Medical student interviewing: a randomized trial of patient-centredness and clinical competence

Transcription

1 Family Practice Vol. 20, No. 2 Oxford University Press 2003, all rights reserved. Doi: /fampra/ , available online at Printed in Great Britain Medical student interviewing: a randomized trial of patient-centredness and clinical competence Maree O Keefe a, Don Roberton a, Michael Sawyer a,b and Peter Baghurst c O Keefe M, Roberton D, Sawyer M and Baghurst P. Medical student interviewing: a randomized trial of patient-centredness and clinical competence. Family Practice 2003; 20: Background. It is widely accepted that the quality of doctor interviewing skills is an important determinant of health care outcomes. Two interviewing skills contributing significantly to optimal health outcomes are the clinical competence of the interviewer and the use of patient-centred techniques. However, the relative importance of these to health outcomes is unknown. Objective. The purpose of this study was to examine the relative effect on maternal recall and satisfaction of medical student clinical competence and use of patient-centred interview techniques. Methods. Sixty-nine mothers of children attending the Paediatric Medical Out-patient Clinic, Women s and Children s Hospital, South Australia agreed to participate in the study, with 60 successfully completing the study. They viewed two of four standardized medical student interview videotapes in which the level of clinical competence and patient-centredness were varied independently. All other interview variables were controlled. Each mother rated the interviews by questionnaire (balanced incomplete block design, each interview rated by 30 mothers). Maternal satisfaction with the student interview was measured using the Medical Interview Satisfaction Scale (MISS) and the Interpersonal Skills Rating Scale (IPS). Maternal recall of interview information was assessed by questionnaire, with maternal answers coded independently before analysis. Results. Significant, independent effects of clinical competence and patient-centredness were observed with both MISS (F = 42.1, P ) and IPS (F = 49.3, P ) scores. The effect of clinical competence was stronger than that of patient-centredness. Maternal recall for specific information was greater for some items following the more clinically competent interviews, but was lower for others. There was no association between maternal recall and the level of patientcentredness demonstrated by the student. Conclusion. Clinical competence was a more significant determinant of maternal evaluations of medical student interviews and maternal recall of information than was patient-centredness. High levels of both patient-centredness and clinical competence were associated with the highest maternal satisfaction. Keywords. Clinical competence, doctor patient relationship, medical, patient-centred interviewing, patient satisfaction, students. Introduction The quality of doctor interviewing skills in medical consultations is important as it influences patient satisfaction, compliance and health outcomes. 1,2 A previous Received 3 January 2002; Revised 12 June 2002; Accepted 4 November a Adelaide University Department of Paediatrics, b Department of Psychiatry and Evaluation Unit and c Public Health Research Unit, Women s and Children s Hospital, North Adelaide, South Australia Correspondence to Dr Maree O Keefe; maree.okeefe@adelaide.edu.au study showed that maternal satisfaction and recall of information following a medical student health consultation varied according to the skills of the student. 3 However, the relative effects of patient-centredness and clinical competence were not assessed. Based on our earlier study, 3 we expected that maternal satisfaction and recall of information would be highest after an interview in which a medical student demonstrated high levels of both patient-centredness and clinical competence. Similarly, it was anticipated that maternal satisfaction and recall of information would be lowest after an interview in which the student demonstrated low levels of both patient-centredness and 213

2 214 clinical competence. We speculated that levels of maternal satisfaction and recall would fall between these extremes in interviews during which the student displayed either high clinical competence and low patientcentredness or, conversely, low clinical competence and high patient-centredness. In this study, comparison of maternal satisfaction and recall following consultations in which student patientcentredness and student clinical competence were varied independently allowed the relative effect of each of these components of medical student interviewing to be assessed. Methods Participants Mothers of children attending the Paediatric Medical Out-patient clinic at the Women s and Children s Hospital, South Australia, between February and December 2000 were invited to participate in the study. Mothers were contacted by letter, and a subsequent telephone call, prior to their scheduled appointment. Mothers were ineligible if the mother (or maternal carer) was not attending with the child, if the mother was not proficient in spoken and written English, if telephone contact was not successful, if the mother had not received the initial letter or if the child s appointment was cancelled subsequently. Mothers who agreed to participate were allocated randomly to view two of the four study videotapes using a balanced incomplete block design. Power calculations based on our previous study 3 indicated that a minimum of four mothers should evaluate each pair of videotapes (80% power to demonstrate a difference at P 0.05). There were 12 possible combinations of videotape pairings (including order of viewing). Recruitment of mothers continued until 60 mothers were enrolled, providing five mothers in each group of videotape pairs. Development of interview videotapes Two transcripts developed from medical student interviews with parents were used to create the clinical scenarios videotaped for this study. The development of these transcripts, together with the independent assessment of clinical competence and patient-centredness, have been described previously. 3 In one transcript, high clinical competence and high patient-centredness were demonstrated. In the other, low clinical competence and low patientcentredness were demonstrated. The two transcripts, whilst differing in the details of student dialogue, had almost identical maternal dialogue and clinical content. The information content regarding diagnosis and management provided by the student in each interview was also the same. The interview scenario was that of a young child who was acutely unwell with a viral illness. Patient-centredness was defined as the extent to which the student sought to elicit the mother s thoughts, Family Practice an international journal feelings, concerns and expectations. 4 Clinical competence was defined as mastery of a body of relevant knowledge and... a range of relevant skills which would include interpersonal, clinical and technical components. 5 In order to study the separate effects of clinical competence and patient-centredness on maternal recall and ratings of medical student interviews, four videotaped interviews were required in which clinical competence and patient-centredness varied separately. Using the Henbest measurement scale for patient-centredness, 4 four new transcripts were developed from the initial two transcripts ( transcript A and transcript B ) by exchanging medical student dialogue in response to 11 discrete maternal offers of information. No change was made to the information content of any of the interviews. The resulting four transcripts were developed as follows: (HCHP) High clinical competence, high patientcentredness transcript (unaltered transcript A) (HCLP) High clinical competence, low patientcentredness transcript The HCLP transcript was created from the HCHP transcript by altering student responses to the 11 discrete maternal offers of information. Student responses from transcript B (low patient-centred) to each of the maternal offers were substituted for those of transcript A (high patient-centred). Student dialogue from transcript A which included checking of the mother s understanding and invitation of questions was deleted from the closing segment. No other dialogue was altered, preserving the high clinical competence component. (LCLP) Low clinical competence, low patientcentredness transcript (unaltered transcript B) (LCHP) Low clinical competence, high patientcentred transcript The LCHP transcript was created from the LCLP transcript by altering student responses to the 11 discrete maternal offers of information. Student responses from transcript A (high patient-centred) to each of the maternal offers were substituted for those of transcript B (low patient-centred). Student dialogue from transcript A which included checking of the mother s understanding and invitation of questions was added to the closing segment. No other dialogue was altered, preserving the low clinical competence component. Minor alterations to the substituted medical student dialogue were required occasionally to maintain the continuity of the interview or to preserve the appropriate clinical competence level. In all cases, the patientcentredness level was not affected by these minor changes. The transcripts were colour coded to prevent viewers identifying or anticipating any differences prior to seeing the videotaped interviews. The final distribution is displayed in Table 1.

3 Medical student interviewing 215 Using the transcripts, four videotapes were produced in a film studio with the assistance of a professional producer. A male actor played the role of the medical student. The transcript of each interview was displayed off-screen and was followed exactly by the actor, including pauses, ums and ahs. When each interview reached the point where the student examined the child, the student s image was replaced with a still shot that read Examination on a plain background. There was no sound transmitted during this still shot which was displayed for 5 s. Following this, the student s image reappeared and the interview was concluded. The actor medical student spoke almost directly to the camera, with his head and shoulders filling the screen. The actor s attire, facial expressions and voice were identical for each videotape to ensure that the verbal content of the interviews was the only variable being assessed. A neutral beige background was used in each videotape, and there were no extraneous noises or other distractions. A technical assistant played the role of the mother, reading her lines off-camera so that only her voice was heard. The videotapes were produced in this way to encourage mothers to identify more closely with the mother in the interview. Videotape screenings The videotaped interviews were screened on a television monitor in the out-patient clinic waiting area. They were preceded by a pre-recorded orientation videotape which encouraged participating mothers to imagine themselves as the mother in the videotaped interview. Each interview ran for 7 min, with the questionnaire completed by mothers after each interview taking ~15 min. Independent assessment of clinical competence Following manipulation of the transcript contents to create the four videotapes (HCHP, HCLP, LCHP and LCLP), independent assessment of each videotaped medical student s clinical competence was undertaken. Eight paediatricians with extensive experience in medical student teaching and assessment individually and independently viewed all four videotapes. TABLE 1 Medical student transcripts High Clinical competence Low Patient-centredness High HCHP LCHP Low HCLP LCLP HCHP, transcript showing high clinical competence, high patientcentredness; HCLP, transcript showing high clinical competence, low patient-centredness; LCHP, transcript showing low clinical competence, high patient-centredness; LCLP, transcript showing low clinical competence, low patient-centredness. The order of viewing was assigned randomly, and viewing was undertaken without prior knowledge of either the videotape content or the study aims. At the conclusion of the last videotape, each paediatrician ranked the interviews according to the clinical competence displayed by the student (as defined above). A score of 1 was awarded to the interview in which the student displayed the highest level of clinical competence, and a score of 4 for the interview in which the student displayed the lowest level of clinical competence. Paediatricians were permitted to rank two or more interviews as demonstrating equivalent clinical competence. If two interviews were both ranked 1, then a final rating of 1.5 was recorded for each. If two interviews were both ranked 2, then a final rating of 2.5 was recorded for each, and a final rating of 3.5 each if two interviews were ranked 3. Independent assessment of patient-centredness An independent rater with extensive experience in coding transcripts for patient-centredness scored each transcript using verbal response modes (VRMs) and a global rating scale. The VRM scales were based on a modification for primary health care of Stiles verbal response mode system in which every verbal utterance is categorized into one of a small number of discrete groupings. 6 Combinations of these groupings can then be developed to describe constructs of interest, and VRM units summed within these groupings. Previous combinations of VRMs describing doctor responsiveness and patient involvement were used, with higher total scores indicating higher patient-centredness. 7 The second measure used was a global scale developed for assessment of general practice trainees, measuring doctor verbal behaviour. This scored the highest demonstrated level of behaviours that sought to understand the patient s perceptions and understanding of their illness, and that involved the patient in management planning. 7 Independent assessment of information content In addition to coding the transcripts for patientcentredness, the independent rater also identified and categorized every information statement made by the student in each interview. Maternal evaluation questionnaires Maternal satisfaction was measured using the Medical Interview Satisfaction Scale (MISS), a 29-item, 7-point scale; 8 and the Interpersonal Skills Rating Scale (IPS), a 13-item 7-point scale. 9 Minor adjustments were made to each scale by substituting the words medical student for doctor, and your child for you to ensure the questions were relevant to the situation portrayed in the videotapes. Mothers also completed a four-item scale rating the realism of the videotape portrayals. On all scales, maternal responses were summed, with higher scores indicating higher satisfaction or realism.

4 216 Family Practice an international journal Finally, mothers were asked to state a preference for one videotape of the two they viewed, together with their reasons for choosing this videotape. Maternal recall questionnaire To assess maternal recall of information provided by the student in the videotape interview, mothers were asked 10 specific questions directly related to either diagnostic or management information provided by the student in all four interviews. The correctness of maternal responses representing recall of student information was coded independently by two psychology graduates blind to the study hypotheses and the type of interview. For each question, maternal responses were coded for correctness after comparison with actual student dialogue from the relevant transcript. There was exact numerical agreement between the two independent coders for 87.7% of the 1320 responses. A consensus rating was determined by mutual agreement where the two coders disagreed. Statistical analyses The experimental design was factorial (two levels of student clinical competence crossed with two levels of patient-centredness). However, because any one mother viewed only two of the four possible videotapes, MISS and IPS scores were compared using an analysis of variance for balanced incomplete block designs, with mothers as the blocking factor. 10 A further term to account for any order effects (first interview versus second) was also included in the analysis. The maternal recall data were summarized in threeway tables, with classification factors consisting of interview (first or second), accuracy of recall (four levels) and clinical competence (high or low) summing over all levels of patient-centredness (or vice versa). Log linear modelling was undertaken to test for homogeneity of the distribution of correctness of recall across high and low levels of clinical competence, or patient-centredness, and across first and second interviews. The analysis provided maximum likelihood chi-square statistics for testing the independence of explanatory factors. For all questions, an adequate fit (as determined by the residual deviance) was obtained after fitting (at most) different distributions for high and low clinical competence, and a clinical competence by interview interaction term. Institutional ethics approval The study was approved by the Research Ethics Committee of the Women s and Children s Hospital, Adelaide. Results Independent assessment of clinical competence Mean paediatrician rankings (scale: 1 = highest; 4 = lowest) of the four videotaped interviews regarding the level of clinical competence displayed by the student were as follows: HCHP 1.3, HCLP 1.8, LCHP 3.4 and LCLP 3.6. The paediatricians all rated HCHP and HCLP as either 1, 2 or equal 1 in terms of clinical competence. The paediatricians all rated LCHP and LCLP as either 3, 4 or equal 3 in terms of clinical competence. Independent assessment of patient centredness Independent rating results are shown in Table 2. The HCHP and LCHP interviews were rated as more patient-centred than the HCLP and LCLP interviews on both the global rating and the doctor responsiveness component of the VRM. Because the maternal dialogue was controlled so carefully in the development of the transcripts, the VRM coding unit patient involvement was not a true reflection of the patient-centredness of the medical student interviews. It was almost identical for the HCHP and HCLP interviews, and for the LCHP and LCLP interviews, reflecting the stability of the maternal dialogue at the two levels of clinical competence. In this particular analysis, only the doctor receptiveness coding unit was appropriate for the measurement of patient-centredness. Independent assessment of information content All four videotaped consultations contained the same amount of specific diagnostic and management information from the student. Maternal evaluations of the videotaped interviews Enrolment. Letters were sent to 125 consecutive eligible mothers with subsequent telephone contact. Sixty-nine agreed to participate (55%), and 56 declined (45%). Nine mothers who agreed to participate did not attend on the day of the videotape screening, or arrived too late to complete the study requirements, leaving the 60 mothers initially sought. There were no significant differences between participating and non-participating TABLE 2 Independent assessment of patient-centredness Interviews HCHP HCLP LCHP LCLP VRM a (total no. of utterences) Doctor responsiveness Patient involvement Global scale b (range 0 20) a Higher total numbers of doctor responsiveness and patient involvement VRMs reflect a more patient-centred interview, but in this study the patient involvement VRMs reflect control of maternal dialogue rather than the level of patient-centredness. b A higher score reflects a more patient-centred interview.

5 Medical student interviewing 217 mothers regarding child gender, age or reason for clinic attendance. The main reason for non-participation cited by mothers was insufficient time available. Three was the maximum number of mothers at any one screening, and mothers were not allowed to confer about the videotape contents. Videotape realism. The mean (±SD) maternal rating regarding the realism of the videotape interview recreations was 18.1 ± 1.8 (maximum achievable score = 20, minimum = 4). There were no significant differences between realism ratings by mothers for each of the four videotapes in the study. Maternal ratings. Mean (±SD) MISS scores for the total of 120 maternal ratings were: HCHP 155 ± 17; HCLP 143 ± 24; LCHP 109 ± 33; LCLP 90 ± 24: (F = P ). Mean IPS scores for the total 120 maternal ratings were: HCHP 78 ± 7; HCLP 68 ± 14; LCHP 53 ± 20; LCLP 38 ± 12: (F = 49.3, P ). Mean MISS and IPS scores for the first and second interviews are presented in graphical form in Figures 1 and 2. The scores on both scales were all substantially higher for the high clinical competence interviews than for the low clinical competence interviews (P 0.001, within subjects). The effect of patient-centredness (shown by the vertical separation of the four lines in Figures 1 and 2) FIGURE 1 Mean MISS scores was highly significant (P 0.001), but not as strong as for clinical competence, which was the only effect found to be statistically significant in both the between- and the within-subject error strata. There was also substantial variation between the first and second interviews, with second interview ratings always being higher than first interview ratings (P 0.001). However, the difference between the ratings for the low clinical competence interviews and the high clinical competence interviews was approximately constant and varied little with either interview order or the level of patient-centredness. This was confirmed in the analysis of variance by the absence of a significant interaction term between interview order and clinical competence, or patient-centredness and clinical competence. The ranking of interviews according to maternal preferences for one interview of each pair viewed was the same as that observed with the satisfaction scores. The HCHP interview was preferred most often and LCLP least often. Reasons cited by mothers for determining their preferences were: better student interpersonal skills (n = 50); more information provided (n = 42); more caring (n = 34); more medical knowledge (n = 30); treated me as an equal (n =32); listened more (n = 50); more confident (n = 42); and included my opinions (n = 44). Maternal recall. There was no significant association between patient-centredness and maternal recall. Maternal recall varied significantly with the level of clinical competence for five of the 10 recall questions (Table 3). However, high clinical competence was not associated consistently with better maternal recall. The unexpected relationship between low clinical competence and better maternal recall observed for two questions was attributed to differences in the actual student dialogue. Although information content was the same in all the interviews, the more competent student gave better explanations that were longer and more complex, therefore harder for mothers to remember accurately. Better recall was observed after the second interview for only two questions, even though the same information was repeated in the second interview. FIGURE 2 Mean IPS scores Discussion This study sought to establish the relative importance to mothers of medical student clinical competence and patient-centredness in child health consultations. The choice of these two consultation elements was based on an extensive literature review seeking key elements of effective medical consultations. The independent assessment clearly supported the successful manipulations of clinical competence and patient-centredness and the designated levels of each of these qualities in each of the created videotapes. The use

6 218 Family Practice an international journal TABLE 3 Summary of the maternal recall questionnaire log-linear modelling Recall question Clinical competence a Interview b by clinical competence a What did the medical student Greater recall of low clinical Greater recall after the second say you should do to treat competence student s interview of low clinical your child s temperature, information (P = 0.004) competence student s coughing, and being off her food? information (P = 0.002) Did the medical student identify Greater recall of high clinical No association any problems with your child s competence student s throat? information (P 0.001) Did the medical student identify Greater recall of high clinical No association any problems with your child s competence student s glands? information (P = 0.027) How long did the medical student Greater partially correct recall Greater recall after the second say it would take your child to for the low clinical competence interview of high clinical recover? student s information competence student s (P = 0.006) information (P = 0.012) Did the medical student tell you Greater incorrect recall of low No association to bring your child back to be clinical competence student s seen again? information (P 0.001) a Statistical significance based on a maximum likelihood chi-squared statistic on 3 degrees of freedom. b Comparison of order of viewing of each interview. of real medical student consultations as the basis of the transcripts provided authenticity to the recreated videotaped interviews. Whilst the patient-centredenss model is conceptually straightforward, identification and measurement of key components have proved challenging in research settings. 11,12 There was some potential difficulty encountered in the preparation and independent analysis of interview transcripts due to the lack of a generally accepted measure of patient-centredness. This was addressed by using three separate measures of patient-centredness: one in the development of the interviews and the other two in their independent assessment. Maternal interview ratings for satisfaction and for interpersonal skills were higher after a more competent interview, and after a more patient-centred interview. There were significant and independent effects of both clinical competence and patient-centredness, with mothers valuing clinical competence more highly. An inconsistent effect of clinical competence on maternal recall was demonstrated, and patient-centredness was not associated with improved recall. Maternal ratings were significantly higher after the second interview. This was not explained either by which particular pair of interviews the mothers viewed or by the order of the interviews. Although the numbers in each of the groups for these comparisons was small, the study design employed ensured all groups had equal numbers of subject mothers so no particular interview combination had a greater influence on the results. A maternal calibration or learning effect following the first interview may explain the order effect observed. The influence of past experiences on maternal ratings requires further exploration, as does the generalizability of the findings of this study using videotaped interviews to actual clinical interview situations. Mothers were asked to attend for an additional 45 min prior to their child s appointment. This was a considerable imposition, and was the main reason cited for nonparticipation. Previous studies of patient evaluations have reported similar participation rates. 8,13,14 The quality of the interviewer s skills may be particularly important when mothers are rushed and possibly distracted. Maternal ratings clearly differentiated between the four interviews. Mothers, as active participants in medical consultations, showed that they are able to provide valuable information regarding medical interviewing skills. They are able to recognize different consultation qualities and practitioner skills, and to evaluate these. The extent to which mothers in this study valued clinical competence over patient-centredness was unexpected. Assessment of clinical competence traditionally is the province of the professional, not the consumer, who by implication is assumed to be less skilled in this area. The inconsistent effect of the study variables on maternal recall indicates the need for a better understanding of factors influencing this complex interview outcome. The findings of this study have clinical applications at all levels of medical practice from medical students, through postgraduate training to experienced professionals. Obtaining patient evaluations of clinical consultations could provide opportunities for continuous, personalized feedback for practitioners. Training programmes which lead to improved clinical competence and patient-centredness in health consultations are likely to lead to better health outcomes and a higher level of patient satisfaction.

7 Acknowledgements This study was supported by an Adelaide University, Faculty of Health Sciences B3 grant. References 1 DiBlasi Z, Harkness E, Ernst E, Georgiou A, Kleijnen J. Influence of context effects on health outcomes: a systematic review. Lancet 2001; 357: Stewart M, Brown JB, Donner A et al. The impact of patientcentered care on outcomes. J Fam Pract 2000; 49: O Keefe M, Sawyer M, Roberton D. Medical student interviewing skills and mother reported satisfaction and recall. Med Educ 2001; 35: Henbest RJ, Stewart MA. Patient-centredness in the consultation. 1: a method for measurement. Fam Pract 1989; 6: Newble D. Assessing clinical competence at the undergraduate level. Med Educ 1992; 26: Stiles W. Verbal response modes and dimensions of interpersonal roles: a method of discourse analysis. J Pers Soc Psychol 1978; 36: Medical student interviewing Winefield H, Murrell T, Clifford J, Farmer E. The search for reliable and valid measures of patient-centredness. Psychol Health 1996; 11: Kinnersley P, Stott N, Peters T, Harvey I, Hackett P. A comparison of methods for measuring patient satisfaction with consultations in primary care. Fam Pract 1996; 13: Schnabl GK, Hassard TH, Kopelow ML. The assessment of interpersonal skills using standardized patients. Acad Med 1991; 66: GenStat Committee of the Statistics Department, IACR-Rothamsted, Harpenden, Hertfordshire, UK. The Guide to GenStat. VSN International Ltd, Stewart M. Methods of scoring patient-centeredness. In Stewart MA, Weston WW, Brown JB (eds). Patient-centered Medicine. Thousand Oaks (CA): Sage, 1995: Mead N, Bower P. Measuring patient-centredness: a comparison of three observation-based instruments. Patient Educ Couns 2000; 39: Haddad S, Potvin L, Roberge D, Pineault R, Remondin M. Patient perceptions of quality following a visit to a doctor in a primary care unit. Fam Pract 2000; 17: Black AE, Church M. Assessing medical student effectiveness from the psychiatric patient s perspective: the medical student interviewing performance questionnaire. Med Educ 1998; 32: