This review protocol was registered with the PROSPERO International Prospective Register of systematic reviews http://www.crd.york.ac.uk/PROSPERO), prior to the start of data collection and extraction.

Objective one: burden of disease from RHD, GAS infection and ARF in SA

This updated review will be designed to best control for the confounding selection and publication biases,17 18 with the findings reported using the Meta-analysis of Observational Studies in Epidemiology guidelines.19 Where available, findings of randomised controlled trials will undergo a separate meta-analysis if there is sufficient data.

Inclusion and exclusion criteria

Similar methods to the previous 2015 systematic review will be employed.20 Any study reporting on the incidence or prevalence of RHD, that had been conducted in SA and published in English with patient recruitment already completed between March 2014 and December 2022 will be considered for inclusion. Studies during this same period that investigate burden of disease from GAS infection and ARF will also be considered.

Although RHD does not affect children younger than 3 years of age, there is no age restriction for the patient population. We will include any study that estimates one or more of the following epidemiological measures of disease burden from RHD, GAS infection and ARF: incidence, prevalence, remission rate, relative risk of mortality (ie, excess mortality) or cause-specific mortality. In addition, we will consider any study of cardiovascular morbidity that quantified the attributable proportion of RHD, ARF or GAS cases, but is limited to disaggregated data pertaining to these conditions. The burden of RHD among pregnant South Africans will be included as estimates were last reported in 2014.21 22

The prevalence of RHD has previously been defined by screening programmes of subclinical disease in asymptomatic populations.23–25 Hospital-based studies, however, focus on clinical disease in symptomatic populations. As far as possible, we intend to elucidate diagnostic methods in both echocardiography screening studies and hospital-based studies. In the context of RHD, improvement in severity of valvular disease is due exclusively to medical, percutaneous or surgical intervention, requiring an exploration of the surgical literature for rates of disease regression versus long-term progression (ie, leading to mortality). Based on clinical experience, the most significant morbid outcomes of RHD include heart failure, ischaemic/thromboembolic or haemorrhagic stroke, AF, IE and valve repair or replacement.

Studies will be excluded if they focus on degenerative heart valve disease or rheumatological conditions other than RHD. Autopsy and necropsy studies will be omitted because consent rates are low for these procedures in SA, and their conclusions about mortality patterns are highly likely to be biased. Editorials, commentaries and case reports will be excluded, but their reference lists will be searched for additional references worth consideration for inclusion. While the WHO recommends considering the inclusion of disease register data in burden-of-disease studies, there is evidence that such data collection is unreliable and not representative of the general population.26 From previous experience, rheumatic fever registers tend to exclude information regarding RHD and are thus excluded from our analysis.20 Although the latest SA Demographic and Health Survey (2016) does not contain primary data on RHD, we will use the mortality data in this review.27

Search strategy

Using predefined search terms that capture relevant disease processes from GAS infection through to the sequelae of RHD (see table 1), the three largest databases relevant to the South African population will be searched: PubMed, ISI Web of Science and Scopus. The search strategy for this review is consistent with the original search strategy employed by Zühlke et al,20 but has been updated to include terms relevant to ARF and GAS and is informed by the strategy employed by Moloi et al for their 2016 systematic review of RHD in Africa.28 Additionally, to identify relevant conference proceedings, theses and abstracts, a search of the following archives at the University of Cape Town’s Health Sciences Library will be performed: Current and Completed Research (SA), SA Heart and Sabinet African Journals (which covers all South African publications, including those not currently indexed). The search process will be managed using the online Rayyan platform and search results will be exported to EndNote V.20 software as reference manager. Vital registration data from Statistics SA will also be collected. Although vital statistics can be flawed, the GBD study considers these an important source of mortality data and incorporates specific methods for handling misclassifications and inconsistencies.29 Finally, a manual snowballing search of all reference lists of studies included in the final review will be performed. Given previous experience with the 2015 review,12 a substantial amount of ‘grey’ literature on RHD is anticipated and so the database search strategy is kept intentionally broad. Contact will be made with other South African cardiovascular disease researchers and practitioners, as well as with international experts on RHD, in order to identify unpublished works or to obtain additional information. All published and unpublished data will be subject to the same quality assessment and data extraction process.

Following an independent review of the titles and abstracts of the search results from all databases by two researchers (SM and LA), a secondary review of the full-text manuscripts of potentially eligible reports will be performed. A tertiary review will be conducted, where potential discrepancies over the final inclusion list will be resolved by consensus discussion between the two primary reviewers (SM and LA), with arbitration by a third reviewer (LJZ) if required.

Quality assessment and risk of bias

All studies deemed eligible for inclusion will undergo quality assessment using study design-specific tools from the Joanna Briggs Institute.30 The purpose of these checklists is to assess the methodological quality of a study and to determine the extent to which a study has addressed the possibility of bias in its design, conduct and analysis. The quality assessment can then inform synthesis and interpretation of the results of the study within the systematic review. The number of questions vary among tools, but cover the sampling of study groups, identification and adjustment for confounders, allocation of interventions, validity and reliability of results and methods to address loss to follow-up, etc. A study must meet at least 80% of applicable criteria in order to be considered of ‘adequate quality’, but studies will not be excluded from this review based on the outcome of the quality assessment.

Data extraction

A standardised data extraction form—developed for this review and based on the template used for the original review by Zühlke et al in 201512—will be used to extract information from included articles that will be independently duplicated (ie, not split between the two authors) in order to improve reliability. The data extraction form will capture basic study characteristics, including objectives, study population, sample size, years and location of study, as well as study design. Disease-related parameters, including hospitalisation, secondary events, surgical interventions and mortality will be recorded too. Where study data are unclear, the original author of the manuscript will be contacted to clarify the findings. Where not provided, CIs will be incorporated into the formula SE=(upper limit−lower limit)/3.92 or calculated using the cii command in Stata V.17. Where not stated, RHD mortality per 100 000 will be calculated as follows: RHD deaths/mid-year population. Where population number is not stated, this will be calculated by using age-specific incidence rates and cases stated in the original paper as follows: 100 000 × (number of cases/incidence per 105).

Data synthesis and analysis

Prevalence data from individual studies will be combined according to the Mantel-Haenszel method using random-effects meta-analysis, given the anticipated heterogeneity with prevalence studies. Heterogeneity will be evaluated using the χ2-based Q statistic (significant for p<0.1) and the I2 statistic (>50% to be indicative of ‘notable’ heterogeneity).31 Meta-analysis will be conducted using the MetaProp command32 using Stata (V.17). We will perform the Freeman-Tukey double arcsine transformation to ensure stabilisation of the variance of study prevalence; the transformation is essential in minimising influence from studies with outliers before data are pooled.32 33 To minimise publication bias, we will conduct an extensive literature search to identify completed studies or ongoing studies with preliminary results. If we include more than 10 studies, we will investigate publication bias by generating a funnel plot and using Egger’s test to assess funnel plot asymmetry.34

Objective two: challenges to implementing a comprehensive national RHD programme

As conceptualised by Wyber in 2013,16 this objective will look at 24 of the 25 specific facets of a comprehensive RHD programme, excluding burden of disease (figure 1). The search strategy will be informed by the systematic review protocol published by Moloi et al,28 which identified stakeholders in RHD in Uganda and Tanzania using a mixed-methods approach that included case series, reports, editorials/commentaries and grey literature.8

Figure 1

A conceptual framework for comprehensive, register-based RHD control programmes (derived from Wyber, 2013). BPG, benzathine penicillin; GAS, Group A Streptococcus; RF, rheumatic fever; RHD, rheumatic heart disease.

Inclusion and exclusion criteria, search strategy and data management

Like objective one, all studies on RHD published between March 2014 and December 2022 (inclusive) are eligible for inclusion, with no age restriction on patient population. Studies during this same period that investigate relevant aspects of GAS infection and ARF, from preventative strategies to diagnosis and management will also be considered. The same limitation to English and human subjects will be applied. However, no restrictions in terms of study or publication type will be made as this objective requires a broad scan of the literature to ascertain information pertinent to the implementation of a national RHD programme. Articles unrelated to the topics of RHD, GAS, ARF, SA and health services will be excluded. Articles already included in burden of disease estimates (ie, objective one) will also be excluded unless they explicitly deal with the remaining 24 facets of a national RHD programme.

The same search strategy for databases as per objective one will be employed (table 1) and results will be managed via the online Rayyan platform and EndNote V.20 software as above. It is anticipated that the same initial results will be returned from each database search as per objective one. The search will be managed using the online Rayyan platform and search results will be exported to EndNote V.20 software as reference manager. The title and abstracts will be reviewed by two independent researchers (SM and LA), with HM acting as arbitrator, using the following categories:

Category one: factors influencing diagnosis.

Category two: factors influencing treatment or referral.

Category three: factors influencing adherence and retention in long-term care.

These three categories reflect the pathway of care that patients undergo from GAS infection through to the sequelae of RHD. Barriers at each point of care highlight issues that can then be captured under the five main strategic areas listed in section Objective two: challenges to implementing a comprehensive national RHD programme. For example, factors influencing diagnosis (category one) could fall under system, inputs or current clinical services which overlap, but require involvement from multiple stakeholders to effect change. All other articles will be excluded. No quality assessment or risk of bias assessment will be performed as this review is intended to provide broad comment on a national RHD strategy in SA, and all available information may be useful at this stage. SA also has a robust research community with much of the literature produced by known stakeholders and experts in the field of RHD. As such, it is anticipated that any grey literature, commentaries and editorials published in peer-reviewed journals will have input from one of these expert sources.

Data extraction, synthesis and analysis

Two reviewers will review the full text of each article and enter relevant information into a template for data extraction under the 5×5 framework for objective two. The information collected for each theme will then be combined following discussion by the two reviewers with any disputes mediated by a third reviewer.

The qualitative data will then undergo inductive analysis for overarching themes and inconsistencies, and reported under the five main strategic areas listed in section Objective two: challenges to implementing a comprehensive national RHD programme. If any numerical estimates are provided, they will be assessed to see if a formal quantitative meta-analysis is feasible, employing similar methods to objective one.

Presenting and reporting of results

For both objectives, flow diagrams summarising the study selection process and detailing reasons for exclusion will be used as per 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines for reporting systematic reviews.35 This protocol as well as the final systematic review will be published in a peer-reviewed journal as was done with the previous 2015 systematic review.12 20 Additionally, this review may help to inform the further development of national RHD programmes, by mapping the current burden of disease from RHD, ARF and GAS infection, associated trends in morbidity and mortality, and challenges faced in South Africa for effective prevention and control of RHD.