Bundled Consent in US Intensive Care Units

We designed a brief survey (provided as a Supplement to this article) to assess how procedural informed consent is used in adult ICUs in the United States. Respondents indicated whether their units used bundled or per-procedure consent and used a 5-point scale (never, rarely, sometimes, often, or always) to estimate how often procedures were performed in their ICUs without documented informed consent. We specified a priori that our principal measure of consent rate would be a self-report of never performing ICU procedures without consent.

Survey respondents also reported whether an intensivist was the attending physician of record in their ICU, which team members routinely obtained consent, and why procedures were performed without prior informed consent in their units. Respondents who indicated that their ICUs used bundled consent were asked a series of follow-up questions to assess why their units adopted this practice and if they would recommend the practice to other ICUs. Respondents whose units used per-procedure consent were asked if they would be interested in implementing bundled consent and, if not, to indicate the reason.

We prospectively surveyed a nationally representative sample of US hospitals from June 2017 through May 2018. We randomly selected hospitals with medical or combined medical/surgical ICUs from the 2017 edition of the AHA [American Hospital Association] Guide to reach providers at both academic and nonacademic institutions.⁸  We gathered additional provider and hospital information for each of the ICUs from the AHA Annual Survey Database.⁹ 

We administered the survey by calling ICUs and requesting to speak to an intensivist physician, bedside nurse, medical director, or nurse manager. We intended to select providers who had direct knowledge of informed consent practices in their unit, spent significant time providing real patient care, and routinely participated in bedside procedures. When we reached appropriate individuals, we asked them to participate in the study. Respondents who agreed received the survey either by telephone or in an email with a personalized link to an online form. We collected and managed study data with electronic data capture tools (REDCap, hosted at the University of Chicago).¹⁰  We obtained informed consent for participation either by telephone or electronically. The study protocol and all materials were deemed exempt by the institutional review board at the University of Chicago (IRB170284).

We powered our study to generate a sufficiently precise estimate of the cross-sectional prevalence of bundled consent use. Assuming that the national prevalence was 15% on the basis of the findings of Stuke et al,⁶  we calculated that surveying 100 hospitals would produce a 95% confidence interval of approximately 10% to 20%. We determined the univariate associations between bundled consent use and various hospital and provider factors, including type of hospital (teaching affiliation, critical access status, and ownership), total number of hospital beds, and number of medical/surgical ICU beds. We compared differences between groups by using t tests of means for normally distributed continuous variables, quantile regression for medians of non-normally distributed continuous variables, and 2-tailed Fisher exact tests for categorical variables. All data analysis was performed with statistics software (Stata 15, StataCorp LLC), and P less than .05 was considered significant.

Of 6307 hospitals in the AHA Guide, 2791 had medical/surgical ICUs and were therefore eligible to be surveyed. We randomly selected 238 hospitals, and respondents from 100 completed the survey (42% response rate; see Figure 1 for flow diagram). Of the respondents, 94% were nurses, 3% were physicians, and 3% were other clinicians. Participating and nonparticipating hospitals were similar in terms of mean number of ICU beds, geographical location, hospital ownership, and teaching affiliation (Table 1).

Of the 100 participating hospitals, respondents from 15 reported using bundled consent in their ICUs (prevalence, 15%; 95% CI, 9%-24%). Hospitals using bundled consent had more medical/surgical ICU beds than did hospitals using per-procedure consent. We found a nonsignificant trend toward more total beds in hospitals using bundled consent than in those using per-procedure consent. Number or presence of full-time intensivists, geographic location, critical access status, teaching affiliation, and hospital ownership did not differ according to type of consent (Table 2). Respondents from both groups agreed on the common ICU procedures that require informed consent (Figure 2). When asked to indicate all of the possible providers who routinely obtained informed consent, respondents most often reported the attending physician of record (68%), followed by bedside nurses (58%), physician assistants and nurse practitioners (30%), and trainees including residents and fellows (27%). Formal training for clinicians on how to discuss informed consent was rarely reported for all ICUs but was more common among units using bundled consent (53% vs 24%, P = .02).

Of respondents using per-procedure consent, 75% self-reported performing invasive procedures without consent, as compared with only 47% of respondents using bundled consent (P = .02; Figure 3). For both groups, the most common reason for performing a procedure without consent was the urgent nature of the procedure (63%). Other reasons were inability of patient to consent with no surrogate available (9%) and unavailability of personnel with necessary skills and credentials to obtain consent (1%).

Respondents from all 15 hospitals using bundled consent said they would recommend the practice to other hospitals. Four common reasons given for using bundled consent were “more efficient for practitioners,” “more efficient for patients/surrogates,” “concern about procedures being performed without consent,” and “belief that this practice is better for physician-patient/surrogate communication” (Table 3). Seven of 15 hospitals using bundled consent had been using it for more than 5 years. Of the respondents using per-procedure consent, 39 of 80 (49%) reported interest in implementing bundled consent (5 respondents using per-procedure consent did not answer this question). Some reasons for lack of interest were “belief that separate consent should be obtained for individual procedures,” “concern that practice may negatively affect physician-patient/surrogate communication,” and “concern regarding legal risk/liability.”

To our knowledge, our study is the first to use a nationally representative sample of both academic and nonacademic adult ICUs to assess the prevalence of bundled consent use for common ICU procedures. We found that bundled consent was used infrequently, with only 15% of US adult medical and combined medical-surgical ICUs using the practice, and that this prevalence has not changed significantly in the past decade. Respondents using bundled consent were less likely than those using per-procedure consent to self-report performing invasive procedures without consent. Notably, all survey respondents using bundled consent recommended the practice to other ICUs, and almost half of respondents using per-procedure consent expressed interest in implementing bundled consent at their institutions.

There are several potential reasons why bundled consent has not been used more widely despite the interest in bundled consent expressed by our survey respondents who were using per-procedure consent. First, providers may worry that bundling together several procedures would reduce a patient’s or surrogate’s understanding of a given procedure. This concern is a reasonable one that has contradictory reports in the literature. One study indicated that bundled consent leads to families missing out on critical updates,¹²  and in another study, only 11% of residents conducted a complete informed consent discussion for each procedure in a bundled consent protocol.⁵  Other single-center studies found that bundled consent did not reduce comprehension by individuals who provided consent³  and even increased family satisfaction with the ICU.⁴  Another advantage of bundled consent over the per-procedure approach is that ICU clinicians need not delay or defer diagnostic or therapeutic procedures that do not meet the criteria for emergency need or implied consent. By enhancing efficiency of the informed consent process, bundled consent may reduce these risks to patients and promote patient autonomy. This factor is of particular relevance considering the increasing use of ICUs without a concomitant increase in the number of critical care providers.¹³ 

Second, a major ethical concern is that bundled consent could result in bombarding the patient or surrogate decision makers at the time of admission with a lengthy and involved conversation about procedures that may need to be done at some point during the patient’s stay. A recent study assessing residents’ perceptions of bundled consent showed that 88% felt there were too many procedures to discuss and that 78% of residents thought bundled consent was awkward and stressful for families.⁵ 

Third, several respondents using per-procedure consent cited unease regarding legal liability if their hospitals were to adopt bundled consent. We found that respondents using bundled consent were unanimous in recommending the practice to other institutions and that almost 50% of these users had been using bundled consent for more than 5 years. This strong endorsement may suggest that these ICUs have not encountered significant medical-legal issues or dissatisfied patients and families, but additional work is needed to identify the precise legal challenges that may be state specific.

We found that respondents from ICUs that used bundled consent were markedly more likely to self-report obtaining informed consent before invasive procedures were performed. Respondents using per-procedure consent were 1.6 times as likely as those using bundled consent to report performing procedures without consent. This observed effect size is consistent with results of the single-center study by Davis et al,³  in which switching from per-procedure to bundled consent increased the rate of consent for procedures. However, this finding from our study is exploratory because it is based on providers’ self-reports. In addition, we did not directly measure the quality of informed consent conversations. Users of bundled consent may have been more inclined to self-report improved consent out of desire to please regulatory agencies and demonstrate that this protocol improves hospital compliance.

Finally, we demonstrate a clear need to enhance clinician training on how to engage in the process of informed consent. Only 30% of respondents in our study reported that their ICUs had a formal training process. This finding aligns with the moderate resident readiness reported in an academic institution using bundled consent. In that study, residents self-rated their comfort level as 3.3 on a scale of 1 to 5 for conducting a consent conversation.⁵  Given the increasing premium placed on patient self-determination and the central role of informed consent within that paradigm, mechanisms to train and assess clinicians’ competency in informed consent practices should not be an afterthought.

Although our response rate was acceptable for this type of survey research, our results may be subject to response bias. Self-reported procedural consent rates might be subject to social desirability bias, which could skew the reported rates higher than the actual rates. However, we would expect hospitals using per-procedure consent and those using bundled consent to experience this bias equally, so it should not invalidate the observed differences between groups.

Although we set out to survey a mix of provider types, most respondents were bedside nurses, and we believe that this is an unexpected strength of our data. Nurses participate in almost all bedside procedures in the ICU and have more direct contact with patients than physicians do.^14,15  The physician explains the risks, benefits, and alternatives of a procedure, whereas the nurse ensures that the consent conversation occurred and, more importantly, acts as an advocate for patients.^15–18  Nurses routinely follow checklists to maintain hospital protocols, such as informed consent, and are able to indicate whether or not informed consent is routinely obtained before procedures. They are ideally positioned to state how often procedures are performed without consent because they are more likely to report what is actually happening in routine clinical care. In our survey, 58% of respondents reported that nurses routinely obtained informed consent in their ICUs, highlighting nurses’ involvement in the informed consent process.

The use of bundled consent for ICU procedures is uncommon in the United States and has not changed since 2010, when it was assessed in academic centers.⁶  Intensive care unit providers, mostly nurses, expressed interest in implementing bundled consent in their units, and respondents using bundled consent unanimously recommended it to other institutions, suggesting a need to provide more information about the practice. Many ethical challenges are involved with bundled consent, including the potential to overwhelm patients and family members by front-loading informed consent conversations or by asking for opinions on procedures that might change if the patient’s condition worsens. However, bundled consent may be effective for increasing rates of informed consent and incorporating patients’ wishes into medical decision-making if better training is provided to the clinicians obtaining informed consent. Future work is needed to assess whether patients, family members, and health care providers prefer bundled consent over individual per-procedure consent.

We acknowledge the Pritzker School of Medicine Scholarship and Discovery team for providing allocated research time and support.