Patient Location Data Research Articles

Geographic Information Systems as Data Sharing Infrastructure for Clinical Data Warehouses

Introduction: The sharing of patient location data is heavily regulated by lawsprotecting patient privacy and institutional review boards that are designed toprotect the rights and wellbeing of human subjects. Patient location data hasincredible value for research due to environmental social determinants ofhealth that heavily influence patient outcomes. The misalignment of sensitivityand utility creates a barrier for researchers wishing to understand the impactof neighborhood-level social determinants of health (SDOH) on health outcomes.Objectives: We aim to describe data warehousing solutions for the safe and ethicalenhancement of patient location data; these enhancements will enable easierdata sharing. Methods: We deploy geographic information systems (GIS) as a complementary datawarehousing service which imports patient addresses, processes the data, andreturns results back to the warehouse for others to use and disseminate.Results: We processed 1.3 million patients and added facets about their locationto our data warehouse where they may later be requested for research. Wefurther added de-identified information about the demographics of patientlocations from the United States Census Bureau which assists researchers inunderstanding the impact of location-driven SDOH.Conclusions: GIS support is a necessary component of a clinical data warehouse andsharing derived, de-identified data is both feasible and useful for research.Open-source software lowers the barrier of adoption and the cost associatedwith deriving these contextual data points.

Twenty Years of the Health Insurance Portability and Accountability Act Safe Harbor Provision: Unsolved Challenges and Ways Forward.

The Health Insurance Portability and Accountability Act (HIPAA) was an important milestone in protecting the privacy of patient data; however, the HIPAA provisions specific to geographic data remain vague and hinder the ways in which epidemiologists and geographers use and share spatial health data. The literature on spatial health and select legal and official guidance documents present scholars with ambiguous guidelines that have led to the use and propagation of multiple interpretations of a single HIPAA safe harbor provision specific to geographic data. Misinterpretation of this standard has resulted in many entities sharing data at overly conservative levels, whereas others offer definitions of safe harbors that potentially put patient data at risk. To promote understanding of, and adherence to, the safe harbor rule, this paper reviews the HIPAA law from its creation to the present day, elucidating common misconceptions and presenting straightforward guidance to scholars. We focus on the 20,000-person population threshold and the 3-digit zip code stipulation of safe harbors, which are central to the confusion surrounding how patient location data can be shared. A comprehensive examination of these 2 stipulations, which integrates various expert perspectives and relevant studies, reveals how alternative methods for safe harbors can offer researchers better data and better data protection. Much has changed in the 20 years since the introduction of the safe harbor provision; however, it continues to be the primary source of guidance (and frustration) for researchers trying to share maps, leaving many waiting for these rules to be revised in accordance with the times.

Effects of patient-level risk factors, departmental allocation and seasonality on intrahospital patient transfer patterns: network analysis applied on a Norwegian single-centre data set

ObjectivesDescribe patient transfer patterns within a large Norwegian hospital. Identify risk factors associated with a high number of transfers. Develop methods to monitor intrahospital patient flows to support capacity management...

Accuracy of Infection Control Surveillance in Identifying Genomically Confirmed Cross Transmission Clusters

Background: Infection prevention surveillance for cross transmission is often performed by manual review of microbiologic culture results to identify geotemporally related clusters. However, the sensitivity and specificity of this approach remains uncertain. Whole-genome sequencing (WGS) analysis can help provide a gold-standard for identifying cross-transmission events. Objective: We employed a published WGS program, the Philips IntelliSpace Epidemiology platform, to compare accuracy of two surveillance methods: (i.) a virtual infection practitioner (VIP) with perfect recall and automated analysis of antibiotic susceptibility testing (AST), sample collection timing, and patient location data and (ii) a novel clinical matching (CM) algorithm that provides cluster suggestions based on a nuanced weighted analysis of AST data, timing of sample collection, and shared location stays between patients. Methods: WGS was performed routinely on inpatient and emergency department isolates of Enterobacter cloacae, Enterococcus faecium, Klebsiella pneumoniae, and Pseudomonas aeruginosa at an academic medical center. Single-nucleotide variants (SNVs) were compared within core genome regions on a per-species basis to determine cross-transmission clusters. Moreover, one unique strain per patient was included within each analysis, and duplicates were excluded from the final results. Results: Between May 2018 and April 2019, clinical data from 121 patients were paired with WGS data from 28 E. cloacae, 21 E. faecium, 61 K. pneumoniae, and 46 P. aeruginosa isolates. Previously published SNV relatedness thresholds were applied to define genomically related isolates. Mapping of genomic relatedness defined clusters as follows: 4 patients in 2 E. faecium clusters and 2 patients in 1 P. aeruginosa cluster. The VIP method identified 12 potential clusters involving 28 patients, all of which were “pseudoclusters.” Importantly, the CM method identified 7 clusters consisting of 27 patients, which included 1 true E. faecium cluster of 2 patients with genomically related isolates. Conclusions: In light of the WGS data, all of the potential clusters identified by the VIP were pseudoclusters, lacking sufficient genomic relatedness. In contrast, the CM method showed increased sensitivity and specificity: it decreased the percentage of pseudoclusters by 14% and it identified a related genomic cluster of E. faecium. These findings suggest that integrating clinical data analytics and WGS is likely to benefit institutions in limiting expenditure of resources on pseudoclusters. Therefore, WGS combined with more sophisticated surveillance approaches, over standard methods as modeled by the VIP, are needed to better identify and address true cross-transmission events.Funding: This study was supported by Philips Healthcare.Disclosures: None

A Bayesian Spatiotemporal Analysis of Pediatric Group A Streptococcal Infections.

BackgroundPharyngitis due to group A Streptococcus (GAS) is a common pediatric infection. Physicians might diagnose GAS pharyngitis more accurately when given biosurveillance information about GAS activity. The availability of geographic GAS testing data may be able to assist with real-time clinical decision-making for children with throat infections. MethodsGAS rapid antigen testing data were obtained from the records of 6086 children at Boston Children’s Hospital and 8648 children at Duke University Medical Center. Records included children tested in outpatient, primary care settings. We constructed Bayesian generalized additive models, in which the outcome variable was the binary result of GAS testing, and predictor variables included smoothed functions of patient location data and both cyclic and longitudinal time data. ResultsWe observed a small degree of geographic heterogeneity, but no convincing clusters of high risk. The probability of a positive test declined during the summer months. ConclusionsFuture work should include geographic data about school catchments to identify whether GAS transmission clusters within schools.

2428. Whole-genome Sequencing to Determine Clostridium difficile Transmission

Background Clostridium difficile is a major problem in healthcare institutions due to its substantial attributable morbidity, mortality and costs. Although traditionally recognized as a nosocomial infection, there is increasing evidence that hospital-based transmission may not be as common as previously thought. Whole-genome sequencing (WGS) has superior discriminatory ability than other previously used techniques for C. difficile typing. This study aimed to investigate whether WGS could help to elucidate C. difficile transmission patterns at The Royal Melbourne Hospital (RMH).MethodsAll C. difficile isolates (N = 138) identified in patients admitted to RMH from November 1, 2015 to October 31, 2016 had molecular typing performed by WGS, multilocus sequence typing (MLST) and PCR ribotyping. Clinical epidemiological data were collected for each episode so that patient locations could be examined together with molecular typing information to determine putative transmissions in the hospital.ResultsAfter combining molecular and clinical epidemiology, a picture of diverse C. difficile emerged. Only 7 (6%) of isolates appeared to have been transmitted from other hospital patients, according to combined WGS and patient location data. However, both PCR ribotyping (33%) and MLST (44%) had significantly higher proportions of putative transmissions in this cohort.ConclusionThis finding has significant implications for the Infection Prevention team as efforts toward prevention of C. difficile infection may need to be redirected away from the current focus on prevention of nosocomial transmission. Future studies are needed to broaden understanding of C. difficile transmission dynamics so that other sources can be identified and targeted for intervention.Disclosures All authors: No reported disclosures.

Measuring access to primary healthcare services after stroke: A spatial analytic approach

AbstractObjective:To determine accessibility of the primary healthcare system for patients with stroke recently discharged from hospital.Methods:This project mapped retrospective patient location data and the location of primary healthcare services in the same region. Patient location data were from all patients with stroke (N = 1595: January 2011–January 2017) discharged from one metropolitan hospital to the local Primary Health Network. Geographic Information System technology was used to map the patient discharge locations and the spatial distribution of primary healthcare services (general practitioner, pharmacy, allied health) across the region. Road network data were used to measure the level of access from each patient’s discharge location to the services.Results:Access to primary healthcare services was variable. Areas with larger proportions of patients with stroke did not necessarily have good service access. With an increase in travel time, the number of services accessible to patients also increased. However, the spatial variation of access to services remained largely unchanged.Conclusion:Access to primary healthcare services for patients with stroke varies spatially, with a trend towards relatively low levels of accessibility for many patients. There is an urgent need for future planning to consider geographical access to primary healthcare services for patients with stroke.

A Study on the Life-Saving Impact of Tsukuba City Government's Sharing Cardiac Arrest Patient Location Data with AED Superintendents

A Study on the Life-Saving Impact of Tsukuba City Government's Sharing Cardiac Arrest Patient Location Data with AED Superintendents

Whole-genome sequencing and epidemiological analysis do not provide evidence for cross-transmission of mycobacterium abscessus in a cohort of pediatric cystic fibrosis patients.

Mycobacterium abscessus has emerged as a major pathogen in cystic fibrosis (CF) patients and has been associated with poor clinical outcomes, particularly following lung transplant. We investigated the acquisition of this bacterium in a cohort of pediatric CF patients. Demographic and patient location data were used to uncover epidemiological links between patients with genetically related strains of M. abscessus that had been previously typed by variable-number tandem repeat profiling. Whole-genome sequencing was applied to 27 M. abscessus isolates from the 20 patients in this cohort to provide definitive data on the genetic relatedness of strains. Whole-genome sequencing data demonstrated that M. abscessus isolates from 16 patients were unrelated, differing by at least 34 single-nucleotide polymorphisms (SNPs) from any other isolate, suggesting that independent acquisition events have occurred. Only 2 clusters of very closely related (<25 SNPs) isolates from different patients were seen. The first cluster contained 8 isolates, differing by a maximum of 17 SNPs, from a sibling pair who had intense exposure to each other both inside and outside the hospital. The second cluster contained 3 isolates, differing by a maximum of 24 SNPs, from 2 individuals with no apparent epidemiological links. We have not demonstrated cross-transmission of M. abscessus within our hospital, except between 1 sibling pair. Alternative routes of acquisition of M. abscessus infection, in particular the environment, require further investigation.

Automatic Detection and Notification of “Wrong Patient—Wrong Location” Errors in the Operating Room

When procedures and processes to assure patient location based on human performance do not work as expected, patients are brought incrementally closer to a possible "wrong patient-wrong procedure'' error. We developed a system for automated patient location monitoring and management. Real-time data from an active infrared/radio frequency identification tracking system provides patient location data that are robust and can be compared with an "expected process'' model to automatically flag wrong-location events as soon as they occur. The system also generates messages that are automatically sent to process managers via the hospital paging system, thus creating an active alerting function to annunciate errors. We deployed the system to detect and annunciate "patient-in-wrong-OR'' events. The system detected all "wrong-operating room (OR)'' events, and all "wrong-OR'' locations were correctly assigned within 0.50+/-0.28 minutes (mean+/-SD). This corresponded to the measured latency of the tracking system. All wrong-OR events were correctly annunciated via the paging function. This experiment demonstrates that current technology can automatically collect sufficient data to remotely monitor patient flow through a hospital, provide decision support based on predefined rules, and automatically notify stakeholders of errors.