Internal capacity, technical expertise, and political mandates vary from one local health department (LHD) to the next. The framework presented in this paper (Figure 1) recognizes the need to tailor engagement around climate and health to the specific needs of individual communities. It, therefore, outlines a tiered approach that meets the needs of both communities that are new to addressing the health effects of climate change, as well as LHDs with sufficient technical capacity and political support to incorporate climate projections into their climate and health planning activities.

The first step in developing local EPHIs for climate change is to identify the highest priority climate change-related hazards in the jurisdiction. Work produced by NCADAC provides an overview of regional climate change-related hazards both according to historical data and future climate projections [17]. Online tools, such as those produced by the Natural Resources Defense Council [18], show historical rates of exposure to climate-related hazards at the state and county levels. Case studies showing examples of the environmental hazards targeted by other US states and cities have been developed by the CDC Climate Ready States and Cities Program [19]. The relevancy of the local set of EPHIs can be maximized by prioritizing the hazards identified in a local climate assessment. For example, LHDs in the Washington, DC, metropolitan area might consider developing EPHIs tracking the health effects of one or more of the following natural hazards, which were identified by the Metropolitan Washington, D.C., Council of Governments as the most significant climate-related threats facing their region: extreme heat, heavy precipitation, severe storms, and sea level rise [20].

In order to become useful decision-support tools, climate change EPHIs at the local level must reflect their political, economic, and social context. It is, therefore, important to balance an initial review of the scientific evidence of local climate-related hazards with existing and planned policies that could influence implementation of interventions tailored to enhance community resilience. For example, a review of the scientific literature and downscaled climate projections might indicate that heat waves and air quality are the most significant environmental hazards in a city. However, if it is located just inland from a highly populated region in a hurricane-prone zone, the political climate might encourage an increased focus on preparing for expected large-scale population displacement in the wake of severe storms.

A local community's hazard mitigation plan may or may not address the health effects of climate change. Many climate change policies in the USA only address mitigation activities (i.e., activities that reduce greenhouse gas emissions [21], such as increasing building energy efficiency or transitioning the city's fleet of vehicles to hybrids). However, even if they do not directly address public health programs, they still indicate policymakers' priorities and objectives. It is, therefore, important to prioritize policy and intervention indicators in the climate change EPHI system with cobenefits for both climate change mitigation and community health.

Initial EHI selection should first involve examination of the state and county indicators developed by the NEHTP [6] and CSTE [16], which point to historical weather, health, and policy data related to climate change hazards such as extreme heat events. Data and step-by-step instructions for assembling these indicators can be found at the NEHTP and CSTE websites: http://ephtracking.cdc.gov/ and http://www.cste.org/group/Indicators/. County data has, in many cases, already been compiled nationally and is available at these websites or from individual state climate or tracking programs [22]. The state health department may be able to provide technical assistance or to collaborate with county health departments to compile indicators that are relevant at the local level. While it is preferable to develop local indicators using national, peer-reviewed datasets, it may be necessary to use a local source—such as data collected through a government epidemiology department, a partner agency, or a local university. In some cases, it may even be necessary to collect new datasets; however, this approach can be time consuming and costly. Proxy indicators may be the most time- and cost-effective mechanism for filling data gaps.

Secondly, where possible, these county-level datasets should be replaced withmore granulated data from local sources, because the aggregation level for some datasets may be too broad to inform local policy. For example, the CSTE climate change indicator, positive test results in sentinels and reservoirs [23], is collected at the state level. However, a vector-borne infectious disease like West Nile Virus will likely spread through a state over a period of time rather than exposing all state residents to equal levels of risk from the beginning of the outbreak. It is important for LHDs to know when the first positive test results approach their jurisdiction, so precautionary measures can be initiated [17, 24]. Furthermore, many of the vulnerability and health outcome indicators outlined from NEHTP and CSTE are relevant to other local surveillance activities, such as asthma reduction programs, chronic disease programs, and flood safety programs. If these programs are already tracking an indicator, the national data source can be replaced with the local source in the final set of local EPHIs. This approach will reduce the likelihood of duplication of efforts across agencies and will help integrate climate readiness considerations into core public health services.

Qualitative data, such as oral histories, can help fill data gaps that are still present at this point. Given the challenges associated with data collection at the local level, it is unlikely that quantitative datasets will paint a comprehensive picture either of population vulnerability or of direct links between public health interventions and community resilience. It is, therefore, important to identify opportunities for incorporating qualitative data into the overall tracking program to capture health determinants that might otherwise be overlooked. For example, in 2009, researchers from the National Center for Atmospheric Research conducted semistructured surveys among populations in Phoenix that had been identified using empirical research methods as potentially vulnerable to negative health outcomes during extreme heat events. The survey gathered information about participants' knowledge of attitude toward and experience with heat waves. It also asked where participants received information and where they turned for help during emergencies (both from within their social network and from community resources). The study results showed that several empirical indicators (such as access to air conditioners) that might be assumed to reduce vulnerability did not necessarily align with the lived reality of these populations, because some participants could not afford to run their air conditioners at a level to sufficiently cool their homes. And many of the survey participants were not aware of community resources designed to address that barrier, such as financial assistance with electric bills and air conditioning repairs [25].

The climate change EPHIs can now be integrated into an agency's overall surveillance program. This step will increase the sustainability of the climate change EPHI program by leveraging existing capacity and data sources. It will also raise awareness throughout the LHD and the community about the many ways that climate change impacts population vulnerability and health outcomes.

When completed, this baseline set of local climate change EPHIs can be shared with agency leadership and the general public. Developing accompanying educational material will assist community engagement efforts in reducing vulnerability, particularly among high-risk populations such as the young, the elderly, and populations with low socioeconomic status.

Once local EPHIs have been developed and incorporated into the local public health surveillance system, they can be used to locate clusters of vulnerable populations. For example, EPHIs for extreme heat can be mapped to identify hot spots where both the built environment and resident populations demonstrate high levels of vulnerability. Sister departments and external partners, such as research institutions and nonprofit organizations, may already have developed geospatially coded data sets that can be incorporated into the vulnerability assessment. And state health departments may be able to provide technical assistance supporting the geospatial analysis required to develop the vulnerability index. For example, the Florida ESF8 Planning Map viewer (http://gis.doh.state.fl.us/ESF8PlanningMap/) allows users to overlay social vulnerability indices and vulnerable population densities on top of flooding and storm surge models.

In order to develop a vulnerability assessment, a definition of exposure must be set for the climatic hazard based on either national standards or regional best practices (such as a definition of an extreme heat event set by the National Environmental Health Tracking Program [26]). The selected definition should be applied to historical extreme weather events to establish trends in climate change-related morbidity and mortality [27–29].

The next step is to gather evidence from a combination of the public health literature [13, 30] and local research to identify populations most likely to suffer negative health outcomes after exposure to the climatic hazards included in the analysis. The vulnerability assessment will combine this information with elements of the natural and built environment, such as exposure to urban heat islands (for extreme heat events) or exposure to air pollution and deforestation (for air quality indicators). The resulting vulnerability index will combine the socioeconomic and demographic data with built and natural environment data into a cumulative vulnerability index. Examples of methodologies for developing vulnerability indices and methods for sensitivity testing have been developed by English et al. [31], Houghton et al. [32], Jerrett et al. [33], Reid et al. [34], and Tate [35, 36].

While developing EPHIs and vulnerability indices using historical data provide a necessary baseline for launching the climate and health planning process, historical trends are not sufficient to prepare for the future effects of climate change. Downscaled climate models must be overlaid on the historical data to develop credible scenarios for how risks, exposures, and vulnerabilities are likely to shift spatially and demographically over time.

Regional downscaled climate change projections can be found through the National Climate Assessment [17]. A number of states and regions have also developed their own, more granular, projections [37]. For example, the New York Climate and Health Project combined heat, air quality, and land-use data to predict the health effects of increasing temperatures [38]. This information can be combined with projected changes to the built environment and demographic shifts to update the vulnerability index according to several likely scenarios. Hayhoe et al. [39], Li et al. [40], and Peng et al. [41] present three methodologies for developing heat wave mortality projections, while Deschênes and Greenstone [42] and Nicholls [43] offer two methodologies for predicting excess all-cause mortality attributable to climate change. Comparisons between baseline vulnerability/mortality and future projections should focus on local policy priorities, such as reducing injury and death rates, bolstering economic prosperity, and reducing health disparities.

Local climate change EPHIs, whether based solely on historical data or incorporating climate projections, provide local health departments with an important tool for implementing key steps of the CDC's Building Resilience Against Climate Effects (BRACE) framework (http://www.cdc.gov/climateandhealth/BRACE.htm/). The BRACE framework relies on data inputs such as EPHIs to demonstrate successful completion of each of its five steps. For example, local environmental exposure and population vulnerability indicators are developed during Step 1, forecasting climate impacts and assessing vulnerabilities. EPHIs and vulnerability indices developed exclusively using historical data are valuable inputs during this step. Similarly, Step 2, projecting the disease burden, mirrors the process of projecting future trends in vulnerability and mortality after overlaying downscaled climate projections on baseline EPHIs and vulnerability indices. Step 3, assessing public health interventions, relies on EPHIs to measure the relative success of policies designed to reduce community vulnerability to climate change. EPHIs should be developed to align with the overarching goals developed in Step 4 of the BRACE framework, developing and implementing a climate and health adaptation plan, so that health indicators are tracked alongside other performance measures managed by sister agencies. Finally, by tracking reductions in population vulnerability and mortality to climate change, EPHIs form the quantitative evidence-base for Step 5, evaluating impact and improving quality of activities.

In addition to forming the quantitative backbone of a climate and health program, climate change EPHIs can be integrated into a local health department's core organizational framework by supporting delivery of several of the 10 essential public health services (EPHS). For example, climate change EPHIs support EPHS 1, monitor health status to identify and solve community health problems, because they are designed to identify trends in exposure of vulnerable populations to climate change-related events. They also support EPHS 2, diagnose and investigate health problems and health hazards in the community, by tracking vulnerabilities and outcomes associated with climate change-related hazards. For example, an air quality EPHI will often track asthma rates associated with increases in ozone formation. Community education and outreach programs developed in accordance with EPHS 3, inform, educate, and empower people about health issues, may rely on climate change EPHIs to share data about the links between climate and health, the dangers associated with exposure to climatic hazards, and policies that are available to assist neighborhoods in reducing vulnerability to climatic events. EPHI data can be used to develop maps and other infographics aimed at mobilizing community support for climate and health policies and interventions (EPHS 4, mobilize community partnerships and action to identify and solve health problems). Visual materials can also be incorporated into interdepartmental and interagency tabletop exercises, such as using a heat vulnerability map and urban forestry/tree canopy map to inform decisions regarding placement of future municipal tree plantings. Climate change EPHIs can also support EPHS 5, develop policies and plans that support individual and community health efforts, by identifying areas that have historically sheltered clusters of vulnerable populations. For example, policymakers might prioritize incentives for constructing vegetated roofs in certain areas of town after reviewing vulnerability indices identifying the neighborhoods that are highly vulnerable to both heat and flooding. Similar to EPHS 5, EPHIs can support EPHS 6, enforce laws and regulations that protect health and ensure safety, by tracking vulnerability and mortality associated with exposure to high risk climatic events. The data generated through the EPHI tracking process can be used to inform assessment of workforce competency, capacity, access to training, and certifications relevant to climate change, supporting delivery of EPHS 8, assure competent public and personal health care workforce. Similarly, EPHIs can be used to generate an evidence base for assessments of local climate adaptation policies designed to protect public health, supporting delivery of EPHS 9, evaluate effectiveness, accessibility, and quality of personal and population-based health services. Finally, the process of developing local EPHIs will inevitably lead to uncovering data and research gaps, supporting EPHS 10, research for new insights and innovative solutions to health problems.