Binary Response Analysis Using Logistic Regression in Dentistry

Professional use of social media platforms by independent dental hygienists in the Netherlands: A quantitative study

AIM

This study examines the prevalence of the professional use of social media platforms by independent DHs in the Netherlands and assesses the associated personal and demographic factors.

METHODS

In this exploratory, observational, cross-sectional study, independent DHs who were members of the Dutch Dental Hygienist Association (Nederlandse Vereniging voor Mondhygiënisten: NVM) were included. Data were collected from the DH practices' websites. Statistics included frequency distributions, percentages, chi-square tests for the relationship between the parameters, and multiple logistic regression for the associations between social media use and the personal and demographic factors.

RESULTS

A total of 830 independent DHs from 670 different practices were included in the study. Of these DHs, 34.4% had a link to a social media platform on their website. DHs with practices in the west or south of the Netherlands were more likely to use Facebook (p = 0.035 and p = 0.002, respectively) than those in the east or north. The likelihood of DHs with 4 years of training using Facebook was 1.910 greater than those with 2 years of training (p = 0.002). Furthermore, DHs who graduated in Utrecht were more likely to use Instagram (p < 0.001).

CONCLUSION

Over a third of the independent DHs in the Netherlands used social media for professional purposes. DHs who trained in Utrecht for 4 years and who had a practice in the west or south of the Netherlands were more likely to use social media for professional purposes.

[Logistic regression]

Logistic regression is a group of statistical techniques that aim to test hypotheses or causal relationships between a categorical dependent variable and other independent variables that can be categorical and quantitative. Through this model we intend to study the probability that the event studied will occur based on some variables that we assume are relevant or influential. In this method it is necessary to detect effect modifier and confounding variables. Its parameters are estimated with the maximum likelihood method through a process with successive iterations.

Give Me a Sign: Using Data Gloves for Static Hand-Shape Recognition

Human-to-human communication via the computer is mainly carried out using a keyboard or microphone. In the field of virtual reality (VR), where the most immersive experience possible is desired, the use of a keyboard contradicts this goal, while the use of a microphone is not always desirable (e.g., silent commands during task-force training) or simply not possible (e.g., if the user has hearing loss). Data gloves help to increase immersion within VR, as they correspond to our natural interaction. At the same time, they offer the possibility of accurately capturing hand shapes, such as those used in non-verbal communication (e.g., thumbs up, okay gesture, …) and in sign language. In this paper, we present a hand-shape recognition system using Manus Prime X data gloves, including data acquisition, data preprocessing, and data classification to enable nonverbal communication within VR. We investigate the impact on accuracy and classification time of using an outlier detection and a feature selection approach in our data preprocessing. To obtain a more generalized approach, we also studied the impact of artificial data augmentation, i.e., we created new artificial data from the recorded and filtered data to augment the training data set. With our approach, 56 different hand shapes could be distinguished with an accuracy of up to 93.28%. With a reduced number of 27 hand shapes, an accuracy of up to 95.55% could be achieved. The voting meta-classifier (VL2) proved to be the most accurate, albeit slowest, classifier. A good alternative is random forest (RF), which was even able to achieve better accuracy values in a few cases and was generally somewhat faster. outlier detection was proven to be an effective approach, especially in improving the classification time. Overall, we have shown that our hand-shape recognition system using data gloves is suitable for communication within VR.

Development of a read-across-derived classification model for the predictions of mutagenicity data and its comparison with traditional QSAR models and expert systems

Mutagenicity is considered an important endpoint from the regulatory, environmental and medical points of view. Due to the wide number of compounds that may be of concern and the enormous expenses (in terms of time, money, and animals) associated with rodent mutagenicity bioassays, this endpoint is a major target for the development of alternative approaches for screening and prediction. The majority of old-aged expert systems and quantitative structure-activity relationship (QSAR) models may show reduced performance over time for their application on newer chemical candidates; thus, researchers constantly try to improve the modeling strategies. In our report, we initially performed traditional classification-based linear discriminant analysis (LDA) QSAR modeling using the benchmark Ames dataset of diverse chemicals (6512 compounds) to recognize the relationship between the molecules and their potential mutagenic behavior. The classical LDA QSAR model is developed from a selected set of 2D descriptors. The LDA QSAR model was developed by using a total of 31 descriptors identified from the analysis of the most discriminating features. Additionally, we have used similarity-derived features obtained from the read-across (RA) to develop an RA-based QSAR model. The developed RA-based LDA QSAR model has better predictivity, transferability, and interpretability compared to the LDA QSAR model, and it uses a very small number of descriptors compared to the classical QSAR model. Different machine learning (ML) models were also developed using the descriptors appearing in the read-across-based LDA QSAR model for comparative studies. We have checked the prediction quality of 216 true external set compounds using the novel similarity-derived RA model. The performance of the OECD toolbox is also compared with the RA-derived LDA QSAR model for a true external set. The current study aimed to explore the significance of the read-across-based algorithm and its application to the most current experimental mutagenicity data to complement already available expert systems.

Application of a developed triple-classification machine learning model for carcinogenic prediction of hazardous organic chemicals to the US, EU, and WHO based on Chinese database

Cancer, the second largest human disease, has become a major public health problem. The prediction of chemicals' carcinogenicity before their synthesis is crucial. In this paper, seven machine learning algorithms (i.e., Random Forest (RF), Logistic Regression (LR), Support Vector Machines (SVM), Complement Naive Bayes (CNB), K-Nearest Neighbor (KNN), XGBoost, and Multilayer Perceptron (MLP)) were used to construct the carcinogenicity triple classification prediction (TCP) model (i.e., 1A, 1B, Category 2). A total of 1444 descriptors of 118 hazardous organic chemicals were calculated by Discovery Studio 2020, Sybyl X-2.0 and PaDEL-Descriptor software. The constructed carcinogenicity TCP model was evaluated through five model evaluation indicators (i.e., Accuracy, Precision, Recall, F1 Score and AUC). The model evaluation results show that Accuracy, Precision, Recall, F1 Score and AUC evaluation indicators meet requirements (greater than 0.6). The accuracy of RF, LR, XGBoost, and MLP models for predicting carcinogenicity of Category 2 is 91.67%, 79.17%, 100%, and 100%, respectively. In addition, the constructed machine learning model in this study has potential for error correction. Taking XGBoost model as an example, the predicted carcinogenicity level of 1,2,3-Trichloropropane (96-18-4) is Category 2, but the actual carcinogenicity level is 1B. But the difference between Category 2 and 1B is only 0.004, indicating that the XGBoost is one optimum model of the seven constructed machine learning models. Besides, results showed that functional groups like chlorine and benzene ring might influence the prediction of carcinogenic classification. Therefore, considering functional group characteristics of chemicals before constructing the carcinogenicity prediction model of organic chemicals is recommended. The predicted carcinogenicity of the organic chemicals using the optimum machine leaning model (i.e., XGBoost) was also evaluated and verified by the toxicokinetics. The RF and XGBoost TCP models constructed in this paper can be used for carcinogenicity detection before synthesizing new organic substances. It also provides technical support for the subsequent management of organic chemicals.

Coverage-level and predictors of maternity continuum of care in Nigeria: implications for maternal, newborn and child health programming

BACKGROUND

Completing maternity continuum of care from pregnancy to postpartum is a core strategy to reduce the burden of maternal and neonatal mortality dominant in sub-Saharan Africa, particularly Nigeria. Thus, we evaluated the level of completion, dropout and predictors of women uptake of optimal antenatal care (ANC) in pregnancy, continuation to use of skilled birth attendants (SBA) at childbirth and postnatal care (PNC) utilization at postpartum in Nigeria.

METHODS

A cross-sectional analysis of nationally representative 21,447 pregnancies that resulted to births within five years preceding the 2018 Nigerian Demographic Health Survey. Maternity continuum of care model pathway based on WHO recommendation was the outcome measure while explanatory variables were classified as; socio-demographic, maternal and birth characteristics, pregnancy care quality, economic and autonomous factors. Descriptive statistics describes the factors, backward stepwise regression initially assessed association (p < 0.10), multivariable binary logistic regression and complementary-log-log model quantifies association at a 95% confidence interval (α = 0.05).

RESULTS

Coverage decrease from 75.1% (turn-up at ANC) to 56.7% (optimal ANC) and to 37.4% (optimal ANC and SBA) while only 6.5% completed the essential continuum of care. Dropout in the model pathway however increase from 17.5% at ANC to 20.2% at SBA and 30.9% at PNC. Continuation and completion of maternity care are positively drive by women; with at least primary education (AOR = 1.27, 95%CI = 1.01-1.62), average wealth index (AOR = 1.83, 95%CI = 1.48 -2.25), southern geopolitical zone (AOR = 1.61, 95%CI = 1.29-2.01), making health decision alone (AOR = 1.39, 95%CI = 1.16-1.66), having nurse as ANC provider (AOR = 3.53, 95%CI = 2.01-6.17) and taking at least two dose of tetanus toxoid vaccine (AOR = 1.25, 95%CI = 1.06-1.62) while women in rural residence (AOR = 0.78, 95%CI = 0.68-0.90) and initiation of ANC as late as third trimester (AOR = 0.44, 95%CI = 0.34-0.58) negatively influenced continuation and completion.

CONCLUSIONS

6.5% coverage in maternity continuum of care completion is very low and far below the WHO recommended level in Nigeria. Women dropout more at postnatal care than at skilled delivery and antenatal. Education, wealth, women health decision power and tetanus toxoid vaccination drives continuation and completion of maternity care. Strategies optimizing these factors in maternity packages will be supreme to strengthen maternal, newborn and child health.

Factors Associated with Self-Report of Type 2 Diabetes Mellitus in Adults Seeking Dental Care in a Developing Country

The aims of the present study were to identify the prevalence and risk indicators of type 2 diabetes mellitus (T2DM) in urban-based Mexican adults seeking care in a university-based triage/intake dental clinic, and to develop a predictive model. A cross-sectional study was conducted on 3354 medical/dental records of adults who sought care at the triage/intake dental clinics of a public university. The dependent variable was self-report of a previous diagnosis of T2DM made by a physician. Several socio-demographic and socioeconomic covariates were included, as well as others related to oral and general health. A multivariate binary logistic regression model was generated. We subsequently calculated well-known statistical measures employed to evaluate discrimination (classification) using an (adjusted) multivariate logistic regression model (goodness-of-fit test). The average age of patients was 42.5 ± 16.1 years old and the majority were female (64.1%). The prevalence of T2DM was 10.7% (95%CI = 9.7−11.8). In the final multivariate model, the variables associated (p < 0.05) with the presence of T2DM were older age (40 to 59 years old, OR = 2.00; 60 to 95 years old, OR = 2.78), having any type of health insurance (OR = 2.33), having high blood pressure (OR = 1.70), being obese (OR = 1.41), and having a functional dentition (OR = 0.68). Although the global fit of the model and the calibration tests were adequate, the sensitivity (0.0%) and positive predictive (0.0%) values were not. The specificity (100%) and negative predictive (89.3%) values, as well as the correctly classified (89.3%) value, were adequate. The area under the ROC curve, close to 0.70, was modest. In conclusion, a prevalence of T2DM of 10.7% in this sample of Mexican adults seeking dental care was similar to national figures. Clinical (blood pressure, BMI and functional dentition), demographic (age), and socioeconomic (health insurance) variables were found to be associated with T2DM. The dental setting could be appropriate for implementing preventive actions focused on identifying and helping to reduce the burden of T2DM in the population.