Embracing Uncertainty: The Value of Partial Identification in Public Health and Clinical Research

Recovery Potential in Patients After Cardiac Arrest Who Die After Limitations or Withdrawal of Life Support

Importance

Understanding the relationship between patients' clinical characteristics and outcomes is fundamental to medicine. When critically ill patients die after withdrawal of life-sustaining therapy (WLST), the inability to observe the potential for recovery with continued aggressive care could bias future clinical decisions and research.

Objective

To quantify the frequency with which experts consider patients who died after WLST following resuscitated cardiac arrest to have had recovery potential if life-sustaining therapy had been continued.

Design, Setting, and Participants

This prospective cohort study included comatose adult patients (aged ≥18 years) treated following resuscitation from cardiac arrest at a single academic medical center between January 1, 2010, and July 31, 2022. Patients with advanced directives limiting critical care or who experienced cardiac arrest of traumatic or neurologic etiology were excluded. An international cohort of experts in post-arrest care based on clinical experience and academic productivity was identified. Experts reviewed the cases between August 24, 2022, and February 11, 2024.

Exposure

Patients who died after WLST.

Main Outcome and Measures

Three or more experts independently estimated recovery potential for each patient had life-sustaining treatment been continued, using a 7-point numerical ordinal scale. In the primary analysis, which involved the patient cases with death after WLST, a 1% or greater estimated recovery potential was considered to be clinically meaningful. In secondary analyses, thresholds of 5% and 10% estimated recovery probability were explored.

Results

A total of 2391 patients (median [IQR] age, 59 [48-69] years; 1455 men [60.9%]) were included, of whom 714 (29.9%) survived to discharge. Cases of uncertain outcome (1431 patients [59.8%]) in which WLST preceded death were reviewed by 38 experts who rendered 4381 estimates of recovery potential. In 518 cases (36.2%; 95% CI, 33.7%-38.7%), all experts believed that recovery potential was less than 1% if life-sustaining therapies had been continued. In the remaining 913 cases (63.8%; 95% CI, 61.3%-66.3%), at least 1 expert believed that recovery potential was at least 1%. In 227 cases (15.9%; 95% CI, 14.0%-17.9%), all experts agreed that recovery potential was at least 1%, and in 686 cases (47.9%; 95% CI, 45.3%-50.6%), expert estimates differed at this threshold.

Conclusions and Relevance

In this cohort study of comatose patients resuscitated from cardiac arrest, most who died after WLST were considered by experts to have had recovery potential. These findings suggest that novel solutions to avoiding deaths based on biased prognostication or incomplete information are needed.

Time-dependent association of grey-white ratio on early brain CT predicting outcomes after cardiac arrest at hospital discharge

BACKGROUND

Cerebral edema after cardiac arrest can be quantified by the ratio of grey matter to white matter radiodensity (GWR) on computed tomography (CT). Severe edema predicts worse outcomes. We hypothesized the sensitivity and false positive rate of GWR predicting outcomes change over the first 24 hours post-arrest.

METHODS

We performed a single-center retrospective cohort study including patients resuscitated from cardiac arrest between January 2010 and December 2023 who were unresponsive to verbal commands. We excluded patients who arrested from a primary traumatic or neurological etiology and those without brain imaging within 24 hours of arrest. We divided patients into groups based on time from arrest to CT, then quantified the performance of GWR dichotomized at <1.10 and <1.20, predicting in-hospital mortality and death by neurologic criteria (DNC).

RESULTS

We included 2,204 patients with mean age 59 (SD 16) years. Overall, 1651 (75%) died in the hospital, of whom 248 (11%) progressed to DNC. Sensitivity of GWR <1.10 and GWR <1.20 for predicting in-hospital mortality increased over the first four hours post-arrest, reaching a maximum of 25% after five hours, while false positive rates remained <5% at all time points. Similar temporal trends were observed with DNC, although absolute values of sensitivity and false positive rate (FPR) varied.

CONCLUSION

The sensitivity and FPR of early GWR predicting in-hospital mortality and DNC after resuscitation from cardiac arrest varies over the initial post-arrest period. Reduced GWR on brain CTs is most sensitive for in-hospital mortality when obtained more than four hours post-arrest and for DNC when obtained between four and five hours. However, FPR remained execellent throughout, making early reductions in GWR a specific marker of poor outcome regardless of timing. While brain CTs obtained within the first 24 hours post-arrest may be indicated to evaluate for neurologic etiologies of arrest, they may be less informative as an independent marker of prognosis.

The economics of adaptations to evidence-based practices

Background

Evidence-based practices (EBPs) are frequently adapted in response to the dynamic contexts in which they are implemented. Adaptation is defined as the degree to which an EBP is altered to fit the setting or to improve fit to local context and can be planned or unplanned. Although adaptations are common and necessary to maximizing the marginal impact of EBPs, little attention has been given to the economic consequences and how adaptations affect marginal costs.

Discussion

In assessing the economic consequences of adaptation, one should consider its impact on core components, the planned adaptive periphery, and the unplanned adaptive periphery. Guided by implementation science frameworks, we examine how various economic evaluation approaches accommodate the influence of adaptations and discuss the pros and cons of these approaches. Using the Framework for Reporting Adaptations and Modifications to Evidence-based interventions (FRAME), mixed methods can elucidate the economic reasons driving the adaptations. Micro-costing approaches are applied in research that integrates the adaptation of EBPs at the planning stage using innovative, adaptive study designs. In contrast, evaluation of unplanned adaptation is subject to confounding and requires sensitivity analysis to address unobservable measures and other uncertainties. A case study is presented using the RE-AIM framework to illustrate the costing of adaptations. In addition to empirical approaches to evaluating adaptation, simulation modeling approaches can be used to overcome limited follow-up in implementation studies.

Conclusions

As implementation science evolves to improve our understanding of the mechanisms and implications of adaptations, it is increasingly important to understand the economic implications of such adaptations, in addition to their impact on clinical effectiveness. Therefore, explicit consideration is warranted of how costs can be evaluated as outcomes of adaptations to the delivery of EBPs.