Summary of findings for the main comparison. Cognitive behavioural therapy compared with control for children and adolescents with recurrent abdominal pain.

Cognitive behavioural therapy compared with control for children and adolescents with recurrent abdominal pain
Patient or population: children and adolescents with recurrent abdominal pain Settings: mixed Intervention: cognitive behavioural therapy Comparison: usual care or wait‐list control
Outcomes	Probable outcome with control or usual care	Probable outcome with CBT	OR (95% CI)	Number of participants (studies)	Quality of the evidence (GRADE)	Comments
Treatment success: postintervention	211 per 1000	494 per 1000	Pooled OR 5.67 (1.18 to 27.32)	175 (4)	⊕⊝⊝⊝ 1 Very low	Varied definitions of 'treatment success' used by authors (pain free; reduction of 10 points on API; Walker 1997).
Treatment success: medium‐term follow‐up (between 3 and 12 months)	349 per 1000	551 per 1000	Pooled OR 3.08 (0.93 to 10.16)	139 (3)	⊕⊕⊝⊝ 2 Low	Varied definitions of 'treatment success' used by authors (pain free; reduction of 10 points on API; Walker 1997).
Pain intensity: postintervention Lower score equals less pain.	The pain score in the CBT groups was, on average, 0.33 SDs lower (95% CI ‐0.74 to 0.08) than in the usual care, wait‐list, or education control groups.		—	405 (7)	⊕⊕⊝⊝ 3 Low	As a rule of thumb, 0.2 SD represents a small difference, 0.5 SD a moderate difference, and 0.8 SD a large difference. Varied measures used to assess pain intensity (FACES Pain Scale (Bieri 1990); visual analogue scale; Likert scale).
Pain intensity: medium‐term follow‐up (between 3 and 12 months) Lower score equals less pain.	The pain score in the CBT groups was, on average, 0.32 SDs lower (95% CI ‐0.85 to 0.20) than in the usual care, wait‐list, or education control groups.		—	301 (4)	⊕⊕⊝⊝ 4 Low	As a rule of thumb, 0.2 SD represents a small difference, 0.5 SD a moderate difference, and 0.8 SD a large difference.
QOL (physical subscale): postintervention Higher score equals better QOL.	The QOL score (physical subscale) in the CBT groups was, on average, 0.71 SDs higher (95% CI ‐0.25 to 1.66) than in the usual care, wait‐list, or education control groups.		—	136 (3)	⊕⊝⊝⊝ 5 Very low	As a rule of thumb, 0.2 SD represents a small difference, 0.5 SD a moderate difference, and 0.8 SD a large difference. 2 studies used PedsQL (Varni 2001), 1 study used KIDSCREEN (Ravens‐Sieberer 2005).
QOL (psychosocial subscale): postintervention Higher score equals better QOL.	The QOL score (psychosocial subscale) in the CBT groups was, on average, 0.43 SDs higher (95% CI ‐0.21 to 1.06) than in the usual care, wait‐list, or education groups.		—	136 (3)	⊕⊕⊝⊝ 6 Low	As a rule of thumb, 0.2 SD represents a small difference, 0.5 SD a moderate difference, and 0.8 SD a large difference. 2 studies used PedsQL (Varni 2001), 1 study used KIDSCREEN (Ravens‐Sieberer 2005).
Functional disability or activity limitations: postintervention Lower score equals less activity disability.	Functional disability in the CBT groups was, on average, 0.57 SDs lower (95% CI ‐1.34 to 0.19) than in the usual care, wait‐list, or education control groups.		—	176 (4)	⊕⊝⊝⊝ 7 Very low	As a rule of thumb, 0.2 SD represents a small difference, 0.5 SD a moderate difference, and 0.8 SD a large difference. 3 different functional disability or activity limitation indices used (KINDL‐R (Ravens‐Sieberer 2005); CALI (Palermo 2004; Palermo 2016); FDI (Walker 1991)).
API: Abdominal Pain Index; CALI: Child Activity Limitations Interview; CBT: cognitive behavioural therapy; CI: confidence interval; FDI: Functional Disability Inventory; KIDSCREEN: Health Related Quality of Life Questionnaire for Children and Young People; KINDL‐R: measure of health‐related quality of life; OR: odds ratio; PedsQL: Pediatric Quality of Life Inventory; QOL: quality of life; SD: standard deviation
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

¹Downgraded three levels: high risk of bias across the studies in study design and outcome assessment; high level of unexplained heterogeneity (> 70%); and a low number of participants included in the analysis, wide CIs.
 ²Downgraded two levels: high risk of bias across the studies in study design and a low number of participants included in the analysis, wide CIs.
 ³Downgraded two levels: high level of unexplained heterogeneity (> 70%) and high risk of bias across the studies, with baseline differences in primary outcomes in the largest study.
 ⁴Downgraded two levels: high level of unexplained heterogeneity (> 70%) and high risk of bias across the studies, with baseline differences in primary outcomes in the largest study.
 ⁵Downgraded three levels: high risk of bias across the studies in study design; high level of unexplained heterogeneity (> 70%); and a low number of participants included in the analysis, wide CIs.
 ⁶Downgraded two levels: high risk of bias across the studies in study design and a low number of participants included in the analysis, wide CIs.
 ⁷Downgraded three levels: high risk of bias across the studies in study design; high level of unexplained heterogeneity (> 70%); and a low number of participants included in the analysis, wide CIs.