The World Health Organization (Luttman, Griefahn & Caffier 2003) defines a work-related musculoskeletal disorder (WRMSD) as a form of ill health ranging from light transitory disorders to irreversible disabling injuries that are induced or aggravated by work and related circumstances. Despite physiotherapists’ specialist knowledge on injury prevention, they are also prone to WRMSD. Several international studies (Bork et al. 1996; Holder et al. 1999) reported a 32% lifetime prevalence of WRMSD in physiotherapists, whilst Cromie, Robertson and Best (2000) and West and Gardner (2001) reported a higher lifetime prevalence of WRMSD of 91% and 55% respectively. African studies (Obembe et al. 2008; Useh, Igumbor & Madzivire 2001) reported a lifetime prevalence of 78% whilst another study (Adegoke, Akodu & Oyeyemi 2008) reported an annual prevalence of 91%.

A hands-on approach of manual therapy is predominantly used in orthopaedic outpatient departments or private practices treating musculoskeletal disorders. Increased use of their hands puts these physiotherapists at risk of developing WRMSD of the wrists and hands (Barnes et al. 2011; Cromie et al. 2000; McMahon, Stiller & Trott 2006; Snodgrass & Rivett 2002; West & Gardner 2001). The lifetime prevalence of work-related thumb problems (WRTP) ranged from 62.5% to 83% in studies focusing on wrist and thumb disorders (Barnes et al. 2011; McMahon et al. 2006; Wajon & Ada 2003).

Barnes et al. (2011) investigated the prevalence of work-related wrist and thumb pain and the contributing risk factors amongst physiotherapists working in Bloemfontein, South Africa. The current study reported on WRTP only and included a national sample of physiotherapists.

Of the 3523 physiotherapists who received the link, 456 responded to the invitation to participate in the study. Of these, 61 were excluded from the study on the grounds of missing data (27), structural deformities (21), diseases of the hand (8) and operations to the forearm and hand as a result of non-work-related factors (5), resulting in a final sample size of 395. This represented 13% (n = 45) more respondents than the calculated (N = 350) sample size.

The demographics of the sample are presented in Table 1. The majority of the respondents who responded to the questionnaire on gender (60%, n = 243) were female (93%, n = 226). The lifetime prevalence of WRTP in the current study was 65.3% (n = 258). The greatest majority respondents with WRTP was found in the age group 20–30 years (70%, n = 84).

Most respondents were employed full time at the time of the survey (89.1%, n = 352). The majority of respondents worked in work areas using manipulative therapy (Figure 1). The greatest prevalence of WRTP was found in physiotherapists working with orthopaedic outpatients and in work areas using manipulative therapy. Preventative education was only received by 47.1% (n = 186) of the respondents. The majority of respondents received education on thumb protection at undergraduate level (28.1%, n = 111).

Table 2 represents the self-reported generalised hypermobility of respondents. Although a minority of physiotherapists reported generalised hypermobility for four out of the five parameters (< 15.4%, n < 61), the majority of the respondents (> 44%, n = 174) reported hypermobility of the joints of the thumbs between 0° and 30°, as shown in Table 3. The highest prevalence of WRTP was found in respondents with interphalangeal (IP) joint hyperextension greater than 30° (dominant: 75.6%, n = 62; non-dominant: 78.5%, n = 62), as shown in Table 4.

The factors that were significantly associated with WRTP in all 395 respondents were put in a regression analysis (Table 5). Only hyperextension > 30° of the non-dominant IP joint of the thumb and cervical spine treatment of up to six patients a day remained significantly associated with WRTP. The factors that were not significantly associated with WRTP were all demographic factors, education factors and work areas of the physiotherapists (p > 0.05).

Of the 258 physiotherapists who reported WRTP, 254 reported on the timing of the initial incident (Table 6). The majority (40.5%, n = 103) of these reported WRTP in their dominant thumbs within 5 years of graduating. Tables 7 and 8 show the prevalence and association of occupational factors with WRTP.

The lifetime prevalence of WRTP for the physiotherapists in the current South African study was 65.3% (n = 258). This is similar to the findings of a study in Australia by McMahon et al. (2006) which also reported on a national sample of physiotherapists and identified a lifetime prevalence of 65.3% (n = 961). Similarly, Barnes et al. (2011) reported a lifetime prevalence of wrist and thumb problems of 62.5% (n = 55) amongst South African respondents. The lifetime prevalence may have been lower if only thumb problems were reported.

Wajon and Ada (2003) reported a higher lifetime prevalence of WRTP (83%, n = 129). The different demographic samples could account for the higher lifetime prevalence. A cross-sectional design including all physiotherapists using manual therapy techniques was used in the current study and in that of McMahon et al. (2006), whereas Wajon and Ada (2003) only included physiotherapists who had postgraduate qualifications in manipulative therapy. If only manipulative therapy graduates were included in this study, it would exclude the newly qualified physiotherapists who are more likely to develop WRMSD (Barnes et al. 2011; Cromie et al. 2000; West & Gardner 2001).

In this study, newly qualified physiotherapists in the age group 20–30 years had a higher prevalence of WRTP (70%, n = 84). The reason for the increased prevalence rates in the younger physiotherapists, according to Cromie et al. (2000), is the reluctance of younger physiotherapists to ask for help. It could also be linked to the increased workload or inability to stabilise their thumbs during the application of manual therapy techniques (Barnes et al. 2011; Buckingham, Das & Trott, 2007; Walsh et al. 2011). The repetitive use of their thumbs during manual therapy techniques, especially on large and muscular patients, may pose a challenge which could possibly result in their thumb joints assuming hyperextended positions. Although this study did not find a significant association between WRTP and metacarpophalangeal (MP) joint hypermobility, McMahon et al. (2006) found a significant association between WRTP and MP and IP joint hyperextension (p < 0.001).

This study showed a significant association between WRTP and non-dominant thumb IP joint hyperextension (p = 0.02). Buckingham et al. (2007) and Wajon, Ada and Retshauge (2007) advocated that the thumb IP joint be positioned in extension or slight flexion to avoid WRTP, which is supportive of the finding. In contrast, Snodgrass and Rivett (2002) theorised that the lack of hyperextension of the IP joint during the application of posterior-anterior (PA) mobilisation to the spine would result in an increase in WRTP because the base of the thumb is positioned further away from the point of contact when the thumb pad is applying the pressure.

An understanding of the biomechanics of the different thumb joints is important in the prevention of WRTP. This is especially important when the biomechanics of the basal joint of the thumb (carpometacarpal joint) is considered. The osseous configuration of this joint allows for maximum mobility for the intricate functions of the hand; however, there are positions where there is incongruity of the joint surfaces (carpometacarpal joint in retroposition), resulting in decreased stability of the joint (Zancolli, Ziadenberg & Zancolli 1987). For this reason, Atkinson and Maher (2004) questioned physiotherapists’ use of the thumbs in the application of longitudinal pressure during manual therapy techniques. In order to lessen stress to the therapists’ thumb joints, Hu, Hsu and Su (2009) recommended a position in which the thumb is supported by the neighbouring index finger. Jull (2011) describes a modified hand placement for the cervical posterior-anterior glide techniques in which the hands and thumbs are fixed and the force is generated by forearm (elbow) flexion and extension. Walsh et al. (2011), Buckingham et al. (2007) and Wajon et al. (2007) recommended support of the thumbs in the optimal position (thumb in opposition, MP and IP joints in extension or slight flexion) by using taping or splints, especially for physiotherapists who have hypermobility of the thumb joints.

Thistlethwaite's (2005) hypothesised a tendency toward hypermobility as a reason for a statistically significant difference (p = 0.005) in race groups; however, the current study found no significant association between WRTP and race. The difference in race demographics could explain this difference. In the current study, 85% of the respondents were white, whilst in Thistlethwaite's (2005) study over 50% of the study population was Indian. She hypothesised that Indian physiotherapists, with a tendency toward hypermobility (Simpson 2006), would increase the tendency for joint instability and the development of WRTP. McMahon et al. (2006) reported a higher prevalence of WRTP in male physiotherapists, whereas the current study found a higher prevalence of WRTP in female physiotherapists. These differences could be explained by the difference in gender distribution in each context. In South Africa, physiotherapy is a female-dominated profession. Statistics obtained from the SASP for 2013 indicated that women represented 87.8% of the physiotherapists registered with the society. In this study, women represented 93% of the sample (n = 226 of the 243 that responded to the gender questionnaire) whilst in the study by McMahon et al. (2006), men represented 22% of the sample. A significant association of WRTP in male physiotherapists working in orthopaedic outpatient departments was also found by McMahon et al. (2006), who attributed this to the fact that male physiotherapists are mostly employed in work areas requiring greater use of manual techniques.

The number of patients treated with cervical spine pathology was significantly associated with WRTP in the physiotherapists in this study. Alternatives such as the use of the pisiform are not usually possible in the cervical region because of the contours in this body region, whereas thoracic and lumbar regions lend themselves to the use of the heel of the hand for the stronger grades of movement. McMahon et al. (2006) as well as the current study also found a significant association between the presence of WRTP and the hands-on activity of manual therapy, trigger point therapy and massage. A greater percentage of respondents (85.4%, n = 217) said that soft tissue techniques aggravated their symptoms when compared to studies by McMahon et al. (2006), Wajon and Ada (2003) and Barnes et al. (2011) (75%, n = 439; 69%, n = 106 and 65.5%, n = 36 respectively). Specific to the technique of passive accessory spinal joint mobilisation, 80% of the respondents in this study reported aggravation of WRTP. Wajon and Ada (2003) reported the same whilst McMahon et al. (2006) reported a slightly lower percentage (75%). Only the current study reported on the grades of the techniques that aggravated WRTP.

A significant association was found between the presence of WRTP and grades III–IV Maitland central posterior-anterior glides, grades II–IV Mulligan techniques as well as all grades of the Maitland unilateral posterior-anterior glides and transverse glides to the spine (p < 0.001). An explanation could be that less force is used by the thumbs during the application of the grade I and II techniques compared to the force used in the stronger grades of movement. All Maitland unilateral posterior-anterior glides and transverse glides to the spine, however, were significantly associated with WRTP, perhaps because of the position of the thumb during the application of the technique. The thumb is less likely to be in a position of congruence (carpometacarpal [CMC] in opposition) when performing all grades of the transverse glides, resulting in more stress to the joint surfaces and surrounding soft tissue (Atkinson & Maher 2004; Zancolli et al. 1987).

Other occupational factors associated with WRTP in physiotherapists treating musculoskeletal conditions using manual therapy techniques included the high, repetitive workload, treating a large numbers of patients daily, working with a current injury, working in sustained positions for prolonged periods, working in uncomfortable positions, inadequate preventative thumb injury training, inadequate rest periods and working at their physical limits. This was reported by 61% – 80% of the respondents in the current study and other studies reporting on WRTP in physiotherapists (Barnes et al. 2011; McMahon et al. 2006; Wajon & Ada 2003).

This study confirms that a high percentage of physiotherapists using manual therapy techniques to treat musculoskeletal conditions are experiencing WRTP. It would seem important that physiotherapists apply the same advice they would give patients who have musculoskeletal disorders. This advice, an important management strategy for WRTP, includes joint protection, job rotation and avoiding sustained postures or repetitive action (Glover 2002).

A longitudinal study which follows newly qualified physiotherapists for at least two years is recommended to investigate a possible cause-effect relationship for WRTP in physiotherapists. Randomised controlled trials that investigate preventative strategies like taping or splints for the physiotherapists’ thumbs are also recommended. If emerging research on this subject is communicated to undergraduate students and practising physiotherapists on an on-going basis, preventative strategies could be implemented. This would prevent physiotherapists from moving to another field of practice or, worse still, leaving the profession as a result of WRTP (McMahon et al. 2006; Snodgrass & Rivett 2002; Wagon & Ada 2003).