In the femoral cohort, site issues and inaccuracies were the causes of two of the removals, and inability to draw was the cause for removal in a single patient. Accidental removal did not occur in the femoral cohort. Mortality data was only monitored in the prospective arm and included only patients with an arterial line in place at the time of death or patients who died shortly after line removal as we did not follow patients beyond line removal. There were 11 deaths in the femoral cohort, with none occurring in patients with prior femoral line failures. There were seven deaths in the radial cohort, with three of those occurring in patients with prior radial line failures. Of the three deaths attributed to patients with radial line failures, one of these failed lines was replaced in the other radial artery , one was replaced in a femoral artery , and one was not replaced.To our knowledge, this is the first study to compare failure rates by arterial line site in the last 35 years. Soderstrom et al. in 1982 also showed differences in “placement duration” and “longevity” when comparing arterial line sites, with results favoring the femoral site. Their data was similar to ours in that it showed a femoral failure rate of 10.6% compared to 26.4% for radial sites, with failures occurring on average 3.5 days sooner in their radial cohort.Our data likewise demonstrate that femoral arterial lines fail at a significantly lower rate than radial arterial lines. Despite making up 34% of the lines placed,cannabis indoor greenhouse they accounted for only 8.6%of the lines that failed.
This difference was consistent in both the retrospective and prospective arms of the study, exhibiting agreement between the two data groups. This gives a number needed to treat of only four patients: four patients preferentially receiving a femoral arterial line prevents one premature line failure. We believe that no meaningful conclusion can be drawn from the data regarding mortality, as we did not follow patients beyond removal of their arterial lines. Additionally, the comorbidity data is likely incomplete given the low prevalence of classical diseases in this critically ill population. This is likely due to limitations of our retrospective review, including incomplete charting and a lack of emphasis on this data during collection. This study was also not designed to evaluate the reason for a provider’s site preference. At the institutional level, radial lines are preferred over femoral, and this likely instilled selection bias, though to what extent is unclear. If radial lines are preferred by default, then femoral lines might have been placed in cases where there were radial site issues or in patients with higher acuity. Femoral lines might also have been placed after multiple radial attempts failed, or even after a placed radial line failed in the first 24 hours, which would not have been caught by our study. These factors should have selected for a sicker population in the femoral cohort, though this did not lead to increased line failure rates. An alternative argument could be made that because the femoral cohort were sicker, they may have gotten more attention by nursing staff and thus better line care. The possibility of this effect is mitigated, however, by the prospective arm: in this phase nursing staff was aware of the study and yet the line failure rates remained consistent to the retrospective data in each site. The length of time in which the lines failed also favors femoral line placement. Of the failed arterial lines, the average times of failure for the femoral and radial cohorts were four and two days after placement, respectively. Still, not all radial lines in our study failed within two days, with some radial lines lasting 17 days; however, it is impossible to know which lines are going to fail in advance, and our data indicates that radial lines on average fail earlier in the course.
Although we would have liked to draw conclusions as to why lines failed, our study with only 58 total failures was not adequately powered to draw meaningful conclusions in this area. The ambispective design also makes it difficult to compare causes of failure between the two groups due to the ambiguity of etiology in the retrospective charting. We attempted to account for this in the prospective arm by having the RAs determine the cause of removal from the nursing staff that had direct care of the patient. Our data between the two groups is comparable in this respect, adding strength to our retrospective methods. Furthermore, the length difference between the arterial lines could have an effect on line longevity, but this does not change the fact that femoral lines failed less often. Longer, wider catheters in the radial artery are likely to increase complications, given the increasing risk for ischemia at this site. Our study suggests that femoral lines were well tolerated with minimal infectious risk. Only one of 93 femoral arterial lines were removed for “site issues”. Note that in designing this study, we wanted to be broad in our attribution of “infection” and thus specified a broader category of “site issue” as a cause of line failure. In retrospect, it would be favorable to define why this one patient’s line was removed. This particular line failed after six days and was replaced with another femoral line. Patients who died early in their disease course may not have had enough time for the line to fail, and this could have led to dilution of the failure rates among sites. Additionally,cannabis growing equipment lines that were not adequate at insertion or were tenuous would increase the number of failures inappropriately as they were likely to get replaced quickly. The retrospective cohort is at greatest risk of being affected by these confounders as providers were unlikely to add additional billing codes for lines replaced rapidly in the same day. We attempted to control for both issues by only including lines surviving greater than 24 hours.
The prospective data matching the retrospective rates also gives confidence that we monitored true failure rates. This was a single-center study performed at a county teaching hospital with only 12 adult ICU beds, which partially accounts for the small volume of patients over the four-year period of the study. Additionally, our retrospective arm relies on a billing database, which certainly does not capture all lines placed during this period. Several factors would contribute to loss of capture, though we suspect, based on observed practice, that lack of appropriate billing code assignment and documentation of placement in individual procedure notes accounted for the majority of lost lines in the database. The prospective arm remedies this in that RAs observed every bed that might have held a patient with an arterial line three times daily. We believe this is why there were 76 patients captured in the nine months of the prospective arm , and only 196 patients captured in the 42 months of the retrospective arm. The external validity of our study may be limited due to the size of our hospital and ICU; however, we would argue that these are in fact the locations where line failures can be the most detrimental, straining a system already stretched thin. Given the observational nature of our study, we were not immune to selection bias and it is impossible to determine why one site was chosen over another site by each individual provider. We noticed that more often, patients were given a femoral line as a rescue from a radial line failure, and thus the lines may have been placed at times when the patient’s illness was further along and possibly improving.Line site preference did change , which may be due to a change in local culture, a change in perception of femoral line risk given new literature, more complete capture of all lines placed, or even a Hawthorne effect. It is also likely that many lines were not captured in our retrospective arm as stated above. A better understanding of the differences between the patient groups , as well as the reasons why a site was chosen, was not captured in our study but would be interesting to study in another, larger trial. Patient care hand offs are a potentially vulnerable time for patient safety.Sub-optimal handoff communication is a common cause of medical errors and malpractice claims.Recently, there has been increased focus on inter-unit hand offs, which occur when patients are transitioned between services, departments, or institutions.9 The inter-unit admission handoff between emergency physicians and inpatient providers is a particularly important example. The emergency department admission process involves changes in the healthcare team and physical location of the patient.10 Unstructured communication, inter-disciplinary conflict, patient throughput pressures, and uncertain assignment of responsibilities may further impede safe care transitioning from the ED to inpatient setting.Survey studies have found that one-third of physicians know of adverse patient events related to the admission handoff process.Standardized handoff communication tools have been shown to improve outcomes for inter-unit hand offs,but they have not been widely implemented for admission hand offs. A survey of 750 physicians at 10 sites found that only 18% of EPs and internal medicine physicians used a standardized admission-handoff tool and only one-third of residents received handoff training.Although many handoff mnemonics exist in the literature,Situation, Background, Assessment, and Recommendation is the most commonly used,18 and is promoted by regulatory and professional organizations, including the Agency for Healthcare Research and Quality and Institute for Healthcare Improvement.In 2012 Beach et. al. published best practice recommendations for ED-to-inpatient handoff communication, including style, form and content. They suggested synchronous, two-way, closed-loop communication, with the goal of constructing a shared mental model of patient care between EPs and IM providers. Rather than rote recitation of data, it was suggested the content of handoff communication should focus on clinical judgment, diagnostic uncertainty, the patient’s clinical trajectory, pending tasks, and any patient- or system-level considerations that may impact care.A study in 2016 found that a standardized method of handoff for patients admitted to a geographically isolated, 35-bed community hospital from the ED resulted in fewer physician-reported “defective” hand offs;however, it is unknown if these findings are applicable to larger academic health centers that have unique complexities, including multiple admitting services with variable processes and trainees at various levels of training.13 The goal of this study was to pilot test a standardized process to improve admission handoff communication between EPs and IM physicians by integrating best-practice recommendations with a modified SBAR format, resulting in the Situation, Background, Assessment, Responsibilities & Risks, Discussion & Disposition, Read-back & Record model.We hypothesized use of SBAR DR would improve the quality of inter-unit handoff communication for patients admitted from the ED.We found that the introduction of a standardized handoff process for patients being admitted from the ED to hospital setting resulted in improvements in verbal handoff quality. The driver of improvement was primarily due to improvements in opportunities for questions and reaching unambiguous agreement regarding patient disposition. Interactive questioning during hand offs is recommended by regulatory agencies and practice guidelines.Not only does this support clarifications and error-correction, but it also facilitates anticipatory guidance, reframing of the clinical picture, and creation of a shared mental model of patient care.Explicit agreement in disposition plan was also an important improvement. Uncertain assignment of responsibility is a known barrier to safe care transitions.EPs and IM physicians are often uncertain when patient care is definitively transferred, especially for a patient boarding in the ED. This leaves the patient in limbo, during which time nursing and ancillary staff do not know where to direct concerns about changes in clinical trajectory or care needs. As part of our intervention, we explicitly tied the disposition decision and assignment of patient care, in which the admitting team assumed responsibility when an admission order was placed. This decision was then clearly delineated in the handoff template, where it was visible to all members of the patient care team. As a result of using the SBAR-DR strategy, over half of surveyed physicians reported personally experiencing improved patient safety during the 60-day study period.