Table 1

An overview of the advantages and disadvantages of available methodologies for the study of doctor–patient communication

MethodLogistical considerationsScientific meritsScientific limitations
Questionnaires
  • Relatively inexpensive and quick to administer.

  • Can suffer from low-response rates.

  • Flexible: can be designed to assess many different aspects of communication, and can be administered to doctors or patients.

  • Analysis can be quantitative or qualitative, depending on the questionnaire design.

  • Questionnaires alone only provide information on reported behaviours: they do not allow for direct observation or measurement of communication behaviours.

  • Although they can have open questions, questionnaires tend not to produce as rich or detailed data as other qualitative methods, such as interviews.

  • There is wide variation in how questionnaires are developed and validated: a range of different questionnaires have been used to assess various aspects of communication, which are of variable quality.

  • Questionnaires can be susceptible to various biases (eg, social desirability bias).

Interviews
  • More time-intensive than questionnaires and require trained interviewers.

  • Generally less expensive than observational, simulated patient or vignette studies.

  • As with questionnaires, interviews are flexible can be used to assess a no of different measures.

  • They allow more detailed exploration of themes compared with questionnaires.

  • Interviews alone can again only provide information on reported behaviours or experiences.

  • There can be inconsistency between interviews with different participants; different interviewers may systematically elicit different responses from participants.

  • Difficulties arise in avoiding potential interviewer and coder biases, and in establishing reliability and validity.

  • In focus groups, group dynamics can influence data collection.

Vignette Studies
  • Can be quicker and less costly than observational studies.

  • Can still incur significant costs, in particular when actors are used in video vignettes.

  • Vignette studies permit greater control and standardisation of communication variables compared with real consultations: they allow experimental manipulation of specific aspects of communication.

  • They provide an ethically acceptable method for studying potentially harmful communication behaviours.

  • The use of analogue patients can overcome ceiling effects.

  • Attempts to isolate specific aspects of communication can result in an oversimplification of complex real-life doctor–patient communications.

  • Internal validity and external validity must be established.

  • Vignette studies cannot capture the influence of the long-term doctor–patient relationship on communication.

  • They cannot be used to study the impact of communication on actual health outcomes.

Simulated Patient Studies
  • Quicker and less expensive to run than observational studies.

  • Require trained actors, and often need significant input from both experts and lay people to develop realistic simulated patient responses.

  • Simulated patient studies allow for the direct study of real healthcare professionals communicating in certain controlled situations.

  • There is greater experimental control over patient variables compared with observational studies of real patients.

  • They are relatively flexible, and scripts can be adapted to be specific to the specialty of interest.

  • Unannounced simulated patients can be used to increase the external validity of the study by making the simulated consultation as realistic as possible.

  • There can be assessment of both verbal and non-verbal behaviours.

  • As with vignette studies, there can be concerns over both internal and external validity.

  • These studies can only give information on how doctors communicate in response to the simulated patients—they cannot be used to study the impact of communication on actual patients or on health outcomes.

  • As with vignette studies, the influence of the long-term doctor–patient relationship on how communication occurs and is perceived cannot be studied.

Direct observation of real consultations
  • Can be time-consuming to run.

  • Involve real patient–doctor interactions, so require additional levels of ethical approval compared with simulated patient or vignette studies.

  • These studies involve real patients and doctors, so reflect actual clinical practice to a greater extent than simulated patient or vignettes studies.

  • They can be tailored to examine communication in particular settings of interest (eg, communication in primary care, outpatient clinic).

  • They permit analysis of consultations, and examination of associations between communication variables and both objective and subjective outcome measures.

  • The use of real patients and doctors can allow for examination of communication in the context of real doctor–patient relationships: the interaction between the length of the relationship and communication can be studied, in contrast to experimental techniques such as vignette or standardised patient studies.

  • There may be systematic differences between patients/doctors who consent to partaking in these studies and those who do not.

  • The Hawthorne effect must be considered.

  • Often, these studies do not allow for controlled manipulation of relevant variables, so are observational rather than interventional. They can provide information about correlations between physician and patient communication behaviours and different outcomes measures, but do not necessarily provide any information about causation.

  • There is a need to carefully control for confounding factors (such as patient’s baseline health).

  • As with simulated patient studies, there is wide variation in how communication behaviours in the observed consultation are measured and analysed.