Critically ill patients require different types of organ support. These include invasive therapies such as orotracheal intubation, mechanical ventilation, inotropic/vasoactive support, invasive monitoring, or continuous renal replacement therapy. This care can usually be implemented only in an intensive care unit (ICU) that has the skills, technology, and human resources to be able to perform them in a safe and effective manner.
Which Outcomes Are We Considering?
When looking at critical care outcomes, there are two aspects to consider: first, we need to select and measure a suitable outcome, and second, we need to compare the intervention with an appropriate alternative.
In recent years we have witnessed many advances, such as the implementation of protective lung strategies and the initiation of early goal-directed therapy (GDT) in septic and surgical patients, and now these are our current standards of care. These are processes where the alternative would be providing care without the same protocols/strategy. In other cases such as life-threatening respiratory failure, airway obstruction or profound hypoxia, or renal failure requiring hemofiltration, there is no alternative.
Traditionally, the principal outcome studied in the ICU is mortality (either at the ICU or hospital discharge). However, there are a number of different endpoints that might also be considered. These alternatives include longer term mortality and morbidity, neurocognitive dysfunction, impaired mental health, poor functional status, decreased quality of life, decreased return to work and usual activities, burden and stress on families, and economic costs to the patient, the family, and society. Many of these items reflect the entire hospital course, the intensive care management of which is only a small part. Quality of life and functional outcome, such as long-term survival, depend on the effectiveness of the entire health-care system, including convalescent care and rehabilitation in the community. The difficulties in measuring longer term outcomes lead us to believe that short-term benefits may persist later, but this is not always true.
Critical illness carries a substantial mortality. Despite medical advances, patients continue to die. This does not mean that our treatments are ineffective. Indeed, outcomes such as mortality must be benchmarked against national and international standards. For example, the mortality of patients with septic shock may have been reduced significantly in recent years. A 2014 meta-analysis showed that observed mortality decreased from 46.9% during years 1991-1995 to 29% during years 2006-2009 (3.0% annual change).
Outcomes from critical care can be viewed from at least three perspectives: that of patients, that of health-care professionals, and that of health managers.
Outcome for Specific Treatments
Mortality is logically an important outcome for patients, but sometimes it may not be the best choice to appropriately assess an intervention. Cardiovascular optimization, for example, requires continuous hemodynamic monitoring, and many studies have demonstrated its efficacy. Some authors have questioned the benefit of pulmonary arterial catheters, and several clinicians prefer to use less invasive monitoring systems (e.g., pulse contour analysis) to follow the trend in a number of variables and to assess the response to therapy. It is not the monitoring system, though, that affects outcome, but rather the therapeutic protocol used and the correct interpretation of the data are paramount.
In 2001, Rivers et al. showed that early GDT reduced mortality in patients with severe sepsis or septic shock compared with patients treated with standard therapy (30.5% vs. 46.5%). This study was very original at that point and focused on an aspect of emergency medicine/critical care therapy that greatly affects outcome: the timing of intervention. Previously, sepsis investigators had up to 48 to 72 hours to enroll patients. Rivers and colleagues reached hemodynamic goals in the first 6 hours of patient assessment in the emergency department (ED). Consequently, international guidelines on sepsis management maintain this advice. In 2014, Peake et al. showed that the mortality of septic patients treated without a specific protocol was lower than that observed by Rivers, but this study was conducted in the “post-Rivers age,” when clinicians realized that timely treatment is very important. Recently, different resuscitation strategies have been compared, and Rivers’ protocol was not found to be superior to usual care. This difference should be viewed in a context that accounts for the overall improvement in the quality of standard ICU care in the interim. A recent meta-analysis demonstrated that reaching resuscitation goals within the first 6 hours and compliance with a resuscitation bundle were associated with a lower overall mortality (29.3% vs. 38.6%, P < .01).
In other cases, mortality can be a very illustrative outcome. In a study of patients with community-acquired pneumonia in the ED, delayed transfer to the ICU was associated with a substantial increase in hospital mortality (odds ratio [OR], 2.07; 95% confidence interval [CI], 1.12 to 3.85). A similar study by Chalfin et al. found that delaying transfer from the ED to the ICU for more than 6 hours increased the risk of hospital death and lengthened hospital stay.
Early GDT may also improve outcome in surgical patients. Three recent meta-analysis revealed that early GDT significantly reduced perioperative complications. A mortality benefit was found, though, only in the highest risk groups, findings that must be tempered by recognition of a mortality reduction in controls, suggesting improvements in the underlying standard care ( Fig. 84-1 ). The implementation of GDT is clinically effective and cost effective . Manecke et al. estimated a cost savings of $569 to $970 per patient when this strategy was implemented.
In recent years the survival of cardiac surgical patients has benefited from improvements in postoperative care. Stamou et al. showed that implementation of a quality improvement program decreased mortality after cardiac surgery (2.6% vs. 5.0%, P < .01). Radbel et al. conducted a 15-year retrospective analysis on mortality in patients with acute respiratory distress syndrome (ARDS). This study identified 174,180 patients from the National Inpatient Sample database between 1996 and 2011 and demonstrated an absolute mortality reduction of 14.6% (from 46.8% to 32.2%) and a relative reduction of 31%. Interestingly, there was an 8.9% absolute reduction from 2000 to 2005. The authors suggested that the improvements in critical care medicine, such as the introduction of low tidal volume ventilation, contributed to this decline.
Impact of Bed Availability on Outcome: Managers’ Outlook
There is a high variability in the provision of ICU beds among different countries even when corrected for population size. In a 2012 study of bed availability in Europe, Germany (29.2/100,000) had the highest, whereas Portugal had the lowest (4.2/100,000) number of ICU beds ( Fig. 84-2 ). The overall number of critical care beds for Europe was 11.5/100,000. This is in contrast to the number for the United States (28/100,000). Although definitional differences make interpretation of these data problematic, there appears to be a relationship between the availability of beds and the outcome. Wunsch et al. demonstrated that patients admitted to U.S. ICUs had lower APACHE (Acute Physiology and Chronic Health Evaluation) II scores and lower rates of mechanical ventilation than patients admitted to ICUs in the United Kingdom. At any given time, approximately two thirds of ICU beds in the United States are occupied, and approximately one third of the total are for patients requiring mechanical ventilation. These results suggest that the United States has a significant excess of ICU beds.
The number of ICU beds per capita is highly correlated with hospital beds across all countries, excluding the United States. It is logical that the patient population in resource-poor areas is sicker, and therefore poorer outcome is to be expected. Quality of care, however, may not differ. Wunsch et al. compared the organizational characteristics of intensive care in England and the United States and found important differences, especially in end-of-life care. This study also demonstrated that the percentage of patients admitted to ICUs in England is substantially lower (2.2% vs. 19.3%) than in the United States. Furthermore, patients in critical care units account for 10.1% of hospital deaths in the United Kingdom but for 47.1% in the United States. This discrepancy reflects minimal use of ICU resources by elderly patients (1.3% in the United Kingdom vs. 11% in the United States). The two countries, though, have similar rates of intensive care use among children and young adults. The practice of discharging ill and ventilator-dependent patients to skilled nursing care facilities or chronic respiratory units outside of hospitals likely reduces the percentage of deaths that occur in U.S. ICUs. The differences between the largely public health-care system in the United Kingdom and the essentially private system in the United States may also explain the differences found by Wunsch et al.
It is difficult to estimate what constitutes optimal provision of ICU beds in a given population. Recently Wunsch described a “Starling curve” for intensive care, speculating that exceeding a certain limit of beds per capita results in poorer outcome and an increase in health-care spending. Therefore, in countries with lower availability of beds, the triage of patients for ICU admission plays a fundamental role.
Impact of Admission/Discharge Criteria on Outcome: Health-Care Professionals’ Outlook
Critically ill patients use resources disproportionately. Expense relates to the high costs of staffing ICUs, where a higher nurse/patient ratio is essential. Thus, ICU admission criteria that identify patients who are truly likely to benefit are a matter of considerable importance. Some national societies have produced guidelines for admission criteria to the ICU. These guidelines suggest that the categories of patients who do not take benefit from the ICU are those “too well to benefit” and those “too sick to benefit.”
The benefit from ICU admission can best be determined by comparing similar patients admitted or not admitted to the ICU. A number of studies demonstrate that there are times when patients are not admitted to the ICU because beds are not available. Sinuff et al. systematically reviewed 10 studies and found that patients denied admission to an ICU were at considerably higher risk of dying (OR, 3.04; 95% CI, 1.49 to 6.17). A report by Simchen et al. demonstrated poor outcomes in patients meeting ICU admission criteria but managed elsewhere because of bed shortages. In this study, only 27% of eligible patients were admitted to the ICU within 24 hours, when survival benefit was greatest. These findings were confirmed by the same authors years later ( Table 84-1 ).