SEARCH · bioRxiv Science

Results for “epidemiology”

Search indexed bioRxiv preprints in genomics, neuroscience, cell biology and bioinformatics. Read source abstracts and check manuscript versions; preprints are not peer reviewed.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 919 records · Page 51Linked to original sources

bioRxiv2017

Infectious Reactivation of Cytomegalovirus Explaining Age-and Sex-Specific Patterns of Seroprevalence

Human cytomegalovirus is a herpes virus with poorly understood transmission dynamics. We here provide quantitative estimates of the transmissibility of primary infection, reactivation, and re-infection using age-and sex-specific antibody response data. The data are optimally described by three distributions of antibody measurements, i.e. uninfected, infected, and infected after reactivation/re-infection. Estimates of seroprevalence increase gradually with age, such that at 80 years 73% (95%CrI: 64%-78%) of females and 62% (95%CrI: 55%-68%) of males is infected, while 57% (95%CrI: 47%-67%) of females and 37% (95%CrI: 28%-46%) of males has experienced a reactivation or re-infection episode. Merging the statistical analyses with transmission models, we find that infectious reactivation is key to provide a good fit fit to the data. Estimated reactivation rates increase from low values in children to 2%-6% per year older women. The results advance a hypothesis in which adult-to-adult transmission after infectious reactivation is the main driver of infection.

epidemiology↗

bioRxiv2017

The Negev hospital-university-based (HUB) autism database

Elucidating the heterogeneous etiologies of autism will require investment in comprehensive longitudinal data acquisition from large community based cohorts. With this in mind, we have established a hospital-university-based (HUB) database of autism which incorporates prospective and retrospective data from a large and ethnically diverse population. Here we present initial findings from 188 children who were diagnosed with autism during the first eighteen months of the study. The unique characteristics of this cohort included: significant differences between Bedouin and Jewish children in different risk factors and clinical characteristics; complete birth records for >90% of the children; and a high frequency of consanguineous marriages. Thus, the Negev HUB autism database comprises a remarkably unique resource to study different aspects of autism.

epidemiology↗

bioRxiv2017

A general goodness-of-fit test for survival analysis

Existing goodness-of-fit tests for survival data are either exclusively graphical in nature or only test specific model assumptions, such as the proportional hazards assumption. We describe a flexible, parameter-free goodness-of-fit test that provides a simple numerical assessment of a models suitability regardless of the structure of the underlying model. Intuitively, the goodness-of-fit test utilizes the fact that for a good model early event occurrence is predicted to be just as likely as late event occurrence, whereas a bad model has a bias towards early or late events. Formally, the goodness-of-fit test is based on a novel generalized Martingale residual which we call the martingale survival residual. The martingale survival residual has a uniform probability density function defined on the interval -0.5 to +0.5 if censoring is either absent or accounted for as one outcome in a competing hazards framework. For a good model, the set of calculated residuals is statistically indistinguishable from the uniform distribution, which is tested using the Kolmogorov-Smirnov statistic.

epidemiology↗

bioRxiv2017

A risk assessment framework for seed degeneration: Informing an integrated seed health strategy for vegetatively-propagated crops

Pathogen build-up in vegetative planting material, termed seed degeneration, is a major problem in many low-income countries. When smallholder farmers use seed produced on-farm or acquired outside certified programs, it is often infected. We introduce a risk assessment framework for seed degeneration, evaluating the relative performance of individual and combined components of an integrated seed health strategy. The frequency distribution of management performance outcomes was evaluated for models incorporating biological and environmental heterogeneity, with the following results. (1) On-farm seed selection can perform as well as certified seed, if the rate of success in selecting healthy plants for seed production is high; (2) When choosing among within-season management strategies, external inoculum can determine the relative usefulness of incidence-altering management (affecting the proportion of diseased plants/seeds) and rate-altering management (affecting the rate of disease transmission in the field); (3) Under severe disease scenarios, where it is difficult to implement management components at high levels of effectiveness, combining management components can produce synergistic benefits and keep seed degeneration below a threshold; (4) Combining management components can also close the yield gap between average and worst-case scenarios. We also illustrate the potential for expert elicitation to provide parameter estimates when data are unavailable.

epidemiology↗

bioRxiv2017

Asian lineage of Zika virus RNA pseudoknot may induce ribosomal frameshift and produce a new neuroinvasive protein ZIKV-NS1’

Zika virus (ZIKV) is a threat to humanity, and understanding its neuroinvasiveness is a major challenge. Microcephaly observed in neonates in Brazil is associated with ZIKV that belongs to the Asian lineage. What distinguishes the neuroinvasiveness between the RNA lineages from Asia and Africa is still unknown. Here we identify an aspect that may explain the different behavior between the two lineages. The distinction between the two groups is the occurrence of an alternative protein NS1 (ZIKV-NS1), which happens through a pseudoknot in the virus RNA that induces a ribosomal frameshift. Presence of NS1 protein is also observed in other Flavivirus that are neuroinvasive, and when NS1 production issuppressed, neuroinvasiveness is reduced.1 This evidence gives grounds to suggest that the ZIKV-NS1 occurring in the Asian lineage is responsible for neuro-tropism, which causes the neuro-pathologies associated with ZIKV infection, of which microcephaly is the most dev astating. The existence of ZIKV-NS1, which only exists in the Asian lineage, was inferred through bioinformatic methods, and it has yet to be experimentally observed. If its occurrence is confirmed, it will be a potential target in fighting the neuro-diseases associated with ZIKV.

epidemiology↗

bioRxiv2017

Education and coronary heart disease: a Mendelian randomization study

ObjectivesTo determine whether educational attainment is a causal risk factor in the development of coronary heart disease.\n\nDesignMendelian randomization study, where genetic data are used as proxies for education, in order to minimize confounding. A two-sample design was applied, where summary level genetic data was analysed from two publically available consortia.\n\nSettingIn the main analysis, we analysed genetic data from two large consortia (CARDIoGRAM and SSGAC), comprising of 112 cohorts from predominantly high-income countries. In addition, we also analysed genetic data from 7 additional large consortia, in order to identify putative causal mediators.\n\nParticipantsThe main analysis was of 589 377 men and women, predominantly of European origin.\n\nExposureA one standard deviation increase in the genetic predisposition towards higher education (i.e. 3.6 years of additional schooling). This was measured by 162 genetic variants that have been previously associated with education.\n\nMain outcomeCombined fatal and nonfatal coronary heart disease (63 746 events).\n\nResults3.6 years of additional education lowered the risk of coronary heart disease by a third (odds ratio = 0.67, 95% confidence interval [CI], 0.59 to 0.77, p=0.01). Equivalent increases in education were also causally associated with reductions in smoking, BMI and improvements in blood lipid profiles.\n\nConclusionsMore time spent in education is causally associated with a large reduction in the risk of coronary heart disease. This may be partly explained by changes to smoking, BMI and a blood lipids. These findings offer support for policy interventions that increase education, in order to also reduce the burden of cardiovascular disease.

epidemiology↗

bioRxiv2017

Disease implications of animal social organization and network structure - a quantitative analysis

O_LIThe disease costs of sociality have largely been understood through the link between group size and transmission. However, infectious disease spread is driven primarily by the social organization of interactions in a group and not its size.\nC_LIO_LIWe used statistical models to review the social network organization of 47 species, including mammals, birds, reptiles, fish and insects by categorizing each species into one of three social systems, relatively solitary, gregarious and socially hierarchical. Additionally, using computational experiments of infection spread, we determined the disease costs of each social system.\nC_LIO_LIWe find that relatively solitary species have large variation in number of social partners, that socially hierarchical species are the least clustered in their interactions, and that social networks of gregarious species tend to be the most fragmented. However, these structural differences are primarily driven by weak connections, which suggests that different social systems have evolved unique strategies to organize weak ties.\nC_LIO_LIOur synthetic disease experiments reveal that social network organization can mitigate the disease costs of group living for socially hierarchical species when the pathogen is highly transmissible. In contrast, highly transmissible pathogens cause frequent and prolonged epidemic outbreaks in gregarious species.\nC_LIO_LIWe evaluate the implications of network organization across social systems despite methodological challenges, and our findings offer new perspective on the debate about the disease costs of group living. Additionally, our study demonstrates the potential of meta-analytic methods in social network analysis to test ecological and evolutionary hypotheses on cooperation, group living, communication, and resilience to extrinsic pressures.\nC_LI

epidemiology↗

bioRxiv2017

Measuring changes in transmission of neglected tropical diseases, malaria, and enteric pathogens from quantitative antibody levels

BackgroundSerologicalantibody levels are a sensitive marker of pathogen exposure, and advances in multiplex assays have created enormous potential for large-scale, integrated infectious disease surveillance. Most methods to analyze antibody measurements reduce quantitative antibody levels to seropositive and seronegative groups, but this can be difficult for many pathogens and may provide lower resolution information than quantitative levels in low transmission settings. Analysis methods have predominantly maintained a single disease focus, yet integrated surveillance platforms would benefit from methodologies that work across diverse pathogens included in multiplex assays.\n\nMethods/Principal FindingsWe developed an approach to measure changes in transmission from quantitative antibody levels that can be applied to diverse pathogens of global importance. We compared age-dependent immunoglobulin G curves in repeated cross-sectional surveys between populations with differences in transmission for multiple pathogens, including: lymphatic filariasis (Wuchereria bancrofti) measured before and after mass drug administration on Mauke, Cook Islands, malaria (Plasmodium falciparum) before and after a combined insecticide and mass drug administration intervention in the Garki project, Nigeria, and enteric protozoans (Cryptosporidium parvum, Giardia intestinalis, Entamoeba histolytica), bacteria (enterotoxigenic Escherichia coli, Salmonella spp.), and viruses (norovirus groups I and II) in children living in Haiti and the USA. Age-dependent antibody curves fit with ensemble machine learning followed a characteristic shape across pathogens that aligned with predictions from basic mechanisms of humoral immunity. Differences in pathogen transmission led to shifts in fitted antibody curves that were remarkably consistent across pathogens, assays, and populations. Mean antibody levels correlated strongly with traditional measures of transmission intensity, such as the entomological inoculation rate for P. falciparum (Spearmans rho=0.75). Seroprevalence estimates recapitulated patterns observed in quantitative antibody levels, albeit with lower resolution.\n\nConclusions/SignificanceAge-dependent antibody curves and summary means provided a robust and sensitive measure of changes in transmission, with greatest sensitivity among young children. The method generalizes to pathogens that can be measured in high-throughput, multiplex serological assays, and scales to surveillance activities that require high spatiotemporal resolution. The approach represents a new opportunity to conduct integrated serological surveillance for neglected tropical diseases, malaria, and other infectious diseases with well-defined antigen targets.\n\nAuthor SummaryGlobal elimination strategies for infectious diseases like neglected tropical diseases and malaria rely on accurate estimates of pathogen transmission to target and evaluate control programs. Circulating antibody levels can be a sensitive measure of recent pathogen exposure, but no broadly applicable method exists to measure changes in transmission directly from quantitative antibody levels. We developed a novel method that applies recent advances in machine learning and data science to flexibly fit age-dependent antibody curves. Shifts in age-dependent antibody curves provided remarkably consistent, sensitive measures of transmission changes when evaluated across many globally important pathogens (filarial worms, malaria, enteric infections). The methods generality and performance in diverse applications demonstrate its broad potential for integrated serological surveillance of infectious diseases.

epidemiology↗

bioRxiv2017

Response Adjusted for Days of Antibiotic Risk (RADAR): evaluation of a novel method to analyze antibiotic stewardship interventions

OBJECTIVESThe Response Adjusted for Days of Antibiotic Risk (RADAR)-statistic was proposed to improve efficiency of antibiotic stewardship trials. We studied the behavior of RADAR in a non-inferiority trial in which a beta-lactam monotherapy strategy (BL, n=656) was non-inferior to fluoroquinolone monotherapy (FQL, n=888) for moderately-severe community-acquired pneumonia (CAP) patients.\n\nMETHODSPatients were ranked according to clinical outcome, using five or eight categories, and antibiotic use. RADAR was calculated as the probability that the BL group had a more favorable ranking than the FQL group. To investigate the sensitivity of RADAR to detrimental clinical outcome we simulated increasing rates of 90-day mortality in the BL group and performed the RADAR and non-inferiority analysis.\n\nRESULTSThe RADAR of the BL-group compared to the FQL group was 60.3% (95% confidence interval 57.9%-62.7%) using five and 58.4% (95% CI 56.0%-60.9%) using eight clinical outcome categories, all in favor of BL. Sample sizes for RADAR were 250 and 580 patients per study arm using five or eight clinical outcome categories, respectively, reflecting 38% and 89% of the original non-inferiority sample size calculation. With simulated mortality rates, loss of non-inferiority of the BL-group occurred at a relative risk of 1.125 in the conventional analysis, whereas using RADAR the BL-group lost superiority at a relative risk of mortality of 1.25 and 1.5, with eight and five clinical outcome categories, respectively.\n\nCONCLUSIONSRADAR favored BL over FQL therapy for CAP. Although RADAR required fewer patients than conventional non-inferiority analysis, the statistic was less sensitive to detrimental outcomes.

epidemiology↗

bioRxiv2017

Inferring a qualitative contact rate index of uncertain epidemics

We will inevitably face new epidemic outbreaks where the mechanisms of transmission are still uncertain, making it difficult to obtain quantitative predictions. Thus we present a novel algorithm that qualitatively predicts the start, relative magnitude and decline of uncertain epidemic outbreaks, requiring to know only a few of its \\macroscopic\" parameters. The algorithm is based on estimating exactly the time-varying contact rate of a canonical but time-varying Susceptible-Infected-Recovered epidemic model parametrized to the particular outbreak. The algorithm can also be extended to other canonical epidemic models. Even if dynamics of the outbreak deviates significantly from the underlying epidemic model, we show the predictions of the algorithm remain robust. We validated our algorithm using real time-series data of measles, dengue and the current zika outbreak, comparing its performance to existing algorithms that also use a few macroscopic parameters (e.g., those estimating reproductive numbers) and to those using a thorough understanding of the mechanisms of the epidemic outbreak. We show our algorithm can outperform existing algorithms using a few macroscopic parameters, providing an informative qualitative evaluation of the outbreak.

epidemiology↗

bioRxiv2017

On the heritability of criminal justice processing

An impressive number of researchers have devoted a great amount of effort toward examining various predictors of criminal justice processing outcomes. Indeed, a vast amount of research has examined various individual- and aggregate-level predictors of arrests, incarceration, and sentencing decisions. To this point, less attention has been devoted toward uncovering the relative contribution of genetic and environmental effects on variation in risk for criminal justice processing. As a result, the current study employs a behavioral genetic design in order to help fill this void in the existing literature. Using twin data from a national sample of youth, the current study produced evidence suggesting that genetic factors accounted for at least a portion of variance in risk for incarceration among female twins and probation among male twins. Shared and nonshared environmental influences accounted for the variance in risk for arrest among both female and male twins, probation among female twins, and incarceration among male twins. Ultimately, it appears that risk for contact with the criminal justice system and criminal justice processing is structured by a combination of factors beyond shared cultural and neighborhood environments, and appear to also include genetic factors as well. Moving forward, continuing to not use genetically sensitive research designs capable of estimating the role of genetic and nonshared environmental influences on criminal justice outcomes may result in misleading results.

epidemiology↗

bioRxiv2017

Cardiac events after macrolides or fluoroquinolones in patients hospitalized for community-acquired pneumonia: post-hoc analysis of a cluster-randomized trial

BackgroundGuidelines recommend macrolides and fluoroquinolones in patients hospitalized with community-acquired pneumonia (CAP), but their use has been associated with cardiac events.\n\nObjectiveTo quantify associations between macrolide and fluoroquinolone use and cardiac events in patients hospitalized with CAP in non-ICU wards.\n\nDesignPost-hoc analysis of a cluster-randomized trial\n\nSettingSix hospitals in the Netherlands\n\nPatientsCAP patients admitted to non-ICU wards and without a cardiac event on admission\n\nMeasurementsCause-specific hazard ratios (HRs) were calculated for effects of time-dependent macrolide and fluoroquinolone exposure on cardiac events, defined as occurrence of new or worsening heart failure, arrhythmia, or myocardial ischemia during hospitalization.\n\nResultsCardiac events occurred in 146 (6.9%) of 2,107 patients and included episodes of heart failure (n=101, 4.8%), arrhythmia (n=53, 2.5%), and myocardial ischemia (n=14, 0.7%). Cardiac events occurred in 11 of 207 (5.3%), 18 of 250 (7.2%), and 31 of 277 (11.2%) patients exposed to azithromycin, clarithromycin, and erythromycin for at least one day, respectively, and in 9 of 234 (3.8%), 5 of 194 (2.6%), and 23 of 566 (4.1%) patients exposed to ciprofloxacin, levofloxacin, and moxifloxacin, respectively. Hazard ratios for any cardiac event, adjusted for confounding, were 0.89 (95% confidence interval (CI) 0.48 to 1.67), 1.06 (95% CI 0.61 to 1.83) and 1.68 (95% CI 1.07 to 2.62) for azithromycin, clarithromycin, and erythromycin, respectively, and adjusted hazard ratios were 0.86 (95% CI 0.47 to 1.57), 0.42 (95% CI 0.18 to 0.96) and 0.62 (95% CI 0.39 to 0.99) for ciprofloxacin, levofloxacin, and moxifloxacin, respectively. Erythromycin was associated with an adjusted hazard ratio of 2.08 (95% CI 1.25 to 3.46) for heart failure.\n\nLimitationsPossibility of confounding by indication and observational bias\n\nConclusionsAmong patients with CAP hospitalized to non-ICU wards, erythromycin use was associated with a 68% increased risk of hospital-acquired cardiac events, mainly heart failure. Levofloxacin and moxifloxacin were associated with a lower risk of heart failure.\n\nRegistrationThe original trial was registered under ClinicalTrials.gov Identifier NCT01660204\n\nFunding SourceThe Netherlands Organization for Health Research and Development (ZONmw, Health care efficiency research, project id: 171202002).

epidemiology↗

bioRxiv2017

Environmental cholera (Vibrio cholerae) dynamics in an estuarine system in southern coastal Ecuador

Cholera emergence is strongly linked to local environmental and ecological context. The 1991-2004 pandemic emerged in Peru and spread north into Ecuadors El Oro province, making this a key site for potential re-emergence. Machala, El Oro, is a port city of 250,000, near the Peruvian border. Many livelihoods depend on the estuarine system, from fishing for subsistence and trade, to domestic water use. In 2014, we conducted biweekly sampling for 10 months in five estuarine locations, across a gradient of human use, and ranging from inland to ocean. We measured water characteristics implicated in V. cholerae growth and persistence: pH, temperature, salinity, and algal concentration, and evaluated samples in five months for pathogenic and non-pathogenic Vibrio cholerae, by polymerase chain reaction (PCR). We found environmental persistence of strains O1 and O139, but no evidence for toxigene presence. V. cholerae presence was coupled to algal and salinity concentration, and sites exhibited considerable seasonal and spatial heterogeneity. This study indicates that environmental conditions in Machala are optimal for human cholera re-emergence, with risk peaking during September, and higher risk near urban periphery low-income communities. This highlights a need for surveillance of this coupled cholera- estuarine system to anticipate potential future outbreaks.

epidemiology↗

bioRxiv2017

TitleSurvey on Prevalence of Canine CutaneousMyiasis in Some Selected Kebeles of DireDawa City Administration

A cross sectional study of canine cutaneous myiasis was conducted in five randomly selected kebeles of Dire Dawa Administrative council from December 2009 up to April 2010 to determine the prevalence of canine cutaneous myiasis and to assess factors that determine the occurrence of the disease specifically in dog. Active questionnaire survey among 60 households were used for which 384 dogs were sampled. From a total of 384 dogs, 162 (42.19%) were found harboring the disease cutaneous myiasis among this 120 (74.07%) were infested with the 3rd and 2nd instar larvae of Cordylobiaantropophaga. whereas the remaining 42(25.93%) observed dogs were found infested with cutaneous myiasis. The larvae were identified in Dire Dawa regional diagnostic veterinary parasitology laboratory. Analysis of active questionnaire survey showed that there is no statistically significance difference in the prevalence of disease among different breeds and sexes (P >0.05). In this study, an overall prevalence rate of 162 (42.19%) was found with a statistically significant association among different age groups, housing system and living area (kebele) (P<0.05). Higher prevalence was recorded at 02 kebele (Sabian area) 59 (54.65%), 03 Kebele (Depo and number-one), 43(44.33%), 04 Keble (Gende kore and Greek camp) 33(37.50%), Addis Ketema. 27(51.92%) and05 Keeble (Dechatu) 0(0%). There was 121(49.59%) confined dogs and 41(29.29%) were stray dogs which let out without any control, and puppies of age less than 6 month old (71.56 %), and dogs of age range between 6 months and 18months (79.03%) while those of greater than 18 months (16.43%), were least affected.

epidemiology↗

bioRxiv2017

Fitting mechanistic epidemic models to data: a comparison of simple Markov chain Monte Carlo approaches

BackgroundSimple mechanistic epidemic models are widely used for forecasting and parameter estimation of infectious diseases based on noisy case reporting data. Despite the widespread application of models to emerging infectious diseases, we know little about the comparative performance of standard computational-statistical frameworks in these contexts. Here we build a simple stochastic, discrete-time, discrete-state epidemic model with both process and observation error and use it to characterize the effectiveness of different flavours of Bayesian Markov chain Monte Carlo (MCMC) techniques. We use fits to simulated data, where parameters (and future behaviour) are known to explore the limitations of different platforms and quantify parameter estimation accuracy, forecasting accuracy, and computational efficiency across combinations of modeling decisions (e.g. discrete vs. continuous latent states, levels of stochasticity) and computational platforms (JAGS, NIMBLE, Stan).\n\nResultsModels incorporating at least one source of population-level variation (i.e., dispersion in either the transmission process or the observation process) provide reasonably good forecasts and parameter estimates, while models that incorporate only individual-level variation can lead to inaccurate (or overconfident) results. Models using continuous approximations to the transmission process showed improved computational efficiency without loss of accuracy.\n\nConclusionSimple models of disease transmission and observation can be fitted reliably to simple simulations, as long as population-level variation is taken into account. Continuous approximations can improve computational efficiency using more advanced MCMC techniques.

epidemiology↗

bioRxiv2017

PHESANT: a tool for performing automated phenome scans in UK Biobank

MotivationEpidemiological cohorts typically contain a diverse set of phenotypes such that automation of phenome scans is non-trivial, because they require highly heterogeneous models. For this reason, phenome scans have to date tended to use a smaller homogeneous set of phenotypes that can be analysed in a consistent fashion. We present PHESANT (PHEnome Scan ANalysis Tool), a software package for performing comprehensive phenome scans in UK Biobank.\n\nGeneral featuresPHESANT tests the association of a specified trait with all continuous, integer and categorical variables in UK Biobank, or a specified subset. PHESANT uses a novel rule-based algorithm to determine how to appropriately test each trait, then performs the analyses and produces plots and summary tables.\n\nImplementationThe PHESANT phenome scan is implemented in R. PHESANT includes a novel Javascript D3.js visualization, and accompanying Java code that converts the phenome scan results to the required JavaScript Object Notation (JSON) format.\n\nAVAILABILITYPHESANT is available on GitHub at [https://github.com/MRCIEU/PHESANT]. Git tag v0.2 corresponds to the version presented here.

epidemiology↗

bioRxiv2017

Emergence and persistence of the Chikungunya virus East-Central-South-African genotype in Northeast Brazil

We investigate an outbreak of exanthematous illness in Maceio, Alagoas, using molecular surveillance. Of 273 samples, 76% tested RT-qPCR positive for Chikungunya virus. Phylogenetic analysis reveals that the outbreak was caused by the East-Central-South-African genotype, and that this lineage has likely persisted since mid-2014 in Northeast Brazil.\n\nArticle summary lineTransmission of the Chikungunya virus East-Central-South-African genotype has been ongoing in the Northeast region of Brazil since mid-2014.

epidemiology↗

bioRxiv2017

Socio-environmental and measurement factors drive spatial variation in influenza-like illness

The mechanisms hypothesized to drive spatial heterogeneity in reported influenza activity include: environmental factors, contact patterns, population age structure, and socioeconomic factors linked to healthcare access and quality of life. Harnessing the large volume and high specificity of diagnosis codes in medical claims data for influenza seasons from 2002-2009, we estimate the importance of socio-environmental determinants and measurement-related factors on observed variation in influenza-like illness (ILI) across United States counties. We found that South Atlantic states tended to have higher ILI seasonal intensity, and a combination of transmission, environmental, influenza subtype, socioeconomic and measurement factors explained the variation in seasonal intensity across our study period. Moreover, our models suggest that sentinel surveillance systems should have fixed report locations across years for the most robust inference and prediction, and high volumes of data can offset measurement biases in opportunistic data samples.

epidemiology↗