Detection Bias Research Articles

Implications for Validation of IMERG Satellite Precipitation in a Complex Mountainous Region

Satellite-based precipitation retrievals such as the Integrated Multi-satellite Retrievals for Global Precipitation Measurement (IMERG), provide alternative data in mountainous regions. In this study, we evaluated IMERG in the Yarlung Tsangbo Grand Canyon (YGC) using ground observations. It was found that IMERG underestimated the total rainfall primarily due to under-detection of rainfall events, with misses being more prevalent than false alarms. We analyzed the relationships between the probability of detection (POD), false alarm ratio (FAR), bias in detection (BID), and Heidke skill score (HSS) and terrain factors. It was found that the POD decreased with elevation, leading to increased underestimation of rainfall events at higher elevations, and the FAR was higher in valley sites. In terms of the hit events, IMERG overestimated the light rainfall events and underestimated the heavy rainfall events and the negative bias in the hit events decreased with elevation. IMERG could capture the early morning peak precipitation in the YGC region but underestimated the amplitude of the diurnal variation. This bias was inherent at the sensor level, and the Global Precipitation Climatology Center (GPCC) calibration partially improved the underestimation. However, this improvement was not sufficient for the YGC region. This study fills the gap in IMERG validation in a complex mountainous region and has implications for users and developers.

Studying bias in visual features through the lens of optimal transport

Computer vision systems are employed in a variety of high-impact applications. However, making them trustworthy requires methods for the detection of potential biases in their training data, before models learn to harm already disadvantaged groups in downstream applications. Image data are typically represented via extracted features, which can be hand-crafted or pre-trained neural network embeddings. In this work, we introduce a framework for bias discovery given such features that is based on optimal transport theory; it uses the (quadratic) Wasserstein distance to quantify disparity between the feature distributions of two demographic groups (e.g., women vs men). In this context, we show that the Kantorovich potentials of the images, which are a byproduct of computing the Wasserstein distance and act as “transportation prices", can serve as bias scores by indicating which images might exhibit distinct biased characteristics. We thus introduce a visual dataset exploration pipeline that helps auditors identify common characteristics across high- or low-scored images as potential sources of bias. We conduct a case study to identify prospective gender biases and demonstrate theoretically-derived properties with experiments on the CelebA and Biased MNIST datasets.

Two super-Earths at the edge of the habitable zone of the nearby M dwarf TOI-2095

The main scientific goal of TESS is to find planets smaller than Neptune around stars that are bright enough to allow for further characterization studies. Given our current instrumentation and detection biases, M dwarfs are prime targets in the search for small planets that are in (or near) the habitable zone of their host star. In this work, we use photometric observations and CARMENES radial velocity (RV) measurements to validate a pair of transiting planet candidates found by TESS. The data were fitted simultaneously, using a Bayesian Markov chain Monte Carlo (MCMC) procedure and taking into account the stellar variability present in the photometric and spectroscopic time series. We confirm the planetary origin of the two transiting candidates orbiting around TOI-2095 (LSPM J1902+7525). The star is a nearby M dwarf (d = 41.90 ± 0.03 pc, Teff = 3759 ± 87 K, V = 12.6 mag), with a stellar mass and radius of M* = 0.44 ± 0.02 M⊙ and R* = 0.44 ± 0.02 R⊙, respectively. The planetary system is composed of two transiting planets: TOI-2095b, with an orbital period of Pb = 17.66484 ± (7 × 10−5) days, and TOI-2095c, with Pc = 28.17232 ± (14 × 10−5) days. Both planets have similar sizes with Rb = 1.25 ± 0.07 R⊕ and Rc = 1.33 ± 0.08 R⊕ for planet b and planet c, respectively. Although we did not detect the induced RV variations of any planet with significance, our CARMENES data allow us to set stringent upper limits on the masses of these objects. We find Mb < 4.1 M⊕ for the inner and Mc < 7.4 M⊕ for the outer planet (95% confidence level). These two planets present equilibrium temperatures in the range of 300–350 K and are close to the inner edge of the habitable zone of their star.

Can Cold Jupiters Sculpt the Edge-of-the-multis?

Compact systems of multiple close-in super-Earths/sub-Neptunes (compact multis) are a ubiquitous outcome of planet formation. It was recently discovered that the outer edges of compact multis are located at smaller orbital periods than expected from geometric and detection biases alone, suggesting some truncation or transition in the outer architectures. Here we test whether this edge-of-the-multis might be explained in any part by distant giant planets in the outer regions (≳1 au) of the systems. We investigate the dynamical stability of observed compact multis in the presence of hypothetical giant (≳0.5 M Jup) perturbing planets. We identify what parameters would be required for hypothetical perturbing planets if they were responsible for dynamically sculpting the outer edges of compact multis. Edge-sculpting perturbers are generally in the range of P ∼ 100–500 days for the average compact multi, with most between P ∼ 200 and 300 days. Given the relatively close separation, we explore the detectability of the hypothetical edge-sculpting perturbing planets, finding that they would be readily detectable in transit and radial velocity data. We compare to observational constraints and find it unlikely that dynamical sculpting from distant giant planets contributes significantly to the edge-of-the-multis. However, this conclusion could be strengthened in future work by a more thorough analysis of the detection yields of the perturbing planets.

Polyautoimmunity in patients with cutaneous lupus erythematosus: A nationwide sex- and age-matched cohort study from Denmark

Polyautoimmunity is defined as having 2 or more autoimmune diseases. Little is known about polyautoimmunity in patients with cutaneous lupus erythematosus (CLE). To estimate prevalence and 5-year incidence of non-lupus erythematosus (LE) autoimmune diseases in patients with CLE. Patients with CLE were identified In the Danish National Patient Registry and each patient was age- and sex-matched with 10 general population controls. Outcome information on non-LE autoimmune diseases was obtained by register-linkage between Danish National Patient Registry and the National Prescription Register. The risk ratio (RR) for prevalent non-LE autoimmune disease at time of CLE diagnosis was calculated in modified Poisson regression; and hazard ratios (HRs) for incident non-LE autoimmune disease were estimated in Cox regression analyses. Overall, 1674 patients with CLE had a higher prevalence of a non-LE autoimmune disease than the comparators (18.5 vs 7.9%; RR 2.4; 95% CI, 2.1 to 2.6). Correspondingly, the cumulative incidence of a non-LE autoimmune disease during 5years of follow-up was increased for the patients with CLE: HR 3.5(95% CI, 3.0 to 4.0). Risk of detection and misclassification bias, mainly pertaining to the CLE group. Patients with CLE had higher prevalence and 5-year cumulative incidence of a non-LE autoimmune disease than the general population.

Multi-Object Tracking Based on Re-Identification Enhancement and Associated Correction

During the Multi-Object Tracking (MOT) process, the effectiveness of Re-Identification (Re-ID) is typically contingent upon the performance of detection. In particular, detection errors may result in the blurring of Re-ID features, thereby diminishing the efficacy of Re-ID. In particular, Re-ID exhibits poor robustness in intricate scenarios such as distance and scale alterations, occlusion, and crowding. Meanwhile, the frequent occurrence of object occlusion can result in the interruption of object tracking, thereby impacting the tracking efficiency. To address these issues, a novel MOT method approach that leverages Re-ID enhancement and association correction is proposed. Through an unanchored, one-stage backbone network with CenterNet as the detector, the object position is detected by predicting the heatmap of the object center point. A constraint on the Predicted Distance Difference Loss (PDDL) is then designed for the response value of the predicted heatmap, effectively enhancing the detection accuracy and reduce the detection bias. To address the issue of Re-ID reliance on detection accuracy and enhance its robustness, an Effective Feature Graph Convolutional Neural network (EFGCN) is proposed. This approach expands learnable effective features by constructing a graph model that enhances them through a GCN model of heatmap true value locations and surrounding response value locations. Additionally, a target trajectory Association Correction Module (ACM) is introduced in the association phase to handle interruptions and trajectory fragments caused by occlusion. By improving data association for tracking judgment when dealing with occluded targets and considering the relationship between trajectory fragmentation and the detection of low-scoring targets, attempts are made to continue with tracking occluded targets to form complete object associations. The proposed method is evaluated on the MOT16 and MOT17 datasets and compared with other algorithms. The results demonstrate that it is highly effective.

Efficacy and safety of atropine at different concentrations in prevention of myopia progression in Asian children: a systematic review and Meta-analysis of randomized clinical trials.

To assess the efficacy versus the adverse effects of various concentrations of atropine in the prevention of myopia in Asian children. Databases (PubMed, EMBASE, the Cochrane Library and Web of science) were comprehensively searched from inception to April 2022. Types of studies included were randomized clinical trials (RCTs). The published languages were limited to English. Two researchers assessed the quality of included studies independently using Cochrane risk of bias tool based on the Cochrane Handbook for Systematic Reviews of Interventions. Funnel plots and Egger's test were used for detection of publication bias. Meta-analyses were conducted using STATA (version 15.0; StataCorp). A total of 15 RCTs involving 2268 patients were included in the study. In the atropine group, spherical equivalent progressed at a significantly lower rate [weighted mean difference (WMD)=0.39, 95% confidence interval (CI): 0.23, 0.54] than in the control group. A WMD of 0.15 mm was associated with less axial elongation (95%CI -0.19, -0.10). Different doses showed statistically significant differences (P<0.05) and an improved effect could result from a higher concentration. Changes in photopic pupil size and mesopic pupil size in atropine group is 0.70 mm (95%CI: 0.33, 1.06) and 0.38 mm (95%CI: 0.22, 0.54) more than the control group. In the present Meta-analysis, no changes in accommodative amplitude (AA) were associated with atropine administration. Atropine administration increased the risk of adverse effects by 1.37 times. Concentrations of less than 1% atropine are able to effectively retard diopter and axis growth of myopia in Asian children in a dose-dependent manner. Meanwhile, it caused pupil enlargement, but induced no change in the AA within this range. Further study is required to determine the dosage needed to achieve maximum efficacy and minimal side effects.

Quantifying ant diversity and community in a subalpine forest mosaic: a comparison of two methods

Sampling efficiency, composition and detection biases associated with pitfall-trap and sample plot (standardised hand-collecting) methods were compared at seven high montane sites at Lasha Mountain, Yunnan, China. Rarefaction-interpolation curves showed total species diversity and sample coverage from each method were undifferentiated. Permutational multivariate analysis of variance showed highly overlapping communities. However, of all species collected, just 53% of all species found were shared between the two methods; 30% were exclusive to sample plot samples and 17% were exclusive to pitfall traps. Modelled comparisons of species richness (alpha) and proportion of species from the population (beta diversity) showed differences between methods. Permutational multivariate analysis of variance blocked by site showed community composition differed between methods, defined largely by the presence of a few dominant species.Implications for insect conservation: Our findings suggest that results from the two methods cannot be directly compared and are imperfect substitutes to one another. For long-term monitoring of biodiversity in complex forest mosaic systems, we suggest integrating a suite of complementary methods to achieve more complete representation of ant composition and diversity.

Reporting quality and risk of bias assessment of animal research on Chaihu-Shugan-San for depression: A systematic review

ObjectiveChaihu-Shugan-San (CSS) is a traditional Chinese medicine formula employed to treat depression. We aim to conduct a reporting quality assessment and risk of bias evaluation of animal research on CSS for depression. MethodsTo acquire eligible studies, two reviewers searched plentiful databases from inception to October 23rd, 2021. Reporting quality assessment and risk of bias assessment of the included animal studies were evaluated by using Animal Research: Reporting In Vivo Experiments (ARRIVE) guidelines and the SYstematic Review Centre for Laboratory animal Experimentation (SYRCLE) risk of bias tool, respectively. ResultsThe initial search identified 720 records, while only 30 studies were included. The result of the reporting quality assessment was inferior, items 17 and 19 were not reported at all. The details of five items (items 3, 6, 7, 10, and 18) were not reported. The outcome of the risk of bias assessment suggested that half of the entries (5/10) displayed an unclear risk of bias and a high risk of bias. Blinding with regard to performance bias and detection bias revealed an unclear risk of bias (100%), followed by baseline characteristics (76.67%) and sequence generation (60%). Random outcome assessment showed a high risk of bias (100%). ConclusionThe included animal studies exhibited methodological defects and imprecise reporting. Hence, the ARRIVE guidelines and SYRCLE's RoB tool should be disseminated among basic medical researchers examining CSS for depression to publish studies with low risk of bias and sufficient reporting so that the animal research can promptly be transformed into clinical research.

Representation of Motion Direction in Visual Area MT Accounts for High Sensitivity to Centripetal Motion, Aligning with Efficient Coding of Retinal Motion Statistics.

The overrepresentation of centrifugal motion in the middle temporal visual area (area MT) has long been thought to provide an efficient coding strategy for optic flow processing. However, this overrepresentation compromises the detection of approaching objects, which is essential for survival. In the present study, we revisited this long-held notion by reanalyzing motion selectivity in area MT of three macaque monkeys (two males, one female) using random-dot stimuli instead of spot stimuli. We found no differences in the number of neurons tuned to centrifugal versus centripetal motion; however, centrifugally tuned neurons showed stronger tuning than centripetally tuned neurons. This was attributed to the heightened suppression of responses in centrifugal neurons to centripetal motion compared with that of centripetal neurons to centrifugal motion. Our modeling implies that this intensified suppression accounts for superior detection performance for weak centripetal motion stimuli. Moreover, through Fisher information analysis, we establish that the population sensitivity to motion direction in peripheral vision corresponds well with retinal motion statistics during forward locomotion. While these results challenge established concepts, considering the interplay of logarithmic Gaussian receptive fields and spot stimuli can shed light on the previously documented overrepresentation of centrifugal motion. Significantly, our findings reconcile a previously found discrepancy between MT activity and human behavior, highlighting the proficiency of peripheral MT neurons in encoding motion direction efficiently.SIGNIFICANCE STATEMENT The efficient coding hypothesis states that sensory neurons are tuned to specific, frequently experienced stimuli. Whereas previous work has found that neurons in the middle temporal (MT) area favor centrifugal motion, which results from forward locomotion, we show here that there is no such bias. Moreover, we found that the response of centrifugal neurons for centripetal motion was more suppressed than that of centripetal neurons for centrifugal motion. Combined with modeling, this provides a solution to a previously known discrepancy between reported centrifugal bias in MT and better detection of centripetal motion by human observers. Additionally, we show that population sensitivity in peripheral MT neurons conforms to an efficient code of retinal motion statistics during forward locomotion.

Adaptive bolus calculators for people with type 1 diabetes: A systematic review.

To conduct a systematic review of studies assessing adaptive insulin bolus calculators for people with type 1 diabetes (T1D). Electronic databases (Medline, Embase and Web of Science) were systematically searched from date of inception to 13 October 2022 for single-arm or randomized controlled studies assessing adaptive bolus calculators only, in children or adults with T1D on multiple daily injections or insulin pumps with glycaemic outcomes reported. The Clinicaltrials.gov registry was searched for recently completed studies evaluating decision support in T1D. The quality of extracted studies was assessed using the Standard Quality Assessment criteria and the Cochrane Risk of Bias assessment tool. Six studies were identified. Extracted data were synthesized in a descriptive review because of heterogeneity. All the studies were small feasibility studies or were not suitably powered, and all were deemed to be at a high risk of performance and detection bias because they were unblinded. Overall, these studies did not show a significant glycaemic improvement. Two studies showed a reduction in postprandial time below range or an incremental change in blood glucose concentration; however, these were in controlled environments over a short duration. There are limited clinical trials evaluating adaptive bolus calculators. Although results from small trials or in-silico data are promising, further studies are required to support personalized and adaptive management of T1D.

Handling Bias in Toxic Speech Detection: A Survey

Detecting online toxicity has always been a challenge due to its inherent subjectivity. Factors such as the context, geography, socio-political climate, and background of the producers and consumers of the posts play a crucial role in determining if the content can be flagged as toxic. Adoption of automated toxicity detection models in production can thus lead to a sidelining of the various groups they aim to help in the first place. It has piqued researchers’ interest in examining unintended biases and their mitigation. Due to the nascent and multi-faceted nature of the work, complete literature is chaotic in its terminologies, techniques, and findings. In this article, we put together a systematic study of the limitations and challenges of existing methods for mitigating bias in toxicity detection. We look closely at proposed methods for evaluating and mitigating bias in toxic speech detection. To examine the limitations of existing methods, we also conduct a case study to introduce the concept of bias shift due to knowledge-based bias mitigation. The survey concludes with an overview of the critical challenges, research gaps, and future directions. While reducing toxicity on online platforms continues to be an active area of research, a systematic study of various biases and their mitigation strategies will help the research community produce robust and fair models. 1

Assessing Biases through Visual Contexts

Bias detection in the computer vision field is a necessary task, to achieve fair models. These biases are usually due to undesirable correlations present in the data and learned by the model. Although explainability can be a way to gain insights into model behavior, reviewing explanations is not straightforward. This work proposes a methodology to analyze the model biases without using explainability. By doing so, we reduce the potential noise arising from explainability methods, and we minimize human noise during the analysis of explanations. The proposed methodology combines images of the original distribution with images of potential context biases and analyzes the effect produced in the model’s output. For this work, we first presented and released three new datasets generated by diffusion models. Next, we used the proposed methodology to analyze the context impact on the model’s prediction. Finally, we verified the reliability of the proposed methodology and the consistency of its results. We hope this tool will help practitioners to detect and mitigate potential biases, allowing them to obtain more reliable models.

Donepezil combined with traditional Chinese medicine has promising efficacy on mild cognitive impairment: a systematic review and meta-analysis.

Prior research has shown mixed results regarding the effectiveness of combining donepezil and traditional Chinese medicine (TCM) to treat mild cognitive impairment (MCI). In light of this, our study aims to examine the efficacy and safety of this treatment approach for patients with MCI. We conducted a comprehensive search of various databases, including Medline (via PubMed), Cochrane, Embase, Web of Science, Chinese National Knowledge Infrastructure, Chinese Biomedical Literature Database, Chinese Scientific Journal Database, and Wanfang Database from their inception to November 16, 2022. The selection of studies, risk of bias assessment, and data extraction were carried out independently by two authors. The statistical analysis was performed using STATA. Our meta-analysis included a total of 35 studies with 2,833 patients, published between 2008 and 2022, with intervention durations ranging from 4 weeks to 12 months. However, most of the studies had a high risk of detection bias. Our findings indicated that the combination of donepezil and TCM significantly improved the Montreal Cognitive Assessment (MoCA) score (weighted mean difference [WMD] = 2.79, 95% confidence interval [CI]: 1.82 to 3.75) and the Barthel Index score (WMD = 9.20, 95% CI: 5.39 to 13.00) compared to donepezil alone. However, subgroup analyses showed that the MoCA score did not increase significantly in patients with MCI resulting from cerebrovascular disease (WMD = 1.47, 95% CI: -0.02 to 2.96). The combination of donepezil and TCM may have a more positive effect on cognitive function and activities of daily living in patients with MCI compared to the use of donepezil alone. However, due to the limited quality of the studies included in our analysis, these findings should be interpreted with caution.

Effectiveness of Telegenetic Counseling for Patients and Families With Suspected Hereditary Cancer: Systematic Review

Background Telegenetic counseling has attracted attention as a preventive measure against the recent COVID-19 pandemic. This systematic review compared telegenetic counseling using videoconferencing versus face-to-face counseling for hereditary cancer. Objective This study aimed to evaluate the effectiveness of telegenetic counseling using videoconferencing versus face-to-face counseling for people with suspected hereditary cancer. Methods A comprehensive literature search was carried out in December 2021 using the databases of the Japan Medical Abstracts Society, PubMed, CINAHL, PsycInfo, EMBASE, and the Cochrane Library. The eligibility criteria were studies in which randomized controlled trials (RCTs) or cluster RCTs were conducted among patients with suspected hereditary cancer and their families, comparing telegenetic counseling using videoconferencing with face-to-face genetic counseling. The Cochrane risk-of-bias tool was used to assess the risk of bias of each RCT study. This study was reported according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) guidelines. Results Two studies comprising a total of 191 participants were included and used for meta-analysis. Participants were individuals who lived in rural areas or in areas without genetic counselors and were suspected to have breast, colon, ovarian, multiple, or genetic cancer. All outcomes were assessed in only 1 study and their effects could not be discussed explicitly. In one study, there were no significant differences in satisfaction, psychological distress, or a number of genetic tests between face-to-face genetic counseling and telegenetic counseling. The cost per patient for genetic counseling was $106.19 for the telegenetics group and $244.33 for the in-person group. The risk of bias was high in both studies, with a high risk of performance bias, detection bias, and attrition bias. Conclusions The results of the two RCTs were described qualitatively. However, the evidence is limited because of the small number of RCT studies on telegenetic counseling and the high risk of bias. Further accumulation of studies is needed in the future. Conflicts of Interest None declared.

Effectiveness of Internet-based Psychosocial Interventions among Family Caregivers of People with Schizophrenia: A Systematic Review

Internet-based interventions have long been available in mental health care, offering alternative options for caring for patients and caregivers. This review aimed to evaluate the effectiveness of internet-based psychosocial interventions on psychological distress, caregiving burden, expressed emotion, knowledge about psychosis among family caregivers of people with schizophrenia, and hospitalization of the patients. A systematic search was conducted on June 20, 2022, in six databases (MEDLINE, EMBASE, CINAHL, Scopus, Cochrane Library, and ProQuest) from 2010 onwards in English and Thai. The review was conducted based on the JBI approach and reported following the PRISMA guidelines. The protocol was registered in PROSPERO as CRD42021255318. Five out of 20 studies were included in the review. Meta-analysis was impossible due to the methodological heterogeneity across the included studies; therefore, narrative synthesis was utilized. Psychological distress was not improved as reported in one study; however, the caregiving burden was improved in another study. Two studies reported expressed emotion; however, they were impossible to combine due to methodological differences, and the results contradicted each other. Knowledge about psychosis was also found to contradict results in two individual studies. Hospitalization was not impacted by the intervention based on one study. Four studies were randomized control trials, and another was a quasi-experimental research study. One out of four of the former did not mention blinding during randomization, which creates potential performance and detection bias. A robust randomized control trial should be carried out with a similar outcome measurement to decrease the heterogeneity and enhance the possibility for further meta-analysis and the ability to recommend for practice.

Efficacy of acupuncture on acute pharynx infections: A systematic review and meta-analysis.

Several clinical reports have focused on acupuncture for the treatment of acute pharyngeal infections. However, the efficacy and safety of acupuncture for the treatment of acute pharyngeal infections are controversial. To systematically assess the efficacy and safety of acupuncture in treating acute pharynx infections, thus providing a reference for clinical decision-making. We searched PubMed, CENTRAL, Embase, Web of Science, China National Knowledge Infrastructure, China Biomedical, clinical research registration platforms, gray literature, and reference lists of the selected studies from inception to October 30, 2022. The risk of bias assessment was performed using RevMan. The meta-analysis was performed using STATA with the Hedges' g value. We also performed a subgroup analysis, meta-regression, and publication bias detection using Harbord's and Egger's tests. We included 19 randomized controlled trials comprising 1701 patients, of which only one study had a high risk of bias. The primary outcome, i.e., the response rate, revealed that acupuncture was more effective than antibiotics. The secondary results revealed that the differences in the reduction of VAS scores, sore throat duration, and white blood cell counts were statistically significant in the acupuncture group compared with the antibiotic group. However, the difference in the modulation of the neutrophil percentage and C-reactive protein levels was insignificant. Moreover, the acupuncture treatment resulted in a lower incidence of adverse events than the antibiotic treatment. Thus, acupuncture therapy for acute pharyngeal infections is safe and its response rate is superior to that of antibiotics. Acupuncture showed positive outcomes for alleviating the sore throat symptoms, shortening the sore throat duration, and improving the immune inflammation index. Nevertheless, owing to the limitations of this study, our conclusions should be interpreted with caution. More high-quality trials are warranted in the future for improving the methodology and reporting quality.

Detecting review manipulation from behavior deviation: A deep learning approach

This research considers the problem of review manipulation by exploring rating distribution deviations and proposing a detection framework that captures review bias. We characterize the severity of review manipulation based on the degree to which the corresponding rating deviates from a normal distribution. Then we utilize deep learning approaches to learn hidden textual representations of unmanipulated reviews for biased review detection. Experiments conducted on 1,396,132 movie reviews demonstrate that highly/low biased reviews yield significant difference through our proposed model. Moreover, sentiment and topic analysis on experimental results reveal that biased reviews differentiate from authentic reviews in covering less movie-related topics and embedding stronger and more inflammatory emotions. The findings provide managerial insights for business owners in detecting review manipulation and suggest practical signals for online consumers to distinguish misleading information online.

Healthy eating interventions delivered in early childhood education and care settings for improving the diet of children aged six months to six years.

Healthy eating interventions delivered in early childhood education and care settings for improving the diet of children aged six months to six years.

Association between lithium treatment and renal, thyroid and parathyroid function: A cohort study of 6659 patients with bipolar disorder.

Although potential adverse effects of lithium treatment on renal and endocrine systems have been extensively investigated, most prior studies are limited by selected populations and short follow-up. Within the Psychiatric Services of the Central Denmark Region, we identified all patients with bipolar disorder and ≥1 serum-lithium (se-Li) measurements between January 1, 2013, and July 20, 2022, and reference patients with bipolar disorder matched on age, sex, and baseline creatinine. Outcomes were diagnoses of renal, thyroid and parathyroid disease, and blood tests measuring creatinine, estimated glomerular filtration rate (eGFR), thyroid-stimulating hormone (TSH), parathyroid hormone (PTH) and calcium. Analyses included unadjusted multilevel regression to describe changes in biochemical markers, and adjusted Cox regression to compare rates of disease/biochemical outcomes between lithium users and reference patients. Among 1646 lithium users (median age 36 years, 63% women) and 5013 reference patients, lithium users had decreasing TSH and eGFR, stable PTH, and increasing calcium levels over time. Lithium use was associated with increased rates of renal, thyroid and parathyroid disease, and levels of biochemical markers outside normal ranges (hazard rate ratios: 1.07-11.22), but the absolute number of severe outcomes was low (e.g., chronic kidney disease: N = 10, 0.6%). Notably, the rate of blood testing was substantially higher among lithium users than among reference patients (e.g., mean number of creatinine tests during the second year of follow-up: lithium users = 2.5, reference patients = 1.4). Severely adverse renal and endocrine outcomes are rare during lithium treatment. Observational studies of long-term lithium treatment are prone to detection bias.