3.3Interviewing Skills & Clinical Competency

Crijnen, A. A. M., Post, G. J., Kraan, H. F., van der Vleuten, C., & Zuidweg, J. (1987). Interviewing skills and medical competence. In H. F. Kraan & A. A. M. Crijnen (Eds.), The Maastricht History-taking and Advice Checklist – studies of instrumental utility (pp. 233–248). Lunbeck, Amsterdam.

The validity of the Maastricht History-taking and Advice Checklist – General is examined by correlating MAAS-scores with other measurements of medical competence. Physicians’ medical knowledge, interpersonal skills, care and concern for the patient and problem-solving skills are assessed in addition to the measurement of medical interviewing skill. Medical interviewing skills are confirmed by measurements of interpersonal skills, by measurements of care and concern and by measurements of problem-solving skills as far as they concern information exchange.

Validity coefficients of medical competency support the validity of MAAS-Medical Interview

The study reveals furthermore, that medical interviewing skills are clearly discerned from medical knowledge and the formulation of a treatment plan. The validity coefficients, therefore, support the validity of the Maastricht History-taking and Advice Checklist.

Validity of Measures of Medical Competency

A plea has been made recently for the study of the validity of measurements of medical competence (Katz, 1982; Gonella, 1985). Some have said that greater attention should be given to the valid assessment of physicians’ actual job performance, whereas others have observed that medical competency cannot be conceptualized as a single variable, since knowledge, problem-solving skills, interviewing skills and attitudes should be distinguished from each other. 

In medical competency, knowledge, problem-solving skills, interviewing skills, attitudes and examination skills can be distinguished

Validity of Measures of Medical Competency – Requirements

Validity studies of methods of measurements of medical competency should be designed according to generally accepted, scientific criteria. On reviewing the literature, the following criteria were detected (Cronbach et al, 1955; Thorndike, 1982). 

Clear Definition and Theory of Competency

The first criterion requires a clear definition of and theory for the competency under study in order to ascertain the match between the conceptualization of the competence and its empirical measures, and to distinguish the competence under study from other distinct competencies: e.g., history-taking skills should be defined clearly in order to achieve agreement among researchers about the content of a test designed to measure history-taking skills. Moreover, the definition of History-taking Skills should augment the distinction from other competences, such as Interpersonal Skills or Medical Knowledge. 

Same Competence Despite Different Measures

The second criterion requires that validity is empirically underscored by different methods of measurement of the same competence. Evidence from different sources gathered in different ways should all point to the same competence domain: e.g., different methods of measurement of a physician’s interpersonal skills are expected to correlate higher with each other than with measurements of his other competencies. 

And Differentiated From Other Competencies

The third criterion requires that methods of measurement of the competence under study are differentiated empirically from measurements of other competences (Cronbach et al, 1955; Kerlinger, 1973). Measurements ought to show the difference between theoretically discernable medical competences: e.g., the skills to Explore the Reasons for Encounter are expected to correlate low with Medical Knowledge or the quality of physician’s Treatment Plan. 

Validity In Practice

These three criteria constitute a scientific model of how the validity of measurements of medical competency ought to be examined. Three validity studies of measurements of medical interviewing skills were scrutinized by applying the criteria (Stillman et al, 1977; Brockway, 1978; Swanson et al, 1981). 

In the first study by Stillman et al (1977), who examined the validity of the Arizona Clinical Interview Rating Scale (ACIR), the 16 interviewing skills measured by the ACIR-scale are grouped into six major subsections which are treated ultimately as if they all contribute to a similar competence domain. Since the items purport to different skills, the correctness of this procedure can be questioned. No methods intending to measure the same competence domain were taken into account. However, the scores on the ACIR-scale were correlated with Medical College Admissions Tests which are supposed to measure a different competence. This study therefore fulfils only one of the three aforementioned criteria.

The second study under consideration is by Brockway (1978), whose analysis of the validity of her interview rating scale shows same shortcomings because the content of the two subscales, respectively Relationship Skills and Problem-solving Skills, is rather heterogeneous and can easily be interchanged. The differences between the subscales are not well defined. No measurements indicating a similar competence domain were taken into account. The study, however, fulfilled one of the three criteria, because the rating scale was compared to two different measurements of medical competence; namely, data collection and problem identification. 

The third study under consideration is by Swanson et al (1981), who compared the validity of three measurements of medical interviewing skills, the ACIR-scale, an Interaction Analysis-rating, and a History and Physical Exam checklist. The heterogeneity of the measures hindered the researchers in adequately interpreting the matrix of correlations between the methods. They concluded that, in essence, no evidence for construct validity through comparison between measurements had been obtained.

By reinterpreting the correlation matrix, it is possible to arrive at a different conclusion. When all measurements are assigned as measurements of interpersonal skills, communicative skills or physical examination skills, each of the three measurements of interpersonal skills appears to correlate significantly with the others: this conclusion supports validity. Moreover, measurements of interpersonal skills can be distinguished from communicative skills and physical examination skills. The same is not true for measurements of communicative skills. The study fulfilled two of three criteria, because measurements with the same and a distinct meaning were applied simultaneously. 

Validity Studies of Medical Competency

• All studies reveal difficulties in defining the measurements under examination;
• Only one provides evidence of validity by applying measurements indicating a similar competence domain;
• Only two studies differentiated medical interviewing skills from measurements of medical knowledge, data collection, diagnosis and treatment. 

Validity MAAS-MI Assessed

The validity of the Maastricht History-taking and Advice Checklist – General is established by taking into account the three criteria mentioned previously.

• Five distinct types of medical interviewing skills distinguished in the MAAS-G, were defined and treated separately.
• Furthermore, measurements with a similar theoretical intention to the MAAS-G-scales and measurements indicating distinct competences were applied simultaneously.

This preliminary study was conducted before the reliability and scalability of the MAAS-G were examined thoroughly by means of the procedures described in Why the Medical Interview? All items are thus taken into account as is the scale Basic Interviewing Skills which had not yet been divided into Interpersonal Skills and Communicative Skills. 

5 Medical Interviewing Skills VS 7 Measures of Medical Competency

The validity of the MAAS-MI General was examined by correlating MAAS measurements of five types of medical interviewing skills with seven distinct measures of medical competence:

• Five types of interviewing skills pertain to:
  • Exploration of Reasons for Encounter
  • History-taking Skills
  • Presenting Solutions
  • Structuring
  • Basic Interviewing Skills.
• Seven measurements of medical competence are:
  • Medical knowledge tests
  • Expert judgments on interpersonal skills
  • Expert judgments of care and concern for the patient
  • Four measurements of medical problem-solving:
    • Subjective Information, 
    • Objective Information,
    • Diagnosis or Problem-definition,
    • Treatment Plan.

The delineation of constituents of medical competence was derived largely from the work of Fabb and Marshall (1983). 

Simulated Patients  Simulated patients, lay-people staging a medical complaint, were used because they provide the opportunity for assessing medical interviewing skills as well as physical examination skills when real patients cannot be used (Stillman, 1983). In this study, the simulated patients were lay people, chosen from a large group of simulated patients who were trained by the Skills Laboratory at Maastricht Medical School to simulate complaints in the undergraduate-medical curriculum.

Two simulated patients presented respectively complaints of fatigue and dyspnoe in exertion, and low back pain without irradiation. The correlations between all measurements are expected to support either the similar meaning or the distinct character of MAAS-G-measurements of medical interviewing skills and, as a secondary goal, to highlight the interrelations between interviewing skills and other domains of medical competence. 

Discussion

In general, the validity coefficients between MAAS-MI measurements of medical interviewing skills and seven other measurements of medical competence support the validity of the MAAS-G, although some unexpected dissonants are found.

Measures of medical competency support the validity of MAAS-MI 

Validity & MAAS-MI

Exploring Reasons for Encounter

The scale Exploring Reasons for Encounter converges with ratings of interpersonal skills and subjective information as expected. The low correlation with medical knowledge and treatment plan, which have also been found in other studies (Stillman et al, 1977; Brockway, 1978), underscore the distinct character of these competence domains. The unexpectedly high validity coefficient with assessment of diagnosis may be explained by the necessity to ask patient-centered questions in order to include in the diagnosis issues defined by the patient as a problem.

Patient-centered information needs to be included in diagnosis and treatment plan

This is especially true for cases with a combination of somatic and psychological problems in which patient’s concerns and real-life circumstances have to be included in the assessment of the diagnosis (Mishler, 1982). The moderate correlation with the care- and concern-variable supports the validity of the scale Exploring Reasons for Encounter because an attitude of caring and reassurance during the physical examination reflects a human dimension in the physician’s approach to the patient, also reflected in the skills to explore the reasons for encounter. 

History-taking

The scale History-taking, which measures the collection of present and past medical data, is highly correlated with measurements of subjective and objective information during medical problem-solving. Medical problem-solving is seen as a collaboration of several competences of which the search for additional data by means of history-taking is considered to be the communicative aspect (Neufeld et al, 1981). Since history-taking skills are determined by the process of medical problem-solving, the strong correlations underscore the validity of the MAAS-MI G-scales.

Medical problem-solving drives the process of History-taking

The low correlations of history-taking with medical knowledge and treatment plan support the distinct character of these measures of medical competence. The moderate correlation between History-taking and assessment of diagnosis suggests that physicians who collect more data have a better chance of establishing an accurate diagnosis. Although this finding is supported by Brockway (1978), other studies reveal that the number of questions asked during history-taking is not unequivocally related to the quality of diagnosis (Kassirer et al, 1978). Specialists collect less information and mention the correct diagnosis earlier in comparison with non-specialists who often revert to a general review of organ systems.

Presenting Solutions

The scale Presenting Solutions does not correlate strongly with any of the other competences. The validity coefficients with measurements of interpersonal skills, care and concern as well as with subjective information are moderate. The strength of these correlations indicates the continuous nature of information-exchange and the human factor of patient-centeredness and reassurance, which are also constituents of the scale Presenting Solutions. The low correlation between this scale and treatment plan was expected because the MAAS-MI G-scale deals only with the process of information-exchange and negotiation and not with the content of the exchanged information, the treatment plan itself. This argument also holds for the low validity coefficients with medical knowledge, objective information and assessment of diagnosis, which underscores once more the distinct character of the scale Presenting Solutions. 

Structuring

The scale Structuring is strongly correlated with the care- and concern-variable, which measures the physician’s care in reducing the patient’s anxiety during the physical examination. Items in this MAAS-MI G-scale measure the quality by which the physician structures the interview into naturel segments enabling the patient to voice his concerns and to understand the goal of certain interview behavior.

Structuring the interview is an important tool to show care and concern and reduce anxiety in your patient

Although not yet studied, it is conceivable that interviewing skills which structure the interview will reduce the patient’s anxiety. In analogy with the reassuring effects of introduction and explanation of the procedures during the physical examination, similar behavior by the physician during the interview may entail similar effects. The low correlation with medical knowledge, objective information and treatment plan indicate a divergency with knowledge and problem solving skills as expected. 

Basic Interviewing Skills

The scale Basic Interviewing Skills accompanies measurements of interpersonal skills and of subjective information during medical problem solving. This agrees with the underlying aim of the scale, which is to measure a physician’s ability to establish an optimal rapport with the patient and to induce an effective exchange of information. This pertains also to the moderate correlation with care and concern. The moderate correlations with objective information and assessment of diagnosis might be explained by the statement of DiMatteo and DiNicola (1982), that a physician’s competence is likely to involve a scientific and technical ability translated into practice through both interpersonal skills and the art of medicine. The low validity coefficients with medical knowledge and treatment plan underscore the distinct character of the pertinent competences. 

Methodological Considerations

Methodologically, the present study posed several problems.

Firstly, the reliability indices of the MAAS-MI G were, unexpectedly, low to moderate: in prior studies, higher inter-observer reliability has been obtained. Further exploration by means of generalizability studies revealed that the coefficients differed markedly for both patients (respectively, .62 and .37). This difference in reliability was attributed at least partly to the controlling communication style of one of the simulated patients. Since MAAS-MI G-items are mainly directed at the physician’s interviewing skills, the observers easily make mistakes when patients take the initiative so intrusively. Item definitions and criteria for scoring cannot deal appropriately with this situation. 

Secondly, the strong impact of the cases on physicians’ interviewing skills influenced the validity study to some extent. The influences of the different cases may be attributed to two factors: the presentation of the case by the simulated patients and the characteristics of the medical problem. Prior studies with the MAAS-MI G provided evidence that the case as medical problem did not influence the physician’s interview behavior significantly (Kraan et al, 1986), whereas others have pointed to a considerable impact of the characteristics of the medical problem on the physician’s interview behavior (Norman et al, 1981). Interpretation of the study presented here suggests that case influences seem to result from the case presentation by the simulated patients.

This conclusion is corroborated by the finding that history-taking skills were influenced least by the difference in cases whereas, at first sight, the greatest influence of the case as medical problem was expected on this type of interviewing skills. 

Conclusion

The MAAS-MI validly measures 5 distinct types of medical interviewing skills, because the validity coefficients confirm the strength:

• With related medical competences, such as interpersonal skills, care and concern, and information-exchange;
• With unrelated medical competences, like knowledge and the quality of a treatment plan, and display the distinct character of MAAS-MI measurements of medical interviewing skills. 

The study supports the model of distinct medical competences as delineated by Fabb and Marshall (1983) and underscores that the evaluation of students’ or physicians’ medical competency can no longer be based solely on the assessment of their medical knowledge as is the case in most examinations. Medical interviewing skills should be taken into account.