Internet-Based Heat Evaluation and Assessment Tool (I-HEAT): Feasibility Analysis of a Visualization and Decision-support Tool for Extreme Heat Preparedness in Detroit, Michigan

1.1. Climate Change, Extreme Heat, Sustainable Living, and Human Health

Climate change will increase the duration, frequency, and intensity of heat waves (IPCC 2007), with portions of the United States (U.S.) experiencing more intense and frequent periods of extreme heat than in previous years (Meehl and Tebaldi 2004). The threat of extreme heat to human health is a real and pressing concern. Despite significant protective technological advances (e.g., forecasting meteorological events, increased access to air conditioning) to reduce exposure to high temperatures, extreme heat remains the leading cause of death related to natural disasters in the U.S. (Borden and Cutter 2008). Due to projected climate change, current and future populations face challenges in maintaining the dynamic relationship between natural systems and human health.

The impacts of extreme heat on human health are often difficult to predict or identify and have been assessed via retrospective epidemiological studies. These studies emphasize that heat-related morbidity and mortality depend on the characteristics of the heat event, as well as the health status and geography of the affected population (Basu 2002; Curriero et al. 2002; Anderson and Bell 2011). Physiologically, short-term exposure to increases in ambient temperature (e.g., heat waves) can cause the human body to reallocate blood flow from vital organs, such as the heart and lungs, to below the skin’s surface in order to thermoregulate. Ineffective thermoregulation, due to aging, medication use, or chronic health conditions (e.g., diabetes), can result in misdirected blood allocation and ultimately contribute to cardiovascular and respiratory stress (Astrand et al. 2003). Population characteristics such as age and health status are commonly used indicators of biological characteristics that contribute to susceptibility to extreme heat.

Climate models suggest that large U.S. urban populations will experience disproportionately extreme heat (Oleson et al. 2013). This is due in part to the urban heat island (UHI) effect where observed ambient temperatures are higher within a city than in the surrounding less-developed, more vegetated suburban and rural areas (Arnfield 2003). Heat is absorbed by cityscape materials like concrete, asphalt and brick, contributing to the UHI effect (Imhoff et al. 2010). U.S. metropolitan centers are projected to grow significantly by the mid to late century; larger and older populations will be living in urban areas (Frumkin 2002) with higher ambient temperatures, potentially leading to a larger-magnitude impact of extreme heat on human health. Some adaptive measures, such as increasing air conditioning prevalence and usage, can limit individual and short-term exposure to extreme heat—but they can also increase ambient temperatures (Sailor and Pavlova 2003) and contribute to global climate change via the release of greenhouse gases and latent heat expenditure (Salamanca et al. 2014). It is critical to consider the net effect of adaptive measures when identifying ways to reduce future risk of extreme heat.

1.2 Identifying Populations Vulnerable to Heat at the Local Level

To prevent heat-related morbidity and mortality, local municipalities face an increasing need to identify populations most vulnerable to extreme heat, and spatial analyses using local data may assist these efforts. National research priorities pertaining to climate change and health have garnered attention from public and private entities, and several researchers have called for the integration of climate and health data to better inform climate change mitigation and adaptation planning decisions (Hess, Schramm, and Luber 2014; Bierbaum et al. 2013; Miles et al. 2006). Heat-related vulnerability, as defined by Wilhemi and Hayden (2010), comprises exposure, sensitivity, and adaptive capacity. Scholars and practitioners have mapped this construct in diverse ways, often using a heat vulnerability index (HVI) (Johnson et al. 2012; Reid et al. 2009; Reid et al. 2012; Wolf and McGregor 2013; Loughnan et al. 2013).

Building on these spatial analyses, organizations that conduct heat preparedness have called for the development of easily accessible tools that can aid in the identification of areas where vulnerable populations reside, as a way to prepare for extreme heat events (O’Neill et al. 2010). Such tools must reach diverse audiences, as heat preparedness efforts are often led by different sectors in different municipalities (White-Newsome et al. 2014), including departments or agencies of aging, sustainability, emergency preparedness, and public health departments. In response, the Internet-based Heat Evaluation and Assessment Tool (I-HEAT) (available at www.biomedware.com/I-Heat/IHeatViewer.html) was recently developed and feasibility tested with diverse stakeholders in Southeast Michigan. I-HEAT is a user-friendly, dynamic mapping tool that concurrently displays HVI data with multi-scale remotely sensed land surface temperatures and measures of vegetation. Users can access I-HEAT online and identify populations at risk from extreme heat in Detroit, Michigan. A series of tabs displays the following information at the census-tract level atop a background map: at-risk areas, satellite and temperature data, demographic data, and HVI values. For further details on the tool and its development, see Conlon et al. (2014).

1.3. Study Purpose and Objectives

City and community leaders must prepare locally for projected climate change, including heat-related threats to health. To this end, scholars and practitioners have mapped heat vulnerability using varied data and scales, and some are beginning to develop decision-making application tools based on these analyses. I-HEAT, a web-based interactive mapping software, is one example. There are few reports of feasibility studies that document how these tools are made usable for and evaluated by end-users, such as emergency managers, public health practitioners, and sustainability coordinators. In this paper, we describe our process for presenting I-HEAT to a diverse group of 28 end users in Southeast Michigan during a day-long workshop, Heat and Health in Michigan: An Interactive Workshop on the Development of Risk Assessment Resources, where we facilitated interactions with the tool and collected feedback through surveys and focus groups. We present the findings of this I-HEAT feasibility study and future directions for development, implementation, and evaluation for I-HEAT and similar tools.

2.1. Climate Change in Detroit, Michigan

Michigan has many climate change adaptation efforts underway. The Michigan Department of Community Health (MDCH) participates in the Climate-Ready States and Cities Initiative of the U.S. Centers for Disease Control and Prevention (CDC). Through this program, MDCH utilizes CDC’s technical and organizational support in developing Michigan-specific climate change adaptation efforts. As part of the collaboration, MDCH works closely with local governments and officials, including representatives from Detroit and surrounding areas, to focus climate adaptation efforts on priority areas, one of which is extreme heat events (Seroka, Kaiser, and Heany 2011). In the Michigan Climate and Health Adaptation Plan, state epidemiologists note that heat-related projections for Detroit are of particular concern, given that more frequent extreme heat events are projected to double or triple in number for Detroit, with temperatures exceeding 90°F approximately 30 to 50 days per year and exceeding 97°F approximately 25 to 50 days per year (Cameron et al. 2011).

2.2. I-HEAT Feasibility Study

In May 2012, the research team held a workshop with local decision makers to introduce and evaluate I-HEAT. This workshop was organized in collaboration with researchers from Michigan State University (MSU), the Michigan Department of Environmental Quality, and the Michigan Department of Community Health. The team was composed of environmental scientists, epidemiologists, behavioral scientists, and ecologists. Participants were selected through snowball sampling (Patton 2002), where members of this research network identified individuals to invite based on their current or potential role in local heat preparedness, and these individuals identified further contacts with relevant knowledge and experience. In particular, we identified and invited individuals who were most likely to make decisions about heat-related health interventions, representing local health, environmental planning and emergency preparedness agencies, departments, and organizations. Detroit’s heat preparedness efforts, however, have not been extensive (White-Newsome et al. 2014). Thus, some participants were not identified as heat preparedness experts; rather they were identified as playing a role that could be relevant in future heat-related climate change planning. Local decision makers from across mid-Michigan were invited, with an over-sampling of Detroit decision makers since the data in the I-HEAT prototype were Detroit-specific.

The workshop design was based largely on an active learning framework (Bens 2005) that emphasized participants’ interactions with I-HEAT and other users. The day’s agenda included icebreakers, large- and small-group activities, presentations, surveys, and focus groups, as well as morning refreshments and lunch. After an introductory welcome, a professional facilitator led a series of activities that required participants to get up and move around the room, physically indicating and discussing their knowledge, attitudes, and concerns about climate change and heat preparedness. This activity enabled participants to recognize their mental models (Morgan et al. 2002) about climate change, heat-related risks, and heat preparedness, which was especially important given the participants’ large range of knowledge. Next, members of the research team gave a brief overview of heat vulnerability and epidemiology, followed by a description and demonstration of I-HEAT for the City of Detroit. After these presentations, participants formed teams of two or three and were provided prompts and scenarios (Table 1) to simulate real-world uses of I-HEAT and guide discussions. Participants interacted with the software online on laptop computers in groups and then completed individual surveys. After lunch, the research team divided the participants into three subgroups and conducted focus groups.

3.1. I-HEAT Feasibility Study Findings

Twenty-eight participants (including researchers) attended the workshop. Twenty-three completed surveys and 26 participated in the focus group. Participants represented state agencies (n = 5), academic institutions (n = 11), non-profit organizations (n = 5), private research organizations (n = 2), local health departments (n = 4) and the National Weather Service (n = 1). As summarized in Table 3, on average, participant satisfaction with I-HEAT fell closest to ‘satisfied’ regarding the types of data (mean = 3.86), user-friendliness of the web interface (mean = 3.64), general appearance of the web interface (mean = 4.09) and the graphics of the application (mean = 3.86). On average, participants were between ‘unsatisfied’ and ‘neither satisfied or unsatisfied’ with the performance, or speed, of the application. During the workshop we did encounter an unforeseen technical problem with the server on which I-HEAT is hosted, which limited the number of simultaneous users. Scores in the performance and speed category likely reflect this experience.

3.2. Strengths and Limitations

This feasibility study could have been improved. The workshop was a day-long event, which may have deterred key decision makers from attending if they were unable to step away from the daily demands of their job for a full work day. As a remedy, researchers could have shortened the workshop or provided additional approaches for testing I-HEAT, such as a self-directed interaction and survey administered online. Furthermore, the sample size of 28 participants yielded informative results, but additional feasibility testing in other cities that have implemented climate adaptation and mitigation plans could be helpful to determine if 1) I-HEAT would be integrated into those efforts and 2) how similarly or differently end-users evaluate I-HEAT. Finally, the range of knowledge and the varied professional roles of survey respondents may complicate the interpretation of survey findings. For instance, participants who do not work directly on heat and health issues may not immediately see how the tool could assist them in their professional work, and they may respond to survey questions accordingly. We attempted to address this concern by providing a basic presentation on heat epidemiology to ensure that participants shared the same baseline knowledge.

Despite these limitations, the feasibility study advanced knowledge among the research team and participants. Specifically, survey and focus group data provided the research team with valuable information about the strengths and weaknesses of I-HEAT to inform next steps in tool development and implementation. The workshop provided new information to local decision makers, many of whom may not have had previous knowledge of heat epidemiology and vulnerability or potential heat preparedness interventions. Through the in-person workshop with diverse stakeholders, the feasibility study also enabled cross-sector, cross-disciplinary discussions that are vital to successful sustainability and climate planning initiatives.

3.3. Future Directions for I-HEAT

Future versions of I-HEAT will continue to build on the knowledge gained through collaboration with state and local experts on extreme heat and health, which may increase the generalizability of the tool. The research team is currently working to identify ways to expand I-HEAT with additional data that reflect community resources and health behaviors (e.g., transportation, social services, air conditioner usage), as these factors also contribute to vulnerability (Sampson et al. 2013). For example, after the workshop, cooling center locations were added to the web interface to help users see if and how current heat preparedness interventions aligned with the locations of vulnerable populations. In Detroit, the Detroit Climate Action Collaborative is now leading development of a Climate Action Plan. If updated, I-HEAT may assist in these efforts by helping decision makers target adaptation and mitigation strategies in ways that protect populations most vulnerable to local climate change. Most heat-related vulnerability studies are published in academic journals and are not usually tailored to communicate the results to an audience of practitioners. I-HEAT uniquely presents research-based information into a tool that can communicate complex environmental health relationships in an accessible manner. Workshop participants stressed that mapping tools like I-HEAT are relevant to audiences beyond environmental health practitioners. For instance, environmental planners, urban planners, and sustainability coordinators could use a tool like I-HEAT to identify fine-scale geographic areas where increased tree or vegetation coverage could result in co-benefits of greenhouse gas mitigation, stormwater management, and the reduction of the local UHI.

3.4 Conclusions

Increasing resiliency to climate impacts on human health requires integrated coordination among climate and health scientists, as well as municipal-level decision makers and planners. With the mass proliferation of climate adaptation tools (e.g., www.cakex.org), efforts to conduct feasibility studies of these tools are vital. As our study illustrates, end-users can offer insights that improve the tool’s usability, clarify its language, expand its applicability, and contextualize its findings. The iterative development process of tools like I-HEAT provides benefits to producers and users alike. Actions to reduce exposure to extreme heat should consider practices that do not exacerbate the problem of climate change, but rather confer co-benefits to natural and human systems. This feasibility analysis suggests the possibility of developing decision-making tools that contribute to developing programs and interventions that support sustainable urban living.