Drug-induced orthostatic hypotension (OH) is common, and its resulting cerebral hypoperfusion is linked to adverse outcomes including falls, strokes, cognitive impairment, and increased mortality. The extent to which specific medications are associated with OH remains unclear.We conducted a systematic review and meta-analysis to evaluate the extent to which specific drug groups are associated with OH. EMBASE, MEDLINE, and Web of Science databases were searched from inception through 23 November 2020. Placebo-controlled randomised controlled trials (RCTs) on any drug reporting on OH as an adverse effect in adults (≥18 years) were eligible. Three authors extracted data on the drug, OH, dose, participant characteristics, and study setting. The revised Cochrane risk-of-bias tool for randomised trials (RoB 2) was used to appraise evidence. Summary odds ratios (ORs) were estimated for OH using fixed effects Mantel-Haenszel statistics. We conducted subgroup analysis on validity of OH measurement, drug dose, risk of bias, age, and comorbidity. The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) tool was used to summarise the certainty of evidence. Of 36,940 citations, 69 eligible RCTs were included in the meta-analysis comprising 27,079 participants. Compared with placebo, beta-blockers and tricyclic antidepressants were associated with increased odds of OH (OR 7.76 [95% CI 2.51, 24.03]; OR 6.30 [95% CI 2.86, 13.91]). Alpha-blockers, antipsychotics, and SGLT-2 inhibitors were associated with up to 2-fold increased odds of OH, compared to placebo. There was no statistically significant difference in odds of OH with vasodilators (CCBs, ACE inhibitors/ARBs, SSRIs), compared to placebo. Limitations of this study are as follows: data limited to placebo-controlled studies, (excluding head-to-head trials), many RCTs excluded older participants; therefore results may be amplified in older patients in the clinical setting. The study protocol is publicly available on PROSPERO (CRD42020168697).Medications prescribed for common conditions (including depression, diabetes, and lower urinary tract symptoms) were associated with significantly increased odds of OH. Drugs causing sympathetic inhibition were associated with significantly increased odds of OH, while most vasodilators were associated with small nonsignificant differences in odds of OH, compared to placebo. Drugs targeting multiple parts of the orthostatic blood pressure (BP) reflex pathway (e.g. sympathetic inhibition, vasodilation, cardio-inhibitory effects) may carry cumulative risk, suggesting that individuals with polypharmacy could benefit from postural BP monitoring.
Precision medicine is the next frontier in pharmaceutical research, aiming to improve the safety and efficacy of therapeutics for patients. However, little progress has been made in this domain. Herein, we combined two cutting-edge technologies to demonstrate their potential for achieving programmable controlled release. A drug delivery system (DDS) was formulated using conductive polymers (CPs) that provided temporal controlled release over drug release. Three-dimensional (3D) printing was used to ensure dimensional control over the design of the DDS. The CP used herein is known to be fragile, and thus was blended with thermoplastic polyurethane (TPU) to achieve a conductive elastomer with sound mechanical properties. Rheological and mechanical analyses were performed, where it was revealed that formulation inks with a storage modulus in the order of 103 - 104 Pa were both extrudable and maintained their structural integrity. Physico-chemical analysis confirmed the presence of the CP functional groups in the 3D printed DDS. Cyclic voltammetry demonstrated that the DDS remained conductive for 100 stimulations. in vitro drug release was performed for 180 mins at varying voltages, where a significant difference (p < 0.05) in cumulative release was observed between either ±1.0 V and passive release. Furthermore, the responsiveness of the DDS to pulsatile stimuli was tested, where it was found to rapidly respond to the voltage stimuli, consequently altering the release mechanism. The study is the first to 3D print electro-active medicines using CPs and paves the way for digitalising DDS that can be integrated into the Internet of Things (IoT) framework.
Presenting many advantages, solid oral dosage forms (SODFs) are widely manufactured and frequently prescribed in older populations regardless of the specific characteristics of patients. Commonly, patients with dysphagia (swallowing disorders) experience difficulties taking SODFs, which may lead to non-adherence or misuse. SODF characteristics (e.g., size, shape, thickness) are likely to influence swallowability. Herein, we used the acceptability reference framework (the ClinSearch acceptability score test (CAST))—a 3D-map juxtaposing two acceptability profiles—to investigate the impact of tablet size on acceptability. We collected 938 observer reports on the tablet intake by patients ≥65 years in hospitals or care homes. As we might expect, tablets could be classified as accepted in older patients without dysphagia (n = 790), while not in those with swallowing disorders (n = 146). However, reducing the tablet size had a significant impact on acceptability in this subpopulation: tablets <6.5 mm appeared to be accepted by patients with swallowing disorders. Among the 309 distinct tablets assessed in this study, ranging in size from 4.7 to 21.5 mm, 83% are ≥6.5 mm and consequently may be poorly accepted by institutionalized older people and older inpatients suffering from dysphagia. This underlines the need to develop and prescribe medicines with the best adapted characteristics to reach an optimal acceptability in targeted users.
Anemia is a prevalent medical condition that typically requires invasive blood tests for diagnosis and monitoring. Electronic health records (EHRs) have emerged as valuable data sources for numerous medical studies. EHR-based hemoglobin level/anemia degree prediction is non-invasive and rapid but still faces some challenges due to the fact that EHR data is typically an irregular multivariate time series containing a significant number of missing values and irregular time intervals. To address these issues, we introduce HgbNet, a machine learning-based prediction model that emulates clinicians' decision-making processes for hemoglobin level/anemia degree prediction. The model incorporates a NanDense layer with a missing indicator to handle missing values and employs attention mechanisms to account for both local irregularity and global irregularity. We evaluate the proposed method using two real-world datasets across two use cases. In our first use case, we predict hemoglobin level/anemia degree at moment T+1 by utilizing records from moments prior to T+1. In our second use case, we integrate all historical records with additional selected test results at moment T+1 to predict hemoglobin level/anemia degree at the same moment, T+1. HgbNet outperforms the best baseline results across all datasets and use cases. These findings demonstrate the feasibility of estimating hemoglobin levels and anemia degree from EHR data, positioning HgbNet as an effective non-invasive anemia diagnosis solution that could potentially enhance the quality of life for millions of affected individuals worldwide. To our knowledge, HgbNet is the first machine learning model leveraging EHR data for hemoglobin level/anemia degree prediction.
Download This Paper Open PDF in Browser Add Paper to My Library Share: Permalink Using these links will ensure access to this page indefinitely Copy URL Copy DOI
Multimodal machine learning with missing modalities is an increasingly relevant challenge arising in various applications such as healthcare. This paper extends the current research into missing modalities to the low-data regime, i.e., a downstream task has both missing modalities and limited sample size issues. This problem setting is particularly challenging and also practical as it is often expensive to get full-modality data and sufficient annotated training samples. We propose to use retrieval-augmented in-context learning to address these two crucial issues by unleashing the potential of a transformer's in-context learning ability. Diverging from existing methods, which primarily belong to the parametric paradigm and often require sufficient training samples, our work exploits the value of the available full-modality data, offering a novel perspective on resolving the challenge. The proposed data-dependent framework exhibits a higher degree of sample efficiency and is empirically demonstrated to enhance the classification model's performance on both full- and missing-modality data in the low-data regime across various multimodal learning tasks. When only 1% of the training data are available, our proposed method demonstrates an average improvement of 6.1% over a recent strong baseline across various datasets and missing states. Notably, our method also reduces the performance gap between full-modality and missing-modality data compared with the baseline.
Parkinson's disease (PD) patients display a combination of motor and non-motor symptoms. The most common non-motor symptom is scent (olfactory) impairment, occurring at least four years prior to motor symptom onset. Recent and growing interest in digital healthcare technology used in PD has resulted in more technologies developed for motor rather than non-motor symptoms. Human-computer interaction (HCI), which uses computer technology to explore human activity and work, could be combined with digital healthcare technologies to better understand and support olfaction via scent training - leading to the development of a scent-delivery device (SDD). In this pilot study, three PD patients were invited to an online focus group to explore the association between PD and olfaction, understand HCI and sensory technologies and were demonstrated a new multichannel SDD with an associated mobile app. Participants had a preconceived link, a result of personal experience, between olfactory impairment and PD. Participants felt that healthcare professionals did not take olfactory dysfunction concerns seriously prior to PD diagnosis. Two were not comfortable with sharing scent loss experiences with others. Participants expected the multichannel SDD to be small, portable and easy-to-use, with customisable cartridges to deliver chosen scents and the mobile app to create a sense of community. None of the participants regularly performed scent training but would consider doing so if some scent function could be regained. Standardised digital SDDs for regular healthcare check-ups may facilitate improvement in olfactory senses in PD patients and potential earlier PD diagnosis, allowing earlier therapeutic and symptomatic PD management.
Abstract Tacrolimus, commonly used after organ transplantation, has a narrow therapeutic index (NTI) and a high risk of dosing errors, with over 50% of patients suffering serious, potentially life-threatening adverse effects. To address this, we developed an interpretable long short-term memory (LSTM) model trained on 14 years of data from 1,774 patients across Spain and the United States. This LSTM predicts tacrolimus levels using clinical data, with approximately 5% mean absolute error, outperforming state-of-the-art models. Addressing data privacy concerns, we demonstrate that synthetic data can replace real data for training without performance loss. Finally, we launched dosetailor.com, a platform providing personalised tacrolimus dosing recommendations for clinicians. Although focused on tacrolimus, this platform forms a foundation for personalising dosage for high-risk, NTI drugs.
'%3e%3cpath fill='%23012169' d='M0 0v30h60V0z'/%3e%3cpath stroke='%23fff' stroke-width='6' d='m0 0 60 30m0-30L0 30'/%3e%3cpath stroke='%23C8102E' stroke-width='4' d='m0 0 60 30m0-30L0 30' clip-path='url(%23b)'/%3e%3cpath stroke='%23fff' stroke-width='10' d='M30 0v30M0 15h60'/%3e%3cpath stroke='%23C8102E' stroke-width='6' d='M30 0v30M0 15h60'/%3e%3c/g%3e%3c/svg%3e)
'/%3e%3cuse xlink:href='%23a' transform='rotate(144 450 300)'/%3e%3cuse xlink:href='%23a' transform='rotate(216 450 300)'/%3e%3cuse xlink:href='%23a' transform='rotate(288 450 300)'/%3e%3c/svg%3e)
