Browsing by Author "Kim, Dohyeong"
Now showing 1 - 16 of 16
- Results Per Page
- Sort Options
Item A Big Data Framework for Unstructured Text Processing With Applications Towards Political Science and Healthcare(2021-12-01T06:00:00.000Z) Salam, Sayeed; Khan, Latifur; Hu, Yang; Bastani, Farokh B.; Kim, Dohyeong; Wu, WeiliMachine learning and deep neural networks have soared in popularity in recent years, allowing us to enhance many aspects of everyday life. While these methods are intuitive, they are very reliant on the dataset being used to build the model. A high-quality dataset boosts the model’s accuracy and validates the model’s output in the context of a real-world scenario. Furthermore, continuous improvement on the dataset contributes in the tuning of the model in a time-consistent way and the mitigation of temporal inconsistencies. However, preparing datasets, particularly for text domains, is difficult due to the inherent unstructured nature of the data and the use of multiple languages. Furthermore, the amount of text produced in the form of news articles or social media posts is massive, necessitating large-scale processing. The velocity at which new texts are produced demands an elastic and scalable system that can accommodate any surge of inputs while remaining resource efficient while not in use. Texts are created in a variety of ways and must be preprocessed and analyzed in order to provide well-structured, consistent data. This can be accomplished through the use of a well-defined domain-specific ontology (rule-based approach) or machine learning approaches. While rule-based systems can provide information that are more precise and are preferred in a variety of circumstances, they lack flexibility as the ontologies are often fixed and does not respond well with the continuous changes in respective domains. We propose associated solutions to the challenges described above in this dissertation. First, we go over a scalable architecture for collecting news stories from around the world and utilizing a rule-based approach with the Conflict and Mediation Event Observation(CAMEO) ontology to generate political events. We present a summary of the generated dataset, as well as some basic analysis, to demonstrate how it relates to the real-world scenario. We present techniques to dynamically adding information to the ontology using a mining approach for discovering new political actors that works as a recommender system and retrieves more than 80% of the missing information including political figures and their roles. We discuss an extended data processing system for processing articles published in several languages, with a focus on translation methodologies and tools developed. In comparison to the English language, we demonstrate the efficacy of the coder in Spanish. When compared to equivalent events in English articles, the revised event coder with translated knowledge-base was able to recognize 83% of information in Spanish. For healthcare, we propose an alternative strategy in which we use several machine learning algorithms and social media, such as tweets, to extract the location and severity of Road Traffic Incidents (RTI). We highlight a pipeline that goes from collecting tweets to summarizing related tweets for an RTI. We also demonstrate how semi-automatic ontology learning can be useful in determining severity and offer a simplified example in which 100% of the target rules were identified using an iterative technique.Item A Meta-Regression Analysis of the Effectiveness of Mosquito Nets for Malaria Control: The Value of Long-Lasting Insecticide Nets(MDPI) Yang, Gi-geun; Kim, Dohyeong; Anh Pham; Paul, Christopher John; 0000-0002-1428-1451 (Kim, D); Kim, DohyeongLong-lasting insecticidal nets (LLINs) have been widely used as an effective alternative to conventional insecticide-treated nets (ITNs) for over a decade. Due to the growing number of field trials and interventions reporting the effectiveness of LLINs in controlling malaria, there is a need to systematically review the literature on LLINs and ITNs to examine the relative effectiveness and characteristics of both insecticide nettings. A systematic review of over 2000 scholarly articles published since the year 2000 was conducted. The odds ratios (ORs) of insecticidal net effectiveness in reducing malaria were recorded. The final dataset included 26 articles for meta- regression analysis, with a sample size of 154 subgroup observations. While there is substantial heterogeneity in study characteristics and effect size, we found that the overall OR for reducing malaria by LLIN use was 0.44 (95% CI = 0.41-0.48, p < 0.01) indicating a risk reduction of 56%, while ITNs were slightly less effective with an OR of 0.59 (95% CI = 0.57-0.61, p < 0.01). A meta-regression model confirms that LLINs are significantly more effective than ITNs in the prevention of malaria, when controlling for other covariates. For both types of nets, protective efficacy was greater in high transmission areas when nets were used for an extended period. However, cross-sectional studies may overestimate the effect of the nets. The results surprisingly suggest that nets are less effective in protecting children under the age of five, which may be due to differences in child behavior or inadequate coverage. Compared to a previous meta-analysis, insecticide-treated nets appear to have improved their efficacy despite the risks of insecticide resistance. These findings have practical implications for policymakers seeking effective malaria control strategies.Item Barriers and Incentives for Sustainable Urban Development: An Analysis of the Adoption of LEED-ND Projects(Academic Press, 2019-05-22) Cease, Brett; Kim, HyoungAh; Kim, Dohyeong; Ko, Y.; Cappel, C.; 0000-0002-1428-1451 (Kim, D); Cease, Brett; Kim, HyoungAh; Kim, DohyeongThe adoption rate for Leadership in Energy and Environmental Design – Neighborhood Development (LEED-ND) projects has varied considerably across the United States. Local governments and developers face variation in the incentives and barriers while implementing LEED-ND projects across four key dimensions – economic, policy, public awareness, and organizational. This paper investigated the drivers of variation using a mixed-methods approach including a two-stage Heckman model, a survey of Texas subdivision developers and interviews with local planning officials. Results indicate that initial public funding may lead to more LEED-ND projects being completed, but with a diminishing return as these projects become established within the region. Support for local programs including tax abatement, public-private partnerships, and other incentives were also demonstrated to help facilitate LEED-ND project adoption. Overall this paper underscored the important role, especially early on, the public sector and local governments play in initiating local LEED-ND projects to inform and motivate the land development industry. © 2019 Elsevier LtdItem Chinese Residential Housing Policy Evaluation(2017-12) Zhang, Yingyuan; Scotch, Richard K.; Elliott, Euel; Kim, Dohyeong; Ho, KarlHousing policies have long been discussed in countries around the world. Every country should use a different type of policy to stabilize its housing market depending on its political system. In this paper, the core question is whether the government’s housing policies actually serve to control prices in China’s housing market as expected. Analysis in this vein is based on new housing market data which other researchers have not used before and employs different customized time series models. Until now, researchers have not been able to obtain very good data to measure such policy effects in China. Hence, one of the contributions of this paper is to evaluate policy effects using more accurate data measurements than had been previously available. This new housing market dataset can be divided into two parts, yearly data and monthly data. The dataset includes housing information from 39 big and medium size cities in China in the form of monthly panel data and from 35 big and medium size cities in China in the form of yearly panel data. The analytical results indicate that housing policies targeting housing prices did successfully control prices, especially in large cities such as Beijing.Item Disparity in Risk Factors for Urban Residential Fire Related Injuries and Deaths(2018-08) Min, Soojin; 0000-0003-4623-258X (Min, S); Scotch, Richard K.; Kim, DohyeongThe location characteristics of neighborhoods and balance in demand and supply capacity may play a role in determining the effectiveness of fire protection service delivery. Spatial accessibility to fire protection services integrates the location characteristics of neighborhoods and the dimensions of demand and supply capacity of fire protection services. Using the two-step floating catchment area (2SFCA) method and logistic regression, this study measures spatial accessibility to fire protection services and examines its association with unintentional residential fire related injuries and deaths in Dallas, Texas. This analysis uses annual public fire incident data from 2012 to 2015, obtained from the U.S. Fire Administration, and census. In addition to fire characteristics and neighborhood demographics, spatial accessibility to fire protection services was significantly associated with unintentional residential fire related injuries with a small effect size. The analysis results suggest that there is disparity in the spatial accessibility score between low-income and non low-income census block groups, mainly in northeast and southwest service areas. The findings can be used to help identify high-risk neighborhoods for implementing fire injury prevention programs and select locations of additional fire stations.Item Examining the Spatial Mismatch in the Supply and Demand for Maternal and Child Health Services in Bangladesh(2017-12) Vyas, Priyanka D.; Kim, DohyeongOver the past decades countries have made remarkable progress in improving population health. Yet gaps in service provision and differences in service utilization often result in within-country disparities in health outcomes. This poses a formidable challenge in the equitable delivery of health care for governments in developing countries. Poor availability of geographic data also attenuates the ability of researchers to examine equity in the spatial distribution of health facilities in relation to the population demand. This dissertation examines spatial equity in access to health services and in the utilization of maternal and reproductive health services in the context of Bangladesh. By examining the spatial distribution of public health facilities in Bangladesh, this research seeks to answer policy relevant questions on factors influencing differencing within regions, which in turn can help the government respond better to the challenges of disparities. This research consists of two components. The first component examines spatial equity in the distribution of tertiary and secondary level health facilities using the lens of central place theory and urban hierarchy, while the second section examines mismatch in the supply and demand for primary health care services. In order to examine spatial equity in the distribution of tertiary and secondary public health services, this research uses geo-spatial data from the Government of Bangladesh in combination with the Demographic and Health Survey data from 2011 for Barisal and Sylhet Division of Bangladesh. Contrary to the assumption of urban hierarchy and central place theory which suggests market areas to be homogeneous for the same type of goods and services, this research finds substantial amount of within region variation in travel distance for the same type of health services. This research finds that inequality in access to public health facilities is influenced by the heterogeneous size of the administrative units to which they are linked. To test the spatial mismatch hypothesis, this dissertation used bivariate kernel density estimation technique to examine whether the distribution of the clinics followed population distribution. This analysis finds a greater concentration of clinics in rural areas than urban areas. This implies lower access to primary health care for the urban population. This methodology to assess spatial inequity can be useful in the context of developing countries where covariate data may not be reliable or available. This research provides recommendation to the Government of Bangladesh to capitalize on its vast distribution of primary health care facilities, with services aimed at improving the overall population health. Overall, it was found that with regard to hospital services, urban consumers in the city corporation had better access while communities living further from the city corporations and district centers were at a disadvantage; whereas in the case of primary health care, rural consumers had better access while urban consumers in the most densely populated places had limited access.Item Exploring Determinants for Recruitment and Retention of Family Doctors for rural Practice in Vietnam: Lessons from a Discrete Choice Experiment(2017-12) Pham, Anh; Kim, DohyeongIn Vietnam, health professional shortage in rural and remote areas is a widely acknowledged problem. Policies promulgated up to now have had very little impact on correcting the critical geographical imbalance of health human resources. This study provides comprehensive and quantitative information to understand the determinants of recruiting and retaining family physicians for rural practice in Vietnam, to examine whether different subgroups (demographic and personal working experiences) put different values on different incentive policies, and how different the values are. To fulfill the objectives, a discrete choice experiment was designed and administered to elicit the job preferences of 315 family doctors. Data collection took place in all the regions of Vietnam (north, central, and south), which allowed the study to address characteristics of family physicians from different regions. An initial qualitative study identified three job attributes – income, career development, and government support for opening private clinic. Mixed logit regression was used for the statistical analysis of relative importance of job attributes and of individual characteristics. The findings of this study suggest that increasing current income by 50 percent has the highest impact on job location decisions for many subgroups (participants aged under 50, assistant doctors, and those with exposure to rural areas). However, the effect of income incentives decreases at a certain threshold. Non-pecuniary interventions have greater impact on some specific subgroups (those already left CHSs to higher level health facilities, individuals currently working at the national level, and individuals working in a primary health care area). It is possible that financial incentive intervention could be complemented with other types of non-financial interventions while still positively impacting family doctors’ deployment to rural areas. This research is the first attempt to provide quantitative information data regarding family physicians’ job preferences and trade-offs between different job attributes in Vietnam. This study is also the first attempt at examining how individual characteristics interact with incentive policies, then which in turn influence family doctors’ job location decisions in Vietnam setting. The findings of this study point out that it is critically important to customize incentive policies based on demographic and personal working experiences.Item Role of Spatial Tools in Public Health Policymaking of Bangladesh: Opportunities and ChallengesKim, Dohyeong; Sarker, Malabika; Vyas, PriyankaIn spite of the increasing efforts to gather spatial data in developing countries, the use of maps is mostly for visualization of health indicators rather than informed decision-making. Various spatial tools can aid policymakers to allocate resources effectively, predict patterns in communicable or infectious diseases, and provide insights into geographical factors which are associated with utilization or adequacy of health services. In Bangladesh, the launch of District Health Information System 2, along with recent efforts to gather spatial data of facilities location, provides an interesting opportunity to study the current landscape and the potential barriers in advancing the use of spatial tools for informed decision making. This study assessed the current level of map usage and spatial tools for health sector planning in Bangladesh, focusing on investigating why map usage and spatial tools remained at a basic level for the purpose of health policy. The study design involved in-depth interviews, followed by an expert survey (n = 39) obtained through snowball sampling. Our survey revealed that assessing areas with shortage of community health workers emerged as the top most for basic map usage or primarily for visualization purpose, while planning for emergency and obstetric care services, and disease mapping was the most frequent category for intermediate and advanced map usage, respectively. Furthermore, we found lack of inter-institutional collaboration, lack of continuous availability of trained personnel, and lack of awareness on the use of geographic information system (GIS) as a decision-making tool as three most critical barriers in the current landscape. Our findings highlight the barriers in increasing the adoption of spatial tools for health policymaking and planning in Bangladesh.Item Spatially Explicit Machine Learning Approaches for House Price Models(May 2023) Chen, Meifang 1989-; Cordell, Rebecca; Chun, Yongwan; Griffith, Daniel A.; Kim, Dohyeong; Qiu, FangSpatial data or georeferenced data are special in that it has spatial reference, meaning that it is linked with geographic coordinates on Earth. The spatial component allows for the identification of spatial patterns, relationships and trends among spatial objects. Spatial objects are usually not randomly or independently distributed, but spatially autocorrelated. In spatial data analysis, spatial autocorrelation has been well recognized with the advocate of spatial statistical techniques, such as spatial clustering, spatial interpolation, spatial regression, and spatial simulation. However, spatial effects or spatial context is largely absent in mainstream machine learning methods. With the popularity of machine learning in various applications in both industry and academia, a new research area has emerged in the spatial community: spatial explicit machine learning. It refers to the use of machine learning algorithms to analyze and predict spatial data with the explicit integration of spatial effects or patterns. It is expected to improve the model accuracy and prediction by incorporating spatial relationships or patterns in the data that have not been captured by traditional machine learning models and, subsequently, to gain better understanding of the data generation mechanism. This research utilizes Franklin County, OH residential house transaction data to explore three different data-driven approaches to integrate spatial perspectives into traditional machine learning algorithms: 1) imposing spatial constraints on unsupervised learning to delineate spatially constrained housing submarkets ; 2) integrating spatial weights into the cost function of supervised learning to improve house price prediction accuracy; and 3) enhancing data input using spatial feature engineering in tree-based ensemble learning for modeling multiscale spatial effects. It intends to contribute new insights for spatially explicit machine learning to the literature. Overall, three studies explore spatially explicit machine learning methods from three different aspects, and the empirical results show that spatially explicit machine learning methods are preferred over traditional machine learning methods when data have strong positive spatial autocorrelation, or more general, data include spatial information that is important for classification, clustering, or prediction tasks.Item Spatio-Temporal Comparison of Pertussis Outbreaks in Olmsted County, Minnesota, 2004-2005 and 2012: A Population-Based Study(BMJ Publishing Group, 2019-05-19) Wi, C. -I; Wheeler, P. H.; Kaur, H.; Ryu, E.; Kim, Dohyeong; Juhn, Y.; 0000-0002-1428-1451 (Kim, D); Kim, DohyeongObjective: Two pertussis outbreaks occurred in Olmsted County, Minnesota, during 2004-2005 and 2012 (5-10 times higher than other years), with significantly higher incidence than for the State. We aimed to assess whether there were similar spatio-temporal patterns between the two outbreaks. Setting Olmsted County, Minnesota, USA Participants: We conducted a population-based retrospective cohort study of all Olmsted County residents during the 2004-2005 and 2012 outbreaks, including laboratory-positive pertussis cases. Primary outcome measure: For each outbreak, we estimated (1) age-specific incidence rate using laboratory-positive pertussis cases (numerator) and the Rochester Epidemiology Project Census (denominator), a medical record-linkage system for virtually all Olmsted County residents, and (2) pertussis case density using kernel density estimation to identify areas with high case density. To account for population size, we calculated relative difference of observed density and expected density based on age-specific incidence. Results: We identified 157 and 195 geocoded cases in 2004-2005 and 2012, respectively. Incidence was the highest among adolescents (ages 11 to <14 years) for both outbreaks (9.6 and 7.9 per 1000). The 2004-2005 pertussis outbreak had higher incidence in winter (52% of cases) versus summer in 2012 (53%). We identified a consistent area with higher incidence at the beginning (ie, first quartile) of two outbreaks, but it was inconsistent for later quartiles. The relative difference maps for the two outbreaks suggest a greater role of neighbourhood population size in 2012 compared with 2004-2005. Conclusions: Comparing spatio-temporal patterns between two pertussis outbreaks identified a consistent geographical area with higher incidence of pertussis at the beginning of outbreaks in this community. This finding can be tested in future outbreaks, and, if confirmed, can be used for identifying epidemiological risk factors clustered in such areas for geographically targeted intervention. © 2018 Author(s) (or their employer(s)).Item Spatiotemporal Association between Temperature and Assaults: A Generalized Linear Mixed-Model Approach(Sage Publications Inc.) Jung, Yeondae; Kim, Dohyeong; Piquero, Alex R.; 0000-0002-1428-1451 (Kim, D); 0000-0003-4198-4985 (Piquero, AR); 2088022 (Piquero, AR); Jung, Yeondae; Kim, Dohyeong; Piquero, Alex R.We aim to analyze the association between temperature and assault at highly disaggregated spatial units with great temporal resolution to investigate their spatiotemporal dynamics. We applied generalized linear mixed models (GLMMs) to assault and weather data from 2015, aggregated weekly at 424 subdistricts in Seoul, South Korea, controlling for various socioeconomic and environmental variables. Analyses revealed a positive and significant linear association between temperature and assaults and a few small but significant interaction effects that relate to an increase in assaults. A more enhanced understanding of the spatiotemporal relationship between temperature and crime would provide useful implications for targeted crime prevention and resource allocations. ©2019 The Author(s).Item Three Essays on Hispanics: the Use of Spanish, Content Analysis and Social Network Analysis(2022-05-01T05:00:00.000Z) Gutierrez Mannix, Carlos Daniel; Ho, Karl; Kim, Dohyeong; Tiefelsdorf, Michael; Santoro, Lauren Ratliff; Gray, ThomasHispanic political representation continues to grow in the United States. Understanding how Hispanic politicians communicate is an important step in having a better picture of how Hispanic politicians operate in Congress. It also allows us to understand what lies behind policies that are most salient for this community. This research adds to our knowledge of how Hispanic Members of Congress communicate by focusing on their use of social media. To do so this research is divided into three main sections and uses original datasets created by the author. Specifically, this research uses Twitter data by the official Twitter Accounts of Hispanic Members of Congress. Using an Application Programming Interface, this research analyzes all tweets by these politicians. Using Content Analysis, in the first section of this dissertation, I explore what factors contribute to the use of Spanish by Hispanic Members of Congress. This section explores the cultural, historical, and political factors that can be enabled to better understand how language is used by minority populations such as the Hispanic. The second section of this research also uses Content Analysis. It focuses on analyzing the connection between the historical background of the three major subpopulations of Hispanics: Puerto Ricans, Cubans and Mexicans and their policy goals. This research shows that Hispanics should not be understood to be a homogenous population but rather a conglomeration of nations and cultures which Speak Spanish. The last section of this dissertation uses Social Network Analysis to understand what factors best explain the dynamics of communication between Hispanic Members of Congress. This last section presents compelling evidence that communication between Hispanic Members of Congress should be understood to respond to political factors.Item Three Essays on Traffic Crashes: Global Health Burdens and Policy Implementations(2019-07-12) Kim, Hyoungah; Kim, DohyeongThe three studies that compose this dissertation explore important issues relating to road traffic accidents in three different geographical areas in the world. The first study in Dallas County, Texas examines community-level social inequity and the risk of pedestrian crashes. Using spatio-temporal Bayesian models, the study identifies the disproportional burden of pedestrian-vehicle crash injuries and fatalities at the community level. However, multilevel models show the disproportional burdens to be statistically suppressed when the information of crash locations is included in the analysis. Findings show that individual crash-level risk factors are more strongly associated with fatal pedestrian crashes than community-level risk factors. The policy implications of the study suggest that municipalities can decrease fatal crashes by installing and maintaining street lights. The second study evaluates the community-based traffic safety program in South Korea. Employing a synthetic control method, the program is found to reduce the number of traffic accidents and injuries, but not the number of traffic accident deaths. All estimated effects are tested with a permutation method and verified by statistical significance. The third study focuses on road traffic accidents in Lagos State, Nigeria. The average travel distance for a car crash victim to reach medical care is established by measuring the distances from a potential crash point to the nearest hospital. Findings show that victims injured at the major-minor road intersections travel shorter distances than ones at the major-residential road intersections. In addition, the average travel distance in suburban areas is almost three and half times longer than in urban areas.Item Understanding Needs and Barriers to Using Geospatial Tools for Public Health Policymaking in ChinaKim, Dohyeong; Zhang, Yingyuan; Lee, Chang Kil; Kim, Dohyeong; Zhang, YingyuanDespite growing popularity of using geographical information systems and geospatial tools in public health fields, these tools are only rarely implemented in health policy management in China. This study examines the barriers that could prevent policy-makers from applying such tools to actual managerial processes related to public health problems that could be assisted by such approaches, e.g. evidence-based policy-making. A questionnaire-based survey of 127 health-related experts and other stakeholders in China revealed that there is a consensus on the needs and demands for the use of geospatial tools, which shows that there is a more unified opinion on the matter than so far reported. Respondents pointed to lack of communication and collaboration among stakeholders as the most significant barrier to the implementation of geospatial tools. Comparison of survey results to those emanating from a similar study in Bangladesh revealed different priorities concerning the use of geospatial tools between the two countries. In addition, the follow- up in-depth interviews highlighted the political culture specific to China as a critical barrier to adopting new tools in policy development. Other barriers included concerns over the limited awareness of the availability of advanced geospatial tools. Taken together, these findings can facilitate a better understanding among policy-makers and practitioners of the challenges and opportunities for widespread adoption and implementation of a geospatial approach to public health policy-making in China.Item Understanding Spatial and Contextual Factors Influencing Intraregional Differences in Child Vaccination Coverage in Bangladesh(Sage Publications Inc, 2018-11-30) Vyas, Priyanka; Kim, Dohyeong; Adams, Alayne; 0000-0002-1428-1451 (Kim, D); Kim, DohyeongIn Bangladesh, policy discourse has mostly focused on regional inequities in health, including child immunization coverage. Knowledge of local geographical and contextual factors within regions, however, becomes pertinent in efforts to address these inequities. We used the Bangladesh Demographic and Health Survey 2011 to examine factors that influence intraregional differences in vaccination coverage using a multilevel analysis. We found that in spite of the provision of health facilities at each level of administrative governance, only distance to the Upazilla Health Complex was a consistent predictor for each dose of vaccine, highlighting the remote locations of the communities that remain underserved. Our analysis demonstrates the value of subregional analyses that identify the characteristics of communities that are vulnerable to incomplete immunization coverage. Unless specific policy actions are taken to increase coverage in these remote areas, geographic inequities are likely to persist within regions, and desired targets will not be achieved.Item Unlocking the Transformative Potential of Health Information Technology (Health IT): Institutional, Ecological & Contextual Influences Shaping Health IT Adoption Regionally(May 2023) Santana, Adela; Hicks, Donald A.; Harris, Michelle; Kim, Dohyeong; Gorina, Evgenia; Scotch, Richard K.Since the 1960s, health IT has evolved into a variety of forms and functions, embedded in all aspects of hospital care. This innovation explosion paved the ways for new ways of delivering care, including wearable and portal technologies that empower the individual to be more knowledgeable of, and involved in their own care. At the same time, adoption of these technologies has been slower than anticipated -- or desired -- among healthcare leaders. In response, President Obama signed into law the Health Information Technology for Economic and Clinical Health (HITECH) Act as part of the American Recovery and Reinvestment Act of 2009 “to promote the adoption and meaningful use of health information technology” (HHS.gov). While this federal policy was undoubtedly successful in spurring adoption and implementation of health IT, some hospitals did so more rapidly than others. Knowing what we know about innovation characteristics and given promising outcomes of health IT adoption, it is reasonable to assume that we may not have yet witnessed the full extent of health IT’s potential and are likely just at the cusp of its explosive growth. It is therefore important to understand the barriers to, and facilitators of, health IT adoption, so we can ensure a nurturing environment for innovation adoption and implementation. Review of the literature shows that overall focus has been on factors that emanate from within the hospital, while few have focused on factors external to hospitals. This study utilizes cross-sectional data to build and test models of the hypothesized influence of hospital-related regional factors on the adoption of computer practitioner order entry (CPOE) technology across metropolitan regions, while controlling for the influence of hospital ecosystem, regional health IT support, regional economic, regional social demographic, regional vitality-impacting, regional healthcare access and population health factors. This study aims to answer the question, “What factors influence the adoption of health information technology?” Part 1 of the analysis examined the influence of static physical capital (measured as average hospital size within the region) on the adoption of CPOE by 2012 (DV1). It revealed that this dimension of hospital-related regional factors, had a positive effect, meaning that as average hospital size for the region increased, adoption of CPOE technology among hospitals in the region also increased. In contrast, Part II analyses, which focused on the next five years (2012- 2017), revealed that hospitals’ health IT investment decision-making across the region was negatively influenced by the region’s growth in physician ranks perhaps because a national Medicaid expansion policy in 2014 signaled growing healthcare utilization. Modeling for DV1 also demonstrated that health IT decision-making by hospitals in the region is exposed to a wider variety of external influences beyond those emanating from the hospital alone, such as the healthcare ecology of a region, regional economic performance, and healthcare access. For healthcare sector and government leaders, this is important to be aware of because it helps define likely patterns of health IT adoption from a regional perspective and the likely regional profiles of healthcare transformation.