Substantial and sustained reduction in under-5 mortality, diarrhea, and pneumonia in Oshikhandass, Pakistan: evidence from two longitudinal cohort studies 15 years apart

BACKGROUND

Oshikhandass is a rural village in northern Pakistan where a 1989-1991 verbal autopsy study showed that diarrhea and pneumonia were the top causes of under-5 mortality. Intensive surveillance, active community health education and child health interventions were delivered in 1989-1996; here we assess improvements in under-5 mortality, diarrhea, and pneumonia over this period and 15 years later.

METHODS

Two prospective open-cohort studies in Oshikhandass from 1989 to 1996 (Study 1) and 2011-2014 (Study 2) enrolled all children under age 60 months. Study staff trained using WHO guidelines, conducted weekly household surveillance and promoted knowledge on causes and management of diarrhea and pneumonia. Information about household characteristics and socioeconomic status was collected. Hurdle models were constructed to examine putative risk factors for diarrhea and pneumonia.

RESULTS

Against a backdrop of considerable change in the socioeconomic status of the community, under-5 mortality, which declined over the course of Study 1 (from 114.3 to 79.5 deaths/1000 live births (LB) between 1989 and 1996), exceeded Sustainable Development Goal 3 by Study 2 (19.8 deaths/ 1000 LB). Reductions in diarrhea prevalence (20.3 to 2.2 days/ Child Year [CY]), incidence (2.1 to 0.5 episodes/ CY), and number of bloody diarrhea episodes (18.6 to 5.2%) seen during Study 1, were sustained in Study 2. Pneumonia incidence was 0.5 episodes /CY in Study 1 and 0.2/CY in Study 2; only 5% of episodes were categorized as severe or very severe in both studies. While no individual factors predicted a statistically significant difference in diarrhea or pneumonia episodes, the combined effect of water, toilet and housing materials was associated with a significant decrease in diarrhea; higher household income was the most protective factor for pneumonia in Study 1.

CONCLUSIONS

We report a 4-fold decrease in overall childhood mortality, and a 2-fold decrease in childhood morbidity from diarrhea and pneumonia in a remote rural village in Pakistan between 1989 and 2014. We conclude that significant, sustainable improvements in child health may be achieved through improved socioeconomic status and promoting interactions between locally engaged health workers and the community, but that continued efforts are needed to improve health worker training, supervision, and the rational use of medications.

TRIAL REGISTRATION

Not Applicable.

Real-time forecasting of epidemic trajectories using computational dynamic ensembles

Forecasting the trajectory of social dynamic processes, such as the spread of infectious diseases, poses significant challenges that call for methods that account for data and model uncertainty. Here we introduce an ensemble model for sequential forecasting that weights a set of plausible models and use a frequentist computational bootstrap approach to evaluate its uncertainty. We demonstrate the feasibility of our approach using simple dynamic differential-equation models and the trajectory of outbreak scenarios of the Ebola Forecasting Challenge. Specifically, we generate sequential short-term forecasts of epidemic outbreaks by combining phenomenological models that incorporate flexible epidemic growth scaling, namely the Generalized-Growth Model (GGM) and the Generalized Logistic Model (GLM). We rely on the root-mean-square error (RMSE) to quantify the quality of the models' fits during the calibration periods for weighting their contribution to the ensemble model while forecasting performance was evaluated using the RMSE of the forecasts. For a given forecasting horizon (1-4 weeks), we report the performance for each model as the percentage of the number of times each model outperforms the other models. The overall mean RMSE performance of the GLM and the GGM-GLM ensemble models outcompeted that of participant models of the Ebola Forecasting Challenge. We also found that the ensemble model provided more accurate forecasts with higher frequency than the GGM and GLM models, but its performance varied across forecasting horizons. For instance, across all of the Ebola Challenge Scenarios, the ensemble model outperformed the other models at horizons of 2 and 3 weeks while the GLM outperformed other models at horizons of 1 and 4 weeks.

Predictability in a highly stochastic system: final size of measles epidemics in small populations

A standard assumption in the modelling of epidemic dynamics is that the population of interest is well mixed, and that no clusters of metapopulations exist. The well-known and oft-used SIR model, arguably the most important compartmental model in theoretical epidemiology, assumes that the disease being modelled is strongly immunizing, directly transmitted and has a well-defined period of infection, in addition to these population mixing assumptions. Childhood infections, such as measles, are prime examples of diseases that fit the SIR-like mechanism. These infections have been well studied for many systems with large, well-mixed populations with endemic infection. Here, we consider a setting where populations are small and isolated. The dynamics of infection are driven by stochastic extinction–recolonization events, producing large, sudden and short-lived epidemics before rapidly dying out from a lack of susceptible hosts. Using a TSIR model, we fit prevaccination measles incidence and demographic data in Bornholm, the Faroe Islands and four districts of Iceland, between 1901 and 1965. The datasets for each of these countries suffer from different levels of data heterogeneity and sparsity. We explore the potential for prediction of this model: given historical incidence data and up-to-date demographic information, and knowing that a new epidemic has just begun, can we predict how large it will be? We show that, despite a lack of significant seasonality in the incidence of measles cases, and potentially severe heterogeneity at the population level, we are able to estimate the size of upcoming epidemics, conditioned on the first time step, to within reasonable confidence. Our results have potential implications for possible control measures for the early stages of new epidemics in small populations.

Seven challenges in modeling vaccine preventable diseases

Vaccination has been one of the most successful public health measures since the introduction of basic sanitation. Substantial mortality and morbidity reductions have been achieved via vaccination against many infections, and the list of diseases that are potentially controllable by vaccines is growing steadily. We introduce key challenges for modeling in shaping our understanding and guiding policy decisions related to vaccine preventable diseases.

Transmission dynamics and control of Ebola virus disease outbreak in Nigeria, July to September 2014

We analyse up-to-date epidemiological data of the Ebola virus disease outbreak in Nigeria as of 1 October 2014 in order to estimate the case fatality rate, the proportion of healthcare workers infected and the transmission tree. We also model the impact of control interventions on the size of the epidemic. Results indicate that Nigeria’s quick and forceful implementation of control interventions was determinant in controlling the outbreak rapidly and avoiding a far worse scenario in this country.

Climatic drivers of seasonal influenza epidemics in French Guiana, 2006-2010

OBJECTIVES

Influenza seasonality remains poorly studied in Equatorial regions. Here we assessed the seasonal characteristics and environmental drivers of influenza epidemics in French Guiana, where influenza surveillance was established in 2006.

METHODS

Sentinel GPs monitored weekly incidence of Influenza-like illnesses (ILI) from January 2006 through December 2010 and collected nasopharyngeal specimens from patients for virological confirmation. Times series analysis was used to investigate relationship between ILI and climatic parameters (rainfall and specific humidity).

RESULTS

Based on 1533 viruses identified during the study period, we observed marked seasonality in the circulation of influenza virus in the pre-pandemic period, followed by year-round activity in the post-pandemic period, with a peak in the rainy season. ILI incidence showed seasonal autoregressive variation based on ARIMA analysis. Multivariate dynamic regression revealed that a 1 mm increase of rainfall resulted in an increase of 0.33% in ILI incidence one week later, adjusting for specific humidity (SH). Conversely, an increase of 1 g/kg of SH resulted in a decrease of 11% in ILI incidence 3 weeks later, adjusting for rainfall.

CONCLUSIONS

Increased rainfall and low levels of specific humidity favour influenza transmission in French Guiana.

The 1918-1920 influenza pandemic in Peru

BACKGROUND

Increasing our knowledge of past influenza pandemic patterns in different regions of the world is crucial to guide preparedness plans against future influenza pandemics. Here, we undertook extensive archival collection efforts from three representative cities of Peru-Lima in the central coast, Iquitos in the northeastern Amazon region, Ica in the southern coast-to characterize the temporal, age and geographic patterns of the 1918-1920 influenza pandemic in this country.

MATERIALS AND METHODS

We analyzed historical documents describing the 1918-1920 influenza pandemic in Peru and retrieved individual mortality records from local provincial archives for quantitative analysis. We applied seasonal excess mortality models to daily and monthly respiratory mortality rates for 1917-1920 and quantified transmissibility estimates based on the daily growth rate in respiratory deaths.

RESULTS

A total of 52,739 individual mortality records were inspected from local provincial archives. We found evidence for an initial mild pandemic wave during July-September 1918 in Lima, identified a synchronized severe pandemic wave of respiratory mortality in all three locations during November 1918-February 1919, and a severe pandemic wave during January 1920-March 1920 in Lima and July-October 1920 in Ica. There was no recrudescent pandemic wave in 1920 in Iquitos. Remarkably, Lima experienced the brunt of the 1918-1920 excess mortality impact during the 1920 recrudescent wave, with all age groups experiencing an increase in all cause excess mortality from 1918-1919 to 1920. Middle age groups experienced the highest excess mortality impact, relative to baseline levels, in the 1918-1919 and 1920 pandemic waves. Cumulative excess mortality rates for the 1918-1920 pandemic period were higher in Iquitos (2.9%) than Lima (1.6%). The mean reproduction number for Lima was estimated in the range 1.3-1.5.

CONCLUSIONS

We identified synchronized pandemic waves of intense excess respiratory mortality during November 1918-February 1919 in Lima, Iquitos, Ica, followed by asynchronous recrudescent waves in 1920. Cumulative data from quantitative studies of the 1918 influenza pandemic in Latin American settings have confirmed the high mortality impact associated with this pandemic. Further historical studies in lesser studied regions of Latin America, Africa, and Asia are warranted for a full understanding of the global impact of the 1918 pandemic virus.

Impact of vaccination and birth rate on the epidemiology of pertussis: a comparative study in 64 countries

Bordetella pertussis infection remains an important public health problem worldwide despite decades of routine vaccination. A key indicator of the impact of vaccination programmes is the inter-epidemic period, which is expected to increase with vaccine uptake if there is significant herd immunity. Based on empirical data from 64 countries across the five continents over the past 30-70 years, we document the observed relationship between the average inter-epidemic period, birth rate and vaccine coverage. We then use a mathematical model to explore the range of scenarios for duration of immunity and transmission resulting from repeat infections that are consistent with empirical evidence. Estimates of pertussis periodicity ranged between 2 and 4.6 years, with a strong association with susceptible recruitment rate, defined as birth rate × (1 - vaccine coverage). Periodicity increased by 1.27 years on average after the introduction of national vaccination programmes (95% CI: 1.13, 1.41 years), indicative of increased herd immunity. Mathematical models suggest that the observed patterns of pertussis periodicity are equally consistent with loss of immunity that is not as rapid as currently thought, or with negligible transmission generated by repeat infections. We conclude that both vaccine coverage and birth rate drive pertussis periodicity globally and that vaccination induces strong herd immunity effects. A better understanding of the role of repeat infections in pertussis transmission is critical to refine existing control strategies.

Risk factors for acute generalized exanthematous pustulosis (AGEP)-results of a multinational case-control study (EuroSCAR)

BACKGROUND

Acute generalized exanthematous pustulosis (AGEP) is a disease characterized by the rapid occurrence of many sterile, nonfollicular pustules usually arising on an oedematous erythema often accompanied by leucocytosis and fever. It is usually attributed to drugs.

OBJECTIVES

To evaluate the risk for different drugs of causing AGEP.

PATIENTS AND METHODS

A multinational case-control study (EuroSCAR) conducted to evaluate the risk for different drugs of causing severe cutaneous adverse reactions; the study included 97 validated community cases of AGEP and 1009 controls. Results Strongly associated drugs, i.e. drugs with a lower bound of the 95% confidence interval (CI) of the odds ratio (OR) > 5 were pristinamycin (CI 26-infinity), ampicillin/amoxicillin (CI 10-infinity), quinolones (CI 8.5-infinity), (hydroxy)chloroquine (CI 8-infinity), anti-infective sulphonamides (CI 7.1-infinity), terbinafine (CI 7.1-infinity) and diltiazem (CI 5.0-infinity). No significant risk was found for infections and a personal or family history of psoriasis (CI 0.7-2.2).

CONCLUSIONS

Medications associated with AGEP differ from those associated with Stevens-Johnson syndrome or toxic epidermal necrolysis. Different timing patterns from drug intake to reaction onset were observed for different drugs. Infections, although possible triggers, played no prominent role in causing AGEP and there was no evidence that AGEP is a variant of pustular psoriasis.

Seasonal influenza in the United States, France, and Australia: transmission and prospects for control

Recurrent epidemics of influenza are observed seasonally around the world with considerable health and economic consequences. A key quantity for the control of infectious diseases is the reproduction number, which measures the transmissibility of a pathogen and determines the magnitude of public health interventions necessary to control epidemics. Here we applied a simple epidemic model to weekly indicators of influenza mortality to estimate the reproduction numbers of seasonal influenza epidemics spanning three decades in the United States, France, and Australia. We found similar distributions of reproduction number estimates in the three countries, with mean value 1.3 and important year-to-year variability (range 0.9-2.1). Estimates derived from two different mortality indicators (pneumonia and influenza excess deaths and influenza-specific deaths) were in close agreement for the United States (correlation=0.61, P60%) in healthy individuals who respond well to vaccine, in addition to periodic re-vaccination due to evolving viral antigens and waning population immunity.

[Prevalence and management of heart failure in France: national study among general practitioners of the Sentinelles network]

BACKGROUND

Epidemiological data on heart failure's epidemiology in France are scarce and mostly hospital based. The present study's objective is to estimate the prevalence of heart failure (HF) and its management, in subjects aged 60 years and older seen by the French general practitioners (GP).

METHODS

A standardised questionnaire was mailed to 900 GPs of the Sentinelles network, requiring answers for any patient aged 60 years and more, seen on a randomly assigned single day of year 2002. National census and health insurance data were used to estimate prevalence.

RESULTS

434 GPs answered, reporting data for 1797 patients aged 60 years and more. The 214 patients with HF, aged 79 years on average, had been seen by a cardiologist in 95% of cases. Results of an echocardiography was available for 58% of HF patients. Compared to non-HF patients, patients with HF were significantly more dependent, more frequently requiring home visit of the GP and more frequently hospitalised (p < 0.001, age adjusted). All the 42% HF patients with a reported left ventricle ejection fraction lower than 40% were treated with an angiotensin converting enzyme inhibitor or an angiotensin receptor inhibitor. The prevalence of HF among patients aged 60 years and older was estimated at 11.9% in general practice (95% confidence interval: 10.5-13.5), and at 2.19% (1.9-2.5) in the general population. The prevalence increased with age, over 20% in persons aged 80 years and more.

CONCLUSION

HF in patients aged 60 years and more seen in general practice in France is characterised by a high prevalence and medical consumption in terms of required number of hospitalisation and GP's home visit. For the GP, the diagnosis of HF relies on the cardiologist more than on an echocardiography. The therapeutic management seems to fit the actual recommendations.

More on RotaShield and intussusception: the role of age at the time of vaccination

BACKGROUND

RotaShield, a vaccine intended to prevent severe rotavirus diarrhea, was withdrawn in July 1999, 9 months after it became available in the United States, because of a temporal association with intussusception events that occurred in vaccinated infants. We explore here the effect of age on the risk of intussusception.

METHODS

We reanalyzed a case-control database of the Centers for Disease Control and Prevention by use of a 21-day window, to define vaccine-associated events. We obtained data on vaccine use from the National Immunization Survey and estimated the age-stratified background incidence of intussusception by use of Healthcare Cost and Utilization Project data. We combined these data to estimate how absolute risk varies with age and to model the projected population-attributable risk associated with 3 different vaccination schedules.

RESULTS

We found that the incidence of intussusception associated with the first dose of vaccine increased with age. Infants > or = 90 days old accounted for 80% of cases of intussusception associated with a first dose but had received only 38% of first doses. Modeling of the recommended schedule of vaccination at ages 2, 4, and 6 months projected 1 intussusception event/11,000-16,000 vaccine recipients; modeling of a 2-dose schedule beginning in the neonatal period projected 1 intussusception event/38,000-59,000 vaccine recipients.

CONCLUSIONS

The practice of initiating immunization after age 90 days, which we call "catch-up" vaccination, contributed disproportionately to the occurrence of intussusception associated with the use of RotaShield. A fully implemented 2-dose vaccination schedule begun during the neonatal period would lead to, at most, a 7% increase in the incidence of intussusception above the annual background incidence.

A Bayesian MCMC approach to study transmission of influenza: application to household longitudinal data

We propose a transmission model to estimate the main characteristics of influenza transmission in households. The model details the risks of infection in the household and in the community at the individual scale. Heterogeneity among subjects is investigated considering both individual susceptibility and infectiousness. The model was applied to a data set consisting of the follow-up of influenza symptoms in 334 households during 15 days after an index case visited a general practitioner with virologically confirmed influenza. Estimating the parameters of the transmission model was challenging because a large part of the infectious process was not observed: only the dates when new cases were detected were observed. For each case, the data were augmented with the unobserved dates of the start and the end of the infectious period. The transmission model was included in a 3-levels hierarchical structure: (i) the observation level ensured that the augmented data were consistent with the observed data, (ii) the transmission level described the underlying epidemic process, (iii) the prior level specified the distribution of the parameters. From a Bayesian perspective, the joint posterior distribution of model parameters and augmented data was explored by Markov chain Monte Carlo (MCMC) sampling. The mean duration of influenza infectious period was estimated at 3.8 days (95 per cent credible interval, 95 per cent CI [3.1,4.6]) with a standard deviation of 2.0 days (95 per cent CI [1.1,2.8]). The instantaneous risk of influenza transmission between an infective and a susceptible within a household was found to decrease with the size of the household, and established at 0.32 person day(-1) (95 per cent CI [0.26,0.39]); the instantaneous risk of infection from the community was 0.0056 day(-1) (95 per cent CI [0.0029,0.0087]). Focusing on the differences in transmission between children (less than 15 years old) and adults, we estimated that the former were more likely to transmit than adults (posterior probability larger than 99 per cent), but that the mean duration of the infectious period was similar in children (3.6 days, 95 per cent CI [2.3,5.2]) and adults (3.9 days, 95 per cent CI [3.2,4.9]). The posterior probability that children had a larger community risk was 76 per cent and the posterior probability that they were more susceptible than adults was 79 per cent.

Modelling responses to a smallpox epidemic taking into account uncertainty

Epidemiology and modelling are currently under pressure to build consistent scenarios of control in case of deliberate release of biological weapons. In order to assess the key parameters for the control of a smallpox outbreak in a large city (2 million inhabitants), we built a stochastic model to simulate the course of an epidemic controlled by ring vaccination and case isolation. Assuming a reference scenario with 100 index cases and implementation of intervention 25 days after the attack, the model forecasts an epidemic of 730 cases with an epidemic duration of 240 days. Setting intervention 20 days later would result in an almost fourfold increase in the epidemic size. A multivariate sensitivity analysis has selected three key parameters: the basic reproduction number (i.e. the number of secondary cases infected by one case in an entirely susceptible population, equal to 3 in the reference scenario), time to intervention, and proportion of traced and vaccinated contacts.

Non-steroidal anti-inflammatory drugs as a risk factor for acute diarrhoea: a case crossover study

BACKGROUND AND AIMS

Several cases of acute colitis induced by non-steroidal anti-inflammatory drugs (NSAIDs) have been reported but the general role of recent NSAID intake as a risk factor for acute diarrhoea has not been studied to date. The aim of our study was to determine whether the risk of acute diarrhoea is increased by NSAIDs in a prospective series of acute diarrhoea cases which were seen by general practitioners in France and were serious enough to require a stool culture. Patients, physicians, and methods: A total of 285 consecutive patients with acute diarrhoea, seen by Sentinel general practitioners (GPs) between December 1998 and July 1999, were enrolled in a case crossover study in which each case served as his/her own control. GPs collected information on exposure to NSAIDs during the four month period preceding the onset of diarrhoea. The relative risk of NSAID related acute diarrhoea was estimated by comparing exposure to NSAIDs during a risk period preceding the onset of diarrhoea with exposure during the first part of the four month observation period. Three risk periods lasting for one, three, and six days before the onset of diarrhoea were considered.

RESULTS

The relative risks of acute diarrhoea due to recent NSAID intake were increased for all three risk periods. These risks and their confidence intervals were 2.9 (1.4-6.1) for the one day risk period, 2.7 (1.4-5.1) for the three day period, and 3.3 (2.0-5.4) for the six day period.

CONCLUSION

Recent NSAID intake emerges as a risk factor for acute diarrhoea. We suggest that acute diarrhoea seen in general practice, and not only acute colitis seen by gastroenterologists, should be considered as a potential complication of recent NSAID intake.

Comparison of the statistical efficiency of case-crossover and case-control designs: application to severe cutaneous adverse reactions

Although case-crossover analyses have lately emerged as an alternative to case-control analyses in epidemiological studies, it is not yet known in which situations they give reliable conclusions. In this work, the case-crossover and the case-control designs were first compared on the basis of a dataset from a published study of severe cutaneous adverse reactions resulting from drug exposures of various durations and prevalences of use (245 cases, 1147 controls, and exposures to 23 drug classes). Next, the statistical efficiency of each design was compared via Monte Carlo simulations. Eight of the 13 risk factors identified by case-control analysis of the published data were also identified by the case-crossover analysis, with fairly good agreement on ranks of risk estimates (Spearman's correlation coefficient = 0.71, P < 0.001 ). Simulation studies showed that for relative risks below 8, the case-crossover design (250 cases, 4 control periods/case) had a higher power than the case-control design (250 cases, 4 controls/case), and that the case-crossover design was more conservative than the case-control design for prevalences of drug use below 10%. We conclude that the case-crossover design is not suitable for long-term exposures, but is an appropriate alternative for assessing rare risks associated with transient to short-term exposures.

Nevirapine and the risk of Stevens-Johnson syndrome or toxic epidermal necrolysis

OBJECTIVE

To draw attention to the many cases of Stevens-Johnson syndrome (SJS) or toxic epidermal necrolysis (TEN) related to nevirapine detected in a multinational case-control study of SJS and TEN.

METHODS

Actively detected cases and matched hospital controls were interviewed for exposure to drugs and other risk factors. Data were analysed with case-control and case-crossover methods.

RESULTS

Between May 1997 and November 1999, a diagnosis of SJS or TEN was established in 246 patients. Eighteen were known to be infected by HIV-1 (7.3%), 15 out of these 18 had been exposed to nevirapine. The reaction began 10-240 days after the introduction of nevirapine (median, 12 days) and all patients had received escalating doses. In 10 patients the reaction occurred with the initial dosage. All but one patients received simultaneously a variety of other antiretroviral agents but no specific drug combination emerged, and nevirapine was the only drug significantly associated with an increased risk of SJS or TEN in HIV-infected persons [odds ratio, 62 (10.4; +infinity) in the case-control analysis; odds ratio, +infinity (2.8; +infinity) in the case-crossover analysis].

CONCLUSIONS

In European countries the risk of SJS or TEN in the context of HIV infection appears to be associated with nevirapine. The respect of a lead-in period does not appear to prevent SJS or TEN. Because of the severity of these reactions and the long elimination half-life of nevirapine, we suggest discontinuation of the drug as soon as any eruption occurs.

Automatic coding of reasons for hospital referral from general medicine free-text reports

Although the coding of medical data is expected to benefit both patients and the health care system, its implementation as a manual process often represents a poorly attractive workload for the physician. For epidemiological purpose, we developed a simple automatic coding system based on string matching, which was designed to process free-text sentences stating reasons for hospital referral, as collected from general practitioners (GPs). This system relied on a look-up table, built up from 2590 reports giving a single reason for referral, which were coded manually according to the International Classification of Primary Care (ICPC). We tested the system by entering 797 new reasons for referral. The match rate was estimated at 77%, and the accuracy rate, at 80% at code level and 92% at chapter level. This simple system is now routinely used by a national epidemiological network of sentinel physicians.