Patient-specific and healthcare real-world costs of atrial fibrillation in individuals treated with direct oral anticoagulant agents or warfarin.

BACKGROUND: Anticoagulant therapies are used to prevent atrial fibrillation-related strokes, with warfarin and direct oral anticoagulant (DOAC) the most common. In this study, we incorporate direct health care costs, drug costs, travel costs, and lost working and leisure time costs to estimate the total costs of the two therapies. METHODS: This retrospective study used individual-level patient data from 4000 atrial fibrillation (AF) patients from North Karelia, Finland. Real-world data on healthcare use was obtained from the regional patient information system and data on reimbursed travel costs from the database of the Social Insurance Institution of Finland. The costs of the therapies were estimated between June 2017 and May 2018. Using a Geographical Information System (GIS), we estimated travel time and costs for each journey related to anticoagulant therapies. We ultimately applied therapy and travel costs to a cost model to reflect real-world expenditures. RESULTS: The costs of anticoagulant therapies were calculated from the standpoint of patient and the healthcare service when considering all costs from AF-related healthcare visits, including major complications arising from atrial fibrillation. On average, the annual cost per patient for healthcare in the form of public expenditure was higher when using DOAC therapy than warfarin therapy (average cost = € 927 vs. € 805). Additionally, the average annual cost for patients was also higher with DOAC therapy (average cost = € 406.5 vs. € 296.7). In warfarin therapy, patients had considerable more travel and time costs due the different implementation practices of therapies. CONCLUSION: The results indicated that DOAC therapy had higher costs over warfarin from the perspectives of the patient and healthcare service in the study area on average. Currently, the cost of the DOAC drug is the largest determinator of total therapy costs from both perspectives. Despite slightly higher costs, the patients on DOAC therapy experienced less AF-related complications during the study period.

Estimating the prevalence and spatial clusters of coal workers' pneumoconiosis cases using medicare claims data, 2011-2014.

BACKGROUND: Workers employed in the coal mining sector are at increased risk of respiratory diseases, including coal workers' pneumoconiosis (CWP). We investigated the prevalence of CWP and its association with sociodemographic factors among Medicare beneficiaries. METHODS: We used 5% Medicare Limited Data Set claims data from 2011 to 2014 to select Medicare beneficiaries with a diagnosis of ICD-9-CM 500 (CWP). We aggregated the data by county and limited our analysis to seven contiguous states: Illinois, Indiana, Kentucky, Ohio, Pennsylvania, Virginia, and West Virginia. We estimated county-level prevalence rates using total Medicare beneficiaries and miners as denominators and performed spatial hotspot analysis. We used negative binomial regression analysis to determine the association of county-wise sociodemographic factors with CWP. RESULTS: There was significant spatial clustering of CWP cases in Kentucky, Virginia, and West Virginia. Spatial clusters of 210 and 605 CWP cases representing an estimated 4200 to 12 100 cases of Medicare beneficiaries with CWP were identified in the three states. Counties with higher poverty levels had a significantly elevated rate of CWP (adjusted rate ratios [RR]: 1.15; 95% CI, 1.12-1.18). There was a small but significant association of CWP with the county-wise catchment area. Rurality was associated with a more than three-fold elevated rate of CWP in the unadjusted analysis (RR: 3.28, 95% CI, 2.22-4.84). However, the rate declined to 1.45 (95% CI, 1.04-2.01) after adjusting for other factors in the analysis. CONCLUSIONS: We found evidence of significant spatial clustering of CWP among Medicare beneficiaries living in the seven states of the USA.

Landscapes of healthcare utilization during a dengue fever outbreak in an urban environment of Colombia.

The well-being of a population and its health are influenced by a myriad of socioeconomic and environmental factors that interact across a wide range of scales, from the individual to the national and global levels. One of these factors is the provision of health services, which is regulated by both demand and supply. Although an adequate provision can significantly improve health outcomes of a population, lopsided flow of patients to specific health centers can result in serious disparities and potentially delay the timeliness of a diagnosis. In this paper, utilization patterns during an epidemic of dengue fever in the city of Cali, Colombia for the year 2010 are investigated. Specifically, the objectives are to (1) identify health facilities that exhibit patterns of over- and underutilization, (2) determine where patients who are being diagnosed at a particular facility originate from, and (3) whether patients are traveling to their closest facility and hence (4) estimate how far patients are willing to travel to be diagnosed and treated for dengue fever. Analysis is further decomposed by age group and by gender, in an attempt to test whether utilization patterns drastically change according to these variables. Answers to these questions can help health authorities plan for future epidemics, for instance, by providing guidelines as to which facilities require more resources and by improving the organization of health prevention campaigns to direct population seeking health assistance to use facilities that are underutilized.

Factors Associated with Travel Time and Distance to Access Hospital Care Among Infants with Spina Bifida.

OBJECTIVES: Using geographic information systems (GIS), we examined travel time and distance to access hospital care for infants with spina bifida (SB). METHODS: This study was a statewide, population-based analysis of Florida-born children with SB, 1998-2007, identified by the Florida Birth Defects Registry and linked to hospitalizations. We geocoded maternal residence at delivery and identified hospital locations for infants (<1 year). Using 2007 Florida Department of Transportation road data, we calculated one-way mean travel time and distance to access hospital care. We used Poisson regression to examine selected factors associated with travel time and distance [≤30 vs. >30 min/miles (reference)], including presence of hydrocephalus and SB type [isolated (no other major birth defect) versus non-isolated SB]. RESULTS: For 612 infants, one-way mean (median) travel time was 45.1 (25.9) min. Infants with both non-isolated SB and hydrocephalus traveled longest to access hospitals (mean 60.8 min/48.5 miles; median 34.2 min/26.9 miles). In adjusted results, infants with non-isolated SB and whose mothers had a rural residence were less likely to travel ≤30 min to hospitals. Infants born to mothers in minority racial/ethnic groups were more likely to travel ≤30 min. CONCLUSIONS: Birth defects registry data and GIS-based methods can be used to evaluate geographic accessibility to hospital care for infants with birth defects. Results can help to identify geographic barriers to accessing hospital care, such as travel time and distance, and inform opportunities to improve access to care for infants with SB or other special needs.

Modeling travel impedance to medical care for children with birth defects using Geographic Information Systems.

BACKGROUND: Children with birth defects may face significant geographic barriers accessing medical care and specialized services. Using a Geographic Information Systems-based approach, one-way travel time and distance to access medical care for children born with spina bifida was estimated. METHODS: Using 2007 road information from the Florida Department of Transportation, we built a topological network of Florida roads. Live-born Florida infants with spina bifida during 1998 to 2007 were identified by the Florida Birth Defects Registry and linked to hospital discharge records. Maternal residence at delivery and hospitalization locations were identified during the first year of life. RESULTS: Of 668 infants with spina bifida, 8.1% (n = 54) could not be linked to inpatient data, resulting in 614 infants. Of those 614 infants, 99.7% (n = 612) of the maternal residential addresses at delivery were successfully geocoded. Infants with spina bifida living in rural areas in Florida experienced travel times almost twice as high compared with those living in urban areas. When aggregated at county levels, one-way network travel times exhibited statistically significant spatial autocorrelation, indicating that families living in some clusters of counties experienced substantially greater travel times compared with families living in other areas of Florida. CONCLUSION: This analysis demonstrates the usefulness of linking birth defects registry and hospital discharge data to examine geographic differences in access to medical care. Geographic Information Systems methods are important in evaluating accessibility and geographic barriers to care and could be used among children with special health care needs, including children with birth defects.

Assessing socioeconomic vulnerability to dengue fever in Cali, Colombia: statistical vs expert-based modeling.

BACKGROUND: As a result of changes in climatic conditions and greater resistance to insecticides, many regions across the globe, including Colombia, have been facing a resurgence of vector-borne diseases, and dengue fever in particular. Timely information on both (1) the spatial distribution of the disease, and (2) prevailing vulnerabilities of the population are needed to adequately plan targeted preventive intervention. We propose a methodology for the spatial assessment of current socioeconomic vulnerabilities to dengue fever in Cali, a tropical urban environment of Colombia. METHODS: Based on a set of socioeconomic and demographic indicators derived from census data and ancillary geospatial datasets, we develop a spatial approach for both expert-based and purely statistical-based modeling of current vulnerability levels across 340 neighborhoods of the city using a Geographic Information System (GIS). The results of both approaches are comparatively evaluated by means of spatial statistics. A web-based approach is proposed to facilitate the visualization and the dissemination of the output vulnerability index to the community. RESULTS: The statistical and the expert-based modeling approach exhibit a high concordance, globally, and spatially. The expert-based approach indicates a slightly higher vulnerability mean (0.53) and vulnerability median (0.56) across all neighborhoods, compared to the purely statistical approach (mean = 0.48; median = 0.49). Both approaches reveal that high values of vulnerability tend to cluster in the eastern, north-eastern, and western part of the city. These are poor neighborhoods with high percentages of young (i.e., < 15 years) and illiterate residents, as well as a high proportion of individuals being either unemployed or doing housework. CONCLUSIONS: Both modeling approaches reveal similar outputs, indicating that in the absence of local expertise, statistical approaches could be used, with caution. By decomposing identified vulnerability "hotspots" into their underlying factors, our approach provides valuable information on both (1) the location of neighborhoods, and (2) vulnerability factors that should be given priority in the context of targeted intervention strategies. The results support decision makers to allocate resources in a manner that may reduce existing susceptibilities and strengthen resilience, and thus help to reduce the burden of vector-borne diseases.

Space-time cluster detection techniques for infectious diseases: A systematic review.

BACKGROUND: Public health organizations have increasingly harnessed geospatial technologies for disease surveillance, health services allocation, and targeting place-based health promotion initiatives. METHODS: We conducted a systematic review around the theme of space-time clustering detection techniques for infectious diseases using PubMed, Web of Science, and Scopus. Two reviewers independently determined inclusion and exclusion. RESULTS: Of 2,887 articles identified, 354 studies met inclusion criteria, the majority of which were application papers. Studies of airborne diseases were dominant, followed by vector-borne diseases. Most research used aggregated data instead of point data, and a significant proportion of articles used a repetition of a spatial clustering method, instead of using a "true" space-time detection approach, potentially leading to the detection of false positives. Noticeably, most articles did not make their data available, limiting replicability. CONCLUSION: This review underlines recent trends in the application of space-time clustering methods to the field of infectious disease, with a rapid increase during the COVID-19 pandemic.

Road environment characteristics and fatal crash injury during the rush and non-rush hour periods in the U.S: Model testing and cluster analysis.

This study aims to assess the relationship between county-level fatal crash injuries and road environmental characteristics at all times of the day and during the rush and non-rush hour periods. We merged eleven-year (2010 - 2020) data from the Fatality Analysis Reporting System. The outcome variable was the county-level fatal crash injury counts. The predictor variables were measures of road types, junction types and work zone, and weather types. We tested the predictiveness of two nested negative binomial models and adjudged that a nested spatial negative binomial regression model outperformed the non-spatial negative binomial model. The median county crash mortality rates at all times of the day and during the rush and non-rush hour periods were 18.4, 7.7, and 10.4 per 100,000 population, respectively. Fatal crash injury rate ratios were significantly elevated on interstates and highways at all times of the day - rush and non-rush hour periods inclusive. Intersections, driveways, and ramps on highways were associated with elevated fatal crash injury rate ratios. Clusters of high fatal crash injury rates were observed in counties located in Montana, Nevada, Colorado, Kansas, New Mexico, Oklahoma, Texas, Arkansas, Mississippi, Alabama, Georgia, and Nevada. The built and natural road environment factors are associated with county-level fatal crash injuries during the rush and non-rush hour periods. Understanding the association of road environment characteristics and the cluster distribution of fatal crash injuries may inform areas in need of focused intervention.

Detecting space-time patterns of disease risk under dynamic background population.

We are able to collect vast quantities of spatiotemporal data due to recent technological advances. Exploratory space-time data analysis approaches can facilitate the detection of patterns and formation of hypotheses about their driving processes. However, geographic patterns of social phenomena like crime or disease are driven by the underlying population. This research aims for incorporating temporal population dynamics into spatial analysis, a key omission of previous methods. As population data are becoming available at finer spatial and temporal granularity, we are increasingly able to capture the dynamic patterns of human activity. In this paper, we modify the space-time kernel density estimation method by accounting for spatially and temporally dynamic background populations (ST-DB), assess the benefits of considering the temporal dimension and finally, compare ST-DB to its purely spatial counterpart. We delineate clusters and compare them, as well as their significance, across multiple parameter configurations. We apply ST-DB to an outbreak of dengue fever in Cali, Colombia during 2010-2011. Our results show that incorporating the temporal dimension improves our ability to delineate significant clusters. This study addresses an urgent need in the spatiotemporal analysis literature by using population data at high spatial and temporal resolutions.

The association of crash response times and deaths at the crash scene: A cross-sectional analysis using the 2019 National Emergency Medical Service Information System.

BACKGROUND: Deaths at the crash scene (DAS) are crash deaths that occur within minutes after a crash. Rapid crash responses may reduce the occurrence of DAS. OBJECTIVES: This study aims to assess the association of crash response time and DAS during the rush and nonrush hour periods by rurality/urbanicity. METHOD: This single-year cross-sectional study used the 2019 National Emergency Medical Services (EMS) Information System. The outcome variable was DAS. The predictor variables were crash response measures: EMS Chute Initiation Time (ECIT) and EMS Travel Time (ETT). Age, gender, substance use, region of the body injured, and the revised trauma score were used as potential confounders. Logistic regression was used to assess the unadjusted and adjusted odds of DAS. RESULTS: A total of 654,675 persons were involved in EMS-activated road crash events, with 49.6% of the population exposed to crash events during the rush hour period. A total of 2,051 persons died at the crash scene. Compared to the baseline of less than 1 minute, ECIT ranging from 1 to 5 minutes was significantly associated with 58% (95% CI: 1.45-1.73) increased odds of DAS. Also, when compared to the baseline of less than 9 minutes, ETT ranging between 9 and 18 minutes was associated with 34% (95% CI: 1.22-1.47) increased odds of DAS. These patterns were consistent during the rush and nonrush hour periods and across rural and urban regions. CONCLUSION: Reducing crash response times may reduce the occurrence of DAS.

An assessment of the non-fatal crash risks associated with substance use during rush and non-rush hour periods in the United States.

BACKGROUND: Understanding how substance use is associated with severe crash injuries may inform emergency care preparedness. OBJECTIVES: This study aims to assess the association of substance use and crash injury severity at all times of the day and during rush (6-9 AM; 3-7 PM) and non-rush-hours. Further, this study assesses the probabilities of occurrence of low acuity, emergent, and critical injuries associated with substance use. METHODS: Crash data were extracted from the 2019 National Emergency Medical Services Information System. The outcome variable was non-fatal crash injury, assessed on an ordinal scale: critical, emergent, low acuity. The predictor variable was the presence of substance use (alcohol or illicit drugs). Age, gender, injured part, revised trauma score, the location of the crash, the road user type, and the geographical region were included as potential confounders. Partially proportional ordinal logistic regression was used to assess the unadjusted and adjusted odds of critical and emergent injuries compared to low acuity injury. RESULTS: Substance use was associated with approximately two-fold adjusted odds of critical and emergent injuries compared to low acuity injury at all times of the day and during the rush and non-rush hours. Although the proportion of substance use was higher during the non-rush hour period, the interaction effect of rush hour and substance use resulted in higher odds of critical and emergent injuries compared to low acuity injury. CONCLUSION: Substance use is associated with increased odds of critical and emergent injury severity. Reducing substance use-related crash injuries may reduce adverse crash injuries.

Uncertainty in geospatial health: challenges and opportunities ahead.

PURPOSE: Uncertainty is not always well captured, understood, or modeled properly, and can bias the robustness of complex relationships, such as the association between the environment and public health through exposure, estimates of geographic accessibility and cluster detection, to name a few. METHODS: We review current challenges and future opportunities as geospatial data and analyses are applied to the field of public health. We are particularly interested in the sources of uncertainty in geospatial data and how this uncertainty may propagate in spatial analysis. RESULTS: We present opportunities to reduce the magnitude and impact of uncertainty. Specifically, we focus on (1) the use of multiple reference data sources to reduce geocoding errors, (2) the validity of online geocoders and how confidentiality (e.g., HIPAA) may be breached, (3) use of multiple reference data sources to reduce geocoding errors, (4) the impact of geoimputation techniques on travel estimates, (5) residential mobility and how it affects accessibility metrics and clustering, and (6) modeling errors in the American Community Survey. Our paper discusses how to communicate spatial and spatiotemporal uncertainty, and high-performance computing to conduct large amounts of simulations to ultimately increase statistical robustness for studies in public health. CONCLUSIONS: Our paper contributes to recent efforts to fill in knowledge gaps at the intersection of spatial uncertainty and public health.

Utilizing prospective space-time scan statistics to discover the dynamics of coronavirus disease 2019 clusters in the State of São Paulo, Brazil.

BACKGROUND: The number of deaths and people infected with coronavirus disease 2019 (COVID-19) in Brazil has steadily increased in the first few months of the pandemic. Despite the underreporting of coronavirus cases by government agencies across the country, São Paulo has the highest rate among all Brazilian states. METHODS: To identify the highest-risk municipalities during the initial outbreak, we utilized daily confirmed case data from official reports between February 25 and May 5, 2020, which were aggregated to the municipality level. A prospective space-time scan statistic was conducted to detect active clusters in three different time periods. RESULTS: Our findings suggest that approximately 4.6 times more municipalities belong to a significant space-time cluster with a relative risk (RR) > 1 on May 5, 2020. CONCLUSIONS: Our study demonstrated the applicability of the space-time scan statistic for the detection of emerging clusters of COVID-19. In particular, we identified the clusters and RR of municipalities in the initial months of the pandemic, explaining the spatiotemporal patterns of COVID-19 transmission in the state of São Paulo. These results can be used to improve disease monitoring and facilitate targeted interventions.

What is the place for space in epidemiology?

At the heart of spatial epidemiology is the need to describe and understand variation in population health. In this review and introduction to the themed issue on "Spatial Analysis and GIS in Epidemiology," we present theoretical foundations and methodological developments in spatial epidemiology, discuss spatial analytical techniques and their public health applications, and identify novel data sources and applications with the potential to make epidemiology more consequential. Challenges with using georeferenced data are also explored, including dealing with small sample sizes, missingness, generalizability, and geographic scale. Given the increasing availability of spatial data and visualization tools, we have an opportunity to overcome traditionally siloed fields and practice settings to advance knowledge and more appropriately respond to emerging public health crises.

Examining Wing Length-Abundance Relationships and Pyrethroid Resistance Mutations among Aedes albopictus in a Rapidly Growing Urban Area with Implications for Mosquito Surveillance and Control.

Aedes albopictus is a cosmopolitan mosquito species capable of transmitting arboviruses such as dengue, chikungunya, and Zika. To control this and similar species, public and private entities often rely on pyrethroid insecticides. In this study, we screened Ae. albopictus collected from June to August 2017 in Mecklenburg County, a rapidly growing urban area of North Carolina, for mutations conferring pyrethroid resistance and examined spatiotemporal patterns of specimen size as measured by wing length, hypothesizing that size variation could be closely linked to local abundance, making this easily measured trait a useful surveillance proxy. The genetic screening results indicated that pyrethroid resistance alleles are not present in this population, meaning that this population is likely to be susceptible to this commonly used insecticide class. We detected no significant associations between size and abundance-related factors, indicating that wing-size is not a useful proxy for abundance, and thus not useful to surveillance in this capacity. However, mosquitoes collected in June were significantly larger than July or August, which may result from meteorological conditions, suggesting that short-term weather cues may modulate morphological traits, which could then affect local fecundity and virus transmission dynamics, as previously reported.

Predicting coliform presence in private wells as a function of well characteristics, parcel size and leachfield soil rating.

Public water systems must be tested frequently for coliform bacteria to determine whether other pathogens may be present, yet no testing or disinfection is required for private wells. In this paper, we identify whether well age, type of well, well depth, parcel size, and soil ratings for a leachfield can predict the probability of detecting coliform bacteria in private wells using a multivariate logistic regression model. Samples from 1163 wells were analyzed for the presence of coliform bacteria between October 2017 and October 2019 across Gaston County, North Carolina, USA. The maximum well age was 30 years, and bored wells (median age = 24 years) were older than drilled wells (median age = 19 years). Bored wells were shallower (mean depth = 18 m) compared to drilled wells (mean depth = 79 m). We found coliform bacteria in 329 samples, including 290 of 1091 drilled wells and 39 of 72 bored wells. The model results showed bored wells were 4.76 times more likely to contain bacteria compared to drilled wells. We found that the likelihood of coliform bacteria significantly increased with well age, suggesting that those constructed before well standards were enforced in 1989 may be at a higher risk. We found no significant association between poorly rated soils for a leachfield, well depth, parcel size and the likelihood of having coliform in wells. These findings can be leveraged to determine areas of concern to encourage well users to take action to reduce their risk of drinking possible pathogens in well water.

Daily surveillance of COVID-19 using the prospective space-time scan statistic in the United States.

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was first discovered in late 2019 in Wuhan City, China. The virus may cause novel coronavirus disease 2019 (COVID-19) in symptomatic individuals. Since December of 2019, there have been over 7,000,000 confirmed cases and over 400,000 confirmed deaths worldwide. In the United States (U.S.), there have been over 2,000,000 confirmed cases and over 110,000 confirmed deaths. COVID-19 case data in the United States has been updated daily at the county level since the first case was reported in January of 2020. There currently lacks a study that showcases the novelty of daily COVID-19 surveillance using space-time cluster detection techniques. In this paper, we utilize a prospective Poisson space-time scan statistic to detect daily clusters of COVID-19 at the county level in the contiguous 48 U.S. and Washington D.C. As the pandemic progresses, we generally find an increase of smaller clusters of remarkably steady relative risk. Daily tracking of significant space-time clusters can facilitate decision-making and public health resource allocation by evaluating and visualizing the size, relative risk, and locations that are identified as COVID-19 hotspots.

Space-Time Conditional Autoregressive Modeling to Estimate Neighborhood-Level Risks for Dengue Fever in Cali, Colombia.

Vector-borne diseases affect more than 1 billion people a year worldwide, causing more than 1 million deaths, and cost hundreds of billions of dollars in societal costs. Mosquitoes are the most common vectors responsible for transmitting a variety of arboviruses. Dengue fever (DENF) has been responsible for nearly 400 million infections annually. Dengue fever is primarily transmitted by female Aedes aegypti and Aedes albopictus mosquitoes. Because both Aedes species are peri-domestic and container-breeding mosquitoes, dengue surveillance should begin at the local level-where a variety of local factors may increase the risk of transmission. Dengue has been endemic in Colombia for decades and is notably hyperendemic in the city of Cali. For this study, we use weekly cases of DENF in Cali, Colombia, from 2015 to 2016 and develop space-time conditional autoregressive models to quantify how DENF risk is influenced by socioeconomic, environmental, and accessibility risk factors, and lagged weather variables. Our models identify high-risk neighborhoods for DENF throughout Cali. Statistical inference is drawn under Bayesian paradigm using Markov chain Monte Carlo techniques. The results provide detailed insight about the spatial heterogeneity of DENF risk and the associated risk factors (such as weather, proximity to Aedes habitats, and socioeconomic classification) at a fine level, informing public health officials to motivate at-risk neighborhoods to take an active role in vector surveillance and control, and improving educational and surveillance resources throughout the city of Cali.

Do socioeconomic factors drive Aedes mosquito vectors and their arboviral diseases? A systematic review of dengue, chikungunya, yellow fever, and Zika Virus.

As the threat of arboviral diseases continues to escalate worldwide, the question of, "What types of human communities are at the greatest risk of infection?" persists as a key gap in the existing knowledge of arboviral diseases transmission dynamics. Here, we comprehensively review the existing literature on the socioeconomic drivers of the most common Aedes mosquito-borne diseases and Aedes mosquito presence/abundance. We reviewed a total of 182 studies on dengue viruses (DENV), chikungunya virus (CHIKV), yellow fever virus (YFVV), Zika virus (ZIKV), and presence of Aedes mosquito vectors. In general, associations between socioeconomic conditions and both Aedes-borne diseases and Aedes mosquitoes are highly variable and often location-specific. Although 50% to 60% of studies found greater presence or prevalence of disease or vectors in areas with lower socioeconomic status, approximately half of the remaining studies found either positive or null associations. We discuss the possible causes of this lack of conclusiveness as well as the implications it holds for future research and prevention efforts.

Second-Phase Sampling Designs for Non-Stationary Spatial Variables.

In spatial sampling, once initial samples of the primary variable have been collected, it is possible to take additional measurements, an approach known as second-phase sampling. Additional samples are usually collected away from observation locations, or where the kriging variance is maximum. However, the kriging variance (also known as prediction error variance) is independent of data values and computed under the assumption of stationary spatial process, which is often violated in practice. In this paper, we weight the kriging variance with another criterion, giving greater sampling importance to locations exhibiting significant spatial roughness that is computed by a spatial moving average window. Additional samples are allocated using a simulated annealing procedure since the weighted objective function is non-linear. A case study using an exhaustive remote sensing image illustrates the procedure. Combinations of first-phase systematic and nested sampling designs (or patterns) of varying densities are generated, while the location of additional observations is guided in a way which optimizes the proposed objective function. The true pixel value at the new points is extracted, the semivariogram model updated, and the image reconstructed. Second-phase sampling patterns optimizing the proposed criterion lead to predictions closer to the true image than when using the kriging variance as the main criterion. This improvement is stronger when there is a low density of first-phase samples, and decreases however as the initial density increases.