Kulldorff's Spatial Scan Statistic Research Articles

Evidence of spatial clustering of childhood acute lymphoblastic leukemia cases in Greater Mexico City: report from the Mexican Inter-Institutional Group for the identification of the causes of childhood leukemia.

A heterogeneous geographic distribution of childhood acute lymphoblastic leukemia (ALL) cases has been described, possibly, related to the presence of different environmental factors. The aim of the present study was to explore the geographical distribution of childhood ALL cases in Greater Mexico City (GMC). A population-based case-control study was conducted. Children <18 years old, newly diagnosed with ALL and residents of GMC were included. Controls were patients without leukemia recruited from second-level public hospitals, frequency-matched by sex, age, and health institution with the cases. The residence address where the patients lived during the last year before diagnosis (cases) or the interview (controls) was used for geolocation. Kulldorff's spatial scan statistic was used to detect spatial clusters (SCs). Relative risks (RR), associated p-value and number of cases included for each cluster were obtained. A total of 1054 cases with ALL were analyzed. Of these, 408 (38.7%) were distributed across eight SCs detected. A relative risk of 1.61 (p<0.0001) was observed for the main cluster. Similar results were noted for the remaining seven ones. Additionally, a proximity between SCs, electrical installations and petrochemical facilities was observed. The identification of SCs in certain regions of GMC suggest the possible role of environmental factors in the etiology of childhood ALL.

Clustering of systemic lupus erythematosus mortality in southwestern Spain.

To analyse time trends in systemic lupus erythematosus (SLE) mortality and explore possible provincial clustering of SLE mortality in Spain (2001-2020). We conducted an ecological study using deaths registered in SLE at the Spanish National Institute of Statistics between 2001 and 2020. Jointpoint regression models have been used to evaluate temporal trends. To analyse the spatial pattern of SLE mortality in men and women in Spain, crude rates, age-standardised mortality rates (ASMRs), smooth relative risk (RR) and posterior probabilities (PP) for RR greater than one for the period 2001-2020 were calculated. The Global Moran I index was used to assess the existence of global spatial autocorrelation. Local indicators of spatial association (LISA) and Kulldorff's spatial scan statistic were used to identify clusters. Over the 20 years analyzed in this study, the SLE average ASMR for the period was 2.7 for women and 0.7 for men, with a sex ratio (female/male) of 3.8. In men, no province showed a RR>1. Conversely, in women, eight provinces showed values of RR> 1 with a PP greater than 0.8 (Seville, Cadiz, Huelva and Murcia in the south, Barcelona, Zaragoza, Huesca and Leon in the north). In men, neither of the two methods detected a clustering of provinces. However, in women, both methods identified a cluster of provinces located in the southwest of the country (Huelva, Cádiz, Seville and Malaga) as a cluster with significant excess mortality. In the second cluster (centred on the province of Huelva) obtained with the Kulldorff method, two more provinces were added (Badajoz and Cordoba, also located in the southwest). We detected a cluster of provinces with an excess risk of female SLE mortality in the southwest of Spain.

Spatial clustering of amyotrophic lateral sclerosis in Sardinia, Italy: The contribution of age, sex, and genetic factors.

Several microgeographic clusters of higher/lower incidence of amyotrophic lateral sclerosis (ALS) have been identified worldwide. Differences in the distribution of local factors were proposed to explain the excess ALS risk, whereas the contribution of known genetic/epigenetic factors remains unclear. The aim is to identify restricted areas of higher risk in Sardinia and to assess whether age, sex, and the most common causative genetic mutations in Sardinia (C9orf72 and TARDBP mutations) contributed to the variation in the ALS risk. We performed an ad hoc analysis of the 10-y population-based incident cohort of ALS cases from a recent study of a large Sardinian area. Cluster analysis was performed by age- and sex-adjusted Kulldorff's spatial scan statistic. We identified a statistically significant cluster of higher ALS incidence in a relatively large area including 34 municipalities and >100,000 individuals. The investigated genetic mutations were more frequent in the cluster area than outside. Regardless of the genetic mutations, the excess of ALS risk was significantly associated with either sex or with age ≥ 65 y. Finally, an additive interaction between older age and male sex contributed to the excess of ALS risk in the cluster area but not outside. Our analysis demonstrated that known genetic factors, age, and sex may contribute to microgeographic variation in ALS incidence. The significant additive interaction between older age and male sex we found in the high-incidence cluster could suggest the presence of a third factor connecting the analyzed risk factors.

Spatial clustering of codeine use and its association with depression: a geospatial analysis of nationally representative South African data

• We spatially identified geographical clusters of codeine use in South Africa, and investigated onset of depression overtime among depression-free individuals in communities with high concentration of codeine use. • The clusters of codeine use were found to be concentrated in the northern parts of South Africa. • Residing in a community with high concentration codeine use was associated with higher subsequent depression incidence. • Whilst opioid crisis has been believed to be epidemic in the global north particularly USA, it is also taking place in sub-Saharan Africa. There is an alarming trend in sub-Saharan Africa in the use of codeine-containing pharmaceuticals, but its risk of common comorbidities, such as mental health, remains unquantified at a national-level. We investigated the relationship between codeine use and onset of depression in South Africa at a population-level. We used nationally representative panel data from the South Africa National Income Dynamic Study to investigate the relationship between exposure to codeine use in the community (i.e. residing in a codeine hotspot) and onset of depression. Geographical hotspots of codeine use were identified using Kulldorff spatial scan statistic in SaTScan. We quantified depression onset at Wave 5 (year 2017) between individuals residing inside and outside a codeine hotspot who were depression-free at enrolment (Wave 4: 2015) using generalised estimating equation (GEE) regression models. We identified four statistically significant hotspots of codeine use, mostly located at the northern part of the country. Among 13,020 participants who were depression-free at enrolment, residing within a codeine use hotspot was significantly associated with higher subsequent onset of depression (aRR=1.21, 95% CI:1.08-1.44). Data on diagnosis of depressive disorder were not available. South Africa, a resource scarce nation with chronically limited mental health services, is not spared from the global opioid epidemic and its impact on depression. Targeted scale-up access to agonist therapy to effectively treat (opioid) addiction in communities at risk for high exposure to codeine use could reduce the risk of subsequent mental health challenges.

Syndemics of intimate partner violence among women in HIV endemic South Africa: geospatial analysis of nationally representative data

Despite some improvement in lowering HIV incidence, HIV-related challenges, such as intimate partner violence (IPV), remain unacceptably high among women in South Africa. For decades, researchers and activists have pointed to the complex and intertwined reality of the substance abuse, violence and AIDS (SAVA) syndemic that endangers women. However, more recent systematic review/meta-analysis evidence points to inconclusive association between IPV and alcohol use. Furthermore, much of the evidence is often non-population-based that focuses on the co-occurrence rather than synergistic SAVA interaction. In this study, using the latest data from the South Africa Demographic and Health Survey (SA-DHS), we identified geographic synergistic clustering of IPV associated with HIV and substance abuse in South Africa as a measure of population-level interactions among these factors. The SA-DHS is a nationally representative sample that includes wide-ranging data on health, social challenges and household geo-locations of 5,874 women who participated in the domestic violence module. First, geographical IPV, harmful alcohol use (as the substance abuse measure available in SA-DHS) and HIV clusters were identified using the Kulldorff spatial scan statistic in SaTScan. Second, synergistic interactions related to recent IPV (i.e. recent physical, sexual, emotional violence during the last 12 months) with harmful alcohol use and HIV challenge were measured using RERI [Relative excess risk due to interaction], AP [attributable proportion] and S [Synergy index]. In our results, we spatially identified geographical physical IPV syndemic interactions in parts of the Eastern Cape/Free State Provinces (RERI = 4.42 [95% CI: 2.34–6.51], AP = 0.56 [95% CI: 0.44–0.68], S = 2.77 [95% CI: 2.01–3.84], but not in other forms of IPV. Although IPV, based on decade old concept of SAVA syndemic, was less common/widespread than expected from the national scale population-based data, we identified population-level physical violence syndemic occurring in South Africa. Our study highlights the need to prioritize public health response targeting vulnerable populations residing in these high-risk areas of syndemic mechanisms linking these synergistic epidemics that women face in South Africa.

Spatial analysis of mental health and suicide clustering among older adults in North Carolina: An exploratory analysis

Late-in-life suicide is a major public health issue, with adults over 65 comprising 18% of all suicide deaths. Yet, little research has identified geographic concentrations of older populations at-risk for the psychological conditions in the pathway of suicide. This research identified spatial clustering of suicide, anxiety, depression, and self-injury among older adults in North Carolina from 2009-18 and the individual and area-level factors influencing high-risk clusters. Kulldorff's spatial scan statistic identified significant spatial clusters of suicides, using data collected from the violent death reporting system, and anxiety, depression and self-injury, using emergency department admissions data for older adults aged 65 years and over. Logistic regression and generalized estimating equations were used to examine the influence of multiple contextual variables (e.g., rural-urban commuting area codes, racial segregation) and compositional variables (e.g., race, insurance) within high-risk cluster locations. High-risk suicide, anxiety, and depression clusters were identified in the western and northwestern regions of the state, whereas self-injury clusters were predominantly located in central and southeastern North Carolina. Low-income communities were associated with a higher risk of clustering for all mental health outcomes, and primarily Black communities were associated with an increased risk of self-injury clustering. Findings can be applied to public policy to focus suicide prevention efforts on older adults, particularly in low-income communities in rural Appalachia and the central, southeastern portion of North Carolina.

Less is more: Developing an approach for assessing clustering at the lower administrative boundaries that increases the yield of active screening for leprosy in Bihar, India.

In India, leprosy clusters at hamlet level but detailed information is lacking. We aim to identify high-incidence hamlets to be targeted for active screening and post-exposure prophylaxis. We paid home visits to a cohort of leprosy patients registered between April 1st, 2020, and March 31st, 2022. Patients were interviewed and household members were screened for leprosy. We used an open-source app(ODK) to collect data on patients' mobility, screening results of household members, and geographic coordinates of their households. Clustering was analysed with Kulldorff's spatial scan statistic(SaTScan). Outlines of hamlets and population estimates were obtained through an open-source high-resolution population density map(https://data.humdata.org), using kernel density estimation in QGIS, an open-source software. We enrolled 169 patients and screened 1,044 household contacts in Bisfi and Benipatti blocks of Bihar. Median number of years of residing in the village was 17, interquartile range(IQR)12-30. There were 11 new leprosy cases among 658 household contacts examined(167 per 10,000), of which seven had paucibacillary leprosy, one was a child under 14 years, and none had visible disabilities. We identified 739 hamlets with a total population of 802,788(median 163, IQR 65-774). There were five high incidence clusters including 12% of the population and 46%(78/169) of the leprosy cases. One highly significant cluster with a relative risk (RR) of 4.7(p<0.0001) included 32 hamlets and 27 cases in 33,609 population. A second highly significant cluster included 32 hamlets and 24 cases in 33,809 population with a RR of 4.1(p<0.001). The third highly significant cluster included 16 hamlets and 17 cases in 19,659 population with a RR of 4.8(p<0.001). High-risk clusters still need to be screened door-to-door. We found a high yield of active household contact screening. Our tools for identifying high-incidence hamlets appear effective. Focusing labour-intensive interventions such as door-to-door screening on such hamlets could increase efficiency.

Spatiotemporal analysis of African swine fever outbreaks on South African smallholder farms, 1993-2018.

African swine fever (ASF) is a contagious viral disease of swine worldwide. ASF in South Africa has for many years been confined to a controlled area in the northeast of the country that was proclaimed in 1935. Since 2012, outbreaks are more likely to occur in the historically ASF-free area. This study aimed to analyse the spatial and spatiotemporal structure of ASF outbreaks in South Africa between 1993 and 2018. Global space-time clustering of ASF outbreaks was investigated by the Diggle space-time K-function while Kulldorff's spatial scan statistic was applied to detect local cluster of ASF outbreaks. Globally, ASF outbreaks exhibit statistically significant spatial clustering. They have shown a significant negative space-time interaction at month scale (p = 0.003) but no significant space-time interaction at year scale (p = 0.577), revealing strong evidence that ASF cases that are close in space occur in months which are close and vice versa. In studying local area space-time clustering at both month and year scale, three significant local clusters associated with high-rate were detected. These clusters are localised in both the ASF-controlled area and outside the controlled area with radius varying from 60.84 km up to 271.43 km and risk ratio varying from 6.61 up to 17.70. At month scale, clusters with more outbreaks were observed between June 2017 and August 2017 and involved 22 outbreaks followed by the cluster that involved 13 outbreaks in January 2012. These results show the need to maintain high biosecurity standards on pig farms in both inside and outside the ASF-controlled areas.

Spatial trends and projections of chronic malnutrition among children under 5 years of age in Ethiopia from 2011 to 2019: a geographically weighted regression analysis

IntroductionUndernutrition is a serious global health issue, and stunting is a key indicator of children's nutritional status which results from long-term deprivation of basic needs. Ethiopia, the largest and most populous country in Sub-Saharan Africa, has the greatest rate of stunting among children under the age of five, yet the problem is unevenly distributed across the country. Thus, we investigate spatial heterogeneity and explore spatial projection of stunting among under-five children. Further, spatial predictors of stunting were assessed using geospatial regression models.MethodsThe Ethiopia Demographic and Health Surveys (EDHS) data from 2011, 2016, and 2019 were examined using a geostatistical technique that took into account spatial autocorrelation. Ordinary kriging was used to interpolate stunting data, and Kulldorff spatial scan statistics were used to identify spatial clusters with high and low stunting prevalence. In spatial regression modeling, the ordinary least square (OLS) model was employed to investigate spatial predictors of stunting and to examine local spatial variations geographically weighted regression (GWR) and multiscale geographically weighted regression (MGWR) models were employed.ResultsOverall, stunting prevalence was decreased from 44.42% [95%, CI: 0.425–0.444] in 2011 to 36.77% [95%, CI: 0.349–0.375] in 2019. Across three waves of EDHS, clusters with a high prevalence of stunting in children under 5 years were consistently observed in northern Ethiopia stretching in Tigray, Amhara, Afar, and Benishangul-Gumuz. Another area of very high stunting incidence was observed in the Southern parts of Ethiopia and the Somali region of Ethiopia. Our spatial regression analysis revealed that the observed geographical variation of under-five stunting significantly correlated with poor sanitation, poor wealth index, inadequate diet, residency, and mothers' education.ConclusionsIn Ethiopia, substantial progress has been made in decreasing stunting among children under the age of 5 years; although disparities varied in some areas and districts between surveys, the pattern generally remained constant over time. These findings suggest a need for region and district-specific policies where priority should be given to children in areas where most likely to exhibit high-risk stunting.

An Exploratory Spatiotemporal Analysis of Socio-Environmental Patterns in Severe Maternal Morbidity.

Severe Maternal Morbidity (SMM) is a group of pregnancy complications in which a woman nearly dies. Despite its increasing prevalence, little research has evaluated geographic patterns of SMM and the underlying social determinants that influence excess risk. This study examined the spatial clustering of SMM across South Carolina, US, and its associations with place-based social and environmental factors. Hospitalized deliveries from 2012 to 2017 were analyzed using Kulldorff's spatial scan statistic to locate areas with abnormally high rates of SMM. SMM patients inside and outside risk clusters were compared using Generalized Estimating Equations (GEE) to determine underlying individual and community-level risk factors. GEE models revealed that the odds of living in a high-risk SMM21 (SMM including blood transfusions) cluster was 2.49 times higher among Black patients (p < .001) compared to those outside of a high-risk cluster. Women residing in a high-risk SMM20 (SMM excluding blood transfusions) cluster were 1.38 times more likely to experience the most number of extremely hot days and 1.70 times more likely to present with obesity than women in a low-risk SMM cluster (p < .001). This study is the first to characterize the geographic clustering of SMM risk in the US. Our geospatial approach contributes a novel understanding to factors which influence SMM beyond patient-level characteristics and identifies the impact of hot ambient temperatureon maternal morbidity. Findings address an important literature gap surrounding place-based risk factors by explaining the contextual social and built environmental factors that drive SMM risk.

Spatial cluster analysis of Plasmodium vivax and P. malariae exposure using serological data among Haitian school children sampled between 2014 and 2016.

Estimation of malaria prevalence in very low transmission settings is difficult by even the most advanced diagnostic tests. Antibodies against malaria antigens provide an indicator of active or past exposure to these parasites. The prominent malaria species within Haiti is Plasmodium falciparum, but P. vivax and P. malariae infections are also known to be endemic. From 2014-2016, 28,681 Haitian children were enrolled in school-based serosurveys and were asked to provide a blood sample for detection of antibodies against multiple infectious diseases. IgG against the P. falciparum, P. vivax, and P. malariae merozoite surface protein 19kD subunit (MSP119) antigens was detected by a multiplex bead assay (MBA). A subset of samples was also tested for Plasmodium DNA by PCR assays, and for Plasmodium antigens by a multiplex antigen detection assay. Geospatial clustering of high seroprevalence areas for P. vivax and P. malariae antigens was assessed by both Ripley's K-function and Kulldorff's spatial scan statistic. Of 21,719 children enrolled in 680 schools in Haiti who provided samples to assay for IgG against PmMSP119, 278 (1.27%) were seropositive. Of 24,559 children enrolled in 788 schools providing samples for PvMSP119 serology, 113 (0.46%) were seropositive. Two significant clusters of seropositivity were identified throughout the country for P. malariae exposure, and two identified for P. vivax. No samples were found to be positive for Plasmodium DNA or antigens. From school-based surveys conducted from 2014 to 2016, very few Haitian children had evidence of exposure to P. vivax or P. malariae, with no children testing positive for active infection. Spatial scan statistics identified non-overlapping areas of the country with higher seroprevalence for these two malarias. Serological data provides useful information of exposure to very low endemic malaria species in a population that is unlikely to present to clinics with symptomatic infections.

Where Are the Newly Diagnosed HIV Positives in Kenya? Time to Consider Geo-Spatially Guided Targeting at a Finer Scale to Reach the "First 90".

Background: The UNAIDS 90-90-90 Fast-Track targets provide a framework for assessing coverage of HIV testing services (HTS) and awareness of HIV status – the “first 90.” In Kenya, the bulk of HIV testing targets are aligned to the five highest HIV-burden counties. However, we do not know if most of the new HIV diagnoses are in these five highest-burden counties or elsewhere.Methods: We analyzed facility-level HTS data in Kenya from 1 October 2015 to 30 September 2016 to assess the spatial distribution of newly diagnosed HIV-positives. We used the Moran's Index (Moran's I) to assess global and local spatial auto-correlation of newly diagnosed HIV-positive tests and Kulldorff spatial scan statistics to detect hotspots of newly diagnosed HIV-positive tests. For aggregated data, we used Kruskal-Wallis equality-of-populations non-parametric rank test to compare absolute numbers across classes.Results: Out of 4,021 HTS sites, 3,969 (98.7%) had geocodes available. Most facilities (3,034, 76.4%), were not spatially autocorrelated for the number of newly diagnosed HIV-positives. For the rest, clustering occurred as follows; 438 (11.0%) were HH, 66 (1.7%) HL, 275 (6.9%) LH, and 156 (3.9%) LL. Of the HH sites, 301 (68.7%) were in high HIV-burden counties. Over half of 123 clusters with a significantly high number of newly diagnosed HIV-infected persons, 73(59.3%) were not in the five highest HIV-burden counties. Clusters with a high number of newly diagnosed persons had twice the number of positives per 1,000,000 tests than clusters with lower numbers (29,856 vs. 14,172).Conclusions: Although high HIV-burden counties contain clusters of sites with a high number of newly diagnosed HIV-infected persons, we detected many such clusters in low-burden counties as well. To expand HTS where most needed and reach the “first 90” targets, geospatial analyses and mapping make it easier to identify and describe localized epidemic patterns in a spatially dispersed epidemic like Kenya's, and consequently, reorient and prioritize HTS strategies.

Time and spatial trends of operated cryptorchidism in France and environmental hypotheses: a nationwide study from 2002 to 2014.

Is there an evolution in the risk of operated cryptorchidism in France and does local geographical environment appear as an important trigger for this defect? We observed an increase of the risk of operated cryptorchidism in boys under the age of 7 years during the period 2002-2014 and a strong spatial heterogeneity, with the detection of spatial clusters suggesting environmental factors. Epidemiologic data on cryptorchidism are scarce and its etiology is poorly understood. As part of the testicular dysgenesis syndrome, cryptorchidism is suspected to be a male genital developmental disorder caused by endocrine disruptor chemical (EDC) exposure during the prenatal period. This was a retrospective and descriptive study using data from the French national hospital discharge database, in the 2002-2014 study period. We built an indicator to reflect incident cases of operated cryptorchidism in boys under the age of 7 years in metropolitan France, with an algorithm using specific codes for diseases (ICD-10 codes) and surgical acts (CCAM codes). The study population was composed of 89 382 new cases of operated cases of cryptorchidism in boys under the age of 7 years. We estimated the temporal evolution of the incidence rate. We fitted a spatial disease-mapping model to describe the risk of cryptorchidism at the postcode scale. We used Kulldorff's spatial scan statistic and Tango's flexibly shaped spatial scan statistic to identify spatial clusters. The estimated increase in the incidence of operated cryptorchidism from 2002 to 2014 was equal to 36.4% (30.8%; 42.1%). Cryptorchidism displayed spatial heterogeneity and 24 clusters (P < 0.0001) were detected. The main cluster was localized in a former coal mining and metallurgic area in northern France, currently an industrial area. The cluster analysis suggests the role of shared socio-economic and environmental factors that may be geographically determined and intertwined. The industrial activities identified in the clusters are potentially the source of persistent environmental pollution by metals, dioxins and polychlorinated biphenyls. The indicator we used reflects operated cases of cryptorchidism, with an under-evaluation of the health problem. We cannot exclude a possible role of the evolution and local differences in surgical practices in the observed trends. Our inclusion of boys under 7 years of age minimized the biases related to differences in practices according to age. Regarding the environmental hypothesis, this is an exploratory study and should be considered as a hypothesis-generating process for future research studies. To our knowledge, this is the first descriptive study to address nationwide trends of operated cryptorchidism with detection of spatial clusters, with a very large sample allowing great statistical power. Our results generate plausible environmental hypotheses, which need to be further tested. This study was entirely funded by Santé publique France, the French National Public Health Agency. All authors declare they have no actual or potential competing financial interest. N/A.

Spatial-temporal clusters of host-seeking Aedes albopictus, Aedes japonicus, and Aedes triseriatuscollections in a La Crosse virus endemic county (Knox County, Tennessee, USA).

A bite from a La Crosse virus (LACV) infected Aedes mosquito can cause La Crosse encephalitis (LACE), which is a neuro-invasive disease that disproportionately affects children under the age of 16 in Southern Appalachia. The three vectors for LACV are Aedes albopictus (Skuse), Ae. japonicus (Theobald), and Ae. triseriatus (Say). Localized maps of the geographic distribution of vectors are practical tools for mosquito management personnel to target areas with high mosquito abundance. This study hypothesized that LACV vectors have unique species-specific spatial and temporal clusters. To test this, 44 sites were identified in Knox County, Tennessee for their land use/type. At each site, host-seeking mosquitoes were collected approximately every other week from May-October 2018. Spatial clusters of host-seeking mosquito collections for each of the three mosquito species were investigated using Kulldorff’s spatial scan statistic, specifying a retrospective space-time Bernoulli model. Most vector clusters were identified in south-central Knox County while the seasonality of clusters varied by mosquito species. Clusters of Ae. albopictus were observed throughout the entire study period while clusters of Ae. japonicus and Ae. triseriatus only occurred May-June. The findings indicate that the relative abundance of LACV vectors were more abundant in south-central Knox County compared to the rest of the county. Of interest, these clusters spatially overlapped with previous LACE diagnosed cases. These findings are useful in guiding decisions on targeted mosquito control in Knox County and may be applied to other counties within Southern Appalachia.

Spatial clustering of food insecurity and its association with depression: a geospatial analysis of nationally representative South African data, 2008\u20132015

While food insecurity is a persistent public health challenge, its long-term association with depression at a national level is unknown. We investigated the spatial heterogeneity of food insecurity and its association with depression in South Africa (SA), using nationally-representative panel data from the South African National Income Dynamics Study (years 2008–2015). Geographical clusters (“hotpots”) of food insecurity were identified using Kulldorff spatial scan statistic in SaTScan. Regression models were fitted to assess association between residing in food insecure hotspot communities and depression. Surprisingly, we found food insecurity hotspots (p < 0.001) in high-suitability agricultural crop and livestock production areas with reliable rainfall and fertile soils. At baseline (N = 15,630), we found greater likelihood of depression in individuals residing in food insecure hotspot communities [adjusted relative risk (aRR) = 1.13, 95% CI:1.01–1.27] using a generalized linear regression model. When the panel analysis was limited to 8,801 participants who were depression free at baseline, residing in a food insecure hotspot community was significantly associated with higher subsequent incidence of depression (aRR = 1.11, 95% CI:1.01–1.22) using a generalized estimating equation regression model. The association persisted even after controlling for multiple socioeconomic factors and household food insecurity. We identified spatial heterogeneity of food insecurity at a national scale in SA, with a demonstrated greater risk of incident depression in hotspots. More importantly, our finding points to the “Food Security Paradox”, food insecurity in areas with high food-producing potential. There is a need for place-based policy interventions that target communities vulnerable to food insecurity, to reduce the burden of depression.

Spatial distribution and characteristics of HIV clusters in Ethiopia.

ObjectivesEthiopia’s HIV prevalence has decreased by 75% in the past 20 years with the implementation of antiretroviral therapy, but HIV transmission continues in high‐risk clusters. Identifying the spatial and temporal trends, and epidemiologic correlates, of these clusters can lead to targeted interventions.MethodsWe used biomarker and survey data from the 2005, 2011 and 2016 Ethiopia Demographic and Health Surveys (DHS). The spatial‐temporal distribution of HIV was estimated using the Kulldorff spatial scan statistic, a likelihood‐based method for determining clustering. Significant clusters (P < 0.05) were identified and compared based on HIV risk factors to non‐cluster areas.ResultsIn 2005, 2011 and 2016, respectively, 219, 568 and 408 individuals tested positive for HIV. Four HIV clusters were identified, representing 17% of the total population and 43% of all HIV cases. The clusters were centred around Addis Ababa (1), Afar (2), Dire Dawa (3) and Gambella (4). Cluster 1 had higher rates of unsafe injections (4.9% vs. 2.2%, P < 0.001) and transactional sex (6.0% vs. 1.6%, P < 0.001) than non‐cluster regions, but more male circumcision (98.5% vs. 91.3%, P < 0.001). Cluster 2 had higher levels of transactional sex (4.9% vs. 1.6%, P < 0.01), but lower levels of unsafe injections (0.8% vs. 2.2%, P < 0.01). Cluster 3 had fewer individuals with> 1 sexual partner (0% vs. 1.7%, P < 0.001) and more male circumcision (100% vs. 91.3%, P < 0.001). Cluster 4 had less male circumcision (59.1% vs. 91.3%, P < 0.01).ConclusionsIn Ethiopia, geographic HIV clusters are driven by different risk factors. Decreasing the HIV burden requires targeted interventions.

Incidence of multiple sclerosis in the province of Catania. A geo-epidemiological study

BackgroundAn increasing incidence of MS in the city of Catania was reported during 1975–2004, with a higher incidence along the south-eastern flank of the Mt.Etna. We evaluated the incidence of MS in the entire province of Catania during 2005–2015 and the spatial distribution of MS-cases using a cluster analysis. MethodsPatients were considered as incident MS-cases if they fulfilled the revised McDonald criteria for MS during 2005–2015 and were residents in the province of Catania at the time of disease onset. Cluster analysis was performed using both LISA and Kulldorff's spatial scan statistic. Residence address at disease onset was considered for each case. Communalities were assessed considering the centroid of their inhabited area. ResultsA total of 973 MS-cases were identified. Mean annual incidence risk was 8.2/100,000 person-years (95%CI 7.7–8.7), significantly higher among women (10.5/100,000 versus 5.7/100,000). LISA identified a spatial aggregation of MS-cases in the eastern side of the province of Catania and Kulldorff's statistics confirmed the existence of a statistically significant spatial cluster in this area (SIR 1.23,95%CI 1.08–1.23, p-value 0.04). ConclusionsOur study confirms a high incidence of MS in the province of Catania and the presence of a spatial cluster along the eastern side of the province. This area is considered the most exposed to volcanogenic ashes due to the prevailing westerly to north-westerly trade winds. Even if such distribution could be related with a greater exposure to volcanogenic metals, further studies are needed to explore possible alternative hypotheses.

Spatiotemporal Analysis and Epidemiology of Kawasaki Disease in Western New York

Kawasaki disease (KD) is one of the leading causes of acquired heart disease in children in developed nations. Epidemiologic evidence suggests that KD is related to an infectious agent; however, the cause remains unknown. Yearly incidence in Japan has been steadily increasing, but few long-term databases of KD cases from North America have been reviewed. We reviewed the epidemiology of local cases over a 16-year period to study incidence with time and temporal and geographic clustering of cases in a representative cohort in North America. The yearly incidence in cases per population <5 years old per 100,000 was 20.2 and 15.9, using International Classification of Disease, ninth revision and detailed chart review, respectively. Using International Classification of Disease, ninth revision alone overestimates our incidence by 27%. We show a distinct seasonality of cases with winter predominance. Applying Kulldorff's spatial scan statistic revealed no significant clustering of cases with either purely spatial or space-time analyses. On purely nonconstrained temporal SaTScan analysis, there was a significant clustering of cases in a 67- to 68-week period in 2000-2001. Our analysis reveals an apparent outbreak of KD in our region in 2000-2001. In contrast to Japan, for the last 14 years, the incidence in our region has been stable.

DClusterm: Model-Based Detection of Disease Clusters

The detection of regions with unusually high risk plays an important role in disease mapping and the analysis of public health data. In particular, the detection of groups of areas (i.e., clusters) where the risk is significantly high is often conducted by public health authorities. Many methods have been proposed for the detection of these disease clusters, most of them based on moving windows, such as Kulldorff's spatial scan statistic. Here we describe a model-based approach for the detection of disease clusters implemented in the DClusterm package. Our model-based approach is based on representing a large number of possible clusters by dummy variables and then fitting many generalized linear models to the data where these covariates are included one at a time. Cluster detection is done by performing a variable or model selection among all fitted models using different criteria. Because of our model-based approach, cluster detection can be performed using different types of likelihoods and latent effects. We cover the detection of spatial and spatiotemporal clusters, as well as how to account for covariates, zero-inflated datasets and overdispersion in the data.

Geographic patterns of prescription opioids and opioid overdose deaths in New York State, 2013-2015

ObjectivesTo examine the relationship between prescription opioid rates and prescription opioid overdose deaths using spatial cluster and regression analyses. MethodsPublicly available county-level data were obtained from the New York State Health Department and the Centers for Disease Control and Prevention, 2013–2015. Kulldorff's spatial scan statistic was used to investigate spatial clustering of New York State opioid prescription overdose death rates, as well as opioid prescription rates. A Poisson regression was used to analyze opioid prescriptions as a predictor of mortality accounting for spatial autocorrelation in the residuals. ResultsWe report 1440 overdose mortalities and 26.8 million opioid prescriptions throughout New York State in 2013–2015. Multiple significant clusters were found for both opioid prescription mortalities as well as prescriptions, although the locations of the elevated rates did not strongly overlap. Poisson regression showed a significant, small, negative relationship between prescriptions and opioid mortalities, wherein for every 10,000 prescriptions increased, the number of opioid mortalities decreased approximately 0.12%; therefore, essentially a null relationship. ConclusionsSimply reducing the number of prescriptions may not be effective in reducing prescription related mortality; although opioid prescription dosing information should be made available to engender a better evaluation of the epidemic. Geographical differences in opioid mortalities exist above and beyond what can be explained by prescription rate data; identifying these locations may help inform and guide public health interventions. Despite the recent reduction in opioid prescription rates, the overall population is still inundated with prescriptions.