The process outlined in the WHO handbook for guideline development (WHO, 2012b) was followed. This included: (1) identifying priority clinical questions and outcomes; (2) retrieving the evidence; (3) assessing the quality of evidence and synthesizing the findings; (4) formulating recommendations, including future research priorities; and (5) planning for dissemination, implementation, impact evaluation and updating the guideline.
The GRADE method was followed to prepare evidence profile tables related to preselected topics, based on up-to-date systematic reviews. The Guideline Development Group comprised content experts, methodologists and representatives of potential stakeholder groups. Some of these experts participated in a WHO technical consultation held in New York City, United States of America, on 17–18 November 2014 to scope questions for the systematic reviews and guideline update.
The full Guideline Development Group met in Geneva, Switzerland on 21–23 October 2015, to review and discuss the evidence, draft the recommendations and agree on their strength, taking into consideration: (1) the desirable and undesirable effects of this intervention; (2) the quality (confidence in estimates of effect) of the available evidence; and (3) the values and preferences related to the intervention as well as to outcomes and specific contextual factors that pertain to various settings. The cost of the options available to health-care workers in various settings was not formally assessed because of lack of primary data in the literature or elsewhere, but the cost and resource implications were considered as part of a general discussion by the Guideline Development Group.
An additional group of external experts and stakeholders reviewed the proposed recommendations following their drafting. All Guideline Development Group members submitted declarations of interest forms before each meeting and also made verbal declarations of interest at the beginning of meetings.
Evidence retrieval, assessment and synthesis
The evidence retrieval process for the priority questions followed the standard outlined in the WHO handbook for guideline development (WHO, 2012b). The acknowledgements provide a list of reviewers. A protocol for each systematic review was developed and included the search terms and strategy and the populations, interventions, comparators and outcomes used to define the inclusion and exclusion criteria. The detailed search strategy for each priority question was agreed on after a series of discussions with the WHO Steering Committee and lead investigators of each review. Each review includes a flow diagram showing the numbers of studies excluded and included. Medline and EMBASE databases were used to identify peer-reviewed publications. The Cochrane Central Register of Controlled Trials, the International Standard Randomised Controlled Trial Number Registry and ClinicalTrials.gov were searched for ongoing studies. The quality of the evidence for each priority question was assessed using the GRADE method (GRADE Working Group, 2016). The quality of the evidence for treatment interventions was graded as high, moderate, low or very low based on the definitions in the WHO handbook for guideline development (WHO, 2012b). The GRADE tables were prepared using the GRADE profiler software (GRADEPro), when appropriate. The reviews are available through URLs as in Annexes 1a, 1b, 1c and 2 and will be published. Based on these reviews, the WHO Steering Committee proposed an initial set of draft recommendations.
WHO Steering Committee
A Steering Committee, with members from the Department of Maternal, Newborn, Child and Adolescent Health, Department of Nutrition for Health and Development and the Department of HIV, has overseen the guideline review process. The acknowledgements list the WHO staff members on the Committee.
Guideline Development Group
WHO convened a 21-member Guideline Development Group consisting of internationally recognized experts in terms of content, methods and regional representation. The acknowledgements list the members. Members were tasked with reviewing and evaluating the quality of the evidence identified through the systematic reviews using the GRADE method (described below) and revising and finalizing the guideline recommendations.
External Peer Review Group
Members of the External Peer Review Group were asked to review the recommendations developed by the Guideline Development Group to ensure that there were no important omissions, contradictions or inconsistencies with scientific evidence or programmatic feasibility and to assist in clarifying the language, especially in relation to implementation and how policy-makers and programme staff might read them.
No additional recommendations were invited from the External Peer Review Group. The WHO Steering Committee collated the queries raised by the External Peer Review Group and discussed them with the chairs to resolve any inconsistencies or contradictions raised.
The acknowledgements list the members, from various countries and disciplines, with their affiliations.
Managing conflicts of interest
All members of the Guideline Development Group, systematic review teams and members of the External Peer Review Group were required to sign and submit a declaration of interests prior to their participation in the meetings. The WHO Steering Committee reviewed the declarations before the Guideline Development Group meeting to determine whether any of the proposed members had a conflict of interest that might have precluded or limited his or her participation. Although the WHO Steering Committee did not identify any conflicts of interest requiring any action, the potential conflicts of interest declared by the Guideline Development Group members are summarized next.
Lesley Bamford works at the National Department of Health, South Africa and declared to have a public position regarding breastfeeding, as stated in the Tshwane Declaration. As an employee of the National Department of Health, she explains, defends and advocates such a position. It was agreed that she could participate fully in the deliberations and decision-making on these recommendations.
Louise Kuhn declared her involvement in projects related to oligosaccharides and cytokines in breast milk, early antiretroviral therapy in neonates and follow-up of children living with HIV, which received grants from the National Institute of Dental and Craniofacial Research, the United States National Institutes of Health and the Eunice Kennedy Shriver National Institute of Child Health and Human Development. It was agreed that she could participate fully in the deliberations and decision-making on this recommendations
Angela Mushavi declared being employed by the Elizabeth Glaser Pediatric AIDS Foundation and seconded to the Ministry of Health and Child Care of Zimbabwe as national coordinator of services for preventing the mother-to-child transmission of HIV and HIV care and treatment for children. In this role, she discusses infant and young child feeding as a child survival strategy and also recognizes and articulates the benefits of breastfeeding, even among women living with HIV. It was also declared that the outcome of this meeting is of importance to the Ministry of Health and Child Care. It was agreed that she could participate fully in the deliberations and decision-making on these recommendations.
Anju Seth declared her role as a co-investigator of a project for the nutritional needs of children made vulnerable by HIV funded by UNICEF. At the time of this meeting, the project was scheduled to be completed by November 2015. It was agreed that she could participate fully in the deliberations and decision-making on these recommendations.
Grading the quality of the evidence
The Guideline Development Group used the GRADE method (GRADE Working Group, 2016) to evaluate the quality of the evidence. The GRADE method is a widely used, standardized approach for characterizing the quality of evidence and distinguishing between the quality of the evidence and the strength of the recommendations (WHO, 2012b). GRADE tables summarize details about the studies included in the systematic review, including study outcomes, limitations (risk of bias), possible inconsistency, indirectness, imprecision and other factors that might affect judgements of the quality of the evidence. Guideline Development Group members then used the information to define the overall quality of the evidence as very low, low, moderate or high as defined below.
Table 8Definition of the quality of the evidence using the GRADE method
View in own window
| QUALITY | DEFINITION | IMPLICATION |
|---|
| High | The Guideline Development Group is very confident that the true effect lies close to that of the estimate of the effect. | Further research is very unlikely to change confidence in the estimate of the effect. |
| Moderate | The Guideline Development Group is moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. | Further research is likely to have an important impact on confidence in the estimate of the effect and may change the estimate. |
| Low | Confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the true effect. | Further research is very likely to have an important impact on confidence in the estimate of the effect and is unlikely to change the estimate. |
| Very low | The Guideline Development Group has very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of the effect. | Any estimate of effect is very uncertain. |
In general, evidence based on randomized controlled trials is given a high-quality rating, and evidence from observational studies is given a low-quality rating. These initial ratings may be adjusted by the following factors:
study limitations or considerations on the risk of bias such as concealment, blinding, attrition and detection bias;
consistency: whether the results from the studies are similar and in the same direction of effect;
directness: whether the population, intervention or comparator are the same as for the clinical question the guideline is considering;
precision: whether data arise from a large sample with high event rates as represented in the breadth of confidence intervals; and
reporting or publication bias: whether the underlying beneficial or harmful effect is systematically underestimated or overestimated because of the selective publication of studies or selective reporting of outcomes.
In non-randomized studies, additional adjustment factors include dose–response gradients, the directional effect of plausible bias and the magnitude of effect.
The authors of one systematic review (Annex 1a) also developed a modified Newcastle-Ottawa Scale to assess the quality of all studies included in the analysis (Wells et al., 2009).
Decision-making process
For each draft recommendation, the WHO Steering Committee presented a synthesis of the evidence, the GRADE tables and the draft null recommendation language. The decision-making tables were drafted, including the benefits and risks of the interventions from a public health perspective; the values, preferences and acceptability to mothers living with HIV and their communities as well as programme managers, policy-makers and health-care providers; and the feasibility of implementing any recommendations (including the resources needed, focusing on national programmes in resource-limited or other settings).
Information from a survey of national health authorities from the 22 priority countries of the Global Plan (UNAIDS, 2011) and from an online survey of other stakeholders was used to inform the values and preferences (Annex 3).
The cost of options available to health-care workers in various settings was not formally assessed because of a lack of primary data in the literature or elsewhere. However, the Guideline Development Group considered the cost implications as part of the general discussion. Comments were therefore restricted to personal experiences and extrapolations from general cost considerations of programmes.
Each Guideline Development Group member was asked to review the material and independently comment on and suggest revisions to the proposed guidance and decision-making tables. They were requested to rank the overall quality of the evidence using the GRADE method (independent of the rating made in the synthesis of the evidence), the balance of benefits versus harm, the values that should be considered in making a recommendation and the applicability of any proposed recommendations to the populations for whom they are intended. Finally, they were asked to assess what strength each recommendation should be given based on the criteria provided in .
Criteria for assessing the strength of recommendations.
The Guideline Development Group used a consensus-building process to finalize the recommendations. Once participants expressed their opinions and suggestions on a recommendation, the chairs and the WHO Steering Committee summarized this information. This summary was presented to the Guideline Development Group members to gauge the degree of consensus and where differences existed. The chairs of the Guideline Development Group facilitated discussions among Guideline Development Group members until there was consensus on the language of each recommendation, the quality of the evidence and the strength. If consensus could not be reached, the Guideline Development Group had agreed at the beginning of the meeting that a simple majority vote would determine a contested decision.
WHO staff members did not express personal opinions on the data, in the discussions or in the decisions on language, the strength of the recommendations or the quality of the evidence. Throughout the meeting, WHO staff members articulated the principles and guidelines of the WHO decision-making process.
The Guideline Development Group reached agreement on all the recommendations following revisions of the text. The recommendations on the duration of breastfeeding by mothers living with HIV were discussed extensively to achieve consensus on the strength of the recommendation. Voting was only used as a straw-poll method to evaluate consensus, and no decisions were required to be subjected to a final vote.
The Guideline Development Group declined to make a recommendation in all cases, but instead formulated guiding practice statements using the same methods.
The WHO Steering Committee circulated the draft recommendations to an External Peer Review Group, which made several suggestions to improve the document. The WHO Steering Committee reviewed all suggestions and incorporated the comments as appropriate following discussion and agreement with the chair and co-chair of the Guideline Development Group. No new recommendations were considered in this round of comments. The Guideline Development Group approved the final version.
