Adding information on multisite and widespread pain to the STarT back screening tool when identifying low back pain patients at risk of worse prognosis

Abbreviations

CWP

chronic widespread pain

EQ-5D

EuroQol-5 domain

FABQ

fear-avoidance beliefs questionnaire

HAD-A and HAD-D

hospital anxiety and depression scale

KSQ

Karolinska sleep questionnaire

LBP

low back pain

MS-CWP

multisite chronic widespread pain

PA

physical activity

RMDQ

Roland-Morris disability questionnaire

SBT

subgroups for targeted treatment back screening tool

1 Introduction

The point prevalence of low back pain (LBP) is around 13% in Europe [1]. LBP is most often transitory, although recurring episodes are common [2]. LBP may contribute to substantial consequences for the individual with activity limitations, participation restrictions, costs due to medical care, and productivity loss [3,4], and LBP is one of the most common causes of disability globally [5]. Most commonly, LBP is non-specific, when no known disease, pathology, or anatomical structure can explain the onset of pain [3], and <10% of the cases are serious, caused by a specific pathophysiological mechanism [6]. Approximately 20% of those suffering from LBP develop chronic LBP (pain present for at least 3 months during the past year) [2,7].

In primary care, history-taking and physical examination triage the patient to the most appropriate management. If the LBP is not serious, non-pharmacological therapies are recommended as the first option in both acute and persistent non-specific LBP [8]. These guidelines include advice to stay active, information about the good prognosis, and the nature of LBP symptoms. The management of persistent LBP includes pain relief using an exercise programme, cognitive behavioural therapy, or multidisciplinary therapy [8,9,10,11,12].

Primary care has a huge mission in the management of patients with LBP, especially for those with recurrent episodes as they have worse prognosis and are more costly [2,13]. Thus, there is a need for strategies early in the management to screen for patients with a high risk of worse prognosis. To better understand an individual’s risk, it is important to have knowledge of prognostic factors of worse prognosis from a biopsychosocial perspective. This includes old age, low general health, radiating pain, stress symptoms, depression, high pain intensity, heavy physical work demands, low physical function [14,15,16,17], previous sickness absence due to LBP, low education level [14], unemployment, low job satisfaction, low social support, kinesiophobia (fear-avoidance of movement), and catastrophising [15,16,18]. Further, having multisite pain, and to have another chronic pain condition are associated with worse prognosis of chronic LBP [18,19] and higher health care utilisation [20]. In the updated chronic widespread pain (CWP) criteria (WP2019), a total pain site score of ≥7 is associated with more severe clinical symptoms [21].

Different prognostic screening tools have been tested by clinicians to identify patients with LBP at risk of worse prognosis [22,23,24]. The STarT (Subgroups for Targeted Treatment) back screening tool (SBT), a validated easy-to-use tool for clinicians, has proven useful in identifying patients at risk. The SBT differentiates patients in low, medium, or high risk of worse prognosis, which can be used as support when planning an appropriate level of treatment [22,25,26]. Low risk implies management with education and general advice to stay active, medium risk physiotherapy, and high risk a combination of physical therapy and cognitive behavioural therapy [22].

It is important to consider risk factors of worse prognosis in patients seeking primary care identify those at high risk of developing chronic conditions and deciding the correct evidence-based treatment level [8,12]. The SBT captures important prognostic risk factors [22,25,26,27] but has shown some limitations in identifying patients at high risk of chronic conditions [24,28]. Adding information on pain distribution from a pain mannequin to refine the screening further has shown a potential to better distinguish between medium and high risk in a general population [29], but has not yet been tested among patients seeking care for LBP in a clinical setting. Using both SBT and information on multisite CWP (MS-CWP) may improve early detection of patients at high risk of worse prognosis, that would otherwise be missed by the use of SBT alone [25,29]. Adding information of MS-CWP may also increase the possibility of identifying those patients who are at risk of high healthcare consumption, including more visits to physiotherapists and sickness absence.

The aim of this work is to study the differences in prognostic factors in patients with LBP classified according to SBT or SBT in combination with MS-CWP, and the 1-year outcome regarding visits to physiotherapist and sickness absence, in relation to risk scorings.

2 Method

3 Results

In all, 95 patients (58 women [61%] and 37 men [39%]) were enrolled in the study. Out of these, 84 (88%) responded to all items in the SBT and on the pain mannequin, which was a condition that had to be fulfilled to enter the analysis. A majority (75%) had sought care for LBP previously (Table 1). When classifying the patients according to the SBT alone, 19 (23%) were considered as low risk, 48 (57%) as medium risk, and 17 (20%) as high risk. Within the high-risk group, 35% reported no chronic pain, 29% reported chronic regional pain, 12% reported non-multisite CWP, and 24% reported MS-CWP.

When classifying according to SBT, there were statistically significant differences in prognostic factors; anxiety, depression, health status, kinesiophobia PA and work, physical function, and sleep disturbances between the three groups (Table 2), with worse scores when classified to a higher risk group. There was no difference in age or sex between the three risk groups.

When using the combined screening tool, seven of the patients in the medium risk group (classified by SBT alone), who reported MS-CWP, were re-classified to the high-risk group of worse prognosis.

The three risk groups, as identified when using the combined screening tool, showed a similar pattern in their scorings of prognostic factors; sex, age, anxiety, depression, health status, kinesiophobia PA and work, physical function, and sleep disturbance as when using SBT alone for classification (Table 2).

4 Discussion

In this 1-year prospective study, we compared the easy-to-use SBT with a combined screening tool including information on MS-CWP in addition to the SBT. The combined screening tool classified more patients as high risk of worse prognosis, than if using SBT alone. The risk groups identified with the combined screening tool showed similar differences in health-related characteristics as when using the SBT alone regarding anxiety and depression, health status, kinesiophobia, physical function, and sleep disturbance. Results of both methods – SBT alone or SBT in combination with MS-CWP – indicate a higher proportion of patients with sickness absence in the medium risk groups. However, the differences when using the SBT alone did not quite reach statistical significance.

The risk group categorisation generated from the SBT has been suggested to be used for standardised triage and decision-making of further treatment. The SBT has been shown to be efficient in referring the patients to management with education and general advice to stay active (low risk); physiotherapy to address symptoms and physical function (medium risk); or a combination of physiotherapy and cognitive behavioural therapy (high risk) [25,42]. Using the SBT in primary care to triage patients with LBP has also been shown to be efficient in reducing symptoms related to physical function (assessed by RMDQ), as compared to treatment as usual in a control group [42], although contrasting results have been reported [43].

Screening with risk stratification is important to better utilise health care resources, including physiotherapy. Early identification of risk could be used as guidance when making decisions for treatment. Both the patient and society benefit from that resources are tailored to the needs. Baseline data on the prognostic health-related factors anxiety, depression, health status, kinesiophobia PA and work, physical function, and sleep disturbance were used to assess risk for worse prognosis in this study. Both when using SBT alone, and when using the combined screening tool, the scores on the health-related factors showed a pattern that followed the risk classification – patients classified as having low risk for worse prognosis scored better health, medium risk slightly worse health, and high risk scored worst health. However, this study indicates that the combined screening tool seems to classify more patients as high risk, and that the reclassified patients report worse health outcomes (anxiety, depression, health status, and sleep) than the patients who remained in the medium risk group. The high proportion of patients classified as medium risk (57% for SBT alone and 49% for the combined screening tool) is in line with previous studies that used the SBT in patients with LBP in primary care [25,42,44].

The relevance of adding information on pain distribution is in line with a previous study that found CWP as a predictor for high health care utilisation [20]. The current study used the ACR 1990 criteria for fibromyalgia according to which pain should be present in both the left and right side of the body, above and below the waist, and in the axial skeleton [35]. In 2020 a new definition of CWP was suggested, where painful sites should be present in at least four regions (regions are divided by [1] left and [2] right shoulder girdle, upper arm, lower arm, [3] left and (4) right hip, upper leg, and lower leg, and [5] axial skeleton), and that the total number of pain sites should be ≥7 [21]. This new definition is similar to the definition of MS-CWP used in the current study, aside from that the new definition (WP2019) includes pain present at four instead of three regions of the body [21,35].

In this study, only 20% of the LBP patients were categorized as high risk according to the SBT. A recent Danish study among LBP patients in primary care settings investigated the SBT as a predictor of physical function (defined by RMDQ), and identified 32% as high risk [45]. The Danish study included patients with LBP of any duration who had not sought care for LBP within the last 3 months, with or without radiating pain. The patients were older than the patients in our study, and scored worse on RMDQ [45]. In a recent Norwegian study, however, 0% of the LBP patients included in the study were classed as high risk, according to the SBT [44]. In a British study in a primary care setting, 28% of the patients were classified as high risk [42]. The diverse proportion of patients allocated to the high-risk group may partly be explained by differences related to the triage. In the current study, the patients were triaged by physiotherapists. Other studies have included patients triaged by general practitioners [42,46].

The results indicate that neither the SBT nor the combined screening tool captures patients at risk for frequent visits to the physiotherapist the following year. The results further indicate that neither the SBT nor the combined screening tool capture patients at risk for sickness absence the year following the initial visit to the primary care. The findings that a larger proportion among the medium risk group had been absent from work due to sickness (>14 consecutive days) during the year following baseline was unexpected. We expected a higher proportion in the high-risk group to have sickness absence. When interpreting these results, one should however consider a few aspects. Approximately three out of four patients included in this study had sought care for their LBP previously, and one out of three had sought care within the last 3 months. Further, regarding sickness absence, we included all registered sickness absences, regardless of diagnosis. The two most common causes for registered sickness absence are mental ill health and musculoskeletal disorders – two disorders that commonly coexist [17]. Furthermore, ongoing sickness absence among the patients at the time of inclusion in the study was not considered. Altogether these aspects may, at least partly, explain the unexpected differences in sickness absence between the risk groups.

In a previous study [29], we found that the combined screening tool had the potential to identify more patients at risk of worse prognosis than when using the SBT alone. However, in that study, patients with LBP were recruited from the general population, rather than patients seeking primary care for LBP [29]. This study adds knowledge on the value of using the combined tool in a healthcare setting, including patients seeking treatment for LBP. In addition, this study includes information on number of visits to a physiotherapist and sickness absence the year following inclusion and categorisation.

In a recently published review of studies examining primary care for LBP when using SBT, the contrasting trial results were highlighted. Some studies had found the approach to be effective, while others found no benefits. The diversity in trial results were discussed in relation to challenges in implementing stratified care (as recommended by the SBT) into clinical practice, and challenges in changing professional practice [43]. Lack of time has been suggested as one potential barrier for using the SBT as intended in primary care [43,47,48]. In the light of this, one need to consider the extra time and effort that the combined screening tool would imply, in relation to using SBT and future studies should investigate physiotherapists’ experiences of using the combined screening tool in primary care. Further, future studies should investigate if the classification (low-, medium- and high-risk) when using the combined tool, in comparison to the SBT, is as effective (regarding treatment outcome) when used for triage to different managements of the patients.

5 Strengths and limitations

A strength of this study is that it was conducted in a clinical setting under real-life circumstances. Another strength of this study is that patients were recruited from five different primary healthcare providers representing different socioeconomic profile areas. This strategy for recruitment aimed to capture patients that were representative of the general Swedish population. However, only 95 patients were recruited during the year, suggesting that a large proportion of the total number of patients seeking care for LBP were not invited to participate in the study even if they met the inclusion criteria. Unfortunately, no data on patients lost to recruitment due to the self-selection among the physiotherapists are available. Performing practice-based research in general practices is challenging and the risk for selection bias has been problematised in a Danish study [49]. We therefore had a limited sample of patients, and the statistical power to detect group differences was low. Furthermore, there is no information available on whether all visits to the physiotherapists targeted the LBP problem in question.

6 Conclusion

The combined screening tool including both the SBT and information on pain distribution (MS-CWP) classified more patients as high risk of worse prognosis than using SBT alone. The combined screening tool identified patients with worse health-related outcomes to the same degree as the SBT alone. The combined screening could be used to capture patients at risk of worse prognosis in a clinical setting. Although the combined screening tool may capture patients at high risk for worse prognosis, the effect when using the tool for triage to a matching treatment, as well as the value of adding a screening tool in the clinical examinations performed by physiotherapists, needs to be further investigated.