2.1. Groups contributing to the guideline development process

Lists of all members of the Guideline Development Group (GDG), External Review Group (ERG), systematic review teams, modelling teams and other contributors are provided in Annex 1, with details of their expertise and affiliations. The WHO Secretariat consisted of staff from various relevant WHO departments, and staff from the International Agency for Research on Cancer (IARC). The Steering Group of the WHO Secretariat led the coordination of the development of this guideline. Members of the Secretariat who were not part of the Steering Group were kept informed of the guideline development process and participated in the discussions, in particular during meetings of the various teams.

The GDG was established during the first half of 2019 to appraise the recommendations in the previous 2013 edition of the guideline (12), prioritize the key PICO questions for which systematic reviews needed to be updated or developed, provide feedback on the evidence reviews, and make recommendations to be presented in the final guideline. There were 52 GDG members (34 women, 18 men), representing all six WHO regions as well as civil society organizations and women’s groups, and women living with HIV. The members brought varied expertise on cervical screening and treatment. Two members acted as co-chairs and moderated the GDG meetings. The WHO Steering Group met regularly with the GDG chairs, the guideline methodologists, and the systematic review and modelling teams to review progress and to ensure evidence presentations and discussions were standardized.

An External Review Group (ERG) was also established. Its 18 members, none of whom was also a member of the GDG, had expertise in research, policy development, programme implementation and clinical care. Once the GDG had agreed on the recommendations, the ERG reviewed the full draft of the guideline and provided feedback.

There were multiple teams preparing evidence:

▪

five teams did the evidence reviews

▪

one team developed two mathematical models

▪

one team ran a survey about the feasibility of screening approaches

▪

one team surveyed women about their values and preferences (Annex 2).

The teams were based at different institutions and worked independently to prepare and present evidence during the GDG meetings. A guideline methodologist with experience of using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach (19) coordinated the presentation of evidence and decision-making processes that facilitated the development of the recommendations, as stipulated in the WHO handbook for guideline development (18).

The WHO Steering Group maintained close communication with the GDG and systematic review teams using multiple platforms:

▪

Zoom Meetings;

▪

email;

▪

surveys and voting on the summaries of the evidence and recommendations using GRADEpro software;

▪

a SharePoint site for access to meeting materials, including slides and evidence summaries, and live documents for comment;

▪

a chat feature in SharePoint to encourage discussion among the GDG members.

2.2. Scoping review and appraisal of the existing recommendations

In October 2019, a subgroup of the GDG met in Geneva to review the previously published recommendations and decide which should be removed, validated, edited or updated based on new evidence, and whether any new recommendations should be made for new interventions. This process was informed by a scoping review of the literature and an assessment of changes in disease burden, practice and policy. These decisions were circulated to all members of the GDG for feedback, and agreement on which recommendations to keep, update and add was reached after additional virtual meetings and electronic correspondence.

The scoping document was initially split between screening and treatment recommendations for women living with HIV (approved by the GRC in September 2019) and the general population of women (approved by the GRC in January 2020). The two scoping documents were then merged and subsequently approved by the GRC in August 2020.

2.3. Priority questions for review of evidence

The GDG identified 14 overarching questions, framed using the population (P), intervention (I), comparator (C), outcomes (O) (PICO) format, as a starting point for formulating recommendations applying to the general population of women and women living with HIV (Table 2.1).

Table 2.1

PICO questions for the recommendations in women (the general population of women and women living with HIV).

2.4. Priority algorithms

Since screening and treatment can be done using different primary screening and triage tests, there are numerous possible combinations or algorithms. In December 2019, GDG members were surveyed to prioritize the screening and/or triage tests and the treatments that should be evaluated. Following this prioritization exercise, a subgroup of GDG members met to review the results from the survey and to agree on the algorithms to be prioritized. They reached a consensus to address seven priority algorithms in this first phase of the guideline update (Table 2.2; for detailed algorithms please refer to Annex 4).

Table 2.2

The seven algorithms considered.

2.5. Outcomes

The GDG agreed that the outcomes previously identified in the 2013 screening and treatment guideline (12) would also be the critical outcomes for the new PICO questions; the critical outcomes are listed in Table 2.3. To ensure coherence in the systematic reviews and modelling, a working group (subgroup of the GDG) developed standardized definitions for these outcomes (see Annex 5). After reviewing the evidence and modelling a limited number of outcomes, the GDG agreed to consider all outcomes together. Adverse events were defined as outcomes that were a direct consequence of pre-cancer treatment and were grouped as one category, with the exception of preterm birth, which was considered a critical outcome (see Table 2.3).

Table 2.3

Critical outcomes for the screening and treatment recommendations.

2.6. Syntheses of evidence

Evidence was synthesized for each PICO question according to the methods in the WHO handbook for guideline development and the Cochrane handbook for systematic reviews of interventions (18,20). The literature review performed for the development of the IARC handbooks of cancer prevention: cervical cancer screening, Vol. 18 (to be published in 2021; referred to in brief throughout this guideline as “IARC handbook”) (21) was also part of the evidence synthesized for the development of this guideline. We used a hierarchical approach to avoid the duplication of reviews that had been previously published. First, we searched for pre-existing systematic reviews to update (including the systematic reviews published at the time of the previous guideline), and then searched for primary studies (including randomized and non-randomized studies) when no systematic reviews were available.

New systematic reviews or updates of systematic reviews were conducted to determine the effects of interventions (including screening tests) on outcomes, and the accuracy of screening tests in the general population of women and in women living with HIV. These systematic reviews and associated details are listed in Table 2.4.

Table 2.4

PICO questions with corresponding systematic reviews and reports, evidence-to-decision (EtD) tables and recommendations or good-practice statements.

2.6.1. Methods used for systematic literature reviews

The detailed methods for each review are reported in the Annex A, Supplementary Materials (Table 2.4). In brief, the systematic review teams applied the following key methods across all systematic reviews for this guideline:

▪

develop a systematic review protocol with inclusion and exclusion criteria for studies based on the finalized PICO questions;

▪

search multiple databases (including MEDLINE, Embase and the Cochrane Library Epistemonikos) and clinical trial registries, contact investigators in the field for potentially relevant systematic reviews, and look at randomized and non-randomized trials to identify studies for new reviews or to update existing reviews;

▪

select literature reviews or studies based on inclusion and exclusion criteria (in duplicate or by one reviewer and verified by a second);

▪

extract data (in duplicate or by one reviewer and verified by another) on the benefits and harms (effects) of screening and treatment, the accuracy of screening tests used, end-user values and preferences, equity, feasibility, resources and acceptability;

▪

contact study authors for missing data or individual patient data when appropriate;

▪

assess the risk of bias in individual studies when available (in duplicate or by one reviewer and verified by another) using an appropriate risk-of-bias tool (e.g. Cochrane Risk of Bias for randomized controlled trials, ROBINS-I tools for non-randomized studies and QUADAS for diagnostic studies);

▪

synthesize the results (narratively or quantitatively) or analyse individual patient data when available;

▪

assess the certainty of the evidence using GRADE methodology (19); the levels of certainty used are summarized in Table 2.5.

Table 2.5

Interpretation of the GRADE levels of certainty of evidence.

2.7. Development of the recommendations

All the GDG meetings that focused on formulation of recommendations were held virtually. Tables to facilitate decision-making for recommendations – evidence-to-decision (EtD) tables – were produced by the guideline methodologist for each recommendation and circulated to the GDG members before each meeting. These tables included a summary of the evidence (benefits and harms), relevant values and preferences information, and other issues, including use of resources and cost, feasibility, equity and acceptability.

During the meeting, the EtD tables and evidence were discussed with the GDG. Following the meeting, all GDG members received an email through GRADEpro that solicited direct individual input. Each GDG member saw the EtD tables several times and had opportunities to ask questions and to comment both during and after the meeting. The methodologist, systematic reviewers, modellers and the WHO Steering Group assessed the GDG input and used it to write the recommendations.

Agreement on the recommendations was made by consensus during the GDG meetings, and the final written recommendations were then approved electronically. The responses solicited via email were either to approve, approve “with the following remarks” or not approve. The GDG had agreed that, if consensus could not be reached, a majority vote of 51% would have been accepted to make recommendations – yet the group did reach a consensus on all the recommendations.

Strong recommendations (worded as “WHO recommends”) were made when all the desirable consequences of the intervention clearly outweighed the undesirable consequences in most settings.

Conditional recommendations (worded as “WHO suggests”) were made when the desirable consequences of the intervention probably outweighed the undesirable consequences in most settings.

Table 2.6 describes how strong and conditional recommendations should be interpreted.

Additionally, the GDG provided good practice statements when it agreed that this guidance was needed, but a review of the literature was not warranted because the balance of desirable and undesirable consequences of an intervention was unequivocal, and no other criteria needed to be considered.

Table 2.6

Interpretation of strong and conditional recommendations.

2.8. Management of the external peer review

The draft guideline document was circulated to the External Review Group (ERG) for comment. The WHO Secretariat prepared a summary table with all ERG responses and sorted the comments by topic or section. The WHO Secretariat then identified comments for discussion and presented these to the GDG, and when these issues had been resolved via email correspondence, the guideline document was finalized.