Novel approaches to evaluate the impact of the SAFE strategy on trachoma and other neglected tropical diseases in Amhara National Regional State, Ethiopia

Background: Trachoma, a blinding bacterial disease of the ocular surface, is the leading
cause of infectious blindness, responsible for the visual impairment of 2.2 million people
worldwide and an estimated economic loss of US$ 5.3 billion annually. Blinding trachoma,
considered a neglected tropical disease, is targeted for global elimination as a public health
problem by the year 2020. To achieve elimination, the World Health Organization (WHO)
recommends implementing an integrated intervention package of surgery, antibiotics, facial
cleanliness, and environmental improvement, known as the SAFE strategy. Surgery aims to
correct trichiasis, the blinding anatomical condition of inward turning lashes touching the eye.
Antibiotics are distributed annually to endemic communities to treat relatively asymptomatic
ocular Chlamydia trachomatis infections to reduce the infectious reservoir. The promotion of
facial cleanliness and environmental improvements, use of water and latrines for hygiene
and sanitation, both target interrupting transmission of the infection. Of the 325 million
persons estimated living in trachoma endemic communities, over 70% reside in sub-Saharan
Africa. Within the region, Ethiopia, Nigeria, and South Sudan are estimated to have the
highest burden of disease. The SAFE strategy in Ethiopia has been implemented since 2007
targeting all 17 million residents of the Amhara National Regional State.
Goals and specific objectives: The goal of this PhD thesis was to investigate novel
approaches of measuring the impact of SAFE interventions on trachoma and other
neglected tropical diseases within the context of a national elimination programme as
implemented in the Amhara National Regional State of Ethiopia. The specific operational
research objectives were to determine whether areas receiving 3-5 years of interventions
had achieved elimination by applying new evaluation guidelines; whether new electronic
data collection technology could facilitate impact evaluation; whether SAFE interventions
had any impact on trachomatous scarring (TS) among children; whether school-based
surveys might serve as an alternative method to assess trachoma; and whether SAFE
interventions have had any impact on intestinal parasitic infections.
Methods: Data collection for the thesis project occurred in three phases. The first phase
involved the implementation of a population-based, cross-sectional study utilising a cluster
random sampling design to survey 360 communities in South Wollo zone to estimate
prevalence of trachoma after three years of SAFE interventions. Data was collected using
standard paper-based questionnaires. The second phase involved the development and
field-testing of a new electronic data collection system in a pilot study utilising a mixed,
quantitative and qualitative, study design. The last phase involved the implementation of
another population-based cross-sectional study in South Gondar zone after receiving five
years of SAFE interventions. The same sampling methodology was used to survey another
360 communities, yet integrating both assessment of trachoma and intestinal parasitic
infections. Additionally, data was collected strictly by the newly developed electronic system.
Clinical signs of trachoma were individually assessed using the WHO simplified trachoma
grading system. Intestinal parasitic infections among children aged 2-15 years were
determined by concentrating preserved stool specimens with ether for microscopic
examination.
Results: From 714 communities in the two zones, 72,452 persons were examined for
trachoma. The prevalence of trachomatous inflammation follicular (TF) among children aged
1-9 years was 26.4% in South Wollo and 25.9% in South Gondar zone. Trachomatous
inflammation intense (TI) was less prevalent than TF; 4.3% and 7.0% in South Wollo and
South Gondar respectively. TT prevalence in the two zones suggest that over 59,000
persons are estimated to have trichiasis and in need of surgery. In South Gondar the
prevalence of TS among children under the age of 11 years has declined from 24.9% in
2000 to 2.2% in 2011. While declines in intense inflammation and scarring were observed
among children since intervention, the WHO targets for elimination have not been achieved.
The android-based tablet computer and the standard paper questionnaire were
comparable in regards to time required to collect data during the pilot study, proportion of
mistakes made while recording data and costs when considering data entry of paper
questionnaires. Data recorders preferred to collect data electronically even though initially
they felt the tablet interrupted their connection with the interview respondents. Electronic
data collection resulted in completion of the large-scale surveys from preparation to
presentation of results in 35% less time (one month earlier) than the standard paper-based
surveys.
An analysis of 75,864 children examined in community-based surveys in Ethiopia
(from field work described in this thesis), Mali, Niger, and Nigeria found that differences
between children who attend and do not attend school varies across survey settings in
regards to age, gender, having a clean face, and participation in antibiotic distribution for
trachoma control. Meta-analysis of the data found that TF was less likely (odds ratio=0.71)
among school-attendees than non-attendees when controlling for age, sex, and clustering at
the household and community levels. Children attending school did not represent the target
age group recommended for assessment of trachoma prevalence.
Stool specimens from 2,338 children aged 2-15 years from 99 communities in South
Gondar were assessed for intestinal parasites. The prevalence of any helminth infection was
24.2%, which represented a 50% reduction from a previously published study prior to the
SAFE interventions. Over 70% of children had at least one type of intestinal protozoan
infection. Significant increases were observed in household latrine ownership, access to
water, use of an improved water source for drinking, and face washing behaviour since the
start of the interventions.
Conclusions: Trachoma remains a public health problem in South Wollo and South Gondar
zones of the Amhara National Regional State of Ethiopia and ongoing interventions are
warranted to control transmission to prevent incident blinding disease and provide surgery
for prevalent and incident cases of trichiasis. The application of new WHO guidelines to
evaluate trachoma at the sub-district level was feasible, but required significant resources.
Electronic data capture facilitates the implementation of such large-scale impact evaluation
surveys for neglected tropical diseases allowing the results to be generated immediately with
as few mistakes as were made with paper-based data collection. Measuring prevalence of
TS among children over time offers an additional way to monitor impact of the SAFE
strategy. Children under the age of 11 years have substantially benefited from having lived in
an environment where the SAFE strategy has been implemented for five years. The use of
school-based sampling approaches for assessing trachoma prevalence risks
underestimating true prevalence in the community. The prevalence of intestinal helminths
among school-aged children has declined alongside significant increases in household-level
indicators of water, sanitation, and hygiene since the implementation of the SAFE strategy in
South Gondar. Yet, there is ongoing transmission of intestinal parasitic infections warranting
improved control interventions. Integrating both the assessment of trachoma and intestinal
parasitic infections in community-based surveys was a feasible approach to evaluate a
broader impact of the SAFE strategy in a programmatic setting. Overall, the operational
research presented in this thesis successfully generated evidence for health system
decisions, contributed new information to the respective scientific fields, and identified areas
warranting additional research.