Types of studies

All randomised controlled trials (RCTs) in which a non‐pharmacological pain intervention is compared concurrently to a placebo/no intervention or another non‐pharmacological pain intervention will be included. If an RCT compares pharmacological and non‐pharmacological interventions to a placebo/no intervention, the study will be included but only the data for a non‐pharmacological intervention will be used.

Types of participants

The criteria for recruiting participants are as follows.

• Patients with any type of orthodontic treatment.

• All age groups will be considered.

• Trials assessing pain relief following orthognathic surgery and/or dental extractions in combination with orthodontic treatment will be excluded.

Types of interventions

The following active interventions to alleviate pain will be assessed.

• Low level laser therapy (LLLT).

• Transcutaneous electric nerve stimulation (TENS).

• Vibrating systems.

• The use of bite wafers.

• The use of chewing gum.

• The application of ice/cryotherapy.

• Acupuncture/acupressure.

• Psychological intervention, which includes verbal follow‐up and re‐assurance in the form of a structured telephone call.

Any combination of these interventions at any time during treatment will be evaluated.

Only studies in which interventions were given once moderate or severe pain developed will be included. Studies with patients receiving pre‐emptive intervention(s) will be excluded.

Types of outcome measures

Electronic searches

For the identification of studies included or considered for this review, a detailed search strategy will be developed for each database searched.

For the MEDLINE search, the subject search will be run with the Cochrane Highly Sensitive Search Strategy (CHSSS) for identifying randomised trials in MEDLINE: sensitivity‐maximizing version (2008 revision) as referenced in Chapter 6.4.11.1 and detailed in box 6.4.a. of the Cochrane Handbook for Systematic Reviews of Interventions version 5.1.0 (updated March 2011) (Higgins 2011).

The following databases will be searched:

• Cochrane Oral Health Group's Trials Register

• Cochrane Pain, Palliative and Supportive Care Group's Trials Register

• Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library current issue)

• MEDLINE (from 1966 to present) (Appendix 1)

• EMBASE (from 1980 to present).

Searching other resources

Management of records produced by the searches

All references will be downloaded into the EndNote reference management software and merged to produce a single database to facilitate retrieval of relevant articles. Non‐electronic references, that cannot be downloaded, will be entered into the database manually after which duplicates will be removed.

Selection of studies

Two review authors (Hardus Strydom (HS) and Piotr Fudalej (PF)) will independently assess the titles and abstracts (when available) of all potential relevant studies identified by the search strategy. For studies with insufficient information in the title and abstract to make a clear decision to exclude, or studies where there is disagreement between the review authors about eligibility, a full report will be obtained. These full reports will then be assessed independently by the two review authors (HS and PF) to establish whether or not the studies meet the inclusion criteria. Disagreements will be resolved by discussion between HS and PF or the involvement of a third review author (Christos Katsaros (CK)). A record of all decisions made about the potentially eligible studies will be kept. Full reports will also be obtained for those studies that will definitely be included in this review. The review authors will not be blinded to trial author(s), institution or site of publication.

Data extraction and management

Data will be extracted independently by the two review authors (HS and PF), using a pre‐designed data collection form, and saved electronically. The data collection form will be pilot tested for clarity, relevance and completeness. The instruction for coding and meanings will be placed adjacent or near to the data field that is to be coded. The coding of the form with a revision date or version number will be applied to ensure ease of update. Any unpublished information, if used, will be written and coded with a specified remark. The comparison of extracted data will be done by each author independently. Any disagreement will be discussed with a third review author (CK), and the final decisions documented.

The following data will be extracted.

i. Trial methods/characteristics.

• Study design (RCT).

• Published/unpublished.

• Conducted in (country).

• Number of centres involved.

• Recruitment period.

• Duration of the trial (including duration of follow‐up).

• Funding source.

• Inclusion criteria.

• Exclusion criteria.

• Sample size.

• • Was the sample size calculation described.

  • Number of patients randomised.

  • Number of patients at the start of the trial.

  • Number of patients evaluated.

ii. Intervention.

• Type of intervention (active and control).

• Mode of intervention administration (active and control).

• Duration of intervention (active and control).

• Dose/intensity of intervention (active and control).

• Frequency of intervention (active or control).

iii. Primary outcome(s) measured.

• Outcome(s) measured.

• Method used to measure the outcome.

• Mean scores (active and control).

• Standard deviation (active and control).

• Time(s) these were measured.

iv. Secondary outcome(s) measured.

• Secondary outcome(s) measured.

• Method used to measure the outcome.

• Mean scores (active and control).

• Standard deviation (active and control).

• Time(s) these were measured.

v. Participant characteristics.

• Age at the start of the trial (mean and range).

• Culture/ethnicity.

• Gender.

• Psychosocial well‐being.

• Previous pain experience.

Assessment of risk of bias in included studies

The two review authors (HS and PF) will independently evaluate the risk of bias for the included studies, using the following domains (as described in the Cochrane Handbook for Systematic Reviews of Interventions version 5.1.0 (Higgins 2011)): random sequence generation (selection bias); allocation concealment (selection bias); blinding of participants and personnel (performance bias), and outcome assessors (detection bias); incomplete outcome data (attrition bias); selective outcome reporting (reporting bias) and other potential sources of bias. The review authors will follow the Cochrane Collaboration's tool for assessing risk of bias in each included study. The first part of the tool involves describing what was reported to have happened in the study and the second part involves assigning a judgment relating to the risk of bias for each entry. The judgment for each entry is categorized as low risk of bias, high risk of bias, or unclear risk of bias.

An assessment of the overall risk of bias will be summarized after considering the relative importance of the different domains.

• Low risk of bias ‐ low risk of bias for all domains.

• Unclear risk of bias ‐ unclear risk of bias for one or more domains.

• High risk of bias ‐ high risk of bias for one or more domains.

Disagreements will be resolved through consensus, or referred to arbitration by the Cochrane Oral Health Group if needed.

Measures of treatment effect

Because patients generally either respond to an analgesic intervention or they do not, and because pain trials generally do not produce results with Gaussian distributions for pain relief measurements, dichotomous outcomes will be preferred. If the dichotomous outcomes are not directly reported in the studies, we will contact the authors and request the data needed to calculate the proportion of patients with at least 50% reduction in pain from baseline. This will be a predefined clinically useful dichotomous outcome.

We will calculate the following.

• Risk ratios (RR) and the number needed to treat (NNT) with 95% confidence intervals (CI) for dichotomous outcomes.

• Mean differences (MD) with 95% CI for continuous data if it is impossible to dichotomize outcomes.

• Standardized mean differences (SMD) with 95% CI for studies that use different numerical scales to measure the same outcome if it is impossible to dichotomize outcomes.

Outcomes in the treatment and control groups will be assessed as follows.

• Pain intensity will be assessed as RR (or MD, if dichotomization of the outcome is impossible).

• Dose/intensity and frequency of pain relief needed will be assessed as RR (or MD, if dichotomization of the outcome is impossible).

• Any rescue (alternative pain relief taken/prescribed, including dose and time, following the last treatment) will be assessed as MD).

• Adverse effects of the intervention will be assessed as RR.

• Quality of life and/or patient satisfaction will be assessed as RR (or MD, if dichotomization of the outcome is impossible).

Dealing with missing data

We will contact trial authors for clarifications or missing information. Where data are unavailable, this will be resolved by estimation on data using the methods described in the section for calculating missing standard deviations of the Cochrane Handbook for Systematic Reviews of Interventions, version 5.1.0 (Higgins 2011). Imputations will be restricted to standard deviations. We will state the assumptions of any methods used to cope with missing data and discuss the potential impact it might have on the findings of the review. An intention‐to‐treat analysis will be conducted wherever it is possible.

In case of the use of 'rescue medication', we will use two imputation methods to calculate estimate of pain relief.

• Baseline observation carried forward (BOCF) ‐ the pain relief score is set to zero for all remaining time points from rescue medication until the end of the observation period.

• Last observation carried forward (LOCF) ‐ the last pain relief measurement, at the observation immediately preceding re‐medication, is used for all remaining assessments.

The most appropriate estimate of pain relief would be between BOCF and LOCF.

Assessment of heterogeneity

We will assess heterogeneity using the I² statistic, where I² values will be interpreted as follows (Higgins 2011):

• 0% to 40%: might not be important;

• 30% to 60%: may represent moderate heterogeneity;

• 50% to 90%: may represent substantial heterogeneity;

• 75% to 100%: considerable heterogeneity.

Assessment of reporting biases

We will address publication bias calculating how much data (trials, participants) would be required both to be unpublished and to have zero treatment effect (relative risk/benefit of 1) to make any result clinically irrelevant. In performing this assessment we will follow the guidelines of the Cochrane Pain, Palliative and Supportive Care Review Group (Cochrane PaPaS 2011).

Data synthesis

Meta‐analysis will only be conducted if there are studies of similar comparisons reporting the same outcome measures. Risk ratios will be combined for dichotomous data, and mean differences for continuous data, using the random‐effects model provided there are more than three studies in the meta‐analysis.

The characteristics and results from all included studies will be summarized qualitatively in tabular form.

Subgroup analysis and investigation of heterogeneity

In case of significant heterogeneity, the following subgroup analyses will be conducted to explore its source.

• Type of interventions.

• Dose/intensity of interventions.

• Characteristics of participants: age, gender, ethnicity, psychological well‐being and previous pain experienced.

• Type of orthodontic appliance used.

Sensitivity analysis

Providing there are sufficient included trials, sensitivity analysis based on risk of bias (low risk of bias versus high or unclear risk of bias) will be done to investigate the robustness of conclusions to decisions made during the review process.