Perceptions of Nurses and Physicians of Their Communication at Night About Intensive Care Patients’ Pain, Agitation, and Delirium

The survey was conducted at Tufts Medical Center, Boston, Massachusetts, a 320-bed academic medical center with two 10-bed medical ICUs (MICUs), and was approved by the appropriate institutional review board.

Because no survey instrument could be found that focuses on communication between ICU caregivers at night that could be adapted for use in the study, existing literature^{1–20,23–25}  on communication in the ICU, particularly articles related to pain, agitation, and delirium, was reviewed to inform the development of a new survey instrument. The review revealed the following constructs:

• The perceived quantity of communication at night between nurses and physicians about pain, agitation, and delirium, and the preferred methods for initiating and delivering this communication, differ between nurses and physicians.

• The perceived urgency and accuracy among nurses and physicians of their communication with each other at night about pain, agitation, and delirium differ.

• Nurses and physicians’ perceptions of how recommendations are communicated and accepted between nurses and physicians at night differ.

In an effort to more thoroughly explore the factors associated within each construct, the survey was designed to include questions about the recent nighttime experiences of each respondent with members of the other profession, the knowledge of each respondent about patients’ pain, agitation, and delirium, responses to vignettes, and attitudes toward both the respondent’s own actions and the actions of the other group (ie, nurse or physician).

A total of 2 instruments were used: 1 survey for physicians and 1 for nurses (see Appendix at the end of this article). The instruments were similar; each was developed via a deliberate series of steps that included generation and construction of items, pilot testing, and clarification. For the purpose of the survey, the nighttime period was defined at the time between 7 pm and 7 am, and a page was defined as each time an ICU nurse (or the nurse’s designate) sent either the ICU telephone number or a descriptive text page to the ICU on-call physician. In order to minimize recall bias, respondents were asked about only the most recent night shift they had worked. The initial survey items were generated on the basis of the survey constructs identified via the literature review by a multidisciplinary group of critical care clinicians at the study site: 2 staff nurses who usually work at night, a nursing manager, 2 physicians (1 member of the house staff and 1 attending physician), and 1 pharmacist. We asked these 6 clinicians to specifically comment on the number of nighttime communications and the preferred communication strategies they use, how urgent and accurate this communication usually is, and how recommendations are usually delivered and accepted during this communication.

A total of 3 members of the medical house staff and 3 ICU nurses (who work at night) were interviewed by using the draft survey instrument, and changes were made on the basis of their comments on survey length and clarity. The instrument was further modified on the basis of the feedback received from 4 national ICU sedation and delirium experts (2 nurses, 1 physician, and 1 pharmacist) about the relevance and clarity of the survey questions. The nurse and physician surveys were organized into 4 sections on the basis of the defined constructs: demographics, number of communications and preferred communication delivery methods, communication urgency and accuracy, and delivery and acceptance of recommendations. For questions on frequency, 5 different responses were provided on the basis of standard survey methods.²⁶  The responses and their representative percentages (frequency) were categorized as follows: never or almost never, 10%; seldom, 30%; sometimes, 50%; often, 70%; and always or almost always, 90%.

The paper-based surveys were administered to house staff physicians who had completed 1 or more MICU rotations in the past year at the medical center and to critical care nurses who currently worked 50% or more of their shifts at night in 1 of the 2 study MICUs. Physicians on staff are on call for 8 to 10 nights per each 4-week MICU rotation; therefore the completion of 1 or more MICU rotations was thought to represent an adequate exposure to night on-call time for the purposes of the survey. Nurses who worked less than half of their shifts at night were thought to represent nurses who might be less exposed to nocturnal communication practices. All responses were anonymous, and no incentives or compensation was offered to survey respondents. Reminders to complete the survey were sent out 1 week and 2 weeks after the first distribution.

All survey results were exported into a research database (SPSS, version 17.0; IBM Corp). Responses were analyzed by using standard statistical methods, including χ² analyses and t tests when appropriate. During factor analysis, multivariate analysis of variance was conducted in those instances in which multiple outcomes were considered in order to avoid a type I error. The value P ≤ .05 was considered significant.

A total of 45 nurses and 75 physicians were sent the surveys. Of these, 30 nurses (67%) and 56 physicians (75%) responded. Most nurses (73%) had worked at night in a MICU for more than 5 years, and most physicians (68%) had completed 3 or more 4-week MICU rotations during which they were on call at night for 8 to 10 nights per rotation.

Among the ICU nurse-physician interactions thought to occur during a typical night (via pager or face to face), the proportion pertaining to pain, agitation, and delirium was similar between nurses (35%) and physicians (31%) (Figure 1). Nurses (63%) and physicians (55%) each perceived that more communications related to pain, agitation, and delirium were initiated via a pager than face to face. When receiving a page from a bedside nurse about a patient’s pain, agitation, or delirium, physicians always preferred to receive a text page containing a brief description of the situation, whereas 27% of nurses preferred to send the ICU telephone number only.

For many common clinical situations related to pain, agitation, or delirium that occur in the ICU at night, nurses and physicians differed in their preference of whether the situation could be initially managed over the telephone or whether the on-call physician should go directly to the patient’s bedside (and not telephone the nurse who placed the page; Figure 2).

A similar proportion of nurses (80%) and physicians (70%) either moderately or strongly agreed that nurses use good judgment when paging physicians at night about issues related to pain, agitation, or delirium (P = .24). However, for common situations related to pain, agitation, and delirium that might occur at night, the perceived degree of urgency of a physician who received the page often differed from that of the nurse who sent the page (Figure 3). After sending a page to the on-call physician at night, nurses (mean, 72%; SD, 21%) were more likely to perceive that their page accurately portrayed the clinical situation than was the physician (mean, 48%; SD, 18%) who received the page (P = .004).

During telephone conversations with the on-call physician after a page, nurses routinely thought that the on-call physician appreciated neither the urgency (mean, 33%; SD, 22%) nor the complexity (mean, 33%; SD, 21%) of the particular clinical situation. More nurses (100%) than physicians (77%) either moderately or strongly agreed (P =.003), that they (ie, the respondents) do a good job in managing pain, agitation, and delirium when caring for patients at night in the ICU. However, nurses thought nearly a quarter of the time (mean, 23%; SD, 15%) that when they contacted the on-call physician at night about a patient’s pain, agitation, or delirium that the physician did not adequately address their concern.

More nurses (mean, 57%; SD, 22%) than physicians (mean, 34%; SD, 20%) perceived that after being paged by the nurse, the on-call physician sometimes delivered a new order to the bedside nurse without seeing the patient (P = .006). Among the new orders received from the on-call physician, nurses (mean, 55%; SD, 22%) thought they question a higher proportion of these orders than the on-call physicians (mean, 32%; SD, 24%) thought that the nurses did (P = .008). A similar percentage of nurses (mean, 53%; SD, 12%) and physicians (mean, 52%; SD, 21%) perceived that nurses provide a recommendation to the physician during the initial communication about a patient’s pain, agitation, or delirium (P = .82). However, during communication episodes in which a nurse delivered a recommendation to the physician, nurses (mean, 49%; SD, 20%) were nearly twice as likely to feel that the physician accepted their recommendation than was the actual physician (mean, 29%; SD, 22%) who received the recommendation (P = .004).

When asked about whether nurses make changes in medications for pain, agitation, or delirium before contacting the on-call physician (eg, administering a sedative at a greater rate or frequency than the rate or frequency that had been ordered), physicians (41%) were more likely than nurses (14%) to either moderately or strongly agree that this practice occurs (P = .008). Similarly, among nocturnal communications with a patient’s bedside nurse about pain, agitation, or delirium, on-call physicians (mean, 32%; SD, 21%) were more likely than nurses (mean, 21%; SD, 21%) to think that the nurse was seeking an order from the physician for something the nurse had already done (P = .03).

Several factors in the ICU at night influence how nurses and physicians work together to optimize care during this period.^27–30  Any factor that weakens communication between an ICU nurse and the on-call physician during this period can adversely affect patients’ safety and outcome.^1,4,7  Our study, the first evaluation of the perceptions and practices of ICU nurses and physicians about the communication they have with each other at night, in the context of pain, agitation, and delirium, raises a number of concerns. Although nurses and physicians communicate frequently with each other at night about patients’ pain, agitation, and delirium, the quality of this communication is often perceived as low. The method by which nurses and physicians prefer to triage common issues associated with pain, agitation, and delirium vary widely, and the process by which recommendations are exchanged between these 2 key groups of nighttime ICU caregivers when they are working together to resolve clinical issues differ substantially. Each of these factors should be explored using qualitative survey techniques.^10,23–25 

Although both MICU nurses who send a page and the on-call physicians who receive the page think that nurses use good clinical judgement when paging physicians about issues related to pain, agitation, and delirium, nurses and physicians often assign a different level of urgency to the same clinical situation. For example, far more nurses than physicians attribute a high degree of urgency to a patient who is agitated despite the administration of 1 as-needed sedative dose. The on-call physician perhaps assumes that the nurse will administer 1 or more additional doses before the nurse pages the physician about this issue, even though an order for the nurse to administer more sedative may not exist. Although almost all physicians stated they would immediately go to the MICU to manage a patient with respirations of 29/min and a decreasing arterial oxygen saturation, a substantial number of nurses stated that they did not expect the physician to do so. Possibly, the nurses would reach out to an in-house respiratory therapist to manage this situation before paging the on-call physician.

Of note, the urgency of a medication-induced coma (ie, a patient’s unresponsiveness to painful stimulus) was not considered either moderate or high among 50% of the nurses and 40% of the physicians despite multiple studies associating sedation-induced coma with both increased mortality and delirium and the establishment in both MICUs well before the survey of a sedation guideline advocating a moderate to light level of sedation and daily awakening.^31,32  Similar to findings in a recent MICU observational study,³³  the nurses in our study do think that deep sedation indicates oversedation. Why a page stating that a patient was confused and not following commands elicited a moderate to high level of urgency among nearly half the nurses and physicians but a page stating a patient had delirium (of which confusion and inattention are the hallmark symptoms) was only perceived as urgent by 10% of the clinicians in each group remains unclear.^21,31 

Reliance on an institutional text paging system when the on-call physician is away from the MICU is a major barrier to effective nurse-physician communication and triage of problems. Physicians clearly placed a premium on receiving a descriptive text page from nurses about a problem related to a patient’s pain, agitation, or delirium and were frustrated when a text page contained nothing more than the ICU telephone number. However, from the nurses’ perspective, delivery of the ICU telephone number may have simply reflected the inability of a nurse to leave the bedside of the nurse’s patient if the patient’s condition was unstable (eg, acute agitation) or the situation was too complex to be communicated to the on-call physician in a text page that allows just a few words.¹⁴ 

The limitations of text paging were reflected by nurses’ perception that one-third of the text pages they send to the on-call physician were not accurate and the on-call physician’s perception that half of the pages were not accurate. Of the pages sent by a nurse to the on-call physician when the nurse wanted the physician to come directly to the MICU to manage an issue related to acute pain, agitation, or delirium, the physician did not realize in 1 of 4 pages that the nurse was requesting the physician’s presence at the patient’s bedside. The finding that nurses often perceive that physicians appreciate neither the urgency nor the complexity of the situation that a nurse is trying to communicate via a text page also highlights the inefficiency in nurse-physician communication that results from using text messages to page an on-call physician at night.

New ICU paging protocols, including a situation-background-assessment-recommendation approach, may be a strategy for overcoming these issues.^9,34  More important is the need to develop and evaluate new communication technologies that will facilitate communication between nurses and the on-call physician when the physician is off the unit. In a recent study,³⁵  use of a hospital-based smartphone system that allowed 2-way nurse-physician communication and the transmission of patient data and pictures was compared with a traditional text-based paging system. The results indicated that use of the smartphone system improved communication efficiency between physicians, nurses, and allied health professionals. However, use of the system was associated with some drawbacks, including a greater incidence of technological failure, greater variability in perceived situational urgency, and more unprofessional behavior. Clearly, more research into innovative communication strategies that can replace traditional text-page systems is warranted.

Several factors might account for the high frequency of MICU nurses’ questioning of orders verbally received from on-call physicians related to patients’ pain, agitation, and delirium. A bedside nurse may have additional information about a patient (eg, earlier administration of a sedative did not work or caused safety concerns) that the nurse does not inform the physician about before the on-call physician contacts the nurse to provide a new order to manage the situation the physician was paged about. Additionally a nurse might have a greater knowledge than the on-call physician does of drug and nondrug strategies to manage pain, agitation, and delirium issues, particularly if the nurse has had extensive work experience in an ICU. Finally, the challenge facing the on-call physician and the bedside nurse in reaching a consensus for a plan when addressing a patient’s pain, agitation, and delirium may simply reflect the current lack of evidence to support many clinical practices in the area of ICU pain, agitation, and delirium.^{21,22,36–39} 

Our study has several limitations. Although the response rate was higher than that of many other surveys of health care providers, response bias still might have influenced the results.^40,41  All data were self-reported, no validation with practice was available, and verification of the accuracy of what was reported was not completed. The answers to some of the questions relied on the memory of the respondent, and thus recall bias might have occurred. On the basis of their clinical experience, respondents may have inferred that additional circumstances existed in one or more of the patient scenarios presented. The physicians and nurses who participated in the study all practice in the same MICU at the same academic center and thus may not represent the perceptions of nurses and physicians who work in other types of ICUs or at other hospitals where communication technology and protocols for treating pain, sedation, and delirium may be different. Last, we did not evaluate other factors that might have affected the quality of communication between on-call physicians and nurses in the ICU at night. These factors include the influence of different communication technologies, guidelines of the Accreditation Council for Graduate Medical Education for reduced duty hours for residents, and the effects of disruption in circadian rhythm on caregivers’ thought process and decision making.^42–44  The limitations highlight the importance of further exploring nighttime communication about pain, agitation, and delirium in a larger cohort of ICU nurses and physicians from multiple institutions and with additional investigative techniques such as qualitative survey methods.^23–25 

Despite possible limitations, our results provide the first extensive evaluation of ICU nurse-physician communication practices at night and suggest that many aspects of communication about pain, agitation, and delirium can be improved. Critical care nurses and physicians should pay close attention to the perceptual differences of members of the other profession when initiating communication to resolve patient care issues at night. New technologies are needed to foster better nighttime communication, particularly when the on-call physician is away from the ICU. Education programs that emphasize proven nurse-physician communication techniques should be developed to foster an ICU environment at night where interprofessional communication is optimized. Unit-wide interdisciplinary dissemination, expert modeling, and reinforcement should be a part of all quality improvement efforts in the area of ICU nocturnal communication. Last, the differences between the way that nurses and physicians interact with each other at night to resolve issues related to pain, agitation, and delirium should be considered when ICU procedures and protocols are being developed or modified.

Our results build on those of other investigations^5,6,13,19,25  that focused on the importance of team-work in the ICU. The ever-increasing evidence to guide ICU pain, agitation, and delirium practices will allow greater protocolization of practice in this area that should decrease the amount of nighttime communication between nurses and physicians about pain, agitation, and delirium. Furthermore, technological advances should facilitate the ability of a bedside nurse to provide an on-call physician (who is not present in the ICU) with the pertinent information in real time that will facilitate better quality decision making and decrease communication. Our results also highlight the importance of further qualitative and quantitative investigations on nocturnal ICU communication. Interventions to improve nighttime communication between ICU clinicians need to be developed, and the impact of the interventions on outcomes such as patient safety and quality of care must be determined.

The authors thank the critical care nurses and medical house staff at Tufts Medical Center who participated in the survey and Brenda Pun, msn, ccrn, Anne Pohlman, msn, ccrn, Richard Riker, md, and Joseph Dasta, msc, who provided invaluable feedback regarding the survey instrument.