Search

Although multiple prospective comparative studies exist for adults with midshaft clavicle fractures, few comparative studies exist in older children and adolescents. This study compares intermediate-term functional, pain, radiographic, and complication outcomes in children and adolescents with midshaft clavicle fractures treated operatively or nonoperatively.MethodsChildren and adolescents who were 10 to 18 years of age with midshaft clavicle fractures treated from 2006 to 2017 were identified through electronic picture archiving and communication system (PACS) radiographic and chart review. Closed injuries with a minimum of 1 year from the time of the injury were included. Patients completed patient-reported outcome measures and a focused questionnaire. The primary outcome was the abbreviated Disabilities of the Arm, Shoulder and Hand Questionnaire (QuickDASH) score; the secondary outcomes included the University of California Los Angeles (UCLA) Activity Scale, Patient-Reported Outcomes Measurement Information System (PROMIS) Pain Interference (PI) and PROMIS Physical Function (PF) Upper Extremity (UE), Hospital for Special Surgery Pediatric Functional Activity Brief Scale (HSS Pedi-FABS), Numeric Rating Scale (NRS) for Pain, and rates of implant removal, nonunion, and refracture. Univariate comparisons between the operatively treated group and the nonoperatively treated group were performed.ResultsIn this study, 302 patients were identified; of these, 79 patients (43 operatively treated and 36 nonoperatively treated) or their parent or legal guardian consented and the patients completed the study protocol. The mean age (and standard deviation) at the time of the injury was 16.0 ± 1.7 years for the operatively treated group and 13.5 ± 1.8 years for the nonoperatively treated group (p < 0.001). The median follow-up was 3.8 years. The patient-reported outcome scores were similar (p > 0.05 for all comparisons) between the operatively treated group and the nonoperatively treated group, even for patients at or near skeletal maturity and those with shortening of ≥15 mm or ≥100% translation. There were no nonunions. The refracture rate was 5% (2 of 43) for the operatively treated group and 3% (1 of 36) for the nonoperatively treated group (p = 1.00). There were 19% (8 of 43) who underwent implant removal. Responders were similar to nonresponders, other than female patients (p < 0.001) and surgically treated patients (p = 0.049) being more likely to respond.ConclusionsIn the subset of this cohort who responded and completed the study, both operatively and nonoperatively treated patients had excellent functional and pain outcomes, similar refracture rates, and no nonunions. Even in widely displaced and translated fractures, and in those at or near skeletal maturity, outcomes were similar. Operative management should potentially be reserved for special circumstances. A prospective comparative study of older adolescents (female patients who are 14 to 18 years of age and male patients who are 16 to 20 years of age) with displaced midshaft clavicle fractures is warranted.Level of evidenceTherapeutic Level III. See Instructions for Authors for a complete description of levels of evidence.

Algorithmic decision making has proliferated and now impacts our daily lives in both mundane and consequential ways. Machine learning practitioners make use of a myriad of algorithms for predictive models in applications as diverse as movie recommendations, medical diagnoses, and parole recommendations without delving into the reasons driving specific predictive decisions. Machine learning algorithms in such applications are often chosen for their superior performance, however popular choices such as random forest and deep neural networks fail to provide an interpretable understanding of the predictive model. In recent years, rule-based algorithms have been used to address this issue. Wang et al. (2017) presented an or-of-and (disjunctive normal form) based classification technique that allows for classification rule mining of a single class in a binary classification; this method is also shown to perform comparably to other modern algorithms. In this work, we extend this idea to provide classification rules for both classes simultaneously. That is, we provide a distinct set of rules for both positive and negative classes. In describing this approach, we also present a novel and complete taxonomy of classifications that clearly capture and quantify the inherent ambiguity in noisy binary classifications in the real world. We show that this approach leads to a more granular formulation of the likelihood model and a simulated-annealing based optimization achieves classification performance competitive with comparable techniques. We apply our method to synthetic as well as real world data sets to compare with other related methods that demonstrate the utility of our proposal.

Background: There are limited data on the incidence and outcomes of open and closed hip reduction in patients with developmental dysplasia of the hip (DDH). The aims of this study were to determine the incidence of open and closed reduction of the hip using population-level data and to assess the rates of subsequent surgery. Methods: Children aged 3 years and younger with DDH who underwent open or closed reduction of the hip between 1997 and 2013 were identified in the New York Statewide Planning and Research Cooperative System (SPARCS) database. Patient age, sex, race, and insurance status as well as concurrent procedures were extracted. Admissions through 2014 were searched for subsequent surgeries, providing a minimum 1-year post-reduction surveillance for all patients. Age-specific incidence rates were calculated using New York State annual population data. The rates of concurrent and subsequent surgeries were calculated. A sensitivity analysis was performed to provide a range for the rates of subsequent surgery. Univariate analyses consisted of chi-square or Fisher exact tests for categorical variables. Results: In total, 897 patients (637 who underwent closed reduction and 260 who underwent open reduction) were identified. The age-specific incidence per 100,000 population was 12.5 for closed reduction and 2.6 for open reduction for

Numerous Machine Learning (ML) bias-related failures in recent years have led to scrutiny of how companies incorporate aspects of transparency and accountability in their ML lifecycles. Companies have a responsibility to monitor ML processes for bias and mitigate any bias detected, ensure business product integrity, preserve customer loyalty, and protect brand image. Challenges specific to industry ML projects can be broadly categorized into principled documentation, human oversight, and need for mechanisms that enable information reuse and improve cost efficiency. We highlight specific roadblocks and propose conceptual solutions on a per-category basis for ML practitioners and organizational subject matter experts. Our systematic approach tackles these challenges by integrating mechanized and human-in-the-loop components in bias detection, mitigation, and documentation of projects at various stages of the ML lifecycle. To motivate the implementation of our system -- SIFT (System to Integrate Fairness Transparently) -- we present its structural primitives with an example real-world use case on how it can be used to identify potential biases and determine appropriate mitigation strategies in a participatory manner., Comment: 14 pages, 4 figures

BackgroundThere are limited data on the incidence and outcomes of open and closed hip reduction in patients with developmental dysplasia of the hip (DDH). The aims of this study were to determine the incidence of open and closed reduction of the hip using population-level data and to assess the rates of subsequent surgery.MethodsChildren aged 3 years and younger with DDH who underwent open or closed reduction of the hip between 1997 and 2013 were identified in the New York Statewide Planning and Research Cooperative System (SPARCS) database. Patient age, sex, race, and insurance status as well as concurrent procedures were extracted. Admissions through 2014 were searched for subsequent surgeries, providing a minimum 1-year post-reduction surveillance for all patients. Age-specific incidence rates were calculated using New York State annual population data. The rates of concurrent and subsequent surgeries were calculated. A sensitivity analysis was performed to provide a range for the rates of subsequent surgery. Univariate analyses consisted of chi-square or Fisher exact tests for categorical variables.ResultsIn total, 897 patients (637 who underwent closed reduction and 260 who underwent open reduction) were identified. The age-specific incidence per 100,000 population was 12.5 for closed reduction and 2.6 for open reduction for

Algorithmic decision making has proliferated and now impacts our daily lives in both mundane and consequential ways. Machine learning practitioners make use of a myriad of algorithms for predictive models in applications as diverse as movie recommendations, medical diagnoses, and parole recommendations without delving into the reasons driving specific predictive decisions. Machine learning algorithms in such applications are often chosen for their superior performance, however popular choices such as random forest and deep neural networks fail to provide an interpretable understanding of the predictive model. In recent years, rule-based algorithms have been used to address this issue. Wang et al. (2017) presented an or-of-and (disjunctive normal form) based classification technique that allows for classification rule mining of a single class in a binary classification; this method is also shown to perform comparably to other modern algorithms. In this work, we extend this idea to provide classification rules for both classes simultaneously. That is, we provide a distinct set of rules for both positive and negative classes. In describing this approach, we also present a novel and complete taxonomy of classifications that clearly capture and quantify the inherent ambiguity in noisy binary classifications in the real world. We show that this approach leads to a more granular formulation of the likelihood model and a simulated-annealing based optimization achieves classification performance competitive with comparable techniques. We apply our method to synthetic as well as real world data sets to compare with other related methods that demonstrate the utility of our proposal.

Backgroundε-Aminocaproic acid (EACA) is an antifibrinolytic agent that has been shown to decrease blood loss and transfusion requirements in several populations undergoing various surgical procedures. However, the efficacy of EACA has not been assessed in pediatric patients with cerebral palsy undergoing bilateral varus rotational femoral osteotomies. The purpose of this study was to assess the efficacy of intravenous EACA in reducing calculated intraoperative blood loss and transfusions in this population.MethodsPatients aged 18 years or younger were eligible. Patients were randomized to receive EACA or placebo (saline), and randomization was stratified based on sex and whether or not additional soft tissue or osseous procedures were performed. On the basis of retrospective data, the calculated sample size was 12 patients per arm to detect a difference of 250-mL blood loss. The primary outcome was calculated intraoperative blood loss. Secondary outcomes included transfusion requirements, 24-hour drain output, length of stay, and incidence of complications.ResultsThe mean age of patients in this study was 8 years (SD: 2.4 y). There were no differences in age, sex, height, weight, type of anesthesia, operative time, and associated procedures between the EACA and placebo groups (P>0.05). Preoperative hematocrit was lower in the EACA group (37.1 vs. 40.0, P=0.04). Calculated intraoperative blood loss was 536 mL in the EACA group and 628 mL in the placebo group (P=0.45). Transfusions were required in 62% of patients in the EACA group and 67% of patients in the placebo group (P=0.68). Total 24-hour drain output was 72.5 mL in the EACA group and 103.3 mL in the placebo group (P=0.37). Length of stay was similar between both groups, and there were no drug or placebo-related complications in either group.ConclusionsThere was no difference in blood loss or transfusion requirements associated with EACA compared with placebo; however, this study is underpowered to detect smaller differences in blood loss. Additional studies with larger sample sizes are needed to confirm these findings and further elucidate the indications for antifibrinolytic agents in pediatric patients.Level of evidenceLevel I.

Objectives: The primary aim of this study was to determine the rate of subsequent ACL reconstruction and additional non-ACL knee surgery in children that had previously undergone ACL reconstruction in New York State. The secondary aim was to assess factors associated with additional ACL and non-ACL knee surgery. Methods: Pediatric patients (age < 21) who underwent ACL reconstruction between 1997 and 2010 in New York State were identified using the Statewide Planning and Research Cooperative System (SPARCS) database. Patients were tracked for subsequent ACL reconstruction and additional non-ACL knee surgery. Each case had a minimum of one year follow up. A Cox proportional hazards model was used to assess time to subsequent surgery, adjusting for age, sex, race, comorbidity index, insurance type, surgeon and hospital ACL volume, and poverty prevalence. Results: 23,912 primary pediatric ACL reconstructions were identified. 1955 patients (8.2%) underwent subsequent ACL reconstruction. 7.4% had one additional ACL reconstruction and 0.7% had two or more additional ACL reconstructions. 3341 patients (14%) had subsequent non-ACL knee surgery with 11.1% having one subsequent surgery and 2.9% having two or more additional knee surgeries. Risk factors for revision ACL surgery were younger age at time of primary ACL surgery (p20 per year; p=0.002), and higher surgeon ACL volume (>10 per year; p=0.003). Risk factors for return to the OR for other non-ACL knee surgeries were younger age at the time of primary ACL reconstruction (p20 per year; p=0.007). Conclusion: This study aims to provide information on the rate of return to the operating room and rate of secondary ACL reconstruction in an adolescent population over the past 20 years in New York State. Males and younger patients had a higher rate of subsequent ACL reconstruction. Higher rates of ACL reconstruction in younger children may be due to greater at-risk activities in younger children, longer follow up, or inherently less reliable reconstructions given the non-anatomic reconstructions that are frequently required in the skeletally immature. The association between socioeconomic factors and increased subsequent knee surgeries likely represents a disparity in access to care for low socioeconomic status patients.