The Validity of Using Analogue Patients in Practitioner–Patient Communication Research: Systematic Review and Meta-Analysis

When studying the patient perspective on communication, some studies rely on analogue patients (patients and healthy subjects) who rate videotaped medical consultations while putting themselves in the shoes of the video-patient. To describe the rationales, methodology, and outcomes of studies using video-vignette designs in which videotaped medical consultations are watched and judged by analogue patients. Pubmed, Embase, Psychinfo and CINAHL databases were systematically searched up to February 2012. Data was extracted on: study characteristics and quality, design, rationales, internal and external validity, limitations and analogue patients’ perceptions of studied communication. A meta-analysis was conducted on the distribution of analogue patients’ evaluations of communication. Thirty-four studies were included, comprising both scripted and clinical studies, of average-to-superior quality. Studies provided unspecific, ethical as well as methodological rationales for conducting video-vignette studies with analogue patients. Scripted studies provided the most specific methodological rationales and tried the most to increase and test internal validity (e.g. by performing manipulation checks) and external validity (e.g. by determining identification with video-patient). Analogue patients’ perceptions of communication largely overlap with clinical patients’ perceptions. The meta-analysis revealed that analogue patients’ evaluations of practitioners’ communication are not subject to ceiling effects. Analogue patients’ evaluations of communication equaled clinical patients’ perceptions, while overcoming ceiling effects. This implies that analogue patients can be included as proxies for clinical patients in studies on communication, taken some described precautions into account. Insights from this review may ease decisions about including analogue patients in video-vignette studies, improve the quality of these studies and increase knowledge on communication from the patient perspective. Electronic supplementary material The online version of this article (doi:10.1007/s11606-012-2111-8) contains supplementary material, which is available to authorized users.

as such. With regard to external validity, the question arises whether results are generalizable to CPs and clinical care, i.e. are APs able to adopt a video-patient's perspective?
To summarize, we lack an understanding of the rationales for conducting video-vignette studies with APs; how both internal and external validity are increased and tested; how APs' perceptions of communication correspond to CPs' perceptions; and whether APs' evaluations of communication overcome ceiling effects. An overview of these elements will provide more insight into when and how APs can be used in future studies. Therefore, a systematic review is conducted with the following research questions: 1. What are the rationales for conducting clinical and scripted video-vignette studies on medical communication with APs ? 2. What have video-vignette studies done to increase and test their internal and external validity? 3. How do APs perceive-affective, instrumental and general-communication elements? 4. Do APs' evaluations of communication overcome ceiling effects?

Identification of Studies
Pubmed, Embase, Psycinfo and CINAHL were searched in February 2012. Searches were not restricted to any parameter and focused on two central concepts: 'analogue patients' and 'video' (see the Online Appendix Supplementary data for search strategies used). Studies were eligible for inclusion if they were about (verbal/nonverbal) communication between physicians/ nurses and patients and: i) used video-vignette designs; ii) included APs (>18 years): healthy subjects, untrained or trained only for this study; patients not judging their own doctor/nurse; standardized patients viewing a videotaped consultation they took part in; and iii) used APs' perceptions of physician's/nurse's communication as outcome measures (e.g., preferences, recall). Studies were excluded if: i) observers were trainers, research assistants, trained/ experienced coders, examiners, medical students or faculty members; ii) APs' comments did not include a quality judgment.

Data
The following data were extracted from each study and summarized in Table 1: study characteristics and quality, design, rationales for conducting video-vignette studies with APs, attempts to increase and test internal and external validity, limitations, and APs' perceptions of the studied communication elements.
Quality of studies was assessed 16 by applying the Research Appraisal Checklist (RAC). 17 The RAC consists of 51 items covering the quality of title, abstract, introduction, methodology, data analysis, discussion, and style/form. Each item is scored on a 1-6 scale, so total scores can vary between 0 and 306 points with three quality categories: i) Below Average (0-103 points), ii) Average (103-204 points), iii) Superior (205-306 points).

Meta-Analysis to Determine Ceiling Effects
To determine whether APs' evaluations of communication (e.g. satisfaction, preferences) overcome ceiling effects, a random-effects multivariate meta-regression analysis 18 was performed using the statistical package MLWIN 2.02. 19 The following quantitative data was abstracted for each evaluation: M, SD, range. For each study the number of participants, videos viewed per participant and available videos was abstracted. For each evaluation, using various scales, the mean score was transformed to a 0-100 score 20 using two formulas; for scales starting at 1: ((mean-1)/ (range-1))x100, for scales starting at 0: ((mean/range))x100. Authors were contacted to provide relevant data not presented in the articles.

RESULTS
The 2950 references initially found were reviewed on title/ abstract (and if necessary on full-text) to determine whether they: a) were about communication, b) used a videovignette design, c) included APs. A random 10 % of the articles were independently checked on these criteria by two authors (LV and JB); interrater agreement exceeded 95 %. Thirty-four articles met these criteria and a forward-and backward reference search was performed. Four hundred and fifty-two new articles were reviewed in the aforementioned manner, resulting in 32 additional articles. These 66 articles were explored full-text on the final criteria: a) a focus on doctor/nurse-patient communication, b) inclusion of APs who viewed videos and judged the communication. Thirty-four articles met all criteria. Their references were In a pilot study APs thought the videos were credible. APs were given information on the medical condition, and were asked to think back of the last time they had this problem.
Courtesy led to higher satisfaction, but competence both to higher satisfaction and intended compliance.
Whether the results of APs' viewing a videotape are the same as CPs' reactions is unclear.

DISCUSSION
This systematic review focused on the rationales, methodology and outcomes of medical video-vignette studies with APs. Scripted studies provided more specific rationales for using video-vignette designs with APs than clinical studies and directed more efforts at increasing/testing internal and external validity. APs' perceptions of communication overlapped generally with CPs' perceptions. Meanwhile, their evaluations overcame ceiling effects. These results have interesting methodological, theoretical and practical relevance.
Scripted studies paid the most attention to increasing the designs' methodological soundness. Specific methodological rationales for conducting video-vignette studies with APs were provided, such as the opportunity to study communication systematically. This fills a gap in clinical care studies, in which only correlations, but no causality between communication and outcomes can be determined. 58,59 Unfortunately, some scripted studies included container-concepts of communication (e.g., patient-centeredness). When positive effects are found, it remains unclear which specific element(s) of communication influenced outcomes. 15,58 Additionally, as argued, when videos are watched by multiple APs, the reliability of assessments increases. 60,61 Another argument for including APs was that their evaluations can overcome ceiling effects. APs' evaluations were indeed not high; averagely 54.28 on a 0-100 scale. By comparison, a meta-analysis of CPs' satisfaction ratings showed an average score of 80.00 (0-100 scale). 20 Moreover, a recent study compared CPs' satisfaction scores with those of APs viewing these videotaped consultations. Mean score (1-6 scale) for CPs was 5.8, while for APs it was 4.0 (p<0.001). 62 APs' ratings thus seem to overcome this limitation of CPs' evaluations. 4,5 Accordingly, these and other methodological rationales provide strong foundations for conducting video-vignette studies with APs.
To achieve internal validity, APs reflected on manipulations in scripted consultations. Unexpectedly, 'experts' (doctors/researchers) were not often asked to comment on manipulations, although they may have insight into the manipulations' (theoretical) success. Moreover, little information was provided on how exactly scripts were created, i.e. it often remained unclear what input researchers used to develop scripts and at what point(s) the scripts were validated.
Focusing on external validity, some studies argued that APs' perceptions overlap with CPs' perceptions. However, none of these studies determined whether APs watching videotaped consultations and CPs in these consultations overlapped on outcome measures. As stated earlier, such a study has recently been performed. 62 In this study-taking into account CPs' skewed satisfaction scores-APs' and CPs' evaluations were correlated. Additionally, a meta-analysis in psychology 63 showed that lay people can make reliable judgments for (non)verbal communication based on brief (clinical and scripted) videotaped interactions.
Theoretical evidence supporting the external validity of APs can be found in simulation theory and is supported by neuro-cognitive studies on empathy. According to simulation theory, we infer other persons' mental states by matching their states with resonant states of one's own mental state. 64 Neuro-cognitive studies show that the brain's mirror neurons fire when a particular action is carried out or observed. 65 They form the basis for empathy, [66][67][68][69] as they are involved in experiencing and observing emotions in others 70 and allow people to adopt another person's perspective. 71 Indeed, some oncological scripted studies included survivors alongside healthy participants. Their perceptions overlapped, indicating that healthy people can put themselves in the shoes of (cancer) patients. 72 However, the methodological and theoretical rationales and advantages of using APs as proxies for CPs are relevant only when APs' perceptions of communication are applicable in clinical practice, which is mainly supported by our results. APs' perceptions of communication overlap mostly with those of CPs. A few-seemingly-contradictory findings were found. APs disliked information-exchange during bad news conversations, while CPs mostly valued this behavior. However, CPs often report receiving too much information during these conversations. [73][74][75][76][77][78] Besides, while most studies point to the positive effects of patientcenteredness, a study with APs 51 and review on CPs 12 found that for purely physical complaints, a patient-centered style may be suboptimal.
Despite these promising results, various aspects should be taken into account when interpreting APs' perceptions for clinical practice. First, in one study APs' perceptions were unrelated to CPs' satisfaction scores. The considerable age difference (students versus seniors) may be responsible for this finding, as age influences communication preferences. [79][80][81] Future studies should take background characteristics influencing preferences-e.g. gender, 81,82 education 83,84 -into account. Consequently, students should not be included as APs merely for convenience. Second, the diversity in APs' evaluations should be kept in mind. The long-term doctor-patient relationship possibly influencing CPs' evaluations cannot be captured by studies using APs. Thus, as video-vignette studies make it possible to disentangle the effect of various communication elements, these elements should afterwards be tested in clinical care.

Limitations
This review has its limitations. First, the literature is inconsistent in the terms used for "analogue patients". To include all relevant articles, both forward and backward reference searches on possible relevant articles were performed and included studies' references were handsearched. Future studies should use the term "analogue patients" consistently. Second, we excluded trained observers, but included lay people trained for this specific study. As studies may have used inconsistent labels, we screened for detailed information on observers. Despite these precautions taken, inadequately indexed and little cited relevant studies may have been missed, as we used a topdown search strategy.

CONCLUSION AND FUTURE STUDIES
Scripted video-vignette studies increased their methodological soundness by providing specific rationales for conducting video-vignette studies with APs and increasing (internal and external) validity. In keeping with simulation theory and neuro-cognitive studies, APs' perceptions of communication overlapped largely with CPs' perceptions-while overcoming ceiling effects. However, it may be necessary to match participants on variables such as age and gender. Moreover, the effect of a long-term doctor-patient relationship on evaluations cannot be studied with APs. This leads to the conclusion-taking these precautions into accountthat APs can provide knowledge on the patient perspective on communication.
Future-scripted-studies may benefit from the described elements to increase their methodological strength and provide more information about the process of ensuring validity. From this review we cannot conclude which communication elements-and outcome measures-can best be studied with APs. Ambady and Rosenthal 63 suggested that communication with an affective component is fastest recognized because its evolutionary importance. 85,86 Future studies could investigate differences between various types of APs. Research could build further on aforementioned work, 62 comparing CPs' perceptions with those from APs watching these consultations, taking into account differences in rating dispersion and focusing on background characteristics. This will raise the level of future studies in this promising research field, aimed at systematically unraveling the patient perspective on communication.