Automated cooling tower detection through deep learning for Legionnaires' disease outbreak investigations: a model development and validation study.

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.

Large Community Outbreak of Legionnaires Disease Potentially Associated with a Cooling Tower - Napa County, California, 2022.

Legionnaires disease is a serious infection acquired by inhalation of water droplets from human-made building water systems that contain Legionella bacteria. On July 11 and 12, 2022, Napa County Public Health (NCPH) in California received reports of three positive urinary antigen tests for Legionella pneumophila serogroup 1 in the town of Napa. By July 21, six Legionnaires disease cases had been confirmed among Napa County residents, compared with a baseline of one or two cases per year. NCPH requested assistance from the California Department of Public Health (CDPH) and CDC to aid in the investigations. Close temporal and geospatial clustering permitted a focused environmental sampling strategy of high-risk facilities which, coupled with whole genome sequencing results from samples and investigation of water system maintenance, facilitated potential linking of the outbreak with an environmental source. NCPH, with technical support from CDC and CDPH, instructed and monitored remediation practices for all environmental locations that tested positive for Legionella. The investigation response to this community outbreak illustrates the importance of interdisciplinary collaboration by public health agencies, laboratory support, timely communication with the public, and cooperation of managers of potentially implicated water systems. Timely identification of possible sources, sampling, and remediation of any facility testing positive for Legionella is crucial to interrupting further transmission.

Potential Association of Legionnaires' Disease with Hot Spring Water, Hot Springs National Park and Hot Springs, Arkansas, USA, 2018-2019.

Legionella pneumophila is the cause of Legionnaires' disease, a life-threatening pneumonia that occurs after inhalation of aerosolized water containing the bacteria. Legionella growth occurs in stagnant, warm-to-hot water (77°F-113°F) that is inadequately disinfected. Piped hot spring water in Hot Springs National Park, Arkansas, USA, has naturally high temperatures (>135°F) that prevent Legionella growth, and Legionnaires' disease has not previously been associated with the park or other hot springs in the United States. During 2018-2019, Legionnaires' disease occurred in 5 persons after they visited the park; 3 of these persons were potentially exposed in spa facilities that used untreated hot spring water. Environmental testing revealed Legionella bacteria in piped spring water, including 134°F stagnant pipe water. These findings underscore the importance of water management programs to reduce Legionella growth in plumbing through control activities such as maintaining hot water temperatures, reducing stored water age, and ensuring adequate water flow.

Legionellosis Cluster Associated With Working at a Racetrack Facility in West Virginia, 2018.

In October 2018, the Centers for Disease Control and Prevention was notified of a cluster of Legionnaires' disease cases in workers at a racetrack facility. The objective of the resulting investigation was to determine the extent of the outbreak and identify potential sources of exposure to halt transmission. Case-finding and interviews were conducted among symptomatic racetrack workers who were known to be at the facility within 14 days prior to symptom onset. An environmental assessment of the facility and surrounding area was conducted for sources of potential Legionella exposure. In total, 17 legionellosis cases were identified. The environmental assessment revealed a poorly maintained hot tub in the jockey locker room as the most likely source. Further investigation identified deficiencies in the facility's ventilation systems, which suggested a transmission mechanism for workers who never entered the locker room floor. Considering indirect exposure routes via air handling systems can be useful for source identification and case-finding in legionellosis outbreaks.

Impact of home remediation and household education on indoor air quality, respiratory visits and symptoms in Alaska Native children.

Alaska Native children experience high rates of lower respiratory tract infections (LRTIs) and lung conditions, which are associated with substandard indoor air quality (IAQ). We conducted an intervention of home remediation and education to assess the impact on IAQ, respiratory symptoms and LRTI visits. We enrolled households of children 1-12 years of age with lung conditions. Home remediation included improving ventilation and replacing leaky woodstoves. We provided education about IAQ and respiratory health. We monitored indoor airborne particles (PM2.5), CO2, relative humidity and volatile organic compounds (VOCs), and interviewed caregivers about children's symptoms before, and for 1 year after intervention. We evaluated the association between children's respiratory visits, symptoms and IAQ indicators using multiple logistic regression. A total of 60 of 63 homes completed the study. VOCs decreased (coefficient = -0.20; p < 0.001); however, PM2.5 (coeff. = -0.010; p = 0.89) did not decrease. Burning wood for heat, VOCs and PM2.5 were associated with respiratory symptoms. After remediation, parents reported decreases in runny nose, cough between colds, wet cough, wheezing with colds, wheezing between colds and school absences. Children had an age-adjusted decrease in LRTI visits (coefficient = -0.33; p = 0.028). Home remediation and education reduced respiratory symptoms, LRTI visits and school absenteeism in children with lung conditions.

Consuming untreated water in four southwestern Alaska Native communities: reasons revealed and recommendations for change.

In this article, the authors provide the first in-depth account of why some Alaska Native people drink untreated water when treated water is available. Their qualitative research was conducted in four Alaska Native village communities that have treated water available from a centralized distribution point. Most respondents (n = 172; 82%) reported that some of their household's drinking water came from an untreated source. Motives for drinking untreated water emerged from analysis of open-ended questions about drinking water practice and could be categorized into six themes: chemicals, taste, health, access, tradition, and cost. Importantly, some residents reported consuming untreated water because they both liked untreated water and disliked treated water. As such, interventions to increase safe water consumption should address this dichotomy by providing education about the benefits of treated water alongside the risks involved with drinking untreated water. Based on the findings, the authors provide specific recommendations for developing behavior change interventions that address influences at multiple social-ecological levels.

Invasive pneumococcal disease in Alaskan children: impact of the seven-valent pneumococcal conjugate vaccine and the role of water supply.

BACKGROUND: Alaska Native (AN) children, especially those in the Yukon-Kuskokwim region (YK-AN children), suffer some of the highest rates of invasive pneumococcal disease (IPD) in the world. Rates of IPD declined after statewide introduction of the 7-valent pneumococcal conjugate vaccine (PCV7) in 2001, but increased in subsequent years. METHODS: Population-based laboratory surveillance data (1986-2007) for invasive Streptococcus pneumoniae infection in Alaskan children <5 years old were used to evaluate the association of IPD rates and serotype distribution with immunization, socioeconomic status, and in-home water service. RESULTS: Introduction of PCV7 vaccine resulted in elimination of IPD caused by vaccine serotypes, but was followed by increasing rates of IPD caused by nonvaccine serotypes. Among YK-AN children IPD rates dropped by 60%, but then rose due to non-PCV7 serotypes to levels 5- to 10-fold higher than rates in non-YK-AN children and non-AN children. IPD rates in YK-AN children were twice as high in villages where <10% of houses had in-home piped water compared with villages where more than 80% of houses had in-home piped water (390 cases/100,000 vs. 146 cases/100,000, P = 0.008). CONCLUSIONS: High IPD rates in Alaska are associated with lack of in-home piped water (controlling for household crowding and per capita income). The effect of in-home piped water is most likely mediated through reduced water supply leading to limitations on handwashing.

The relationship between in-home water service and the risk of respiratory tract, skin, and gastrointestinal tract infections among rural Alaska natives.

OBJECTIVES: We investigated the relationship between the presence of in-home piped water and wastewater services and hospitalization rates for respiratory tract, skin, and gastrointestinal tract infections in rural Alaska. METHODS: We determined in-home water service and hospitalizations for selected infectious diseases among Alaska Natives by region during 2000 to 2004. Within 1 region, infant respiratory hospitalizations and skin infections for all ages were compared by village-level water services. RESULTS: Regions with a lower proportion of home water service had significantly higher hospitalization rates for pneumonia and influenza (rate ratio [RR] = 2.5), skin or soft tissue infection (RR = 1.9), and respiratory syncytial virus (RR = 3.4 among those younger than 5 years) than did higher-service regions. Within 1 region, infants from villages with less than 10% of homes served had higher hospitalization rates for pneumonia (RR = 1.3) and respiratory syncytial virus (RR = 1.2) than did infants from villages with more than 80% served. Outpatient Staphylococcus aureus infections (RR = 5.1, all ages) and skin infection hospitalizations (RR = 2.7, all ages) were higher in low-service than in high-service villages. CONCLUSIONS: Higher respiratory and skin infection rates were associated with a lack of in-home water service. This disparity should be addressed through sanitation infrastructure improvements.

Sharing environmental health practice in the North American Arctic: a focus on water and wastewater service.

While providing water and wastewater services in rural Alaska has traditionally been difficult, the task has become even more challenging in the new millennium. A combination of factors has increased the cost and complexity of water and wastewater systems while leaving residents with diminished ability to pay for services. Alaskans may be better able to meet these challenges by learning more about and adopting some approaches used in similar regions. In March 2005, the author visited six communities in the Canadian Northwest Territories (NWT). Made possible by NEHA's Sabbatical Exchange Ambassador Award, these visits provided the author an opportunity to learn about the NWT approach to providing water and wastewater services. He then compiled comprehensive descriptions of each region's service delivery model. Comparisons were made, and features of interest were identified and used to develop recommendations for improving service delivery in rural Alaska.