The aim of the current review is to give the state of the art on the social gradient in doctor-patient communication, to describe which aspects of the consultation are affected by this social gradient, and whether an evolution over time can be noticed comparing the results of older studies with those of newer studies. In this review we found that patients from lower social classes (measured by income, education or occupation) receive less socio-emotional talk, a more directive and a less participatory consulting style characterised by for example less involvement in treatment decisions; a higher percentage of biomedical talk and physicians' question asking; lower patient control over communication; less diagnostic and treatment information and more physical examination. Doctors give more information, more explanations, more (emotional) support and adapt more often a shared decision making style with higher SES participants.
This review also indicates that the literature on the social gradient in doctor-patient communication that was published after 2002, at least addresses new issues and themes. Firstly, in the period 1965-2002, 42 articles were selected for this review, while for the period 2002-2011, 87 articles were selected. These numbers indicate that doctor-patient communication becomes a more emerging topic in the research on delivering qualitative care. Secondly, most of the more recent studies emphasize the importance of the reciprocity of communication: the doctor might communicate differently according to the social status of the patient, and patients may adapt a different communication style according their social class. Patients with a high SES tend to ask more questions, ask for explanations, are more expressive and have a higher level of being opinionated than their lower SES counterparts [33
Furthermore, there seems to be a growing interest in patient's perception of doctor-patient communication. While in the past, patient's perception was not taking into account or no differences in perception were found, more recent studies show that low SES patients have the feeling doctors fail to explain things in a way they can understand and spend less time with them [21
These findings emphasise that doctor-patient communication is a complex interactional system. To depict this complexity, Street et al. (2007) applied an ecological model that takes into account the interplay of multiple physician, patient and contextual factors that collectively influence doctor-patient interactions [56
]. The influence of any variable (e.g. ethnicity) may vary depending on the presence of other factors (e.g., the patients' level of education, income, doctors' communication style) [53
]. The ecological approach recognizes that within the context of any medical encounter, a number of processes affect the way physicians and patients communicate and perceive one another. There are four important sources of potential influence: the physician's communication style, patients' characteristics, physician-patient demographic concordance and the patients' communication. First, how a physician communicates with a patient may depend on his or her style. Some physicians provide more information, ask more questions, are more supportive and use more partnership-building than other physicians [33
]. Second, variability in physicians' communication and perceptions may be related to the patients' demographic characteristics (education, income, occupation) [57
]. Finally, the patients' communication style can have a strong effect on physician behaviour and beliefs [18
Important in this model is that patient interaction not only depends on the physician's behaviour but also on patients' characteristics and preferences. Patients from lower social classes more often suffer from (multiple) chronic conditions and more severe acute conditions [58
]. But also they often have lower levels of health literacy-the degree to which persons have the capacity to obtain, process and understand basic health information and services needed to make appropriate health decisions [59
]. Furthermore lower social class is associated with a lower sense of personal control also known as external locus of control. This means that the person perceives that certain events such as health and sickness are beyond his/her control [61
]. This might explain why low SES people show lower levels of participation. Also, because they are less used to or feel less capable to interact during consultation, they might prefer a more directive consultation style. Recently, an international consortium of research teams in the UK, the Netherlands, Italy and Belgium set up the Gulliver study which focuses on the patient's preferences in doctor-patient communication. Analysing a possible social gradient in these preferences is hereby one of the points of attention of the researchers.
Limitations of the study
Many of the limitations the review of 2005 encountered are still applicable today. Affective and non-verbal behaviour are important aspects in physician-patient communication e.g. through their influence on patient satisfaction [44
]. Still a limited number of studies described the interaction between social class and non-verbal physician-patient communication. All studies indicates the difficulties measuring and coding non-verbal behaviour. Therefore, these limited number of studies entails important methodological difficulties and does not allow us to draw conclusions concerning non-verbal and affective behaviour. Further research on this topic is still needed.
Secondly, it is very difficult to compare the results of the studies due to the great diversity of measurements and frameworks organising these measurements in the different studies. Socio-economic status of the patient was measured by means of educational level, income or occupation [3
]. An alternative to determine SES is to use "proxy" measures e.g. the insurance status, house tenure, car ownership, socio-demographic measures (race, etc.). Articles using proxy-variables as the only measure for SES were excluded. However, some of the selected articles used these variables in combination with educational level, income or occupational class. Next to the SES of the patient, also communication variables can be classified in many different ways. The variables used in these classifications are not always comparable, making if very difficult to compare the studies using different classification systems. We chose to categorise most of the communication variables according to the axis verbal/non-verbal behaviour. The determinants of communication that did not fit into the categories of this axis were related to patient centeredness.
In order to improve the comparability of future research, the use of a uniform definition and classification of communication variables is indispensable.