Category: Statistics

Clin Neurol Neurosurg. 2021 Mar 25;205:106610. doi: 10.1016/j.clineuro.2021.106610. Online ahead of print.

ABSTRACT

INTRODUCTION: The relationship between degree of surgical overlap and adverse postoperative outcomes remains poorly defined. This study aims to evaluate the impact of increasing duration of overlap on lumbar fusion outcomes.

PATIENTS AND METHODS: 1302 adult patients undergoing overlapping surgery during single-level, posterior-only lumbar fusion at a multi-hospital, university health system were retrospectively assessed. Amount of overlap was calculated as a percentage of total overlap time. Patients were separated into groups with the most (top 10% of patients) and least amounts of overlap (bottom 40% of patients). Using Coarsened Exact Matching, patients with the most and least amounts of overlap were matched on demographics alone, then on both demographics and attending surgeon. Univariate analysis was performed for the whole population and both matched cohorts to compare amount of overlap to risk of adverse postsurgical events. Significance for all analyses was p-value < 0.05.

RESULTS: Duration of overlap was not associated with outcomes in the whole population, demographic-matched, or surgeon-matched analyses. Before exact matching, patients with the most amount of overlap had a significantly higher CCI score (p = 0.031) and shorter length of surgery (p = 0.006). In the demographic matched cohort, patients with increased overlap had a significantly shorter length of surgery (p = 0.001) only. In the surgeon matched cohort, there were no differences in length of surgery or CCI score.

CONCLUSIONS: Duration of surgical overlap does not predict adverse outcomes following lumbar fusion. These results suggest that overlapping surgery is a safe practice within this common neurosurgical indication.

PMID:33845404 | DOI:10.1016/j.clineuro.2021.106610

J Surg Res. 2021 Apr 9;264:316-320. doi: 10.1016/j.jss.2021.03.016. Online ahead of print.

ABSTRACT

BACKGROUND: Current thyroid hormone replacement therapy (THRT) is built on weight-based standard calculation of dose. A novel Poisson regression model, which accounts for seven clinical variables, was recently proposed to improve accuracy of THRT. We aimed to compare the accuracy of estimated THRT dose to reach euthyroid and the difference in predicted dose between the Poisson (scheme A) and the weight-based standard (scheme B) in patients following total thyroidectomy for benign disease.

METHODS: We retrospectively reviewed medical record of patients who underwent total or completion thyroidectomy for benign disease at a single institution between 2011 and 2019. The THRT dose was calculated using both schemes. We compared the difference between calculated THRT and prediction rates for optimal THRT dosing needed to achieve a euthyroid state between dosing schemes. Patients were evaluated for achieving euthyroid state, defined as TSH 0.45-4.5 mIU/L. We also compared dosing error rates (> 25 mcg over- and underdosing) between schemes. Prediction rates were compared by BMI tertiles to account for the effect of BMI extremes in achieving euthyroid state. The difference in predicted dose between schemes was calculated in both the total sample size and patients that met euthyroid. A measure of agreement, Kappa, was used to estimate agreement between dosing schemes.

RESULTS: A total of 406 patients underwent total thyroidectomy for benign disease, with 184 having sufficient follow up data confirming euthyroid state. Of the 184 patients, 85.9% (n = 158) were women, 81% (n = 149) were Hispanic, and 56.5% (n = 104) were obese with a median BMI of 30.8 kg/m2. Scheme A resulted in a higher, but not statistically significant, accuracy rate (A: 60.3%, n = 111 versus B: 53.8%, n = 99; P = 0.21). Overdosing errors were lower with Scheme A (A:17.9% versus B: 32.1%; P = 0.0025) and less extreme > 25 µg (A: 17.9% versus B: 26.1%; P = 0.08). A trend in improved accuracy in patients with a BMI > 35 kg/m2 was noted (A: 46.9% versus B: 34.4%; P = 0.20). Scheme A also resulted in less overdosing errors in obese patients compared to Scheme B (A: 19.2% versus 45.2%; P = 0.0006). The average difference in predicted dose between schemes was an entire dose difference, mean of 16.0 µg and 15.8 µg for the total and euthyroid samples respectively. Furthermore, for the majority of patients the predicted dose did not match between the two dosing schemes for total and euthyroid samples, 76% (n = 311) and 76% (n = 141) respectively. In patients that achieved euthyroid, agreement between dosing schemes was low to moderate (Kappa = 0.360).

CONCLUSIONS: Lower rates of overdosing were found for scheme A, particularly with obese patients. No statistically significant differences in predicted THRT dose was observed between schemes. The difference in predicted dose between schemes was on average 15 ug, correlating with an entire dose. The consideration of clinical variables other than weight (scheme A) when determining optimal THRT dosing may be of importance to prevent overdoses, with particular clinical relevance in patients with higher BMIs.

PMID:33845415 | DOI:10.1016/j.jss.2021.03.016

Am J Emerg Med. 2021 Apr 3;48:73-78. doi: 10.1016/j.ajem.2021.04.005. Online ahead of print.

ABSTRACT

BACKGROUND: Emergency Department (ED) patients who leave without being seen (LWBS) are associated with adverse safety and medico-legal consequences. While LWBS risk has been previously tied to demographic and acuity related factors, there is limited research examining crowding-related risk in the pediatric setting. The primary objective of this study was to determine the association between LWBS risk and crowding, using the National Emergency Department Overcrowding Score (NEDOCS) and occupancy rate as crowding metrics.

METHODS: We performed a retrospective observational study on electronic health record (EHR) data from the ED of a quaternary care children’s hospital and trauma center during the 14-month study period. NEDOCS and occupancy rate were calculated for 15-min windows and matched to patient arrival time. We leveraged multiple logistic regression analyses to demonstrate the relationship between patientlevel LWBS risk and each crowding metric, controlling for characteristics drawn from the pre-arrival state. We performed a chi-squared test to determine whether a difference existed between the receiver operating characteristic (ROC) curves in the two models. Finally, we executed a dominance analysis using McFadden’s pseudo-R 2 to determine the relative importance of each crowding metric in the models.

RESULTS: A total of 54,890 patient encounters were studied, 1.22% of whom LWBS. The odds ratio for LWBS risk was 1.30 (95% CI 1.27-1.33) per 10-point increase in NEDOCS and 1.23 (95% CI 1.21-1.25). per 10% increase in occupancy rate. Area under the curve (AUC) was 86.9% for the NEDOCS model and 86.7% for the occupancy rate model. There was no statistically significant difference between the AUCs of the two models (p-value 0.27). Dominance analysis revealed that in each model, the most important variable studied was its respective crowding metric; NEDOCS accounted for 55.6% and occupancy rate accounted for 53.9% of predicted variance in LWBS.

CONCLUSION: Not only was ED overcrowding positively and significantly associated with individual LWBS risk, but it was the single most important factor that determined a patient’s likelihood of LWBS in the pediatric ED. Because occupancy rate and NEDOCS are available in real time, each could serve as a monitor for individual LWBS risk in the pediatric ED.

PMID:33845424 | DOI:10.1016/j.ajem.2021.04.005

Epilepsy Behav. 2021 Apr 9;118:107945. doi: 10.1016/j.yebeh.2021.107945. Online ahead of print.

ABSTRACT

BACKGROUND: The identification of hyperperfusion on ictal single-photon emission computed tomography (SPECT) scan is a technique for the localization of the epileptogenic zone (EZ) in patients with focal epilepsy undergoing presurgical evaluation. The accuracy of this technique has been improved by subtraction from an interictal image and coregistration with magnetic resonance imaging (MRI) (subtraction ictal SPECT coregistered to MRI (SISCOM)), and subsequently by the development of Statistical Ictal SPECT Co-registered to MRI (STATISCOM) which is reported to further improve localization accuracy by statistically accounting for random variation between images. However, the use of ictal SPECT is limited by the necessity for rapid injection of the radiotracer. The purpose of this study was to investigate the effect of tracer injection time on EZ localization rates using both STATISCOM and SISCOM.

METHODS: Consecutive patients with drug-resistant focal epilepsy who had an ictal SPECT scan while admitted to the video-electroencephalography (EEG) monitoring unit at the Royal Melbourne Hospital, Victoria, Australia, and a subsequent interictal scan, between 2009 and 2017 were included. The information collected included age, sex, seizure type, epilepsy diagnosis, and injection time. Statistical Ictal SPECT Co-registered to MRI and SISCOM images were generated and reviewed by two blinded reviewers. The rates of potential localization of the EZ, and the agreement with the EEG, were determined for each scan. Localization rates were compared between ictal scans with different radiotracer injection time windows (<30 s, 30-45 s, 45-60 s, 60-90 s, 90-120 s, >120 s).

RESULTS: Seventy patients (male = 32, 16-67 years) were included in the study. Overall agreement between the primary raters was moderate for STATISCOM (k = 0.44) and SISCOM (k = 0.57). The ability of SPECT to localize the potential EZ was 69% (48/70) for STATISCOM and 59% (41/70) for SISCOM. Injection time was not associated with the rate of localizing the potential EZ for STATISCOM (p = 0.64), whereas for SISCOM there was a trend that shorter injection times were associated with better ability to localize the potential EZ (p = 0.06). Agreement between SPECT and video-EEG data was 54% (38/70) for STATISCOM and 39% (27/70) for SISCOM. Statistical Ictal SPECT Co-registered to MRI did not show any difference of agreement across injection time groups (p = 0.42) whereas SISCOM showed better agreement with video-EEG data in the earlier injection time groups (p = 0.02). No differences in agreement between SPECT and video-EEG data were seen between patients with and without MRI lesions for either STATISCOM or SISCOM. Statistical Ictal SPECT Co-registered to MRI showed significantly better agreement for temporal than extratemporal seizures, with no difference of agreement between early (<45 s) and late (>45 s) injections.

CONCLUSION: Statistical Ictal SPECT Co-registered to MRI showed overall higher agreement rates with EZ localization by video-EEG than SISCOM, which was not affected by the injection times. Statistical Ictal SPECT Co-registered to MRI may provide localizing information for “late” injections where visual reads and SISCOM are inconclusive.

PMID:33845344 | DOI:10.1016/j.yebeh.2021.107945

Clin Biomech (Bristol, Avon). 2021 Mar 24;84:105333. doi: 10.1016/j.clinbiomech.2021.105333. Online ahead of print.

ABSTRACT

BACKGROUND: Patellar instability is a considerable problem that leads to pain and anxiety during simple functional tasks. Femoral derotational osteotomy has become a common surgical procedure to improve patella mechanics, stability and loading. However, it remains unclear if static (MRI measured) femoral anteversion is sufficient to capture the dynamic femoral rotation during walking and represents a good indication for the surgical procedure. This research investigates the relationship between static femoral anteversion and internally rotated gait in adolescents with patellofemoral instability.

METHODS: This retrospective study included 30 adolescents with recurrent patella instability (minimum three patella dislocations) aged 12 to 18 years (28 female/2 male; 22 unilateral/8 bilateral). All participants were assessed with 3D gait analysis and the femoral anteversion was examined using a rotational MRI. Multiple kinematic parameter were correlated with the ipsilateral femoral anteversion and tibia torsion using the Pearson coefficient.

FINDINGS: The correlation between parameters of dynamic hip rotation (e.g. maximum and mean internal hip rotation in stance and swing) and MRI measured femoral anteversion (mean 26.5° ± 9°) was weak and did not reach statistical significance. We found 47% (14 out of 30) subjects with increased femoral anteversion but normal hip rotation in stance.

INTERPRETATION: There was no relationship between increased femoral anteversion and dynamic hip rotation. Consequently, femoral anteversion should not be used as the only indication for femoral derotational osteotomy. Three-dimensional gait analysis might be necessary to assess the appropriate surgical intervention in adolescents with patello femoral instability.

PMID:33845347 | DOI:10.1016/j.clinbiomech.2021.105333

Biomed Pharmacother. 2021 Apr 9;139:111559. doi: 10.1016/j.biopha.2021.111559. Online ahead of print.

ABSTRACT

BACKGROUND: In a previous drug-drug interaction (DDI) screening study intended to generate hypotheses, clopidogrel + either eszopiclone or zolpidem (vs. clopidogrel alone) were associated with serious bleeding.

OBJECTIVES: To confirm or refute these DDI signals and examine associations with other hypnotics in an independent population of United States Medicaid beneficiaries METHODS: We employed a bi-directional self-controlled case series design in eligible individuals concomitantly exposed to one of 12 hypnotics (precipitants, exposures of interest) plus either clopidogrel (the object drug) or pravastatin (the negative control object drug). The outcome was hospital presentation with serious bleeding. Using conditional Poisson regression, we calculated confounder-adjusted rate ratios (RRs) and 95% confidence intervals for serious bleeding during clopidogrel + precipitant use (vs. clopidogrel alone). To distinguish a DDI from a precipitant’s inherent effect on bleeding, we divided effect measures by the adjusted RR for the corresponding pravastatin + precipitant pair to obtain ratios of RR (RRRs).

RESULTS: Among 23,194 users of clopidogrel and 3824 of pravastatin who experienced serious bleeding during an active prescription for one of these agents, confounder-adjusted RRRs for serious bleeding were 6.63 (0.39-113.01) and 0.77 (0.53-1.11) with eszopiclone and zolpidem, respectively, whereas confounder-adjusted RRRs for other hypnotics ranged from 0.18 (0.04-0.85) for triazolam to 1.79 (0.16-20.44) for zaleplon. Statistical imprecision therefore precluded us from confirming or refuting these prior signals with eszopiclone and zolpidem.

CONCLUSIONS: While we could not confirm or refute previously identified DDI signals, numerically elevated RRRs for serious bleeding with several clopidogrel + hypnotic pairs warrant further examination.

PMID:33845372 | DOI:10.1016/j.biopha.2021.111559

J Pediatr Nurs. 2021 Apr 9;59:103-109. doi: 10.1016/j.pedn.2021.03.019. Online ahead of print.

ABSTRACT

PURPOSE: The aim of this study was to establish content validity of a developmentally based assessment tool of readiness for medical independence for specialty providers.

DESIGN AND METHODS: The validation process used expert panel evaluation to assess the items believed to measure the desired content in the nine age-based scales within the RAISE (Readiness Assessment of Independence for Specialty Encounters) tool. Experts in child development and transition rated items on relevance, clarity and developmental appropriateness via electronic survey. Statistical analyses included calculation of interrater agreement (IRA), content validity indices (CVIs), and factorial validity indices (FVI).

RESULTS: A total of 135 items were rated by 36 experts. Mean I-CVIs for 123 items across nine developmental scales met criteria for retention, ranging from 0.76 (threshold) to 1.00 (excellent). Mean I-CVIs for all 25 items across the five psychosocial stressor scales met criteria for retention, ranging from 0.92 to 1.00 (excellent).

CONCLUSIONS: Findings from the current content validation study suggest that items on the revised RAISE tool are relevant, clear, and developmentally-appropriate as rated by experts in the fields of child development and transition. The tool, consisting of age based scales (ages birth-2, 3-4, 5-6, 7-8, 9-11, 12-13, 14-15, 16-17, 18-21), is shown to have content validity of the retained items meeting criteria.

PRACTICE IMPLICATIONS: With content validity of the RAISE tool established by experts, this developmentally based assessment tool can be integrated into practice to assist providers in educating patients around skills of medical independence which could improve transition outcomes.

PMID:33845322 | DOI:10.1016/j.pedn.2021.03.019

Plant Physiol Biochem. 2021 Apr 4;163:139-153. doi: 10.1016/j.plaphy.2021.03.059. Online ahead of print.

ABSTRACT

In this study, the effects of three halotolerant rhizobacterial isolates AL, HR, and SB, which are able to grow at a salinity level of 1600 mM NaCl, with multiple plant growth promoting (PGP) traits on some seed and forage quality attributes, and vegetative, reproductive, biochemical and physiological characteristics of wheat plant irrigated with saline water (0, 40, 80, and 160 mM NaCl) were investigated. The ability of halotolerant bacterial isolates to produce PGP traits was affected by salinity levels, depending upon the bacterial isolates. Salinity stress significantly affected the yield, quality, and growth of wheat by modifying the morpho-physiological and biochemical traits of the exposed plants. However, all three bacterial isolates, especially isolate AL, significantly improved the biochemical (an increase in K⁺/Na⁺ ratio by 55%, plant P content by 50%, and plant Ca content by 31%), morphological (an increase in stem dry weight by 52%, root dry weight by 44%, spike dry weight by 34%, and grain dry weight by 43%), and physiological (an increase in leaf proline content by 50% and total phenol in leaf by 42%) attributes of wheat and aided the plant to tolerate salinity stress in contrast to un-inoculated plant. Plants inoculated with bacterial isolates showed significantly improved seed amylose by 36%, leaf crude protein by 30%, leaf metabolic energy by 37%, and leaf water-soluble sugar content by 34%. Among the measured PGP and plant attributes, bacterial auxin and plant K content were of key importance in increasing reproductive performance of wheat. The bacterial isolates AL, HR, and SB were identified as Bacillus safensis, B. pumilus, and Zhihengliuella halotolerans, respectively, based on 16 S rDNA sequence. The study reveals that application of halotolerant plant growth-promoting rhizobacteria isolated from halophytic rangeland plants can be a cost effective and ecological sustainable method to improve wheat productivity, especially the attributes related to seed and forage quality, under salinity stress conditions.

PMID:33845330 | DOI:10.1016/j.plaphy.2021.03.059

J Mech Behav Biomed Mater. 2021 Apr 3;119:104503. doi: 10.1016/j.jmbbm.2021.104503. Online ahead of print.

ABSTRACT

This in vitro study used the same frequency and duration of acid contact as a previous in situ/in vivo study to evaluate the effect of erosion on CAD/CAM restorative materials and human enamel and to compare the effects of in vitro and in situ/in vivo acid challenges on CAD/CAM restorative materials and human enamel. The CAD/CAM restorative materials (IPS e.max CAD, Lava Ultimate, and PMMA block) and human enamel were eroded by immersion in 150 ml of cola drink for 14 days (4 × 5 min/day). The surface microhardness and surface roughness of the specimens were measured at baseline (T1), day 7 (T2), and day 14 (T3). The substance losses were measured at T2 and T3. The data were statistically analyzed using repeated measures ANOVA and Bonferroni’s test (α = 0.05). Erosion significantly decreased the surface microhardness of the CAD/CAM restorative materials and human enamel (all P < 0.001). The overall percentage of surface microhardness loss (%SMH_l) of the PMMA block and enamel due to in vitro erosion was significantly higher than that due to in situ/in vivo erosion (P = 0.02 and P < 0.001, respectively). Consistent with in situ/in vivo erosion, the surface roughness and profile of the tested restorative materials remained unchanged after in vitro erosion. A significant increase in the surface roughness and substance loss was observed for enamel after in vitro erosion (all P < 0.001). The overall substance loss of enamel due to in vitro erosion was significantly higher than that due to in situ/in vivo erosion (P < 0.001). In conclusion, erosion decreased the surface microhardness of the CAD/CAM restorative materials and human enamel. Moreover, erosion negatively influenced the substance loss and surface roughness of human enamel. For the substance loss of enamel and %SMH_l of PMMA block and enamel, the in vitro erosive effects were approximately 1-2 times greater than the in situ/in vivo effects. However, for the surface roughness and profile of the CAD/CAM restorative materials, no significant difference was found between in vitro and in situ/in vivo erosion.

PMID:33845297 | DOI:10.1016/j.jmbbm.2021.104503

Neural Netw. 2021 Mar 26;141:11-29. doi: 10.1016/j.neunet.2021.03.025. Online ahead of print.

ABSTRACT

In deep learning tasks, the update step size determined by the learning rate at each iteration plays a critical role in gradient-based optimization. However, determining the appropriate learning rate in practice typically relies on subjective judgment. In this work, we propose a novel optimization method based on local quadratic approximation (LQA). In each update step, we locally approximate the loss function along the gradient direction by using a standard quadratic function of the learning rate. Subsequently, we propose an approximation step to obtain a nearly optimal learning rate in a computationally efficient manner. The proposed LQA method has three important features. First, the learning rate is automatically determined in each update step. Second, it is dynamically adjusted according to the current loss function value and parameter estimates. Third, with the gradient direction fixed, the proposed method attains a nearly maximum reduction in the loss function. Extensive experiments were conducted to prove the effectiveness of the proposed LQA method.

PMID:33845311 | DOI:10.1016/j.neunet.2021.03.025