Insights from mathematical modelling and quantitative analysis on the proposed 2030 goals for trachoma

Disclaimer

The views and opinions expressed in this article are those of the author(s) and do not necessarily reflect the views of the World Health Organization. Publication in Gates Open Research does not imply endorsement by the Gates Foundation.

Background

Trachoma is a neglected tropical disease caused by infection with the bacterium Chlamydia trachomatis. During an infection episode, conjunctival inflammation occurs, which leads to the presence of follicles on the eyelids (Trachomatous inflammation-follicular (TF))¹. Repeated infection with the bacteria over time, which results in scarring of the eyelids, leads to in-turning of the eyelashes. This is known as trachomatous trichiasis (TT), which traumatizes the eye surface leading to superinfection and blindness^1,2.

The World Health Organization leads an Alliance that aims to achieve the elimination of trachoma as a public health problem (EPHP) in all endemic districts by 2020¹. This is defined by the achievement of three goals: 1) reduction of TF prevalence in 1–9 year olds to <5% in each formerly endemic district, 2) a TT prevalence “unknown to the health system” in >=15-year-olds of <0.2% , and 3) the presence of a system to identify and manage incident cases of TT. In order to eliminate trachoma, the WHO endorses the implementation of the SAFE strategy which consists of four components: (S) surgery to correct trichiasis; (A) mass distribution of antibiotics to clear infection in the community (topical tetracycline is used in very young children or other individuals unable to take azithromycin), (F) promotion of facial cleanliness in order to reduce transmission via eye discharge and (E) environmental improvement to ensure that the environment no longer helps to facilitate the transmission of infection. Currently, the prevalence of clinically active trachoma is assessed by trained graders’ clinical examination. In the context of low endemicity, other methods are being considered, including photographic and laboratory assessment. MDA is recommended to all districts where TF is >5%. A course of one, three, or five annual rounds is recommended where TF is between 5–10%, 10–30%, and over 30% respectively. Following completion of the prescribed MDA, a follow-up survey is conducted to assess whether further rounds of MDA are required. To date, ten countries have been validated by WHO as having achieved EPHP³.

Both mathematical and statistical models have been developed to gain insight into the transmission dynamics of infection. Such models have been used to try and understand the potential impact of different intervention strategies that could help to accelerate elimination efforts⁴, as well as understanding likely elimination timelines through forecasting. In addition, a recent review on the contribution of mathematical modelling to trachoma research and elimination efforts was published by the two teams in the first iteration of the NTD modelling consortium⁵. Furthermore, a multi-group forecast comparison was also conducted to look at the strengths and limitations of different modelling approaches for forecasting the future prevalence of TF at the district level⁶.

Moving forward past the current 2020 goals, whilst substantial progress has been made towards achieving EPHP of trachoma, it has become apparent that a number of endemic regions will not achieve this target by 2020. It is important to note that the original 2020 goals were political and aspirational, and thanks to the effort of ministries of health and countless donors and partners over the years can now be formally assessed with data. Therefore, WHO has revised the timeline, with the aim of validating EPHP in all endemic countries by 2030. Using the insights that have been gained from recent modelling work on trachoma, in this article we highlight the practical considerations of EPHP (the timelines required, sufficiency of current surveillance diagnostics and feasibility of achieving it) and the future considerations that may be needed following EPHP to maintain the gains (Table 1 provides a summary of the key issues).

What have we learned from the past 10 years that we can apply for the next 10 years?

There has been a substantial amount of programmatic success in trachoma elimination over the last 10 years, with global prevalence falling dramatically as a result of successful intervention programs. Whilst EPHP includes both TF and TT, most research focuses on changes, modelling and monitoring of TF and infection prevalence. Limited analysis and forecasting of TT prevalence to date has occurred in part because the trajectory of changes in observed TT prevalence depends not only on the incidence of TT (a chronic condition and stochastic process related to an individual’s past number of infections⁷), but also due to demography and health service access. Thus the observed prevalent number of TT cases is partly determined by the speed and efficiency of active case finding and surgical service delivery, which are inherently more challenging and uncertain to model.

Mathematical modelling and current surveillance data suggests that EPHP is feasible, and indeed has already been achieved by a number of endemic countries³. However, in health districts with long-term persistence, such as a few high prevalence districts in Ethiopia (>40% baseline prevalence), annual MDA alone is not sufficient to achieve EPHP and must be supplemented with additional tools to reduce transmission^6,8–10. In particular, more intensive facial cleanliness and environmental improvement (F&E) or more intensive antibiotics are measures that may be necessary in a select few hot spots^4,8,11. Similarly, statistical analysis of data provided and collected by trachoma endemic countries has indicated that the vast majority of endemic health districts are on track to achieve TF <5% for EPHP by 2020^6,8. These findings are consistent across both dynamic and statistical modelling frameworks that were independently developed by the different partners of our consortium.

In health districts that remain problematic, to understand how EPHP may be achieved by 2030, dynamic modelling work has explored a range of alternative and more intensive antibiotic distribution strategies that could be implemented^4,12. To date it has been challenging to measure the true impact of F&E and its potential role in helping to reduce transmission, and thus it has been challenging to model. A few field studies that have assessed F&E were unable to find a significant effect^13–15. However, an on-going clinical trial is seeking to help address this gap in knowledge (Stronger SAFE). Even if annual mass antibiotic treatment is insufficient to achieve EPHP goals in certain hyperendemic areas, it may prevent resurgence of infection¹⁶.

Modelling has also been used to investigate whether targeting a residual core group of children with additional antibiotic treatment, while continuing annual MDA to the entire population would be more effective at clearing infection from the community¹⁷. The study suggested that if average duration of infection per group and dominant eigenvalue of a next generation matrix of the transmission model are defined, then a sufficient core group can be determined and used to find the absolute minimum sized core group, based on a fully specified model or even from epidemiological data. A separate mathematical model of a double-dose antibiotic treatment strategy where two doses of antibiotics are given two weeks apart, in combination with enhanced F&E suggested that feasibility of EPHP may be increased in high transmission settings¹⁸. This modelling suggested that sustained F&E could help maintain the gains initially achieved through intense antibiotic distribution¹⁸.

A number of RCTs informed by modelling are currently underway in Ethiopia, with the aim of assessing the potential impact of alternative and intensive treatment strategies. One RCT (KETFO) is assessing whether quaterly treatment of children alone can lead to EPHP in severely affected communities¹⁹. Additionally, an RCT looking at intensive WASH (SWIFT-WUHA)²⁰ and an RCT looking at the distribution of two doses of antibiotics one week apart (TESFA²¹) are in progress.

What are the practical considerations of the currently proposed goals?

What risks need to be mitigated to achieve and maintain the 2030 goals?

There are a number of practical factors that may directly impact on-going program implementation. Firstly, both empirical data and dynamic modelling have suggested that in areas of high prevalence, annual MDA alone is not sufficient to reach the goal^5,11,33. As previously described, a number of alternative intervention strategies are currently being evaluated within RCTs to identify solutions to this problem. Secondly, maintaining and optimising the frequency of antibiotic use is of paramount importance in order for gains to be achieved and maintained. Coverage is often reported to be high³⁴, but in practice this can be hard to measure in the field³⁵. Equally, systematic variation within Health districts leading to local pockets of undelivered MDA and exclusion from TF surveys, may limit progress by leaving reservoir sources of infection in communities that are deemed to have been treated³⁶. Thirdly, no resistance to azithromycin has been reported, however careful monitoring for suboptimal treatment effects is needed. If resistance does emerge, EPHP success will be severely undermined⁵. Fourthly, as prevalence begins to decline in many endemic regions, movement of individuals between infected and uninfected areas may facilitate persistence of infection or re-introduction into formerly infection-free areas.

A number of risks remain for surveillance in terms of classifying and continuing to confirm elimination. First of all, it is currently uncertain whether or not TF prevalence alone is sufficient to classify health districts that have achieved EPHP^28,37,38. Meanwhile, the non-linear relationship between viral load and TF complicates our understanding of how PCR detection and TF relate to one another. In fact, TF has been detected in some areas of the world without the bacterial organism being identified, suggesting that other factors besides trachoma may also cause TF. In short, following EPHP, it is unclear how to conduct surveillance to ensure that EPHP is maintained. Serology has been suggested as one potential option, although sero-surveillance data in EPHP settings are only starting to become available.

There are a number of risks that we need to be mindful of with respect to modelling trachoma and interpreting model outputs. It is typically assumed that the accuracy of TF detection will remain constant. However, this is an optimistic assumption, as we expect the ability to recognize TF to decrease as the disease becomes rarer. This issue will become particularly important as modelling surveillance in very low transmission scenarios receives greater attention. Importantly, there are no high-resolution empirical studies on dynamics of infection in areas with hypo-endemic disease, which means that simulations and models of low-level prevalence are likely to have a large number of uncertainties. Further empirical studies are needed in order to understand how to more accurately model transmission at low prevalence.

Future directions

Data availability

No data are associated with this article.