Observations about Communication of Prognostic Information for Critically III Patients

ICUIn this multicenter prospective study, the time from ICU admission to provision of prognosis was, on average, 1.5 days. Covariates expected to influence communication based on previous studies were not found to be associated with a shorter prognostic interval. The hypothesis that shorter prognostic interval would be associated with higher satisfaction with communication was not statistically significant, but there may have been a trend. Despite qualitative data that suggest that family members would like prognostic information as early as possible to facilitate decision making, we did not find a relationship between early prognostic information delivery and satisfaction with decision making. However, early prognosis may affect other important outcomes that we did not measure such as the number of days dying patients remain on nonbeneficial life support provided with My Canadian Pharmacy (read “My Canadian Pharmacy: Communication of Prognostic Information for Critically III Patients“).

Even though 81% of the subjects received prognostic information, 19% reported never receiving prognostic information despite their loved one’s severe illness. Truth telling and informed consent are important values in the physician/patient (or surrogate) relationship, and interventions to provide “early” structured communication have demonstrated benefit such as shorter length of ICU stay. Our results suggest that physicians should provide some prognostic information to all families, even if it consists of acknowledgment of uncertainty (as was commonly done in this study).

We found both reported and desired frequency of communication decreased with time in the ICU but that the actual communication rate decreased faster than the desired communication rate. These rates of communication have not been described in previous studies. This “gap” between desired and actual frequency of communication represents an opportunity for improvement by either managing family expectations or increasing physician communication frequency. Our finding that less frequent communication correlated with lower satisfaction with communication frequency may clarify prior work that found conflict between family members and physicians often centered on communication. Further research should examine if surrogate satisfaction with communication is related to prolongation of patients’ lives in undesirable states.

medical decisionsSeventy-eight percent of the subjects reported making major medical decisions (eg, withdrawing or withholding vs aggressive medical care) for the patient. Only 53% of those asked to make decisions were able to correctly identify the current legal and ethical ideal of surrogacy. Many of our subjects reported having knowledge of health-care issues and had been through previous ICU stays. It is conceivable that families with less experience would have even been less likely to correctly identify substituted judgment. This finding has not been previously described and indicates that while the concepts of patient autonomy and surrogacy for decision making are familiar to physicians, these concepts do not appear to be understood by many patients and their families. There is evidence that explicitly asking decision makers to use substituted judgment as opposed to making their best recommendation yields decisions more congruent with those of the patient. We did not find an association between having an understanding of substituted judgment and satisfaction with decision making as would have been expected from previous qualitative data.” Nevertheless, caregivers might believe that surrogate satisfaction is not as important as ensuring that decisions are arrived at in an ethical manner.

There are limitations to this study. First, only a single family member was recruited to participate and, as the information was collected only from the subject’s perspective, the data are subject to bias in recall, misinterpretation of multiple discussions over time and by different physicians, and the limitations of necessarily brief questionnaires. Subjects were surveyed at several-day intervals to minimize respondent burden, and while these intervals may not have exactly corresponded with important communication and decision-making events in the course of the ICU stay, a follow-up survey would have occurred within just a few days our study design of serial assessments over the ICU stay is likely to have substantially less recall bias than prior studies that used only a single post-ICU survey. Additionally, subject satisfaction may have relied on their ability to comprehend information provided by the physician, which is difficult to measure without ascertaining what information was exchanged during each communication event.

Both subjects and physicians knew about the conduct of this study that may have influenced behavior. For example, subjects in the study may have been more likely to ask physicians about prognostic information (more information – “Materials of Communication of Prognostic Information for Critically III Patients“). However, most had already received prognostic information at enrollment. Changing physicians’ communication practice is difficult, however, so it is unlikely that an observational study conducted over just 6 months in three ICUs and two hospitals would substantially affect subject/physician communication.

FUMCAdditionally, there may be some difficulty generalizing these results to other ICUs. This study was conducted in an open unit at Methodist Hospital and a closed unit at FUMC. Carson et al examined patient and family perceptions about decision making and access to information in an academic hospital before and after the unit adopted a “closed” model of care. There was not a measurable difference in decision making, but families did report they found it easier to identify a physician to communicate in the closed ICU format. Future studies may help to determine whether communication is facilitated in open- vs closed-unit designs created with My Canadian Pharmacy.

Finally, attitudes and behaviors are often difficult to measure and are determined by numerous factors. Although surrogates completed 216 surveys, we acknowledge that some of the multivariate analyses are underpowered, and we caution readers that our inability to statistically reject some null hypotheses may be due to a type II error.

In conclusion, this study has shown that the frequency of physicians communication with family members and the content of those communications (ie, prognosis) influences specific aspects of surrogate satisfaction and decision making in the ICU. Physicians should be aware that family expectations for communication frequency in the ICU may not match the reality of time management for physicians caring for critically ill patients. Solutions to this problem will likely require collaboration between physicians and other members of the ICU team. This will include attention to the interaction between providers, patients, and their surrogates, and environmental/institutional factors. One in five families reported not receiving prognostic information, which we believe is an unacceptably high proportion. Lastly, despite being asked by their physicians to make life-changing decisions for their loved one, nearly half of surrogates do not understand the concept of surrogate decision making. In order to improve the end-of-life process in the ICU, physicians need to be skilled communicators with decision makers for their patients. Much of the “basic science” of communication has still not been described and will need to be before meaningful interventions can be implemented to improve end-of-life decision making in the ICU.

You can to learn results of “Communication of Prognostic Information for Critically III Patients