Novel Metrics for Quality of Hospital Surge Capacity

In this issue of Academic Emergency Medicine, Abir et al.1 report evidence that care of hospital inpatients was affected when their facilities were called upon to treat large numbers of patients from a mass casualty incident (MCI). Although they did not observe a change in clinical parameters such as mortality, patients hospitalized during the response to the MCI experienced a significantly prolonged average length of stay (LOS) and incurred higher charges. The authors considered the patients who were admitted from the MCI separately from those whose hospitalization was unrelated to the incident and compared each of them to their reference population. They rightly point out that the increased LOS and cost associated with the cohort of patients from the MCI are consistent with several previous reports in the literature and may reflect the nature of the injuries sustained more than strained hospital resources. Their finding that both LOS and costs incurred were also significantly increased for inpatients who were coincidentally present in the hospitals while the MCI patients were receiving care there is potentially important for a number of reasons. First, they represent new metrics that might be used to assess how well a hospital is able to absorb a surge of trauma patients without adversely affecting concurrent patient care. Some measures that have traditionally been cited in anecdotal reports of hospital responses to MCIs include time to definitive care and total numbers of MCI victims managed by the facilities. These are difficult data to accurately collect in real time, during response operations. It is also problematic to try to compare them over different incidents and disparate circumstances when attempting to gauge levels of preparedness and effectiveness of response. The results reported by Abir et al.1 offer some alternative metrics that might be useful for evaluation of hospital preparedness and response following an MCI. Comparing markers of efficiency, and perhaps clinical outcomes, of hospitalized patients who were not part of the MCI population with those of patients hospitalized in the same facility when there was no MCI should allow a more robust assessment of each hospital’s response capacity than has been seen in the past. Ultimately, it may be possible to create easily measurable standards for response capacity using such parameters. That process might be hastened, since the larger sample sizes, more reliable baseline data, and use of valid measures that are independent of the nature of the MCI would facilitate comparative evaluation of hospital responses to every incident. There are potentially great advantages to using data that can be retrospectively derived from established state or national databases, as was done in this study, rather than figures that must be collected during the emergency. By aggregating hospital data, it should also be possible to examine the effect of large-scale MCIs on a variety of health-related issues at the community or regional level that might otherwise be unapparent and unanticipated. Strategies for generating hospital surge capacity to manage patients from MCIs are currently superimposed on expanding evidence of the adverse, clinical effects of hospital overcrowding. Both of these issues have been the topic of Academic Emergency Medicine consensus conferences in recent years (specifically the consensus conferences from 2006 and 2011, available free of charge at www.aemj.org).2, 3 Discussions of surge capacity typically incorporate disaster triage assumptions in the development of staffing patterns, flow routines, and dedication of other resources, including hospital beds, intensive care unit space, ventilators, and other considerations. The orderly transition to surge routines, enabled by legislation in some states and envisioned for responses to catastrophic incidents, is designed to increase patient throughput, but is not expected to result in the quality of care that is provided under noncrisis conditions. The metrics used to demonstrate inferior outcomes associated with hospital overcrowding have not generally been applied to MCI operations. Does the work by Abir et al.1 provide impetus for development and refinement of measures that can be used to more accurately assess hospital preparedness and response? Hick et al.4 distinguish surge capability from surge capacity in their definitions. They define surge capability as the “ability of the health care system to manage patients who require specialized evaluation or interventions (e.g., contaminated, highly contagious, or burn patients).” Under this construct, general hospital surge capacity should be required for MCI patients needing specialty care only until they can be transported to appropriate facilities. The typical mechanisms for generating surge capacity include expedited discharge of ambulatory emergency department and inpatients, redistribution of stable patients to other facilities, cancellation of elective admissions and procedures, adding clinical and support staff, increasing the number of patients cared for by each staff member, adding beds to existing clinical spaces, and converting both clinical and nonclinical areas to care for the patients involved. Based on both exercises and published experience, it is generally estimated that between 10 and 30% of a hospital’s operating beds can be made available to manage patients from an MCI over 8 to 24 hours.5 Some of these interventions may require temporary waivers of regulations, (e.g. EMTALA, California’s nurse:patient staffing ratios, and others). Intuitively, some of these measures may be less effective at generating surge capability than surge capacity. While it is widely agreed that the care patients receive following an MCI may not be comparable to the care expected under less challenging circumstances, the goal of the emergency medical response is to approximate normal standards of care.6, 7 Of the two hospitals studied by Abir et al., one had a burn unit and one did not. The increased LOS and charges incurred in the non-MCI cohort during the surge period were observed at both facilities.1 Thus, these “spillover” effects may be relevant in assessing surge capability as well as surge capacity. It is also significant that these effects were observed in the setting of a surge that was estimated to be “>5%” of operating beds, not in the 20% range of surge capacity that is considered consistent with hospital preparedness. The paper by Abir et al.1 opens a productive area for further studies in measurement of hospital preparedness and emergency response. Their methods can be used to identify additional markers of effective surge capacity and to define goals and limits of capability based on such indicators that can be tracked over incidents of various types and magnitude. Such analysis will help distinguish whether the observed changes correlate linearly with simple overcrowding or align better with requirements for specialty services associated with specific MCIs. Of course, additional study is also needed to confirm that their observations are operationally, as well as statistically, significant. For example, it is not known which mechanisms were used by the two hospitals to generate their surge capacity. Did they institute the typical interventions of cancelling electives and discharging stable patients early? If so, this may have eliminated enough short-stay patients to alter the average LOS and cost data for the non-MCI surge cohort. The presumed spillover effect of the MCI on the other patients in the hospital might then be substantially less or absent if the cohort could be normalized to exclude this phenomenon. It is also unclear to what extent the extended LOS accounted for the increased charges incurred by the non-MCI surge cohort or whether the costs reflect additional inefficiencies imposed by the hospitals’ responses to the MCI.