Table 2 Characteristics of included reviews.

Fibromyalgia [14]

August 2007

A confirmed diagnosis of fibromyalgia

34/2276

Mean van Tulder scores^a: 7.33 for aerobic training; 4 for strength training)

Met: all

Low back pain [22]

August 2008

Non-specific low back pain (with or without leg pain) 12 weeks or longer

41/4485

Mean PEDro scores^b: 6.33 for exercise versus no treatment), and 5.82 for exercise versus minimal intervention

Met: 2, 3, 4, 5, 6, 7, 8 and 9; not met: 10, 11; unclear: 1

Neck pain [21]

February 2008

Non-specific neck pain in adults

33/6218

Mean PEDro scores: 5.67

Met: 2, 6, 7, 11; not met: 3, 4, 5, 8, 9, 10; unclear: 1

Shoulder pain [20]

March 2011

Primary symptom of shoulder pain in adults

17/1436

Mean PEDro scores: 5.87

Met: 5, 7, 11; not met: 3, 4, 8, 10; unclear: 1, 2, 6, 9

Osteoporosis [15]

December 2010

Healthy postmenopausal aged women between 45 and 70 years

43/4320

Most trials were scored as 'unclear' for randomization, allocation concealment, attrition, selective reporting, and assessor blinding

Met: all

Osteoarthritis (knee) [16]

December 2007

Knee OA based on accepted criteria or self-reported knee OA on the basis of chronic joint pain

32/3616

Of 43 trials, 9 (28 % ) studies had low risk of bias

Met: 1, 2, 4, 5, 6, 7, 8, 9, 11; not met: 3, 10

Osteoarthritis (hip) [17]

August 2008

Hip OA based on accepted criteria or self-reportedhip OA

5/267

Of 5 trials, 3 had low risk of bias and 2 trials had moderate risk of bias

Met: 1, 2, 4, 5, 6, 7, 8, 9, 11; not met: 3, 10

Rheumatoid Arthritis [18]

December 2008

Adults with classic or definite RA based on established criteria

8/575

4 of 8 trials fulfilled at least 8 of 10 methodological criteria

Met: all

Ankylosing spondylitis [19]

January 2007

AS based on the modified New York criteria

11/763

Of 11 trials, 3 had low risk of bias; 3 had moderate risk of bias, and 5 had high risk of bias

Met: all