4. MAAS-MI / Evidence Base – MAAS-Mental Health

4.4Construct Validity MAAS-Mental Health

The medical interview has three main goals: to gather information from the patient, to establish a relationship of trust and acceptance with the patient, and to convey information to the patient.

How successful is the MAAS Medical Interview Mental Health in helping to achieving these goals?

Construct validity is the degree to which a measurement method like MAAS is actually measuring what it sets out to measure. Testing the construct validity of MAAS Medical Interview Mental Health scales – Exploring Reasons for Encounter, History-Taking, Psychiatric Examination, and Socio-emotional Exploration – confirms that they all support the establishment of a climate of trust and acceptance, important values for the patient and for treatment success.

In This Chapter

Kraan, H. F., & Crijnen, A. A. M. (1987). Construct validity studies with the MAAS-Mental Health. In H. F. Kraan & A. A. M. Crijnen (Eds.), The Maastricht History-taking and Advice Checklist – studies on instrumental utility (pp. 331–353). Lundbeck, Amsterdam.

Construct validity

In general, the construct validity of a measurement method is supported when theoretical constructs underlying the method can be confirmed empirically. In this respect, the construct validity of the MAAS-MH is supported when this method can measure interviewing skills that contribute to the achievement of the three main functions of the medical interview in primary mental health care (Schouten, 1982).

We restate once more these functions which have already been discussed in Why The Medical Interview?

Function 1: Collection of information from the patient necessary for diagnostics and clinical problem-solving;
Function 2: Conveyance of information in order to inform the patient and to enhance his insight and compliance;
Function 3: Establishment and maintenance of a physician-patient relationship of trust and acceptance for the achievement of both previous functions.

In this chapter, an exploratory correlative study is carried out to investigate how the physician’s interviewing skills as measured with the MAAS-MI MH are related to the achievement of these functions.

Three Research Questions

From these functions of the medical interview, we derive three research questions. Are there significant validity coefficients between the physician’s interviewing skills measured with MAAS-MH variables and:

The quality of diagnosis and clinical problem-solving?
The degree of insight the patient has into his problems and his intention to comply?
The quality of the physician-patient relationship as experienced by the patient?

Construct Validity and the Medical Interview

In the literature, only one study is found reporting construct validity research that relates measured interviewing skills with outcome variables (Inui et al., 1982; Carter et al., 1982). The authors compare three interaction analysis instruments of medical interviewing (Bales, 1950; Roter, 1977; Stiles, 1978; see also Clinical Competency) to assess their capability for prediction of variance in several outcome measures: insight of patients in their problems, patients’ compliance and satisfaction. In general, 35-40% of the variance in these outcome variables can be predicted from interviewing skills. Later in this chapter, we refer to some findings of this study.

The first research question is studied by determining validity coefficients between measurements of diagnostics and clinical problem-solving and interviewing skills as measured with the MAAS-MH.

The second and third research questions are investigated by determining validity coefficients between measurements of the MAAS-MH variables and variables measuring the effect of the medical interview on the patient:

These patient variables are:
- the recall of information conveyed by the physician;
- the degree to which the patient feels facilitated to tell his story;
- the degree to which the patient feels disrupted in the communication by the physician;
- the degree to which the patient feels directed towards ideas and solutions of the physician;
- his insight into his own problems and his intention to comply with medical advice.
Both validity studies were carried out during the simulated consultation hour which has been elaborated extensively in Matrix Examined, but which we briefly recapitulate further.

First Function

The importance of the medical interview for diagnostics and, in a broader sense, for clinical problem-solving, has repeatedly been reported. Strong associations between the history-taking sections of the initial interview and clinical problem solving have been established in somatic problems (Elstein et al., 1978; Kassirer et al., 1978) as well as for minor psychiatric disorders (Goldberg et al., 1980), as we have elaborated in Medical Interview & Related Skills.

We expect significant validity coefficients between the physician’s scores on the MAAS-MH scales History-taking, Psychiatric Examination and Socio-emotional Exploration with measurements of diagnostics and clinical problem solving. High scores on these MAAS-MH scales tend to be indicative of the quality of data collection for hypotheses generation and testing.

Furthermore, it is expected that the physician’s scores on the scale Communicative Skills also has a significant validity coefficient with measures of diagnostics and clinical problem-solving. The physician’s communicative skills are to establish an effective and accurate exchange of data between physician and patient.

For communication to be effective, physician and patient should be mutually aware of the meaning one attaches to the message sent by the other

Consequently, physician and patient will be mutually aware of the meaning that one attaches to the messages sent by the other (Schouten, 1987).

Methods & results

Research setting, instruments, data-analyses and results are presented.

Research setting

Forty residents in general practice each interviewed two simulated patients, the first representing a major depression, the second a panic disorder. In this simulated consultation hour, each role was played by two patients who were randomly assigned to the residents. The interviews were rated live by trained observers and were videotaped. After the interview, the residents filled in a questionnaire measuring diagnostics and clinical problem-solving (see below), whereas the simulated patients filled out the Patient Satisfaction with Communication and responded to 6 open questions pertaining to the information conveyed by the physician (see below). After 4 months, the videotaped interviews were rescored by the same pool of trained observers.

Instruments

Residents’ interviewing skills were rated twice with the MAAS-MH; the first time, live and the second time, from the videotaped interviews to which the same pool of observers were randomly assigned. In this study, both sets of scores were summated to improve the reliability. The summated scores of the items per scale were used as indices for the residents’ interviewing skills.

Diagnostics and clinical problem-solving were measured by an instrument called Problem-Solving in Primary Mental Health Care (PS-PMHC), see TOOLS Cont’d. On the PS-PMHC, a semi-structured questionnaire, the physician has to respond to open-ended questions about differential diagnosis, explanatory hypotheses and hypotheses about further management. Scoring took place by comparing the physicians’ responses on the questions with criteria predefined by the panel of experts who also designed the roles of both simulated patients. Two raters scored the PS-PMHC according to the criteria and attained high reliability (Kappa=0.88, Cbhen, 1960). According to a calculation rule with preset weight factors, the correct answers have been summated into a numerical score on two variables, diagnosis and aetiology and patient management plan.

The variable diagnosis and aetiology (DIAG) is a summation score of the correctness of the diagnosis, the elaboration of the differential diagnosis and correctness of explanatory hypotheses. The variable patient management plan (HELP) is a summation of the correctness of the hypotheses about further patient management and the agreement between the request for help and the proposed patient management plan.

Data analysis

First, the descriptive statistics of both variables DIAG and HELP are given.

Secondly, the relationship of the residents’ interviewing skills to his diagnostics and clinical problem-solving is determined by Pearson’s correlations of both variables DIAG and HELP with the 8 MAAS-MH scale scores. To estimate the “case influence”, a dummy variable, taking the value 1 of the “depression case” and the value 2 for the “anxiety case”, are included in the correlation matrix.

In order to calculate validity coefficients from these correlations, a correction for attenuation is applied to the MAAS-scores and to the scores of diagnostics and clinical problem-solving (Guilford et al., 1982).

The validity coefficients are calculated from the Pearson correlations of 8 MAAS-MH scale scores with scores on both variables DIAG and HELP. In the formula of correction for attenuation, the inter-rater reliability (Kappa = 0.88, Cohen, 1960) of the variables DIAG and HELP are taken. As reliabilities for the MAAS-scales, the generalizability coefficients (random physician, observers; fixed items) for 2 observers are used (Table 4), because a summated set of scores over 2 observers are taken for these validity studies.

Results

Table 1 – 6 are presented at the end of the chapter.

Descriptive statistics of the variables DIAG and HELP over 40 residents and 2 cases are given in Table 1.

The distribution of these scores is almost normal whereas their variances are sufficient to allow further analyses.

The matrix of the validity coefficients of 8 MAAS-MH scale scores with the scores on both variables DIAG and HELP is given in Table 2. Moreover, the Pearson correlation of case influence with the other variables is included in the matrix.

We find moderate validity coefficients between DIAG with the scales History-taking and Socio-emotional Exploration. They are fairly high between DIAG and Psychiatric Examination (positive) and Communicative Skills (negative).

The variable HELP has moderate validity coefficients with the scale Psychiatric Examination and Presenting Solutions and (again) negatively with communicative skills. The variable DIAG and HELP both have considerable correlations with the dummy variable “case influence”.

Finally, we also encounter significant and moderate correlations of “case influence” with all MAAS-MH scales, except for Structuring the interview and Interpersonal Skills.

Table 3 -- Questions asked the patient about the information conveyed by the physician

Table 4 -- Recall of Conveyed Information (RCI), a rating scale to test the patiënt’s recall of information conveyed by the physician and it’s interrater reliability

Obtaining Information From Your Patient

Although already noted in Content Validity Mental Health, the ‘case influence’ in the scores on the MAAS-MH scales is again striking. Not only scales with many content categories, but also scales measuring process skills (Presenting Solutions and Communicative Skills) suffer from case specificity. The ‘case influence’ on the variables DIAG and HELP is also very notable as in all measures of medical problem-solving up until now (a.o. Elstein et al., 1978).

According to our expectation, the validity of the scales History-taking, Psychiatric Examination and Socio-emotional Exploration is supported by the variable DIAG. Salient is the lack of validity of Exploration of the Reasons for Encounter and the magnitude of the validity coefficient of Psychiatric Examination.

History-taking, Psychiatric Examination and Socio-emotional Examination all contribute to establishing diagnoses

Diagnosis as Symptom & as Aetiological Understanding

A closer look at the variable DIAG provides further insight into this validity issue: DIAG combines the quality of diagnosis on the symptom level (see also Medical Interview & Related Skills) with that on an aetiological explanatory level.

Diagnosis on the symptom level is brought about by information from Psychiatric Examination.
Diagnosis on the aetiological level is achieved by information-gathering during History-taking and Socio- emotional Exploration.

Apparently, in our study, the latter relationship cannot be so firmly established which may indicate a lack of construct validity in the scales History-taking and Socio- emotional Exploration and its absence in the scale Exploration of the Reasons for Encounter.

The validity coefficient between Communicative Skills and the variable DIAG, being negative and fairly high, is a conspicuous, but difficult-to-explain finding. It might point to deficient validity of this scale, probably based on a lack of reliability.

Patient Management Plan

With measurements of hypotheses concerning the patient-management plan and of their accordance with the request for help (HELP), the MAAS-MH scales show two moderate validity coefficients. In accordance with expectations, the scale Psychiatric Examination has a significant validity coefficient but at least similar findings should be expected from the scales History-taking and Socio-emotional Exploration which purport to measure the data collection necessary for treatment hypotheses.

A second significant validity coefficient is found between Presenting Solutions and the variable HELP. It might support the validity of the negotiation aspect in the scale Presenting Solutions. Negotiation is necessary to attain an accord between physician and patient on the treatment proposal. For this negotiation process, the skills to Explore the Reasons for Encounter are also considered to be necessary to attune the treatment proposal to the patient’s desires and needs (Tuckett, 1985). Therefore, a significant validity coefficient between the scale Exploration of the Reasons for Encounter with the variable HELP is missing.

Conclusion

The construct validity with respect to diagnostics and clinical problem-solving turned out to be supported for the MAAS-MH scales History-taking, Socio-emotional Exploration and Psychiatric Examination. Moreover, there is slight support in validity of the MAAS-scale Presenting Solutions with respect to measurements of patient management hypotheses and their accordance to the patient’s request for help. Also in this context, the MAAS-scale Psychiatric Examination takes the predominant role. Conspicuous is the lack of validity of the MAAS-scales measuring process skills of interviewing in respect to diagnostics and clinical problem solving.

Second & Third Function – Outcomes For The Patient

In this section, validity of the MAAS-MH in terms of relationships with variables measuring the effects of the interview on patients, specifically the conveyance of information from physician to patient, is assessed. In this respect, we heavily lean on the theory, described in Medical Interview & Related Skills, from which we derive hypotheses about significant validity coefficients with MAAS-variables.

We recapitulate the following six sets of theoretical relationships

For the Second Function

The patient’s Recall of conveyed information is considered to be enhanced by the scales Present Solutions and Interpersonal and Communicative Skills. Recall of information conveyed by the physician is improved by the use of appropriate methods to inform the patient (Ley, 1983), by physician’s addressing the patient’s concerns and expectations (Bartlett, 1981) and by an explanation suited to the patient’s frame of reference (Tuckett et al.1985).
The patient’s Insight is considered to increase by means of the information conveyed by the physician, by means of the Communicative and Interpersonal Skills and the skills to Present Solutions (Wolff et al, 1978; Putnam et al. 1985; Pendleton 1983).
The patient’s Intention to Comply with advice is influenced by the physician’s Interpersonal and Communicative Skills (Bartlett, 1981; Davis, 1968) and by the physician’s skills to Present Solutions (Eisenthal et al. 1976).

For the Third Function

Interviewing skills pertaining to the scales Exploring Reasons for Encounter, Interpersonal Skills and Communicative Skills and, to a lesser extent, Socio-emotional Exploration, enhance the patient’s feeling of being Facilitated by the physician. These feelings of Facilitation are elicited when the patient is encouraged to describe his problems and coexisting emotions and attributions, and to give additional information in an interviewing climate of trust and acceptance (Korsch et. al. 1972; Eisenthal et al., 1976; Wolf et al. 1978; Stiles et al., 1979; Putnam et al., 1985).
By contrast, the patient’s perception of Disrupted Communication might be induced when he is not able to tell ‘the story’ from his own frame of reference. The variable Disrupted Communication is thus considered as the counterpart of Facilitation or affective satisfaction (Wolf et al., 1978).
The patient’s perception of being directed towards ideas and solutions in the frame of reference (Directivity) may be influenced by interviewing skills as measured by the scales History-taking, Psychiatric Examination, and Socio-emotional Exploration. Moreover, process skills as measured by the scale Structuring the interview may have a relation to this patient variable.

These theoretical relationships are investigated in an explorative study with validity coefficients.

Methods & Results

In this section, we discuss the instruments for measuring the variables and, subsequently, the analysis of the data.

Instruments

In addition to the MAAS-MH, two other instruments are used: the Patient Satisfaction with Communication Checklist (PSCC) and an instrument to measure the patient’s recall of information conveyed by the physician.

With the 19-item PSCC, extensively described and studied in Measuring Patient Satisfaction, the following variables are measured: Facilitation (3 items), Disrupted Communication (4 items), Directivity (3 items), Insight (5 items) and Intention to Comply (4 items).

After the completion of the PSCC, the simulated patients were asked to respond on paper to 6 open questions pertaining to the information conveyed by the physician. These questions are stated in Table 3.

The written responses are taken as the information conveyed by the physician and recalled by the patient. To measure the quality of recall, this written recalled information is compared with the information the physician actually conveyed during the interview. This comparison is carried out with a specially constructed instrument, the Recall of Conveyed Information (RCI). The most important part of this instrument is given in Table 4.

The RCI enables the researcher to split the total amount of conveyed information into 5 categories: information conveyed about diagnosis, aetiological conditions, treatment-related advice, further examination/tests and referral. Two observers independently compared the information conveyed by the physician by reviewing the videotaped interviews with the responses given by the simulated patient concerning the information conveyed to thee. Both observers rated the degree of agreement on the RCI scales.

Data analysis

Firstly, descriptive statistics and reliabilities of the 5 PSCC variables and of the variable recall of conveyed information are calculated.

Secondly, the effects of the residents’ interviewing skills on the simulated patients are explored by Pearson correlations. The 8 scores on the MAAS-MH scales are correlated with each of the 5 PSCC variables facilitation, disruptive communication, directivity, insight and intention to comply. From these correlations, validity coefficients are calculated after correction for attenuation.

As reliability to be used in the formula for correction for attenuation, the alphas of the 5 (composed) variables of the PSCC are taken. For the same purpose, we take the averaged inter-rater reliability of the (composed) variables Recall of Conveyed Information for the correction for attenuation.

To estimate the case influence, a dummy variable, taking the value 1 for the ‘depression case’ and the value 2 for the ‘anxiety case’, is included in the correlation matrix.

Results

Descriptive statistics and reliabilities of recall of conveyed information (see Table 4) the PSCC variables facilitation, disruptive communication, directivity, insight and intention to comply (see Table 5) permit further research, although the distributions of disrupted communication and intention to comply are somewhat skewed.

The validity coefficients between the scores on the MAAS-MH scales with the five PSCC variables and recall of conveyed information are presented in Table 6.

From general inspection of the matrix, it appears that case influence significantly (p<.01) and sometimes substantially correlates with variables such as directivity (.59), insight (.55) and recall of conveyed information (.80).

Furthermore, there is a considerable amount of moderate to reasonable validity coefficients between MAAS-scales and PSCC variables, except with the variable “intention to comply”. We discuss these validity coefficients in the next section.

Table 7 -- Overview of theoretically relevant validity coefficients of MAAS-MI MH scales with Patient Satisfaction with the Communication Checklist (PSCC) variables and the variable Recall of Conveyed Information

Second Function: Conveying Information

In this study we examine the validity of the MAAS-MI MH scales by means of their validity coefficients with variables measuring the effect of the interview on the patient, more specifically the conveyance of information from physician to patient. In previous paragraphs we postulated 6 groups of hypotheses concerning these validity coefficients. These are subsequently discussed.

Recall of Information

The patient’s recall of information conveyed by the physician, has a modest, positive validity coefficient with the MAAS- scale Structuring the interview and a fairly high, negative validity coefficient with History-taking.

The modest validity coefficient with Structuring the interview may point to a positive effect on the recall and retention of information when the interview is structured according to the criteria implied in the MAAS-MH. However, we lack validity coefficients with Presenting Solutions and Interpersonal and Communicative Skills and neither does the negative validity coefficient with History-taking contribute much to construct validity. This relationship is understandable from the assumption that extensive History-taking has been carried out at the cost of the physician’s conveyance of information and its subsequent recall by the patient.

Extensive History-taking has a cost: you will convey less information and your patient can recall less of what you have said

In summary, we are not able to confirm with the MAAS-MH the recommendations about effective conveyance of information: for example, simplification and explicitation of information (Ley, 1983), checking of comprehension (Tuckett, 1985) and coping with defensive mechanisms after conveyance of bad news (a.o. Schouten, 1982).

The underlying reason for the poor contribution of recall to the construct validity of the MAAS-MH might be deficiencies in the method to measure the recall of information conveyed by the physician to the simulated patient (RCI). It is questionable whether the information simulated patients write down in response to the open questions (see Table 3) reflects in a valid way the information the physician has conveyed. The response may be confounded with pre-existent knowledge about illness conditions and treatment from his simulated role. Furthermore, the response may be influenced by the simulated patient’s ability to express himself in written language.

In addition, the strong case influence (r=.80) in recall might act as a confounder acting in suppressing potentially valid correlative relationships.

Insight

The patient’s Insight resulting from the information conveyed by the physician has significant validity coefficients with the MAAS-MH scale Presenting Solutions (r=.63). This fairly high validity coefficient is a logical consequence of the nature of this scale which measures the conveyance, the discussion and the bargaining of relevant information. When Presenting Solutions has been well achieved then, of course, it entails increased insight for the patient into his problems.

Presenting Solutions entails increased Insight into diagnosis and underlying problems

The strong influence of case specificity is also here notable (r=.55), which is self-evident. Patients differ in their knowledge of mental health and disorder, often disposing of very idiosyncratic fabrics of knowledge. Tuckett (1985) therefore argues that the patient’s insight can only be fostered when the physician himself has some insight into the patient’s frame of reference pertaining to his case.

We here also lack positive validity coefficients with Interpersonal and Communicative Skills. The reliability of these scales may be too low to form a necessary condition for validity.

The negative validity coefficient of Insight with History-taking may be due to the counteractive effects of extensive History-taking on improving insight into the problems by the patient. In this respect, we have already assumed, in the previous subsection, a similar counteractive influence of History-taking on the recall of conveyed information.

Intention to comply

Disappointingly, the patient’s intention to comply with the proposed medical advice does not show any positive validity coefficient with one of the MAAS-MI MH scale scores: or, at any events, we lack positive validity coefficients with Presenting Solutions, Interpersonal and Communicative Skills. Surprisingly, a modest negative validity coefficient is found with Communicative Skills. This finding, which is rather inexplicable, is probably again due to the low reliability of this scale.

We do not confirm the often-claimed relationship of Interpersonal and Communicative Skills with Intention to comply in this study (a.o. Davis, 1968; Bartlett, 1981; DiMatteo and DiNicola, 1982; Ley, 1983). Again, the lack of reliability is probably the explanation for the insufficient validity with respect to Intention to Comply.

Third Function: Climate of Trust & Acceptance

Facilitation

The feelings of the patient of being facilitated to tell his own story (Facilitation) has a significant validity coeficient with all MAAS-MI-scales, except Communicative Skills. The coefficients with the scales Socio-emotional Exploration and Interpersonal Skills are even considerable to fairly high. This finding agrees more or less with our expectations because these latter scales pertain to the interviewing skills leading to facilitation and fostering the expression of patient-centered information. Nevertheless, we expected the validity-coefficient with the scale Exploring Reasons for Encounter (.33) to be higher.

Patients with Mental Health issues experience systematic and directive questioning from History-taking and Psychiatric Examination as facilitative – contrary to patients with somatic complaints

Significant validity coefficients with the scales History-taking and Psychiatric Examination point to the conclusion that patients may consider periods of systematic and directive questioning by the physician as facilitative. However, these findings also may be due to the validity problem of the scale Facilitation in which patients tend to evaluate their relationship with the physician as overly satisfactory. Nevertheless, a significant validity coefficient with Communicative Skills is missing.

Disrupted Communication

The patient’s feeling of being disrupted in the narration of his story (Disrupted Communication) shows a correlative pattern with MAAS-variables which is globally the reverse of that of Facilitation. This is to be expected because the patient-variable Disrupted Communication is theoretically the opposite of Facilitation, the measure of the patient’s affective satisfaction. We find negative and significant validity coefficients with the MAAS-scales Exploring Reasons for Encounter, Structuring the interview (both moderate) and with Socio-emotional Exploration (modest), according to expectation.

We miss (negative) correlations with Interpersonal and Communicative Skills. This may indicate that the patients feel more disrupted when hampered in the expression of certain factual or emotional information, than by the way they are allowed to express it. The negative correlation with Structuring may indicate that this MAAS-scale measures a way of structuring the interview in phases that is generally considered as facilitative and satisfactory by the patient.

All these findings agree with related reports that systematic questioning must neither suppress the expression of feelings nor must it be experienced by informants as unduly intrusive or lacking in understanding (Cox et al., 1981).

Directivity

The patients perception of being directed towards ideas and solutions in the physician’s frame of reference (Directivity) has moderate to fairly high validity coefficients with the MAAS-scales Exploring Reasons for Encounter, History-taking, Psychiatric Examination and Socio-emotional Exploration. These negative validity coefficients may indicate that the physician’s skills pertaining to these scales are experienced by the patient as ‘not directive’. This is also in accordance with a previous remark that the influence of systematic questioning and structuring does not influence patient satisfaction negatively. The variable Directivity, not really described in the literature, does not indicate negative effects on the patient. In this context, Carter et al. (1982) has also reported positive effects of physician’s imperative direction on the degree to which patients feel more informed about their problems.

Furthermore, it is conspicious that Directivity has similar negative validity coefficients to Exploring Reasons for Encounter and to the scales History-taking, Psychiatric Examination and Socio-emotional Exploration. This finding confirms a conclusion drawn in the previous chapter: it is difficult to draw a distinction between the variables Exploration of the Reasons for Encounter and History-taking, both pertaining to patient-centered information.

In Directivity, a strong case influence is finally notable (r=.59). This finding means that the directivity perceived by the patient is very dependent on the nature of the problem and the way the patient presents the problem. Certain problems, or modes of their presentation, may apparently elicit more directivity from the physician than other problems.

Conclusions

In Table 7, an overview of the theoretically relevant validity coefficients of the MAAS-MI MH scales with Patient Satisfaction with Communication variables, as well as with the variable Recall of conveyed information, is presented.

In summary, the construct validity of the MAAS-MI MH scales Exploring Reasons for Encounter, History-taking, Psychiatric Examination and Socio-emotional Exploration is supported by the Patient Satisfaction with Communication variables as a criterion. To a lesser extent, the same holds for the scales Presenting Solutions and Structuring the interview. The construct validity of the scales Interpersonal and Communicative Skills seems modest to low.

Construct validity of MAAS-MI MH scales Exploring Reasons for Encounter, History-taking, Psychiatric Examination and Socio-emotional Exploration support the establishment of a climate of trust and acceptance, important values for the patient

Some important validity coefficients with the MAAS-MI MH scales could not be established such as Recall of Conveyed Information and, in particular, Intention to Comply. The former my be partially due to a lack in validity in the way the recall of information is measured. The main cause, however, seems to be a deficit reliability and therefore a lack of validity of the scales Interpersonal and Communicative Skills.

Furthermore, the case influence in both MAAS-MI-scales and PSCC variables may also play a confounding role in causing spurious significant and suppressed non-significant correlations (Guilford et al., 1982).

Finally, our statements about the validity coefficients should be treated with caution and are definitely not be taken as causal relationships. Within the framework of this study, the validity coefficients are no more than inductive and hypothesis-generating in character. They are insufficient to confirm or refute hypotheses about construct validity.

Restrictions in the research situation also attenuate our validity statements.

The simulated consultation hour might be a threat to the ‘ecological validity’ of our findings. In particular, the PSCC-scores of simulated patients may be questionable, especially in the more affective variables, such as Facilitation. Furthermore, the validity of the information recalled by simulated patients is questionable, irrespective of the measurement method itself.

Construct Validity On Functions Of Interviewing In Mental Health

Three Functions

In this explorative, correlative study, issues of construct validity of the MAAS-MH have been investigated. Underlying theoretical relationships for construct validity have been derived from the three functions of initial medical interviews, as stated by Schouten et al. (1982) and further elaborated in Medical Interview & Related Skills:

Medical interviewing skills are intended to collect information from the patient enabling diagnostics and clinical problem-solving.
Furthermore, the physician should convey information on problems and possible solutions, resulting in increased insight and compliance.
Finally, the physician should build and maintain a relationship of trust and acceptance in which the previously mentioned functions of the medical interview can be performed.

Construct Validity

It has been stated that the construct validity of the MAAS-MH is supported if this method is able to measure interviewing skills which have validity coefficients with measured variables such as:

Quality of diagnostics and clinical problem solving;
Patient’s recall of conveyed information, insight into their problems and intention to comply with advice;
Patients’ feelings of being facilitated or disrupted in their communication.

Diagnostics & Problem-solving

With respect to diagnostics and problem-solving, the MAAS-scales History-taking, Socio-emotional Exploration and, in particular, Psychiatric Examination, show significant validity coefficients indicative of their construct validity. Measurements of the quality of the patient-management hypotheses and their accordance with the patient’s request for help give some support to the scales Psychiatric Examination and Presenting Solutions.

Conveyance of Information

With respect to the conveyance of information from physician to patient, some validity coefficients with the MAAS-MH scales could not be established, such as recall of conveyed information and, in particular, intention to comply. The former may be partially due to a lack in validity in the way the recall of conveyed information is measured in this simulated consultation hour. The main cause, however, seems to be a deficit of reliability and therefore a lack of validity of the scales Interpersonal and Communicative Skills.

Relationship of Trust and Acceptance

With respect to the relationship of trust and respect, the construct validity of the MAAS-MI MH scales Exploration of the Reason for Encounter, History-taking, Psychiatric Examination and Socio-emotional Exploration is supported by the Patient Satisfaction with Communication variables as a criterion. To a lesser extent, the same holds for the scales Presenting Solutions and Structuring the interview.

The methodology used in this study only permits cautious statements about construct validity. The validity coefficients found in this exploratory study are indicative of construct validity, but are not confirmative. Our results therefore have only a hypothesis generating character. It is necessary to carry out further research to establish whether confirmation or refutation of these hypotheses is possible. This will require methodology with regression and path analysis and the opportunity for cross-validation in more samples.

Furthermore, it is necessary to point to other restrictive considerations concerning cur findings:

Firstly, the simulated consultation hour procures standardization and comparability at the cost of ecological validity. In particular, the PSCC-scores of simulated patients may be questionable, especially in the more affective variables such as facilitation. Furthermore, the validity of the information recalled by simulated patients is questionable, irrespective of the measurement method itself.
Secondly, the low to moderate reliability of same scales such as Presenting Solutions and Communicative and Interpersonal Skills, has put restraints on these validity studies, the correction for attenuation.
Thirdly, case influence contributes much variance to the MAAS-MH scales and to same variables such as insight, recall of information and directivity. It is plausible that this variable may provoke spurious correlations or suppress potentially valid correlations (Guilford et al., 1982).

Table 1 -- Descriptive statistics of the variables Diagnosis and Aetiology and Patient Management Plan of 40 residents, interviewing two simulated patients

Table 2 -- Matrix of validity coefficients of scores on MAAS-MI MH scales with measurements of diagnostics and problem-solving (after correction for attenuation)

Table 6 -- Matrix of validity coefficients between scores on MAAS-MI MH scales with patient variables measuring Facilitation, Disrupted Communication, Directivity, Insight, Intention to Comply, and Recall of Conveyed Information

References

Selected Reading

Eisenthal S, Koopman C, Lazare A. Process analysis of two dimensions of the negotiated approach in relation to satisfaction in the initial interview. Journal of Nervous and Mental Diseases, 1983; 171: 49-54.

Elstein AS, ShuIman LS, Sprafka SA. Medical problem solving. An analysis of clinical reasoning. Harvard Univ. Press, Cambridge Mass., 1978. Goldberg D, Huxley P. Mental illness in the community; the pathway to psychiatric care. Tavistock Publ., London/New York, 1980.

Inui TS, Carter WB, Kukull WA, Haigh VH. Outcome-based doctor- patient interaction analysis. I. Comparison of techniques. Medical Care, 1982; 20: 537-549.

Ley P. Patients’ understanding and recall in clinical communication failure. In: Pendleton D, Hasler J (Eds.). Doctor-patient communication. Academic Press, London, 1983.

Schouten JAM. Anamnese en advies. Stafleu, Alphen a/d Rijn/Brussel, 1982.

Stiles WB. Verbal response mades and dimensions of interpersonal roles: a method of discourse analysis. Journal of Personality and Social Psychology, 1978; 36: 693-703.

All References

Bales RF. Interaction process analysis. Addison Wesley, Cambridge, 1950. Bartlett EE. The contribution of consumers’ health education to primary care practice: a review. Medical Care, 1980; 18: 862-871.

Carter WB, Inui TS, Kukull WA, Haigh VH. Outcome-based doctor- patient interaction analysis. II. Identifying effective provider and patient behavior. Medical Care, 1982; 20: 550-566.

Cohen J. A coefficient of agreement for nominal scales. Educational and Psychological Méasurement, 1960; 20: 37-46.

Cox A, Rutter M, Holbrook D. Psychiatric interviewing techniques V. Experimental study: eliciting factual information. British Journal of Psychiatry, 1981; 139: 29-37.

Davis MS. Variations in patients’ compliance with doctor’s advice: an empirical analysis of patterns of communication. American Journal of Public Health, 1968; 58: 274-288.

DiMatteo MR, DiNicola DD. Achieving patient compliance. The psychology of the medical practitioner’s role. Pergamon Press, New York, 1982.

Guilford JP, Fruchter B. Fundamental statistics in psychology and education. McGraw-Hill, Auckland, 1982 (6th ed.).

Inui TS, Carter WB, Kukull WA, Haigh VH. Outcome-based doctor- patient interaction analysis. I. Comparison of techniques. Medical Care, 1982; 20: 537-549.

Kassirer JP, Gorry GA. Clinical problem solving: a behavioral analysis. Annals of Internal Medicine, 1978; 89: 245-255.

Korsch BM, Negrete VF. Doctor-patient communication. Scientific American, 1972: 66-74.

353

Ley P. Patients’ understanding and recall in clinical communication failure. In: Pendleton D, Hasler J (Eds.). Doctor-patient communication. Academic Press, London, 1983.

Pendleton D. Doctor-patient communication; a review. In: Pendleton D, Hasler J (Eds.). Doctor-patient communication. Acadend_c Press, London, 1983.

Putnam SM, Stiles WB, Jacob MC, James SA. Patient exposition and physician explanation in initial medial interviews and outcomes of clinic visits. Medical Care, 1985; 23: 74-83.

Roter DL. Patient participation in the patient-provider interaction: the effects of patient question asking on the quality of interaction, satisfaction and compliance. Health Education Monographies, 1977; 5: 281-315.

Schouten JAM. Anamnese en advies. Stafleu, Alphen a/d Rijn/Brussel, 1982.

Stiles WB. Verbal response mades and dimensions of interpersonal roles: a method of discourse analysis. Journal of Personality and Social Psychology, 1978; 36: 693-703.

Stiles WB, Putnam SM, Wolf MH, et al. Interaction exchange structure and patient satisfaction with medical interviews. Medical Care, 1979; 17: 667-679.

Tuckett D. Meeting between experts: an approach to sharing ideas in medical consultations. Tavistock, London, 1985.

TUckett D, Williams A. An approach to the measurement of explanation and information giving in medical consultations: a review of empirical studies. Social Science in Medicine, 1984; 18: 571-580.

Wolf MH, Putnam SM, James SA, Stiles WB. The medical interview satisfaction scale: development of a scale to measure patient perceptions of physician behavior. Journal of Behavioral Medicine, 1978; 1: 391-401.