Unsupervised Feature Selection to Identify Important ICD-10 and ATC Codes for Machine Learning on a Cohort of Patients With Coronary Heart Disease: Retrospective Study.

Background: The application of machine learning in health care often necessitates the use of hierarchical codes such as the International Classification of Diseases (ICD) and Anatomical Therapeutic Chemical (ATC) systems. These codes classify diseases and medications, respectively, thereby forming extensive data dimensions. Unsupervised feature selection tackles the "curse of dimensionality" and helps to improve the accuracy and performance of supervised learning models by reducing the number of irrelevant or redundant features and avoiding overfitting. Techniques for unsupervised feature selection, such as filter, wrapper, and embedded methods, are implemented to select the most important features with the most intrinsic information. However, they face challenges due to the sheer volume of ICD and ATC codes and the hierarchical structures of these systems. Objective: The objective of this study was to compare several unsupervised feature selection methods for ICD and ATC code databases of patients with coronary artery disease in different aspects of performance and complexity and select the best set of features representing these patients. Methods: We compared several unsupervised feature selection methods for 2 ICD and 1 ATC code databases of 51,506 patients with coronary artery disease in Alberta, Canada. Specifically, we used the Laplacian score, unsupervised feature selection for multicluster data, autoencoder-inspired unsupervised feature selection, principal feature analysis, and concrete autoencoders with and without ICD or ATC tree weight adjustment to select the 100 best features from over 9000 ICD and 2000 ATC codes. We assessed the selected features based on their ability to reconstruct the initial feature space and predict 90-day mortality following discharge. We also compared the complexity of the selected features by mean code level in the ICD or ATC tree and the interpretability of the features in the mortality prediction task using Shapley analysis. Results: In feature space reconstruction and mortality prediction, the concrete autoencoder-based methods outperformed other techniques. Particularly, a weight-adjusted concrete autoencoder variant demonstrated improved reconstruction accuracy and significant predictive performance enhancement, confirmed by DeLong and McNemar tests (P<.05). Concrete autoencoders preferred more general codes, and they consistently reconstructed all features accurately. Additionally, features selected by weight-adjusted concrete autoencoders yielded higher Shapley values in mortality prediction than most alternatives. Conclusions: This study scrutinized 5 feature selection methods in ICD and ATC code data sets in an unsupervised context. Our findings underscore the superiority of the concrete autoencoder method in selecting salient features that represent the entire data set, offering a potential asset for subsequent machine learning research. We also present a novel weight adjustment approach for the concrete autoencoders specifically tailored for ICD and ATC code data sets to enhance the generalizability and interpretability of the selected features.

Assessing Google Searches for Toothache during COVID-19 Lockdowns.

Background: Lockdowns due to the coronavirus disease 2019 (COVID-19) pandemic forced many dental offices to be closed. This study aims to investigate the association between COVID-19 imposed lockdowns and online searches for toothache using Google Trends (GT). Methods: We investigated GT online searches for the term "toothache" within the past 5 years. The time frame for data gathering was considered as the initiation and end dates of national/regional lockdowns in each country. We used 1-way analysis of variance to identify statistical differences in relative search volumes (RSVs) between 2020 and 2016-2019 for each country. Results: Overall, 16 countries were included in our analyses. Among all countries, Indonesia (n = 100), Jamaica (n = 56), Philippines (n = 56), Iran (n = 52), and Turkey (47) had the highest RSVs for toothache in the specified period. Compared with the previous 4 years, higher RSVs were seen in the world (as a whole) (2020 RSVs, 94.4; vs 2019 RSVs, 77.8 [ P < 0.001]) and 13 countries (81.3% of the included countries). Conclusion: Generally, searching for the term "toothache" showed an increase during the COVID-19 lockdowns in 2020 compared with the past 4 years. This can imply the importance of dental care as urgent medical care during public health emergencies such as COVID-19.

Effect of four-week home-based exercise program on immune response, fat and muscle mass in subjects recovered from COVID-19.

OBJECTIVES: A rehabilitation program for patients with COVID-19 may be required to improve related side effects. This study aimed to investigate the effect of a four-week home workout on body composition and serum levels of IL-6 and cortisol in men recovering from COVID-19. METHODS: The present study is quasi-experimental. To this end, 45 healthy individuals in Tehran were purposefully divided into three groups: those who recovered after COVID-19 (n=30), including two groups of exercise and non-exercise, and individuals who did not have COVID-19 (control) (15=n). The four-week training program included three days of training per week with Traband resistance stretches, weight body training, and cardio exercises. The smirnov-Kolmogorov test was employed for data normality. A one-way analysis of variance test will be used to compare both the mean of variables in groups and the mean of variables before and after the exercise, correlated t-test was employed at a significance level of 0.05. RESULTS: The results showed that serum levels of interleukin 6 and cortisol were significantly reduced in the recovered training group (p=0.001) and the non-trained recovery group (p=0.001) and a significant difference was observed between the groups (p=0.001). Furthermore, only in the recovered training group, did the fat percentage (p=0.001) and muscle mass (p=0.001) respectively decreased and increased. CONCLUSIONS: Four-week home training improves body composition by reducing body fat percentage and increasing muscle mass. In addition, lowering interleukin-6 and cortisol leads to reduced inflammation, faster recovery, and increased immunity.

Effects of Sambucus ebulus Extract on Cell Proliferation and Viability of Triple- Negative Breast Cancer: An In Vitro and In Vivo Study.

BACKGROUND: Triple-negative breast cancer (TNBC) accounts for 15% of all breast cancer (BC) cases and is a severe type of BC. Since medicinal herbs containing biocompatible substances that are accepted by patient more than chemical therapeutics, they can be considered a safe option for treating BC. OBJECTIVE: This study evaluated the effect of Sambucus Ebulus (S. ebulus) extract on a model of TNBC. METHODS: S. ebulus extract was prepared using petroleum ether, ethyl acetate, and methanol. The petroleum ether extract was fractionated and analyzed using vacuum liquid chromatography and GC-MS, respectively. MDAMB- 231 and MCF-10A were used as TNBC and normal breast cells, respectively. Flowcytometry and MTT assays were performed to evaluate cell cycle, apoptosis, and viability of the cells. Gene expression analysis was performed using RT-qPCR. Nude mouse allograft tumor models were used, and pathological sections were evaluated. RESULTS: The findings indicated that S. ebulus extract remarkably decreased cell proliferation and viability. The extract had no toxicity to the normal breast cells but efficiently killed the cancer cells. Cell cycle- and apoptosisrelated gene expression showed that fraction 4 of S. ebulus extract significantly increased the expression of Bax, Bak, P53, and c-MYC. CONCLUSION: This study showed satisfactory results of the effect of S. ebulus extract on clearing BC cells both in vitro and in vivo. Thus, S. ebulus extract may be a safe herbal compound for eliminating BC cells without toxicity to host cells.

Clinicians and Older Adults' Perceptions of the Utility of Patient-Generated Health Data in Caring for Older Adults: Exploratory Mixed Methods Study.

BACKGROUND: Many people are motivated to self-track their health and optimize their well-being through mobile health apps and wearable devices. The diversity and complexity of these systems have evolved over time, resulting in a large amount of data referred to as patient-generated health data (PGHD), which has recently emerged as a useful set of data elements in health care systems around the world. Despite the increased interest in PGHD, clinicians and older adults' perceptions of PGHD are poorly understood. In particular, although some clinician barriers to using PGHD have been identified, such as concerns about data quality, ease of use, reliability, privacy, and regulatory issues, little is known from the perspectives of older adults. OBJECTIVE: This study aims to explore the similarities and differences in the perceptions of older adults and clinicians with regard to how various types of PGHD can be used to care for older adults. METHODS: A mixed methods study was conducted to explore clinicians and older adults' perceptions of PGHD. Focus groups were conducted with older adults and health care providers from the Greater Toronto area and the Kitchener-Waterloo region. The participants were asked to discuss their perceptions of PGHD, including facilitators and barriers. A questionnaire aimed at exploring the perceived usefulness of a range of different PGHD was also embedded in the study design. Focus group interviews were transcribed for thematic analysis, whereas the questionnaire results were analyzed using descriptive statistics. RESULTS: Of the 9 participants, 4 (44%) were clinicians (average age 38.3 years, SD 7 years), and 5 (56%) were older adults (average age 81.0 years, SD 9.1 years). Four main themes were identified from the focus group interviews: influence of PGHD on patient-provider trust, reliability of PGHD, meaningful use of PGHD and PGHD-based decision support systems, and perceived clinical benefits and intrusiveness of PGHD. The questionnaire results were significantly correlated with the frequency of PGHD mentioned in the focus group interviews (r=0.42; P=.03) and demonstrated that older adults and clinicians perceived blood glucose, step count, physical activity, sleep, blood pressure, and stress level as the most useful data for managing health and delivering high-quality care. CONCLUSIONS: This embedded mixed methods study generated several important findings about older adults and clinicians' perceptions and perceived usefulness of a range of PGHD. Owing to the exploratory nature of this study, further research is needed to understand the concerns about data privacy, potential negative impact on the trust between older adults and clinicians, data quality and quantity, and usability of PGHD-related technologies for older adults.

Association of COVID-19-imposed lockdown and online searches for toothache in Iran.

BACKGROUND: Novel coronavirus disease-2019 (COVID-19) has impacted populations in many ways worldwide, including access to oral health services. This study aims to assess the association between lockdown due to COVID-19 and online searches for toothache in Iran using Google Trends (GT). METHODS: We investigated GT online searches for toothache within the past five years. The time frame for data gathering was considered as the initiation and end dates of the national lockdown in Iran. We performed one-way ANOVA statistical test to compare relative search volumes (RSVs) between the year 2020 and 2016-2019 for the whole country. Then we investigated the possible association of RSVs in provinces with dentists' density, prevalence of current daily smokers, Human Development Index (HDI), Internet access, and fluoride concentration in water using linear regression. RESULTS: When comparing 2020 with the previous four years, there was a rise of 2020 RSVs versus all previous years combined and each year (P < 0.001 for all of them). In the linear model for the year 2020, HDI (B = - 3.29, 95% CI: (- 5.80, - 0.78), P = 0.012) had a strong negative relationship with provincial RSVs. HDI mostly had strong positive relationship with provincial RSVs in prior years. Fluoride concentration (B = - 0.13, 95% CI: (- 0.24, - 0.03), P = 0.017) and dentists' density (B = - 0.04, 95% CI: (- 0.25, 0.17), P = 0.669) were also negatively associated with RSVs in 2020. These associations were mostly negative in the previous years as well. Internet access (B = 0.36, 95% CI: (- 0.38, 1.09), P = 0.325) and prevalence of daily smokers (B = 0.33, 95% CI: (0.13, 0.53), P = 0.002) were positively associated with RSVs. CONCLUSION: The RSVs for toothache in 2020 have increased due to COVID-19-imposed lockdown compared with the same period in the past four years. This increase was related to socioeconomic factors.

Effect of 8 Weeks of Hospital-Based Resistance Training Program on TCD4+ Cell Count and Anthropometric Characteristic of Patients With HIV in Tehran, Iran: A Randomized Controlled Trial.

Ghayomzadeh, M, SeyedAlinaghi, S, Shamsi, MM, Rezaei, S, Earnest, CP, Akbarnejad, S, Taj, L, Mohraz, M, Navalta, JW, Ghasemi, P, and Voltarelli, FA. Effect of 8 weeks of hospital-based resistance raining program on TCD4+ cell count and anthropometric characteristic of patients with HIV in Tehran, Iran: A randomized controlled trial. J Strength Cond Res 33(4): 1146-1155, 2019-We examined the effects of an 8-W circuit resistance training (RES) program using elastic bands and body weight on TCD4+ cell counts and anthropometry in patients with HIV. Patients (N = 21) receiving antiretroviral therapy were randomly assigned to resistance training (RES; n = 14) or control (CON; n = 7) groups. RES (3/W) consisted of training with elastic bands and bodyweight training focusing on major muscle groups. CON received standard care. Statistical analyses were performed using general linear models adjusted for age, sex, length of infection, and respective baseline measures. The primary outcome was TCD4+, and secondary outcomes were anthropometry indices. Tertiary assessments explored Pearson correlations surrounding the relationship between changes in anthropometry and TCD4+. We observed significant increases in TCD4+ count accompanying RES training (105.50 cells·mm, 95% confidence interval [CI], 47.42-163.59), whereas CON significantly decreased (-41.01 cells·mm, 95% CI, -126.78 to 44.76). Significant between-group differences were noted (p < 0.02; n = 0.42). We also observed significant reductions in fat mass for RES (1.18 kg, 95% CI = 1.80 to -0.56) vs. increased fat mass for CON (1.21 kg, 95% CI, 0.31 to 2.11). Significant between-group differences were noted (p = 0.001, n = 0.64). Similar effects were noted for lean body mass. No significant changes were observed for body mass. Significant correlations were observed for fat mass (r = -0.699, p = 0.001) and lean mass (r = 0.553, p = 0.017), but not body mass (r = -0.390, p = 0.109) vs. changes in TCD4+. Our results suggest that the RES program used in this study is effective for improving TCD4+ status and body composition in patients with HIV.