Criteria

Description

Random sequence generation (selection bias ‐ biased allocation to interventions ‐ due to inadequate generation of a randomised sequence)

1. Criteria for a judgement of low risk of bias: the investigators describe a random component in the sequence generation process such as:

   1. referring to a random number table;

   2. using a computerised random number generator;

   3. coin tossing;

   4. shuffling cards or envelopes;

   5. throwing dice; or

   6. drawing of lots and minimisation.

2. Criteria for a judgement of high risk of bias: the investigators describe a non‐random component in the sequence generation process. Usually the description would involve some systematic, non‐random approach, for example:

   1. sequence generated by odd or even date of birth;

   2. sequence generated by some rule based on date (or day) of admission;

   3. sequence generated by some rule based on hospital or clinic record number. Other non‐random approaches happen much less frequently than the systematic approaches mentioned above and tend to be obvious, for example: allocation by judgement of the clinician; allocation by preference of the participant; allocation based on the results of a laboratory test or a series of tests and allocation by availability of the intervention.

3. Criteria for a judgement of unclear risk of bias: insufficient information about the sequence generation process to permit a judgement of low or high risk of bias.

Allocation concealment (selection bias ‐ biased allocation to interventions ‐ due to inadequate concealment of allocations prior to assignment)

1. Criteria for a judgement of low risk of bias: participants and investigators enrolling participants could not foresee assignment because one of the following, or an equivalent method, was used to conceal allocation:

   1. central allocation;

   2. sequentially numbered drug containers of identical appearance; or

   3. sequentially numbered, opaque, sealed envelopes.

2. Criteria for a judgement of high risk of bias: participants or investigators enrolling participants could possibly foresee assignments and thus introduce selection bias such as allocation based on:

   1. an open random allocation schedule (for example, a list of random numbers);

   2. assignment of envelopes without appropriate safeguards (for example, if envelopes were unsealed or non­opaque or not sequentially numbered);

   3. alternation or rotation;

   4. date of birth;

   5. case record number; or

   6. any other explicitly unconcealed procedure.

3. Criteria for a judgement of unclear risk of bias: insufficient information to permit a judgement of low or high risk of bias. This is usually the case if the method of concealment is not described or not described in sufficient detail to allow a definite judgement; for example, if the use of assignment envelopes is described but it is unclear whether envelopes were sequentially numbered, opaque and sealed.

Blinding of participants and personnel (performance bias due to knowledge of the allocated interventions by participants and personnel during the study)

1. Criteria for a judgement of low risk of bias may involve any one of the following:

   1. no blinding or incomplete blinding, but the review authors judge that the results are unlikely to be influenced by lack of blinding; or

   2. blinding of participants and key study personnel ensured, and it is unlikely that the blinding could have been broken.

2. Criteria for a judgement of high risk of bias may involve any one of the following:

   1. no blinding or incomplete blinding, and the results are likely to be influenced by lack of blinding; or

   2. blinding of key study participants and personnel attempted, but it is likely that the blinding could have been broken, and the results are likely to be influenced by lack of blinding.

3. Criteria for a judgement of unclear risk of bias may involve any one of the following:

   1. insufficient information to permit a judgement of low or high risk of bias; or

   2. the study did not address this outcome.

Blinding of outcome assessment (detection bias due to knowledge of the allocated interventions by outcome assessors)

1. Criteria for a judgement of low risk of bias may involve any one of the following:

   1. no blinding of outcome or outcome assessment (or both), but the review authors judge that the outcome and its measurement are unlikely to be influenced by lack of blinding; or

   2. blinding of outcome and outcome assessment ensured, and it is unlikely that the blinding could have been broken.

2. Criteria for a judgement of high risk of bias may involve any one of the following:

   1. no blinding of outcome or outcome assessment (or both), and the outcome and its measurement are likely to be influenced by lack of blinding; or

   2. blinding of outcome or assessment (or both), but it is likely that the blinding could have been broken, and the outcome and its measurement are likely to be influenced by lack of blinding.

3. Criteria for a judgement of unclear risk of bias may involve any one of the following:

   1. insufficient information to permit a judgement of low or high risk of bias; or

   2. the study did not address this outcome.

Incomplete outcome data (attrition bias due to the amount, nature or handling of incomplete outcome data)

1. Criteria for a judgement of low risk of bias may involve any one of the following:

   1. no missing outcome data;

   2. reasons for missing outcome data are unlikely to be related to true outcome (for survival data, censoring unlikely to be introducing bias);

   3. missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups;

   4. for dichotomous outcome data, the proportion of missing outcomes compared with the observed event risk is not enough to have a clinically relevant impact on the intervention effect estimate;

   5. for continuous outcome data, the plausible effect size (difference in means or standardised difference in means) among missing outcomes is not enough to have a clinically relevant impact on the observed effect size; or

   6. missing data have been imputed using appropriate methods.

2. Criteria for a judgement of high risk of bias may involve any one of the following:

   1. the reason for missing outcome data is likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups;

   2. for dichotomous outcome data, the proportion of missing outcomes compared with the observed event risk is enough to induce clinically relevant bias in the intervention effect estimate;

   3. for continuous outcome data, the plausible effect size (difference in means or standardised difference in means) among missing outcomes is enough to induce clinically relevant bias in the observed effect size;

   4. 'as‐treated' analysis done with substantial departure of the intervention received from that assigned at randomisation; or

   5. potentially inappropriate application of simple imputation.

3. Criteria for a judgement of unclear risk of bias may involve any one of the following:

   1. insufficient reporting of attrition or exclusions (or both) to permit a judgement of low or high risk of bias (for example, number randomised not stated, no reasons for missing data provided); or

   2. the study did not address this outcome.

Selective reporting (reporting bias due to selective outcome reporting)

1. Criteria for a judgement of low risk of bias may involve any of the following:

   1. the study protocol is available and all of the study's pre‐specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre‐specified way; or

   2. the study protocol is not available but it is clear that the published reports include all expected outcomes, including those that were pre‐specified (convincing text of this nature may be uncommon).

2. Criteria for a judgement of high risk of bias may involve any one of the following:

   1. not all of the study's pre‐specified primary outcomes have been reported;

   2. one or more primary outcomes is reported using measurements, analysis methods or subsets of the data (for example, subscales) that were not pre‐specified;

   3. one or more reported primary outcomes were not pre‐specified (unless clear justification for their reporting is provided, such as an unexpected adverse effect);

   4. one or more outcomes of interest in the review were reported incompletely so that they could not be entered in a meta‐analysis; or

   5. the study report failed to include results for a key outcome that was expected to have been reported for such a study.

3. Criteria for a judgement of unclear risk of bias: insufficient information to permit a judgement of low or high risk of bias. It was likely that most studies would fall into this category.

Other bias (other sources of bias not captured by the other domains)

1. Criteria for a judgement of low risk of bias: the study appears to be free of other sources of bias.

2. Criteria for a judgement of high risk of bias: at least one important risk of bias exists. For example, the study:

   1. had a potential source of bias related to the specific study design used; or

   2. has been accused of bring fraudulent; or

   3. had some other problem.

3. Criteria for a judgement of unclear risk of bias: there is a risk of bias, but there is either:

   1. insufficient information to assess whether an important risk of bias exists; or

   2. insufficient rationale or evidence that an identified problem introduced bias.

Method

Approach

Measures of treatment effect

Dichotomous data

We did not find dichotomous data. Should these data become available in future updates of this review, we will calculate the risk ratio (RR) with 95% confidence intervals (CI), as most readers find it easier to understand the RR and 95% CI than the odds ratio and risk difference.

Unit of analysis issues

Cluster‐RCTs

We did not find cluster‐RCTs. This design is uncommon in this field. Should such studies become available in the future, and if the investigators report cluster‐randomised trial data as though the randomisation was performed on individuals rather than on clusters, we will request individual participant data to calculate an estimate of the intra‐cluster correlation coefficient (ICC). If the individual participant data are not available, we will obtain external estimates of the ICC from similar studies or available resources (Campbell 2000). Once established, we will use the ICC to re‐analyse the trial data to obtain approximate, correct analyses as described in section 16.3.4 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We will combine the effect estimates and their corrected standard errors from cluster‐RCTs with those from parallel‐group designs using the generic inverse variance method (Higgins 2011). If the available information is insufficient to control for clustering in this manner, we will enter the data into Review Manager 5 (RevMan 2014), using individuals as the unit of analysis. We planned to perform sensitivity analyses to assess the potential bias that may occur as a result of the inadequately controlled clustered trials. Additionally, if the ICCs are obtained from external sources, we will perform sensitivity analyses to assess the potential biasing effects of inadequately controlled cluster‐randomised trials (Donner 2001).

Cross‐over trials

We did not find cross‐over trials. Should we find these types of studies in future updates of this review, and as the duration of any effect of CBT is unknown, we will use only first‐period data from any cross‐over trials that fit the inclusion criteria to avoid a possible carry‐over effect.

Dealing with missing data

If the studies do not report the standard deviation (SD), we will calculate it from the P values, t values, CIs or standard errors (as described in section 7.7.3.3 of the Cochrane Handbook for Systematic Reviews of Interventions; Higgins 2011). If this information is not reported or is unattainable, we will impute the SD from the study with the highest SD for that outcome, and assess the effects of this assumption on the analysis by conducting a sensitivity analysis. If the outcome data are reported as a median, a range or as a mean without a variance, we will report the data in additional tables.

Sensitivity analysis

We will conduct sensitivity analyses to:

1. assess the potential bias that may have occurred as a result of inadequately controlled clustered‐RCTs;

2. assess the effect of different values of the ICC;

3. compare the results of the analyses with our imputed 'highest SD' versus analyses that used an SD imputed from the study with the lowest SD; and

4. assess the effects of eventual missing dichotomous data on our primary meta‐analyses by assuming, on the one hand, that all missing data were successes and, on the other hand, that all missing data were failures (best versus worst‐case scenario analyses).

Outcomes

Studies and instruments

ADHD symptom severity

1. Emilsson 2011: Kiddie Schedule for Affective Disorder and Schizophrenia (SADS) ‐ Present and Lifetime Version (K‐SADS‐PL) (Kaufman 1996)/Barkley ADHD Current Symptoms Scale (BCS‐Total Score) (Barkley 1998b)/R&R2 ADHD Training Evaluation Self‐Report Scale (RATE‐S) (Young 2007b)

2. Fleming 2015: Brown Attention Deficit Disorder Scale – Adult Version (BADDS) (Brown 1996)

3. Gu 2017: Conners' Adult ADHD Rating Scale ‐ Self‐Reported (CAARS‐S) (Conners 1999c)

4. Hepark 2015: Conners' Adult ADHD Rating Scale ‐ Investigator Rated (CAARS‐INV) (Conners 1999c)/Conners' Adult ADHD Rating Scale ‐ Self‐Reported (CAARS‐S) (Conners 1999c)

5. Hirvikoski 2011: current ADHD Symptoms Scale (self‐report form) (Barkley 1998a)

6. Pettersson 2017: current ADHD Symptoms Scale (self‐report) (Barkley 2006b)

7. Safren 2005: current ADHD Symptoms Scale (self‐report and clinician rating versions) (Barkley 1998a)

8. Safren 2010: ADHD Rating Scale (DuPaul 1998)/current ADHD Symptoms Scale (self‐report) (Barkley 2006b)

9. Schoenberg 2014: Conners' Adult ADHD Rating Scale ‐ Self‐Reported: Screening Version (CAARS: SV) (Conners 2008)

10. Solanto 2010: Conners Adult ADHD Rating Scales – Observer: Long Version, inattention/memory subscale (T‐score) (Conners 1999b; Erhardt 1999)/Brown Attention‐Deficit Disorder Scale, total score (T‐score) (Brown 1996)

11. Stevenson 2002: DSM‐III‐R ADHD Checklist (DSM‐III‐R 1989; Gittelman 1985)

12. Vidal Estrada 2013: ADHD Rating Scale (ADHD‐RS) (DuPaul 1990)

13. Virta 2010: Brown Attention Deficit Disorder Scale – Adult Version (BADDS) (Brown 1996)/WHO Adult ADHD Self‐Report Scale (ASRS) (Kessler 2005)

Inattention

1. Emilsson 2011: Barkley ADHD Current Symptoms Scale (BCS‐Inattention) (Barkley 1998b)

2. Fleming 2015: Barkley Adult ADHD Rating Scale ‐ Fourth Edition (BAARS‐IV) (Barkley 2011)

3. Hepark 2015: Conners' Adult ADHD Rating Scales ‐ Investigator Rated (CAARS‐INV) (Conners 1999c)/Conners' Adult ADHD Rating Scale ‐ Self‐Reported (CAARS‐S) (Conners 1999c)

4. Moëll 2015: the WHO Adult ADHD Self‐Report Scale (ASRS) (Kessler 2005)

5. Schoenberg 2014: Conners' Adult ADHD Rating Scale ‐ Self‐Reported: Screening Version (CAARS‐S: SV) (Conners 2008)

6. Solanto 2010: Adult ADHD Investigator Symptom Rating Scale Inattention subscale (AISRS) (Adler 2003)

Hyperactivity

1. Moëll 2015: the WHO Adult ADHD Self‐Report Scale (ASRS) (Kessler 2005)

2. Vidal Estrada 2013: Conners' Adult ADHD Rating Scale ‐ Self‐Report (CAARS‐S) (Amador‐Campos 2014)

Impulsivity

1. Vidal Estrada 2013: Conners' Adult ADHD Rating Scale ‐ Self‐Report (CAARS‐S) (Amador‐Campos 2014)

Hyperactivity‐impulsivity

1. Emilsson 2011: Barkley ADHD Current Symptoms Scale (BCS‐Hyperactivity/Impulsivity) (Barkley 1998b)

2. Hepark 2015: Conners' Adult ADHD Rating Scale ‐ Investigator Rated (CAARS‐INV) (Conners 1999c)/Conners' Adult ADHD Rating Scale ‐ Self‐Reported (CAARS‐S) (Conners 1999c)

3. Schoenberg 2014: Conners' Adult ADHD Rating Scale ‐ Self‐Report: Screening Version (CAARS‐S: SV) (Conners 2008)

Clinical Global Impression

1. Emilsson 2011: Clinical Global Impression Scale ‐ NIMH (CGI) (NIMH 1985)

2. Safren 2005: Clinical Global Impression Scale ‐ NIMH (CGI) (NIMH 1985)

3. Safren 2010: Clinical Global Impression Scale ‐ NIMH (CGI) (NIMH 1985)

4. Vidal Estrada 2013: Clinical Global Impression Scale ‐ NIMH (CGI) (NIMH 1985)

5. Virta 2010: Clinical Global Impression Scale ‐ NIMH (CGI) (Guy 1976)

Outcomes

Studies and instruments

Depression

1. Emilsson 2011: Beck Depression Inventory (BDI) (Beck 1961)

2. Fleming 2015: Beck Depression Inventory ‐ Second Edition (BDI‐II) (Beck 1996)

3. Gu 2017: Beck Depression Inventory ‐ Second Edition (BDI‐II) (Beck 1996)

4. Hepark 2015: Beck Depression Inventory ‐ Second Edition (BDI‐II) (Beck 1996)

5. Moëll 2015: Hospital Anxiety and Depression Scale (HADS) (Zigmond 1983)

6. Pettersson 2017: Beck Depression Inventory ‐ Second Edition (BDI‐II) (Beck 1996)

7. Safren 2005: Hamilton Depression Scale (HAM‐D) (Hamilton 1960)/Beck Depression Inventory (BDI) (Beck 1961)

8. Solanto 2010: Beck Depression Inventory (BDI) (Beck 1996)

9. Vidal Estrada 2013: Beck Depression Inventory (BDI) (Beck 1961)

10. Virta 2010: Beck Depression Inventory ‐ Second Edition (BDI‐II) (Beck 1996)

Anxiety

1. Emilsson 2011: Beck Anxiety Inventory (BAI) (Beck 1993)

2. Fleming 2015: Beck Anxiety Inventory (BAI) (Beck 1993)

3. Gu 2017: Beck Anxiety Inventory (BAI) (Beck 1993)

4. Hepark 2015: State‐Trait Anxiety Inventory (Van der Ploeg 2000)

5. Moëll 2015: Hospital Anxiety and Depression Scale (HADS) (Zigmond 1983)

6. Pettersson 2017: Beck Anxiety Inventory (BAI) (Beck 2012)

7. Safren 2005: Hamilton Anxiety Scale (HAM‐A) (Hamilton 1959)/Beck Anxiety Inventory (BAI) (Beck 1988)

8. Solanto 2010: Hamilton Anxiety Scale (HAM‐A) (Shear 2001)

9. Vidal Estrada 2013: State‐Trait Anxiety Inventory‐State subscale (Spielberger 1986)

State anger

1. Stevenson 2002: State‐Trait Anger Expression Inventory (STAXI) (Speilberger 1991)

Trait anger

1. Stevenson 2002: State‐Trait Anger Expression Inventory (STAXI) (Speilberger 1991)

Self‐esteem

1. Solanto 2010: Rosenberg Self‐Esteem Inventory (Rosenberg 1965b)

2. Stevenson 2002: Davidson and Lang Self‐Esteem Measure (Davidson 1960)

Quality of life

1. Fleming 2015: ADHD Quality of Life Questionnaire (AAQoL) (Brod 2006)

2. Pettersson 2017: ADHD Impact Module ‐ Adult (AIM‐A) (Landgraf 2007)

3. Vidal Estrada 2013: Quality of Life Enjoyment and Satisfaction Questionnaire (QLESQ) (Endicott 1993)

4. Virta 2010: Quality of Life Enjoyment and Satisfaction Questionnaire (QLESQ) (Endicott 1993)

Employment status

No study included this outcome.

Section of Review

Protocol

Full review

Description of the condition

"The International Classification of Diseases (ICD‐10) offers a similar definition for hyperkinetic disorders (WHO 1993). Along with these three main symptomatic clusters, people with ADHD also present with deficits in executive functions, behaviour and emotion regulation, and motivation (Brown 2000; Wender 2001; Davidson 2008; Torrente 2011). There is a high prevalence of comorbid disorders, estimated at 50% to 75% (Kessler 2006), including anxiety, depression and substance abuse (Biederman 1993; Murphy 1996). Epidemiological studies estimate that the prevalence of ADHD is approximately 5% in childhood (Polanczyk 2007) and approximately 2.5% in adulthood (Simon 2009)."

"The International Classification of Diseases (ICD‐10) offers a similar definition for hyperkinetic disorders (WHO 1993), but the required number of symptoms and the age of onset are different. Along with these three main symptomatic clusters, people with ADHD also present with deficits in executive functions, behaviour and emotion regulation, and motivation (Brown 2000; Davidson 2008; Torrente 2011; Wender 2001). There is a high prevalence of comorbid disorders, estimated at 50% to 75% (Kessler 2006), including anxiety, depression and substance abuse (Biederman 1993; Murphy 1996). Epidemiological studies estimate that the prevalence of ADHD is approximately 5% in childhood and around 2.5% in adulthood (Polanczyk 2014; Simon 2009)."

Why it is important to do this review

"Between 20% and 50% of people with ADHD do not respond to drug treatment (Wilens 2002)."

"Between 20% and 50% of people with ADHD do not respond to drug treatment (Wilens 2002). Also, pharmacological treatment is frequently associated with relevant side effects in both children and adults (AJCD 2001; Castells 2013; Cunill 2013; Graham 2011; King 2006; Lim 2006; Morton 2000; Perrin 2008; Prescrire 2007). Due to these concerns, it is important to have non‐pharmacological interventions for treating adults with ADHD."

"To date, no systematic review has examined the effects of CBT in adults with ADHD. The growing number of randomised controlled trials assessing the efficacy of CBT for this population (Knouse 2008) suggest that this review is timely."

"To our knowledge, three systematic reviews have compared the effects of CBT in adults with ADHD (Jensen 2016; Knouse 2017; Young 2016). However, there are important methodological differences between them, also with respect to our review. Both Jensen 2016 and Young 2016 employed more restrictive criteria for defining CBT treatments that excluded relevant CBT variants such as mindfulness‐based cognitive therapy and dialectical behavioural therapy. Knouse 2017 did not report grades of quality of evidence of the included studies."

Types of interventions

We considered the following comparisons:

Monotherapy

1. CBT versus control (supportive psychotherapies, placebo interventions, waiting list or no treatment)

2. CBT versus usual treatment (other specific psychotherapies for ADHD)

Combined therapy

1. CBT combined with pharmacotherapy versus pharmacotherapy alone

We considered the following comparisons:

1. "CBT versus unspecific control conditions (supportive psychotherapies, waiting list or no treatment).

2. CBT plus pharmacotherapy versus pharmacotherapy alone.

3. CBT versus other specific interventions (control interventions that include therapeutic ingredients specifically targeted to ADHD)."

Types of outcome measures

We considered psychometrically validated self‐report measures or those completed by an independent rater or relative. The measures were considered short‐ (up to six months), medium‐ (six months to 12 months) and long‐term (more than 12 months)."

"We considered psychometrically validated self‐report measures or those completed by an independent rater or relative.

"We presented clinical and self‐reported outcomes separately, as do most studies about this topic, because assessing ADHD is more accurate when symptom information comes from more than one source (Barkley 1998a).

"We considered the measures as short term (up to 6 months), medium term (6 months to 12 months) and long term (more than 12 months).

"We included studies that assessed at least one primary outcome or at least one secondary outcome."

"We will include studies that have assessed at least one primary or secondary outcome."

"We included studies that assessed at least one primary outcome or at least one secondary outcome."

The safety outcome 'All‐cause treatment discontinuation' was considered as secondary outcome

We considered the safety outcome 'All‐cause treatment discontinuation' to be a primary outcome.

Electronic searches

We planned to search Ovid MEDLINE

We used MEDLINE PubMed because of the availability of this interface.

Searching other resources

"Additionally, we searched dissertations and abstracts from the following.

1. Association for Behavioural and Cognitive Therapies (ABCT)

2. World Congress on ADHD, organised by the World Federation of ADHD

3. Annual Meeting ‐ American Psychiatric Association (APA)"

"Additionally, we searched dissertations and abstracts from the following.

1. Association for Behavioural and Cognitive Therapies (ABCT) Convention, 2008 to 2017 (www.abct.org/Conventions/?m=mConvention&fa=PastFutureConvention).

2. World Congress on ADHD, organised by the World Federation of ADHD, 2007 to 2017 (www.adhd‐federation.org/congresshistory).

3. Annual Meeting ‐ American Psychiatric Association (APA), 1973 to 2016 (www.psychiatry.org/psychiatrists/search‐directories‐databases/library‐and‐archive)."

Assessment of risk of bias in included studies

"We independently evaluated the risk of bias using EROS. This process followed the six criteria described in Table 8.5.d, "Criteria for judging risk of bias in the 'Risk of bias' assessment tool" of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) (See Table 1)."

"We evaluated the risk of bias in each included trial using the seven criteria described in Table 8.5.d ('Criteria for judging risk of bias in the "Risk of bias' assessment tool") of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Two review authors (PL and FT) independently assessed each included study as being at low, high or unclear (uncertain) risk of bias for each domain, using EROS software (Ciapponi 2011; Glujovsky 2011; Glujovsky 2010); see Table 1. If there were discrepancies between their assessments, and the two review authors were unable to reach a consensus, a third review author (AC) joined the decision‐making process. All three review authors discussed the issue and made a final decision."

Assessment of heterogeneity

The bands that we reported for I² were:

1. "0% to 30%: may not be important

2. 30% to 60%: may represent moderate heterogeneity

3. more than 60%: may represent substantial or considerable heterogeneity"

The bands that we reported for I² were:

1. "0% to 40%: might not be important.

2. 30% to 60%: may represent moderate heterogeneity.

3. 50% to 90%: may represent substantial heterogeneity.

4. 75% to 100%: considerable heterogeneity."

Data synthesis

"When we considered studies to be sufficiently homogenous in terms of participants, interventions and outcomes, we synthesised the results in a meta‐analysis using RevMan (RevMan 2014)."

"We synthesised the results in a meta‐analysis using Review Manager 5 (RevMan 5) when we considered studies to be sufficiently homogenous in terms of population (regarding sex, age and diagnosis), interventions (comparable modalities of CBT) and comparisons (as a monotherapy or a part of a combined treatment) to avoid clinical heterogeneity, and in terms of outcome measurement methods to avoid methodological heterogeneity (RevMan 2014). Two authors assessed homogeneity independently and solved discrepancies by consensus."

We did not include a subsection about summary of findings

We included a 'Summary of findings table' subsection:

"We prepared a 'Summary of findings' table for our three main comparisons (see Types of interventions) according to GRADE methodology (Atkins 2004; Guyatt 2011), using GRADEpro GDT software (GRADEpro 2015). We included our primary outcome, the core symptoms of ADHD (self‐, clinician‐ or observer‐reported), in the tables.

"Two review authors (AC and PL) independently assessed the quality of the evidence as high, moderate, low or very low, downgrading the rating according to the presence of the following criteria: study limitations, in which we considered the studies' 'Risk of bias' level; imprecision; inconsistency of results; indirectness of evidence; and likely publication bias.

"To assess the magnitude of effect for continuous outcomes, we used the criteria suggested in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011): 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect."

Subgroup analysis and investigation of heterogeneity

We considered the type of CBT as a possible subgroup analysis

We did not include the type of CBT as a possible subgroup analysis after we redefined the comparisons.

Sensitivity analysis

For this review, we had planned to undertake sensitivity analyses to determine the effect of restricting the analysis to: "(a) only studies with low risk of selection bias (associated with sequence generation or allocation concealment), (b) only studies with low risk of performance bias (associated with issues of blinding), and (c) only studies with low risk of attrition bias (associated with completeness of data). In addition, we will assess the sensitivity of findings to any imputed data within a study."

"[W]e undertook sensitivity analyses to determine the effect of removing from the analysis: studies with high risk of selection bias (associated with sequence generation or allocation concealment); studies with high risk of performance bias (associated with issues of blinding); and studies with high risk of attrition bias (associated with completeness of data). In addition, we assessed the sensitivity of findings to any imputed data within a study."

Effects of interventions

The transformation of the continuous results to relative percentage changes had not been foreseen.

We included the transformation of continuous results to relative percentage changes in Appendix 2.