The Application of a Hospital Medical Surge Preparedness Index to Assess National Pandemic and Other Mass Casualty Readiness

Abstract

EXECUTIVE SUMMARY 

This article describes the use and findings of the Hospital Medical Surge Preparedness Index (HMSPI) tool to improve the understanding of hospitals’ ability to respond to mass casualty events such as the COVID-19 pandemic. For this investigation, data from the U.S. Census Bureau, the Dartmouth Atlas Project, and the 2005 to 2014 annual surveys of the American Hospital Association (AHA) were analyzed. The HMSPI tool uses variables from the AHA survey and the other two sources to allow facility, county, and referral area index calculations. Using the three data sets, the HMSPI also allows for an index calculation for per capita ratios and by political (state or county) boundaries. In this use case, the results demonstrated increases in county and state HMSPI scores through the period of analysis; however, no statistically significant difference was found in HMSPI scores between 2013 and 2014. The HMSPI builds on the limited scientific foundation of medical surge preparedness and could serve as an objective and standardized measure to assess the nation’s medical readiness for crises such as the COVID-19 pandemic and other large-scale emergencies such as mass shootings. Future studies are encouraged to refine the score, assess the validity of the HMSPI, and evaluate its relevance in response to future legislative and executive policies that affect preparedness measures.

INTRODUCTION

Background and Rationale

Events and threats that test the United States’ national resilience and population health are increasingly common, and it is essential that America’s healthcare system be ready to respond to crises. At the time of this writing, the world is in the midst of a pandemic that threatens the health of the human race, particularly people with comorbidities. As an individual’s health depends on access to and delivery of quality healthcare, mass casualty emergencies require a hospital to “medically surge” to save lives (^{Committee on Quality Measures, 2013}).

The ^{Agency for Healthcare Research and Quality (2006)} defines medical surge capacity as “a healthcare system’s ability to expand quickly beyond normal services to meet an increased demand for medical care in the event of bioterrorism or other large-scale public health emergencies.” The ^{American College of Emergency Physicians (2005)} describes medical surge capacity as a measurable representation of a healthcare system’s ability to manage a sudden or rapidly progressing influx of patients with the available resources at a given point in time. In 2006, ^{Barbisch and Koenig (2006)} identified four elements of surge capacity: staff, stuff, structure, and system. Building from this, ^{Kelen and McCarthy (2006)} then created a framework based on these four metrics—substituting “supplies” for “stuff” and “space” for “structure”—to estimate an institution’s capacity to address surge events. The metrics are defined as follows:

• Staff refers to personnel such as nurses, physicians, pharmacists, respiratory therapists, and technicians as well as nonmedical personnel who are necessary for the efficient functioning of a healthcare facility or entity, including clerical support personnel, security specialists, and physical plant specialists.
• Supplies include durable equipment, such as cardiac monitors, defibrillators, intravenous (IV) pumps, ventilators, blood glucose monitors, and laboratory equipment. Supplies also include consumable materials such as medications, oxygen, sterile dressings, intravenous fluids, IV catheters, syringes, sutures, and personal protective equipment.
• Space includes total beds, staffed beds (number of beds that staff is available to attend to), available spaces and opportunities to house patients, and the percentage of beds occupied.
• Systems for healthcare organizations include integrated policies and procedures that operationally and financially link multiple hospitals and individual departments within a healthcare setting.

Until recently, there was no objective, dedicated tool for assessing a hospital’s capacity to respond to a mass casualty event such as a pandemic. The Hospital Medical Surge Preparedness Index (HMSPI), based on ^{Kelen and McCarthy}’s (2006) framework, defines a methodology to quantify a hospital’s surge capacity, or the ability to care for many during a public health emergency (^{Marcozzi et al., 2020}). The HMSPI cross-references variables from the American Hospital Association (AHA) database to the categories described in Kelen and McCarthy’s surge framework. This standardized assessment index gives hospitals and regions more precise and easily understood metrics to assess medical surge and enable more informed decision-making and preparedness efforts for pandemics and mass casualty incidents.

It should also be noted that healthcare policy and reimbursement changes most likely affect surge capacity, including the ability of a hospital to mount an effective pandemic response. While reform efforts vary, concepts that focus on expanding insurance coverage, increasing provider coordination, modernizing electronic health records, and decreasing costs have prevailed in the past decade (^{McHugh, 2010; Obama, 2016}). One well-known health policy reform was the Affordable Care Act (ACA), signed into law on March 23, 2010 (^{U.S. Centers for Medicare & Medicaid Services, 2010}). The objective impact of such health policies is poorly understood in regard to medical surge capacity.

At multiple levels, including congressional oversight, presidential directives, federal guidelines, payment mechanisms, and metrics of success, the delivery of healthcare during public health emergencies or mass casualties—and the financing thereof—is considered separate and distinct from care delivered during periods of normalcy. This disconnect results in a fragmented and underdeveloped scientific foundation, inadequate population-based quality measures, ineffective patient- centered care, inadequate training standards, and a lack of readiness of the U.S. healthcare system to deliver care during a crisis. In the face of inadequate public health preparedness resources and measures, public health officials, healthcare professionals, and emergency managers continue to seek better measures and approaches to “prevent, protect against, mitigate, respond to, and recover from” all forms of disasters (^{U.S. Department of Homeland Security, 2015}).

Relevant to the nation’s current COVID-19 response, New York’s response to the Ebola outbreak demonstrated that equipping U.S. healthcare systems for medical crises requires time, resources, cooperation, and collaboration at all levels of care (^{Zucker et al., 2017}). Although policy revisions and investments in healthcare preparedness have been made, the nation’s medical preparedness is still insufficient (^{Zucker et al., 2017}). Now, almost two decades after 9/11, no objective score to assess a hospital’s readiness for a surge of patients from a pandemic or other disaster exists.

The intention of this analysis was to apply HMSPI methodology and objectively measure U.S. hospitals’ surge capacity based on reported healthcare survey data to quantify the ability to respond to mass casualty events such as the current pandemic, natural disasters, and other public health emergencies. Specifically, this study

• used healthcare data to assess and score a hospital’s ability to medically surge and respond to a mass casualty incident,
• aggregated these hospital-level data to build county- and state-level HMSPI scores,
• used geographic data to calculate HMSPI per capita ratios and normalize weaknesses and strengths across the nation, and
• monitored the HMSPI scores and ratios over time to analyze the effect of national healthcare policies on hospital-, county-, and state-level medical surge preparedness.

METHODS

Study Design

Secondary data analysis was conducted to evaluate the impact of defined criteria of hospital capacity, operations, staffing, and financial information on medical surge capacity for all 50 states, based on population. This analysis used 2005 to 2014 data from the annual survey conducted by the AHA (2018) as well as data from the ^{U.S. Census Bureau (2008)} and the Dartmouth Atlas Project (^{Wennberg, 1996}). This study was conducted in accordance with Strengthening the Reporting of Observational Studies in Epidemiology guidelines (^{Von Elm et al., 2007}). The institutional review board at the University of Maryland School of Medicine approved this study.

Data Sets

The AHA annual survey collects a comprehensive data set from community hospitals in the United States, with a response rate above 75%. Community hospitals include all nonfederal, short-term general and specialty hospitals as well as nonfederal, short-term academic medical centers, and other teaching hospitals. The resulting database allows analysis of trends in personnel, revenues, and expenses across local, regional, and national markets (AHA, 2018). The analysis involved data from community hospitals operating in the United States between 2005 and 2014. The AHA data set included 6,966 hospitals; 727 of them were excluded because of incomplete data, leaving 6,239 hospitals for this analysis. The variables used to create the HMSPI came from the AHA survey (AHA, 2018, Table S1, provided as Appendix 1 to this article, published as Supplemental Digital Content at https://links.lww.com/JHM/A56).

The U.S. Census Bureau sponsors the American Community Survey (ACS), an ongoing nationwide project “designed to provide communities with reliable and timely demographic, housing, social, and economic data” (^{U.S. Census Bureau, 2008}). The Census Bureau releases data from the ACS as 1-, 3-, and 5-year estimates. The ACS replaces the decennial census long form in providing detailed state and local data (^{U.S. Census Bureau, 2008}). The ACS data were linked to the AHA survey using a combination of local zip codes and Federal Information Processing Standards county codes. County-specific demographic variables were then extracted from the ACS and used to create models that could evaluate the impact of the ACA (described later).

The Dartmouth Atlas Project compares healthcare effectiveness and efficiency of regions, states, individual hospitals, and associated physicians in treating chronically ill patients (^{Wennberg, 1996}). In this study, the hospital service area zip codes provided through Dartmouth’s database were used to establish the geographic service area of each hospital in the AHA database.

The HMSPI

The HMSPI is based on ^{Kelen and McCarthy}’s (2006) theoretical framework and contains four metrics of surge capacity, summarized as follows:

• Staff (e.g., nurses, physicians, clerical support)
• Supplies (including durable equipment [defibrillators, ventilators, wheelchairs] and consumables [medications, oxygen, syringes])
• Space (e.g., bed counts, staffed bed counts, percentage of beds occupied)
• Systems (e.g., policies and procedures that integrate departments within the healthcare facility as well as link the facility with other providers of healthcare, such as physicians’ offices)

All four metrics were previously verified using exploratory and confirmatory factor analyses that demonstrated high internal reliability (^{Marcozzi et al., 2020}).

The HMSPI and Geolocation

The longitude and latitude of each hospital facility (available from the AHA survey) were used to calculate HMSPI per capita ratios for each geographic unit: hospital service area, county, and state. The numerator was the HMSPI score per unit (in the case of a county or state, the sum of all hospital-level HMSPI scores in that area). The denominator was the total population living in the geographic unit (hospital service area, county, or state). Population estimates were calculated using data from the Dartmouth database and the ACS. Trends were derived by observing ratios across all years in the sample (2005–2014) and differences between 2 years (2014 minus 2013 for ACA evaluation and 2014 minus 2005 for evaluation of the entire study period).

Evaluation of Hospital Surge Capacity Before and After the ACA

To evaluate whether a healthcare policy change affects the HMSPI, the passage of the ACA was examined and a series of difference-in-differences (DD) and difference-in-difference-in-differences (DDD) analyses were conducted. Specifically, county-level data on 2012 to 2014 insurance coverage rates from the Small Area Health Insurance Estimates Program (^{U.S. Census Bureau, 2018}) were used to estimate a DD using a generalized linear model in which

• HMSPI ratio is the sum of individual hospital scores for a given county divided by the population covered by health insurance in a particular year,
• year is an indicator variable equal to 1 in 2014 and 0 in either 2013 or 2012 (separate models), and
• Medicaid is an indicator variable equal to 1 if a given state expanded its Medicaid program under ACA and 0 if it did not.

In addition, the DDD analysis included a vector of county-level categorical variables indicating proportions for gender, age, race, ethnicity, and education groups among that population. All analyses were performed using the R language (^{R Foundation, 2015}).

RESULTS

Sample Characteristics

The data set included 6,239 hospitals from the AHA national survey, excluding the 727 hospitals with missing data. Among the 6,239 hospitals, 3,123 had 82 or fewer beds and 3,116 had more than 82 beds. Comparisons were performed using chi-square tests for categorical variables and t tests and one-way analysis of variance tests for continuous variables. As expected, statistically significant differences were observed for all the evaluated hospital metrics, with larger hospitals (i.e., more than 82 beds) having a significantly higher number of total facility admissions, surgical operations, emergency department visits, operating rooms, adult and pediatric beds, neonatal intermediate and pediatric intensive care beds, and burn beds (p < .001; ^{Marcozzi et al., 2020}).

Outcomes, Analyses, and Trends

The evaluation revealed a steady increase in HMSPI scores from 2005 to 2014 in every state. The degree to which the HMSPI scores increased per year varied across states, with trends suggesting some leveling off in the last year of the study. This plateau was, however, not statistically significant.

Figure 1 shows the increase in hospital HMSPI scores throughout the nation over the study period, stratified by the number of beds. Of note, Alaska and Hawaii were included in the analysis but removed from graphical map results for clarity. Figure 1 reflects the trend in increasing county-level HMSPI scores throughout the nation over the study period, demonstrated through the progressive transition from light to dark in all regions. Figures 2 and 3 display changes in state-level HMSPI scores between 2005 and 2014 and 2013 and 2014, respectively. Between 2005 and 2014, Montana experienced the largest increase in HMSPI scores, followed by Kansas; the smallest increase in HMSPI scores occurred in Nevada. Results between 2013 and 2014 were mixed. The greatest increase in HMSPI scores occurred in North Dakota, the greatest decreases were observed in Maine and Arkansas, and many states had little or no differences over this period.

Figure 1: Trends Over Time in Hospital Medical Surge Preparedness Index Scores for Hospitals, by Number of Beds

Figure 2: Change in State Hospital Medical Surge Preparedness Index Scores Between 2005 and 2014

Figure 3: Change in State Hospital Medical Surge Preparedness Index Scores Between 2013 and 2014

From the analysis, and as might be expected, larger hospitals (those with more than 82 beds) had significantly higher surge capacity in terms of the space metric. This evaluation revealed HMSPI scores increasing from 2005 to 2014 at both county and state levels. However, there was no statistically significant difference in HMSPI scores between 2013 and 2014 or between 2012 and 2014.

Findings From the DD and DDD Analyses

To further investigate the varied effects of the ACA on HMSPI scores, DD and DDD models were estimated as a series. As shown in Table 1, regardless of which model was employed, there was no statistically significant difference in HMSPI scores between 2013 and 2014 or between 2012 and 2014.

TABLE 1 - Estimated Differences in Hospital Medical Surge Preparedness Index Scores Before and After Patient Protection and Affordable Care Act Implementation

Model	Years	Score difference (p value)
DD	2013–2014	19.96 (0.69)
DDD	2013–2014	19.96 (0.79)
DD	2012–2014	19.83 (0.58)
DDD	2012–2014	19.83 (0.09)

Note. Analysis of 2005–2014 ^{American Hospital Association (2018),U.S. Census Bureau (2008)}, and Dartmouth Atlas (^{Wennberg, 1996}) data. DD = difference-in-differences; DDD = difference-in-difference-in-differences.

DISCUSSION

The scientific foundation for estimating a hospital or healthcare system’s ability to deliver medical care during a public health emergency is limited, even though it is frequently cited in the medical literature as a challenge in achieving optimal outcomes during and after disasters. After more than $5 billion of federal investments in disaster preparedness since the 9/11 attacks, the nation’s hospitals and healthcare systems continue to struggle with disaster readiness, which at its core is defined by access to quality healthcare during large-scale crises (^{Lister, 2014}). The intention of this study was to better link and understand the intersection between routine hospital data and a score that quantifies the ability to respond to a mass casualty event or pandemic. Knowing or not knowing whether (and which) hospitals are able to respond to an influx of patients above normal volumes affects resource allocation, response planning, communications strategies, and, ultimately, lives.

Scientific understanding of and advances in medical care during crises have been improving modestly, and most efforts in this area are not proactive. Unfortunately, any scientific advancements that improve healthcare delivery result mostly from intense attention during a disaster response or in its immediate aftermath—when policymakers, researchers, resources, hospitals, and the public are mobilized to address the challenge. Typically, this energy and prioritization fade with the passing of the event and, although much is written on how to achieve change, little is done to assess and improve a hospital’s operational readiness for a surge associated with a mass casualty incident or pandemic.

This study is unique in several dimensions:

• It creates an objective index and score for previously conceived but unsolidified concepts.
• It uses contemporary healthcare data to assess and score a hospital’s ability to medically surge and respond to a mass casualty incident.
• It aggregates hospital-level data to construct county- and state-level scores.
• It studies HMSPI scores over time and analyzes the impact of national healthcare policies on hospital-, county-, and state-specific medical surge preparedness.
• It assesses per capita medical surge preparedness, which can be applied to population health initiatives and thereby assist decision-making for future funding and policies.

Study Limitations

Despite filling an important and timely gap in the literature, this study has some limitations. First, although the methodology behind the technical properties of the HMSPI has been reported in an earlier article (^{Marcozzi et al., 2020}), the index has not yet been validated in relation to hospital performance in the face of actual disasters. Therefore, it should not yet be used to estimate victims’ access to healthcare or outcomes in a given U.S. hospital during a mass casualty and medical surge situation. Second, this study analyzed data up to 2014, which provides only one year of data from the time the ACA was fully implemented. Given the dynamic nature of hospital infrastructures, analysis of data from later years might lead to different results. Third, the use of three data sets meant that any missing values affected the overall analysis. Finally, the HMSPI calculation is limited with respect to the supplies metric because the AHA database lacks granularity pertaining to durable medical equipment.

Fundamentally, mass casualty events such as the current pandemic show us the strength and weakness of the U.S. healthcare system. The resilience of a community and a nation are enhanced by strengthening the healthcare system and improving health outcomes, as measured by health indicators, the response to health needs, protection from economic and social risks, and the efficiency of the system (^{Barnea et al., 2020}). While optimal day-to-day delivery of healthcare is foundational to this system, mass casualty incidents uncover and amplify healthcare inefficiencies and pose population health challenges. By creating a standardized index to evaluate both the local and national ability to respond to mass casualty events, this measure could become a much-needed tool to guide future discussions among policymakers, emergency preparedness planners, hospital administrators, and health services researchers in the field. For example, the HMSPI scores could be incorporated into accreditation, quality, and other ratings from governmental or industry bodies. Future studies should continue to refine the score, assess the validity of the HMSPI, and evaluate its performance in response to future crises and policy changes.

CONCLUSION

The HMSPI builds on the limited scientific foundation of medical surge preparedness. As threats increase in frequency and magnitude, it is imperative that the United States advance its understanding of healthcare delivery during crises to avoid substantial morbidity and mortality. An improved ability to measure and understand a hospital’s capacity to respond to medical surges could have resulted in better planning and decision making in response to the surge caused by the coronavirus pandemic, possibly lessening the staggering number of COVID-19-related deaths. In the future, the HMSPI could serve as an objective and standardized measurement to assess the ability of hospitals, counties, and states to manage public health emergencies, such as the current coronavirus pandemic, and improve planning for mass casualty medical surges.

NOTE

This study was funded by the Bipartisan Commission on Biodefense and the Department of Emergency Medicine, University of Maryland School of Medicine.

Acknowledgments

Deborah M. Stein, ELS, and Linda J. Kesselring, ELS, provided language and technical editing of the manuscript.