Category: Statistics

Methods Inf Med. 2024 Aug 13. doi: 10.1055/a-2385-1355. Online ahead of print.

ABSTRACT

BACKGROUND: Synthetic data has been proposed as a solution for sharing anonymized versions of sensitive biomedical datasets. Ideally, synthetic data should preserve the structure and statistical properties of the original data, while protecting the privacy of the individual subjects. Differential privacy (DP) is currently considered the gold standard approach for balancing this trade-off.

OBJECTIVES: The aim of this study is to investigate how trustworthy are group differences discovered by independent sample tests from DP-synthetic data. The evaluation is carried out in terms of the tests’ Type I and Type II errors. With the former, we can quantify the tests’ validity i.e. whether the probability of false discoveries is indeed below the significance level, and the latter indicates the tests’ power in making real discoveries.

METHODS: We evaluate the Mann-Whitney U test, Student’s t-test, chi-squared test and median test on DP-synthetic data. The private synthetic datasets are generated from real-world data, including a prostate cancer dataset (n=500) and a cardiovascular dataset (n=70 000), as well as on bivariate and multivariate simulated data. Five different DP-synthetic data generation methods are evaluated, including two basic DP histogram release methods and MWEM, Private-PGM, and DP GAN algorithms.

CONCLUSION: A large portion of the evaluation results expressed dramatically inflated Type I errors, especially at levels of ϵ≤1. This result calls for caution when releasing and analyzing DP-synthetic data: low p-values may be obtained in statistical tests simply as a byproduct of the noise added to protect privacy. A DP smoothed histogram-based synthetic data generation method was shown to produce valid Type I error for all privacy levels tested but required a large original dataset size and a modest privacy budget (ϵ≥ 5) in order to have reasonable Type II error levels.

PMID:39137913 | DOI:10.1055/a-2385-1355

Am J Perinatol. 2024 Aug 13. doi: 10.1055/s-0044-1789018. Online ahead of print.

ABSTRACT

OBJECTIVE: We aim to examine the population-level rates of induction, stillbirth, perinatal mortality, and neonatal death before and after the ARRIVE (A Randomized Trial of Induction Versus Expectant Management) trial.

STUDY DESIGN: This study was a cross-sectional analysis of publicly available U.S. Live Birth data linked with Infant Death and Fetal Death certificate data from National Vital Statistics Online. We limited analyses to nulliparous individuals with singleton pregnancy and cephalic presentation who delivered at 39 weeks or greater. The pre- and post-ARRIVE periods spanned from August 2016 to July 2018, and from January 2019 to December 2020, respectively. Our primary outcome was a stillbirth. Secondary outcomes included induction of labor, perinatal mortality, and neonatal death. Outcomes were compared between the pre- and post-ARRIVE periods. Modified Poisson regression was used to calculate adjusted relative risks (aRRs).

RESULTS: Of 2,817,071 births, there were 1,454,346 births in the pre-ARRIVE period and 1,362,725 in the post-ARRIVE period; there were 1,196 and 1,062 stillbirths in the pre- and post-ARRIVE periods, respectively. Compared to the pre-ARRIVE period, the post-ARRIVE period was not associated with a significant decrease in the risk of stillbirth at 39 weeks or greater (aRR 0.92 [95% confidence interval {95% CI} 0.85-1.00]) and stillbirth at 40 weeks or greater (aRR 0.92 [95% CI 0.82-1.04]). Compared to the pre-ARRIVE trial period, the post-ARRIVE trial was associated with increased rates of induction of labor at 39 weeks (aRR 1.37 [95% CI 1.37-1.38]) and 40 weeks (aRR 1.24 [95% CI 1.24-1.25]. Similar to stillbirth, there was no significant decrease in the risk of perinatal mortality at 39 weeks or greater or 40 weeks or greater. There was also no statistically significant change in neonatal death rates at 39 weeks or greater or at 40 weeks or greater.

CONCLUSION: The increase in induction of labor at 39 weeks was not large enough to impact the stillbirth rate at 39 weeks or greater.

KEY POINTS: · Post-ARRIVE trial, rate of induction of labor increased at 39 and 40 weeks.. · Post-ARRIVE trial, stillbirth and perinatal mortality rates remained unchanged.. · Induction rate rise post-ARRIVE trial did not impact neonatal death rates..

PMID:39137898 | DOI:10.1055/s-0044-1789018

Curr Probl Cardiol. 2024 Aug 11:102780. doi: 10.1016/j.cpcardiol.2024.102780. Online ahead of print.

ABSTRACT

BACKGROUND: Acute coronary syndrome characterizes the spectrum of myocardial ischemia states, which include non-ST elevated myocardial infarction (MI), and angina.

OBJECTIVE: The aim of the present study is to determine the effectiveness of foot massage on selected physiological parameters among patients with acute coronary syndrome.

METHODS: A quasi- experimental study was conducted in Karbala center for cardiac diseases and surgery from December 25th, 2023, to May 7th, 2024. A nonprobability purposive sampling consisted of 60 patients with acute coronary syndrome in the intervention group were instructed to performed four-step foot massage similarly 5 minutes for each foot. While patients in the control group just received routine medical treatment. The physiological parameters were checked in the two groups before, after 5 min, and 10 min after the foot massage. The study instrument consisted of two main parts: part one included patient’s socio demographics and clinical data, and the second part was used to assess the physiological parameters .Statistical tests were conducted using the software SPSS, version 24, with a level of significance of 5% (p value < 0.05).

RESULTS: that are a significant statistical differences between the mean of the selected physiological parameters readings for the study group except reading of pulse pressure, while there is no significant statistical difference between the mean of the readings of the selected physiological parameters for the control group except the reading of heart rate.

CONCLUSION: The study found that the foot massage is effective for improving the blood pressure, heart rate, mean arterial pressure.

PMID:39137879 | DOI:10.1016/j.cpcardiol.2024.102780

Brain Res. 2024 Aug 11:149158. doi: 10.1016/j.brainres.2024.149158. Online ahead of print.

ABSTRACT

The intersection of neuroscience and technology hinges on the development of wearable devices and electrodes that can augment brain networks to improve cognitive capabilities such as learning and concentration. The capacity to enhance networks associated with these functions above baseline capabilities, holds the potential to benefit numerous individuals. The purpose of this study was to determine if electromagnetic field exposure modeled from physiological data would increase instances of flow in participants playing a computer game. The flow state refers to a subjective state of optimal performance experienced by individuals during a variety of tasks. For this study, participants (n = 39, 18-65 years, nfemale = 20) played the arcade game Snake for two ten-minute periods (each with a ten-minute rest period immediately following). For one of the trials, an electromagnetic field was applied bilaterally to the temporal lobes, with the other serving as the control. Brain activity was measured using quantitative electroencephalography, flow experience was measured using the Flow Short Scale and game play scores were also recorded. Results showed deceased beta 1 (12-16 Hz) activity in the left cuneus [t = 4.650, p < 0.01] and left precuneus [t = 4.603, p < 0.01], left posterior cingulate [t = 4.521, p < 0.05], insula [t = 4.234, p < 0.05], and parahippocampal gyrus [t = 4.113, p < 0.05] for trials when the field was active, compared to controls during rest periods. Results from the Flow Short Scale showed a statistically significant difference in mean “concentration ease” scores across electromagnetic field conditions, irrespective of difficulty [t = 2.131, p < 0.05]. In the EMF exposure trials, there was no discernible experience effect; participants with prior experience in the game Snake did not exhibit significantly better performance compared to those without prior experience. This anticipated effect was observed in control conditions. The comparable performance observed between novices and experienced players in the EMF condition indicate a noteworthy learning curve for novices. In all, these results provide evidence supporting the ability of EMF patterned from amygdaloid firing (6-20 Hz) to elicit neurological correlates of flow in brain regions previously reported in the literature, facilitate concentration, and subtly improve game scores. The possibility for wearable devices to support learning, concentration, and focus are discussed.

PMID:39137825 | DOI:10.1016/j.brainres.2024.149158

J Korean Med Sci. 2024 Aug 12;39(31):e222. doi: 10.3346/jkms.2024.39.e222.

ABSTRACT

BACKGROUND: Migraine presents a significant global health problem that emphasizes the need for efficient acute treatment options. Triptans, introduced in the early 1990s, have substantially advanced migraine management owing to their effectiveness compared to that of traditional medications. However, data on triptan use in migraine management from Asian countries, where migraines tend to have milder symptoms than those in European and North American countries, are limited. This study aimed to identify the trends in triptan usage in Korea.

METHODS: This retrospective cohort study used data from the Korean National Health Insurance Service-National Sample Cohort spanning from 2002 to 2019. Patients with migraine were identified using the International Classification of Diseases 10th revision codes, and triptan prescriptions were evaluated annually in terms of quantity, pills per patient, and associated costs. The distribution of triptan prescriptions across different medical specialties was also examined. Factors contributing to the odds of triptan use were analyzed using multivariable logistic regression.

RESULTS: From 2002 to 2019, the total number of triptan tablets, prescriptions, and patients using triptans increased by 24.0, 17.1, and 13.6 times, respectively, with sumatriptan being the most frequently prescribed type of triptan. Additionally, the number of prescriptions and related costs have consistently increased despite stable pricing because of government regulation. By 2019, only approximately one-tenth of all patients with migraines had been prescribed triptans, although there was a notable increase in prescriptions over the study period. These prescription patterns varied according to the physician’s specialty. After adjusting for patient-specific factors including age and sex, the odds of prescribing triptans were higher for neurologists than for internal medicine physicians (odds ratio 2.875, P < 0.001), while they were lower for general practitioners (odds ratio 0.220, P < 0.001).

CONCLUSION: The findings revealed an increasing trend in triptan use among individuals with migraines in Korea, aligning with global usage patterns. Despite these increases, the overall prescription rate of triptans remains low, indicating potential underutilization and highlighting the need for improved migraine management strategies across all medical fields. Further efforts are necessary to optimize the use of triptans in treating migraines effectively.

PMID:39137809 | DOI:10.3346/jkms.2024.39.e222

Phys Med Biol. 2024 Aug 13. doi: 10.1088/1361-6560/ad6eda. Online ahead of print.

ABSTRACT

OBJECTIVE: Multi-energy CT conducted by photon-counting detector has a wide range of applications, especially in multiple contrast agents imaging. However, static multi-energy (SME) CT imaging suffers from higher statistical noise because of increased energy bins with static energy thresholds. Our team has proposed a dynamic dual-energy (DDE) CT detector model and the corresponding iterative reconstruction algorithm to solve this problem. However, rigorous and detailed analysis of the statistical noise characterization in this DDE CT was lacked.

APPROACH: Starting from the properties of the Poisson random variable, this paper analyzes the noise characterization of the DDE CT and compares it with the SME CT. It is proved that the multi-energy CT projections and reconstruction images calculated from the proposed DDE CT algorithm have less statistical noise than that of the SME CT.

MAIN RESULTS: Simulations and experiments verify that the expectations of the multi-energy CT projections calculated from DDE CT are the same as those of the SME projections. Still, the variance of the former is smaller. We further analyze the convergence of the iterative DDE CT algorithm through simulations and prove that the derived noise characterization can be realized under different CT imaging configurations.

SIGNIFICANCE: The low statistical noise characteristics demonstrate the value of DDE CT imaging technology.

PMID:39137803 | DOI:10.1088/1361-6560/ad6eda

Biomed Phys Eng Express. 2024 Aug 13. doi: 10.1088/2057-1976/ad6e87. Online ahead of print.

ABSTRACT

Investigating U-Net model robustness in medical image synthesis against adversarial perturbations, this study introduces RobMedNAS, a neural architecture search strategy for identifying resilient U-Net configurations. Through retrospective analysis of synthesized CT from MRI data, employing Dice coefficient and mean absolute error metrics across critical anatomical areas, the study evaluates traditional U-Net models and RobMedNAS-optimized models under adversarial attacks. Findings demonstrate RobMedNAS’s efficacy in enhancing U-Net resilience without compromising on accuracy, proposing a novel pathway for robust medical image processing.&#xD.

PMID:39137798 | DOI:10.1088/2057-1976/ad6e87

Lancet Psychiatry. 2024 Aug 5:S2215-0366(24)00188-3. doi: 10.1016/S2215-0366(24)00188-3. Online ahead of print.

ABSTRACT

BACKGROUND: Older people admitted to hospital in an emergency often have prolonged inpatient stays that worsen their outcomes, increase health-care costs, and reduce bed availability. Growing evidence suggests that the biopsychosocial complexity of their problems, which include cognitive impairment, depression, anxiety, multiple medical illnesses, and care needs resulting from functional dependency, prolongs hospital stays by making medical treatment less efficient and the planning of post-discharge care more difficult. We aimed to assess the effects of enhancing older inpatients’ care with Proactive Integrated Consultation-Liaison Psychiatry (PICLP) in The HOME Study. We have previously described the benefits of PICLP reported by patients and clinicians. In this Article, we report the effectiveness and cost-effectiveness of PICLP-enhanced care, compared with usual care alone, in reducing time in hospital.

METHODS: We did a parallel-group, multicentre, randomised controlled trial in 24 medical wards of three English acute general hospitals. Patients were eligible to take part if they were 65 years or older, had been admitted in an emergency, and were expected to remain in hospital for at least 2 days from the time of enrolment. Participants were randomly allocated to PICLP or usual care in a 1:1 ratio by a database software algorithm that used stratification by hospital, sex, and age, and randomly selected block sizes to ensure allocation concealment. PICLP clinicians (consultation-liaison psychiatrists supported by assisting clinicians) made proactive biopsychosocial assessments of patients’ problems, then delivered discharge-focused care as integrated members of ward teams. The primary outcome was time spent as an inpatient (during the index admission and any emergency readmissions) in the 30 days post-randomisation. Secondary outcomes were the rate of discharge from hospital for the total length of the index admission; discharge destination; the length of the index admission after random allocation truncated at 30 days; the number of emergency readmissions to hospital, the number of days spent as an inpatient in an acute general hospital, and the rate of death in the year after random allocation; the patient’s experience of the hospital stay; their view on the length of the hospital stay; anxiety (Generalized Anxiety Disorder-2); depression (Patient Health Questionnaire-2); cognitive function (Montreal Cognitive Assessment-Telephone version); independent functioning (Barthel Index of Activities of Daily Living); health-related quality of life (five-level EuroQol five-dimension questionnaire); and overall quality of life. Statisticians and data collectors were masked to treatment allocation; participants and ward staff could not be. Analyses were intention-to-treat. The trial had a patient and public involvement panel and was registered with ISRTCN (ISRCTN86120296).

FINDINGS: 2744 participants (1399 [51·0%] male and 1345 [49·0%] female) were enrolled between May 2, 2018, and March 5, 2020; 1373 were allocated to PICLP and 1371 to usual care. Participants’ mean age was 82·3 years (SD 8·2) and 2565 (93·5%) participants were White. The mean time spent in hospital in the 30 days post-randomisation (analysed for 2710 [98·8%] participants) was 11·37 days (SD 8·74) with PICLP and 11·85 days (SD 9·00) with usual care; adjusted mean difference -0·45 (95% CI -1·11 to 0·21; p=0·18). The only statistically and clinically significant difference in secondary outcomes was the rate of discharge, which was 8.5% higher (rate ratio 1·09 [95% CI 1·00 to 1·17]; p=0·042) with PICLP-a difference most apparent in patients who stayed for more than 2 weeks. Compared with usual care, PICLP was estimated to be modestly cost-saving and cost-effective over 1 and 3, but not 12, months. No intervention-related serious adverse events occurred.

INTERPRETATION: This is the first randomised controlled trial of PICLP. PICLP is experienced by older medical inpatients and ward staff as enhancing medical care. It is also likely to be cost-saving in the short-term. Although the trial does not provide strong evidence that PICLP reduces time in hospital, it does support and inform its future development and evaluation.

FUNDING: UK National Institute for Health and Care Research.

PMID:39137790 | DOI:10.1016/S2215-0366(24)00188-3

Am J Trop Med Hyg. 2024 Aug 13:tpmd230891. doi: 10.4269/ajtmh.23-0891. Online ahead of print.

ABSTRACT

The Ministry of Public Health and Population in Haiti is committed to malaria elimination. In 2017, we used novel methods to conduct a census, monitor progress, and return to sampled households (HH) before a cross-sectional survey in La Chapelle and Verrettes communes in Artibonite department (“the 2017 Artibonite HH census”). Geospatial PDFs with digitized structures and basemaps were loaded onto tablets. Enumerators captured GPS coordinates and details of each HH and points of interest. The census used 1 km2 enumeration areas (EAs) to draw a representative sample. Three remote sampling frames were compared with the 2017 Artibonite HH census. First, 2003 census EAs with 2012 population estimates from the Haitian Institute of Statistics and Informatics were standardized to the study EAs. The second sampling frame used the 2016 LandScanTM population estimates and study EAs. The third sampling frame used structures ≥3 m2 manually digitized using Maxar satellite images. In each study EA, 70% of structures were estimated to be inhabited with 4.5 persons/HH. The census identified 33,060 inhabited HHs with an estimated population of 121,593 and 6,126 points of interest. Using daily coverage maps and including digitized structures were novel methods that improved the census quality. Manual digitization was closest to the census sampling frame results with 30,514 digitized structures in the study area. The LandScanTM method performed better in urban areas; however, it produced the highest number of HHs to sample. If a census is not possible, when feasible, remotely digitizing structures and estimating occupancy may provide a close estimate.

PMID:39137752 | DOI:10.4269/ajtmh.23-0891

J Psychiatr Res. 2024 Jul 25;178:125-129. doi: 10.1016/j.jpsychires.2024.07.033. Online ahead of print.

ABSTRACT

Co-occurrence between gambling disorder (GD) and other mental disorders is common, but its association with problematic pornography use (PPU) remains unexplored. This study aimed to investigate relationships between sociodemographic variables, personality measures, psychopathology, emotional regulation, and impulsivity and the co-occurrence of GD and PPU using structural equation modeling (SEM). The sample consisted of 359 adults seeking treatment for GD. The short version of the Problematic Pornography Consumption Scale (PPCS-6) identified patients with GD + PPU. Psychopathology, impulsivity, emotional regulation, and personality were also assessed. Higher impulsivity levels statistically predicted co-occurrence between GD and PPU. Impulsivity mediated the relationship between younger age, maladaptive personality features, and emotional dysregulation and co-occurrence. Psychopathological distress did not directly associate with GD + PPU co-occurrence. Impulsivity relates importantly to the co-occurrence of GD and PPU. Younger age, maladaptive personality, and emotional dysregulation contribute to increased impulsivity levels and co-occurrence. The findings highlight the importance of addressing impulsivity in understanding and treating co-occurring GD and PPU.

PMID:39137720 | DOI:10.1016/j.jpsychires.2024.07.033