Problem Statement

While there are now multiple examples of qualitative and semi-quantitative assessments being used for some degree of predicting disaster-related health risk, few models express health risk in terms of the actual probability of health outcomes and uncertainty of evidence.Reference Tracie ¹ This is particularly problematic since risk is defined as the probability that a specific outcome will occur. This absence of metrics related to chance and uncertainty limit the utility of most public health disaster risk assessments. This logically results in an acceptably accurate hazard identification and prioritization, but inaccurate and non-reproducible estimations of health-related impact. This lack of information then complicates the ability to prioritize and allocate health resources effectively, before and after the disaster.

Ideally, disaster risk management is based on a prioritization process. Once hazards have been identified, they are assessed in terms of the probability and impact in terms of losses. The hazards associated with the greatest probability and impact are prioritized. In addition to prioritization, risk assessment also offers a process for ongoing research involving the interaction of health determinants, risk, and protective factors that may contribute to future adverse health outcomes. Finally, risk assessment provides a framework for monitoring and evaluating the performance of interventions intended to reduce adverse health outcomes.

More recently, assessments of health risk have become an integral part of public health and medical emergency preparedness programs. ² One of the strengths of these assessments lies in that these processes typically bring together multi-sectoral input for public health decision making and plans. However, this diversity of input also creates challenges in development of a standardized approach for assessment, as well as a common nomenclature for assessing and communicating the characteristics of this risk.

The utility of existing models for risk assessment is attenuated by challenges with both first- and second-order uncertainty. This uncertainty is often due to a lack of predictability (reflected in the standard deviation) and accuracy (reflected in the standard error) of results from these assessments. Most risk assessment tools currently being used by the public health and medical sectors lack predictive value that would guide accurate cost accounting necessary for large-scale resource allocation or investment. Some apply risk-based models that offer indices derived from semi-quantitative estimation of select disaster risk components (namely hazards, vulnerability, and capacity). Other models offer retrospective analysis of the association between disaster incidence and select risk factors (eg, social vulnerability and capacity).Reference Tracie ^{1 ,} Reference Cutter, Boruff and Shirley ^{3 - 5}

However, to date, risk assessments do not yet offer accurate quantification of exposure as a risk factor for disaster-related adverse health outcomes. To date, none are validated as predictive of quantifiable health risk. As a result, a relatively wide variety of semi-quantitative risk assessments are being performed in various locations throughout the US and the world. Measures in many of these assessments are estimated subjectively and indicators are often sufficiently ambiguous so as to limit the reproducibility of the assessment over time or among users. Few include quantifiable representations of uncertainty.Reference Tracie ¹

Clear definitions of terminology are essential to the performance of a logical and evidence-based risk assessment. Lack of a standardized nomenclature and approach has the potential to increase variation in the accuracy and reproducibility of the assessment (resulting in a second-order uncertainty).

Perhaps part of the justification for the current level of acceptance of these tools (known to be associated with relatively high degree of uncertainty) may stem from the currently limited applications of disaster risk assessments themselves. Typically, disaster risk assessments are performed in order to gain a prioritized list of hazards before writing a plan that includes an all-hazard approach to preparedness and response. Thus, much of the value and time spent in prioritization of the risk is lost when this level of specificity is not necessarily needed for plans involving risk acceptance, which is largely comprised of general emergency response functions, which are not hazard specific. On the contrary, disaster risk management programs that include risk reduction measures require hazard-specific risk prioritization in order to evaluate cost effectiveness of risk reduction measures being considered. Accuracy and validity of the risk assessment are therefore of key importance for guiding the effectiveness of disaster risk management practice and policy.

These challenges in understanding disaster-related risk assessment have also narrowed scientific progress on risk assessment, in general. In absence of measurable indicators, risk assessments are not monitored and evaluated for quality. There are few linkages between the development of risk assessment methodology and subsequent epidemiological validation or other measures of effectiveness.

For the past decade, most disaster health studies have extensively investigated those risk factors associated with human vulnerability, while very few epidemiological studies include the widely-recognized influences of capacity and exposure in disaster risk. Epidemiological assessments of disaster-related health risk associated with environmental disasters have also typically under-reported key risk factors related to the environment and agent (exposure and hazard), while focusing almost entirely on characteristics of the host (vulnerability). Figure 1 ^{6 , 7} illustrates how disease is caused by a complex interaction between the person (host), the disease agent (hazard), and the environment (exposure).

Figure 1 Causal Factors for Disease. ^{6 , 7}

There remains a need for a reproducible, well-validated, disaster-related health risk assessment process that is suitable for accommodating the current gaps in certainty. The purpose of this report is to offer a practical framework and nomenclature for assessing disaster-related health risk that is: (1) accurate; (2) based upon historical evidence; (3) quantifiable in public health terms; and (4) inclusive of uncertainty. ⁷

Risk

In simplest terms, risk is the probability that an outcome will occur. This outcome may be beneficial or adverse. This relationship may be represented as:

$${\it p}\left( {{\rm risk}} \right){\equals}{\int} {{\it p}({\rm outcome})} .$$

Risk is the effect of uncertainty on outcomes. ^{7 ,} Reference Wharton ⁸ Uncertainty is a state or condition that involves a deficiency of information and leads to inadequate or incomplete knowledge or understanding. Uncertainty exists whenever the knowledge or understanding of an event, consequence, or likelihood is inadequate or incomplete. Expressions of risk therefore include estimations of uncertainty:

$${\it p}\left( {{\rm risk}} \right){\equals}{\int} {{\it p}\left( {{\rm outcome}} \right)} \,\pm\,{\rm uncertainty}{\rm .}$$

Impact

In general terms, risk is conceptualized as the probability of events and the severity of outcomes (consequences) that would arise if the events take place. The severity of consequences (usually conceptualized as losses) is frequently described in terms of impact. Correspondingly, the United Nations International Strategy for Disaster Reduction (Geneva, Switzerland) defines impact as “the degree of severity associated with … consequences….”Reference Keim ⁹ Thus, the term “consequence” is not synonymous with “impact.” Consequence is a qualitative description of the loss, while impact is a quantitative measure of that loss.

Risk Assessment

Risk is assessed as a function of the probability that an adverse event (referred to as a hazard) and its resultant impact will occur during a given timeframe. This relationship is commonly described as follows in what is commonly referred to as the “risk equation:”Reference Keim ¹⁰

$${\it p}\left( { R} \right){\equals}{\int} {{\it p}\left( {{\rm H}\,{\rm x}\,{\rm I}} \right)} $$

where, R= risk; H=hazard incidence; and I=degree of impact.

The process for risk assessment is described in Figure 2 and Table 1.

Figure 2 Schematic Overview of Disaster Risk Management Process. ⁶

Table 1 Key Components of Disaster Risk Management, as Applied to Health ^{6 , 11}

Exposure Assessment

The first step of impact analysis is to estimate the degree of hazard exposure that the population may be expected to receive over the time in question. ^{11 - 13} Exposures are assessed for each of the hazards identified in the hazard analysis. Exposures are assessed in terms of dose, which is a representation of the relative magnitude of the hazard and its contact rate for the population over time.

When data are available, the hazard analysis includes a hazard characterization step intended to present quantitative information in terms of dose-response relationship (the relationship between hazard dose and health response). The probability of exposures is estimated through dose reconstruction and dose-response modeling of many physical and biological hazards. It now appears possible to accomplish the same for natural and human-induced disaster hazards.

Vulnerability Assessment

Host-related factors of the target human population can influence susceptibility to the particular hazard, taking into account host intrinsic and acquired traits that modify the likelihood of contracting disease (susceptibility) and the likelihood of complications (severity). Vulnerability assessments identify those (inherent and acquired) risk and protective factors known to influence health outcomes, given exposure to disaster hazard has occurred. Susceptibility is expressed as the slope of the dose-response curve. Severity of physiological response (eg, disease) is correspondingly expressed as a peak value on the x-axis of the dose-response curve.

Vulnerability to the same hazard may differ widely among individuals within the same population. For example, people that have been immunized are much less susceptible to the same degree of exposures to viruses as compared to those who have not been immunized. And when disease does occur, some individuals in the population (eg, elderly or immunocompromised persons) are more likely to suffer a more-severe course of illness, resulting in a higher degree of impairment, disability, or even death.

Capacity Assessment

Capacity is defined as “the combination of all the strengths, attributes, and resources available in a community, society, or organization that can be used to minimize [adverse outcomes] following exposure to a hazard.”Reference Keim ⁹ Capacity is therefore considered as a measure of resources available to accomplish an objective. Populations apply individual, household, community, and societal capacity to reduce risk at every of possible stage of intervention (eg, avoidance, reduction, transfer, and acceptance). Capacity is quite complex and assessments commonly include economic, material, behavioral, and sociopolitical resources for reducing disaster risk.

Capacity assessments are most commonly represented as an asset inventory (ie, number of meals, amount of water, and number of tents). However, this application recognizes only the resources of the population. Such inventories do not take into consideration the strengths and attributes that are also necessary in order to operate efficiently and effectively achieve the intended outcome over time.

Capacity is expressed as the rate of outputs over time (ie, number of meals delivered per day; liters of water delivered per person - per day; or number of tents erected per day). Thus, the maximum capacity of the population represents a rate limiting step for reducing disaster impact.

Capacity assessments identify those resources (eg, capabilities with corresponding capacities) that are required to reduce the incidence of disease for each of the hazards. The process is as follows:

• ∙ Capability inventory - identify those capabilities that are required to reduce disease incidence;

• ∙ Expected capacity - estimate the expected capacity for each capability in the event of disaster;

• ∙ Current capacity - estimate the current capacity of the population to reduce disease incidence; and

• ∙ Gap analysis - analyze the gap between expected and current capacity to identify mismatches.

Capacity is applied in order to treat risk at all stages of risk management. Capacity assessment should include evaluations of capacity to be implemented before, during, and after the disaster event.

Table 2 ¹⁴ provides examples of capabilities that can be applied for managing health risk from environmental (eg, hydro-meteorological, seismic, and technological), societal, and biological disaster hazards.

Table 2 Examples of Societal Capabilities for Prevention of Disaster-Related Mortality ¹⁴ Abbreviations: PPE, personal protective equipment; WASH, water, sanitation, and hygiene.