Types of studies

We included all randomised controlled trials (RCTs) in which participants who were provided with one‐to‐one oral hygiene advice (OHA) by a member of the dental team in the dental setting were compared to participants not provided with OHA. We also included RCTs which compared different formats (e.g. DVD, leaflet, video, etc.) of one‐to‐one OHA provided within the dental setting or OHA provided by different members of the dental team within the dental setting. We only included studies with a minimum follow‐up period of 8 weeks. The unit of randomisation could be individual participants, clusters of individuals, or dental setting.

Types of participants

We included studies that involved children (with or without guardian/parent) or adult participants who were fit and healthy or groups of individuals who had well defined medical conditions. Studies involving participants who wore a removable prosthesis or orthodontic appliance of any kind were excluded.

Types of interventions

We included studies that involved OHA being provided by any member of the dental team within the dental setting environment (i.e. within a general dental practice, community dental setting or dental hospital setting) on a one‐to‐one basis in which the comparator was no advice or advice in an alternative format or advice from a different member of the dental team. Dental team members included dentist, dental nurse, dental hygienist and dental therapist. Studies were included if the OHA took place alongside an intervention aiming to change dietary behaviour or smoking behaviour, although these multi‐intervention studies were planned to be subjected to a subgroup analysis if any such studies had met the inclusion criteria.

We excluded any studies where additional measures were provided to the intervention group but not the control group (e.g. dental prophylaxis, dental scaling, application or provision of fluoride containing products, provision of toothbrushes, etc.).

Types of outcome measures

Outcome measures considered in this review included clinical status, patient‐centred and economic factors. The outcome measure must have been assessed at least 8 weeks following the intervention.

Searching other resources

The following trial registries were searched for ongoing studies:

• US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (clinicaltrials.gov; searched 10 November 2017) (Appendix 5);

• World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch; searched 10 November 2017) (Appendix 6).

Bibliographic references of identified RCTs and articles identified as relevant to the review retrieved in the electronic search were also searched (e.g. review articles and systematic reviews). In addition, emails were sent to authors of identified potentially eligible RCTs asking them for other known unpublished or ongoing research. Relevant sections of the reports were translated of non‐English publications identified from title, abstract, or keyword, to be of interest.

Selection of studies

The search was designed to be sensitive and included controlled clinical trials, these were filtered out early in the selection process if they were not randomised.

At least two review authors reviewed and analysed independently all reports identified by the search on the basis of title, keywords, and abstract (where this was available) to see if the study was likely to be relevant. Where it was not possible to classify an article based on title, keywords and abstract, we obtained the full article. The full report was obtained of all potentially eligible studies, and also of studies where the title/abstract provided insufficient information to make a decision on eligibility.

The review authors were not blinded with respect to report authors, journals, date of publication, sources of financial support, or results. The inclusion criteria, as outlined above, were applied and those studies deemed suitable for inclusion by at least two review authors were included. We contacted study authors for missing information or clarity regarding methods, results, etc. where required and feasible. We linked multiple publications of the same study under one single study title. Those studies excluded are cited with reasons for exclusion reported.

Two review authors independently and in duplicate assessed the eligibility of the non‐English language reports. Relevant sections of the reports were translated with the assistance of Cochrane Oral Health and non‐English reports that met the inclusion criteria had data extraction and risk of bias completed.

Data extraction and management

A data extraction form was designed and piloted prior to full use. At least two review authors independently extracted data from each of the potentially eligible studies; any disagreements between the two review authors undertaking data extraction was resolved by discussion and if necessary the involvement of a third review author. If necessary we contacted the study author, where possible, to clarify any unclear or inadequate characteristics before a final decision on inclusion was made. Data recorded included the following:

• general study information: authors, title, year research completed, year research published, country, ethics and consent process, financial support and conflicts of interest, country, contact address;

• methods: research objective, sample size calculation, allocation procedures, follow‐up period, degree of blindness in outcome assessment;

• participants: inclusion/exclusion criteria, age, medical factors, baseline periodontal health/caries status/behaviours/knowledge/attitudes/quality of life;

• intervention and comparators: type of OHA, format of OHA, advice duration, advice frequency, personnel providing advice;

• outcomes: primary and secondary outcomes and outcome measures.

Assessment of risk of bias in included studies

The risk of bias of included trials for the seven domains of random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other potential sources for bias was undertaken independently and in duplicate by at least two review authors as part of the data extraction process and in accordance with the guidelines in theCochrane Handbook for Systematic Reviews of Interventions 5.1.0 (Higgins 2011). Where possible, we contacted study authors for missing information or clarification of their study methods. Any disagreements were resolved by discussion between the review authors with Cochrane Oral Health being consulted where there was continuing disagreement.

We allocated the level of bias for each domain as high or low as per the criteria below; for those domains where insufficient information was available or not described in sufficient detail to allow a definitive judgement then the level of bias was recorded as unclear.

• Random sequence generation

• • Low: the investigators describe a random component in the sequence generation process such as referring to a random number table, using a computer random number generator, etc..

  • High: the investigators describe a non‐random component in the sequence generation process. Usually, the description would involve some systematic, non‐random approach, (e.g. sequence generated by odd or even date of birth, etc.).

• Allocation concealment

• • Low: participants and investigators enrolling participants could not foresee assignment because an appropriate method was used to conceal allocation (e.g. central allocation, etc.).

  • High: participants or investigators enrolling participants could possibly foresee assignments and thus introduce selection bias, such as allocation based on using an open random allocation schedule, etc..

• Blinding of participants and personnel

• • Low: no blinding or incomplete blinding, but the review authors judge that the outcome is not likely to be influenced by lack of blinding.

  • High: no blinding or incomplete blinding, and the outcome is likely to be influenced by lack of blinding.

• Blinding of outcome assessment

• • Low: no blinding of outcome assessment, but the review authors judge that the outcome measurement is not likely to be influenced by lack of blinding; blinding of outcome assessment ensured, and unlikely that the blinding could have been broken.

  • High: no blinding of outcome assessment, and the outcome measurement is likely to be influenced by lack of blinding; blinding of outcome assessment, but likely that the blinding could have been broken, and the outcome measurement is likely to be influenced by lack of blinding.

• Incomplete outcome data

• • Low: no missing outcome data; reasons for missing outcome data unlikely to be related to true outcome (for survival data, censoring unlikely to be introducing bias); missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk not enough to have a clinically relevant impact on the intervention effect estimate; for continuous outcome data, plausible effect size (difference in means or standardized difference in means) among missing outcomes not enough to have a clinically relevant impact on observed effect size; missing data have been imputed using appropriate methods.

  • High: reason for missing outcome data likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk enough to induce clinically relevant bias in intervention effect estimate; for continuous outcome data, plausible effect size (difference in means or standardized difference in means) among missing outcomes enough to induce clinically relevant bias in observed effect size; 'as‐treated' analysis done with substantial departure of the intervention received from that assigned at randomisation; potentially inappropriate application of simple imputation.

• Selective reporting

• • Low: 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.

  • High: not all of this study's pre‐specified primary outcomes have been reported; one or more primary outcomes is reported using measurements, analysis methods or subsets of the data (e.g. subscales) that were not pre‐specified; one or more reported primary outcomes were not pre‐specified (unless clear justification for their reporting is provided, such as an unexpected adverse effect); one or more outcomes of interest in the review are reported incompletely so that they cannot be entered in a meta‐analysis; the study report fails to include results for a key outcome that would be expected to have been reported for such a study.

• Other potential sources for bias

• • Low: the study appears to be free of other sources of bias.

  • High: there is as least one important risk of bias, for example a potential source of bias related to the specific study design used.

We summarised the risk of bias as follows.

Measures of treatment effect

For continuous outcomes (e.g. plaque/gingivitis scores), where studies used the same scale, we used the mean values and standard deviations reported in the studies in order to express the estimate of effect of the intervention as mean difference (MD) with 95% confidence interval (CI). Where different scales were used, we expressed the treatment effect as standardised mean difference (SMD) and 95% CI.

For dichotomous outcomes (e.g. attachment loss/no attachment loss), we expressed the estimate of effect as a risk ratio (RR) with 95% CI. Had we included any cross‐over studies, we would have extracted appropriate data following the methods outlined by Elbourne 2002, and would have used the generic inverse variance method to enter log RRs or MD/SMD and standard error into Review Manager 5 (Review Manager 2014).

Unit of analysis issues

The participant was the unit of analysis. Had we included any cross‐over studies, these should have analysed data using a paired t‐test, or other appropriate statistical test, to take into account the paired nature of the data. Cluster‐RCTs should have analysed results taking account of the clustering present in the data, otherwise we would have used the methods outlined in Section 16.3.4 of the Cochrane Handbook for Systematic Reviews of Interventions in order to perform an approximately correct analysis (Higgins 2011).

Dealing with missing data

We attempted, where feasible, to contact the author(s) of studies to obtain missing data or for clarification. We did not use any further statistical methods or carry out any further imputation to account for missing data.

Assessment of heterogeneity

Heterogeneity was assessed by examining the types of participants, interventions and outcomes in each study. It was agreed in advance that meta‐analysis would only be attempted if studies of similar comparisons reporting the same outcome measures were included in the review.

If it had been appropriate to perform meta‐analysis, we would have assessed the possible presence of heterogeneity visually by inspecting the point estimates and CIs on the forest plots; if the CIs had poor overlap then heterogeneity would have been considered to be present. We would also have assessed heterogeneity statistically using a Chi² test, where a P value < 0.1 would have been considered to indicate statistically significant heterogeneity. Furthermore, we would have quantified heterogeneity using the I² statistic. A guide to interpretation of the I² statistic given in Section 9.5.2 of the Cochrane Handbook for Systematic Reviews of Interventions is 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

Assessment of reporting bias within studies has already been described in the section Assessment of risk of bias in included studies. Reporting biases can occur when reporting (or not reporting) research findings is related to the results of the research (e.g. a study that did not find a statistically significant difference/result may not be published). Reporting bias can also occur if ongoing studies are missed (but that may be published by the time the systematic review is published), or if multiple reports of the same study are published, or if studies are not included in a systematic review due to not being reported in the language of the review authors. If there had been more than 10 studies included in a meta‐analysis, we would have assessed the possible presence of reporting bias by testing for asymmetry in a funnel plot. If present, we would have carried out statistical analysis using the methods described by Egger 1997 for continuous outcomes and Rücker 2008 for dichotomous outcomes. However, we did attempt to limit reporting bias in the first instance by conducting a detailed, sensitive search, including searching for ongoing studies, and any studies not reported in English were translated by a member of Cochrane Oral Health.

Data synthesis

We would only have carried out a meta‐analysis where studies of similar comparisons reported the same outcomes. We would have combined MDs (we would have used SMD where studies had used different scales) for continuous outcomes, and would have combined RRs for dichotomous outcomes, using a fixed‐effect model if there were only two or three studies, or a random‐effects model if there were four or more studies.

We would have used the generic inverse variance method to include data from cross‐over studies in meta‐analyses as described in Section 16.4 of the Cochrane Handbook for Systematic Reviews of Interventions (Elbourne 2002; Higgins 2011). Where appropriate, we would have combined the results from cross‐over studies with parallel group studies, using the methods described by Elbourne 2002.

Subgroup analysis and investigation of heterogeneity

If it had been required, subgroup analyses would have been undertaken where OHA had taken place alongside an intervention aiming to change dietary behaviour or smoking behaviour, personnel providing advice and frequency of advice.

If there had been a number of similar studies, sensitivity analysis would have been considered to determine whether conclusions reached would be affected by different inclusion criteria. However, given the heterogeneity of studies included this was not required.