Validation of the generic medical interview satisfaction scale: the G-MISS questionnaire

A patient has about seven doctors’ consultations per year [1]. Physicians are getting more involved in the quality assessment of daily practice [2], but few questionnaires have been validated to assess patients’ experiences with medical interviews [3, 4]. Patient-reported outcomes are considered valuable measures of healthcare and a step in the development of patient-centered care [5]. The assessment of patients’ experiences could enhance comparisons of strategies about physicians’ communication [6], treatment [7], or accountable care [8]. Available questionnaires in the field are not psychometrically sound [9], rely on expert generated items [3], and focus on specific physicians’ specialties [10] or specific patient courses [11, 12]. The Medical Interview Satisfaction Scale (MISS) was developed to assess patients’ experiences with interviews in primary care [4]. The authors used a rigorous method for item generation with patient interviews, but the factorial structure relied on the classical test theory (CTT) used at the time of questionnaire development [13]. Despite its predicted use in general practice, the original questionnaire tended to be a reference for the evaluation of patient-centered consultations [14‐16]. Some questions were raised about the internal validity and acceptability of its 29 item form [17], and other authors suggested the factorial structure may differ across populations, stressing the need for a new validation process [18].

The main objective of the Generic Medical Interview Satisfaction Scale (G-MISS) study was to validate a generic version of the MISS questionnaire in general practice, medical specialties, and surgical specialties. The secondary objectives were to reduce the number of items and to explore the determinants of experience and satisfaction across patient groups, medical conditions, and consultation settings. The null hypothesis, defined as the lack of difference with the original questionnaire structure, was ruled out using Item Response Theory (IRT) with exploratory factorial analysis to assess patients’ experiences and satisfaction.

A total of 2055 patients were included in the study between May 2016 and July 2016 (see Fig. 2 Flow diagram). The baseline characteristics of the patients and types of consultations appear in Table 1. The psychometric validation resulted in a final version comprising 16 items structured into three dimensions, depending on their content: Relief (eight items), Communication (six items), Compliance (two items), Table 2. This short form explained 54.5% of the total variance.

Patient-reported outcome evaluations have become mandatory [27, 28]. Most regulatory authorities in Organisation for Economic Co-operation and Development (OECD) have included such outcomes in their quality assessment framework. In the United States, the Primary Care Assessment Survey (PCAS) was initially developed to measure the quality of service in seven domains of primary care through 11 scales [29] and was further developed and tested to assess performance [30]. The Consumer Assessment of Healthcare Providers & Systems (CAHPS) is used in the hospital setting to discriminate between strategies in Medicare Accountable Care Organizations [8].

In Europe, the UK’s Quality of Outcomes Framework was introduced in 2004 [31]. The General Practice Assessment Questionnaire (GPAQ) was derived from the original PCAS [32‐34] and its psychometrical properties were retrospectively tested [35]. The GPAQ is used by the UK department of Public Health and Primary Care as a survey for general practitioners’ revalidations and practices [36]. These tools are oriented toward quality of service and performance assessments, but there is no generic tools forecasted to the evaluation of patient experience itself in various practice. However, patient satisfaction could be misleading if evaluated with tools that are validated to assess quality of service or inherited from the consumer field [5]. Taking into account patient expectations should help physicians with patient interactions and shared decision-making [37, 38].

In our study, the null hypothesis was rejected according to the differences that emerged between the original MISS-29 structure and that of the G-MISS short form reported here. The MISS-29 was specifically developed to assess patients’ experiences and expectations about physician consultations. The item generation relied on patient interviews to assess the cognitive, affective, and behavioral dimensions of patient-physician interactions [4]. The original set of items demonstrated good wording and comprehension compared with the Consultation Satisfaction Questionnaire [3]. The 29 item form was criticized for its acceptability, with a mean filling time of 12 min. This encouraged some authors to develop a short form of the MISS-29, for use only in general practice [17].

The purpose of the present study was to develop a questionnaire suitable for most consultation settings, by using IRT to select items and structure them into dimensions of experience.

One particular advantage of IRT over CTT is its independence in regard to the population being tested. The G-MISS final version is a self-reported questionnaire of 16 items structured into three dimensions of patients’ experiences and satisfaction with doctors’ consultations. One strength of the questionnaire is its large validation sample of 1822 patients, who were consulted in various general practices as well as medical and surgical specialties. The process of selection deleted items with equivocal loading into the factorial structure. There was no item switch from a dimension to another. The items belonging to the original “Rapport” and “Communication comfort” dimensions were merged into a single dimension, “Communication”, after the exploratory factorial analysis. The G-MISS questionnaire reported high levels of internal validity across its three dimensions, confirming that patients’ experiences are a multidimensional concept. Accordingly, the reliability indices of the three dimensions were satisfactory.

No notable differential item functioning was reported, ensuring that the questionnaire is equally reliable in every setting of consultation it was forecasted for, i.e. general practice, medical specialties and surgical specialties. The three-dimensional structure assesses different aspects of patients’ experiences. The G-MISS dimensions were named after the original MISS-29 questionnaire [13].

The “Relief” dimension assesses the alleviation of illness-related stress. It has been reported that patients experiencing symptoms of chronic or acute conditions worry about the potential impact of their disease on their life [39]. By relieving stress, physicians could improve health status and well-being [40]. The “Communication” dimension assesses the communication comfort between the patient and doctor.

It has been previously reported that patients consider communication to be one of the most important physician skills [41], but even if technically sound, the physician’s communication may appear inappropriate to the patient [42]. A recent systematic review stressed the lack of tools to assess physician’s communication [43]. The “Compliance” dimension reports the patient’s intent to follow doctor’s recommendations. It is well known that medical consultation influences patient compliance; [44] for example, specific attention should be given to verbal interactions, especially in long-term treatments like blood pressure control programs [45].

The items from the “Compliance” dimension were below the usual range of acceptance for the Infit in the general practice population [25]. One could hypothesize that there is an overlap between compliance and communication or relief, because of an indirect connection between these dimensions. A systematic review pointed out the link between physician-patient communication and outcomes in primary care [43].

The level of explanation given by the physician has been correlated with patient compliance [45] and could have influenced the “Compliance” and “Relief” or “Communication” dimensions assessed by the G-MISS scale. In contrast, unadapted behavior from a physician demonstrating nervousness or anger has been associated with less compliance from the patient and could also influence several dimensions of the patient’s experience [46]. The level of Infit was satisfactory in the global population, medical specialties, and surgical specialties, for all dimensions.

The level of satisfaction in the “Relief” dimension was significantly lower in general practice than medical or surgical specialties.

Some patients may not feel sufficiently reassured by their general practitioner’s explanations and need a referral to a specialist [39]. Specialists could be more likely to use situation-specific reassurance strategies while standard communication skills tend to stay generic [47]. These “Relief” dimension findings translated into the results for the global index with specialists’ interviews reporting higher satisfaction scores. Lack of practitioner’s time is classically reported as a cause of dissatisfaction for patients [48].

Although the perceived consultation duration has been associated with better patient satisfaction, the actual length of consultation has produced controversial results [49, 50]. One could hypothesize that the use of non-validated questionnaires may have explained the lack of difference [51]. In our study, interviews that lasted less than 5 min were significantly correlated with lower satisfaction scores in the “Relief” and “Communication” dimensions and global index. The G-MISS correlated with patients’ type of employment which is concordant with the original questionnaire. The external validity of the G-MISS scale was further emphasized by correlations with the emergency context, the educational level and the number of consultations in the last 6 months.

Among patients with a rate of missing values >20%, the number of responded items dramatically fell from item 15 to 29. This reinforces the interest for a short-form to improve the applicability of an everyday practice questionnaire. The final version of the G-MISS questionnaire is 16 items long and took about 6 min to complete. Despite this short form, the G-MISS questionnaire still explained 54.5% of the total variance.

The level of education wasn’t correlated with the rate of missing answers, suggesting that the questionnaire is widely administrable to patients able to read and complete an online survey. Application abilities in various medical fields of primary care underline the interest for the questionnaire in a patient-centered care approach.

The G-MISS questionnaire is a valid and reliable short-form questionnaire to assess patients’ experiences and satisfaction with physician consultations in general practices, medical specialties, and surgical specialties. The multidimensional structure relies on IRT and assesses different aspects of patients’ experiences that could be useful in clinical practice and research settings.