Abstract
Climate change, the greatest public health threat of the 21st century, will uniquely affect rural areas that are geographically isolated and experience greater health inequities. This systematic review describes and evaluates interventions to lessen the effects of climate change on human health in the rural United States, including interventions on air pollution, vector ecology, water quality, severe weather, extreme heat, allergens, and water and food supply. Searches were constructed based on the eight domains of the Centers for Disease Control and Prevention (CDC) Framework "Impact of Climate Change on Human Health." Searches were conducted in EBSCO Environment Complete, EBSCO GreenFILE, Embase.com, MEDLINE via PubMed, and Web of Science. Duplicate citations were removed, abstracts were screened for initial inclusion, and full texts were screened for final inclusion. Pertinent data were extracted and synthesized across the eight domains. Article quality was assessed using the Mixed Methods Appraisal Tool. Of 8471 studies screened, 297 were identified for full text review, and a total 49 studies were included in this review. Across the domains, 34 unique interventions addressed health outcomes due to air pollution (n = 8), changes in vector ecology (n = 6), water quality (n = 5), severe weather (n = 3), extreme heat (n = 2) increasing allergens (n = 1), water and food supply (n = 1), and across multiple CDC domains (n = 8). Participatory action research methodology was commonly used and strived to mobilize/empower communities to tackle climate change. Our review identified three randomized controlled trials, with two of these three published in the last five years. While original research on the impact of climate change on health has increased in the past decade, randomized control trials may not be ethical, cost effective, or feasible. There is a need for time-efficient and high-quality scholarship that investigates intervention efficacy and effectiveness for reducing health impacts of climate change upon rural populations.
Export citation and abstract BibTeX RIS
Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 license. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Corrections were made to this article on 21 July 2023. An update was made to the institution name in affiliation 3.
Climate change is one of the greatest public health threats of the 21st century and is the result of increases in greenhouse gas concentrations, including CO2 levels in the earth's atmosphere, leading to increasing temperatures, extreme weather events, and rising sea levels [1]. In the past two decades, non-optimal temperatures due to climate change have been linked to an estimated 5 million deaths globally per year [2]. The World Health Organization (WHO) estimates that between 2030 and 2050, climate change will cause an additional 250 000 deaths globally per year, from heat stress, diarrhea, and malnutrition [3]. Without mitigation efforts, the cost of climate-related health impacts in the States is expected to reach tens to hundreds of billions of dollars annually by the end of this century [4]. Climate change will likely widen already existing health inequities, since lower-income and other marginalized communities face greater challenges preparing for, and coping with, extreme weather and climate-related events [5].
Compared to urban areas within the States, climate change is expected to disproportionately impact rural areas due to a great presence of outdoor workers, [6, 7] longer transportation times to hospitals, [8] and limited heat warnings or community cooling centers [9]. Additionally, rural inhabitants tend to be older and have lower incomes, further decreasing their capacity to respond to climate change related health impacts [10]. Increased intensity and frequency of extreme heat and weather events [11] coupled with decreased food production in rural areas, necessitate climate change mitigation and adaptation efforts that target rural residents [12]. Furthermore, the rural States is home to at least 19% of the States population and occupies 95% of the country's land area [13]. Most research to date has focused on the impacts of climate change in urban settings. A 2021 review of systematic reviews of 94 articles found only 1 article focused on the health impacts of climate change within rural populations [14]. Another systematic review of 14 studies, [15] which included 2 studies in the States and examined heat-related mortality risk in rural areas, found that rural populations are no less vulnerable to heat-related mortality than populations living in urban 'heat islands' [15]. A 2013 systematic review of 33 systematic reviews examined public health interventions to address vector-borne and water-borne diseases and heat stress [16]. The authors of this review found only three unique health interventions for heat stress, two of which specifically targeted urban areas [16]. Despite evidence describing the impact of increasing global temperatures on rural communities, evaluations of broader climate change-related interventions within rural areas are lacking.
1. Purpose
The purpose of this review is to appraise the evidence, the quality of the available evidence, and to inform gaps in knowledge on intervention studies that aimed to address the effects of climate change within the rural States. The review was guided by the Centers for Disease Control and Prevention (CDC) Impact of Climate Change on Human Health framework, which lists eight domains of climate change impacts that will affect human health (extreme heat, severe weather, air pollution, changes in vector ecology, increasing allergens, water quality impacts, water and food supply impacts, and environmental degradation) [1]. This review aims to identify gaps in the current peer reviewed literature related to climate change and health to identify areas for future research to better support the health and well-being of rural communities.
2. Methods
2.1. Protocol and registration
This review was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement for systematic reviews [17]. The review protocol, including design details, analysis and inclusion and exclusion criteria were preregistered with the International Prospective Register of Systematic Reviews (PROSPERO) 25 July 2020, (registration number CRD42020194172) [18].
2.2. Research questions
The research questions that guided this review and our discussion are as follows: (a) what interventions have been developed to decrease the impact of climate change on rural health and to improve these health outcomes (either at a population level or individual level) within the States? (b) Are these interventions effective at reducing negative health outcomes related to climate change? The original second research question guiding this review, as published in PROSPERO, was 'are these interventions feasible, effective, and scalable?' However, the scope of the review was not narrow enough to allow for a thorough discussion of this question as proposed. Thus, the second question was limited to focusing on the interventions' effectiveness.
2.3. Study eligibility criteria
Studies meeting the following criteria were considered for inclusion: interventional methodology; human participants in rural areas, including vulnerable, marginalized or medically underserved populations, such as farmworkers, older adults, and those who are incarcerated; and published in the English or Spanish language between 1 January 1988, and 31 May 2021. The starting date for the search corresponds with the year 1988 when the Intergovernmental Panel on Climate Change (IPCC) was created by the World Meteorological Organization and the Nations Environment Programme. Studies were excluded if they were conducted outside the States, had no intervention, were animal studies, or were related to heat injury or cooling interventions in athletes or sports.
2.4. Search strategy and information sources
A comprehensive literature search was undertaken to identify relevant published studies meeting the inclusion and exclusion criteria. The search strategies were developed and conducted by an experienced medical librarian (SL) with input from the research team. The CDC Impact of Climate Change on Human Health framework was used as the basis for design of the search strategy and the visual representation of the framework is presented in figure 1 [1]. The CDC, through its Climate and Health Program, created this framework that delineates the health disruptions that arise from climate change events. Because the CDC is one of the leading US governmental agencies that support programs to strengthen climate change adaptation programs at the local, state, and federal level, a comprehensive framework such as this is a useful tool for organizing the search terms for this systematic review. The CDC framework is divided into four sections identified as major disruptions from climate change: increasing CO2 levels, rising sea levels, extreme weather, and rising temperatures. The framework further divides these disruptions into eight specific categories: extreme heat, severe weather, air pollution, changes in vector ecology, increasing allergens, water quality impacts, water and food supply impacts, and environmental degradation. Within the framework, each of the domains highlights potential health outcomes of interest. Extreme heat is associated with heat related illness and death and cardiovascular failure; severe weather is associated with injuries, fatalities, and mental health impacts; air pollution is associated with asthma and cardiovascular disease; changes in vector ecology is associated with multiple disease causing pathogens; increasing allergens is associated with the health impacts of respiratory allergies and asthma; water quality impacts is associated with waterborne illnesses and the negative effects of algal blooms; water and food supply impacts is associated with malnutrition and diarrheal diseases; and environmental degradation is associated with the negative mental health impacts stemming from forced migration and civil conflict.
Figure 1. Impact of climate change on human health framework. The Centers for Disease Control (CDC) created this framework that delineates the health disruptions that arise from climate change events. Centers of Disease Control and Prevention. Climate Effects on Health. Published 3 March 2021. Accessed 6 February 2023. www.cdc.gov/climateandhealth/effects/default.htm Reproduced from CDC. CC BY 3.0.
Download figure:
Standard image High-resolution imageThe database searches used controlled vocabulary supplemented with keywords related to the concepts of climate change (e.g. greenhouse effect; global warming); rural residence (e.g. NOT urban; medically underserved; farmworker); and intervention (e.g. prevention; protective factors), which were then paired with terms representing each of the four main CDC framework sections (increasing CO2 levels; rising sea levels; extreme weather; and rising temperatures). Pre-identified sentinel articles were also hand-searched for keywords relating to the study objectives. The draft search results were assessed by the team and terms were revised and again tested in PubMed.
The search terms were then translated for each literature database identified as appropriate for the study topic. A total of five bibliographic databases were searched: EBSCO Environment Complete, EBSCO GreenFILE, Embase.com, MEDLINE via PubMed, and Web of Science Classic Core Collection (Science Citation Index Expanded, Social Sciences Citation Index, Arts & Humanities Citation Index, Conference Proceedings Citation Index-Science, Conference Proceedings Citation Index-Social Sciences & Humanities, Book Citation Index—Science, Book Citation Index—Social Sciences & Humanities, Emerging Sources Citation Index, Current Chemical Reactions, Index Chemicus). The EBSCO Environment Complete and EBSCO GreenFILE searches were run simultaneously. All searches were initially undertaken 21 July 2020 and rerun for updates on 1 March 2021. The final screening of all articles was completed on 11 October 2021, with the extraction of data and writing of the manuscript taking place until submission to the journal on 15 August 2022. Full search strategies for each database can be found in the supplementary material.
2.5. Study selection
A total of 12 427 studies identified through the database searches were uploaded to EndNote X9 (Clarivate Analytics, PA, USA). Duplicates were manually removed leaving 8300 records. We manually removed an additional 253 records with the following locations in the title: Africa, Australia, Canada, Europe, China, Japan; but not if the title also contained USA, States, African American or African-American. Additionally, 124 Morbidity and Mortality Weekly Report review articles were removed, leaving 7923 records. These were uploaded to the systematic review software Covidence [19]. When the searches were rerun an additional 779 records were uploaded to Covidence.
In the combined 8702 records, 231 additional records were identified by Covidence as duplicates and automatically removed. The title and abstract of the remaining 8471 records were screened by seven independent reviewers (DS, EM, SL, SS, ER, LT, GL). Of these, 8174 were excluded for irrelevancy, leaving 297 eligible for full-text review. During the full-text review, four reviewers (DS, EM, PS, GL) independently evaluated each article. Disagreements between reviewers were resolved by discussion and consensus. An additional 248 articles were excluded that did not meet the inclusion criteria, leaving 49 articles that met all eligibility criteria for inclusion in this study. The review and selection processes for the studies are summarized in the diagram in figure 2.
Figure 2. Flow diagram of literature search.
Download figure:
Standard image High-resolution image
2.6. Quality evaluation
Articles included in this review underwent a quality assessment using the 2018 Mixed Methods Appraisal Tool (MMAT) [20] by seven reviewers (DS, EM, AA, PS, GL, RML, LT). The MMAT tool has been used in multiple systematic reviews, such as ours, that included unique quantitative, qualitative, and mixed-methods research articles [21–25]. The MMAT tool is commonly used to evaluate studies focusing on the acceptability, satisfaction with, or barriers to implementation of interventions in the health literature [26]. The tool can be used to appraise five different study designs (qualitative, randomized controlled, nonrandomized, quantitative descriptive, and mixed methods) and includes 25 criteria and two screening questions [20]. Each article was independently rated by two reviewers with disagreements between reviewers resolved by discussion and consensus. If the pair was unable to reach a consensus, a third reviewer (DS) served as the tiebreaker. The results of the MMAT tool were presented using asterisks that correspond to the total score out of five, with scores closer to five indicating higher quality studies.
2.7. Data extraction and synthesis
Data were extracted by five reviewers (DS, EM, GL, RML, LT) with each reviewer assigned between 8 and 14 articles to extract. Data points extracted from each included study were specific location with the U.S., sample size and demographics, the domain from the CDC framework, a brief description of the intervention/policy/simulation reported in the study, study design, health outcome(s), statistical analysis (if any), and confounder assessment. If a reviewer was unsure of which data points to extract from a study, the concern was brought to the larger review group with a final decision made by DS, EM, and LT. Given that this study utilized the CDC Impact of Climate Change on Human Health framework [1] to drive the search process, the interventions were descriptively synthesized based on the domain of the CDC framework to which the study belonged.
3. Results
Across 49 studies, 34 unique interventions were identified to potentially reduce the negative health effects of climate change on human health in the rural States (table 1). The majority of studies included in this review (n = 40) could be classified using the MMAT tool as quantitative (n = 29), qualitative (n = 6) or mixed-methods (n = 5). Three of the 29 quantitative studies were randomized control trials [27–29]. The majority of quantitative studies were either descriptive or quasi-experimental. The remaining nine articles included in this review could not be rated/classified using the MMAT tool based on the responses to the first two screening questions (i.e. response of 'no' or 'cannot tell' to the questions 'are there clear research questions?' and 'do the collected data allow to address the research questions?') [30–38]. These studies were primarily policy [35] or programmatic case reports [30, 31, 34, 36] and narrative reviews of current policies [32, 37], proposed policies [38], or proposed interventions [33]; however, they were included in the review because they described interventions and policies designed to decrease climate change's impact on rural health.
Table 1. Summary table of intervention studies to mitigate the impact of climate change on health in rural areas of the States.
| Quantitative studies | ||||||
|---|---|---|---|---|---|---|
| Author | Climate change's impact on health (domain of CDC wheel). | Name | Health outcomes impacted by intervention | Statistical significance of findings | Study quality based on MMAT tool, * to ***** with ***** indicating highest quality | |
| Study location | Study design | Confounder assessment/covariate assessment/sensitivity analysis | ||||
| Sample size and demographics (i.e. age, sex, ethnicity) | Intervention/policy/simulation | |||||
| Arcipowski et al [56] Rural Appalachia n = 30 environmental assessments were conducted n = 1020 community members were trained Estimated n = 4080 people served | Water quality impacts | Water, Sanitation, and Hygiene Education Program Quantitative descriptive Intervention | WASH-related knowledge | No statistical analysis conducted. Only reported the interventions developed and estimated people reached for the education intervention Confounders assessed: none | *** | |
| Bersamin et al [59] Southwest Alaska Native communities n = 76 middle and high school students | Water and food supply impacts | Neqa Elicarvigmun or the Fish-to-School Program Quantitative non-RCT Intervention | Improved overall diet quality and fish intake among participants in the intervention community after time two while diet quality declined among participants in the comparison community | Significant improvements in diet quality compared to the comparison community (Beta = 4.57; p < .05) Fish intake, measured using the stable nitrogen isotope ratio of hair, a validated biomarker increased significantly in the experimental community (Beta = 0.16; p < .05) Confounders assessed: age, sex, traditional way of life, and white way of life | **** | |
| Brewer et al [27] Rural Kentucky communities n = 18 white, female, middle/older-aged | Air pollution | Nutrition-Based Lesson Series (Body Balance Lesson Series) Quantitative RCT Intervention | Increased knowledge and awareness of the effects of environmental pollution on health and the protective role of dietary strategies. | Environmental pollution concern increased significantly (2.8 ± 1.2–4.7 ± 0.6, p < 0.001) when measured using a five-point scale. Confounders assessed: none. | **** | |
| Chicas et al [28] Agricultural regions in Florida n = 78 agriculture workers mean age of 42, 66% female | Extreme heat | Cooling clothing (i.e. vests and bandanas) Quantitative RCT Intervention | Core body temperature and heat-related illness symptoms | Workers who received a cooling bandana had lower odds of core body temperature exceeding 38.0 °C (OR = 0.7, 90% CI = [0.2, 3.1]). Workers who received cooling vests had higher odds of core body temperature exceeding 38.0 °C (OR = 1.8, 90% CI = [0.4, 7.9]). Workers who received both the cooling vest and bandana in combination showed no difference from the control group in core body temperature exceeding 38.0 °C (OR = 1.3, 90% CI = [0.3, 5.6]). Confounders Assessed: BMI, years worked in agriculture, work duration, heat index, and physical activity. | **** | |
| Connally et al [49] Three, rural health districts Connecticut n = 364 cases from 24 towns where Lyme disease is endemic | Changes vector ecology | Peridomestic Prevention Measures for Lyme Disease Quantitative non-RCT Intervention | Detectable cases of Lyme disease | Performing tick checks within 36 h after spending time in the yard (OR 0.55; CI 0.32, 0.94) and bathing within 2 h after spending time in the yard (OR 0.42; CI 0.23, 0.78), having a fence on the property (OR 0.54; CI 0.33, 0.90) were all protective against Lyme disease. Confounders assessed: occupation, participation in recreational activities, pet or livestock ownership, prophylactic antibiotic therapy, gender, history or having contracted Lyme disease | ***** | |
| Cushing et al [64] California n = 23 190 neighborhoods | Air pollution | Cap-and-Trade Quantitative non-RCT Policy | No direct outcome measured. However, look at reductions in emissions of greenhouse gasses and co-pollutant emissions. | Simple descriptive statistics showed that neighborhoods <2.5 miles away from facilities regulated under the cap-and-trade program had a 34% higher proportion of residents of color, 59% more population density, 23% higher proportion of poor residents, 64% higher proportions of residents with low educational attainment, and 80% higher proportion of linguistically isolated households. Facilities regulated by the program also had 52% higher annual average local greenhouse gas emissions after implementation of the cap-and-trade program. Emissions of greenhouse gasses and co-pollutants (PM2.5, NOx, SOx, and volatile organic compounds) all had positive correlations (p < 0.001). Confounders assessed: none | **** | |
| Drexler et al [48] Arizona n = 576 households (2012) n = 558 households (2013) Approximately 1000 dogs | Changes in vector ecology | Rocky Mountain Spotted Fever (RMSF) Quantitative non-RCT Intervention | Detectable cases of Rocky Mountain Spotted Fever (RMSF) in Rodeo community as well as the number of ticks in the community. | 99% of dogs in Rodeo community were tick-free in 2012, compared to 36% of dogs in non-Rodeo community. Only 2% of people in Rodeo community reported seeing ticks in their homes, where 20% of people in non-Rodeo community reported ticks in their homes. Sub-analysis of homes with at least one dog found that factors associated with tick infestations included lack of tick collars (RR = 5.4, p < 0.05), and having more than two dogs (p < 0.05). Prior to the pilot project, annual incidence of human RMSF cases was 1.2 per 1000 people in both Rodeo and non-Rodeo communities; in the subsequent two years, average incidence of RMSF fell to 0.71 per 1000 persons in Rodeo community and 0.9 per 1000 in non-Rodeo community. Confounders assessed: none | *** | |
| Fernandes et al [41] Montana n = 428 children | Air pollution | Home Visiting Program for Children with Asthma Quantitative non-RCT Intervention | Asthma related activity limitations, number of days with asthma symptoms in the last month, unscheduled emergency department or office visits due to asthma in the last sixth months, use of short acting beta-agonist (SABA) medication every day for the last 30 d. | Those with Medicaid were more likely to have a reduced number of days with asthma symptoms than those with private insurance (adjusted OR 2.56. 95% CI 1.04–6.30). Those with Medicaid were more likely to have decreased number of SABA use days in the last 30 d than those with private insurance (adjusted OR 2.35, 95% CI 1.07–5.15). Those aged 4–11 years old were more likely to have reduced activity limitation than those 0–3 or 12–17 years old (adjusted OR 2.41, 95% CI 1.06–5.47). Participants added 12–17 years were more likely to have fewer parental missed workdays than those 0–3 or 4–11 years old (adjusted OR 0.11, 95% CI 0.02–0.61). Confounders Assessed: Age, sex, race, and insurance type. | ***** | |
| Fox et al [50] Kansas n = 534 residents | Changes in vector ecology | Educational Materials for West Nile Virus Quantitative descriptive Intervention | Health behaviors measured: use of insect repellent, wearing long pants and sleeves when outside near dawn or dusk, not allowing water to be standing, and inspection and repair of window screens. | More likely to action if heard about the West Nile Virus if information from newspaper compared to those who did not get information from newspaper (OR 1.808, 95% CI 1.115–2.931), more likely if information was seen on internet compared to those who did not get information from internet (OR 3.152, 95% CI 1.434–6.926), and more likely to take action through word of mouth information source (OR 2.167, 95% CI 1.358–3.460). Confounders Assessed: Age, education, residence, race, gender, and information source. | **** | |
| Godfred-Cato et al [51] States Virgin Islands (USVI) n = 148 families with children born to mothers exposed to Zika virus during pregnancy. n = 290 infants have been born to mothers with confirmed or probable Zika. | Changes in vector ecology | Pediatric Health Brigade Quantitative Descriptive Intervention | Concluded that the health brigade model is a useful tool for any disaster-affected region or underserved areas with limited access to adequate healthcare. Indirect health outcomes included in this study were improved care and access to care for patients and children exposed to Zika virus and a better understanding of Zika's impact on the community. | No statistical analysis conducted. Confounders assessed: none | **** | |
| Green et al [55] San Joaquin Valley, California No true sample | Water quality impacts | Water filtration system for agricultural drainage Quantitative descriptive Intervention | No direct health outcome measure; however, indirect through the removal of water contaminants. | Algal bacterial selenium removal facility at Panoche removed 95% of the influent nitrogen load and 80% of the influent selenium load. Recent results from the algal bacterial selenium removal Facility including dissolved air flotation and slow sand filter units indicated that removals of 90% or greater are possible. Average total soluble selenium mass removal of >76% was achieved. Confounders assessed: none | ***** | |
| Greenberg et al [73] New Jersey No true sample | Water quality impacts | Planning Support System Quantitative Descriptive Model | Multiple acute and chronic illnesses through drinking water (waterborne outbreaks, carcinogenic, toxic metal, and hormonal disrupter burden), contact recreation, and fish consumption | No inferential statistical analysis conducted. Confounders assessed: none | * | |
| Hansel et al [60] Rural Louisiana, US n = 157 school-aged children, average age 14, 20% reporting trauma from natural disasters. | Environmental degradation | Crisis Counseling Quantitative non-RCT Intervention | Multiple trauma related symptoms (i.e. intrusion, avoidance, arousal, etc) and diagnosis of post-traumatic stress disorder, depression, and anxiety. | Checklist scores revealed that students' follow-up scores were statistically significantly lower on anxiety, depression, and posttraumatic stress compared to baseline scores representing a disaster impacted state (p < 0.01). Students also had less symptoms of intrusion, avoidance, arousal, and under response (p < 0.01). Confounders assessed: age, gender, length of treatment, and degree of trauma exposure. | **** | |
| Horner et al [29] Unspecified rural areas located 20–45 miles from a densely populated urban center in Texas n = 183 elementary school children, mean age 9, 46% Hispanic | Increasing allergens | Asthma Education Classes Quantitative RCT Intervention | Changes in rural children's asthma self-management improvements in children's asthma knowledge, asthma self-management, self-efficacy for managing asthma symptoms, and metered dose inhaler technique, and significant group interaction effects for the treatment intervention on the measures of children's asthma knowledge, asthma self-management, and metered dose inhaler techniques | The treatment group showed the following gains relative to the attention-control groups: for asthma knowledge [−9.325 (SE = 1.392) v −4.853 (SE = 1.486)] yielding −4.47% greater mean change for the treatment group; self-efficacy [−.380 (SE = .109), v −.139 (SE = .124)] yielding −.24% greater mean change on the five-point scale for the treatment group; asthma management [−.208 (SE =. 064) v .011 (SE = .069)] yielding .22% greater mean change on the six-point scale for the treatment group; and MDI skill [−1.758 (SE = .184) v −.530 (SE = .212)] yielding 1.23% greater mean change on the nine-point scale for the treatment group. Confounders assessed: none | ** | |
| Krol et al [58] Southern Mississippi n = 1205 patients from n = 23 mobile medical clinics after hurricane Katrina, 62% African American. | Severe weather | Mobile Medical Clinics Quantitative descriptive Intervention | Multiple health outcomes were influenced by the use of mobile medical clinics after hurricane Katrina. Common reasons for seeking care were chronic medical conditions (e.g. hypertension, diabetes, and asthma), vaccine administration, common respiratory illnesses, skin conditions, and minor injuries. | Simple descriptive statistics showed that vaccine administration was the most common reason for seeking care (53.7%) with tetanus being the most requested vaccine. The top five diagnosis categories were respiratory disorders (27.8%), circulatory disorders (27.8%), minor injury (19.2%), skin conditions (18.8%), and endocrine disorders (10.9). Confounders Assessed: None | ***** | |
| Lin et al [65] New York n = total residents of New York State from 1997 to 2006 | Air pollution | Cap-and-trade Quantitative non-RCT Policy | Number of hospital admissions for respiratory diagnoses and concentrations of atmospheric ozone (a known respiratory irritant) | Statistically significant reduction in statewide ozone concentration in parts per billion (−2.47, α = 0.05). Statistically significant reduction in respiratory-related hospital admissions (measured in percent change) in certain regions: Central (−10.18, α = 0.05), Lower Hudson (−11.05, α = 0.05), NYC Metro (−5.71, α = 0.05). Confounders assessed: geographic region, race/ethnicity, age, disease subgroups, insurance coverage, and urbanicity. | **** | |
| McCabe et al [61] Maryland, Illinois, and Iowa n = 391 participants | Environmental degradation and severe weather | Crisis Counseling Quantitative non-RCT Intervention | Mental health preparedness, community resilience | Statistically significant improvement in knowledge, skills, and attitudes as measured by pre- and post-test after undergoing training (p < 0.001)—for both psychological first aid and guided preparedness planning trainings. Confounders assessed: none | **** | |
| Miller et al [46] Rhode Island n = 2518 ha areas in southern Rhode Island, blacklegged tick densities and Lyme disease risk | Changes in vector ecology | Deer-Targeted Acaricide Applicator Quantitative Non-RCT. Intervention | Density of blacklegged ticks at larvae, nymph and adult life cycles in ticks min−1 (calculated using modified Abbott's formula) | No statistical analysis conducted. Only reported raw change in tick population density over the period of the study. Confounders assessed: none | **** | |
| Muller and Yin [72] Colorado, US n = 2 assessed dimensions (e.g. wildfire risk accumulation and land planning use assessment) | Air pollution and severe weather | Planning Support System Quantitative descriptive Simulation | No direct health outcomes, but useful in planning for wildfires and their multiple recognized health impacts. | Descriptive statistics showed that the majority (e.g. >80% of development sites) in the cities of Boulder, Douglas, Jefferson, El Paso, Larmier, and El Pueblo, Colorado, were at either a medium-low or medium-high annual risk for wildfires with the risk model showing that following planning and hazard zone scenarios, 100% of regions would be at medium and high risk for wildfires. Confounders assessed: none | ***** | |
| Noonan et al [42] Libby, Montana n = 1147 households with 920 children in grades 1–12, between the years of 2003–2009 | Air pollution | Wood Stove Modernization Quantitative non-RCT Intervention | Parent-report of child respiratory symptoms, including wheeze | 26.7% (95% CI = 3.0–44.6) lower odds of reported wheeze per 5 µg m−3 decrease in winter PM2.5. 8.9% lower illness-related school absence (95% CI = 4.0–13.6). No difference in children's respiratory illnesses between homes using wood stoves versus other types of heating. Confounders assessed: school grade, community levels of influenza, presence of wood stove in home | **** | |
| Orengo-Aguayo et al [62] South Carolina n = 70 trauma exposed youth | Severe weather | Crisis Counseling Quantitative descriptive Intervention | Symptoms of PTSD | 88.6% of participants completed a full course of trauma informed cognitive behavioral therapy and after the intervention 96.8% no longer met the diagnostic criteria for PTSD. Effect size between the pre- and post-tests was large with both child-reported symptoms (d = 2.04) and caregiver reported symptoms (d = 1.50). Confounders assessed: none | **** | |
| Postma et al [43] Yakima Valley, Washington n = 866 children, mean age 6, 91% Hispanic, and 61% had caregivers that were migrant or seasonal farmworkers | Air pollution | Childhood Asthma Project Quantitative non-RCT Intervention | Primary program outcomes included number of emergency room admissions, number hospital admissions, and number of urgent care visits due to asthma exacerbations. Intermediate program outcomes included medication and device management, changing behaviors to reduce triggers specific to the child's bedroom, reducing indoor asthma triggers, cleaning throughout the house to reduce triggers, and reducing outdoor asthma triggers. | There was a decrease in the number of emergency room admissions (p < 0.0005), hospital admissions (p < 0.0005), and the number of urgent care visits (p < 0.0005) in participants. There were also improvements in medication and device management (p < 0.0005) and behaviors reducing asthma triggers specific to the child's bedroom (p < 0.0005), reducing indoor asthma triggers (p< 0.0005), and cleaning throughout the house to reduce triggers (p < 0.0005). However, there was no improvement in behaviors related to reducing outdoor asthma triggers (p = .332). Confounders assessed: clinic site and housing differences. | **** | |
| Rappold et al [39] Pocosin Lakes National Wildlife Refuge, North Carolina Simulated sample n = ∼2800 Asthma ED visits and ∼2300 CHF visits. | Air pollution | Smoke Forecast Based Intervention Quantitative non-RCT Simulation | ED visits for asthma and heart failure | There was a high correlation (ρ = 0.8) between smoke PM2.5 estimated using the smoke forecasting system reanalysis and 24 h ahead daily forecasts of smoke PM2.5 and a somewhat lower correlation with 48 h ahead daily forecast of smoke PM2.5 (ρ = 0.6). RR per 10 μg m−3 (95% confidence intervals) Asthma 1.049 (1.022 − 1.076) CHF 1.035 (1.004 − 1.066) Confounders assessed: seasonal variation | ***** | |
| Rosen et al [63] Regions affected by Hurricane Katrina 703 000 crisis counseling encounters 3–18 months after Hurricane Katrina. | Severe weather | Crisis Counseling Program Quantitative descriptive Intervention | Mental health; referrals to disaster relief and social services. | The strongest predictors were need (disaster-related losses), age (adults rather than children), urbanicity, and prior physical disability (potentially indicating greater need for tangible assistance). Individuals with preexisting mental illness were less likely to receive referrals for disaster relief, suggesting that counselors may have not fully responded to their instrumental rather than psychological needs. Over half (58.4%, n = 410 500) resulted in referrals to disaster relief and other social services. Months 3 and 18 post-disaster, 159 543 people (22.7%) were referred for additional crisis counseling. Referrals for psychological services were uncommon (6.6%, n = 46 500). ** factors influencing need for further referrals are included Confounders assessed: age, gender, urbanicity, and disaster declaration status. | ***** | |
| Running et al [57] Southeastern Idaho n = 265 farmers | Water quality impacts | 2015 Water Policy and Agreement of the Eastern Snake Plain Aquifer Quantitative descriptive Policy | No direct health outcome impacts were reported. However, food production is essential to human survival. | The policy led to reduced spending on water, installation of more efficient irrigation systems, and changing crop rotations. Farmers also reported losing 7.6% of their yield and 8.4% of their income during the first two years of the policy. Confounders assessed: none | **** | |
| Savage and Ribaudo [53] Chesapeake Bay Watershed n = 6 states located within the watershed n = 18 states located outside of the watershed | Water quality impacts | Watershed Nutrient Management Quantitative non-RCT Intervention | No formal measure of health impacts. However, indirect impacts of clean water on health are apparent. | Animal operations in the Bay states are 13.1% more likely to have implemented a comprehensive nutrient management plan on average than animal farms operating outside the Bay. Operations in the Bay remove 7.5% more manure off-farm than operations outside the Bay. Confounders assessed: none | *** | |
| Ward et al [45] Rocky Mountain Valley n = 16 homes | Air pollution | Wood Stove Modernization Quantitative descriptive Intervention | No direct health outcome impacts measured—however, the study examined the effect of this intervention on indoor and ambient levels of fine particulate matter, which can cause respiratory illness. | An observed 76% reduction in average concentration of fine particulate matter (p = 0.0001) in homes after wood stoves were replaced. An observed 133% increase in average concentration of resin acids (p = 0.0001) after wood stoves were replaced. Confounders assessed: none | **** | |
| Whetten et al [40] New Mexico n = 1 academic center n = 12 rural hospitals | Air pollution | Telemedicine Quantitative descriptive Intervention | No direct health outcomes measured. However, air pollution is associated with multiple negative health outcomes. | The total number of travel miles avoided by implementing the telemedicine program was 477 932. This equated to avoiding 618.77 metric tons CO2 emissions. Confounders assessed: none | ***** | |
| Woodward et al [66] Ohio n = 771 farmers | Water quality impacts | Cap-and-Trade Quantitative Descriptive Intervention | No direct health outcomes measured. But water quality impacts health outcomes such as incidence and prevalence of communicable diseases, etc. | No statistical evidence that participation in Great Miami Water Quality Trading Program affected the number of conservation practices utilized by surveyed farmers. Confounders assessed: none | **** | |
| Qualitative studies (as rated by the MMAT tool) | ||||||
|---|---|---|---|---|---|---|
| Author | Climate change's impact on health (domain of CDC wheel). | Name | Health outcomes impacted by intervention | Important results | Study quality based on MMAT tool, * to ***** with ***** indicating highest quality | |
| Study location | Intervention/policy/simulation | |||||
| Sample size and demographics (i.e. age, sex, ethnicity) | ||||||
| Arnold et al [68] Marquette County, Michigan n = 31 survey participants | Severe weather, changes in vector ecology, and air pollution | Deliberation with Analysis Intervention | No direct outcome measured. However, multiple health outcomes potentially impact by increasing severe weather, changes in vector ecology (e.g. Lyme Disease), and air pollution from wildfires. | The deliberation with analysis model of public participation was used effectively to overcome barriers to climate change adaptation and mitigation related to lack of public awareness of climate change's threat to human health with relation to increased flooding, increased tick habitats, and the impact of wildfires. | ***** | |
| Enloe et al [52] Boone River Watershed n = 31 interviews with 33 stakeholders | Water quality impacts | Watershed Nutrient Management Intervention | Indirect through reducing nutrient load in the Boone River watershed. | Multi-stakeholder collaboration enabled partners to overcome traditional barriers for watershed management. However, scale mismatches due to socio-economic and ecological forces pose significant obstacles for the program's resilience. Issues with public funding often cause issues with monitoring of water quality and farmer engagement. Monitoring and farmer engagement. | ***** | |
| Laursen et al [70] Hawaii n = 29 managed natural resource managers | Environmental degradation and severe weather | Participatory Action Research Intervention | Indirect through creating a healthy environment to sustain life. | Resource managers' primary source of information were other professionals. Managers desire to widen their networks to build upon their management capacities during the climate change era. Co-development of knowledge between scientists and local managers strengthens collaborations. | ***** | |
| Matlock et al [47] Central California n = 8 representatives of public health agencies in five California counties | Changes in vector ecology | Health Communication Intervention | Seasonal Valley Fever prevalence | Climate change's impact on seasonal Valley Fever prevalence is not currently included in any of the eight agencies included in the study. Politics were commonly cited as the reason behind not including climate change in health communication. | ***** | |
| Miller-Hesed et al [69] Chesapeake Bay, Maryland n = 111, from local faith-based organizations, local government employees/policy makers, and employees of local environmental non-profits. | Environmental degradation | Participatory Action Research Intervention | Mental health impacts | Collaborative learning led to greater understanding of capacities and limitations in addressing environmental challenges, increased trust and social networks, expanded engagement with a greater diversity of stakeholders, increased opportunities for new conversations, new pathways toward interventions, and stakeholder empowerment. | ***** | |
| Preece et al [54] Anacostia Watershed n = 10 cohorts of 76 adults who had started 72 community-driven educational projects. | Water quality impacts | NatureNet technology platform to support the community-driven environmental projects Intervention | Indirect through ensuring healthy water management to support clean water. | Community-driven environmental projects should have six pillars for success: (a) common endeavor that relates to the identity of community members, (b) facilitation of collaborations between new and existing community members, (c) supporting a broad ranges of knowledge sources, (d) multiple engagement methods so participants can express themselves, (e) different methods for expressing status of projects, and (f) support for changing leadership roles. | ** | |
| Mixed-methods studies (as rated by the MMAT tool) | ||||||
|---|---|---|---|---|---|---|
| Author | Climate change's impact on health (domain of CDC wheel). | Name | Health outcomes impacted by intervention | Quantitative findings | Study quality based on MMAT tool, * to ***** with ***** indicating highest quality | |
| Study location | Intervention/policy/simulation | Qualitative findings | ||||
| Sample size and demographics (i.e. age, sex, ethnicity) | ||||||
| Jurjonas et al [74] Albemarle Pamlico Peninsula, North Carolina n = 37 focus groups comprised of farmers, fishers, loggers, local business owners, tourism guides, and NGO employees | Environmental degradation and severe weather | Rural Coastal Community Resilience (RCCR) framework Intervention | Boost rural coastal community resilience by focusing on locally perceived resilience needs as targets for capacity building workshops, management interventions, and climate action planning Reduce/mitigate negative effects on wildlife habitats, agriculture (agricultural abandonment), and land and development (reduction in suitable land and development options) | Quantitative findings: Participants had increased agreement that sea level rise and saltwater intrusion are threatening to communities, p = 0.13 and p = 0.27 respectively. There were no significant changes in perception of community threat to sea level rise, flooding, and saltwater intrusion. The intervention also increased participant disagreement with the statement that their community has access to the resources needed to plan for climate change (no p-value provided; raw change from −0.17 to −0.75 signifying increased disagreement) Confounders assessed: none Qualitative findings: Commonly expressed direct experience with sea level rise, college education and physically leaving the peninsula was seen as a way to secure the future, and inequity and unsustainable development acknowledged by participants. | ***** | |
| Roesch-McNally et al [67] Corn Belt States (Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, Nebraska, Ohio, South Dakota, and Wisconsin) Quantitative data: n = 4778 farmers Qualitative data: n = 159 farmers | Environmental degradation and water quality impacts | Crop Rotations Intervention | Adaptation to cropping system diversity will decrease the need for herbicides, thus decreasing the direct exposure to humans and through (water pollution). | Quantitative findings 46% of farmers utilized diversified crop rotations with 205 intending to increase their use of crop rotations. Farmers with cattle (standardized logit coefficient 0.34, p < 0.001), those who farmed increased hectares of highly erodible land (0.07, p < 0.01), farmers with positive attitudes toward crop rotations (0.36, p < 0.01), and farmers in agreement that profitable markets for small grains should be developed as a climate adaptation tool (0.43, p < 0.001) were more likely to use crop rotations. As farm revenue increased, there was a decrease in the use of crop rotations (−0.24, p < 0.01) Confounders assessed: farmer productivity, environmental stewardship, cattle farming, crop insurance, corn markets, water concerns, farming of highly erodible land, farmer agreement to adaptation, education level, and farm revenue. Quantitative findings: Historically, crop rotation used to be more common due to the diversified crops produced in the region. However, 'corn is king' and is now the predominant crop produced. Farmers also discussed the lack of markets for crops other than corn as being a driver for solely producing corn in the region. They also recognized the importance of diverse crop rotations for achieving broader conservation goals. Farmers recognized the importance of crop rotations as a viable mechanism of responding to a changing climate. The loss of livestock and transition to only 'crop farming' was recognized as a reason for not rotating crops. Farmers also identified the high cost of rented farmland and the need for ensuring profit, which limits the rotation of less profitable crops. | **** | |
| Thomas et al [75] Indian Health Service Great Plains Area n = 30 semi-structured interviews across three different communities. n = 22 community members participated in a workshop on arsenic in drinking water. | Water quality impacts | Participatory Action ResearchIntervention | Demonstrated positive acceptance of health protective behaviors. No direct health impacts measured. | Quantitative findings: During the workshops, the issues ranked the highest with regards to water issues were communication of water test results (n = 36*), need for education and information (n = 34), and contamination from herbicides/pesticides (n = 30). The top three channels of communication for intervention delivery were utility bills (n = 22), mailings (n = 17), and public events/meetings (n = 16). Confounders assessed: none *Description of workshop reported n = 22 participants. However, n = 36 reported as number of votes tallied. Qualitative findings: Water taste, temperature, smell, and appearance were important preferences for drinking and cooking water (at the individual level) with the sacredness of water being identified (at the population/tribal level). Concerns around water safety (i.e. arsenic) were identified at the individual level with private well users expressing varying levels of familiarity with arsenic in well water. There was also a recognition of water on health outcomes at the individual level and there was a desire for community level awareness of 'what is in the water'. The Environmental Protection Agency and tribal organizations were identified as preferred sources of information about water issues. | **** | |
| Tovar-Aguilar et al [71] Florida n = 301 male citrus farmers | Extreme heat and severe weather | Participatory Action Research Intervention | Eye safety (personal protective equipment, PPE) | Quantitative findings: Work crews with community health workers had significantly higher use of PPE than those without community health workers (t = − 3.070, p = 0.012). Baseline use of PPE for intervention crews was 11.2% with a high of 35.5% rate of PPE use on the last day of observation after the intervention. Confounders assessed: none Qualitative findings: Workers have high risk perceptions and frequently refuse to use PPE. Workers recognized that PPE could protect their eyes but had negative attitudes due to the cost of obtaining safety glasses. Even when provided safety glasses, participants did not use PPE due to fears that the glasses would decrease their efficiency. | **** | |
| Vogt et al [44] Salinas, CA n = 4925 adults, from 2004 to 2006 | Air pollution | Salinas Asthma Project Intervention | Prevalence of asthma | Quantitative findings: Two strongest modifiable risk factors associated with asthma were: (a) inability to afford prescription medications in last 12 months (p < 0.001) and (b) morbid obesity (p < 0.001) Confounders assessed: none Qualitative findings: Never present the results of the semi-structured interviews. | * | |
| Model and program descriptions (not classified as either quantitative or qualitative research). | ||||||
|---|---|---|---|---|---|---|
| Author and study location | Sample size and demographics (i.e. age, sex, ethnicity) | Climate change's impact on health (domain of CDC wheel). | Name Intervention/policy/simulation | Health outcomes impacted by intervention | Important results | Study quality Studies that are purely descriptive of models and policies are not able to be assessed using the MMAT tool and this column is purposefully left blank. |
| Connor et al [30] Rural, southern Georgia n = 1 community-based program that serves over 1000 migrant farm workers and their families annually. | Severe weather, extreme heat, and changes in vector ecology | Farmworker Family Health Program Intervention | Multiple indirect health outcomes through increasing the delivery of health care services. | When replicating this model, persons should consider the importance of developing community and academic partnerships which are based on mutual trust and connections among the partners. Similar models must also ensure that they are tailored to meet the needs of the community being served. | ||
| Deal et al [31] Coastal Texas n = 21 nursing school faculty, 100% female. n = 57 students, 81% female | Severe weather | Sheltering for Medical Special Needs Populations Intervention | Multiple indirect health outcomes through maintaining the delivery of healthcare services after Hurricane Gustav | Nursing schools can and should respond to the needs of medical complex individuals during natural disasters. To do so, schools can rely on the unique skills of nurse faculty and students to meet community needs while also serving the interests of faculty and students involved. | ||
| Ganesh and Smith [32] California n = 4 policy principles that are critical to meaningful climate change policy. | Severe weather | San Joaquin Valley Policy Evaluation Policy | Evaluate and improve health policies focusing on climate change to improve the health of communities, with case study example of California CC policy and improving mental health | The article provides examples of the application of the four key policy principles in San Joaquin valley, aka the "nations salad bowl", when considering mental health of the larger determinants of health. | ||
| Gibson et al [33] Examines legal regulatory systems in the States and approaches to preparedness for hurricanes and severe weather. | Severe weather | Participatory Action Research Intervention | No direct measurement; however, multiple indirect health outcomes impacted by hurricanes and severe weather. | Proposed increasing heritage professional's preparedness for hurricanes and severe weather events by utilizing public engagement mechanisms to position the community as a partner in the disaster response. | ||
| Lee et al [34] Southeast Alabama n = 100 rising fifth- and six-grade students, African-Americans, and their teachers from three elementary schools. | Water quality impacts | Water, Sanitation, and Hygiene Education Program Intervention | No direct measurement; however, multiple indirect health outcomes through contaminated ground water. | Water education is best delivered when students are active participants in lessons. Students have high motivation to learn independent of the setting of the learning environment. Students were also concerned about water quality and this program was a beneficial way to introduce students to the scientific method surrounding questions of water quality. | ||
| Riley et al [35] California n = 1 Labor and Occupational Safety and Health Program | Extreme heat | UCLA Labor OccupationalSafety and Health Program Intervention | No direct measure of health impacts. Program description targeting heat-related illness. | Promotores used in a statewide heat awareness campaign expanded the state program to multiple heat-exposed workers. The article also highlights that proper enforcement of the heat standard is needed in order for the policy to be successful. | ||
| Shendell et al [36] San Joaquin Valley agricultural region California n = 402 public schools, n = 22 private schools, and n = 22 other locations (e.g. hospitals, clinics, health departments, and daycare centers). | Air pollution | Air Quality Flag Intervention | No direct measure of health impacts. Program description only. | Educational interventions that are simple, low-cost, and science based can help prevent exposure to environmental hazards in socioeconomically and linguistically diverse agricultural communities. Local environmental justice groups can also successfully lead school-based environmental educational interventions to improve children's quality of life related to environmental hazards. | ||
| Shortle et al [37] Agricultural communities Multiple policies | Water quality impacts | Watershed Nutrient Management Policy | No direct measurement; however, indirect through exposure to pesticide and herbicides in contaminated water. | Polluter pay approaches that have been used in tackling issues of point source pollution would not be feasible for agricultural run-off (which is inherently non-point source). Improving the connections between farmers and water quality programs will be paramount to the success of any legislation aimed at tackling the issue of nonpoint agricultural pollution. | ||
| Sweeney et al [38] Waterways Proposal of utilizing incentives to increase watershed stewardship. | Water quality impacts | Watershed Nutrient Management Policy | No direct measurement; however, indirect through exposure to pesticide and herbicides in contaminated water. | Incentivization for best management practices of watersheds should accompany education and legislation that help to promote and adopt best management practices of watersheds. | ||
3.1. Summary of interventions by domain
Using the CDC framework domains, eight studies focused on air pollution, six on changes in vector ecology, five on water quality impacts, three on severe weather, two on extreme heat, one on increasing allergens, and one on water and food supply impacts. No interventions focused solely on environmental degradation as these interventions were used across multiple domains. Eight interventions intended to protect human health across multiple domains (table 1). Because unique interventions were used across multiple domains in several studies, the total number of interventions identified is greater than the total number of studies identified.
3.1.1. Air pollution
Eight unique interventions were identified to address air pollution risk in the rural U.S. (table 2). Three of these interventions targeted populations (e.g. schools, communities, counties) [36, 39, 40] and five interventions targeted individuals [27, 41–45]. For example, one of the population-based interventions focused on the use of flags to alert populations of air quality risks to minimize exposure to air pollution and supported broader use of low-cost, educational interventions to alert persons to environmental risk [36]. This study [36] along with the findings of Rappold et al [39] support the broader use of low-cost, educational interventions to alert persons to environmental risk stemming from air pollution. At the individual level, two studies replaced old wood burning stoves with newer ones to minimize the pollution from stoves used for heating inside the home [42, 45]. These studies found that the use of modernized wood stoves decreased the concentration of fine particulate matter in homes [45] and also decreased odds of wheezing and illness-related school absences [42].
Table 2. Description of interventions grouped by domain.
| Intervention | Description of intervention | Level at which the intervention was deployed (i.e. population vs. individual) | Reviewed studies that utilized this intervention |
|---|---|---|---|
| Air pollution | |||
| Air Quality Flag | Used flags to represent the colors of the U.S. AQI system to alert school communities of ambient air quality risk for children with asthma. This program evaluation described a community-wide program to influence policy related to ambient air quality and educate about air quality and potential health risks. | Population | Shendell et al [36] |
| Childhood Asthma Project | Community health workers trained in environmental assessment conducted a home-visitation program in which they assess homes of migrant children and make recommendations to families on measures that can reduce the children's exposures to environmental triggers that can potentially exacerbate asthma symptoms. A total of eight home visits occurred approximately one month apart from each other. Families were provided education modules that are designed to increase awareness of asthma triggers, asthma care plans, and medication management. | Individual | Postma et al [43] |
| Montana Home Visiting Program for children with asthma | A home visiting program was designed to address residential asthma triggers, provide participants with asthma self-management education, and coordinate asthma care in school, childcare, and the clinical setting. Participants received six contacts with a registered nurse to increase asthma control. | Individual | Fernandes et al [41] |
| Nutrition Based Lesson Series (Body Balance Lesson Series) | A nine-lesson nutrition-based lesson series, 'Body Balance: Protect Your Body from Pollution with a Healthy Lifestyle' was developed based on a needs-based assessment to enhance the understanding of environmental health and protective actions. | Individual | Brewer et al [27] |
| Salinas Asthma Intervention | A multi-site asthma intervention to increase asthma education and treatment/control activities among providers/teachers, the community, and patients/individuals. The intervention focused on training healthcare providers and patients on how to best control asthma symptoms. | Individual | Vogt et al [44] |
| Smoke Forecast Based Intervention | Simulation of forecasted-based interventions implemented at county level to reduce personal exposure to populations susceptible to complications from smoke. Interventions took place when a 24 h-ahead or 48 h-ahead smoke forecast predicted PM2.5 concentration exceeding a threshold and the population assumed a level of adherence. | Population | Rappold et al [39] |
| Telemedicine | A telemedicine network was implemented to provide remote providers to rural hospitals from a central academic medical center. Such a network allows for remote assessment of patients and helps to reduce the number of patients transported from rural hospitals to urban centers for treatment. | Population | Whetten et al [40] |
| Wood Stove Modernization | Older wood stoves used for heating were replaced with new, less-polluting wood stoves in homes using wood stoves for heating. | Individual | Noonan et al [42], Ward et al [45] |
| Changes in vector ecology | |||
| Deer-Targeted Acaricide Applicator | Introduction of baited four-poster acaricide applicators to area of high deer/blacklegged tick population density. | Population | Miller et al [46] |
| Educational Materials for West Nile Virus | An education and awareness campaign was designed to increase community awareness about West Nile Virus and measures to protect oneself from the virus. Educational materials included news releases (written and video), newsletter articles, public service announcements (newspaper and radio), pre-written letters to editors, mailings to veterinarians, and brochures. | Population | Fox et al [50] |
| Health Communication | Central California public health agencies' risk communication on Valley Fever. | Population | Matlock et al [47] |
| Pediatric Health Brigade | Pediatric specialists were recruited to provide the recommended evaluations and screenings for children born to mothers with confirmed or probable Zika virus during pregnancy in the US Virgin Islands. | Individual | Godfred-Cato et al [51] |
| Rocky Mountain Spotted Fever Rodeo | A program that was enacted to decrease the prevalence of brown dog ticks. The program focused on the systematic registration of homes and dogs, application of dog collars containing insecticide, lawn spraying of insecticide, dog population control via spay-neuter opportunities. | Population | Drexler et al [48] |
| Peridomestic Prevention Measures for Lyme Disease | Personal protective measures included landscape features/modifications, personal protection, and chemical control practices. | Population | Connally et al [49] |
| Environmental degradation | |||
| See Interventions that Spanned Multiple Domains, below. | |||
| Extreme heat | |||
| Cooling Clothing (i.e. vests and bandanas) | Cooling vests or cooling bandanas were provided to agricultural workers during the workday to decrease core body temperature. | Individual | Chicas et al [28] |
| UCLA Labor Occupational Safety and Health Program (LOSH) | Specific strategies included community-based outreach, popular education, and organizational capacity building. Integration of health promotores into core program planning and training activities and the expansion of campaign activities to a wide variety of rural and urban workers. | Individual | Riley et al [35] |
| Increasing allergens | |||
| Asthma education classes | Lay health educators (LHE) offered asthma education/general health promotion classes to elementary school students grades 2 through 5 with the aim of seeing improvements in children's asthma knowledge, asthma self-management, and self-efficacy for managing asthma symptoms. | Individual | Horner et al [29] |
| Severe weather | |||
| Improvements to Climate Change policy | Four key principles to evaluate health policies focused on climate change: mainstreaming, linked approach, population perspective, and coordination. | Population | Ganesh and Smith [32] |
| Mobile Medical Clinics | Mobile medical clinics were deployed to provide clinical services to areas impacted by severe weather extremes. | Individual | Krol et al [58] |
| Shelters for Medically Complex Persons During Hurricane | Emergency sheltering of medical special needs populations evacuated from the Texas coast during Hurricane Gustav. | Population | Deal et al [31] |
| Water and food supply impacts | |||
| Neqa Elicarvigmun or the Fish-to-School Program | A school-based, multilevel intervention that included activities in the cafeteria, classroom, and community was designed to improve diet quality and increase intake of traditional foods. It was evaluated using a pre—and post-comparison group design with data collection occurring at three time points: baseline, four months, and nine months. | Individual | Bersamin et al [59] |
| Water quality impacts | |||
| 2015 Water Policy and Agreement of the Eastern Snake Plain Aquifer | In 2015, Idaho overhauled its water policy and created a water use agreement for farmers located in the region of the Eastern Snake Plain Aquifer. The new legislation had the goal of reducing water consumption from the aquifer by 13% per year. | Population | Running et al [57] |
| Water, Sanitation, and Hygiene Education Programs | A team of professionals provided interactive, culturally and literacy appropriate WASH education to target community groups within the school, senior center, and larger community. Arcipowski et al further adapted WASH programs and also provided a clean water kiosk (2017). | Population | Arcipowski et al [56], Lee et al [34] |
| Watershed Nutrient Management | Water monitoring, watershed planning, and outreach efforts by private and public partners, including cost-share incentives for farmers, in order to manage nutrients and engage farmers in water quality management. | Population | Enloe et al [52], Sweeney et al [38], Savage and Ribaudo [53], Shortle et al [37] |
| Water filtration system for agricultural drainage. | ABSR Facility at Panoche removed 95% of the influent nitrogen load and 80% of the influent selenium load. Recent results from the ABSR Facility including DAF and slow sand filter units indicated that removals of 90% or greater are possible. Average total soluble selenium mass removal of >76% was achieved. | Population | Green et al [55] |
| NatureNet technology platform to support the CDEPs ('community-driven environmental projects') | NatureNet consisted of a website, an interactive display in a community center and for use in WSA classes and on computers at home, and iOS and Android mobile 'apps' that supported collection of data and community interaction for CDEPs. Options to 'Explore', go to 'Projects,' submit 'Design Ideas' to help tailor the software, go to the 'Communities' page, and submit 'Contributions'. Also includes likes and comments, maps, members can store data, notes, or project plans. | Population | Preece et al [54] |
| Interventions used across multiple domains | |||
| Cap-and-Trade | Cap-and-trade programs regulate various pollutants and serve as a mechanism for pollution pricing. Under such regulation, facilities are allowed to produce up to a set amount of emissions before they need to buy more 'allowances' from other facilities that have not met their emissions caps. | Population | Cushing et al [64], Lin et al [65], Woodward et al [66] |
| Crisis Counseling | Crisis counseling and psychological/psychiatric programs designed to attend to the mental health needs of individuals after a natural disaster. | Individual | Hansel et al [60], McCabe et al [61], Orengo-Aguayo [62], Rosen et al [63] |
| Crop Rotations | Farmer used diversified rotations that include small grains, forages, or other crops on land they own and/or rent (other than corn). | Population | Roesch-McNally et al [67] |
| Deliberation with Analysis | Deliberation with analysis model of public participation conducted during community meetings with the purpose of building community resilience and ability to adapt to climate change in Marquette County, Michigan. | Population | Arnold et al 2020 |
| Farm Worker Family Health Program (FWFHP) | Students and faculty members from departments of nursing, physical therapy, dental hygiene, and psychology worked as an interdisciplinary team for two weeks to deliver health care services to 500 migrant farm workers and 500 migrant children. | Population | Connor et al [30] |
| Participatory Action Research | Participatory action research methods aimed to empower community members and involve them in the systematic design of projects to impact the planning, response, and recovery from various climate related health disasters and health effects. | Population | Miller-Hesed et al [69], Gibson et al [33], Laursen et al [70], Thomas et al [75], Tovar-Aguilar et al [71] |
| Planning Support System | Models or information systems designed to help local governments and other organizations plan various programs to help communities adapt and mitigate the impacts of climate change on health. | Population | Muller and Yin [72], Greenberg et al [73] |
| Rural Coastal Community Resilience (RCCR) framework | Framework implemented into rural coastal communities to analyze their level of vulnerability or resilience to climate change effects. Administered questionnaires to participants and held discussions among participants regarding the following themes: livelihood dependency and livelihood diversity, poverty and prosperity, unsustainable development and sustainable development, community disengagement and community cohesion, and rigidity and agency. RCCR framework intended to boost resilience by stimulating dialogue between community members, resource managers, planners, and other stakeholders necessary to maintain their ways of life, subsequently promoting future implementations of intervention and climate change planning strategies. | Population | Jurjonas et al [74] |
3.1.2. Changes in vector ecology
Six distinct interventions addressed changes in vector ecology (table 2). Five interventions targeted populations (e.g. states, communities, counties) [46], [47–50], and one targeted individuals [51]. To illustrate, two population-based studies focused on decreasing tick prevalence in high-density rural areas [48, 49]. These studies provided evidence that targeting behaviors such as performing tick checks [49], bathing within 2 h of being outdoors, [30] and applying tick collars to family dogs [48] help to reduce the incidence of Lyme disease. The individual-based intervention provided screening and recommendations to pregnant women with confirmed or probable Zika virus infection, which found that their health brigade model of community engagement was successful and should be replicated to successful respond to future infectious disease outbreaks in resource limited settings [51].
3.1.3. Water quality impacts
Five unique population-based interventions were identified across multiple studies [34, 37, 38, 52–57] to improve water quality (table 2). The most common intervention in this group was the implementation of a watershed nutrient management intervention that aimed to regulate nutrients and engage farmers through water monitoring, watershed planning, and outreach efforts by private and public partners [37, 38, 52, 53]. Specifically, the work of Shortle et al [37] problematizes the use of 'the polluter pays' to decrease non-point source agricultural pollution. Their work and the work of Sweeney and Blaine [38] highlight the need for appropriate incentive programs and outreach to tackle non-point source pollution in agricultural communities [37, 38]. Another example of an intervention was a community-based intervention that developed the NatureNet technology platform consisting of a website, interactive display, and application to support data sharing and planning for community-driven environmental projects [54].
3.1.4. Severe weather
Three unique interventions were identified to evaluate and implement medical services in response to severe weather events (table 2). Two interventions targeted populations [31, 32] while the other intervention provided clinical services to individuals in areas affected by severe weather [58]. For example, one population-based intervention established an emergency shelter for special needs populations who evacuated the Texas coast during Hurricane Gustav [31]. While the primary focus was on evaluating educational outcomes, the authors concluded that interventions that mobilize the skills of academic, practice professions in meeting the needs of communities during disaster response should be supported [31].
3.1.5. Extreme heat
Two interventions were identified to address exposures to extreme heat (table 2). Each of these interventions were implemented at the individual level [28, 35]. One Florida-based intervention provided cooling vests and bandanas to agricultural workers during the day to decrease core body temperature, and found that bandanas alone decreased core body temperature compared to control group participants [28]. A Southern California-based labor and occupational safety program trained health promotores in planning and training activities to reduce heat-related illness but did not describe health outcomes related to the training [35].
3.1.6. Increasing allergens
Only one educational intervention was identified that provided education in response to increased allergens [29]. In this study, lay health educators offered asthma and general health education courses to students in second through fifth grade to promote their understanding of asthma. The intervention was successful in increasing knowledge of and self-management of asthma symptoms [29].
3.1.7. Water and food supply impacts
One intervention was identified to address food and water supply at the individual level (table 2). This school-based intervention held activities at various locations (e.g. cafeteria, classroom, community) that were designed to improve diet quality and to increase the intake of traditional foods among two rural, remote Indigenous communities located within Alaska [59]. This intervention successfully increased diet quality and fish intake in school-aged, Alaskan Native, children [59].
3.1.8. Multiple domains
Eight interventions were identified to address issues that encompassed multiple CDC domains (table 2). Four interventions established crisis counseling and psychiatric programs for mental health needs of individuals following a natural disaster [60–63]. Seven of these unique interventions encompassing multiple domains were conducted at the population level [30, 33, 64–75]. One example of an intervention spanning multiple domains is the Farmworker Family Health Program, [30] which serves migrant farmworkers in the Southeastern US in an attempt to decrease multiple negative health impacts related to extreme heat and changes in vector ecology. The program also highlights the importance of community collaborations when providing innovative clinical services and the need to meet the community where they stand [30]. In five studies, participatory action research was used to engage and mobilize community members on the planning, response, and recovery from various climate-related disasters in Hawaii, [70] Maryland, [69] and more broadly, across the U.S. [33] as well as to improve water quality on tribal lands, [75] and the use of personal protection equipment in citrus growers in Florida [71].
3.2. Quality appraisal
Using the MMAT, most articles included in this review were rated as high quality. The majority of quantitative studies received ratings of four or five stars, out of a possible five stars (table 1). Only five studies received ratings of three stars or less. Most of the qualitative studies received high ratings on the MMAT. Only one qualitative study did not receive five stars due to flaws in the data collection methods and analysis of the qualitative data [54]. Similarly, mixed-methods studies included in this review were generally rated as high quality (either four or five stars). Only one study received a rating of one out of five stars because the authors used a mixed-methods design but never presented the qualitative findings [44].
4. Discussion
This systematic review details the existing interventions that have been developed to decrease the impact of climate change on rural health outcomes in the United States. Across 49 articles, published from 2003 to 2021, 34 unique interventions were described. Overall, there is a lack of high-quality, randomized controlled trials that aim to evaluate the effect of interventions, and lack of interventions to address the health consequences of severe weather, heat stress, increasing allergens, and environmental degradation. Given the impending, severe consequences of climate change, we expected more research would have been published over the past decades since the IPCC report was published in 1988. In fact, across major databases we searched, the first intervention-based article we were able to retrieve was published in 2003. However, in this discussion, we suggest that despite the lack of published academic research, health interventions are likely being conducted at the grassroots level and not published in the peer-reviewed literature.
Our review only identified three randomized controlled trials, with two of these three published in the last five years [27, 28]. This is highly concerning given that randomized controlled trials are considered the gold standard for establishing causality and proof of concept. However, it would not be ethical to conduct randomized control trials if there is not equipoise, or a state of uncertainty about an intervention's efficacy, given the concerns about intentional withholding of potentially beneficial programs among those randomized to a control group [76]. Furthermore, many climate-related health-affecting events cannot be randomized (e.g. severe weather events). Randomized control trials are also resource intensive. However, with the WHO predicting less than ten years to act on climate [77], researchers must consider the urgency to provide solutions. Despite the lack of randomized controlled trials, original research on climate change's impact on health has increased 11-fold from 2007 to 2020 [78]. This is consistent with our review: despite searching for articles published as early as 1988 in our search criteria, we only found one article published before 2007 [73]. Even with the recent surge in research, there is still a dire need for time-efficient and high-quality scholarship (e.g. systematic program evaluations and epidemiological studies, such as retrospective cohort studies) that investigates intervention efficacy and effectiveness for reducing the impacts of climate change on human health.
While we expected to have found more research conducted in rural US areas, there may be a dissociation between what is being done on the local, programmatic level and what is being investigated through an academic lens. It has been suggested that the efforts of grassroots organizations that are implementing strategies to reduce the burden of climate-related impacts on health do not intersect with traditional biomedical and academic scholarship [78–81]. It is highly possible that this is one reason why we did not find as much research as expected in this review. There are many grassroots organizations throughout the US that are committed to combating the effects of climate change on their communities. For example, the Women's Earth and Climate Action Network, International (WECAN), supports many grassroots efforts to promote a sustainable environment in a changing climate [82]. In the Gulf South, WECAN supports the Okla Hina Ikhish Holo: People of the Sacred Medicine Trail. This effort is described as 'a network of femme and non-binary Indigenous gardeners growing Food Sovereignty in the Gulf South' [83]. However, mistrust felt by grassroots organizations, and hesitancy to partner with academic institutions is not unfounded. Many research universities have a history of contentious, exploitative relationships with the communities in which they exist [84–86]. It is the responsibility of these institutions to establish community partnerships centered on the needs of communities and their priorities related to protecting their health in a changing climate.
Interestingly, in this review, participatory action research, also known as community-based participatory research, was frequently used as an intervention across multiple studies reviewed [33, 69–71, 75]. Participatory action research is traditionally thought of as a research method. However, the use of this method as an intervention can mobilize communities to tackle issues related to climate change. This is encouraging because engaged community members have been effective in reducing health inequities [87]. Co-creating research plans, data collection, analysis and interpretation with engaged community members are key to addressing complex problems like climate change [87]. A possible explanation for the popularity of this design could also be the 'bottom-up' approach to climate change mitigation through grassroots movements [88]. It is exciting to have found examples of academic-community partnerships tackling climate change's impact on rural health. However, there is still a pressing need to conduct high-quality, community-based research in rural areas of the US to evaluate interventions, programs, and policies designed to decrease climate change's impact on rural health. Academic scholars may gain important insights from the participatory action research articles included in this review.
Although using community-based participatory action research design could reduce the health burden of climate change, this method is time-consuming, and requires a long-term commitment from research team members and community members. Climate mitigation is time sensitive. Van Aalst et al [88] suggest building on pre-existing community risk assessments performed by governmental agencies or community organizations. In rural communities, focusing on climate-related disease burdens in crop and livestock workers and those working in forestry and fishing industries would likely develop interventions transferable to other populations within rural areas.
A surprising finding from this review were the few articles published on interventions to protect against the health effects of extreme heat on rural populations. Confirmed by another systematic review on climate change health interventions, there appears to be a gap in what is known about the health effects of heat stress, and what is being done about it [16]. It was promising that the two articles from this review focused on outdoor workers [28, 35], including farmworkers, a group that is extremely vulnerable to heat-related illnesses [6, 89]. However, no studies were identified that focused on other vulnerable groups to heat-related illnesses, such as rural older adults who live in areas that lack indoor cooling during extreme heat events [90]. Another gap found in this review was the lack of interventions related to protecting rural population health from severe weather patterns leading to droughts and water shortages, both of which are increasing due to climate change [77]. Developing, implementing, and investigating the effectiveness of interventions designed to protect rural populations who will experience water scarcity is a priority area for future research. This review did identify four studies focused on improving health outcomes after hurricanes in rural areas [31, 33, 58, 63]. Hurricane intensity is an exceptionally challenging climate event for rural, vulnerable populations throughout the southeastern United States. There are approximately 60 million persons living in the Atlantic and Gulf of Mexico regions of the United States [91] and many in these regions live within low-lying areas along the coast [11]. Of these four studies, only two were able to be methodologically rated for quality in this review [58, 63], further showing the need for timely, high-quality scholarship that explicitly evaluates the impact that interventions have on extreme weather events.
Three policy studies included in this review focused on cap-and-trade programs, in which emitters of pollution are charged for their emissions either through taxes or through tradable means, such as permits and allowances[92]. While these programs were enacted at the state level in California [64], New York [65], and Ohio [66], they non-selectively impact both rural and urban populations within a state. Even though interventions should be tailored to the unique needs of rural areas, rural areas will still benefit from interventions that non-selectively focus on decreasing the impact of climate change on health across rural and urban populations. It is worth noting that not all cap-and-trade policies have been successful. The cap-and-trade programs in California and New York were successful in reducing emissions of carbon dioxide (CO2), [64] methane (CH4), [64] nitrous oxide (N2O), [64] fluorinated greenhouse gases [65] and nitrous oxides (NOx ) [65], respectively. Ohio's cap-and-trade program was not successful in supporting farmers' uptake of conservation practices to reduce non-point source agricultural pollution [66]. Investigating the reasons behind cap-and-trade success and the factors influencing their impact on climate change's impact on rural health areas remains an area of further investigation.
4.1. Resources for climate change researchers
For future studies on the effect of climate change on human health research, a One Health approach should be employed. The One Health approach is holistic and stems from an ecosystem approach to thinking about human health that recognizes the inherent linkage between the environment, human health, animal health, and the social context in which all living beings live [93]. Many of the articles included in this review that focused on water quality were included due to the linkage between water quality and human health, even when the article did not explicitly measure health outcomes related to herbicides and/or pesticides used in rural crops that were present in drinking water. For example, one article focused on improving the management of animal manure produced and used as fertilizers [53]. Because farmers have had to grow more food on less arable land, there has been an uptick in fertilizer use in agriculture [94]. Increased food production, land-use change, and fertilizer use, particularly of nitrogen-based fertilizers, has accelerated climate change and has multiple impacts on the environment and human health through groundwater contamination [95, 96]. Researchers working to understand the impacts of climate change on human health are encouraged to consider the concept of One Health when designing and conducting studies.
Researchers could also refer to the National Institute of Health (NIH)'s Climate Change and Health Initiative to further understand the needs for climate change and health research. Seven of NIH's institutes and centers are working together to address the climate crisis through research with 'urgency, foresight, innovation, and collaborative spirit' (p 7) [97]. Some of the research needs included in the Initiatives' Strategic Framework are climate change's effect on mental health and food quality. Interestingly, another research need highlighted was protecting vulnerable populations from the impacts of extreme temperatures, especially heat in urban settings. Literature and results presented in this review suggest there is a great need for heat-health research in rural settings, as well.
4.2. Strengths and limitations
To our knowledge this is the first review to specifically investigate interventions conducted in the rural US to reduce the impacts of climate change on human health. The use of the CDC 'Impact of Climate Change on Human Health' framework highlights the broad range of academic disciplines that have worked across the framework's eight domains to better understand how to protect human health due to a changing climate. Our broad research question, and the narrow focus on specific interventions, limited our ability to perform a meta-analysis of clusters of interventions on human health outcomes. In order to quantify the impacts, future reviews will need more precise research questions that can be compared within domains and focus on a specific health outcome of choice. We did not review the gray literature, which may have limited our results to academically published articles describing interventions related to climate change on health. Future reviews may find gray literature to be a rich source of information about local actions to protect health due to a changing climate.
5. Conclusion
Several studies have been published on health interventions to reduce the health impacts of climate change effects onf air pollution, vector ecology, and water quality on the rural US population. There is a need for high-quality, timely scholarship to implement and evaluate efficacy and effectiveness of health interventions, as well as interventions to combat severe weather, heat stress, increasing allergens, and environmental degradation. Our review highlights some specific gaps in health research in rural spaces. Effective interventions that target the health of rural Americans must be prioritized as climate change continues to impact human mobility and mortality.
Acknowledgments
The authors wish to acknowledge the work of Bradley Frueh during the screening process.
Data availability statement
All data that support the findings of this study are included within the article (and any supplementary information files).
Funding
This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.
Conflict of interest
The authors do not have any conflicts of interest to declare.