Diagnostic criteria for musculoskeletal disorders for use in occupational healthcare or research: a scoping review of consensus- and synthesised-based case definitions

A PRISMA flow diagram of the study selection process is shown in Fig. 1. In total, 2404 references were retrieved from the two databases and assessed on title and abstract. The full texts of 104 potentially eligible articles were then examined, of which 39 studies met the inclusion criteria, of which 2 were included from experts. Reasons for excluding studies based on full text are presented in Additional file 2.

In total, 39 articles [7, 13, 21‐57] described diagnostic criteria for MSD case definitions for non-specific LBP (n = 15, 38%) [25, 27, 29, 30, 37‐41, 44, 47‐49, 52, 54], LRS (n = 6, 15%) [26, 32, 37, 40, 47, 56], SAPS (n = 5, 13%) [7, 13, 21, 31, 53], CTS (n = 8, 21%) [7, 23, 24, 33, 34, 36, 50], lateral and medial elbow tendinopathy (n = 1, 3%) [7], and knee (n = 8, 21%) [22, 35, 42, 43, 45, 46, 55, 57] and hip OA (n = 4,10%) [42, 45, 51, 55]. The charting of these articles is presented in Additional file 3.

Most studies were conducted in Europe (n = 21, 54%) [7, 13, 21, 23, 25, 26, 32, 33, 37‐41, 45, 47, 50‐52, 55‐57], reported a synthesis of the literature on case definitions (n = 15, 38%) [7, 21, 22, 24, 27, 28, 36, 39, 41, 46, 48‐51, 57], and/or had a clinical aim (n = 32, 82%) [13, 21‐27, 31‐41, 43‐50, 52, 53, 55‐57], while almost half of the studies reported three or more disciplines that were involved in reaching expert consensus regarding the case definition (n = 19, 49%) [13, 21, 24‐26, 29, 31, 32, 37, 41‐43, 45‐48, 52, 55, 56] (Additional file 3).

Regarding non-specific LBP (Table 1, Additional file 3), there was general agreement between the studies included that diagnostic criteria for non-specific LBP should consist of pain in the low back without leg pain and absence of specific pathology of the LBP [25, 27, 30, 37‐41, 47‐49, 52]. Some of the studies referred to a specific duration of the symptoms, which varied from < 1 month to < 12 weeks [30, 37, 38, 40, 47]. Most studies also agreed that imaging should not be recommended [25, 27, 38‐40, 49, 52]. The same was true for LRS with the specification that leg pain should be reported [26, 29, 32, 39, 40, 47, 54, 56]. Additionally, the same was found for chronic LBP, with the duration of pain in the low back referring to chronicity varying from > 6 weeks to > 12 weeks [37‐41, 44, 47, 49].

The five studies included on SAPS reported that pain in the shoulder in combination with results from tests are needed to identify a patient with SAPS (Table 1, Additional file 3) [7, 13, 21, 31] and that retraction of tendon(s) to the glenoid rim indicate a massive rotator cuff tear [53]. The articles varied on which tests were needed (e.g., painful arc test, Neer’s sign test, Hawkins-Kennedy test, empty can, lift-off and drop arm test) and varied on the use of imaging. For example, some studies were of the opinion that X-ray was needed to assess degenerative changes, while others were of the opinion that ultrasound was needed to exclude a rotator cuff rupture.

More variation was found between the studies included on CTS (Table 1, Additional file 3) [7, 23, 24, 28, 33, 34, 36, 50]. Signs varied from none to numbness, pain, tingling, paraesthesia, brachialgia, and nocturnal numbness/paraesthesia in the fingers or hand and/or a combination of these. Determination of symptoms varied from conducting provocation tests such as Tinel’s sign and Phalen’s test [7, 34, 36], to not conducting these tests [33], or the assessment of nerve conduction speed, nerve cross-sectional area, and/or latency. Recommended imaging included electrodiagnostic and ultrasound to determine the cross-sectional area, while MRI was not recommended [23, 24, 36, 50].

The only article included on lateral and medial elbow tendinopathy showed that the symptom pain should be directly located around the lateral or medial epicondyle [7]. The pain should be intermittent and activity dependent. Furthermore, the pain should be present at assessment or on at least 4 days during the last 7 days. The signs of lateral elbow tendinopathy were reported as local pain on resisted wrist extension and the signs of medial elbow tendinopathy were reported as resisted wrist flexion (Table 1, Additional file 3).

All of the studies included on hip OA reported hip pain with three out of four of the studies reporting hip pain combined with joint space narrowing. All recommended X-ray to assess degeneration using the Kellgren-Lawrence scale (Table 1, Additional file 3) [42, 45, 51, 55]. More variation was found on knee OA. Symptoms ranged from knee pain to knee pain combined with stiffness, reduced function, swelling, cracking and/or grinding movement [22, 35, 42, 43, 45, 46, 55, 57]. Signs ranged from none to an extensive list of alterations in the knee, such as bony enlargement. Five of the eight studies on knee OA agreed on using X-ray to assess degeneration using the Kellgren-Lawrence scale [22, 35, 42, 43, 45, 46].

In one study [44] discussing chronic LBP the following working exposure criteria were reported: long-term spinal heavy burden, excessive rotation, or vibration.

In another study [7] discussing case definitions for SAPS, CTS and lateral and medial elbow tendinopathy, work-related exposure criteria were included in the clinical assessment of the case definition. Work-related exposure criteria for SAPS included: 1) postures: i) with hands behind the back, ii) where hand reaches to opposite part of the trunk, iii) with extreme rotation, iv) where the arm is unsupported by the body for several minutes; 2) elevation movements of the upper arm in comparison with those of the trunk; 3) high repetitions of movements of the upper extremity; and 4) a combination of more than average force and one of the movements or postures mentioned.

For CTS, work-related exposure criteria included: extreme wrist postures, handling (vibrating) tools, high repetition of wrist movements, high forces for the hands, and combination of postures/movements and forces [7].

Finally, for lateral and medial elbow tendinopathy work-related exposure criteria included: extreme flexion of the elbow, posture with extended elbows, posture with extreme pronation or supination of the elbow, high repetition of movements of the elbow, grasping or lifting of objects with high forces and/or combination of postures, movements and forces [7].

While this scoping review identified various clinical and research settings in which consensus has been reached or the literature has been synthesised regarding case definitions, some knowledge gaps were found. Case definitions of diagnostic criteria for MSDs may differ depending on the setting and purpose. A setting may be a clinical or an occupational epidemiological research context, while the purpose may consist of prevention (including screening activities), prognosis or treatment. When considering a clinical setting we agree with Genevay et al. that case definitions for clinical practice require high sensitivity and high specificity, while case definitions for occupational epidemiological research mainly require higher specificity. High specificity is required, given that the inclusion of false positive cases need to be avoided in occupational epidemiological research [32] as it may influence the identification of personal- and work-related risk factors. However, prevention-related studies among workers may require just higher sensitivity for precautious reasons.

Case definitions for research and for workers’ health surveillance also need to consider practicalities such as costs, the burden of completing questionnaires, and the availability of resources. This need for balance resulted in some studies providing a minimal and an optimal case definition depending on its research purpose [30]. For example, Dionne et al. reported a minimal case definition consisting of two questions, one on back pain and one on severity, and an optimal case definition that was based on the minimal definition with additional questions covering frequency, duration of symptoms and severity, as well as a question covering sciatica and a question excluding other causes [30].

Case definitions aiming to assess the work-relatedness of a disorder/disease in association with workers’ health surveillance [58] or financial compensation may need higher sensitivity, given the financial costs involved. In both situations, a symptom questionnaire with high disorder/disease sensitivity and a follow-up medical examination may be the best option [58].

Considering the purpose of the case definition, those aimed at prevention in clinical care or research could encompass greater heterogeneity since sensitivity is less important given that this heterogeneity has a limited effect on identifying personal and work-related risk factors [59, 60]. For treatment at an individual level, a higher sensitivity is needed, especially in the case of high costs, detrimental side effects or the limited availability of resources. This is reflected, for example in the fact that case definitions used in a clinical care more often included imaging (Additional file 3).

Work-related exposure criteria in the clinical assessment of a case definition require more attention in future research [61, 62]. We only found two studies that reached consensus for work-related MSDs diagnostic case definitions [7, 44], although recently there are more studies available that address work-related risk factors for specific MSDs (e.g., [2, 11, 28, 63‐67]). The International Classification of Diseases (ICD-11) of the World Health Organisation (WHO) [68] includes expert-based criteria for work-related diseases [68, 69]. These expert-based criteria, in combination with the data charted in this scoping review, could serve as a basis for a Delphi study aiming for the harmonisation of these case definitions, focussing on research aimed at prevention of work-related disorders/diseases.

In comparison with previous research on expert consensus methods in other fields such as medical education, we found that consensus methods were not standardised and not transparently reported; for example, no detailed data were provided about the participating experts [70]. Furthermore, we also found those studies which reported on a consensus method, most often did not provide a definition of consensus a priori, in terms of envisioned content and format of the intended outcome [70, 71]. Finally, previous research found that rating scales influence the outcome of these consensus methods [72] but less is known about which factors, such as tacit professional collective knowledge, affect the outcome of these consensus methods among experts. To that end, more research is needed on consensus methods themselves as well as criteria to improve the reporting of these consensus methods [71]. When that is achieved, future reviews on this topic might be able to only include high quality studies.

One strength of this review is that it encompassed a variety in reported consensus or synthesis approaches providing an overview of case definitions for the MSDs included, as reported in peer-reviewed medical journals. This appeared especially important for our research question, as we found that heterogeneous study designs were used to reach consensus or to synthesise the literature on case definitions. Another strength was the international research team of whom most have been involved in these case definition developments. Four limitations should be noted. First, we may have overlooked some relevant studies, since our search strategy was limited to one database and the Web of Science. Second, we may have overlooked some relevant studies since in some studies it was arbitrary whether or not the authors applied a consensus- or synthesized-based method. Third, in a scoping review, quality assessments are not performed, because the literature is charted without critical appraisal of the studies included [10]. However, additional file 3 shows that the studies included differed in the amount of information they provided regarding their research methods. In addition, it was not always made clear whether consensus was based on the consultation of experts from various disciplines, although the involvement of multiple disciplines was considered an essential element in reaching consensus regarding case definition [73]. Finally, when reviewing the studies included there was a large variety in consensus methods used and how well these methods were described [71].