Translation and validation of the Finnish version of the Fear-Avoidance Beliefs Questionnaire (FABQ)

1 Introduction

Low back pain (LBP) is a common problem worldwide, with a point prevalence of up to 33% [1]. It is the number one debilitating condition affecting daily function, and can diminish the quality of life, contribute to an increased likelihood of work disability, and increase health care costs [2]. Psychosocial risk factors are related to slower recovery from acute nonspecific LBP [3,4,5, 6,7]. These include a number of different conditions, such as depression or anxiety, social environment and work characteristics. Fear-avoidance behaviour is a psychosocial risk factor in the form of fear of pain due to physical activity. Some patients may associate their pain with certain situations and movements, leading to inactivation [8]. This may lead not only to prolonged sick leaves and early disability pensions, but can also postpone recovery from LBP [7].

The Fear Avoidance Beliefs Questionnaire (FABQ) was developed on the basis of the assumption that fear-avoidance beliefs play a major role in disability, especially among people with LBP [9]. To date, FABQ validation studies have been made in numerous languages [10,11,12,13,14,15]. The FABQ has not so far been validated in Finnish. Our objective was to translate and cross-culturally adapt the FABQ and to study its psychometric properties among Finnish patients with chronic LBP.

2 Material and methods

3 Results

The Finnish version of the FABQ was adapted using a process of forward-backward translation. Some semantic issues in the translation process were debated, mainly concerning cultural adaptations related to the different accident insurance systems of the countries (items 6 and 8) and word choices (items 1–5), to guarantee that the meanings of the original items were adequately captured in Finnish. The questionnaire was further pre-tested with five patients with back problems. The questionnaire proved highly acceptable, easy to understand, and capable of being self-administered. No suggestions for improving the wording were given.

Altogether 66 patients (92% of the participants) returned a complete baseline FABQ questionnaire and 62 patients (86%) returned both a complete baseline and follow-up FABQ questionnaire. The response rates for the FABQ items varied from 96% to 100% at baseline and from 92% to 96% at follow-up. The time between the first and second questionnaire varied from 2 days to 59 days, with a median follow-up of 16 days (IQR 12–21).

Mean age was 46 (SD 13) years and the majority of the patients were females (58%). Half of the patients (50%) were working or studying, 27% were on sick leave, and 9% had retired. Almost all patients reported pain in the lower back (97%), whereas pain in the legs, neck, shoulders, and upper back was reported less frequently. Most patients had chronic pain lasting over 6 months (61%) or 3–6 months (15%) whereas 11 patients (17%) had had LBP (current pain episode) for less than 3 months. The mean intensity of back pain was 58 mm (SD 23) on VAS. The mean ODI score among patients was 29 (SD 14). The detailed demographic and clinical information are shown in Table 1.

During the time between the baseline and follow-up questionnaire, only one (1.5%) patient reported incident pain in the left ankle. Two (3.0%) of the patients had experienced an accident during follow-up and reported these as follows: “Foot slipped, resulting in bruising of the right leg”, and “fell down an icy stairway but needed no medical attention”. Three (4.5%) of the patients had surgery during the follow-up (haemorroid; canalis carpi; knee prosthesis). Fifty-eight (87.9%) of the patients had no treatment during follow-up, whereas 10 (15.2%) reported some kind of treatment during follow-up. The most common treatments were massage (N = 6), glucocorticoid-injections (N = 2) and physiotherapy (N = 2).

The mean values of all 16 FABQ items are shown in Table 2. The highest mean scores were found in items related to physical activity: “Physical activity makes my pain worse” (4.5 at baseline and 4.6 at follow-up), “I should not do physical activities which make my pain worse” (4.4,4.4), and “My pain is caused by physical activity” (4.2, 4.1). The item “My work aggravated my pain” had high scores (4.2, 3.9).

The item “I should not do physical activities which (might) make my pain worse” had a strong ceiling effect both at baseline (34.9%) and follow-up (39.5%). At baseline, a strong ceiling effect was also observed in the items “Physical activity makes my pain worse” (41.9%), “My work aggravated my pain” (32.6%), and “I do not think that I will be back to my normal work within 3 months” (30.2%). A strong floor effect was observed for item 8 “I have a claim for compensation for my pain” (at baseline 83.7%, at follow-up 79.1%).

The ICC values for the items related to physical activity ranged from 0.43 to 0.66, displaying moderate to good reliability between the baseline and follow-up questionnaires. The items regarding work showed high or excellent reliability throughout, with ICC values ranging from 0.63 to 0.89 (Table 2).

The mean values of the total score and two sub-scores are shown in Table 3. The internal consistency was excellent for the Total score (α = 0.92 and 0.93) and Sub-score 1 (work; α = 0.89 and 0.94), and good for Sub-score 2 (physical activity; α = 0.71). Reliability between the baseline and follow-up scores was also excellent for the Total score and Sub-score 1, and good for Sub-score 2. Item analysis of the FABQ showed that all items had a high corrected item correlation except item 8 (Fig. 1).

Fig. 1

Item analysis.

The first factor analysis revealed a four-factor solution with low loadings for items “I cannot do physical activities which might make my pain worse” and “I have a claim for compensation for my pain”. After removing these two items the final factor analysis revealed a three-factor solution with good internal consistency (Table 4, Fig. 1 ). Factors 1 and 3 were characterized by the fear of pain being aggravated by working or physical activity, and named “Work” and “Physical activity”. Factor 2 included items with the fear-avoidance beliefs concerning work, and was named “Beliefs related to work”. The mean scores of the items loaded into different factors were 3.8 (95% CI 3.4–4.2) for Work, 2.8 (2.3–3.3) for Beliefs related to work and 4.2 (3.9–4.5) for Physical activity.

The total FABQ score associated significantly with intensity of LBP (Table 5). When adjusted for age and sex, the strength of associations between Sub-scores 1 and 2 and LBP increased slightly. All the FABQscores associated significantly with ODI, both unadjusted and adjusted.

The two diagnostic groups (radicular vs. non-radicular LBP) did not differ in FABQ scores, LBP intensity, or disability (Group 1 vs. Group 2: Total score 52.6 vs. 51.8, P= 0.898; Sub-score 1 24.1 vs. 22.7, P = 0.664; Sub-score 2 15.3 vs. 16.8, P = 0.212; LBP intensity 59.5 vs. 54.9, P = 0.485, ODI 30.7 vs. 27.9, P = 0.420).

4 Discussion

The results of the present study on the cross-cultural translation and adaptation of the Finnish version of the FABQ indicate that the expert committee successfully created an applicable questionnaire.

Our results are well in line with other validation studies of the FABQ. Our factor loading was parallel to the German validation study [10]. Internal consistency was excellent for Total score and Sub-score 1 (work), and good for Sub-score 2 (physical activity). The ICC values for work showed a moderate to good correlation, whereas items regarding work showed high or excellent reliability parallel to the original version [9]. Item 8 did not load into any factor, as visualized in item analysis in Fig. 1, which has also been noted in other studies [9,10,11,12,13].

Numerous studies aiming to culturally validate the FABQ have implied that the psychometric properties and predictive value of FABQ is either good or similar to the original version among people with sub-acute and chronic LBP [10,11,12,13,14,15]. In Turkish and Swiss-German FABQvalidation studies two factors were identified parallel to the original version by Waddell et al. [9,11,12]. Pfingsten et al. validated the German version of the FABQ and stated that it is a valid instrument for patients with LBP [10]. They also noticed different factor structure. In their study the Sub-scale ‘work’ was divided into two therefore introducing three Subscores: ‘Work as a cause’; ‘Prognosis work’ and; ‘Physical activity’. Three factors were also identified by our study. In final factor analysis our factor loading differed only in the case of item 9 “My work is too heavy for me” and item 5 “I cannot do physical activities which (might) make my pain worse”. In the analysis by the Germans item 9 was loaded in ‘Work’ whereas in our study it was included in ‘Beliefs related to work’. Our results revealed low loading for item 5 and it was therefore removed from ‘Physical activity’. As mentioned the ‘Work’ subscale introduced by Waddell et al. was divided into two in our study [9]. These represent two variant beliefs: (1) ‘Work’ represent subjects’ beliefs about their ability to work and possible aggravation of pain due to work environment and; (2) ‘Beliefs related to work’ represents expectations of subjects’ ability to return to their work and pain as a possible defining agent of their working ability.

According to systematic reviews, fear-avoidance beliefs are not related to the evolution of LBP, but predict persistent disability [3,4,7] including work disability [5,20]. Other known prognostic factors related to slower recovery and to the development of disabling LBP, in addition to abnormal fear-avoidance beliefs, include radiating pain, high pain intensity or disability at baseline, and psychiatric comorbidities such as depression [21].

According to our results, item 8 (“I have a claim for compensation for my pain”) had a strongly skewed distribution with a large amount of zero-values and it correlated very weakly with other items. Similar observation has been documented in several European countries [9,10,11,12,13]. Therefore, it can be stated that item 8 has different properties compared to the other items, which is potentially related to the differences in the design of compensation contracts across countries, and we suggest that item 8 should be considered to be excluded from the future FABQ analyses. In addition, our factor structure accords with those of the German [10] and Italian [15] validation studies and suggests that subscale ‘Work’ should be optimally divided into two work-related subscales describing different aspects of thoughts about work and work place. In the view of these results, the factor structure of FABQ needs a thorough re-evaluation. This could best be conducted using a large patient sample from European countries with an available translation of the FABQ.

Strength of our study was that we controlled possible changes in health status between the baseline and follow-up examinations. Patients reported changes in their health status such as changes in symptoms, accidents, surgery, and treatments received during the follow-up period. To avoid bias these were adjusted for in the reliability analysis. The variation of follow-up time may be considered a limitation of this study, which was also noted in the reliability analysis. Because of the incompletely filled questionnaires, the number of patients included was also somewhat small.

In conclusion, the Finnish FABQis a valid and reliable instrument for assessing LBP patients’ fear avoidance behaviour and has similar properties to those found in other validation studies of the FABQ. Thus it can be used for assessing the risk of disability due to fear avoidance behaviour in Finnish-speaking patients with LBP for both clinical and scientific purposes.

Highlights

• The Fear-Avoidance Beliefs Questionnaire (FABQ) measures fear avoidance due to pain.

• The FABQ has not been validated in Finnish.

• This study proposed the Finnish translation of the FABQ as reliable and acceptable for use.