We have come a long way but, I still have questions

In this month's edition of the journal, Chiem et al. present the results of a continuous opioid-sparing quality improvement initiative spanning 9 years. They report decreased perioperative opioid use with no negative impacts on pain or postoperative nausea and vomiting (PONV) in some 5600 children undergoing tonsillectomy.1 Throughout the evolution of surgical tonsillectomy techniques, visualization and speed represent key challenges which prompted innovations such as mouth-gags, guillotines, and drug development. The development of anesthetic agents, pharmacokinetic knowledge, intubation, and preformed curved tubes further contributed to the refinement of perioperative tonsillectomy care. One can make the case that such refinements represent multiple PDSA (Plan-Do-Study-Act) cycles across a centuries' long “tonsillectomy improvement project” that continues to play out on a daily basis in operating rooms around the world. Chiem et al have made an important contribution to the literature focused on determining the ideal anesthetic-management strategy for pediatric tonsillectomy. Methodological rigor is the invisible work which ensures a QI project can reliably address the clinical concern in question. Examples of such include: inviting early stakeholder involvement; best-evidence review; reliable data collection, analysis, and rapid turnover in outcomes reporting to facilitate timely interventions. The current publication asked the correct questions early in the planning phase and made these central to a long-term large-scale CQI (continuous QI) project that went through 16 different PDSA (Plan-Do-Study-Act) cycles. Those questions are addressed in the study's outcome measures: the percent of patients requiring intravenous opioids in postanesthesia care unit (PACU) post tonsillectomy (Primary Measure); Maximum pain scores in PACU, PONV, and discharge opioid prescription counts (Secondary Measures); and PACU length of stay (LOS) and rates of reoperation (Balancing Measures). Questions addressing quality of care have always been baked into the simple tonsillectomy: how to do it, how to do it quickly, how to do it well, how to do it safely, and how to do it better? To these questions, we have more recently added: who can we safely discharge home from recovery, who should we admit, and how long do we admit them for? The indications for tonsillar surgery have changed with many centers now performing significantly more tonsillectomies for sleep-disturbed breathing than recurrent tonsillar infections. We can reflect on the 2013 publication by Cote et al (“Death or neurologic injury after tonsillectomy in children with a focus on obstructive sleep apnea: Houston, we have a problem”2) as a transitional moment in how pediatric anesthesiologists currently consider and approach tonsillectomies. That publication, blessed with exquisite timing, warns how the scale is tipping from recurrent infection to obstructive sleep apnea (OSA), and the incidence of postoperative bleeding is ceding prominence to perioperative respiratory adverse events (PRAEs). The questions subtly change, and we are forced to ask in an age where we are spoiled-for-choice, which agents are safe, and which should anesthesiologists avoid? There is a further subtlety within that question, namely, if we over-correct toward safety and “drug-avoidance” is there a point where we are faced with a different type of patient safety question? For example, if we practice total avoidance of opioids (opioid-free protocols to avoid PRAE) alongside avoidance of nonsteroidal agents (to prevent postoperative bleeding), we run the risk of patients experiencing severe early pain (+/− emergence delirium) in PACU. Such problems result in rescue medications (often opioids) being administered in PACU and the reintroduction of PRAE-concerns. The subtle question that emerges—is it feasible to administer low-dose long-acting opioids intraoperatively to patients with mild to moderate obstructive sleep apnea undergoing tonsillectomy without introducing negative effects such as PONV, postoperative apneas and obstructions, increased PACU LOS and increased unanticipated ICU admissions—preempts a bigger concern for QI: what are the long-term effects of perioperative anesthesia protocols on patient-reported outcomes and surgical outcomes? Chiem et al are to be commended for anticipating and addressing both questions through appropriate QI methodology, rigorous data collection and processing, creative solutions, and their implementation thereof. Of course, such commitments may result in delays in bringing successful outcomes to publication; however, this ultimately moves the bar forward. Many QI publications currently describe 12–24-month projects with prolonged duration of each PDSA cycle due to low patient numbers. In addition, uncertainty remains regarding long-term feasibility of sustaining the reported improvements. The current publication addresses those concerns in full and moves the bar closer to par with large volume multicenter randomized controlled trials (RCTs). The issue of data acquisition in CQI recedes in the rear-mirror as more institutions adopt hospital-wide electronic health record (EHR) systems with fully integrated anesthesia information platforms. It is reasonably straightforward to create reports within an EHR so one might extract relevant data from surgical patients. For tonsillectomies, one could write a report to extract data from the initial surgical consultation (e.g., grade of tonsillar hypertrophy, clinical symptoms and signs), follow-up investigations (e.g., polysomnography [PSG]), preoperative anesthesia assessment (e.g., comorbidities, allergies, and suggestions for optimization), perioperative care (e.g., preoperative sedation, induction scores, desaturations and obstructions, and medications), and PACU (e.g., pain scores, PONV, medications, and postoperative disposition) in order to build a database relevant to current and future QI/research examining tonsillectomies. This data is readily available and its addition to QI methodology can reliably confirm or refute the effects of interventions on important metrics intended to improve meaningful patient-related outcomes. However, as patient numbers grow, so too does the volume of data we collect. In this study, Chiem et al implemented and employed a stand-alone data management software system which extracts patient-related data from EHR. This data is deidentified, aggregated, and updated in real time. It is presented as statistical process control charts which reliably measure success of intervention and are easily appreciated by clinicians irrelevant of QI-experience. The software described is commercially available but is not the only solution. A recently published alternative describes a less expensive, equally granular solution to data collection, collation, and reporting.3 In an era of Multicenter Perioperative Outcomes Group and The Children's Surgery Verification QI program, it is important we collect and report data rapidly so we might benchmark pediatric anesthesia management and improve quality of delivered care. With the advent of near-universal EHR, practitioners of pediatric anesthesia are at a tipping point where the questions we pose are changing—correction, they were always there—however, as we put older questions to bed, ongoing concerns emerge into plain sight. The current publication speaks to the experience of many but not all. It describes over 5600 mostly healthy patients in a single center. As an opioid-free protocol administered to otherwise healthy children expected to be discharged home from recovery, I would posit this as the new gold standard: Multiple PDSA cycles are well explained and feasibly implementable, the data are reliable and sustainable in the long term. However, the reported patient population likely differs significantly from other centers in terms of clinical acuity and comorbidities. The reported severity of OSA is severe (OAHI 10–15) but the patients are almost entirely ASA I-II status. There is no reporting of complex surgery for sleep-disturbed breathing (e.g., redo tonsillectomy, tongue reduction surgery, and turbinate reductions) with similar expectations for same-day discharge (likely because it is not done in the ambulatory institution where this study was performed). For these patients, an opioid-free protocol may not be the best anesthetic. Similarly, the “simple tonsillectomy” keeps changing, for example, co-incident flexible bronchoscopy to rule out asthma and upper endoscopy to rule out eosinophilic esophagitis, therefore, we should keep changing too. What is the “best anesthetic” in those circumstances? We should no longer be losing out on the complex patient's contribution to our knowledge and collective dataset simply because they are so complex as to be almost singular. Instead, we can (and should) collect collate and collaborate in this era of EHR and data management. This is neither criticism of Chiem et al's publication, nor caveat to its relevance; this is a signal boost: opportunities to marry EHR, data management, and CQI methodology, so we might deliver better quality of perioperative care and share our own long-term patient-reported outcomes. Does this swing the “cult of QI” closer to the “culture of research?” Might there come a time when QI methodologies, such as the one presented by Chiem et al, are submitted to national grant funding proposals and are not returned with the comment, “where is your power analysis?” How many patients, how many years would it take to answer just one of the questions addressed in this study via conventional RCT? A rigorous RCT design would likely mandate lack of PSG as an exclusion criterion, therefore significantly inhibiting patient recruitment. Inclusion/exclusion criteria exist to provide an important assurance, these patients are all the same; however, the quest for such guarantees creates two important gaps. The first is a lack of diversity in the patient population studied and an exclusion of those who may stand to benefit the most, that is, language comprehension, socioeconomic challenges and poor access to services such as preoperative PSG. The second gap ignores the possibility that these patients are similar enough that we can rigorously interrogate the integrity of current process of care and examine possible benefits of introducing an intervention into that process. As mentioned by Chiem et al, another group describe incorporating suprazygomatic maxillary nerve block as part of an opioid-free protocol, albeit with limited data.4 What is the best way to test this intervention? How many patients are required for such an RCT to be suitably powered, and which patients will be excluded along the way? Should we not instead implement the nerve block as the latest PDSA intervention cycle and add it to an ongoing QI project already 5600 patients strong? Might such data even be utilized in systematic reviews and meta-analyses? Is there a future where data such as that reported by Chiem et al is applied to a mixed-methodology review which includes both RCTs and QI studies? The authors of this publication are to be congratulated for applying CQI over many years in a rapidly evolving data management environment to address many of the important considerations that go into making a safe, pleasant tonsillectomy experience for our patients. These findings, especially when attached to the methodology reported, represent a significant step forward for patient care and, maybe, a guide to bridging the gap between CQI and “research.” Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.