Study designs of evaluations included in the review
Randomised controlled trials (RCTs), controlled trials, case reports, case series, case-control studies, surveys or surveillance studies were eligible for inclusion.
Specific interventions included in the review
Studies of spinal manipulation were eligible for inclusion. Inclusion criteria were not limited by the comparison intervention. Where reported, a variety of types of spinal manipulation were used: strong rotation; flexion, extension, axial loading; rapid manual rotations of the head with flexion and extension; assisted passive motion of cervical spine; high velocity, low-amplitude; and Gonstead technique. Where reported, chiropractors (including students) mainly performed the manipulations; other practitioners included physiotherapists and medical doctors.
Participants included in the review
Studies that included children aged 18 years or younger were eligible for inclusion. Where reported, children were undergoing spinal manipulation for a variety of conditions, such as congenital torticollis, spinal cord astrocytoma, osteogenesis imperfects, head and neck and back pain, nocturnal enuresis and chronic otitis media.
Outcomes assessed in the review
Studies that reported adverse events were eligible for inclusion. The review assessed adverse events associated with the delivery of spinal manipulation and those associated with delayed or missed diagnoses. Adverse events were classified as severe (hospitalisation, permanent disability, mortality), moderate (transient disability with the seeking of medical aid but no hospitalisation), minor (self-limiting, requiring no additional medial care) and delayed diagnosis or treatment (moderate to severe adverse events as defined previously, as a result of delayed diagnosis or treatment of a medical condition).
How were decisions on the relevance of primary studies made?
Two reviewers independently screened titles and abstracts, while three reviewers independently selected the studies.