Differences in Swedish and Australian medical student attitudes and beliefs about chronic pain, its management, and the way it is taught

1.1 Aims

Considering that medical education is the basis of knowledge for physicians, it is important to explore the attitudes and beliefs of medical students with respect both to chronic pain management and to their views on current pain education [19], [20]. Hence this study compared Australian and Swedish medical students’ attitudes towards chronic pain. In particular, it asked:

1. What are medical students’ attitudes and beliefs about patients with chronic pain?

2. Are there any differences in Swedish medical students and Australian medical students’ attitudes and management towards chronic pain?

3. What are Swedish and Australian medical students’ perceptions regarding chronic pain and their education.

2.1 Medical education setting

Two universities participated in this study. The medical programme at the Swedish university extends over a 5½-year period consisting of 11 semesters, leading to the MD degree. Graduates then enter a 2-year general post graduate medical training programme “allmäntjänstgöring” (AT), to acquire a licence to practise medicine. Speciality training, “specialisttjänstgöring” (ST), can be initiated after this licence is obtained. At the Umeå university medical programme, clinical training including treating patients, starts in semester 6 (year 3; Swedish Universities use a two-term academic year). During semesters 9–11, topics such as psychiatry, paediatrics and research in the form of a master thesis are taught. As Chronic pain is not taught as a specific topic during these last three semesters at Umeå University, they were the ideal choice for our project. Basal pharmacology is taught as a discrete subject during 4 weeks in semester 5. In year 4, chronic pain education is given during semester eight (within the pain rehabilitation course). Two days are scheduled with clinical placements when the students participate in assessment of chronic pain patients together with an interdisciplinary team (including: physician, specialist in rehabilitation medicine, physiotherapist, occupational therapist, psychologist and social worker). The students also participate in case seminars about chronic pain and in group-based sessions about treatment methods and multimodal rehabilitation.

In Australia, the medical programme extends over 4–6 years depending on whether the student undertakes undergraduate training (5–6 years) or commences other studies before entering postgraduate training (4 years). The medical graduate then undergoes a 1-year internship followed by a residency (length determined by graduate) before undertaking specialist training after being accepted into a training programme. Depending on the medical school, students may be in the hospital from year 1 or from their penultimate year. At the School of Medicine at Notre Dame, Sydney, rotations start in year 3, then the students see patients alone and with supervising doctors. Experience specifically in a health care central setting is introduced during the 1st year during three afternoons with a GP, but is followed by a rotation of 5 weeks, also with a GP during the 3rd year. The remaining rotations occur in a hospital setting where one morning is scheduled at a hospital pain clinic during psychiatry rotation. Full responsibility for treating patients starts at the internship. Pharmacology is not taught as a discrete subject, instead pharmacology lectures are given throughout the courses. Similarly to Umeå University, problem based learning (PBL) using fictive patient cases, are also used.

2.2 Participant selection

The target group for the survey was medical students in their last three semesters at the school of medicine at Umeå University, Sweden, and final year (Year 4) medical students at the School of Medicine at Notre Dame, Sydney, Australia. The survey was available as an online questionnaire in SurveyMonkey^® (SurveyMonkey, San Mateo, CA, USA). Multiple distribution channels were used to maximise the participation rate. For the Swedish arm of the study, an invitation letter and web link to the survey was sent to 327 medical students by email in December 2016. Two reminders were sent out during the 2-month period the survey was open. For the Australian arm of the study, students were notified in a broadcast via their education software (blackboard) with information about the study and a link to the online survey was sent in February 2017. A flyer providing details about the project and survey with contact details was also displayed in venues used by the students. Australian students were also given the option to enter a draw for a gift voucher upon completion of the survey.

2.3 Questionnaire development

There is no gold standard tool to evaluate medical student attitudes and knowledge about chronic (non-malignant) pain management. In consequence, the survey instrument was developed using key themes from the literature [19], [21], [22]. The initial version was in English and this was translated into Swedish then back-translated to English to ensure that the two versions were internally consistent. The Swedish version consisted of 27 questions and the Australian version of 25 questions. The extra two questions in section four of the questionnaire (described below) in the Swedish version were not relevant to the situation in Australia.

This questionnaire had four sections. Section one assessed attitudes towards chronic pain using a validated instrument; the Health Care Providers’ Pain and Impairment Relationship Scale (HC-PAIRS) [23]. HC-PAIRS is a widely used instrument for assessment of attitudes of health care professionals, and it has been shown be a reliable and valid tool for measuring health care providers’ attitudes and beliefs about the relationship between pain and impairment [22]. Although HC-PAIRS is a quite commonly used instrument to explore both health care students and professionals, scores between different studies are not always directly comparable due to different application of the instrument. Some studies use a seven-point rather than a six-point Likert-scale and there is also variation in how many items are included, particularly with respect to the questions relating to projected cognition (see below).

The HC-PAIRS questionnaire originally consisted of 15 items (see Supplementary Table S1 for the questions as used here) rated on a Likert scale ranging from “totally disagree” to “totally agree”. In the original study, the items were classified into four groups on the basis of a principal components analysis. These items, reflecting different dimensions of attitudes were classified under the following headings: 1, “functional expectations” (items 1–3, 6–9, 11–12); 2, “social expectations” (items 5, 7, 11, 14); 3, “need for cure” (items 4, 9, 15); and 4, “projected cognition” (items 10 and 13). Thus, items 7, 9 and 11 are included in two of the dimensions, due to the method in which they were derived (orthogonal principal component analyses, see [23]). In the original study, Rainville et al. [23] validated the method by comparing the scores for 144 community health care providers with 66 functional restoration providers, who would be expected to have lower scores. This was the case for the factors 1, 2 and 3 and the total score, but not for factor 4 [23]. Modified versions of the instrument can be found throughout the literature, and some authors only use the items in the first three factors [24], [25], [26], [27], [28]. A Swedish version has been published by Overmeer et al. [28]. This version consists of 15 statements with a six-point Likert scale where 1=Strongly disagree, 2=Disagree, 3=Slightly disagree, 4=Slightly agree, 4=Agree, 5=Strongly agree, with a final score ranging from 15 to 90. This version was used, following minor modifications (a few word changes were made after pilot testing better to suit current guideline terminology. For example, the term “disabled” was changed to “a person with a disability”). Higher HC-PAIRS scores reflect a stronger belief that chronic pain is in close relation to impairment and limitation of activities, and as a consequence, the person answering the questionnaire has less beneficial attitudes towards chronic pain. Internal consistency has been evaluated by Houben et al. [22] (Chronbach’s α=0.78–0.84, a good outcome) and Briggs et al. [29] showed that HC-PAIRS scores were well in line with clinical recommendations.

Section two of the questionnaire was a Best-Worst Scaling Experiment to compare factors affecting chronic (non-malignant) pain management (to be reported separately).

Section 3 included three open-ended questions about chronic pain education; Students were asked what areas they thought were important for knowledge and skills regarding chronic pain treatment, if there was any part of their chronic pain education that could be changed to improve their skill and knowledge, and about their overall view on chronic pain management and education.

The fourth section consisted of seven (Australian version) or nine (Swedish version) demographic questions such as age, gender, what speciality they would choose if they were to choose today and in which sector they would prefer to work. It also asked if they knew someone (a friend or a family member) who suffered from chronic pain (and in that case, if they thought that their treatment was satisfactory or not) and if they themselves had ever experienced chronic pain. The Swedish survey included two additional questions asking what semester they were currently in and what type of health care facility they would like to work in the future. These questions were not included in the Australian version since we only investigated final year medical students in Australia and since the health care facilities are not directly comparable between countries. Responses to sections 1, 3 and 4 are reported in this paper.

2.4 Pilot testing

The Swedish survey was tested by a former medical student who recently graduated from Umeå University and by one final year medical student from another university in Sweden. The Australian survey was tested by two former University of Notre Dame school of medicine students undertaking their intern year. A feed-back questionnaire was sent out together with an online link to the survey. Pilot testers where asked to give critical feedback in relation to language congruity, survey design and medical plausibility. The Australian pilot testers were given a $20 gift card as incentive to participate. Minor changes were made as result of the pilot testing. Furthermore, specialist medical doctors with specialities strongly linked to chronic pain in both Sweden, UK and Australia were consulted to assure general medical authenticity of the survey.

2.5 Data analysis

3.1 Demographic characteristic

The total number of Swedish students who completed the survey was 80 (of 327 invited to participate, response rate 25%) and the number of Australian students that completed the survey was 30 (of 110 reached, response rate 27%). Figure 1 shows the number of participants from each country that completed each individual section of the survey. There was a considerable attrition rate, particularly at the level of Section 2, with the Swedish students showing a higher attrition rate at this (and subsequent levels) than the Australian students (p<0.01, Fisher’s exact test).

Figure 1:

Number of Australian medical students and Swedish medical students completing each section of the survey.

The characteristics of the respondents are described in Table 1. With one exception (future workplace), there was no statistical significant difference between the two samples. In both cohorts, the majority of the respondents were female and aged between 25 and 29 years. Approximately 70% of the students knew someone who suffered from chronic pain and most of these pain sufferers believed that their treatment was inadequate. Approximately one in six of the responders had experienced chronic pain themselves.

The public sector was the most popular choice for future workplace amongst the Swedish students while 15/21 Australian students chose both private and public sector. The Swedish students were asked in which type of health care facility they would prefer to work, with teaching hospital being the most popular choice (43%) followed by regional hospital (23%). Only a small proportion of students in both countries (~10%) expressed the desire to pursue general medicine as their medical speciality.

3.2 Attitudes and beliefs towards chronic (non-malignant) pain patients

Attitudes and beliefs were assessed using HC-PAIRS scores. Median scores on the six point Likert scale for the 15 individual items are reported in Supplementary Table S1. Analysis of Likert scores is a somewhat contentious issue, but there are good arguments for analysing the data parametrically given that the score consists of a sufficient number of items and that the internal consistency of the scores is good [33]. This is certainly the case for the total scores for all items (15 items) and for those included in factors 1–3 (13 items) as well as for factor 1 (nine items), and possibly factor 2 (four items), but whether it holds for factors 3 (three items) and 4 (two items) is more open to debate. In consequence, we have presented the data for the scores both as means±SD in Table 2, and as medians and interquartile ranges in Supplementary Table S2 to allow the reader to make her/his own decision, and analysed these accordingly using a 5% false discovery rate to assign critical values of p [31]. Scatterplots showing the individual scores are shown in Supplementary Fig. S1.

The Australian medical students show statistically significantly higher total HC-PAIRS scores than the Swedish medical students, see Table 2. Australian students had statistically significantly higher scores than the Swedish students for two of four factors: functional expectations and need for cure. The effect sizes (Cohen’s d) were 0.80, 0.86 and 0.77 for total score, factor 1 and factor 3, respectively. The differences between the groups for factor 2 (social expectations) and factor 4 (projected cognition) were not significant.

For the Swedish students, we investigated whether the mean scores were significantly different for the individuals completing vs. not completing the whole questionnaire (shown as filled and unfilled symbols, respectively, in Supplementary Fig. 1). For the total score and factors 1, 3 and 4, there were no significant differences between the two groups (p>0.36, two-tailed t-test with Welch’s correction). However, the score for factor 2 was higher for the completers than the non-completers (15.3±3.4, n=30 and 13.0±2.8, n=27, respectively, means±SD; p=0.0098, two-tailed t-test with Welch’s correction).

The demographic data for students completing the survey was well balanced between countries with respect to gender and whether they knew someone who suffered from chronic pain (Table 1). In order to determine whether or not these factors were associated with differences in the scores, linear models were constructed for all 51 students completing the surveys. For factor 1, the p-values for the main effects of land, gender and knowing someone suffering from chronic pain were 0.027, 0.18 and 0.30, respectively (Supplementary Table S2). A full interaction model gave no significant interactions between the parameters (p>0.5). Similar results were seen for factor 3 (Supplementary Table S2). Thus, gender and knowing someone suffering from chronic pain can be ruled out as factors affecting the observed scores of the students completing the survey.

3.3 Student assessment of their chronic pain education

4.1 HC-PAIRS

The considerable difference in total HC-PAIRS score between our cohorts may be due to the fact that Swedish medical students gain experience in treating patients earlier than the Australian students, resulting in the Australian students having less experience with patients. But it could also be due to possible differences in other education aspects of the medical programme, national or regional clinical guidelines and cultural differences.

To allow us to compare the score results from other studies, we have normalised the final scores to a percentage of the attainable maximum score (Table 4). The normalised mean scores range from 41 to 67%, with our scores of 51 and 46% (Australian and Swedish, respectively) being in the middle of the range. Interestingly, the normalised mean score for the Australian students is higher than the British medical students, who have similar education system and health care system. As a caveat, it should be pointed out that scores will be dependent upon the items selected for the study, since some use all 15 items, whereas others do not. It would be beneficial to align the implementation of the HC-PAIRS instrument in order to achieve comparable results between future studies.

Investigating different items, we found the most negative scores for item 10 “When their pain gets worse, chronic pain patients find it very hard to concentrate on anything else” and for item 13 “Chronic pain patients find themselves frequently thinking about their pain and what it has done to their life”. Swedish and Australian students believe that for chronic pain patients, when the pain gets worse, they find it very hard to concentrate on anything else; and that the patients find themselves frequently thinking about their pain and what it has done to their life. These views are potentially stigmatising: it appears students expect that chronic pain patients in general dwell on their pain, which is a highly negative generalised view that may have a negative impact on chronic pain treatment. The association between negative attitudes and beliefs of health care professional has repeatedly been linked with more negative patient treatment outcomes [37], [38]. It has also been shown that attitudes can be changed through education. Thus, Jacobs et al. demonstrated an improvement in HC-PAIRS scores of physiotherapists following a day course raising awareness of the biopsychosocial model for chronic pain [25]. Morris et al. [34] found a progression in positive attitudes towards chronic pain patients between 1st and final year medical students. It is thus clear how important education and knowledge is to improve health care students’ attitudes towards chronic pain.

4.2 Chronic pain education

It is reassuring that both Australian and Swedish final year medical students have a clear notion of current recommendations of chronic pain management. The top five subjects within chronic pain that need the most attention according to AAPM [18] can be recognised in our thematic analysis for both cohorts through themes such as “Patient attentiveness” and “Pain assessment”. Moreover, key chronic pain concepts such as the biopsychosocial model and pain management in multidisciplinary teams were also repeatedly acknowledged by our students within the theme “Biopsychosocial model”. Observed differences between cohorts could similarly be explained due to differences in the medical programmes. The Swedish medical programme is longer, currently comprising 11 semesters (5.5 years of studies) before entering the post graduate training, whilst the Australian programme ranges over 4 years before post graduate training. It appears that Swedish students obtain more experience seeing patients during this time which could be the reason for their willingness to have even more experience with chronic patients compared to Australian students who more frequently mentioned wanting more clinical vignette-based learning. Furthermore, Australian students seem to be assigned to rotations in a way that will result in differences in gained chronic pain education between individual students.

Regarding pharmacology training, there was a tendency for Australian students asking for more basal pharmacology training compared to the Swedish students who requested more clinical pharmacology in terms of treatment plans. Students at Umeå university obtain a 5-week course in pharmacology in the fifth semester of their studies whilst students at Notre Dame receive their pharmacology lectures throughout the programme, and this might have an impact on their respective needs. The worldwide issue with increased opioid prescribing in chronic pain management was repeatedly addressed by the Australian students but not the Swedish, this might be due to the current large focus on this subject in Australia exemplified here by a review from the National drug and Alcohol Research Centre [12]. Other reasons for differences in open-end answers can naturally be differences in expression between different languages and cultural differences.

Both cohorts described that chronic pain education and management is complex but a very important topic. The majority of students stated a general need for more knowledge on the subject including many suggestions for improvements such as increased training with other health care practitioners, more focus on the biopsychosocial model, practical experience with patients, pharmacology and treatment plans, more exposure to pain clinics and increased knowledge about non-pharmacological treatment. We consider our findings to be in line with those of Briggs et al. [17] who concluded that chronic pain training is scattered and unsatisfactory.

4.3 What this study contributes

It is important to know more about health care professionals and medical students’ attitudes towards chronic pain. Attitudes and beliefs are formed during medical education, and our study exploring attitudes of medical students towards chronic pain and how it is taught have provided valuable information. Our survey provided detailed and cohesive suggestions for education improvement that also are in line with current clinical guidelines. This study highlights several areas of interest that warrant further investigation, for example, the impact of a changed medical curriculum in alignment with these clinical guidelines requested by students in this survey, and correspondingly if their attitudes towards chronic pain patients can be improved through education. Further, it would be valuable to align the implementation of the HC-PAIRS instrument in order to achieve comparable results between future studies and we consider this an important subject to be addressed by future research.

4.4 Strengths and limitations

Recruiting students was challenging in both countries and a higher response rate would have been favourable. Further there was a considerable attrition rate throughout, and these factors should be taken into consideration with respect to the interpretation of the data. Nonetheless, our results are in line with the results of other studies such as those on medical education across Europe [17], and our HC-PAIRS are within the range reported in other studies (see Table 4). We thus add HC-PAIRS scores from two key cohorts in near future chronic pain management, from two different countries to the literature, and the need for more chronic pain education amongst both student cohorts is evident through our thematic analysis.

There were more females than males who completed the study, and many of them knew someone suffering to chronic pain, thus suggesting there might be risk of selection bias. However, our linear model analysis would suggest that these factors did not significantly contribute to the outcome of the HC-PAIRS scores (see Supplementary Table 3).