Background

Statistical models to preoperatively predict patients' risk of death and major complications after total joint arthroplasty (TJA) could improve the quality of preoperative management and informed consent. Although risk models for TJA exist, they have limitations including poor transparency and/or unknown or poor performance. Thus, it is currently impossible to know how well currently available models predict short-term complications after TJA, or if newly developed models are more accurate. We sought to develop and conduct cross-validation of predictive risk models, and report details and performance metrics as benchmarks.

Methods

Over 90 preoperative variables were used as candidate predictors of death and major complications within 30 days for Veterans Health Administration patients with osteoarthritis who underwent TJA. Data were split into 3 samples-for selection of model tuning parameters, model development, and cross-validation. C-indexes (discrimination) and calibration plots were produced.

Results

A total of 70,569 patients diagnosed with osteoarthritis who received primary TJA were included. C-statistics and bootstrapped confidence intervals for the cross-validation of the boosted regression models were highest for cardiac complications (0.75; 0.71-0.79) and 30-day mortality (0.73; 0.66-0.79) and lowest for deep vein thrombosis (0.59; 0.55-0.64) and return to the operating room (0.60; 0.57-0.63).

Conclusions

Moderately accurate predictive models of 30-day mortality and cardiac complications after TJA in Veterans Health Administration patients were developed and internally cross-validated. By reporting model coefficients and performance metrics, other model developers can test these models on new samples and have a procedure and indication-specific benchmark to surpass.

The Epilepsy Bioinformatics Study for Antiepileptogenic Therapy (EpiBioS4Rx) is an international, multicenter, multidisciplinary study aimed at preventing epileptogenesis (EpiBioS4Rx: https://epibios.loni.usc.edu/). One of the study's major objectives is the discovery of diagnostic, prognostic, and predictive plasma protein and microRNA (miRNA) biomarkers that are sensitive, specific, and translatable to the human condition. Epilepsy due to structural brain abnormalities, secondary to neurological insults such as traumatic brain injury (TBI), currently represents ∼50% of all epilepsy cases. In the preclinical EpiBioS4Rx study, TBI was induced in adult male Sprague Dawley rats using a standardized protocol for lateral fluid-percussion injury. Whole blood was collected from the tail vein at baseline and 2, 9 and 30 days post-injury and processed for plasma separation. Biomaterial properties, sample preparation and integrity, and choice of analysis platform can significantly impact measured marker levels and, in turn, interpretation with respect to injury and/or other variables. We present here the results of procedural harmonization for the first 320 rats included in the EpiBioS4Rx study study, from three international research centers, and preliminary proteomic and miRNA analyses. We also discuss experimental considerations for establishing rigorous quality controls with the goal of harmonizing operating procedures across study sites, and delivering high-quality specimens for preclinical biomarker discovery in a rat model of post-traumatic epilepsy (PTE).

We describe the outcome of a large international interlaboratory study of the measurement of particle number concentration of colloidal nanoparticles, project 10 of the technical working area 34, "Nanoparticle Populations" of the Versailles Project on Advanced Materials and Standards (VAMAS). A total of 50 laboratories delivered results for the number concentration of 30 nm gold colloidal nanoparticles measured using particle tracking analysis (PTA), single particle inductively coupled plasma mass spectrometry (spICP-MS), ultraviolet-visible (UV-Vis) light spectroscopy, centrifugal liquid sedimentation (CLS) and small angle X-ray scattering (SAXS). The study provides quantitative data to evaluate the repeatability of these methods and their reproducibility in the measurement of number concentration of model nanoparticle systems following a common measurement protocol. We find that the population-averaging methods of SAXS, CLS and UV-Vis have high measurement repeatability and reproducibility, with between-labs variability of 2.6%, 11% and 1.4% respectively. However, results may be significantly biased for reasons including inaccurate material properties whose values are used to compute the number concentration. Particle-counting method results are less reproducibile than population-averaging methods, with measured between-labs variability of 68% and 46% for PTA and spICP-MS respectively. This study provides the stakeholder community with important comparative data to underpin measurement reproducibility and method validation for number concentration of nanoparticles.