vary in their instrumental communication based on SES.

Affective communication

Affective communication of the doctor

Affective communication utterances of the doctor (n = 2,390) were divided into four subcategories (RQ1b): verbal attention, supportive talk, social talk, and negative talk. The verbal attention category contained most utterances (n = 1,103; 46.2%). An example of a verbal attention quote of the doctor is: “Uhu, yes, I understand.” Supportive talk had a total of 729 (30.5%) utterances. An example of supportive talk is: “You do not need to be worried about this.” The third category was social talk (n = 532; 22.6%): “Oh yes, all those COVID-measures, it is really something.” Negative talk included the least utterances (n = 42; 1.8%). An example is best illustrated with the following quote: “Yes, but again, we look at the upper pressure, OK?.” Although this quote might not sound negatively perse, the doctor interrupted the patient abruptly. Table 7 presents more information about these subcategories across the different SES levels.

Table 7. Total utterances of affective communication by the doctor.

	Low-SES (n = 16)	Middle-SES (n = 16)	High-SES (n = 13)	Total (n = 45)
Affective	720 (30.1)	858 (35.9)	812 (34.0)	2,390 (100)
Verbal attention	369 (33.5)	415 (37.6)	319 (28.9)	1,103 (100)
Supportive talk	206 (28.6)	242 (33.2)	281 (38.5)	729 (100)
Social talk	136 (25.7)	197 (37.0)	199 (37.4)	532 (100)
Negative talk	9 (21.4)	11 (26.0)	22 (52.0)	42 (100)

Note: Descriptives represent absolute frequencies with percentages between parentheses.

There were no significant SES differences (RQ2b) for the proportion of affective communication of the doctor in total, F(2, 42) = 1.03, p = .365, η_p² = 0.57, nor for the affective communication subcategories: verbal attention: F(2, 42) = 0.96, p = .390, η_p² = .04; supportive talk: F(2, 42) = 0.67, p = .549, η_p² = .03; social talk: F(2, 42) = 0.65, p = .527, η_p² = .03; and negative talk: F(2, 42) = 2.31, p = .112, η_p² = .10. This may suggest that doctors did not vary in their affective communication to patients across different levels of SES.

Affective communication of the patient

Affective communication utterances of the patients (n = 2,106) could be subdivided into five different subcategories (RQ1b): verbal attention, social talk, concern expression, supportive talk, and negative talk. Most utterances were expressed in the verbal attention category, with the following quote illustrating verbal attention (n = 1,157; 54.9%): “Uhu, yes, yes.” The second subcategory, social talk, contained 520 utterances (24.7%). An example of social talk is: “Ha ha, yeah, no problem when you have an electric bike, right?” The third most commonly expressed affective utterances were those regarding concern expression, which encompassed 235 (11.1%) utterances in total. An example of a quote in which the patient expressed a concern is: “And that really scared me, you know.” Supportive talk of the patient (n = 183; 8.7%) is best illustrated with the following utterance: “Oh that is great, thank you.” The final category, negative talk, had the lowest number of utterances (n = 27; 1.3%). An example of negative talk is: “well, I think differently.” See Table 8 for an overview of the subcategories across the various SES levels.

Table 8. Total utterances of affective communication by the patient.

	Low-SES (n = 16)	Middle-SES (n = 16)	High-SES (n = 13)	Total (n = 45)
Affective	544 (24.8)	780 (37.0)	782 (37.1)	2,106 (100)
Verbal attention	364 (31.7)	419 (36.2)	374 (32.3)	1,157 (100)
Social talk	95 (18.3)	203 (39.0)	222 (42.7)	520 (100)
Concern expression	43 (18.3)	86 (36.6)	106 (45.1)	235 (100)
Positive talk	37 (20.2)	69 (3.3)	77 (42.1)	183 (100)
Negative talk	8 (29.6)	7 (25.9)	12 (44.5)	27 (100)

Note: Descriptives represent absolute frequencies with percentages between parentheses.

No significant differences were found across the three SES groups (RQ2b) for the proportion of affective communication by the patient in total, F(2, 42) = 0.56, p = .576, η_p² = .03. Nonetheless, a significant difference was revealed for concern expression across SES level, F(2, 42) = 3.36, p = .04, η_p² = .14. Specifically, a significant difference was found between low-SES (M = 0.02, SD = 0.02) and high-SES patient groups (M = 0.05, SD = 0.04), p = .043, demonstrating that low-SES patients expressed less of their concerns during consultations compared to high-SES patients. No significant differences were found for the different SES levels and the proportion of verbal attention, F(2, 42) = 1.77, p = .183, η_p² = .08, social talk, F(2, 42) = 0.28, p = .759, η_p² = .01, positive talk, F(2, 42) = 0.19, p = .829, η_p² = .01, and negative talk, F(2, 42) = 1.05, p = .360, η_p² = .05.

Patient-centeredness

Patient-centeredness of the doctor

Patient-centeredness of the doctor (RQ1c) included partnership building and SDM, but we first looked at question asking in general, as the type of questions asked also provided insight in how patient-centered the consultations were. With regard to question asking, 909 (15.7%) of all utterances were questions. Of these questions, 713 utterances (78.4%) were closed questions such as “Do you experience this often?,” and 196 (21.6%) were open questions such as “And if that happens, how does the pressure on your heart feel?.” With regard to patient-centeredness, 776 (13.4%) utterances reflected partnership building, which are illustrated with the quote: “And what did your husband say about this?.” A total of 290 (5.0%) utterances were classified as SDM. An example quote is: “What shall we agree upon?.” Table 9 presents an overview of these data.

Table 9. Question asking and patient-centeredness (i.e., partnership building, and shared decision making) by the doctor.

	Low-SES (n = 16)	Middle-SES (n = 16)	High-SES (n = 13)	Total (n = 45)
Questions	352 (38.7)	299 (32.9)	258 (28.4)	909 (100)
Closed questions	279 (39.1)	233 (32.7)	201 (28.2)	713 (100)
Open questions	73 (37.2)	66 (33.7)	57 (29.1)	196 (100)
Patient-centeredness	339 (31.8)	362 (34.0)	365 (34.2)	1,066 (100)
Partnership building	224 (28.8)	284 (36.4)	268 (34.5)	776 (100)
Shared decision making	115 (39.7)	78 (26.9)	97 (33.4)	290 (100)

Note: Descriptives represent absolute frequencies with percentages between parentheses.

No significant differences were found across the three SES groups (RQ2c) for the proportion of patient-centeredness by the doctor in total, F(2, 42) = 0.60, p = .556. No significant SES differences were found for the proportion of partnership building utterances by the doctor, F(2, 42) = 0.47, p = .626, η_p²= .02, nor for the proportion of SDM utterances by the doctor, F(2, 42) = 1.34, p = .272, η_p²= .06. As such, doctors did not vary partnership building and SDM based on their patient’s level of SES. Regarding question asking by the doctor, there were also no significant differences for the different SES levels, F(2, 42) = 0.54, p = .588, η_p² = .03.

Patient-participation of the patient

Patient-participation of the patient was captured by patients’ question asking and answering of questions by the doctor (RQ1c). Of all utterances of the patient (N = 4,645), 170 (3.7%) were questions. Of these questions, 54 (31.7%) were open and 116 (68.3%) were closed questions. An example of an open question is “What if I think of questions after this?.” A closed question can be illustrated by the following quote: “Is that a one-day admission?.” It should be noted that, overall, few questions were asked by patients, as 96.3% of the utterances (n = 4,475) did not contain questions. Regarding answers, patients provided 180 (3.9%) one-word answers (i.e., yes or no). Table 10 presents an overview of questions asked and one-word answers by patients across SES level.

Table 10. Question asking and one-word answers by the patient (patient-participation).

	Low-SES (n = 16)	Middle-SES (n = 16)	High-SES (n = 13)	Total (N = 45)
Questions	37 (21.8)	63 (37.0)	70 (41.2)	170 (100)
Closed questions	26 (22.4)	48 (41.4)	42 (36.2)	116 (100)
Open questions	11 (20.4)	15 (27.8)	28 (51.8)	54 (100)
One-word answers	82 (45.6)	57 (31.7)	41 (22.7)	180 (100)

Note: Descriptives represent absolute frequencies with percentages between parentheses.

No significant differences were found across the three SES groups for the proportion of questions asked by the patient, F(2, 42) = 0.34, p = .711, η_p² = .02. However, a significant difference was found across SES groups regarding how often a patient answered a question of the doctor with one word, F(2, 42) = 3.81, p = .030, η_p² = .15. Nonetheless, the Bonferroni post-hoc test gave no convincing support for potential SES-level differences, as only marginally significant differences between SES groups were found. Thus, the following results should be interpreted with caution. Low-SES patients (M = 0.07; SD = 0.06) answered questions of their doctor more often with one word than middle-SES patients (M = 0.03; SD = 0.03), p = .059 and high-SES patients (M = 0.03; SD = 0.03), p = .078. Middle-SES and high-SES patients did not differ in this regard, p = 1.000. Although these findings only approached significance, it suggests that low-SES patients are more likely to answer questions of the doctor with one word compared to higher-SES patients.

Discussion

The main aim of our study was to examine how communication between doctors and low-SES patients organically unfolds, representing a crucial step in understanding and ultimately improving communication between doctors and low-SES patients. Our first aim was to understand how doctors and low-SES patients interact regarding instrumental, affective, and patient-centered aspects of communication. In line with earlier research, our results showed that most doctor-patient communication is instrumental (Desjarlais-deKlerk & Wallace, 2013). Most instrumental utterances were about the diagnosis of the patient. Doctors provided little lifestyle and psychosocial counseling. However, for low-SES patients specifically, who already tend to have less healthy lifestyle conditions and habits (Chan et al., 2018; Stormacq et al., 2019), attending to these subjects could be of great importance.

Affect is another important factor for adequate communication (Desjarlais-deKlerk & Wallace, 2013). Affective communication was present for about half of the time in the observed consultations. However, when we solely focus on explicitly expressing concerns or positive emotions, only one sixth of the consultation remained affective. Despite previous studies showing similar results, it is unfortunate that there is little affective communication in consultations between doctors and low-SES patients. Research has shown that patients with a lower education level highly value the affective part of the consultation (Jessup et al., 2017; Ong et al., 1995). For disadvantaged patients, having an affective medical environment, which potentially leads to a more trusting relationship, can counteract unfavorable communication effects, such as stress and impaired recall, which is beneficial for patients’ health (Walker et al., 2014).

We also demonstrated that doctors’ expressions of SDM were very rare within our observations. We assume that this result is due to the nature of the consultations, which were first-time appointments. Contrary to patients who are further along in their patient journey or those diagnosed with different diseases needing preference-sensitive decisions (e.g., cancer; Stiggelbout et al., 2015), we believe there might be less need for an extensive SDM process in the consultations we observed.

Our second aim was to examine how low- and high-SES patient consultations differed from each other. There are two general conclusions we can draw from our findings. First, compared to high-SES patients, low-SES patients are more passive communicators. Low-SES patients answered questions of the doctor more often with one word, voiced less concerns, and expressed fewer utterances during the consultation in general. These findings indicate that low-SES patients were less vocal, showed less patient-participation, and thus behaved more passively in consultations compared to high-SES patients (Halabi et al., 2020; Street & Haidet, 2011; Willems et al., 2005).

Since low-SES patients expressed fewer utterances in consultations, this naturally means that doctors talk a larger share of consultations with low-SES patients compared to consultations with high-SES patients. Two lines of thought can be expressed here. First, the passive communication style of low-SES patients gives doctors more room to speak. Second, doctors take more room to speak as they, logically, feel more confident and more like the expert in communicating health issues (Siminoff et al., 2006). Specifically, in medical encounters with low-SES patients, the (already) unequal doctor-patient relationship could be exacerbated by inequalities due to SES. As a result, besides taking up less space, patients with low SES are given less room to speak. Future research could assess which reasoning holds by, for example, conducting interviews zooming in on specific motivations of patients and doctors for their communication efforts.

Another interesting finding, contrary to previous research, is that our results pointed to low-SES patients receiving the greatest amount of instrumental communication, whereas it is generally assumed that high-SES patients receive more information from doctors (Willems et al., 2005). Although not statistically significant, this finding allows us to speculate the following: doctors might give a substantially large amount of biomedical information to low-SES patients simply because there is substantially less affective communication in the consultation. Our results revealed that low-SES patients voice less of their concerns, hereby possibly unintentionally encouraging a medical environment in which there are fewer opportunities to talk about affective themes (Finset, 2012; Willems et al., 2005).

This reasoning promptly supports our second major conclusion that patients’ communication style seems to account for more of the communication differences than doctors’ communication. To further illustrate, no differences were found for doctors’ instrumental, affective, and patient-centered communication. Consider affective communication, previous research has shown that doctors generally express less affective behavior toward low-SES patients, also because low-SES patients are more passive communicators (Willems et al., 2005). Our study, however, was unable to confirm such findings, even though we found that low-SES patients were less expressive.

As we only discovered differences in patients’ communication, our study might imply that doctors are already adequate communicators. On the one hand, these results might suggest that doctors have already learned to adapt their communication to the different needs and abilities of the patients. Given there has been growing attention for patients’ individual needs and health literacy levels, such an explanation is plausible. On the other hand, it can be hypothesized that other contextual factors, such as doctors’ attitude or tone of voice still affect patients’ communication (Street et al., 2007). Although we only found SES differences in patients’ communication, this does not necessarily mean that other factors, such as doctors’ non-verbal communication or patients’ sensitivity to the unequal doctor-patient relationship, do not play a role in patients’ communication (Siminoff et al., 2006). Future research should further investigate factors that contribute to understanding SES differences in patients’ communication to be more conclusive.

Implications for theory and practice

This study has several important implications. First, this study was able to highlight the importance of simultaneously considering the doctor’s and patient’s perspectives in doctor-patient communication research. Although there has been much scientific attention for patient outcomes of communication (e.g., satisfaction, information recall; Chandra & Mohammadnezhad, 2021; Haskard et al., 2009; Ong et al., 1995), most studies that examined communication styles tend to focus on doctors’ communication (Allen, 2020). Less attention has been paid to patients’ communication in the interaction. This study was able to highlight the importance of patients’ impact on communication. By equally involving both parties, nuanced conclusions regarding SES communication differences could be made.

While our results suggest that patients’ communication plays a significant role in consultation differences, it is important to recognize that it is a shared responsibility of patients and doctors to reach adequate communication (Hall, 2003). Patients should sufficiently explain their symptoms, concerns, and preferences (Ong et al., 1995), but if the patient has a hard time doing so, doctors should encourage patients to actively share their health issues and concerns. As doctors are the experts, they are inclined to control the medical encounter and can therefore influence the consultation to a large extent (Siminoff et al., 2006). Patients can be empowered by doctors using clear communication techniques (e.g., plain speech, no jargon), asking for confirmation of understanding (e.g., teach-back), increasing their level of patient-centeredness, using multiple modalities of explaining, offering easy-to-read materials and forms, and by teaching patients to ask questions about their health problem (Sudore & Schillinger, 2009). Such efforts enable patients to take a more active role in their health care, by making the patient feel more at ease and confident (Sudore & Schillinger, 2009). Ultimately, doctors could become even more aware of SES differences, the substantial encouragement that low-SES patients might need to voice their concerns, and how they can support patients in doing so.

Finally, this study provides important methodological implications for measuring SES. As previous scholars have stated, SES is often insufficiently approached and measured (Allen et al., 2019; Gagné & Ghenadenik, 2018). By constructing a comprehensive SES measurement (i.e., by combining education, income, occupation, and neighborhood), we were able to capture patients’ SES beyond education level or other related concepts (Smith, 2021). So, it is believed that the communication differences found in our study are indeed related to SES and not, for instance, merely education level. Although generalizability issues within the fields of health communication and SES remain, this study takes a step in the right direction: by being aware of the importance to construct comprehensive SES measurements and by providing in-depth information on how SES was operationalized. Furthermore, this study shows the importance of splitting SES into three categories, as we demonstrated communication differences between low- and high SES patients, but not between low- and middle-SES or middle- and high-SES patients. Thus, it is useful to approach middle-SES as a separate level instead of dividing SES into high and low, as this may result in drawing incorrect conclusions about low- and high-SES (patient) populations.

Limitations

While this study makes valuable contributions to the literature on doctor-patient communication and SES, our findings should also be viewed in light of its limitations. We should first acknowledge that causal inferences cannot be made based on this study (Hall, 2003). Because of its exploratory nature, this observational study cannot draw conclusions about specific causes and effects of communication. For instance, we cannot make statements about whether the communication of the patient influences the doctor or vice versa.

We should also recognize that the study results do not provide any information about patient outcomes, such as satisfaction or information recall. Even though communication differences were found, it is unclear how patients experienced these differences and how they may have affected their understanding of information. Although beyond the scope of this study, it is important to continue including patient outcomes in future studies to determine the topics of interest regarding communication differences. Furthermore, by relying solely on verbal coding measurements, we did not assess nonverbal cues. Since much affective behavior is often non-verbal instead of verbal (Ong et al., 1995), it is possible this study may have underestimated the amount of affective communication in our doctor-patient interactions. In addition, knowing about patients’ experiences would have been highly useful, especially when analyzing affect. For future studies, we recommend having the patient rate the communication via a survey or interview afterward, as this will provide a more comprehensive overall score of the communication.

It is also important to note that this observational study only included cardiovascular patients and cardiologists from one hospital. The organization of this study’s hospital, its policies, and how health care is embedded in its country’s society all impact health care experiences of patients and the doctor-patient communication (Kalwij, 2019). Furthermore, patients diagnosed with different diseases than cardiovascular disease could have had different needs and expectations, which might have been reflected in patients’ communication (Ong et al., 1995). Future research could examine whether same results are also found for other disease contexts, in other hospitals, and in different countries.

Conclusion

Communication differences based on patients’ socioeconomic status differences exist. Doctors seem to talk more in consultations with low-SES patients. Low-SES patients seem to be more passive communicators. Although no differences were found for doctors’ instrumental, affective, and patient-centered communication, we believe that communication between doctors and low-SES patients is still suboptimal. Being an active communicator is needed to effectively take part in the interaction which is essential if we want to improve quality of care. As low-SES patients may be more passive in their communication, doctors could try to encourage them to be more active to close the socioeconomic communication gap.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

We were financially supported by the collaborative initiative We Care, supported by Tilburg University and Elisabeth-Tweesteden hospital.