Table 2 Secondary analyses:

(1)

Assessing equivalence in clinical outcome using the MYMOP2 score.

(2)

Examining clinical outcome at 6 weeks using the SF-36 PCS

(3)

Comparing the proportion of patients who 'respond to treatment' in each arm, in line with the Outcomes Measures in Rheumatology Clinical Trials- Osteoarthritis Research Society International (OMERACT-OARSI) recommendations [30], using the SF-36 physical function and bodily pain scales and the global improvement score

(4)

Repeating the primary analysis adjusting also for any variables exhibiting marked imbalance at baseline to check that this does not influence the findings

(5)

Analyses as above for secondary outcomes (where p-values will be adjusted to account for multiple testing)

(6)

Investigating the effectiveness and cost-effectiveness of PhysioDirect for patients of different age-groups, or with different presenting problems

(7)

Investigating the effectiveness and cost-effectiveness of PhysioDirect in each of the four PCTs, in exploratory sub-group analyses

(8)

Investigation of process measures such as physiotherapy consultation rates, physiotherapy 'did-not-attend' rates and consultation rates with other health care services in the NHS and private sectors