Table 2. Summary of findings.
| Evidence treatment options across regional musculoskeletal pain presentations | |||||
|---|---|---|---|---|---|
| Treatment Options | Service Model | Evidence base | Outcomes / Effects | Comments | Overall Strength of evidence (Grade) |
| Clinical outcome (pain and disability) | Open access | 1 Trial (Holdsworth 2007, Overman et al 1988), 2 Cohorts (Badke et al 2014, Denninger 2018); 2 Service evaluations (Ojha 2015, Mallet 2014). | Small differences between groups (e.g., Mean functional improvement score at discharge 15.2 ±11.7 for self-referred patients vs 14.6 ±10.6 for GP led care; p = 0.77) on a 0–100 scale for function) and (e.g., percent decrease in pain 64.6% for self-referrers vs 66.6% for Physician referred patients; p = 0.76), Badke et al. 2014; Mean improvement in function from baseline, 54%; 95% CI: 46%, 62%) and pain (mean difference, 4 points; 95% CI: 1, 7 points), with no differences between groups (P>.05), Denninger 2018). | Overall, patients displayed good clinical improvement in disability and pain, with no differences between groups (P >.05). Between group differences in pain and function were also not sustained in the long term (>12 months). | ** Limited evidence |
| Combination | 2 cross-sectional analysis of patient cohort. Ludvigsson 2012; Phillips 2012 | Mean (SD) summary index (EQ VAS) of self-rated health including pain and functional disability on a scale from 0 to 100: 67 (18) for self-referred patients vs. 56 (19) for GP-led care; p = 0.006). Ludvigsson 2012. Mean pain intensity (VAS (SD)) 6.91 (9.4), p<0.001 at 3 months follow up. | Significant differences were found between groups. Relatively small data-set (n = 93) from a patient cohort. Philips et al 2012 was compared to baseline but did not include comparison group data. | ||
| Service based pathway | 1 cluster randomised trial. Bishop et al 2017 | Perceived change from baseline:4% of self-referred patients vs. 6.5% of GP-led care patients reported complete recovery at 6 months | Evidence from pilot trial. (cluster randomisation based on GP practices). | ||
| Clinical outcome (Quality of life) | Open access | 2 Cohort (Denninger 2018; Goodwin 2016/Moffatt 2017). | Beneficial effects demonstrated. Small, statistically insignificant differences between groups at follow-up (e.g. percent change in pre-post EQ 5D mean (SD) at 6 months 0.13 (0.27) Goodwin 2016). | Comparable improvements (slightly better among self-refers) in QoL outcomes for up to 2 years across studies. | **Limited evidence |
| Combination | 2 cross-sectional analysis of patient cohort. Ludvigsson 2012; Phillips 2012 | e.g., mean EQ 5D (SD) 0.65 (0.22) for self-referred groups vs. 0.51 (0.30) for GP led care at 3 months, p = 0.014 Ludvigsson et al; and 0.82 (0.2) at 3months, p<0.001 Phillips et al 2012. | Unadjusted analysis | ||
| Service based pathway | 1 cluster randomised trial. Bishop et al 2017 | Mean EQ 5D score (SD) for control practices vs intervention practices respectively: @ baseline: 0.565 (0.246) vs. 0.544 (0.262) @ 6 months 0.602 (0.251) vs. 0.594 (0.262) @ 12 months 0.615 (0.254) vs. 0.606 (0.258) | Quality of life increased similarly in both arms compared to baseline across all follow-up time points | ||
| Safety outcomes (adverse effects and missed red-flag diagnoses) | Open access | 2 Cohort (Denninger 2018, Goodwin 2016) 2 service evaluation (Mintken 2015, Moore 2005). Other studies without safety as a priori outcomes: (McGill et al 2013, Ojha 2015, Pendergast et al 2012, Holdsworth 2007, Greenfield 1975, Boissonnault 2010, 2016, Desjardins-Charbonneau et al 2016) | No adverse events/effects, missed red flag diagnoses due to accessing care through MSK triage and direct access/self-referral across all included studies. | MSK triage/direct access presented no higher risks to patients. However, most services included specially trained and/or more senior professionals. | *** Moderate evidence |
| Combination | Other studies without safety as a priori outcomes: Ferguson et al 1999 | Informal liaison with GPs, access to patient medical notes, and use of pre-defined protocol/checklists for minimising mis-diagnosis. | |||
| Service based pathway | 1 cluster randomised trial. Bishop et al 2017 | No evidence that the direct access pathway led to adverse events, missed diagnosis of serious pathologies. No comparison with control practices without direct access services. | |||
| Socio-economic outcomes (work absence and sickness certification) | Open access | 1 Trial (Holdsworth 2007) 1 cross-sectional analysis (McGill et al 2013) | (Mean MSK related work absence, S.D., range (days): 2.5, ±10.6, 0 to 120 for self-referrers; vs. 6.0, ±19.6, 0 to 300; p = 0.048). Holdsworth et al 2007 94% drop in lost time from work due to MSK related condition over 12 months. | Consistently large differences in favour of direct across/self-referral for up to 12 months across studies. | *** Moderate evidence |
| Combination | 1 cross sectional analysis (Bornhoft 2015) 1 analysis of patient cohort (Phillips 2012). | N (%) of sick-leave recommendations for direct access and GP led care respectively. 82 (14.1%) vs. 369 (23.2%) @ 6months 73 (15.1%) vs. 338 (23.5%) @ 12 months. Bornhft 2015. Mean (SD) Sickness absence @ baseline and @ 3months 4.6 (12.6) vs. 1.45 (9.7); p <0.05 Mean (SD) Work performance @ baseline and @ 3months 75.9 (19.6) vs. 87.8 (13.2); p <0.001. Phillips et al 2012 | Significant differences in work related outcomes relative to baseline. | ||
| Service based pathway | 1 cluster randomised trial. Bishop et al 2017 | Mean (SD) work related costs associated with MSK conditions: £740.30 (2084.75) for control practices vs £ 539.36 (2069.43) for intervention practices who accessed care via MSK triage/ direct self-referrals. | Work related absence costs were significantly higher for patients without direct access. Outcome over 12 month period. | ||
| Health care utilisation (costs, further consultations, prescriptions, tests, referrals, and impact on GP workload/services) | Open access | 2 Trial (Holdsworth 2007, Greenfield 1975), 3 Cohorts (Badke et al 2014, Denninger 2018, Goodwin 2016); 2 Service evaluations (Ojha 2015, Swinkels 2014). 4 cross-sectional analysis (McGill et al 2013, Mitchell et al 1997, Pendergast et al 2012) | Badke- Mean total cost of care per patient (SD): $2423.5 (2555.3). Mean total cost of care per patient (SD): $3878.7 (2923.8) Denninger 2014. Total cost care per patient (SD): 1542 (108, 2976). For DA vs 3085 (1939, 4224) McGill et al 2013: Medication use: Medication use: 24.07% for DA compared to 90.53% for GP led care. Radiology use: 11.11% compared to 82.11% for GP led care. | Overall, consistently significant differences in health care utilisation costs (higher for usual GP-led care compared to MSK triage and direct access/self-referral) | ***Moderate evidence |
| Combination | 1 cross sectional analysis (Bornhoft 2015) 2 analysis of patient cohort (Phillips 2012; Ludvigsson 2012. 1 service evaluation (Mallet 2014) | ||||
| Service based pathway | 1 cluster randomised trial. Bishop et al 2017 | ||||
*Very weak evidence: Perspective / opinions only/ Absence of empirical data (from qualitative or quantitative studies).
** Limited evidence: Some empirical evidence from cohort and cross-sectional observational studies, lacking comparisons with usual GP led care, AND when there were small, inconsistent, or non-significant differences in patient related outcomes, OR without.
*** Moderate evidence: Some empirical evidence from trials, good quality cohort and cross-sectional analyses of large data sets including, comparisons with usual GP led care, and /or with small to moderate but consistent effects on patient related outcomes.
**** Strong evidence: Evidence from good quality trials, cohort and cross-sectional analyses of large data sets including direct access, comparisons with usual GP led care, and /or with moderate to strong consistent effects on patient related outcomes.