Modifiable temporal unit problem (MTUP) and its effect on space-time cluster detection

A Bayesian maximum entropy model for predicting tsetse ecological distributions

BACKGROUND

African trypanosomiasis is a tsetse-borne parasitic infection that affects humans, wildlife, and domesticated animals. Tsetse flies are endemic to much of Sub-Saharan Africa and a spatial and temporal understanding of tsetse habitat can aid surveillance and support disease risk management. Problematically, current fine spatial resolution remote sensing data are delivered with a temporal lag and are relatively coarse temporal resolution (e.g., 16 days), which results in disease control models often targeting incorrect places. The goal of this study was to devise a heuristic for identifying tsetse habitat (at a fine spatial resolution) into the future and in the temporal gaps where remote sensing and proximal data fail to supply information.

METHODS

This paper introduces a generalizable and scalable open-access version of the tsetse ecological distribution (TED) model used to predict tsetse distributions across space and time, and contributes a geospatial Bayesian Maximum Entropy (BME) prediction model trained by TED output data to forecast where, herein the Morsitans group of tsetse, persist in Kenya, a method that mitigates the temporal lag problem. This model facilitates identification of tsetse habitat and provides critical information to control tsetse, mitigate the impact of trypanosomiasis on vulnerable human and animal populations, and guide disease minimization in places with ephemeral tsetse. Moreover, this BME analysis is one of the first to utilize cluster and parallel computing along with a Monte Carlo analysis to optimize BME computations. This allows for the analysis of an exceptionally large dataset (over 2 billion data points) at a finer resolution and larger spatiotemporal scale than what had previously been possible.

RESULTS

Under the most conservative assessment for Kenya, the BME kriging analysis showed an overall prediction accuracy of 74.8% (limited to the maximum suitability extent). In predicting tsetse distribution outcomes for the entire country the BME kriging analysis was 97% accurate in its forecasts.

CONCLUSIONS

This work offers a solution to the persistent temporal data gap in accurate and spatially precise rainfall predictions and the delayed processing of remotely sensed data collectively in the - 45 days past to + 180 days future temporal window. As is shown here, the BME model is a reliable alternative for forecasting future tsetse distributions to allow preplanning for tsetse control. Furthermore, this model provides guidance on disease control that would otherwise not be available. These 'big data' BME methods are particularly useful for large domain studies. Considering that past BME studies required reduction of the spatiotemporal grid to facilitate analysis. Both the GEE-TED and the BME libraries have been made open source to enable reproducibility and offer continual updates into the future as new remotely sensed data become available.

Making Space in Geographical Analysis

In this commentary we reflect on the potential and power of geographical analysis, as a set of methods, theoretical approaches, and perspectives, to increase our understanding of how space and place matter for all. We emphasize key aspects of the field, including accessibility, urban change, and spatial interaction and behavior, providing a high-level research agenda that indicates a variety of gaps and routes for future research that will not only lead to more equitable and aware solutions to local and global challenges, but also innovative and novel research methods, concepts, and data. We close with a set of representation and inclusion challenges to our discipline, researchers, and publication outlets.

Methodological Challenges in Spatial and Contextual Exposome-Health Studies

The concept of the exposome encompasses the totality of exposures from a variety of external and internal sources across an individual's life course. The wealth of existing spatial and contextual data makes it appealing to characterize individuals' external exposome to advance our understanding of environmental determinants of health. However, the spatial and contextual exposome is very different from other exposome factors measured at the individual-level as spatial and contextual exposome data are more heterogenous with unique correlation structures and various spatiotemporal scales. These distinctive characteristics lead to multiple unique methodological challenges across different stages of a study. This article provides a review of the existing resources, methods, and tools in the new and developing field for spatial and contextual exposome-health studies focusing on four areas: (1) data engineering, (2) spatiotemporal data linkage, (3) statistical methods for exposome-health association studies, and (4) machine- and deep-learning methods to use spatial and contextual exposome data for disease prediction. A critical analysis of the methodological challenges involved in each of these areas is performed to identify knowledge gaps and address future research needs.

Optimization Framework for Spatiotemporal Analysis Units Based on Floating Car Data

Spatiotemporal scale is a basic component of geographical problems because the size of spatiotemporal units may have a significant impact on the aggregation of spatial data and the corresponding analysis results. However, there is no clear standard for measuring the representativeness of conclusions when geographical data with different temporal and spatial units are used in geographical calculations. Therefore, a spatiotemporal analysis unit optimization framework is proposed to evaluate candidate analysis units using the distribution patterns of spatiotemporal data. The framework relies on Pareto optimality to select the spatiotemporal analysis unit, thereby overcoming the subjectivity and randomness of traditional unit setting methods and mitigating the influence of the modifiable areal unit problem (MAUP) to a certain extent. The framework is used to analyze floating car trajectory data, and the spatiotemporal analysis unit is optimized by using a combination of global spatial autocorrelation coefficients and the coefficients of variation of local spatial autocorrelation. Moreover, based on urban hotspot calculations, the effectiveness of the framework is further verified. The proposed optimization framework for spatiotemporal analysis units based on multiple criteria can provide suitable spatiotemporal analysis scales for studies of geographical phenomena.

Accelerating ethics, empathy, and equity in geographic information science

There has been a groundswell in the support needed to center ethics, empathy, and equity in scientific thought and practice. Drawing on our experience from GIScience, our goal is to accelerate ethical, empathetic, and equitable scientific practices. Many of the opportunities and challenges we outline are broadly applicable and will stimulate the conversations needed to accelerate transformation of science practice and culture. With an emphasis on practical suggestions for reshaping science, we invite all scientists to join in a fundamentally different approach. This paper is a step toward mobilizing the scientific community toward ethics, empathy, and equity by inviting humility, broader measures of excellence and success, diversity in our networks, and the creation of pathways to inclusive education.

Science has traditionally been driven by curiosity and followed one goal: the pursuit of truth and the advancement of knowledge. Recently, ethics, empathy, and equity, which we term “the 3Es,” are emerging as new drivers of research and disrupting established practices. Drawing on our own field of GIScience (geographic information science), our goal is to use the geographic approach to accelerate the response to the 3Es by identifying priority issues and research needs that, if addressed, will advance ethical, empathic, and equitable GIScience. We also aim to stimulate similar responses in other disciplines. Organized around the 3Es we discuss ethical issues arising from locational privacy and cartographic integrity, how our ability to build knowledge that will lead to empathy can be curbed by data that lack representativeness and by inadvertent inferential error, and how GIScientists can lead toward equity by supporting social justice efforts and democratizing access to spatial science and its tools. We conclude with a call to action and invite all scientists to join in a fundamentally different science that responds to the 3Es and mobilizes for change by engaging in humility, broadening measures of excellences and success, diversifying our networks, and creating pathways to inclusive education. Science united around the 3Es is the right response to this unique moment where society and the planet are facing a vast array of challenges that require knowledge, truth, and action.

Identifying a spatial scale for the analysis of residential burglary: An empirical framework based on point pattern analysis

A key issue in the spatial and temporal analysis of residential burglary is the choice of scale: spatial patterns might differ appreciably for different time periods and vary across geographic units of analysis. Based on point pattern analysis of burglary incidents in Columbus, Ohio during a 9-year period, this study develops an empirical framework to identify a useful spatial scale and its dependence on temporal aggregation. Our analysis reveals that residential burglary in Columbus clusters at a characteristic scale of 2.2 km. An ANOVA test shows no significant impact of temporal aggregation on spatial scale of clustering. This study demonstrates the value of point pattern analysis in identifying a scale for the analysis of crime patterns. Furthermore, the characteristic scale of clustering determined using our method has great potential applications: (1) it can reflect the spatial environment of criminogenic processes and thus be used to define the spatial boundary for place-based policing; (2) it can serve as a candidate for the bandwidth (search radius) for hot spot policing; (3) its independence of temporal aggregation implies that police officials need not be concerned about the shifting sizes of risk-areas depending on the time of the year.

Spatiotemporal epidemiology of COVID-19 from an epidemic course perspective

Although coronavirus disease 2019 (COVID-19) remains rampant in many countries, it has recently waned in Sichuan, China. This study examined spatiotemporal variations of the epidemiological characteristics of COVID-19 across its course. Three approaches, i.e. calendar-based, measure-driven and data-driven ones, were applied to all individual cases reported as of 30th November 2020, dividing the COVID-19 pandemic into five periods. A total of 808 people with confirmed diagnosis and 279 asymptomatic cases were reported, the majority of whom were aged 30-49 and <30 years, respectively. The highest risk was seen in Chengdu (capital city), with 411 confirmed and 195 asymptomatic cases. The main sources of infection changed from importation from Hubei Province to importation from other provinces, then local transmission and ultimately importation from foreign countries. The periods highlighted by the three methods presented different epidemic patterns and trends. The calendar-based periods were even with most cases aggregated in the first period, which did not reflect various transmission patterns of COVID-19 due to various sources of infection; the measure-driven and data-driven periods were not consistent with each other, revealing that the effects of implementing prevention measures were reflected on the epidemic trend with a time lag. For example, the decreasing trends of new cases occurred 7, 3 and 4 days later than the firstlevel emergency response, the district-level prevention measures and the second-level emergency response, respectively. This study has advanced our understanding of epidemic course and foreshown all stages of COVID-19 epidemic. Many countries can learn from our findings about what will occur next in their timelines and how to be better prepared.

COVID-19 Deaths in the United States: Shifts in Hot Spots over the Three Phases of the Pandemic and the Spatiotemporally Varying Impact of Pandemic Vulnerability

The geographic areas most impacted by COVID-19 may not remain static because public health measures/behaviors change dynamically, and the impacts of pandemic vulnerability also may vary geographically and temporally. The nature of the pandemic makes spatiotemporal methods essential to understanding the distribution of COVID-19 deaths and developing interventions. This study examines the spatiotemporal trends in COVID-19 death rates in the United States from March 2020 to May 2021 by performing an emerging hot spot analysis (EHSA). It then investigates the effects of the COVID-19 time-dependent and basic social vulnerability factors on COVID-19 death rates using geographically and temporally weighted regression (GTWR). The EHSA results demonstrate that over the three phases of the pandemic (first wave, second wave, and post-vaccine deployment), hot spots have shifted from densely populated cities and the states with a high percentage of socially vulnerable individuals to the states with relatively relaxed social distancing requirements, and then to the states with low vaccination rates. The GTWR results suggest that local infection and testing rates, social distancing interventions, and other social, environmental, and health risk factors show significant associations with COVID-19 death rates, but these associations vary over time and space. These findings can inform public health planning.

Crime Risk Stations: Examining Spatiotemporal Influence of Urban Features through Distance-Aware Risk Signal Functions

Static indicators may fail to capture spatiotemporal differences in the spatial influence of urban features on different crime types. In this study, with a base station analogy, we introduced crime risk stations that conceptualize the spatial influence of urban features as crime risk signals broadcasted throughout a coverage area. We operationalized these risk signals with two novel risk scores, risk strength and risk intensity, obtained from novel distance-aware risk signal functions. With a crime-specific spatiotemporal approach, through a spatiotemporal influence analysis we examined and compared these risk scores for different crime types across various spatiotemporal models. Using a correlation analysis, we examined their relationships with concentrated disadvantage. The results showed that bus stops had relatively lower risk intensity, but higher risk strength, while fast-food restaurants had a higher risk intensity, but a lower risk strength. The correlation analysis identified elevated risk intensity and strength around gas stations in disadvantaged areas during late-night hours and weekends. The results provided empirical evidence for a dynamic spatial influence that changes across space, time, and crime type. The proposed risk functions and risk scores could help in the creation of spatiotemporal crime hotspot maps across cities by accurately quantifying crime risk around urban features.

Do Mobile Phone Data Provide a Better Denominator in Crime Rates and Improve Spatiotemporal Predictions of Crime?

This article assesses whether ambient population is a more suitable population-at-risk measure for crime types with mobile targets than residential population for the purpose of intelligence-led policing applications. Specifically, the potential use of ambient population as a crime rate denominator and predictor for predictive policing models is evaluated, using mobile phone data (with a total of 9,397,473 data points) as a proxy. The results show that ambient population correlates more strongly with crime than residential population. Crime rates based on ambient population designate different problem areas than crime rates based on residential population. The prediction performance of predictive policing models can be improved by using ambient population instead of residential population. These findings support that ambient population is a more suitable population-at-risk measure, as it better reflects the underlying dynamics in spatiotemporal crime trends. Its use has therefore much as-of-yet unused potential not only for criminal research and theory testing, but also for intelligence-led policy and practice.

The impact of the COVID-19 pandemic on people's mobility: A longitudinal study of the U.S. from March to September of 2020

This paper examines changes in people's mobility over a 7-month period (from March 1st to September 30th, 2020) during the COVID-19 pandemic in the U.S. using longitudinal models and county-level mobility data obtained from people's anonymized mobile phone signals. It differentiates two distinct waves of the study period: Wave 1 (March-June) and Wave 2 (June-September). It also analyzes the relationships of these mobility changes with various social, spatial, policy, and political factors. The results indicate that mobility changes in Wave 1 have a V-shaped trend: people's mobility first declined at the early stage of the COVID-19 pandemic (March-April) but quickly recovered to the pre-pandemic mobility levels from April to June. The rates of mobility changes during this period are significantly associated with most of our key variables, including political partisanship, poverty level, and the strictness of mobility restriction policies. For Wave 2, there was very little mobility decline despite the existence of mobility restriction policies and the COVID-19 pandemic becoming more severe. Our findings suggest that restricting people's mobility to control the pandemic may be effective only for a short period, especially in liberal democratic societies. Further, since poor people (who are mostly essential workers) kept traveling during the pandemic, health authorities should pay special attention to these people by implementing policies to mitigate their high COVID-19 exposure risk.

Type 1 Diabetes incidence among youth in Utah: A geographical analysis

Type 1 Diabetes (T1D) poses an increasing threat to public health, as incidence rates continue to rise globally. However, the etiology of T1D is still poorly understood, especially from the perspective of geography. The objective of this research is to examine the incidence of T1D among youth and to identify high-risk clusters and their association with socio-demographic and geographic variables. The study area was the entire state of Utah and included youth with T1D from birth to 19 years of age from 1998 to 2015 (n = 4161). Spatial clustering was measured both globally and locally using the Moran's I statistic and spatial scan statistic. Ordinary least squares (OLS) regression was used to measure the association of high-risk clusters with certain risk factors at the Census Block Group (CBG) level. The mean age at diagnosis was 9.3 years old. The mean incidence rate was 25.67 per 100,000 person-years (95% CI, 24.57-26.75). The incidence rate increased by 14%, from 23.94 per100,000 person-years in 1998 to 27.98 per 100,000 person-years in 2015, with an annual increase of 0.80%. The results of the spatial scan statistic found 42 high-risk clusters throughout the state. OLS regression analysis found a significant association with median household income, population density, and latitude. This study provides evidence that incidence rates of T1D are increasing annually in the state of Utah and that significant geographic high-risk clusters are associated with socio-demographic and geographic factors.

Time to address the spatiotemporal uncertainties in COVID-19 research: Concerns and challenges

In this correspondence, we emphasize methodological caveats of ecological studies assessing associations between COVID-19 and its physical and social environmental determinants. First, we stress that inference is error-prone due to the modifiable areal unit problem and the modifiable temporal unit problem. The possibility of confounding from using aggregated data is substantial due to the neglect of person-level factors. Second, studying the viral transmission of COVID-19 solely on people's residential neighborhoods is problematic because people are also exposed to nonhome locations and environments en-route along their daily mobility path. We caution against an uncritical application of aggregated data and reiterate the importance of stronger research designs (e.g., case-control studies) on an individual level. To address environmental contextual uncertainties due to people's day-to-day mobility, we call for people-centered studies with mobile phone data.

Modelling the Relationship between Rainfall and Mental Health Using Different Spatial and Temporal Units

Drought is thought to impact upon the mental health of agricultural communities, but studies of this relationship have reported inconsistent results. A source of inconsistency could be the aggregation of data by a single spatiotemporal unit of analysis, which induces the modifiable areal and temporal unit problems. To investigate this, mental health-related emergency department (MHED) presentations among residents of the Wheat Belt region of Western Australia, between 2002 and 2017, were examined. Average daily rainfall was used as a measure of drought. Associations between MHED presentations and rainfall were estimated based on various spatial aggregations of underlying data, at multiple temporal windows. Wide variation amongst results was observed. Despite this, two key features were found: Associations between MHED presentations and rainfall were generally positive when rainfall was measured in summer months (rate ratios up to 1.05 per 0.5 mm of daily rainfall) and generally negative when rainfall was measured in winter months (rate ratios as low as 0.96 per 0.5 mm of daily rainfall). These results demonstrate that the association between drought and mental health is quantifiable; however, the effect size is small and varies depending on the spatial and temporal arrangement of the underlying data. To improve understanding of this association, more studies should be undertaken with longer time spans and examining specific mental health outcomes, using a wide variety of spatiotemporal units.

Methodological limitations in studies assessing the effects of environmental and socioeconomic variables on the spread of COVID-19: a systematic review

BACKGROUND

While numerous studies have assessed the effects of environmental (meteorological variables and air pollutants) and socioeconomic variables on the spread of the COVID-19 pandemic, many of them, however, have significant methodological limitations and errors that could call their results into question. Our main objective in this paper is to assess the methodological limitations in studies that evaluated the effects of environmental and socioeconomic variables on the spread of COVID-19.

MAIN BODY

We carried out a systematic review by conducting searches in the online databases PubMed, Web of Science and Scopus up to December 31, 2020. We first excluded those studies that did not deal with SAR-CoV-2 or COVID-19, preprints, comments, opinion or purely narrative papers, reviews and systematic literature reviews. Among the eligible full-text articles, we then excluded articles that were purely descriptive and those that did not include any type of regression model. We evaluated the risk of bias in six domains: confounding bias, control for population, control of spatial and/or temporal dependence, control of non-linearities, measurement errors and statistical model. Of the 5631 abstracts initially identified, we were left with 132 studies on which to carry out the qualitative synthesis. Of the 132 eligible studies, we evaluated 63.64% of the studies as high risk of bias, 19.70% as moderate risk of bias and 16.67% as low risk of bias.

CONCLUSIONS

All the studies we have reviewed, to a greater or lesser extent, have methodological limitations. These limitations prevent conclusions being drawn concerning the effects environmental (meteorological and air pollutants) and socioeconomic variables have had on COVID-19 outcomes. However, we dare to argue that the effects of these variables, if they exist, would be indirect, based on their relationship with social contact.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1186/s12302-021-00550-7.

The impact of modelling choices on modelling outcomes: a spatio-temporal study of the association between COVID-19 spread and environmental conditions in Catalonia (Spain)

The choices that researchers make while conducting a statistical analysis usually have a notable impact on the results. This fact has become evident in the ongoing research of the association between the environment and the evolution of the coronavirus disease 2019 (COVID-19) pandemic, in light of the hundreds of contradictory studies that have already been published on this issue in just a few months. In this paper, a COVID-19 dataset containing the number of daily cases registered in the regions of Catalonia (Spain) since the start of the pandemic to the end of August 2020 is analysed using statistical models of diverse levels of complexity. Specifically, the possible effect of several environmental variables (solar exposure, mean temperature, and wind speed) on the number of cases is assessed. Thus, the first objective of the paper is to show how the choice of a certain type of statistical model to conduct the analysis can have a severe impact on the associations that are inferred between the covariates and the response variable. Secondly, it is shown how the use of spatio-temporal models accounting for the nature of the data allows understanding the evolution of the pandemic in space and time. The results suggest that even though the models fitted to the data correctly capture the evolution of COVID-19 in space and time, determining whether there is an association between the spread of the pandemic and certain environmental conditions is complex, as it is severely affected by the choice of the model.

Understanding the Epidemic Course in Order to Improve Epidemic Forecasting

The epidemic course of the severe acute respiratory syndrome (SARS) has been differently divided according to its transmission pattern and the infection and mortality status. Unfortunately, such efforts for the coronavirus disease 2019 (COVID-19) have been lacking. Does every epidemic have a unique epidemic course? Can we coordinate two arbitrary courses into an integrated course, which could better reflect a common real-world progression pattern of the epidemics? To what degree can such arbitrary divisions help predict future trends of the COVID-19 pandemic and future epidemics? Spatial lifecourse epidemiology provides a new perspective to understand the course of epidemics, especially pandemics, and a new toolkit to predict the course of future epidemics on the basis of big data. In the present data-driven era, data should be integrated to inform us how the epidemic is transmitting at the present moment, how it will transmit at the next moment, and which interventions would be most cost-effective to curb the epidemic. Both national and international legislations are needed to facilitate the integration of relevant policies of data sharing and confidentiality protection into the current pandemic preparedness guidelines.

Effects of Data Aggregation on Time Series Analysis of Seasonal Infections

Time series analysis in epidemiological studies is typically conducted on aggregated counts, although data tend to be collected at finer temporal resolutions. The decision to aggregate data is rarely discussed in epidemiological literature although it has been shown to impact model results. We present a critical thinking process for making decisions about data aggregation in time series analysis of seasonal infections. We systematically build a harmonic regression model to characterize peak timing and amplitude of three respiratory and enteric infections that have different seasonal patterns and incidence. We show that irregularities introduced when aggregating data must be controlled during modeling to prevent erroneous results. Aggregation irregularities had a minimal impact on the estimates of trend, amplitude, and peak timing for daily and weekly data regardless of the disease. However, estimates of peak timing of the more common infections changed by as much as 2.5 months when controlling for monthly data irregularities. Building a systematic model that controls for data irregularities is essential to accurately characterize temporal patterns of infections. With the urgent need to characterize temporal patterns of novel infections, such as COVID-19, this tutorial is timely and highly valuable for experts in many disciplines.

Using time-series analysis techniques to enhance the understanding of musculoskeletal injury in Thoroughbred racehorses

BACKGROUND

Many racing jurisdictions monitor race-day musculoskeletal injury (MSI) but fail to evaluate injuries occurring during training. Additionally, previous risk factor analyses have failed to explore temporal trends in injury occurrence.

OBJECTIVES

To use time-series analysis techniques to identify trends, cyclicity and peaks in MSI incidence, in racehorses training and racing at the Hong Kong Jockey Club (HKJC) from July 2010 to June 2018.

STUDY DESIGN

Retrospective longitudinal study.

METHODS

The monthly incidence of all MSI, superficial digital flexor tendon (SDFT) injury, suspensory ligament injury and appendicular skeletal fracture occurring in training and racing were collected from veterinary records. The number of horses in training was collated monthly from trainer records. Time-series analysis techniques were used to describe trends and cyclical patterns for injury types. For each injury, incidence risks above the 90th percentile were identified as peaks in incidence.

RESULTS

A total of 1471 injuries were recorded over eight racing seasons; 605 fractures (41.1%), 550 SDFT injuries (37.4%), and 316 suspensory ligament injuries (21.5%). Evidence of seasonality was detected in fracture incidence risk; increasing from October (median 0.25 per 1000 horses) until May (median 0.71 per 1000 horses), coinciding with the racing season (ending mid-July). Elevated incidence of MSI occurred throughout 2012; however, the greatest incidence risks of SDFT (14.8 per 1000 horses) and fracture (1.3 per 1000 horses) occurred since 2017.

MAIN LIMITATIONS

Monthly (opposed to daily) incidence risk of injury reduced the resolution of the data. Additionally, fracture was not described according to bone or fracture type, which may have confounded overall trends.

CONCLUSIONS

Evidence for seasonal variation in the incidence of fracture occurrence has been demonstrated. Based on using time-series techniques, further epidemiological studies, retrospectively targeting periods of high peaks in injury incidence risk could be used to aid identification of risk factors for injury.

Space-time clustering of childhood leukemia in Colombia: a nationwide study

BACKGROUND

Leukemia is the most common cancer in childhood. The estimated incidence rate of childhood leukemia in Colombia is one of the highest in America and little is known about its spatial distribution.

PURPOSE

To explore the presence of space-time clustering of childhood leukemia in Colombia.

METHODS

We included children less than 15 years of age with confirmed diagnosis of acute leukemia reported to the national surveillance system for cancer between 2009 and 2017. Kulldorff's spatio-temporal scan statistics were used with municipality and year of diagnosis as units for spatial and temporal analysis.

RESULTS

There were 3846 cases of childhood leukemia between 2009 and 2017 with a specific mean incidence rate of 33 cases per million person-years in children aged 0-14 years. We identified five spatial clusters of childhood leukemia in different regions of the country and specific time clustering during the study period.

CONCLUSION

Childhood leukemia seems to cluster in space and time in some regions of Colombia suggesting a common etiologic factor or conditions to be studied.

Public health in genetic spaces: a statistical framework to optimize cluster-based outbreak detection

Genetic clustering is a popular method for characterizing variation in transmission rates for rapidly evolving viruses, and could potentially be used to detect outbreaks in 'near real time'. However, the statistical properties of clustering are poorly understood in this context, and there are no objective guidelines for setting clustering criteria. Here, we develop a new statistical framework to optimize a genetic clustering method based on the ability to forecast new cases. We analysed the pairwise Tamura-Nei (TN93) genetic distances for anonymized HIV-1 subtype B pol sequences from Seattle (n = 1,653) and Middle Tennessee, USA (n = 2,779), and northern Alberta, Canada (n = 809). Under varying TN93 thresholds, we fit two models to the distributions of new cases relative to clusters of known cases: 1, a null model that assumes cluster growth is strictly proportional to cluster size, i.e. no variation in transmission rates among individuals; and 2, a weighted model that incorporates individual-level covariates, such as recency of diagnosis. The optimal threshold maximizes the difference in information loss between models, where covariates are used most effectively. Optimal TN93 thresholds varied substantially between data sets, e.g. 0.0104 in Alberta and 0.016 in Seattle and Tennessee, such that the optimum for one population would potentially misdirect prevention efforts in another. For a given population, the range of thresholds where the weighted model conferred greater predictive accuracy tended to be narrow (±0.005 units), and the optimal threshold tended to be stable over time. Our framework also indicated that variation in the recency of HIV diagnosis among clusters was significantly more predictive of new cases than sample collection dates (ΔAIC > 50). These results suggest that one cannot rely on historical precedence or convention to configure genetic clustering methods for public health applications, especially when translating methods between settings of low-level and generalized epidemics. Our framework not only enables investigators to calibrate a clustering method to a specific public health setting, but also provides a variable selection procedure to evaluate different predictive models of cluster growth.

Dynamic Changes of NDVI in the Growing Season of the Tibetan Plateau During the Past 17 Years and Its Response to Climate Change

The fragile alpine vegetation in the Tibetan Plateau (TP) is very sensitive to environmental changes, making TP one of the hotspots for studying the response of vegetation to climate change. Existing studies lack detailed description of the response of vegetation to different climatic factors using the method of multiple nested time series analysis and the method of grey correlation analysis. In this paper, based on the Normalized Difference Vegetation Index (NDVI) of TP in the growing season calculated from the MOD09A1 data product of Moderate-resolution Imaging Spectroradiometer (MODIS), the method of multiple nested time series analysis is adopted to study the variation trends of NDVI in recent 17 years, and the lag time of NDVI to climate change is analyzed using the method of Grey Relational Analysis (GRA). Finally, the characteristics of temporal and spatial differences of NDVI to different climate factors are summarized. The results indicate that: (1) the spatial distribution of NDVI values in the growing season shows a trend of decreasing from east to west, and from north to south, with a change rate of −0.13/10° E and −0.30/10° N, respectively. (2) From 2001 to 2017, the NDVI in the TP shows a slight trend of increase, with a growth rate of 0.01/10a. (3) The lag time of NDVI to air temperature is not obvious, while the NDVI response lags behind cumulative precipitation by zero to one month, relative humidity by two months, and sunshine duration by three months. (4) The effects of different climatic factors on NDVI are significantly different with the increase of the study period.

Decomposition of Repulsive Clusters in Complex Point Processes with Heterogeneous Components

The decomposition of a point process is useful for the analysis of spatial patterns and in the discovery of potential mechanisms of geographic phenomena. However, when a local repulsive cluster is present in a complex heterogeneous point process, the traditional solution, which is based on clustering, may be invalid for decomposition because a repulsive pattern is not subject to a specific probability distribution function and the effects of aggregative and repulsive components may be counterbalanced. To solve this problem, this paper proposes a method of decomposing repulsive clusters in complex point processes with multiple heterogeneous components. A repulsive cluster is defined as a set of repulsive density-connected points that are separated by a certain distance at a small scale and aggregated at a large scale simultaneously. The H-function is used to identify repulsive clusters by determining the repulsive distance and extracting repulsive points for further clustering. Through simulation experiments based on three datasets, the proposed method has been shown to effectively perform repulsive cluster decomposition in heterogeneous point processes. A case study of the point of interest (POI) dataset in Beijing also indicates that the method can identify meaningful repulsive clusters from types of POIs that represent different service characteristics of shops in different local regions.

Pathogen seasonality and links with weather in England and Wales: a big data time series analysis

Background

Many infectious diseases of public health importance display annual seasonal patterns in their incidence. We aimed to systematically document the seasonality of several human infectious disease pathogens in England and Wales, highlighting those organisms that appear weather-sensitive and therefore may be influenced by climate change in the future.

Methods

Data on infections in England and Wales from 1989 to 2014 were extracted from the Public Health England (PHE) SGSS surveillance database. We conducted a weekly, monthly and quarterly time series analysis of 277 pathogen serotypes. Each organism’s time series was forecasted using the TBATS package in R, with seasonality detected using model fit statistics. Meteorological data hosted on the MEDMI Platform were extracted at a monthly resolution for 2001–2011. The organisms were then clustered by K-means into two groups based on cross correlation coefficients with the weather variables.

Results

Examination of 12.9 million infection episodes found seasonal components in 91/277 (33%) organism serotypes. Salmonella showed seasonal and non-seasonal serotypes. These results were visualised in an online Rshiny application. Seasonal organisms were then clustered into two groups based on their correlations with weather. Group 1 had positive correlations with temperature (max, mean and min), sunshine and vapour pressure and inverse correlations with mean wind speed, relative humidity, ground frost and air frost. Group 2 had the opposite but also slight positive correlations with rainfall (mm, > 1 mm, > 10 mm).

Conclusions

The detection of seasonality in pathogen time series data and the identification of relevant weather predictors can improve forecasting and public health planning. Big data analytics and online visualisation allow the relationship between pathogen incidence and weather patterns to be clarified.

Electronic supplementary material

The online version of this article (10.1186/s12889-018-5931-6) contains supplementary material, which is available to authorized users.

Spatiotemporal Modeling for Fine-Scale Maps of Regional Malaria Endemicity and Its Implications for Transitional Complexities in a Routine Surveillance Network in Western Cambodia

Due to the associated and substantial efforts of many stakeholders involved in malaria containment, the disease burden of malaria has dramatically decreased in many malaria-endemic countries in recent years. Some decades after the past efforts of the global malaria eradication program, malaria elimination has again featured on the global health agenda. While risk distribution modeling and a mapping approach are effective tools to assist with the efficient allocation of limited health-care resources, these methods need some adjustment and reexamination in accordance with changes occurring in relation to malaria elimination. Limited available data, fine-scale data inaccessibility (for example, household or individual case data), and the lack of reliable data due to inefficiencies within the routine surveillance system, make it difficult to create reliable risk maps for decision-makers or health-care practitioners in the field. Furthermore, the risk of malaria may dynamically change due to various factors such as the progress of containment interventions and environmental changes. To address the complex and dynamic nature of situations in low-to-moderate malaria transmission settings, we built a spatiotemporal model of a standardized morbidity ratio (SMR) of malaria incidence, calculated through annual parasite incidence, using routinely reported surveillance data in combination with environmental indices such as remote sensing data, and the non-environmental regional containment status, to create fine-scale risk maps. A hierarchical Bayesian frame was employed to fit the transitioning malaria risk data onto the map. The model was set to estimate the SMRs of every study location at specific time intervals within its uncertainty range. Using the spatial interpolation of estimated SMRs at village level, we created fine-scale maps of two provinces in western Cambodia at specific time intervals. The maps presented different patterns of malaria risk distribution at specific time intervals. Moreover, the visualized weights estimated using the risk model, and the structure of the routine surveillance network, represent the transitional complexities emerging from ever-changing regional endemic situations.

An empirical method to cluster objective nebulizer adherence data among adults with cystic fibrosis

BACKGROUND

The purpose of using preventative inhaled treatments in cystic fibrosis is to improve health outcomes. Therefore, understanding the relationship between adherence to treatment and health outcome is crucial. Temporal variability, as well as absolute magnitude of adherence affects health outcomes, and there is likely to be a threshold effect in the relationship between adherence and outcomes. We therefore propose a pragmatic algorithm-based clustering method of objective nebulizer adherence data to better understand this relationship, and potentially, to guide clinical decisions.

METHODS TO CLUSTER ADHERENCE DATA

This clustering method consists of three related steps. The first step is to split adherence data for the previous 12 months into four 3-monthly sections. The second step is to calculate mean adherence for each section and to score the section based on mean adherence. The third step is to aggregate the individual scores to determine the final cluster ("cluster 1" = very low adherence; "cluster 2" = low adherence; "cluster 3" = moderate adherence; "cluster 4" = high adherence), and taking into account adherence trend as represented by sequential individual scores. The individual scores should be displayed along with the final cluster for clinicians to fully understand the adherence data.

THREE ILLUSTRATIVE CASES

We present three cases to illustrate the use of the proposed clustering method.

CONCLUSION

This pragmatic clustering method can deal with adherence data of variable duration (ie, can be used even if 12 months' worth of data are unavailable) and can cluster adherence data in real time. Empirical support for some of the clustering parameters is not yet available, but the suggested classifications provide a structure to investigate parameters in future prospective datasets in which there are accurate measurements of nebulizer adherence and health outcomes.

Fine-Scale Mapping by Spatial Risk Distribution Modeling for Regional Malaria Endemicity and Its Implications under the Low-to-Moderate Transmission Setting in Western Cambodia

The disease burden of malaria has decreased as malaria elimination efforts progress. The mapping approach that uses spatial risk distribution modeling needs some adjustment and reinvestigation in accordance with situational changes. Here we applied a mathematical modeling approach for standardized morbidity ratio (SMR) calculated by annual parasite incidence using routinely aggregated surveillance reports, environmental data such as remote sensing data, and non-environmental anthropogenic data to create fine-scale spatial risk distribution maps of western Cambodia. Furthermore, we incorporated a combination of containment status indicators into the model to demonstrate spatial heterogeneities of the relationship between containment status and risks. The explanatory model was fitted to estimate the SMR of each area (adjusted Pearson correlation coefficient R² = 0.774; Akaike information criterion AIC = 149.423). A Bayesian modeling framework was applied to estimate the uncertainty of the model and cross-scale predictions. Fine-scale maps were created by the spatial interpolation of estimated SMRs at each village. Compared with geocoded case data, corresponding predicted values showed conformity [Spearman’s rank correlation r = 0.662 in the inverse distance weighed interpolation and 0.645 in ordinal kriging (95% confidence intervals of 0.414–0.827 and 0.368–0.813, respectively), Welch’s t-test; Not significant]. The proposed approach successfully explained regional malaria risks and fine-scale risk maps were created under low-to-moderate malaria transmission settings where reinvestigations of existing risk modeling approaches were needed. Moreover, different representations of simulated outcomes of containment status indicators for respective areas provided useful insights for tailored interventional planning, considering regional malaria endemicity.