Machine learning approach to predict postoperative opioid requirements in ambulatory surgery patients.
ABSTRACT: Opioids play a critical role in acute postoperative pain management. Our objective was to develop machine learning models to predict postoperative opioid requirements in patients undergoing ambulatory surgery. To develop the models, we used a perioperative dataset of 13,700 patients (? 18 years) undergoing ambulatory surgery between the years 2016-2018. The data, comprising of patient, procedure and provider factors that could influence postoperative pain and opioid requirements, was randomly split into training (80%) and validation (20%) datasets. Machine learning models of different classes were developed to predict categorized levels of postoperative opioid requirements using the training dataset and then evaluated on the validation dataset. Prediction accuracy was used to differentiate model performances. The five types of models that were developed returned the following accuracies at two different stages of surgery: 1) Prior to surgery-Multinomial Logistic Regression: 71%, Naïve Bayes: 67%, Neural Network: 30%, Random Forest: 72%, Extreme Gradient Boost: 71% and 2) End of surgery-Multinomial Logistic Regression: 71%, Naïve Bayes: 63%, Neural Network: 32%, Random Forest: 72%, Extreme Gradient Boost: 70%. Analyzing the sensitivities of the best performing Random Forest model showed that the lower opioid requirements are predicted with better accuracy (89%) as compared with higher opioid requirements (43%). Feature importance (% relative importance) of model predictions showed that the type of procedure (15.4%), medical history (12.9%) and procedure duration (12.0%) were the top three features contributing to model predictions. Overall, the contribution of patient and procedure features towards model predictions were 65% and 35% respectively. Machine learning models could be used to predict postoperative opioid requirements in ambulatory surgery patients and could potentially assist in better management of their postoperative acute pain.
Project description:BACKGROUND AND OBJECTIVES:Percutaneous peripheral nerve stimulation (PNS) is an analgesic modality involving the insertion of a lead through an introducing needle followed by the delivery of electric current. This modality has been reported to treat chronic pain as well as postoperative pain the day following knee surgery. However, it remains unknown if this analgesic technique may be used in ambulatory subjects following foot procedures beginning within the recovery room immediately following surgery, and with only short series of patients reported to date, the only available data are derived from strictly observational studies. The purposes of this proof-of-concept study were to demonstrate the feasibility of using percutaneous sciatic nerve PNS to treat postoperative pain following ambulatory foot surgery in the immediate postoperative period and provide the first available data from a randomized controlled study design to provide evidence of analgesic effect. METHODS:Preoperatively, an electrical lead (SPRINT; SPR Therapeutics, Inc, Cleveland, Ohio) was percutaneously inserted posterior to the sciatic nerve between the subgluteal region and bifurcation with ultrasound guidance. Following hallux valgus osteotomy, subjects received 5 minutes of either stimulation or sham in a randomized, double-masked fashion followed by a 5-minute crossover period and then continuous stimulation until lead removal on postoperative days 14 to 28. RESULTS:During the initial 5-minute treatment period, subjects randomized to stimulation (n = 4) experienced a downward trajectory in their pain over the 5 minutes of treatment, whereas those receiving sham (n = 3) reported no such change until their subsequent 5-minute stimulation crossover. During the subsequent 30 minutes of stimulation, pain scores decreased to 52% of baseline (n = 7). Three subjects (43%) used a continuous popliteal nerve block for rescue analgesia during postoperative days 0 to 3. Overall, resting and dynamic pain scores averaged less than 1 on the numeric rating scale, and opioid use averaged less than 1 tablet daily with active stimulation. One lead dislodged, 2 fractured during use, and 1 fractured during intentional withdrawal. CONCLUSIONS:This proof-of-concept study demonstrates that percutaneous sciatic nerve PNS is feasible for ambulatory foot surgery and suggests that this modality provides analgesia and decreases opioid requirements following hallux valgus procedures. However, lead dislodgement and fracture are concerns. CLINICAL TRIAL REGISTRATION:This study was registered at Clinicaltrials.gov, identifier NCT02898103.
Project description:Importance:Opioids are routinely prescribed for postoperative home pain management for most patients in the United States, with limited evidence of the amount needed to be dispensed. Opioid-based treatment often adversely affects recovery. Prescribed opioids increase the risk of chronic opioid use, abuse, and diversion and contribute to the current opioid epidemic. Objective:To evaluate whether after hospital discharge, postsurgical acute pain can be effectively managed with a markedly reduced number of opioid doses. Design, Setting, and Participants:In this case-control cohort study, an ultrarestrictive opioid prescription protocol (UROPP) was designed and implemented from June 26, 2017, through June 30, 2018, at a single tertiary-care comprehensive cancer center. All patients undergoing gynecologic oncology surgery were included. Patients undergoing ambulatory or minimally invasive surgery (laparoscopic or robotic approach) were not prescribed opioids at discharge unless they required more than 5 doses of oral or intravenous opioids while in the hospital. Patients who underwent a laparotomy were provided a 3-day opioid pain medication supply at discharge. Main Outcomes and Measures:Total number of opioid pain medications prescribed in the 60-day perioperative period, requests for opioid prescription refills, and postoperative pain scores and complications were evaluated. Factors associated with increased postoperative pain, preoperative and postoperative pain scores, inpatient status, prior opioid use, and all opioid prescriptions within the 60-day perioperative window were monitored among the case patients and compared with those from consecutive control patients treated at the center in the 12 months before the UROPP was implemented. Results:Patient demographics and procedure characteristics were not statistically different between the 2 cohorts of women (605 cases: mean [SD] age, 56.3 [14.5] years; 626 controls: mean [SD] age, 55.5 [13.9] years). The mean (SD) number of opioid tablets given at discharge after a laparotomy was 43.6 (17.0) before implementation of the UROPP and 12.1 (8.9) after implementation (P?<?.001). For patients who underwent laparoscopic or robotic surgery, the mean (SD) number of opioid tablets given at discharge was 38.4 (17.4) before implementation of the UROPP and 1.3 (3.7) after implementation (P?<?.001). After ambulatory surgery, the mean (SD) number of opioid tablets given at discharge was 13.9 (16.6) before implementation of the UROPP and 0.2 (2.1) after implementation (P?<?.001). The mean (SD) perioperative oral morphine equivalent dose was reduced to 64.3 (207.2) mg from 339.4 (674.4) mg the year prior for all opioid-naive patients (P?<?.001). The significant reduction in the number of dispensed opioids was not associated with an increase the number of refill requests (104 patients [16.6%] in the pre-UROPP group vs 100 patients [16.5%] in the post-UROPP group; P?=?.99), the mean (SD) postoperative visit pain scores (1.1 [2.2] for the post-UROPP group vs 1.4 [2.3] for pre-UROPP group; P?=?.06), or the number of complications (29 cases [4.8%] in the post-UROPP group vs 42 cases [6.7%] in the pre-UROPP group; P?=?.15). Conclusions and Relevance:Implementation of a UROPP was associated with a significant decrease in the overall amount of opioids prescribed to patients after gynecologic and abdominal surgery at the time of discharge for all patients, and for the entire perioperative time for opioid-naive patients without changes in pain scores, complications, or medication refill requests.
Project description:Primary cleft palate repair may result in significant pain in the immediate postoperative period, which can lead to vigorous crying resulting in wound dehiscence and pulmonary complications. Effective pain control with opioids is the mainstay but administration on the floor has to be countered with the complications associated with their use, chiefly respiratory depression and sedation. We retrospectively examined the efficacies of intraoperative administration of intravenous (IV) dexmedetomidine (DEX) and ketamine (KET) to prevent early postoperative pain in children undergoing primary cleft palate repair and compared the results against relevant literature. The Texas Children's Hospital anesthesia database was queried to identify children undergoing a palatal surgery from December 2011 to December 2012. Inclusion criteria permitted completed primary palatal surgery without major complications and intraoperative administration of DEX or KET. The control group (CTRL) received no additional drug. A comprehensive literature review was performed. A total of 71 pediatric patients underwent palatal surgery during the study period with 46 patients qualifying for analysis. Although results were not significant, consistent trends were observed with regards to lower opioid requirements during the first 24 hours for both medications compared with the CTRL. KET also had shorter time to discharge. The literature review resulted in several studies supporting decreased postoperative pain end points for both DEX and KET. In our sample, DEX and KET reduced postoperative opioid requirements. KET seems to have the added benefit of a shorter hospital stay. These finding are supported in the literature. With further investigation, the addition of these drugs may serve to provide improved pain relief without over sedation in patients undergoing cleft palate repair.
Project description:The local infiltration analgesia (LIA) technique has been widely used to reduce opioid requirements and to improve postoperative mobilization following total hip arthroplasty (THA). However, the evidence for the efficacy of LIA in THA is not yet clear. We determined whether single-shot LIA in addition to a multimodal analgesic regimen would reduce acute postoperative pain and opioid requirements after THA.116 patients undergoing primary THA under spinal anesthesia were included in this randomized, double-blind, placebo-controlled trial. All patients received oral opioid-sparing multimodal analgesia: etoricoxib, acetaminophen, and glucocorticoid. The patients were randomized to receive either 150 mL ropivacaine (2 mg/mL) and 0.5 mL epinephrine (1 mg/mL) or 150 mL 0.9% saline. Rescue analgesic consisted of morphine and oxycodone as needed. The primary endpoint was pain during mobilization in the recovery unit. Secondary endpoints were pain during mobilization on the day after surgery and total postoperative opioid requirements on the first postoperative day.The levels of pain during mobilization-both in the recovery unit and on the day after surgery-and consumption of opioids on the first postoperative day were similar in the 2 groups.LIA did not provide any extra analgesic effect after THA over and above that from the multimodal analgesic regimen used in this study.
Project description:STUDY DESIGN:Retrospective analysis on a national longitudinal database (2007-2014). OBJECTIVE:To determine the association between arthrodesis and complication rates, costs, surgical revision, and postoperative opioid prescription. SUMMARY OF BACKGROUND DATA:Arthrodesis in patients receiving laminectomy for lumbar spondylolisthesis remains controversial. However, population-level evidence to support the use of arthrodesis remains limited. METHODS:We identified 73,176 patient records and used coarsened exact matching to create comparable populations of patients who received laminectomy or laminectomy with arthrodesis. We use linear and logistic regression models to analyze the relationship between arthrodesis and postoperative complications, length of stay, costs, readmissions, surgical revisions, and postoperative opioid prescribing. RESULTS:Patients who underwent arthrodesis spent 1 more day in the hospital on average (P?<?0.01), and had higher costs of care at their index visit ($24,126, P?<?0.01), which were partially offset by lower costs of care over the 2 years following their procedure ($14,667 less in arthrodesis patients, P?=?0.01). Patients with arthrodesis were less likely to have a surgical revision (odds ratio?=?0.66, P?<?0.01). Patients with arthrodesis used more opioids in the first 2 months following their procedure, but had comparable opioid use to patients undergoing laminectomy without arthrodesis in all other postoperative months over the next 2 years, and were not more or less likely to convert to chronic opioid use. Postoperative opioid prescription varied dramatically across states (P?<?0.01); geographic variation in opioid use is substantially greater than differences in opioid use based on procedure performed. CONCLUSION:Arthrodesis is associated with reduced likelihood of surgical revision and increased use of opioids in the first 2 months following surgery, but not associated with greater or lesser opioid use beyond the initial 2 postoperative months. Geographic variation in opioid use is substantial even after accounting for patient characteristics and for whether patients underwent arthrodesis. LEVEL OF EVIDENCE:3.
Project description:Techniques that reduce injury to the knee extensor mechanism may cause less pain and allow faster recovery of knee function after primary total knee arthroplasty. A quadriceps-sparing (QS) subvastus technique of total knee arthroplasty was compared with medial parapatellar arthrotomy (MPPA) to determine which surgical technique led to better patient-reported function and less postoperative pain and opioid utilization.In this prospective, double-blind study, 129 patients undergoing total knee arthroplasty were randomized to the QS or the MPPA group after skin incision. All surgical procedures utilized minimally invasive surgery principles and standardized anesthesia, implants, analgesia, and rehabilitation. The Knee Society Score (KSS) was obtained at baseline and one and three months after surgery. Weekly telephone interviews were used to collect patient-reported outcomes including ambulatory device use, the UCLA (University of California Los Angeles) activity score, performance of daily living activities, and opioid utilization.No differences between groups were seen in opioid utilization, either during the acute hospitalization or in the eight weeks after surgery. The QS group reported significantly less pain at rest on postoperative day one and with activity on day three (p = 0.04 for each). Compared with baseline, both groups showed significant improvements in the KSS at one month (MPPA, p = 0.0278; QS, p = 0.0021) and three months (p < 0.0001 for each) as well as week-to-week gains in walking independence through five weeks after surgery. Independence from ambulatory devices outside the home lagged behind independence indoors by about two weeks in both groups.When primary total knee arthroplasty was performed with contemporary minimally invasive surgery principles and standardized implants, anesthesia, and postoperative pathways, the QS technique yielded no significant early functional advantages or differences in opioid utilization compared with the MPPA technique. However, the mean pain scores reported by patients in the QS group were slightly lower at rest on postoperative day one and during activity on day three.Therapeutic Level I. See Instructions for Authors for a complete description of levels of evidence.
Project description:BACKGROUND:Patients recovering from opioid use disorders (OUD) may be prone to relapse and opioid misuse in the postoperative period due to re-exposure to prescription opioids for pain control. This retrospective study analyzed the incidence of confirmed opioid misuse in the postoperative period in patients with OUDs enrolled in an opioid agonist treatment (OAT) program. METHODS:The study population was US veterans with a diagnosis of OUD who enrolled in the OAT program at VA Maryland Health Care System (Baltimore, Maryland, USA) between 1/1/2000 and 12/31/2016. The patients were excluded if they were enrolled in OAT for less than a year, or if they had surgery within the first 180 days after OAT admission. The surgical group consisted of veterans who had surgery or an invasive procedure during their enrollment in the OAT program. The control (reference) group consisted of enrolled veterans who did not have any invasive procedure. The primary outcome was the first opioid misuse within 365?days after surgery date in the surgical group or a randomly assigned sham surgery date in controls. Opioid misuse was defined as either inappropriate use of opioids detected via urinalysis or admission with a diagnosis of an opioid overdose. RESULTS:From a total of 1352 patients enrolled in the OAT program, 413 were excluded because they were enrolled for less than a year, and 26 were excluded because they had surgery within the first 180?days after admission to the OAT program. Of the 923 eligible patients, 87 had surgery while enrolled and 836 did not. Using propensity scores, all 87 of the surgical cases were matched to 249 of the control cases. In the matched groups, surgery was positively associated with postoperative opioid misuse (odds ratio (OR) of 1.91, 95% CI 1.05-3.48, p?=?0.034) in logistic regression. CONCLUSION:Among patients with a history of opioid use disorders, the postoperative period was associated with an increased risk of opioid misuse. Moreover, opioid misuse among patients in an opioid agonist treatment program may well be considered a surgical hazard.
Project description:Analgesic protocols used to treat pain after breast surgery vary significantly. The aim of this systematic review was to evaluate the available literature on this topic and develop recommendations for optimal pain management after oncological breast surgery. A systematic review using preferred reporting items for systematic reviews and meta-analysis guidance with procedure-specific postoperative pain management (PROSPECT) methodology was undertaken. Randomised controlled trials assessing postoperative pain using analgesic, anaesthetic or surgical interventions were identified. Seven hundred and forty-nine studies were found, of which 53 randomised controlled trials and nine meta-analyses met the inclusion criteria and were included in this review. Quantitative analysis suggests that dexamethasone and gabapentin reduced postoperative pain. The use of paravertebral blocks also reduced postoperative pain scores, analgesia consumption and the incidence of postoperative nausea and vomiting. Intra-operative opioid requirements were documented to be lower when a pectoral nerves block was performed, which also reduced postoperative pain scores and opioid consumption. We recommend basic analgesics (i.e. paracetamol and non-steroidal anti-inflammatory drugs) administered pre-operatively or intra-operatively and continued postoperatively. In addition, pre-operative gabapentin and dexamethasone are also recommended. In major breast surgery, a regional anaesthetic technique such as paravertebral block or pectoral nerves block and/or local anaesthetic wound infiltration may be considered for additional pain relief. Paravertebral block may be continued postoperatively using catheter techniques. Opioids should be reserved as rescue analgesics in the postoperative period. Research is needed to evaluate the role of novel regional analgesic techniques such as erector spinae plane or retrolaminar plane blocks combined with basic analgesics in an enhanced recovery setting.
Project description:OBJECTIVES:The purpose of this prospective proof of concept study was to investigate the feasibility of using percutaneous peripheral nerve stimulation of the femoral nerve to treat pain in the immediate postoperative period following ambulatory anterior cruciate ligament reconstruction with a patellar autograft. MATERIALS AND METHODS:Preoperatively, an electrical lead (SPRINT, SPR Therapeutics, Inc., Cleveland, OH, USA) was percutaneously implanted with ultrasound guidance anterior to the femoral nerve caudad to the inguinal crease. Within the recovery room, subjects received 5 min of either stimulation or sham in a randomized, double-masked fashion followed by a 5-min crossover period, and then continuous active stimulation until lead removal postoperative Day 14-28. Statistics were not applied to the data due to the small sample size of this feasibility study. RESULTS:During the initial 5-min treatment period, subjects randomized to stimulation (n = 5) experienced a slight downward trajectory (decrease of 7%) in their pain over the 5 min of treatment, while those receiving sham (n = 5) reported a slight upward trajectory (increase of 4%) until their subsequent 5-min stimulation crossover, during which time they also experienced a slight downward trajectory (decrease of 11% from baseline). A majority of subjects (80%) used a continuous adductor canal nerve block for rescue analgesia (in addition to stimulation) during postoperative Days 1-3, after which the median resting and dynamic pain scores remained equal or less than 1.5 on the numeric rating scale, respectively, and the median daily opioid consumption was less than 1.0 tablet. CONCLUSIONS:This proof of concept study demonstrates that percutaneous femoral nerve stimulation is feasible for ambulatory knee surgery; and suggests that this modality may be effective in providing analgesia and decreasing opioid requirements following anterior cruciate ligament reconstruction. clinicaltrials.gov: NCT02898103.
Project description:BACKGROUND: Gabapentin is an antiepileptic drug used in a variety of chronic pain conditions. Increasing numbers of randomized trials indicate that gabapentin is effective as a postoperative analgesic. This procedure-specific systematic review aims to analyse the 24-hour postoperative effect of gabapentin on acute pain in adults. METHODS: Medline, The Cochrane Library and Google Scholar were searched for double-blind randomized placebo controlled trials of gabapentin for postoperative pain relief compared with placebo, in adults undergoing a surgical procedure. Qualitative analysis of postoperative effectiveness was evaluated by assessment of significant difference (P < 0.05) in pain relief using consumption of supplemental analgesic and pain scores between study groups. Quantitative analyses of combined data from similar procedures, were performed by calculating the weighted mean difference (WMD) of 24-hour cumulated opioid requirements, and the WMD for visual analogue scale (VAS) pain, (early (6 h) and late (24 h) postoperatively), between study groups. Side-effects (nausea, vomiting, dizziness and sedation) were extracted for calculation of their relative risk (RR). RESULTS: Twenty-three trials with 1529 patients were included. In 12 of 16 studies with data on postoperative opioid requirement, the reported 24-hour opioid consumption was significantly reduced with gabapentin. Quantitative analysis of five trials in abdominal hysterectomy showed a significant reduction in morphine consumption (WMD - 13 mg, 95% confidence interval (CI) -19 to -8 mg), and in early pain scores at rest (WMD - 11 mm on the VAS, 95% CI -12 to -2 mm) and during activity (WMD -8 mm on the VAS; 95% CI -13 to -3 mm), favouring gabapentin. In spinal surgery, (4 trials), analyses demonstrated a significant reduction in morphine consumption (WMD of - 31 mg (95%CI - 53 to -10 mg) and pain scores, early (WMD - 17 mm on the VAS; 95 % CI -31 to -3 mm) and late (WMD -12 mm on the VAS; 95% CI -23 to -1 mm) also favouring gabapentin treatment. Nausea was improved with gabapentin in abdominal hysterectomy (RR 0.7; 95 % CI 0.5 to 0.9). Other side-effects were unaffected. CONCLUSION: Perioperative use of gabapentin has a significant 24-hour opioid sparing effect and improves pain score for both abdominal hysterectomy and spinal surgery. Nausea may be reduced in abdominal hysterectomy.