Risk factors and events in the adult intensive care unit associated with pain as self-reported at the end of the intensive care unit stay

Critically ill patients in intensive care units (ICUs) are exposed to stressful conditions and experience discomfort from multiple sources, such as the environment or the treatment being provided, which depends on the organization (essentially the nurse-to-patient ratio) and the patient’s health status [1‐6]. Among these discomforts, pain is one of the main sources of psychological stress during an ICU stay and after hospital discharge [7, 8]. A program for reducing ICU patients’ discomfort has resulted in adult critically ill patients reporting decreased self-perceived discomfort at the end of their ICU stay. This program effectively reduced the overall discomfort scores and the scores for each discomfort item on the questionnaire “Inconforts des Patients de REAnimation” (IPREA), except those for pain [9].

Several hypotheses may explain why the program was ineffective for reducing pain scores. First, previously implemented health policies for pain management may have previously been be prioritized over managing other discomforts. Second, measures to reduce pain reported by critically ill patients may require more time to implement, because of complexity of pain management in adult critically ill patients due to highly individual pain patterns, and unique features in the ICU such as impaired communication and altered mental status, potentially painful procedures and use of invasive devices [10]. Moreover, a consistent and effective approach to pain assessment and management may require reorganization of care and redefinition of the roles of nurses and physicians in relieving procedural pain.

Better identification of procedures that cause the pain reported at the end of ICU stays would enable developing more effective strategies against pain in the ICU. Frequent ICU-associated procedures that can cause pain include arterial blood gas sampling and tracheal suctioning in mechanically ventilated patients, which lead to more frequent arterial catheter insertions and greater care for tracheal suctioning [2, 11]. Patients’ perceptions and responses to procedural pain were assessed after common ICU procedures such as turning in bed, wound drain removal, tracheal suctioning, femoral catheter removal, central venous catheter (CVC) insertion and nonburn wound dressing changes [12]. More recently, a prospective, multicenter study of pain intensity associated with ICU procedures identified chest drain removal, wound drain removal, and arterial line insertion as being associated with greater procedural pain intensity. The study also described risk factors associated with greater procedural pain in adult ICU patients [8]. Several studies have focused on preventing pain related to chest drain removal [13]. In past studies that focused on pain experienced by critically ill patients, pain was the only discomfort assessed. Pain was often rated immediately after the procedure and sometimes before the procedure, but not at the end of the ICU stay, which would allow considering the entire ICU stay and patients’ perceptions of possible repeated painful procedures. Moreover, no previous studies have evaluated procedures such as intra-hospital transport (e.g., for computed tomography). We proposed to identify the risk factors and events that have occurred during the ICU stay associated with pain-related discomfort perceived by critically ill patients during the whole ICU stay, as self-reported by critically ill patients at the end of their ICU stay, using a global ICU-related self-perceived discomfort tool and examining a large panel of frequent pain-inducing procedures, including intra-hospital transport, chest drain insertion, and chest drain removal, from a large patient population in a multicenter prospective study.

We used a large sample of 2130 French ICU patients treated at academic tertiary care and community hospitals to determine factors modulating pain-related discomfort during ICU stays. All the patients were previously included in the IPREA3 study, whose primary endpoint was the overall discomfort score derived from the French 16-item IPREA questionnaire. This first version of the IPREA questionnaire was validated (several years before the start of the IPREA3 study) through a process based on international guidelines using a large sample of critically ill patients hospitalized in different ICU types (medical, surgical, and mixed ICUs) at different institutions (university and nonuniversity hospitals) [15]. However, during the IPREA3 study, patients were questioned with a 18-item version of the IPREA questionnaire resulting from the addition of two discomfort items, item 17 and item 18 corresponding to dyspnea and feeling depressed, respectively, to the previous 16-item version. Since then, an ancillary study of the IPREA3 study yielded the psychometric validation of the 18-item version of the IPREA questionnaire [16].

To our knowledge, this was the first study exploring the weight of a large panel of potentially pain- or discomfort-inducing procedures or treatments used to treat critically ill patients.

The four procedures undergone by patients during the ICU stay that were associated with greater pain-related discomfort were chest drain removal, bladder catheter use, CVC insertion, and intra-hospital transport.

Chest drain-related procedures (insertion or removal) seemed to more strongly affect pain perception during ICU stays than did most other procedures considered invasive or discomfort-inducing (e.g., arterial catheter insertion or bronchoscopy). It is also important to note that no specific mandatory measure to reduce pain associated with either chest drain insertion or chest drain removal has been implemented during the study period in the participating ICUs. Previous studies that found chest drain-related procedures to be painful [21, 22] often did not differentiate between insertion and removal times. We showed that the two procedures did not induce the same pain-related discomfort: chest drain removal appeared to be more painful than chest drain insertion. This information may help intensivists decide the timing of analgesics prior to removing a chest drain [23]. Greater attention should be given to pain relief before and during this procedure, leading to the use of previously described nonpharmacological methods [24‐26]. Of note, critically ill surgical patients may be admitted to the ICU with a chest drain having been previously inserted in the operating room under general anesthesia and therefore will not require this procedure in the ICU. Thus, chest drain removal is more common than chest drain insertion during ICU stays. However, chest drains may be left in place during transfers from the ICU to the hospital surgical ward. Because of the large number of surgical patients, the number of patients who underwent at least one chest drain removal during their ICU stay was more than double that of patients who had at least one chest drain inserted. Remarkably, of the 2130 patients included here, many more patients underwent at least one chest drain removal under spontaneous ventilation or NIV than under MV.

Intra-hospital transport during ICU stays was also associated with self-reported pain-related discomfort. Intra-hospital transport of critically ill patients is a well-known risk factor for unexpected adverse events and complications, such as pulmonary complications, hemodynamic complications, infection, interruption of critical drugs, etc. [27‐31], but an association between intra-hospital transport and higher pain-related discomfort scores being reported at the end of the ICU stay is surprising. We did not record the reasons for intra-hospital transport. We hypothesize that most patients were transferred to the radiology department for computed tomography scans, and moving patients from the critical care bed to the examination table may be comparable to turning them in bed, which was previously reported by critically ill patients to be one of the most painful procedures [12, 32].

As previously reported, younger people were also independently associated with higher self-reported pain-related discomfort scores compared with those of older people [33].

Sex, patient type (i.e., medical vs. surgical), prior health status, and ICU stay duration were not independently associated with pain scores.

We used a comprehensive approach because the patients interviewed at the end of the ICU stay were not specifically informed of precisely what happened to them during their ICU stay before the discomfort assessment. Moreover, the interviewed patients were not asked if they remembered having undergone a particular procedure but only to score each of the 18 discomfort items. For example, patients may be unaware of having a chest drain inserted or withdrawn, a bladder catheter, a CVC insertion or being transported outside the ICU at any time during their stay. The healthcare team did not control or influence patients’ memories, thus reducing any suggestive bias. Moreover, the patients were not interviewed immediately after the potentially painful event as is the case in most studies aiming to assess or prevent procedural pain [34].

Our study had several limitations. First, even if the bedside nurse asked patients to rate the severity of each discomfort for the entire stay in the ICU, patients may have reported discomforts, including pain, experienced only during the final part of the ICU stay if they could not remember what they experienced at the beginning of their ICU stay. Thus, events occurring during the last days of their ICU stay, such as chest drain removal after the clinical situation improved, may have had more influence on the pain-related discomfort score than did events that occurred at the beginning of their ICU stay, such as arterial catheter insertion or bronchoscopies. Future studies should record and analyze the time course of all procedures and life support therapies undertaken during the ICU stay. Second, all discomfort item scores, particularly the pain-related discomfort scores, were obtained with the help of a nonblinded bedside nurse (or assistant nurse), which could introduce bias. We accepted that conducting a pragmatic trial to explore potential benefits in the ICU with such an easily reproducible program would require participation from bedside nurses. Third, we could not assess the link between potential painful or discomfort-inducing procedures or life support therapies and the overall ICU-induced pain burden experienced in nonsurvivors during their ICU stay because ICU nurses administered the IPREA only to survivors on the day of discharge. This limitation also applied to patients who could not be questioned such as those with diminished mental capacity, those with language barriers, those who were transferred to another ICU while mechanically ventilated, or those who were urgently discharged. The issue of representativeness of this sample of ICU survivors among the general population of critically ill patients remains uncertain. Fourth, we could not analyze the impact of analgesic or sedative drug dosages or nonpharmacological therapies used during the procedures or the implementation of life support therapies. Fifth, we did not record details concerning chest drain-related procedures [35, 36]. Drain size (e.g., indwelling small catheter, small-bore chest drain, large-bore chest drain), use of premedication and/or local anesthesia, confirmation of the insertion site via ultrasonography or needle suction, chest wall dissection (by finger, curved hemostat or clamp, or trocar), and parietal refection and sutures (none, “mattress” suture, or “purse string” suture) would enable better understanding the self-perceived pain. However, the goal of the study was not to determine the exact role of any factor characterizing each potential painful or discomfort-inducing procedure nor was it to propose recommendations for preventing pain or discomfort when performing these procedures. Based on the standard procedures and treatments used in the ICU, our study constituted a first approach to detect those likely to remember pain during their ICU stay to design further studies to understand the involved mechanisms in greater detail. Sixth, recent recommendations on variable selection have been proposed [37]. However, we assumed that our approach to identify variables associated with recalled pain, based on consensual and well-recognized selection procedure, remains robust and valid. Finally, we acknowledged that our study was not designed to assess and distinguish recalled pain associated with procedures from recalled pain not associated with them. We only sought an association of the achievement of some commonly used procedures in the ICU with pain experienced during the entire ICU stay as it was self-reported at the end of the ICU stay, whether pain was exacerbated—or caused—by procedures or whether pain was already present between the procedures due to other causes.

However, despite these limitations and without being able to affirm the causal link between these events and the painful memorization of the ICU stay at discharge, these first results showing an association between the occurrence of such potentially painful procedures and the pain-related discomfort score should encourage us to properly assess the balance of the benefits and risks of these common procedures such as maintenance of a bladder catheter, intra-hospital transport for an iterative CT scan, CVC insertion or thoracic drainage.

Our study showed that adult ICU patients who underwent chest drain removal, bladder catheter, CVC insertion, and intra-hospital transport during their ICU stay reported higher pain-related discomfort scores (regarding the entire ICU stay and assessed at the end of the ICU stay) than patients who did not experience these events. This study may pave the way for further targeted studies aiming at investigating a causal link between these common procedures undergone by adult critically ill patients during their ICU stay and recalled pain at ICU discharge.