We study bias‐reduced estimators of exponentially transformed parameters in general linear models (GLMs) and show how they can be used to obtain bias‐reduced conditional (or unconditional) odds ...ratios in matched case‐control studies. Two options are considered and compared: the explicit approach and the implicit approach. The implicit approach is based on the modified score function where bias‐reduced estimates are obtained by using iterative procedures to solve the modified score equations. The explicit approach is shown to be a one‐step approximation of this iterative procedure. To apply these approaches for the conditional analysis of matched case‐control studies, with potentially unmatched confounding and with several exposures, we utilize the relation between the conditional likelihood and the likelihood of the unconditional logit binomial GLM for matched pairs and Cox partial likelihood for matched sets with appropriately setup data. The properties of the estimators are evaluated by using a large Monte Carlo simulation study and an illustration of a real dataset is shown. Researchers reporting the results on the exponentiated scale should use bias‐reduced estimators since otherwise the effects can be under or overestimated, where the magnitude of the bias is especially large in studies with smaller sample sizes.
Prediction models are used in clinical research to develop rules that can be used to accurately predict the outcome of the patients based on some of their characteristics. They represent a valuable ...tool in the decision making process of clinicians and health policy makers, as they enable them to estimate the probability that patients have or will develop a disease, will respond to a treatment, or that their disease will recur. The interest devoted to prediction models in the biomedical community has been growing in the last few years. Often the data used to develop the prediction models are class-imbalanced as only few patients experience the event (and therefore belong to minority class).
Prediction models developed using class-imbalanced data tend to achieve sub-optimal predictive accuracy in the minority class. This problem can be diminished by using sampling techniques aimed at balancing the class distribution. These techniques include under- and oversampling, where a fraction of the majority class samples are retained in the analysis or new samples from the minority class are generated. The correct assessment of how the prediction model is likely to perform on independent data is of crucial importance; in the absence of an independent data set, cross-validation is normally used. While the importance of correct cross-validation is well documented in the biomedical literature, the challenges posed by the joint use of sampling techniques and cross-validation have not been addressed.
We show that care must be taken to ensure that cross-validation is performed correctly on sampled data, and that the risk of overestimating the predictive accuracy is greater when oversampling techniques are used. Examples based on the re-analysis of real datasets and simulation studies are provided. We identify some results from the biomedical literature where the incorrect cross-validation was performed, where we expect that the performance of oversampling techniques was heavily overestimated.
The goal of class prediction studies is to develop rules to accurately predict the class membership of new samples. The rules are derived using the values of the variables available for each subject: ...the main characteristic of high-dimensional data is that the number of variables greatly exceeds the number of samples. Frequently the classifiers are developed using class-imbalanced data, i.e., data sets where the number of samples in each class is not equal. Standard classification methods used on class-imbalanced data often produce classifiers that do not accurately predict the minority class; the prediction is biased towards the majority class. In this paper we investigate if the high-dimensionality poses additional challenges when dealing with class-imbalanced prediction. We evaluate the performance of six types of classifiers on class-imbalanced data, using simulated data and a publicly available data set from a breast cancer gene-expression microarray study. We also investigate the effectiveness of some strategies that are available to overcome the effect of class imbalance.
Our results show that the evaluated classifiers are highly sensitive to class imbalance and that variable selection introduces an additional bias towards classification into the majority class. Most new samples are assigned to the majority class from the training set, unless the difference between the classes is very large. As a consequence, the class-specific predictive accuracies differ considerably. When the class imbalance is not too severe, down-sizing and asymmetric bagging embedding variable selection work well, while over-sampling does not. Variable normalization can further worsen the performance of the classifiers.
Our results show that matching the prevalence of the classes in training and test set does not guarantee good performance of classifiers and that the problems related to classification with class-imbalanced data are exacerbated when dealing with high-dimensional data. Researchers using class-imbalanced data should be careful in assessing the predictive accuracy of the classifiers and, unless the class imbalance is mild, they should always use an appropriate method for dealing with the class imbalance problem.
Blagus, R, Jurak, G, Starc, G, and Leskošek, B. Centile reference curves of the SLOfit physical fitness tests for school-aged children and adolescents. J Strength Cond Res 37(2): 328-336, 2023-The ...study provides sex- and age-specific centile norms of Slovenian children and youth. Physical fitness was assessed using the SLOfit test battery on population data, including 185,222 children, aged 6-19 years, measured in April and May 2018. Centile curves for both sexes and 12 test items were constructed using the generalized additive models for location, scale, and shape (GAMLSS). Boys generally achieved higher scores in most of the physical fitness tests, except in stand and reach, but this was not consistent throughout childhood and adolescence, nor did it pertain to the entire range of performance. Girls outperformed boys in the arm-plate tapping test throughout childhood; the poorest performing girls outperformed the poorest performing boys in the 600-m run, 60-m dash, backward obstacle course, and standing broad jump. The shapes and trends of physical fitness curves adequately reflect the effects of growth and development on boys' and girls' physical performance. Comparing the existing reference fitness curves showed that Slovenian children and adolescents display higher fitness levels than their peers from other countries. This study provides the most up-to-date sex- and age-specific reference fitness centile curves of Slovenian children, which can be used as benchmark values for health and fitness monitoring and surveillance systems.
Abstract
Background
For finite samples with binary outcomes penalized logistic regression such as ridge logistic regression has the potential of achieving smaller mean squared errors (MSE) of ...coefficients and predictions than maximum likelihood estimation. There is evidence, however, that ridge logistic regression can result in highly variable calibration slopes in small or sparse data situations.
Methods
In this paper, we elaborate this issue further by performing a comprehensive simulation study, investigating the performance of ridge logistic regression in terms of coefficients and predictions and comparing it to Firth’s correction that has been shown to perform well in low-dimensional settings. In addition to tuned ridge regression where the penalty strength is estimated from the data by minimizing some measure of the out-of-sample prediction error or information criterion, we also considered ridge regression with pre-specified degree of shrinkage. We included ‘oracle’ models in the simulation study in which the complexity parameter was chosen based on the true event probabilities (prediction oracle) or regression coefficients (explanation oracle) to demonstrate the capability of ridge regression if truth was known.
Results
Performance of ridge regression strongly depends on the choice of complexity parameter. As shown in our simulation and illustrated by a data example, values optimized in small or sparse datasets are negatively correlated with optimal values and suffer from substantial variability which translates into large MSE of coefficients and large variability of calibration slopes. In contrast, in our simulations pre-specifying the degree of shrinkage prior to fitting led to accurate coefficients and predictions even in non-ideal settings such as encountered in the context of rare outcomes or sparse predictors.
Conclusions
Applying tuned ridge regression in small or sparse datasets is problematic as it results in unstable coefficients and predictions. In contrast, determining the degree of shrinkage according to some meaningful prior assumptions about true effects has the potential to reduce bias and stabilize the estimates.
Model checking plays an important role in parametric regression as model misspecification seriously affects the validity and efficiency of regression analysis. Model checks can be performed by ...constructing an empirical process from the model's fitted values and residuals. Due to a complex covariance function of the process obtaining the exact distribution of the test statistic is, however, intractable. Several solutions to overcome this have been proposed. It was shown that the simulation and bootstrap-based approaches are asymptotically valid, however, we show by using simulations that the rate of convergence can be slow. We, therefore, propose to estimate the null distribution by using a novel permutation-based procedure. We prove, under some mild assumptions, that this yields consistent tests under the null and some alternative hypotheses. Small sample properties of the proposed approach are studied in an extensive Monte Carlo simulation study and real data illustration is also provided.
In binary logistic regression data are 'separable' if there exists a linear combination of explanatory variables which perfectly predicts the observed outcome, leading to non-existence of some of the ...maximum likelihood coefficient estimates. A popular solution to obtain finite estimates even with separable data is Firth's logistic regression (FL), which was originally proposed to reduce the bias in coefficient estimates. The question of convergence becomes more involved when analyzing clustered data as frequently encountered in clinical research, e.g. data collected in several study centers or when individuals contribute multiple observations, using marginal logistic regression models fitted by generalized estimating equations (GEE). From our experience we suspect that separable data are a sufficient, but not a necessary condition for non-convergence of GEE. Thus, we expect that generalizations of approaches that can handle separable uncorrelated data may reduce but not fully remove the non-convergence issues of GEE.
We investigate one recently proposed and two new extensions of FL to GEE. With 'penalized GEE' the GEE are treated as score equations, i.e. as derivatives of a log-likelihood set to zero, which are then modified as in FL. We introduce two approaches motivated by the equivalence of FL and maximum likelihood estimation with iteratively augmented data. Specifically, we consider fully iterated and single-step versions of this 'augmented GEE' approach. We compare the three approaches with respect to convergence behavior, practical applicability and performance using simulated data and a real data example.
Our simulations indicate that all three extensions of FL to GEE substantially improve convergence compared to ordinary GEE, while showing a similar or even better performance in terms of accuracy of coefficient estimates and predictions. Penalized GEE often slightly outperforms the augmented GEE approaches, but this comes at the cost of a higher burden of implementation.
When fitting marginal logistic regression models using GEE on sparse data we recommend to apply penalized GEE if one has access to a suitable software implementation and single-step augmented GEE otherwise.
To date, few data on how the COVID-19 pandemic and restrictions affected children's physical activity in Europe have been published. This study examined the prevalence and correlates of physical ...activity and screen time from a large sample of European children during the COVID-19 pandemic to inform strategies and provide adequate mitigation measures. An online survey was conducted using convenience sampling from 15 May to 22 June, 2020. Parents were eligible if they resided in one of the survey countries and their children aged 6-18 years. 8395 children were included (median age IQR, 13 10-15 years; 47% boys; 57.6% urban residents; 15.5% in self-isolation). Approximately two-thirds followed structured routines (66.4% 95%CI, 65.4-67.4), and more than half were active during online P.E. (56.6% 95%CI, 55.5-57.6). 19.0% (95%CI, 18.2-19.9) met the WHO Global physical activity recommendation. Total screen time in excess of 2 h/day was highly prevalent (weekdays: 69.5% 95%CI, 68.5-70.5; weekend: 63.8% 95%CI, 62.7-64.8). Playing outdoors more than 2 h/day, following a daily routine and being active in online P.E. increased the odds of healthy levels of physical activity and screen time, particularly in mildly affected countries. In severely affected countries, online P.E. contributed most to meet screen time recommendation, whereas outdoor play was most important for adequate physical activity. Promoting safe and responsible outdoor activities, safeguarding P.E. lessons during distance learning and setting pre-planned, consistent daily routines are important in helping children maintain healthy active lifestyle in pandemic situation. These factors should be prioritised by policymakers, schools and parents.
Highlights
To our knowledge, our data provide the first multi-national estimates on physical activity and total screen time in European children roughly two months after COVID-19 was declared a global pandemic.
Only 1 in 5 children met the WHO Global physical activity recommendations.
Under pandemic conditions, parents should set pre-planned, consistent daily routines and integrate at least 2-hours outdoor activities into the daily schedule, preferable on each day. Schools should make P.E. lessons a priority. Decision makers should mandate online P.E. be delivered by schools during distance learning. Closing outdoor facilities for PA should be considered only as the last resort during lockdowns.
Lyme borreliosis is the most prevalent vector-borne disease in Europe and the USA. Doxycycline for 10 days is the primary treatment recommendation for erythema migrans. To reduce potentially harmful ...antibiotic overuse by identifying shorter effective treatments, we aimed to assess whether oral doxycycline for 7 days is non-inferior to 14 days in adults with solitary erythema migrans.
In this randomised open-label non-inferiority trial, we enrolled patients with a solitary erythema at the University Medical Centre in Ljubljana, Slovenia. Patients were excluded if they were pregnant or lactating, immunosuppressed, allergic to doxycycline, or had received antibiotics with anti-borrelial activity within 10 days preceding enrolment or had additional manifestations of Lyme borreliosis Adults were randomly allocated 1:1 to receive oral doxycycline 100 mg twice a day for 7 days or 14 days. The primary efficacy endpoint was the difference in proportion of patients with treatment failure, defined as persistent erythema, new objective signs of Lyme borreliosis, or borrelial isolation on skin re-biopsy at 2 months, in a per-protocol analysis (the population that completed the assigned doxycycline regimen according to the study protocol and did not receive any other antibiotics with anti-borrelial activity until the 2-month visit). The non-inferiority margin was 6 percentage points. Safety was assessed in all randomly assigned patients who followed the study protocol and were evaluable at the 14-day visit. This study is registered with ClinicalTrials.gov, NCT03153267.
Between July 3, 2017, and Oct 2, 2018, we enrolled 300 patients (150 per group: median age 56 years IQR 47–65; 126 45% of 300 male; skin culture positive 72 30% of 239 assessed). 295 patients completed antibiotic therapy as per protocol and 294 (98%) patients were evaluable 2 months post-enrolment. Five (3%) of 147 patients from the 7-day group versus 3 (2%) of 147 patients from the 14-day group (one patient did not attend the 2-month visit and was unreachable by telephone) had treatment failure manifesting as persistence of erythema (difference 1·4 percentage points; upper limit of one-sided 95% CI 5·2 percentage points; p=0·64). No patients developed new objective manifestations of Lyme borreliosis during follow-up or had positive repeat skin biopsies. Two (1%) of 150 patients in the 7-day and one (1%) of 150 patients in the 14-day group discontinued therapy due to adverse events.
Our data support 7 days of oral doxycycline for adult European patients with solitary erythema migrans, permitting less antibiotic exposure than current guideline-driven therapy.
Slovenian Research Agency and the University Medical Centre Ljubljana.
Background and Objective. Knee osteoarthritis is a serious epidemiological problem that causes severe pain and impairs abilities. We investigated the effects of adductor canal blockade (ACB) on ...chronic osteoarthritis knee pain, motor function, and mobility. Methods. Seventy-seven patients with chronic knee osteoarthritis pain received ultrasound-guided ACB with 14 ml 0.25% levobupivacaine and 100 mcg clonidine. At baseline and 1 month after the blockade, we assessed maximal and minimal pain intensity in the knee using a numeric rating scale (NRS) and the Knee Injury and Osteoarthritis Outcome Score (KOOS). The range of motion in extension and flexion (ROMext and ROMflex) and quadriceps muscle strength of both knees (QS), Timed Up and Go Test (TUG), and 30-Second Chair Stand Test (30CST) results were determined at baseline, 1 hour, 1 week, and 1 month after the blockade. Results. ACB with levobupivacaine and clonidine appeared to decrease pain severity (NRSmax 8.13 to 4.2, p<0.001 and NRSmin 3.32 to 1.40, p<0.001). Similarly, knee ROMext decreased from 3.90 preintervention to 2.89 postintervention at 1 month, p<0.001; ROMflex decreased from 5.70 to 3.29, p<0.001; TUG time decreased from 3.22 to 2.93, <0.001; QS increased from 18.43 to 22.77, p<0.001; CST increased from 8.23 to 10.74, p<0.001. The KOOS for pain (36.40 to 58.34), symptoms (52.55 to 64.32), activities of daily living functions (ADLs, 36.36 to 60.77), and quality of life (QoL, 17.87 to 30.97) also increased, all p<0.001. Conclusion. ACB appeared to decrease pain and increase ambulation. If our preliminary results are reproducible in a planned randomized controlled trial, ACB could be a useful adjunctive pain therapy in patients with disabling pain due to knee OA.