Pub Date : 2015-06-01Epub Date: 2016-01-14DOI: 10.1109/GEOINFORMATICS.2015.7378608
Xuan Shi, Bowei Xue, Imam Xierali
The spatial distribution of physicians has a significant impact in public health research. It is critical to clarify whether the addresses provided by the physicians are the home addresses or the practice addresses, since the practice address is the key to understand relevant issues of maldistribution, accessibility and disparity. Through a pilot study as partial effort of the research project "Reducing Physician Distribution Uncertainty in Spatial Accessibility Research" sponsored by the National Institutes of Health (NIH award number 1R21CA182874-01), appropriate solutions were developed to differentiate the home addresses from practice addresses. This paper introduces how to understand the clustering patterns in physician distribution through Affinity Propagation, a relatively new clustering algorithm, to derive the potential extent of the practice locations for those physicians who provided home addresses. The physician data is derived from the 2014 American Medical Association (AMA) Physician Masterfile, while two counties (Fulton and DeKalb) in the metropolitan area of Atlanta, Georgia were selected as the study area. Both Euclidian distance and driving distance were applied in the AP algorithm, while gravity models based AP calculation were applied in comparison to the clustering of individual physicians. By justifying preference and similarity parameters in the AP calculation, hierarchical clustering patterns can be derived and perceived. Future research challenges in AP clustering are identified, while this pilot study can be extended with broader impact in public health research.
医生的空间分布对公共卫生研究具有重要影响。澄清医生提供的地址是家庭地址还是执业地址至关重要，因为执业地址是了解相关分布不当、可及性和差异问题的关键。作为美国国立卫生研究院资助的研究项目 "减少空间可及性研究中医生分布不确定性"（NIH 奖项编号 1R21CA182874-01）的一部分工作，通过试点研究，开发出了适当的解决方案来区分家庭地址和执业地址。本文介绍了如何通过亲和传播（一种相对较新的聚类算法）了解医生分布的聚类模式，从而推导出提供家庭地址的医生的潜在执业地点范围。医生数据来自 2014 年美国医学会（AMA）医生主档案，研究区域则选取了佐治亚州亚特兰大大都会区的两个县（富尔顿县和德卡尔布县）。在 AP 算法中应用了欧几里得距离和驾车距离，而基于重力模型的 AP 计算则应用于单个医生的聚类比较。通过证明 AP 计算中的偏好和相似性参数，可以得出并感知分层聚类模式。我们确定了 AP 聚类的未来研究挑战，而这项试验性研究可以在公共卫生研究中产生更广泛的影响。
{"title":"Understanding the Clustering Patterns in Physician Distribution Through Affinity Propagation.","authors":"Xuan Shi, Bowei Xue, Imam Xierali","doi":"10.1109/GEOINFORMATICS.2015.7378608","DOIUrl":"10.1109/GEOINFORMATICS.2015.7378608","url":null,"abstract":"<p><p>The spatial distribution of physicians has a significant impact in public health research. It is critical to clarify whether the addresses provided by the physicians are the home addresses or the practice addresses, since the practice address is the key to understand relevant issues of maldistribution, accessibility and disparity. Through a pilot study as partial effort of the research project \"Reducing Physician Distribution Uncertainty in Spatial Accessibility Research\" sponsored by the National Institutes of Health (NIH award number 1R21CA182874-01), appropriate solutions were developed to differentiate the home addresses from practice addresses. This paper introduces how to understand the clustering patterns in physician distribution through Affinity Propagation, a relatively new clustering algorithm, to derive the potential extent of the practice locations for those physicians who provided home addresses. The physician data is derived from the 2014 American Medical Association (AMA) Physician Masterfile, while two counties (Fulton and DeKalb) in the metropolitan area of Atlanta, Georgia were selected as the study area. Both Euclidian distance and driving distance were applied in the AP algorithm, while gravity models based AP calculation were applied in comparison to the clustering of individual physicians. By justifying preference and similarity parameters in the AP calculation, hierarchical clustering patterns can be derived and perceived. Future research challenges in AP clustering are identified, while this pilot study can be extended with broader impact in public health research.</p>","PeriodicalId":92217,"journal":{"name":"International Conference on Geoinformatics : [proceedings]. International Conference on Geoinformatics","volume":"2015 ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2015-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5791758/pdf/nihms872288.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"35792911","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号