Journal Article > ResearchFull Text
PLoS Comput Biol. 23 May 2022; Volume 18 (Issue 5); e1008800.; DOI: 10.1371/journal.pcbi.1008800
Jarvis CI, Gimma A, Finger F, Morris TP, Thompson JA, et al.
PLoS Comput Biol. 23 May 2022; Volume 18 (Issue 5); e1008800.; DOI: 10.1371/journal.pcbi.1008800
The fraction of cases reported, known as 'reporting', is a key performance indicator in an outbreak response, and an essential factor to consider when modelling epidemics and assessing their impact on populations. Unfortunately, its estimation is inherently difficult, as it relates to the part of an epidemic which is, by definition, not observed. We introduce a simple statistical method for estimating reporting, initially developed for the response to Ebola in Eastern Democratic Republic of the Congo (DRC), 2018-2020. This approach uses transmission chain data typically gathered through case investigation and contact tracing, and uses the proportion of investigated cases with a known, reported infector as a proxy for reporting. Using simulated epidemics, we study how this method performs for different outbreak sizes and reporting levels. Results suggest that our method has low bias, reasonable precision, and despite sub-optimal coverage, usually provides estimates within close range (5-10%) of the true value. Being fast and simple, this method could be useful for estimating reporting in real-time in settings where person-to-person transmission is the main driver of the epidemic, and where case investigation is routinely performed as part of surveillance and contact tracing activities.
Journal Article > ResearchFull Text
PLOS Med. 1 April 2021; Volume 18 (Issue 4); e1003587.; DOI:10.1371/journal.pmed.1003587
Polonsky JA, Ivey M, Mazhar KA, Rahman Z, le Polain de Waroux O, et al.
PLOS Med. 1 April 2021; Volume 18 (Issue 4); e1003587.; DOI:10.1371/journal.pmed.1003587
BACKGROUND
Unrest in Myanmar in August 2017 resulted in the movement of over 700,000 Rohingya refugees to overcrowded camps in Cox's Bazar, Bangladesh. A large outbreak of diphtheria subsequently began in this population.
METHODS AND FINDINGS
Data were collected during mass vaccination campaigns (MVCs), contact tracing activities, and from 9 Diphtheria Treatment Centers (DTCs) operated by national and international organizations. These data were used to describe the epidemiological and clinical features and the control measures to prevent transmission, during the first 2 years of the outbreak. Between November 10, 2017 and November 9, 2019, 7,064 cases were reported: 285 (4.0%) laboratory-confirmed, 3,610 (51.1%) probable, and 3,169 (44.9%) suspected cases. The crude attack rate was 51.5 cases per 10,000 person-years, and epidemic doubling time was 4.4 days (95% confidence interval [CI] 4.2-4.7) during the exponential growth phase. The median age was 10 years (range 0-85), and 3,126 (44.3%) were male. The typical symptoms were sore throat (93.5%), fever (86.0%), pseudomembrane (34.7%), and gross cervical lymphadenopathy (GCL; 30.6%). Diphtheria antitoxin (DAT) was administered to 1,062 (89.0%) out of 1,193 eligible patients, with adverse reactions following among 229 (21.6%). There were 45 deaths (case fatality ratio [CFR] 0.6%). Household contacts for 5,702 (80.7%) of 7,064 cases were successfully traced. A total of 41,452 contacts were identified, of whom 40,364 (97.4%) consented to begin chemoprophylaxis; adherence was 55.0% (N = 22,218) at 3-day follow-up. Unvaccinated household contacts were vaccinated with 3 doses (with 4-week interval), while a booster dose was administered if the primary vaccination schedule had been completed. The proportion of contacts vaccinated was 64.7% overall. Three MVC rounds were conducted, with administrative coverage varying between 88.5% and 110.4%. Pentavalent vaccine was administered to those aged 6 weeks to 6 years, while tetanus and diphtheria (Td) vaccine was administered to those aged 7 years and older. Lack of adequate diagnostic capacity to confirm cases was the main limitation, with a majority of cases unconfirmed and the proportion of true diphtheria cases unknown.
CONCLUSIONS
To our knowledge, this is the largest reported diphtheria outbreak in refugee settings. We observed that high population density, poor living conditions, and fast growth rate were associated with explosive expansion of the outbreak during the initial exponential growth phase. Three rounds of mass vaccinations targeting those aged 6 weeks to 14 years were associated with only modestly reduced transmission, and additional public health measures were necessary to end the outbreak. This outbreak has a long-lasting tail, with Rt oscillating at around 1 for an extended period. An adequate global DAT stockpile needs to be maintained. All populations must have access to health services and routine vaccination, and this access must be maintained during humanitarian crises.
Unrest in Myanmar in August 2017 resulted in the movement of over 700,000 Rohingya refugees to overcrowded camps in Cox's Bazar, Bangladesh. A large outbreak of diphtheria subsequently began in this population.
METHODS AND FINDINGS
Data were collected during mass vaccination campaigns (MVCs), contact tracing activities, and from 9 Diphtheria Treatment Centers (DTCs) operated by national and international organizations. These data were used to describe the epidemiological and clinical features and the control measures to prevent transmission, during the first 2 years of the outbreak. Between November 10, 2017 and November 9, 2019, 7,064 cases were reported: 285 (4.0%) laboratory-confirmed, 3,610 (51.1%) probable, and 3,169 (44.9%) suspected cases. The crude attack rate was 51.5 cases per 10,000 person-years, and epidemic doubling time was 4.4 days (95% confidence interval [CI] 4.2-4.7) during the exponential growth phase. The median age was 10 years (range 0-85), and 3,126 (44.3%) were male. The typical symptoms were sore throat (93.5%), fever (86.0%), pseudomembrane (34.7%), and gross cervical lymphadenopathy (GCL; 30.6%). Diphtheria antitoxin (DAT) was administered to 1,062 (89.0%) out of 1,193 eligible patients, with adverse reactions following among 229 (21.6%). There were 45 deaths (case fatality ratio [CFR] 0.6%). Household contacts for 5,702 (80.7%) of 7,064 cases were successfully traced. A total of 41,452 contacts were identified, of whom 40,364 (97.4%) consented to begin chemoprophylaxis; adherence was 55.0% (N = 22,218) at 3-day follow-up. Unvaccinated household contacts were vaccinated with 3 doses (with 4-week interval), while a booster dose was administered if the primary vaccination schedule had been completed. The proportion of contacts vaccinated was 64.7% overall. Three MVC rounds were conducted, with administrative coverage varying between 88.5% and 110.4%. Pentavalent vaccine was administered to those aged 6 weeks to 6 years, while tetanus and diphtheria (Td) vaccine was administered to those aged 7 years and older. Lack of adequate diagnostic capacity to confirm cases was the main limitation, with a majority of cases unconfirmed and the proportion of true diphtheria cases unknown.
CONCLUSIONS
To our knowledge, this is the largest reported diphtheria outbreak in refugee settings. We observed that high population density, poor living conditions, and fast growth rate were associated with explosive expansion of the outbreak during the initial exponential growth phase. Three rounds of mass vaccinations targeting those aged 6 weeks to 14 years were associated with only modestly reduced transmission, and additional public health measures were necessary to end the outbreak. This outbreak has a long-lasting tail, with Rt oscillating at around 1 for an extended period. An adequate global DAT stockpile needs to be maintained. All populations must have access to health services and routine vaccination, and this access must be maintained during humanitarian crises.
Journal Article > ResearchFull Text
Epidemics. 1 December 2018; Volume 25; 72-79.; DOI:10.1016/j.epidem.2018.05.008
le Polain de Waroux O, Flasche S, Kucharski AJ, Langendorf C, Ndazima D, et al.
Epidemics. 1 December 2018; Volume 25; 72-79.; DOI:10.1016/j.epidem.2018.05.008
Although patterns of social contacts are believed to be an important determinant of infectious disease transmission, it remains unclear how the frequency and nature of human interactions shape an individual's risk of infection. We analysed data on daily social encounters individually matched to data on S. pneumoniae carriage and acute respiratory symptoms (ARS), from 566 individuals who took part in a survey in South-West Uganda. We found that the frequency of physical (i.e. skin-to-skin), long (≥1 h) and household contacts - which capture some measure of close (i.e. relatively intimate) contact - was higher among pneumococcal carriers than non-carriers, and among people with ARS compared to those without, irrespective of their age. With each additional physical encounter the age-adjusted risk of carriage and ARS increased by 6% (95%CI 2-9%) and 7% (2-13%) respectively. In contrast, the number of casual contacts (<5 min long) was not associated with either pneumococcal carriage or ARS. A detailed analysis by age of contacts showed that the number of close contacts with young children (<5 years) was particularly higher among older children and adult carriers than non-carriers, while the higher number of contacts among people suffering from ARS was more homogeneous across contacts of all ages. Our findings provide key evidence that the frequency of close interpersonal contact is important for transmission of respiratory infections, but not that of casual contacts. Those results are essential for both improving disease prevention and control efforts as well as informing research on infectious disease dynamics and transmission models, and more studies should be undertaken to further validate our results.
Journal Article > ResearchFull Text
BMC Infect Dis. 11 April 2018; Volume 18 (Issue 1); 172.; DOI:10.1186/s12879-018-3073-1
le Polain de Waroux O, Cohuet S, Ndazima D, Kucharski AJ, Juan-Giner A, et al.
BMC Infect Dis. 11 April 2018; Volume 18 (Issue 1); 172.; DOI:10.1186/s12879-018-3073-1
BACKGROUND
Quantification of human interactions relevant to infectious disease transmission through social contact is central to predict disease dynamics, yet data from low-resource settings remain scarce.
METHODS
We undertook a social contact survey in rural Uganda, whereby participants were asked to recall details about the frequency, type, and socio-demographic characteristics of any conversational encounter that lasted for ≥5 min (henceforth defined as 'contacts') during the previous day. An estimate of the number of 'casual contacts' (i.e. < 5 min) was also obtained.
RESULTS
In total, 566 individuals were included in the study. On average participants reported having routine contact with 7.2 individuals (range 1-25). Children aged 5-14 years had the highest frequency of contacts and the elderly (≥65 years) the fewest (P < 0.001). A strong age-assortative pattern was seen, particularly outside the household and increasingly so for contacts occurring further away from home. Adults aged 25-64 years tended to travel more often and further than others, and males travelled more frequently than females.
CONCLUSION
Our study provides detailed information on contact patterns and their spatial characteristics in an African setting. It therefore fills an important knowledge gap that will help more accurately predict transmission dynamics and the impact of control strategies in such areas.
Quantification of human interactions relevant to infectious disease transmission through social contact is central to predict disease dynamics, yet data from low-resource settings remain scarce.
METHODS
We undertook a social contact survey in rural Uganda, whereby participants were asked to recall details about the frequency, type, and socio-demographic characteristics of any conversational encounter that lasted for ≥5 min (henceforth defined as 'contacts') during the previous day. An estimate of the number of 'casual contacts' (i.e. < 5 min) was also obtained.
RESULTS
In total, 566 individuals were included in the study. On average participants reported having routine contact with 7.2 individuals (range 1-25). Children aged 5-14 years had the highest frequency of contacts and the elderly (≥65 years) the fewest (P < 0.001). A strong age-assortative pattern was seen, particularly outside the household and increasingly so for contacts occurring further away from home. Adults aged 25-64 years tended to travel more often and further than others, and males travelled more frequently than females.
CONCLUSION
Our study provides detailed information on contact patterns and their spatial characteristics in an African setting. It therefore fills an important knowledge gap that will help more accurately predict transmission dynamics and the impact of control strategies in such areas.
Journal Article > ResearchFull Text
BMJ Glob Health. 1 August 2021; Volume 6 (Issue 8); e006736.
Carter SE, Ahuka-Mundeke S, Pfaffmann Zambruni J, Navarro-Colorado C, van Kleef E, et al.
BMJ Glob Health. 1 August 2021; Volume 6 (Issue 8); e006736.
The emerging field of outbreak analytics calls attention to the need for data from multiple sources to inform evidence-based decision making in managing infectious diseases outbreaks. To date, these approaches have not systematically integrated evidence from social and behavioural sciences. During the 2018-2020 Ebola outbreak in Eastern Democratic Republic of the Congo, an innovative solution to systematic and timely generation of integrated and actionable social science evidence emerged in the form of the Cellulle d'Analyse en Sciences Sociales (Social Sciences Analytics Cell) (CASS), a social science analytical cell. CASS worked closely with data scientists and epidemiologists operating under the Epidemiological Cell to produce integrated outbreak analytics (IOA), where quantitative epidemiological analyses were complemented by behavioural field studies and social science analyses to help better explain and understand drivers and barriers to outbreak dynamics. The primary activity of the CASS was to conduct operational social science analyses that were useful to decision makers. This included ensuring that research questions were relevant, driven by epidemiological data from the field, that research could be conducted rapidly (ie, often within days), that findings were regularly and systematically presented to partners and that recommendations were co-developed with response actors. The implementation of the recommendations based on CASS analytics was also monitored over time, to measure their impact on response operations. This practice paper presents the CASS logic model, developed through a field-based externally led consultation, and documents key factors contributing to the usefulness and adaption of CASS and IOA to guide replication for future outbreaks.
Journal Article > ResearchFull Text
Vaccine. 18 September 2017; Volume 35 (Issue 39); 5271-5277.; DOI:10.1016/j.vaccine.2017.07.081
Nackers F, Cohuet S, le Polain de Waroux O, Langendorf C, Nyehangane D, et al.
Vaccine. 18 September 2017; Volume 35 (Issue 39); 5271-5277.; DOI:10.1016/j.vaccine.2017.07.081
BACKGROUND
Information on Streptococcus pneumoniae nasopharyngeal (NP) carriage before the pneumococcal conjugate vaccine (PCV) introduction is essential to monitor impact. The 10-valent PCV (PCV10) was officially introduced throughout Ugandan national childhood immunization programs in 2013 and rolled-out countrywide during 2014. We aimed to measure the age-specific Streptococcus pneumoniae carriage and serotype distribution across all population age groups in the pre-PCV10 era in South Western Uganda.
METHODS
We conducted a two-stage cluster, age-stratified, cross-sectional community-based study in Sheema North sub-district between January and March 2014. One NP swab was collected and analyzed for each participant in accordance with World Health Organization guidelines.
RESULTS
NP carriage of any pneumococcal serotype was higher among children <2years old (77%; n=387) than among participants aged ≥15years (8.5%; n=325) (chi2 p<0.001). Of the 623 positive cultures, we identified 49 serotypes among 610 (97.9%) isolates; thirteen (2.1%) isolates were non-typeable. Among <2years old, serotypes 6A, 6B, 14, 15B, 19F and 23F accounted for half of all carriers. Carriage prevalence with PCV10 serotypes was 29.4% among individuals aged <2years (n=387), 23.4% in children aged 2-4years (n=217), 11.4% in 5-14years (n=417), and 0.4% among individuals ≥15years of age (n=325). The proportion of carried pneumococci serotypes contained in PCV10 was 38.1% (n=291), 32.8% (n=154), 29.4% (n=156), and 4.4% (n=22) among carriers aged <2years, 2-4years, 5-14years and ≥15years, respectively.
DISCUSSION
In Sheema district, the proportion of PCV10 serotypes was low (<40%), across all age groups, especially among individuals aged 15years or older (<5%). PCV10 introduction is likely to impact transmission among children and to older individuals, but less likely to substantially modify pneumococcal NP ecology among individuals aged 15years or older.
Information on Streptococcus pneumoniae nasopharyngeal (NP) carriage before the pneumococcal conjugate vaccine (PCV) introduction is essential to monitor impact. The 10-valent PCV (PCV10) was officially introduced throughout Ugandan national childhood immunization programs in 2013 and rolled-out countrywide during 2014. We aimed to measure the age-specific Streptococcus pneumoniae carriage and serotype distribution across all population age groups in the pre-PCV10 era in South Western Uganda.
METHODS
We conducted a two-stage cluster, age-stratified, cross-sectional community-based study in Sheema North sub-district between January and March 2014. One NP swab was collected and analyzed for each participant in accordance with World Health Organization guidelines.
RESULTS
NP carriage of any pneumococcal serotype was higher among children <2years old (77%; n=387) than among participants aged ≥15years (8.5%; n=325) (chi2 p<0.001). Of the 623 positive cultures, we identified 49 serotypes among 610 (97.9%) isolates; thirteen (2.1%) isolates were non-typeable. Among <2years old, serotypes 6A, 6B, 14, 15B, 19F and 23F accounted for half of all carriers. Carriage prevalence with PCV10 serotypes was 29.4% among individuals aged <2years (n=387), 23.4% in children aged 2-4years (n=217), 11.4% in 5-14years (n=417), and 0.4% among individuals ≥15years of age (n=325). The proportion of carried pneumococci serotypes contained in PCV10 was 38.1% (n=291), 32.8% (n=154), 29.4% (n=156), and 4.4% (n=22) among carriers aged <2years, 2-4years, 5-14years and ≥15years, respectively.
DISCUSSION
In Sheema district, the proportion of PCV10 serotypes was low (<40%), across all age groups, especially among individuals aged 15years or older (<5%). PCV10 introduction is likely to impact transmission among children and to older individuals, but less likely to substantially modify pneumococcal NP ecology among individuals aged 15years or older.
Journal Article > CommentaryFull Text
Vaccine. 23 October 2019; Volume 37; DOI:10.1016/j.vaccine.2019.09.038
van Zandvoort K, Checchi F, Diggle E, Eggo RM, Gadroen K, et al.
Vaccine. 23 October 2019; Volume 37; DOI:10.1016/j.vaccine.2019.09.038
Streptococcus pneumoniae is a common human commensal that causes a sizeable part of the overall childhood mortality in low income settings. Populations affected by humanitarian crises are at especially high risk, because a multitude of risk factors that are enhanced during crises increase pneumococcal transmission and disease severity. Pneumococcal conjugate vaccines (PCVs) provide effective protection and have been introduced into the majority of routine childhood immunisation programmes globally, though several barriers have hitherto limited their uptake during humanitarian crises. When PCV coverage cannot be sustained during crises or when PCV has not been part of routine programmes, mass vaccination campaigns offer a quick acting and programmatically feasible bridging solution until services can be restored. However, we currently face a paucity of evidence on which to base the structure of such campaigns. We believe that, now that PCV can be procured at a substantially reduced price through the Humanitarian Mechanism, this lack of information is a remaining hurdle to PCV use in humanitarian crises. Considering the difficulties in conducting research in crises, we propose an evidence generation pathway consisting of primary data collection in combination with mathematical modelling followed by quasi-experimental evaluation of a PCV intervention, which can inform on optimal vaccination strategies that consider age targeting, dosing regimens and impact duration.
Journal Article > ResearchFull Text
PLOS One. 29 April 2021; Volume 16 (Issue 4); e0250505.; DOI:10.1371/journal.pone.0250505
Mazhar KA, Finger F, Evers ES, Kuehne A, Ivey M, et al.
PLOS One. 29 April 2021; Volume 16 (Issue 4); e0250505.; DOI:10.1371/journal.pone.0250505
In the summer of 2017, an estimated 745,000 Rohingya fled to Bangladesh in what has been described as one of the largest and fastest growing refugee crises in the world. Among numerous health concerns, an outbreak of acute jaundice syndrome (AJS) was detected by the disease surveillance system in early 2018 among the refugee population. This paper describes the investigation into the increase in AJS cases, the process and results of the investigation, which were strongly suggestive of a large outbreak due to hepatitis A virus (HAV). An enhanced serological investigation was conducted between 28 February to 26 March 2018 to determine the etiologies and risk factors associated with the outbreak. A total of 275 samples were collected from 18 health facilities reporting AJS cases. Blood samples were collected from all patients fulfilling the study specific case definition and inclusion criteria, and tested for antibody responses using enzyme-linked immunosorbent assay (ELISA). Out of the 275 samples, 206 were positive for one of the agents tested. The laboratory results confirmed multiple etiologies including 154 (56%) samples tested positive for hepatitis A, 1 (0.4%) positive for hepatitis E, 36 (13%) positive for hepatitis B, 25 (9%) positive for hepatitis C, and 14 (5%) positive for leptospirosis. Among all specimens tested 24 (9%) showed evidence of co-infections with multiple etiologies. Hepatitis A and E are commonly found in refugee camps and have similar clinical presentations. In the absence of robust testing capacity when the epidemic was identified through syndromic reporting, a particular concern was that of a hepatitis E outbreak, for which immunity tends to be limited, and which may be particularly severe among pregnant women. This report highlights the challenges of identifying causative agents in such settings and the resources required to do so. Results from the month-long enhanced investigation did not point out widespread hepatitis E virus (HEV) transmission, but instead strongly suggested a large-scale hepatitis A outbreak of milder consequences, and highlighted a number of other concomitant causes of AJS (acute hepatitis B, hepatitis C, Leptospirosis), albeit most likely at sporadic level. Results strengthen the need for further water and sanitation interventions and are a stark reminder of the risk of other epidemics transmitted through similar routes in such settings, particularly dysentery and cholera. It also highlights the need to ensure clinical management capacity for potentially chronic conditions in this vulnerable population.
Journal Article > Pre-PrintFull Text
bioRxiv. 17 February 2021; DOI:10.1101/2021.02.17.431606
Jarvis CI, Gimma A, Finger F, Morris TP, Thompson JA, et al.
bioRxiv. 17 February 2021; DOI:10.1101/2021.02.17.431606
The fraction of cases reported, known as ‘reporting’, is a key performance indicator in an outbreak response, and an essential factor to consider when modelling epidemics and assessing their impact on populations. Unfortunately, its estimation is inherently difficult, as it relates to the part of an epidemic which is, by definition, not observed.
We introduce a simple statistical method for estimating reporting, initially developed for the response to Ebola in Eastern Democratic Republic of the Congo (DRC), 2018-2020. This approach uses transmission chain data typically gathered through case investigation and contact tracing, and uses the proportion of investigated cases with a known, reported infector as a proxy for reporting. Using simulated epidemics, we study how this method performs for different outbreak sizes and reporting levels. Results suggest that our method has low bias, reasonable precision, and despite sub-optimal coverage, usually provides estimates within close range (5-10%) of the true value.
Being fast and simple, this method could be useful for estimating reporting in real-time in settings where person-to-person transmission is the main driver of the epidemic, and where case investigation is routinely performed as part of surveillance and contact tracing activities.
We introduce a simple statistical method for estimating reporting, initially developed for the response to Ebola in Eastern Democratic Republic of the Congo (DRC), 2018-2020. This approach uses transmission chain data typically gathered through case investigation and contact tracing, and uses the proportion of investigated cases with a known, reported infector as a proxy for reporting. Using simulated epidemics, we study how this method performs for different outbreak sizes and reporting levels. Results suggest that our method has low bias, reasonable precision, and despite sub-optimal coverage, usually provides estimates within close range (5-10%) of the true value.
Being fast and simple, this method could be useful for estimating reporting in real-time in settings where person-to-person transmission is the main driver of the epidemic, and where case investigation is routinely performed as part of surveillance and contact tracing activities.