1
|
Wong KK, Segura T, Mein G, Lu J, Hannapel EJ, Kunz JM, Ritter T, Smith JC, Todeschini A, Nugen F, Edens C. Automated cooling tower detection through deep learning for Legionnaires' disease outbreak investigations: a model development and validation study. Lancet Digit Health 2024; 6:e500-e506. [PMID: 38906615 DOI: 10.1016/s2589-7500(24)00094-3] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/23/2023] [Revised: 04/10/2024] [Accepted: 05/01/2024] [Indexed: 06/23/2024]
Abstract
BACKGROUND Cooling towers containing Legionella spp are a high-risk source of Legionnaires' disease outbreaks. Manually locating cooling towers from aerial imagery during outbreak investigations requires expertise, is labour intensive, and can be prone to errors. We aimed to train a deep learning computer vision model to automatically detect cooling towers that are aerially visible. METHODS Between Jan 1 and 31, 2021, we extracted satellite view images of Philadelphia (PN, USA) and New York state (NY, USA) from Google Maps and annotated cooling towers to create training datasets. We augmented training data with synthetic data and model-assisted labelling of additional cities. Using 2051 images containing 7292 cooling towers, we trained a two-stage model using YOLOv5, a model that detects objects in images, and EfficientNet-b5, a model that classifies images. We assessed the primary outcomes of sensitivity and positive predictive value (PPV) of the model against manual labelling on test datasets of 548 images, including from two cities not seen in training (Boston [MA, USA] and Athens [GA, USA]). We compared the search speed of the model with that of manual searching by four epidemiologists. FINDINGS The model identified visible cooling towers with 95·1% sensitivity (95% CI 94·0-96·1) and a PPV of 90·1% (95% CI 90·0-90·2) in New York City and Philadelphia. In Boston, sensitivity was 91·6% (89·2-93·7) and PPV was 80·8% (80·5-81·2). In Athens, sensitivity was 86·9% (75·8-94·2) and PPV was 85·5% (84·2-86·7). For an area of New York City encompassing 45 blocks (0·26 square miles), the model searched more than 600 times faster (7·6 s; 351 potential cooling towers identified) than did human investigators (mean 83·75 min [SD 29·5]; mean 310·8 cooling towers [42·2]). INTERPRETATION The model could be used to accelerate investigation and source control during outbreaks of Legionnaires' disease through the identification of cooling towers from aerial imagery, potentially preventing additional disease spread. The model has already been used by public health teams for outbreak investigations and to initialise cooling tower registries, which are considered best practice for preventing and responding to outbreaks of Legionnaires' disease. FUNDING None.
Collapse
Affiliation(s)
- Karen K Wong
- Centers for Disease Control and Prevention, Atlanta, GA, USA; University of California, Berkeley, CA, USA.
| | | | | | - Jia Lu
- University of California, Berkeley, CA, USA
| | | | - Jasen M Kunz
- Centers for Disease Control and Prevention, Atlanta, GA, USA
| | - Troy Ritter
- Centers for Disease Control and Prevention, Atlanta, GA, USA
| | - Jessica C Smith
- Centers for Disease Control and Prevention, Atlanta, GA, USA
| | | | - Fred Nugen
- University of California, Berkeley, CA, USA
| | - Chris Edens
- Centers for Disease Control and Prevention, Atlanta, GA, USA
| |
Collapse
|
2
|
Mandelbaum J, Almeda J, Blackwell S, Hopkins JW, Myers K, Hicks S, Daguise VG. An Analysis of the Social Determinants of Health in South Carolina's I-95 Corridor. Health Promot Pract 2024; 25:335-345. [PMID: 36546686 DOI: 10.1177/15248399221142517] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 02/17/2024]
Abstract
BACKGROUND One in four South Carolinians lives in a county along a nearly 200-mile stretch of Interstate 95 (I-95). Stretching from North Carolina to Georgia, this region is among the most rural, economically depressed, and racially/ethnically diverse in the state. Research is needed to identify social factors contributing to adverse health outcomes along the I-95 corridor, guide interventions, and establish a baseline for measuring progress. This study assessed social determinants of health in counties in South Carolina's I-95 corridor relative to the rest of the state. METHOD Data for South Carolina's 46 counties were extracted from the Centers for Disease Control and Prevention Minority Health Social Vulnerability Index (SVI), which grouped 34 census variables into six themes: socioeconomic status, household composition and disability, minority status and language, housing type and transportation, health care infrastructure, and medical vulnerability. Each theme was ranked from 0 (least vulnerable) to 1 (most vulnerable). Measures between regions were compared using the Wilcoxon-Mann-Whitney test. RESULTS Compared with counties outside the I-95 corridor (n = 29), counties in the corridor (n = 17) scored higher on socioeconomic status vulnerability (.67 and .82, respectively) and medical vulnerability (.65 and .79, respectively). No statistically significant differences were found across other themes. CONCLUSION Identifying social determinants of health in South Carolina's I-95 corridor is a crucial first step toward alleviating health disparities in this region. Interventions and policies should be developed in collaboration with local stakeholders to address distal social factors that create and reinforce health disparities.
Collapse
Affiliation(s)
- Jennifer Mandelbaum
- South Carolina Department of Health and Environmental Control, Columbia, SC, USA
- University of South Carolina, Columbia, SC, USA
| | - Jennifer Almeda
- South Carolina Department of Health and Environmental Control, Columbia, SC, USA
| | - Shanikque Blackwell
- South Carolina Department of Health and Environmental Control, Columbia, SC, USA
| | - John W Hopkins
- South Carolina Department of Health and Environmental Control, Columbia, SC, USA
| | - Kristian Myers
- South Carolina Department of Health and Environmental Control, Columbia, SC, USA
| | - Shauna Hicks
- South Carolina Department of Health and Environmental Control, Columbia, SC, USA
| | - Virginie G Daguise
- South Carolina Department of Health and Environmental Control, Columbia, SC, USA
| |
Collapse
|
3
|
Yu F, Nair AA, Lauper U, Luo G, Herb J, Morse M, Savage B, Zartarian M, Wang M, Lin S. Mysteriously rapid rise in Legionnaires' disease incidence correlates with declining atmospheric sulfur dioxide. PNAS NEXUS 2024; 3:pgae085. [PMID: 38476666 PMCID: PMC10929586 DOI: 10.1093/pnasnexus/pgae085] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Figures] [Subscribe] [Scholar Register] [Received: 12/27/2023] [Accepted: 02/09/2024] [Indexed: 03/14/2024]
Abstract
Legionnaires' disease (LD) is a severe form of pneumonia (∼10-25% fatality rate) caused by inhalation of aerosols containing Legionella, a pathogenic gram-negative bacteria. These bacteria can grow, spread, and aerosolize through building water systems. A recent dramatic increase in LD incidence has been observed globally, with a 9-fold increase in the United States from 2000 to 2018, and with disproportionately higher burden for socioeconomically vulnerable subgroups. Despite the focus of decades of research since the infamous 1976 outbreak, substantial knowledge gaps remain with regard to source of exposure and the reason(s) for the dramatic increase in LD incidence. Here, we rule out factors indicated in literature to contribute to its long-term increases and identify a hitherto unexplored explanatory factor. We also provide an epidemiological demonstration that the occurrence of LD is linked with exposure to cooling towers (CTs). Our results suggest that declining sulfur dioxide air pollution, which has many well-established health benefits, results in reduced acidity of aerosols emitted from CTs, which may prolong the survival duration of Legionella in contaminated CT droplets and contribute to the increase in LD incidence. Mechanistically associating decreasing aerosol acidity with this respiratory disease has implications for better understanding its transmission, predicting future risks, and informed design of preventive and interventional strategies that consider the complex impacts of continued sulfur dioxide changes.
Collapse
Affiliation(s)
- Fangqun Yu
- Atmospheric Sciences Research Center, University at Albany, State University of New York, Albany, NY 12226, USA
| | - Arshad A Nair
- Atmospheric Sciences Research Center, University at Albany, State University of New York, Albany, NY 12226, USA
| | - Ursula Lauper
- New York State Department of Health, Bureau of Water Supply Protection, Albany, NY 12223, USA
| | - Gan Luo
- Atmospheric Sciences Research Center, University at Albany, State University of New York, Albany, NY 12226, USA
| | - Jason Herb
- Atmospheric Sciences Research Center, University at Albany, State University of New York, Albany, NY 12226, USA
| | - Matthew Morse
- New York State Department of Health, Bureau of Water Supply Protection, Albany, NY 12223, USA
| | - Braden Savage
- New York State Department of Health, Bureau of Water Supply Protection, Albany, NY 12223, USA
| | - Martin Zartarian
- New York State Department of Health, Bureau of Water Supply Protection, Albany, NY 12223, USA
| | - Meng Wang
- School of Public Health and Health Professions, University at Buffalo, State University of New York, Buffalo, NY 14214, USA
| | - Shao Lin
- School of Public Health, University at Albany, State University of New York, Albany, NY 12144, USA
| |
Collapse
|
4
|
Moffa MA, Rock C, Galiatsatos P, Gamage SD, Schwab KJ, Exum NG. Legionellosis on the rise: A scoping review of sporadic, community-acquired incidence in the United States. Epidemiol Infect 2023; 151:e133. [PMID: 37503568 PMCID: PMC10540183 DOI: 10.1017/s0950268823001206] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/06/2023] [Revised: 06/14/2023] [Accepted: 07/19/2023] [Indexed: 07/29/2023] Open
Abstract
Over the past two decades, the incidence of legionellosis has been steadily increasing in the United States though there is noclear explanation for the main factors driving the increase. While legionellosis is the leading cause of waterborne outbreaks in the US, most cases are sporadic and acquired in community settings where the environmental source is never identified. This scoping review aimed to summarise the drivers of infections in the USA and determine the magnitude of impact each potential driver may have. A total of 1,738 titles were screened, and 18 articles were identified that met the inclusion criteria. Strong evidence was found for precipitation as a major driver, and both temperature and relative humidity were found to be moderate drivers of incidence. Increased testing and improved diagnostic methods were classified as moderate drivers, and the ageing U.S. population was a minor driver of increasing incidence. Racial and socioeconomic inequities and water and housing infrastructure were found to be potential factors explaining the increasing incidence though they were largely understudied in the context of non-outbreak cases. Understanding the complex relationships between environmental, infrastructure, and population factors driving legionellosis incidence is important to optimise mitigation strategies and public policy.
Collapse
Affiliation(s)
- Michelle A. Moffa
- Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD, USA
| | - Clare Rock
- Division of Infectious Diseases, Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD, USA
- Department of Hospital Epidemiology and Infection Control, Johns Hopkins Hospital, Baltimore, MD, USA
| | - Panagis Galiatsatos
- Medicine for the Greater Good, Johns Hopkins University School of Medicine, Baltimore, MD, USA
- Division of Pulmonary and Critical Care Medicine, Johns Hopkins University School of Medicine, Baltimore, MD, USA
| | - Shantini D. Gamage
- U.S. Department of Veterans Affairs, National Infectious Diseases Service, Veterans Health Administration, Washington, DC, USA
- Division of Infectious Diseases, Department of Internal Medicine, University of Cincinnati College of Medicine, Cincinnati, OH, USA
| | - Kellogg J. Schwab
- Department of Environmental Health and Engineering, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA
| | - Natalie G. Exum
- Department of Environmental Health and Engineering, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA
| |
Collapse
|
5
|
Barskey AE, Derado G, Edens C. Rising Incidence of Legionnaires' Disease and Associated Epidemiologic Patterns, United States, 1992-2018. Emerg Infect Dis 2022; 28:527-538. [PMID: 35195513 PMCID: PMC8888234 DOI: 10.3201/eid2803.211435] [Citation(s) in RCA: 8] [Impact Index Per Article: 4.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/22/2023] Open
Abstract
Reported Legionnaires' disease (LD) cases began increasing in the United States in 2003 after relatively stable numbers for >10 years; reasons for the rise are unclear. We compared epidemiologic patterns associated with cases reported to the Centers for Disease Control and Prevention before and during the rise. The age-standardized average incidence was 0.48 cases/100,000 population during 1992-2002 compared with 2.71 cases/100,000 in 2018. Reported LD incidence increased in nearly every demographic, but increases tended to be larger in demographic groups with higher incidence. During both periods, the largest number of cases occurred among White persons, but the highest incidence was in Black or African American persons. Incidence and increases in incidence were generally largest in the East North Central, Middle Atlantic, and New England divisions. Seasonality was more pronounced during 2003-2018, especially in the Northeast and Midwest. Rising incidence was most notably associated with increasing racial disparities, geographic focus, and seasonality.
Collapse
|
6
|
Li S, Yin Z, Lesser J, Li C, Choi BY, Parra-Medina D, Flores B, Dennis B, Wang J. A Community Health Worker-Led mHealth-Enabled Diabetes Self-Management Education and Support Intervention in Rural Latino Adults: Single-Arm Feasibility Trial (Preprint). JMIR Diabetes 2022; 7:e37534. [PMID: 35635752 PMCID: PMC9153909 DOI: 10.2196/37534] [Citation(s) in RCA: 4] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/24/2022] [Revised: 04/19/2022] [Accepted: 05/02/2022] [Indexed: 11/13/2022] Open
Abstract
Background Latinos living in rural South Texas have a higher prevalence of diabetes, but their access to diabetes self-management education and support (DSMES) is limited. Objective We aimed to test the feasibility of a community health worker-led, mobile health (mHealth)-based DSMES intervention to reduce disparities in accessing DSMES in underserved rural Latino residents in South Texas. Methods This 12-week, single-arm, pre-post trial was delivered by trained community health workers to 15 adults with type 2 diabetes. The intervention consisted of digital diabetes education, self-monitoring, a cloud-based connected platform, and community health worker support. Feasibility was evaluated as retention, actual intervention use, program satisfaction, and barriers to implementation. We also explored the intervention’s effect on weight loss and hemoglobin A1c (HbA1c). Results All 15 participants were Latino (mean age 61.87 years, SD 10.67; 9/15 female, 60%). The retention rate at posttest was 14 of 15 (93%). On average, the participants completed 37 of 42 (88%) digital diabetes education lessons with 8 participants completing all lessons. Participants spent 81/91 days (89%) step tracking, 71/91 days (78%) food logging, 43/91 days (47%) blood glucose self-monitoring, and 74/91 days (81%) weight self-monitoring. The level of program satisfaction was high. On average, participants lost 3.5 (SD 3.2) kg of body weight (P=.001), while HbA1c level remained unchanged from baseline (6.91%, SD 1.28%) to posttest (7.04%, SD 1.66%; P=.668). Conclusions A community health worker-led mHealth-based intervention was feasible and acceptable to improve access to DSMES services for Latino adults living in rural communities. Future randomized controlled trials are needed to test intervention efficacy on weight loss and glycemic control.
Collapse
Affiliation(s)
- Shiyu Li
- Center on Smart and Connected Health Technologies, School of Nursing, University of Texas Health Science Center at San Antonio, San Antonio, TX, United States
| | - Zenong Yin
- Department of Public Health, The University of Texas at San Antonio, San Antonio, TX, United States
| | - Janna Lesser
- School of Nursing, University of Texas Health Science Center at San Antonio, San Antonio, TX, United States
| | - Chengdong Li
- College of Nursing, Florida State University, Tallahassee, FL, United States
| | - Byeong Yeob Choi
- Department of Population Health Sciences, University of Texas Health Science Center at San Antonio, San Antonio, TX, United States
| | - Deborah Parra-Medina
- Latino Research Institute, The University of Texas at Austin, Austin, TX, United States
| | - Belinda Flores
- South Coastal Area Health Education Center, Corpus Christi, TX, United States
| | - Brittany Dennis
- Center on Smart and Connected Health Technologies, School of Nursing, University of Texas Health Science Center at San Antonio, San Antonio, TX, United States
| | - Jing Wang
- College of Nursing, Florida State University, Tallahassee, FL, United States
| |
Collapse
|