Patient Priorities and ICU Care in Adults Receiving Mechanical Ventilation

A recent study evaluated whether ICU care delivery aligns with patient treatment priorities among adults receiving invasive mechanical ventilation. Goal-concordant care, defined as medical care that reflects an individual’s goals and priorities, is a central principle of critical care. However, measuring goal concordance is difficult because patient priorities may change during illness, are often absent from electronic health records, and are commonly assessed retrospectively, introducing recall bias. Since critically ill patients are frequently unable to communicate their wishes, surrogate decision-makers often represent patient preferences despite evidence that surrogate assessments are not always accurate. The study therefore sought to determine whether ICU care corresponded with treatment priorities reported in real time by surrogates and whether these priorities were associated with patient characteristics.

This was a secondary analysis of the Successful Clinical Response in Pneumonia Therapy (SCRIPT) study, a single-centre prospective cohort involving adults admitted to a medical ICU who required invasive mechanical ventilation and bronchoalveolar lavage for suspected pneumonia. Patients enrolled more than 72 hours after ICU admission or whose surrogates declined to answer questions regarding treatment priorities were excluded. During enrolment, surrogates were first asked whether they had previously discussed unacceptable medical treatments with the patient. Regardless of the response, they were then asked to identify the patient’s overall treatment priority by selecting one of three options: maximising survival despite burdensome treatments, adopting a middle-ground approach that balanced survival with some treatment limitations, or maximising comfort and quality of life even if this shortened life expectancy. Demographic characteristics, comorbidities, ICU interventions, hospital outcomes, and end-of-life care were extracted from the electronic health record. Statistical analyses included χ² tests, Kruskal-Wallis tests, linear regression, and ordinal logistic regression.

Of the 393 eligible patients enrolled within 72 hours of ICU admission, 217 surrogates (55.2%) completed the treatment priority questionnaire. The most common priority was the middle-ground approach (48.4%), followed by maximising survival (34.6%) and maximising comfort (17.1%). Median patient age increased across the three categories, although the difference was not statistically significant. Most patients identified as White, and racial distribution was similar across groups. Nearly all patients were admitted with a full-code status, although this was less common among those whose reported priority was maximising comfort. Importantly, patients in the maximise-comfort group were significantly more likely to have previously discussed treatment preferences with their surrogate than those in the other two groups.

The primary finding was that surrogate-reported treatment priorities were not strongly associated with ICU care delivery. Median ICU length of stay and duration of mechanical ventilation were numerically shorter among patients prioritising comfort, but these differences were not statistically significant. Rates of tracheostomy, extracorporeal membrane oxygenation, specialist palliative care consultation, and ethics consultation were also similar across all treatment-priority groups. Likewise, the frequency of invasive interventions did not differ meaningfully according to patient priorities.

Hospital outcomes were similarly comparable. In-hospital mortality was approximately one-third across all three groups, with no statistically significant differences. Discharge disposition and the need for ongoing life-sustaining therapies, including invasive mechanical ventilation, renal replacement therapy, or artificial nutrition, were also similar regardless of treatment priority. The one notable exception was hospital length of stay. Patients whose reported priority was maximising comfort experienced significantly shorter hospital stays than those prioritising survival or adopting a middle-ground approach. This association remained significant after adjustment for age, race, and Charlson Comorbidity Index score, suggesting that comfort-focused priorities may have influenced the overall duration of hospitalisation even if they did not substantially alter other aspects of care.

Among patients who died in hospital, end-of-life care was also remarkably similar across treatment-priority groups. Most patients, including those whose priority was maximising comfort, continued to receive life-sustaining therapies during the 48 hours preceding death. Rates of cardiopulmonary resuscitation at the time of death and family presence were likewise comparable. These findings indicate that reported patient priorities had little observable influence on treatment intensity near the end of life.

The investigators also explored associations between patient characteristics and reported priorities. Younger patients were more likely to prioritise maximising survival, whereas middle-ground preferences predominated among older patients, although all age groups demonstrated considerable variation. When analysed by race, middle-ground priorities remained the most common overall, except among Black patients, who most frequently prioritised maximising survival. In unadjusted analyses, Black race was associated with lower odds of comfort-focused priorities, but this relationship narrowly lost statistical significance after adjustment for age and comorbidity. The authors emphasised that demographic characteristics should not be used to infer individual treatment preferences because substantial variability existed within every subgroup.

The authors proposed several explanations for the limited association between patient priorities and care delivery. Clinicians may simply have been unaware of patients’ priorities early in the ICU admission, as these discussions often occur only after life-sustaining treatments have already been initiated. Many patients whose priority was maximising comfort nevertheless had full-code orders, illustrating a disconnect between documented resuscitation status and overall treatment goals. Surrogates may also hesitate to limit aggressive treatments because of concerns that clinicians could withdraw other beneficial care or because they feel unable to challenge existing treatment plans. These findings suggest a need for more structured and reliable opportunities for discussions between clinicians, patients, and surrogates regarding goals of care early during critical illness.

Overall, the study found little evidence that ICU care delivery differed according to surrogate-reported patient treatment priorities. Although patients prioritising comfort had shorter hospital stays, most ICU interventions, hospital outcomes, and end-of-life practices remained similar regardless of stated goals. These findings highlight the ongoing challenge of achieving truly goal-concordant care in critical illness and underscore the need for systematic processes that reliably incorporate patient priorities into clinical decision-making throughout the ICU stay.

Source: Chest
Image Credit: iStock