Background: Major depressive disorder (MDD) or depression is among the most prevalent psychiatric disorders, affecting more than 300 million people globally. Early detection is critical for rapid intervention, which can potentially reduce the escalation of the disorder. Objective: This study used data from social media networks to explore various methods of early detection of MDDs based on machine learning. We performed a thorough analysis of the dataset to characterize the subjects’ behavior based on different aspects of their writings: textual spreading, time gap, and time span. Methods: We proposed 2 different approaches based on machine learning singleton and dual. The former uses 1 random forest (RF) classifier with 2 threshold functions, whereas the latter uses 2 independent RF classifiers, one to detect depressed subjects and another to identify nondepressed individuals. In both cases, features are defined from textual, semantic, and writing similarities. Results: The evaluation follows a time-aware approach that rewards early detections and penalizes late detections. The results show how a dual model performs significantly better than the singleton model and is able to improve current state-of-the-art detection models by more than 10%. Conclusions: Given the results, we consider that this study can help in the development of new solutions to deal with the early detection of depression on social networks.

Early Detection of Depression: Social Network Analysis and Random Forest Techniques

Background

The Global Burden of Disease Study 2013 (GBD 2013) aims to bring together all available epidemiological data using a coherent measurement framework, standardised estimation methods, and transparent data sources to enable comparisons of health loss over time and across causes, age–sex groups, and countries. The GBD can be used to generate summary measures such as disability-adjusted life-years (DALYs) and healthy life expectancy (HALE) that make possible comparative assessments of broad epidemiological patterns across countries and time. These summary measures can also be used to quantify the component of variation in epidemiology that is related to sociodemographic development.

Methods

We used the published GBD 2013 data for age-specific mortality, years of life lost due to premature mortality (YLLs), and years lived with disability (YLDs) to calculate DALYs and HALE for 1990, 1995, 2000, 2005, 2010, and 2013 for 188 countries. We calculated HALE using the Sullivan method; 95% uncertainty intervals (UIs) represent uncertainty in age-specific death rates and YLDs per person for each country, age, sex, and year. We estimated DALYs for 306 causes for each country as the sum of YLLs and YLDs; 95% UIs represent uncertainty in YLL and YLD rates. We quantified patterns of the epidemiological transition with a composite indicator of sociodemographic status, which we constructed from income per person, average years of schooling after age 15 years, and the total fertility rate and mean age of the population. We applied hierarchical regression to DALY rates by cause across countries to decompose variance related to the sociodemographic status variable, country, and time.

Findings

Worldwide, from 1990 to 2013, life expectancy at birth rose by 6·2 years (95% UI 5·6–6·6), from 65·3 years (65·0–65·6) in 1990 to 71·5 years (71·0–71·9) in 2013, HALE at birth rose by 5·4 years (4·9–5·8), from 56·9 years (54·5–59·1) to 62·3 years (59·7–64·8), total DALYs fell by 3·6% (0·3–7·4), and age-standardised DALY rates per 100 000 people fell by 26·7% (24·6–29·1). For communicable, maternal, neonatal, and nutritional disorders, global DALY numbers, crude rates, and age-standardised rates have all declined between 1990 and 2013, whereas for non–communicable diseases, global DALYs have been increasing, DALY rates have remained nearly constant, and age-standardised DALY rates declined during the same period. From 2005 to 2013, the number of DALYs increased for most specific non-communicable diseases, including cardiovascular diseases and neoplasms, in addition to dengue, food-borne trematodes, and leishmaniasis; DALYs decreased for nearly all other causes. By 2013, the five leading causes of DALYs were ischaemic heart disease, lower respiratory infections, cerebrovascular disease, low back and neck pain, and road injuries. Sociodemographic status explained more than 50% of the variance between countries and over time for diarrhoea, lower respiratory infections, and other common infectious diseases; maternal disorders; neonatal disorders; nutritional deficiencies; other communicable, maternal, neonatal, and nutritional diseases; musculoskeletal disorders; and other non-communicable diseases. However, sociodemographic status explained less than 10% of the variance in DALY rates for cardiovascular diseases; chronic respiratory diseases; cirrhosis; diabetes, urogenital, blood, and endocrine diseases; unintentional injuries; and self-harm and interpersonal violence. Predictably, increased sociodemographic status was associated with a shift in burden from YLLs to YLDs, driven by declines in YLLs and increases in YLDs from musculoskeletal disorders, neurological disorders, and mental and substance use disorders. In most country-specific estimates, the increase in life expectancy was greater than that in HALE. Leading causes of DALYs are highly variable across countries.

Interpretation

Global health is improving. Population growth and ageing have driven up numbers of DALYs, but crude rates have remained relatively constant, showing that progress in health does not mean fewer demands on health systems. The notion of an epidemiological transition—in which increasing sociodemographic status brings structured change in disease burden—is useful, but there is tremendous variation in burden of disease that is not associated with sociodemographic status. This further underscores the need for country-specific assessments of DALYs and HALE to appropriately inform health policy decisions and attendant actions.

Global, regional, and national disability-adjusted life years (DALYs) for 306 diseases and injuries and healthy life expectancy (HALE) for 188 countries, 1990–2013: Quantifying the epidemiological transition

Depression is a worldwide public health concern. The World Health Organization (WHO) has recently recommended the implementation of programs for strengthening subjective well-being (SWB) to reduce mental disorders, including depression. Also, in 2013, European member-states agreed on a single measure of SWB, i.e., life satisfaction, for monitoring the progress of SWB in the WHO health policy framework, "Health 2020." Life satisfaction is strongly associated with depression; therefore, its use as health indicator could be suitable to identify individuals at risk for depression. Critical to this use of life satisfaction to target also depression is knowledge on the nature of the association between the two throughout the lifespan and by gender. This study aims at contributing to the knowledge about this association in a sample of 51 individuals screened for major depressive disorder (MDD) and dysthymic disorder (Dys). All individuals were administered the Primary Care Screener for Affective Disorders and the Satisfaction With Life Scale (SWLS). Among individuals negative for MDD or Dys, women displayed similar satisfaction compared with men, whereas among individuals positive for MDD or Dys, women showed greater satisfaction compared with men, whose score denoted life dissatisfaction. Consistently, the regression model for SWLS revealed a significant main effect of positivity for MDD or Dys on life satisfaction as well as a significant interaction between positivity for MDD or Dys and gender. The results of this study do not support the notion that satisfaction with life and depressive symptoms could belong to highly related dimensions, at least among female individuals.

/pdf/the-relationship-between-satisfaction-with-life-and-rvf5l0es6q.pdf

The Relationship Between Satisfaction With Life and Depression Symptoms by Gender.

The early detection or identification of emotional states plays a vital role in today’s world, where the number of internet and social media users are increasing at an unprecedented rate. The psychiatric disorders are very dangerous and it is affecting 300 million people. This is the motivation behind addressing the research problem with novel research articles. Early detection is the key to reduce the number affected individuals due to this disorder potentially. This research study performs an analysis of a standard dataset obtained from online social media, where detection can be based on a machine learning algorithm. This research article proposes a machine-learning algorithm to develop an early prediction from their depression mode, which can be protected from mental illness and suicide state of affairs. The combination of support vector machine and Naïve Bayes algorithm will be used to provide a good accuracy level. The classification model contains many cumulative distribution parameters, which should be classified and identified dynamically. This identification or detection is the features obtained from textual, semantic, and writing content. The evaluation of various Deep Learning (DL) approaches is identifying the early prediction. The sensitivity and accuracy of the method are providing the significant conditions for early detection and late detection. The proposed hybrid Journal of trends in Computer Science and Smart technology (TCSST) (2021) Vol.03/ No. 01 Pages: 24-39 https://www.irojournals.com/tcsst/ DOI: https://doi.org/10.36548/jtcsst.2021.1.003 25 ISSN: 2582-4104 Submitted:26.02.2021 Revised: 05.04.2021 Accepted: 15.04.2021 Published: 1.05.2021 method provides better results for early detection and retained good sensitivity and better accuracy of existing methods. The study from results can help to develop a new idea to develop a early prediction of various emotions of people present in social media.

https://irojournals.com/tcsst/V3/I1/03.pdf

Analysis of Deep Learning Techniques for Early Detection of Depression on Social Media Network - A Comparative Study

More than 300 million people have been affected by depression all over the world. Due to the medical equipment and knowledge limitations, most of them are not diagnosed at the early stages. Recent work attempts to use social media to detect depression since the patterns of opinions and thoughts expression of the posted text and images, can reflect users' mental state to some extent. In this work, we design a system dubbed SenseMood to demonstrate that the users with depression can be efficiently detected and analyzed by using proposed system. A deep visual-textual multimodal learning approach has been proposed to reveal the psychological state of the users on social networks. The posted images and tweets data from users with/without depression on Twitter have been collected and used for depression detection. CNN-based classifier and Bert are applied to extract the deep features from the pictures and text posted by users respectively. Then visual and textual features are combined to reflect the emotional expression of users. Finally our system classifies the users with depression and normal users through a neural network and the analysis report is generated automatically.

SenseMood: Depression Detection on Social Media

A randomised controlled trial of the effectiveness of a program for early detection and treatment of depression in primary care

Objective: There is considerable uncertainty about whether depression screening programs in primary care may improve outcomes and what speciﬁc features of such programs may contribute to success. We tested the effectiveness of a program involving substantial commitment from local mental health services.
Methods: Prospective, randomised, patient- and evaluator-masked, parallel-group, controlled study. Participants were recruited in several urban primary care practices where they completed the PC-SAD screener and WHOQOL-Bref. Those who screened positive and did not report suicidal ideation (N¼115) were randomised to an intervention group (communication of the result and offer of psychiatric eva-luation and treatment free of charge; N¼56) or a control group (no feedback on test result for 3 months; N¼59). After 3 months, 100 patients agreed to a follow-up telephone interview including the admin-istration of the PC-SAD5 and WHOQOL-Bref.
Results: Depression severity and quality of life improved signiﬁcantly in both groups. Intent-to-treat analysis showed no effect of the intervention. As only 37% of patients randomised to the intervention group actually contacted the study outpatient clinic, we performed a per-protocol analysis to determine whether the intervention, if delivered as planned, had been effective. This analysis revealed a signiﬁcant positive effect of the intervention on severity of depressive symptoms, and on response and remission rate. Complier average causal effect analysis yielded similar results.
Conclusion: Due to the relatively small sample size, our ﬁndings should be regarded as preliminary and have limited generalizability. They suggest that there are considerable barriers on the part of many patients to the implementation of depression screening programs in primary care. While such programs can be effective, they should be designed based on the understanding of patients' perspectives.

F. Spandonaro

Papers

A randomised controlled trial of the effectiveness of a program for early detection and treatment of depression in primary care

A randomised controlled trial of the effectiveness of a program for early detection and treatment of depression in primary care