Attempt of Bayesian Estimation from Left-censored Data Using the Markov Chain Monte Carlo Method: Exploring Cr(VI) Concentrations in Mineral Water Products

Egg protein exposure estimation in risk assessment for Japanese food allergy labeling

To assess the risk of food allergies in foods processed under the Japanese food labeling system, estimating exposure to hidden allergens is necessary. We assessed exposure to egg protein in foods processed according to the Japanese food labeling system. First, we estimated the concentration distribution of egg protein by Bayesian methods using data from the literature and the measurement of food products with precautional declarations in the labeling margin. We then estimated the food-intake portion-size distribution under two scenarios: soft drink consumption as an example of single, high-intake consumption, and confections, which are frequently consumed by children, as a realistic example of low-intake consumption. Finally, we estimated the distribution of unexpected intake of egg proteins in the form of single consumption. The mean exposure to egg protein under the high-intake scenario was estimated to be 0.0164 mg for 1-15-year-olds, 0.0171 mg for 4-15-year-olds, 0.0181 mg for 7-15-year-olds, and ≥0.0188 mg for 16-year-olds. The mean exposure to egg protein under the low-intake scenario was estimated to be 0.0018 mg for 1-15-year-olds, 0.0019 mg for 4-15-year-olds, 0.0020 mg for 7-15-year-olds, and ≥0.0022 mg for 16-year-olds. Compared to the reference dose of 2.0 mg proposed by the Joint the Food and Agriculture Organization (FAO)/World Health Organization (WHO) Expert Committee, the risk of onset of food allergies due to egg protein contamination from foods without egg labeling is considered to be extremely low under the current Japanese food labeling system.

Frequentist and Bayesian tolerance intervals for setting specification limits for left-censored gamma distributed drug quality attributes

Tolerance intervals from quality attribute measurements are used to establish specification limits for drug products. Some attribute measurements may be below the reporting limits, that is, left-censored data. When data has a long, right-skew tail, a gamma distribution may be applicable. This paper compares maximum likelihood estimation (MLE) and Bayesian methods to estimate shape and scale parameters of censored gamma distributions and to calculate tolerance intervals under varying sample sizes and extents of censoring. The noninformative reference prior and the maximal data information prior (MDIP) are used to compare the impact of prior choice. Metrics used are bias and root mean square error for the parameter estimation and average length and confidence coefficient for the tolerance interval evaluation. It will be shown that Bayesian method using a reference prior overall performs better than MLE for the scenarios evaluated. When sample size is small, the Bayesian method using MDIP yields conservatively too wide tolerance intervals that are unsuitable basis for specification setting. The metrics for all methods worsened with increasing extent of censoring but improved with increasing sample size, as expected. This study demonstrates that although MLE is relatively simple and available in user-friendly statistical software, it falls short in accurately and precisely producing tolerance limits that maintain the stated confidence depending on the scenario. The Bayesian method using noninformative prior, even though computationally intensive and requires considerable statistical programming, produces tolerance limits which are practically useful for specification setting. Real-world examples are provided to illustrate the findings from the simulation study.

Bayesian-Based Probabilistic Risk Assessment of Fipronil in Food: A Case Study in Taiwan

Fipronil, a broad-spectrum insecticide, is widely used in agriculture and veterinary practices. Fipronil-induced neurotoxicity and potential adverse effects on humans and aquatic organisms have raised health concerns. Monitoring programs have been implemented globally to assess fipronil residues in food, including fruits, vegetables, and animal products. However, previous exposure assessments have often focused on specific food categories or subsets of items, resulting in limited insights into the overall health risks. Additionally, the large number of non-detect fipronil residues in food has introduced uncertainties in exposure assessment. To address these issues, a probabilistic exposure assessment and dose-response analysis were adopted in this study, considering the sample distribution below the detection limit to better characterize uncertainties and population variability in health risk assessments. The estimated fipronil exposure to the general public ranges from 6.38 × 10-6 ± 0.00017 mg/kg/day to 9.83 × 10-6 ± 0.00034 mg/kg/day. Only one out of 200,000 simulated individuals had a fipronil dose exceeding the probabilistic reference dose (0.048 mg/kg/day, pRfD), which aims to protect 99% of the population with effects less than 10% extra risk. By incorporating uncertainties in exposure and dose-response data, a more comprehensive understanding of the health risks associated with fipronil exposure in the Taiwanese population has been achieved.

Multisensory Integration Dominates Hypnotisability and Expectations in the Rubber Hand Illusion

Some recent papers by P. Lush and colleagues have argued that the rubber hand illusion (RHI), where participants can feel a rubber hand as their own under appropriate multisensory stimulation, may be caused mainly by hypnotic suggestibility and expectations (demand characteristics). These papers rely primarily on a study with 353 participants who took part in a RHI experiment carried out in a classical way with brush stroking. Participants experienced a synchronous condition where the rubber hand was seen to be touched in synchrony with touch felt on their corresponding hidden real hand, or the touches were applied asynchronously as a control. Each participant had a related measure of their hypnotisability on a scale known as the Sussex-Waterloo Scale of Hypnotisability (SWASH). The authors found a correlation between the questionnaire ratings of the RHI in the synchronous condition and the SWASH score. From this, they concluded that the RHI is largely driven by suggestibility and further proposed that suggestibility and expectations may even entirely explain the RHI. Here we examine their claims in a series of extensive new analyses of their data. We find that at every level of SWASH, the synchronous stimulation results in greater levels of the illusion than the asynchronous condition; moreover, proprioceptive drift is greater in the synchronous case at every level of SWASH. Thus, while the level of hypnotisability does modestly influence the subjective reports (higher SWASH is associated with somewhat higher illusion ratings), the major difference between the synchronous and asynchronous stimulation is always present. Furthermore, by including in the model the participants' expectancy ratings of how strongly they initially believed they would experience the RHI in the two conditions, we show that expectations had a very small effect on the illusion ratings; model comparisons further demonstrate that the multisensory condition is two-to-three-times as dominant as the other factors, with hypnotisability contributing modestly and expectations negligibly. Thus, although the results indicate that trait suggestibility may modulate the RHI, presumably through intersubject variations in top-down factors, the findings also suggest that the primary explanation for the RHI is as a multisensory bodily illusion.