Context

This quality improvement project was conducted between October 2021 and January 2024 at an inpatient and outpatient pediatric sedation clinic in an urban academic quaternary care children’s hospital, wherein the children’s services occupy a separate but attached pediatric hospital, and MRI scanners and technologists also serve adult patients in the neighboring university hospital. Our clinic provides minimal, moderate, and deep (i.e., propofol) sedation defined as a medically induced alteration in consciousness with the intent to maintain a natural airway. Patients deemed at risk of airway instrumentation with the administration of sedatives are usually referred to GA. The project followed Plan-Do-Study-Act (PDSA) cycles: an iterative problem-solving model used for improving health care processes. The cycle begins with making predictions about the outcome (Plan), conducting the plan and collecting data (Do), comparing the predictions to the data collected after specific interventions (Study), and acting based on the new knowledge, validating whether the hypothesis for improvement is correct (Act) [14]. The sedation (physicians, schedulers, certified child life specialist, and nurse manager) and radiology teams (physicians, radiology manager, and MRI technologists) were involved in these cycles. Patients referred to the pediatric sedation clinic for a sedated MRI were screened for the study. Patients with visual impairment, severe developmental delay, or severe autism spectrum disorder that would preclude understanding of the AVD technology or tolerance of the confined MRI space were excluded from the study. The inclusion criteria for this study changed, broadening successively for diagnosis, age, and scanning time, with each PDSA cycle undertaken by the stakeholder team. This project followed the Revised Standards for Quality Improvement Reporting Excellence (SQUIRE 2.0) and was exempt by the local institutional review board [15].

Planning the intervention

In October 2021, a multidisciplinary stakeholder team, including the sedation clinic director and radiology personnel, convened to develop a new workflow for an awake MRI program (Fig. 1 and Table 1).

Fig. 1Table 1 Detailed description of Plan-Do-Study-Act cycles undertaken by stakeholder team

Various MRI AVD technologies were evaluated in terms of cost and convenience, and a decision was made to trial an open-bore technology (MRI in-bore video system, PDC Inc. Hartland, WI) for the MRI scanner located in the pediatric hospital. This technology consists of a video projector located outside the MRI bore that projects a movie onto the upper inner surface of the bore. This technology was suitable only for head-first positioning in the MRI scanner, the most common study type performed at the imaging center located in the pediatric hospital. The AVD technology was installed and available for clinical use at the children’s hospital MRI scanner on January 27, 2022.

Targeting the intervention (PDSA cycle 1)

The patients receiving minimal and moderate sedation (i.e., non-propofol sedation) were targeted for the awake MRI program because patients receiving propofol (deep sedation) usually have anxiety or are developmentally unable to cooperate with the use of the AVD technology. The team created an “awake MRI” workflow to outline the criteria to use the AVD technology in patient care (Fig. 1). Patients referred to the sedation clinic for a sedated MRI were identified as eligible candidates by the sedation staff for our awake MRI program before arrival at the clinic, by review of clinical information and direct communication with the family by phone. At the clinic visit, on the day of the sedation/MRI exam, a CCLS, a nurse, and a sedation provider performed an in-person assessment on all eligible patients to trial awake versus sedated MRI (Fig. 1 and Table 2). Patients who failed an awake MRI were immediately sedated at the same-day appointment.

Table 2 Certified child life specialist (CCLS) assessment for pediatric awake versus sedated MRI

Criteria to use AVD were age 7 years and older, diagnosis of central nervous system tumor, needing follow-up brain MRI scans, head-first MRI position only, less than 60 min scan duration, and MRI scheduled at the children’s hospital. This was primarily to maximize the diagnostic quality of MRI imaging and to minimize the likelihood of a patient needing to convert to a sedated scan, thus prolonging the allotted exam time, defined as scan time plus 15 min to account for patient and room set-up.

Broadening the scope for diagnosis (PDSA cycle 2)

The sedation team (sedation provider, nurse, and CCLS) noted that their in-person assessments of patients were yielding more patients who would likely benefit from AVD than the initial narrow criteria allowed. Therefore, the criteria to use AVD were broadened to cases for non-CNS indications. However, patients with visual impairment, severe developmental delay, or severe autism spectrum disorder that would prevent understanding of AVD technology and tolerance of confined MRI space were still excluded from AVD use. This change in diagnostic criteria allowed the use of AVD for any type of head-first MRI with a scan duration of less than 60 min, scheduled at the children’s hospital.

Broadening the scope for age (PDSA cycle 3)

This PDSA cycle reduced the restrictions for age to any patient 4 years and older who was found to be cooperative, calm, and likely to remain still without sedation on the day of their MRI based on CCLS, nurse, and physician in-person assessment.

Increasing likelihood of a diagnostic study (PDSA cycle 4)

This PDSA cycle removed restrictions on scan duration, allowing examinations of any length. In addition, alongside the ongoing work to reduce sedation for MRI scans in the summer of 2022, the pediatric neuroradiology team developed an array of accelerated MRI protocols that reduced both the time to acquire each sequence and the total time spent in the MRI scanner, thus optimizing the chance for success. Several complex protocols, however, remained, involving detailed imaging that required longer scan and sequence times, particularly for patients undergoing evaluation for seizure focus resection. Because of the importance of these images for surgical planning, and because individual sequences in these protocols could stretch beyond 10 min in duration, the pediatric neuroradiology team was very interested in minimizing the number of patients with motion artifact on these exquisitely motion-sensitive studies. In the spring of 2023, they developed and distributed a motion sensitivity score (Table 3) awarding each protocol a “difficulty” score based on scan parameters by which the sedation team could triangulate the motion sensitivity score of the MRI protocol alongside a bedside assessment of the patient’s temperament and cooperability on the day of their appointment. After these scoring systems were integrated into the sedation workflow, they were integrated by the pediatric radiology department supervising MRI protocols for pediatric patients, unifying the use of the motion sensitivity score, which was also used to help decide on patients that were appropriate for AVD. In addition, an average of 30 min was added to the allotted exam time at the time of scheduling to allow sedation of patients who failed an awake MRI due to motion artifact.

Table 3 Pediatric MRI exams, allotted exam times, and assigned motion sensitivity scores. Scanning times include a 15-min allowance for room set-up prior to scanningReinforcing the use of AVD at the bedside (PDSA cycle 5)

Although the study team was able to remove barriers that limited patients’ access to AVD during previous PDSA cycles, it was noted in the spring of 2023 that the percentage of patients placed in the AVD pathway had fallen. Sedation team leaders and pediatric radiologists convened and shared run chart data with the CCLS and sedation nurses, directing their attention to very high success rates of patients distracted with AVD alongside the unexpectedly decreased rates of use of this tool in recent months. A culture change was encouraged to reinforce the value of awake AVD MRI among stakeholders.

In September 2023, the MRI in-bore video system, PDC Inc. Hartland, WI, was replaced with a new open-bore technology (MRI view, MRI Audio, Carlsbad, CA). This technology consisted of a self-standing mirror and projector and allowed the patient to watch a movie during head-first and feetfirst body position MRIs, which expanded the availability of AVD during MRI scanning.

Study of the improvement

The quality improvement team continually reviewed prospectively collected data of all patients 4 years and older who were referred to the sedation clinic for MRI scanning at the pediatric hospital. We reviewed patient demographics, diagnosis, type of MRI, location, sedative use, AVD technology use, MRI duration compared to allotted exam time, defined as scan time plus 15 min to account for patient and room set-up, and study outcome (i.e., diagnostic or non-diagnostic). The data were reviewed monthly by the multidisciplinary team, and any changes were discussed and approved by all stakeholders.

Measures

The primary outcome measure was the percentage of referrals who completed an awake MRI with AVD and without the need for sedatives, analyzed on a statistical process control (SPC) P-Chart, a type of control chart where each item can be categorized as either “pass” or “fail”—typically when subgroup sizes vary across data collection points [14]. This percentage was calculated using “all eligible patients for AVD” (i.e., minimal and moderate sedations for MRI and awake patients undergoing MRIs with AVD without sedatives) as the denominator and awake patients undergoing MRIs with AVD without sedatives as the numerator. Patients who received propofol sedation for MRI were excluded from the calculation of the primary outcome because they usually had developmental or behavioral statuses that would prevent their success with awake MRI. The process measure to track compliance with our awake MRI program workflow was the number of patients deemed eligible for AVD per month by inclusion criteria, analyzed on an SPC C-Chart, a type of control chart used to count items in a process at regular intervals [14]. A diagnostic quality awake MRI was defined as a study that did not require sedation to achieve diagnostic images. This was determined by the interpreting radiologist at the time of imaging, and as part of this quality improvement project, this was recorded in each dictation report’s technique section for all pediatric imaging studies during the study period, regardless of whether the exam was done with sedation or while awake. The balance measures, the metrics used to assess the unintended consequences of changes with the implementation of our awake MRI program, were the number of MRIs that were not of diagnostic quality or which required longer than the allotted exam time (Table 3). We chose this balancing measure because unanticipated MRI delays that prolong allotted exam time can significantly impact MRI workflow. Exam beginning and ending times were acquired from routine records created by the MRI technologists in the electronic health record.

Caregiver and patient satisfaction data were collected in a subset of patients following the conclusion of an awake MRI appointment. Patients were verbally asked, in language appropriate for their age, to rate whether their experience was “easy” (i.e., high satisfaction), “medium” (i.e., moderate satisfaction), or “hard” (i.e., low satisfaction). Caregivers were asked to rate their level of satisfaction on a Likert scale, with “1” representing “very dissatisfied” and “5” representing “very satisfied.”

Analysis

Outcome and process measures were assessed with statistical process control chart methodology. Statistical process control charts were constructed with QI Charts (version 2.0.23; Performance Improvement Products, Austin, TX) and Microsoft Excel (version 16.73; Microsoft Corporation, Redmond, WA). Processes that are “in control” have stable and predictable behavior over time, and they are indicated by a stable centerline, which depicts the mean of the data over time. When a system or a process changes, either by project interventions or by external forces, the process is said to be no longer “in control,” and the centerline shifts to accommodate a new mean. Special cause variations are causes of variations in a process that are not inherent to the process. Means were shifted when corresponding to interventions that signaled special cause variation. Any of the following rules indicated special cause variation in SPC charts: a single data point outside of the control limits, eight consecutive points above or below the mean line, or a trend of six consecutive points all moving in the same direction, and two out of three points falling in the outer third of a control limit [16]. Continuous characteristics of the MRI referrals were summarized in terms of the range, mean, and standard deviation, or the median and interquartile range (IQR); categorical factors were described using frequencies and percentages. Data were described separately according to group (awake or sedated during the MRI), and comparison of characteristics between groups was made using t-tests, rank-sum tests, or chi-square tests. A P-value of less than 0.05 was considered statistically significant.