Using natural language processing to facilitate the harmonization of mental health questionnaires: a validation study using real-world data DOI Open Access
Eoin McElroy, Thomas Andrew Wood, Raymond Bond

et al.

Published: Sept. 1, 2023

Pooling data from different sources may help further mental health research by providing larger sample sizes and allowing cross-study comparisons; however, the heterogeneity in how variables are measured across studies poses a significant challenge to this process. This study explores potential of natural language processing (NLP) harmonize questionnaires matching similar items based on their semantic content. Using Sentence-BERT model, content 39 questions 5 scales was converted into numeric vectors representing These enabled calculation cosine similarity scores, which served as measure between (N=741 item pairs). representative UK adults (N=2,058), Spearman rank correlations were also calculated for same pairs items. We then tested Pearson these two indices, found moderate overall correlation (r = .48, p <.001) scores coefficients. In holdout sample, exhibited ability predict real-world with mean error +/- 0.05, suggesting utility NLP identifying pooling. indicates that score can actual participants would answer However, struggled replicate more complex correlational structures (i.e. latent factors) data. contributes burgeoning field retrospective harmonization highlighting facilitate pooling research. Nevertheless, researchers cautioned verify psychometric equivalence matched items, not fully capture intricate structures.

Language: Английский

Using natural language processing to facilitate the harmonisation of mental health questionnaires: a validation study using real-world data DOI Creative Commons
Eoin McElroy, Thomas Andrew Wood, Raymond Bond

et al.

BMC Psychiatry, Journal Year: 2024, Volume and Issue: 24(1)

Published: July 24, 2024

Pooling data from different sources will advance mental health research by providing larger sample sizes and allowing cross-study comparisons; however, the heterogeneity in how variables are measured across studies poses a challenge to this process.

Language: Английский

Citations

4
Modifiable risk factors of vaccine hesitancy: insights from a mixed methods multiple population study combining machine learning and thematic analysis during the COVID-19 pandemic DOI Creative Commons
Omid V. Ebrahimi, Ella Marie Sandbakken, Sigrun Marie Moss

et al.

BMC Medicine, Journal Year: 2025, Volume and Issue: 23(1)

Published: March 12, 2025

Abstract Background Vaccine hesitancy, the delay in acceptance or reluctance to vaccinate, ranks among top threats global health. Identifying modifiable factors contributing vaccine hesitancy is crucial for developing targeted interventions increase vaccination uptake. Methods This mixed-methods multiple population study utilized gradient boosting machines and thematic analysis identify predictors of during COVID-19 pandemic. Predictors were investigated 2926 Norwegian adults ( M age = 37.91, 79.69% female), before predictive utility these variables was an independent sample 734 UK 40.34, 57.08% female). Two teams authors conducted machine learning analyses, blind each other’s analytic procedures results. Results The model performed well discerning hesitant n 248, 8.48% 109, 14.85%, Norway UK, respectively) from uptaking individuals 2678, 91.52% 625, 85.15%), achieving AUC 0.94 (AUPRC: 0.72; balanced accuracy: 86%; sensitivity 0.81; specificity 0.98) sample, 0.98 0.89; 89%; 0.83; 0.97) out-of-sample replication UK. mixed methods investigation identified five categories risk tied including illusion invulnerability, doubts about efficacy, mistrust official entities, minimization societal impact COVID-19, health-related fears vaccination. portrayal rare incidents across alternative media platforms as fear amplifiers, mainstream media’s stigmatizing presentation unvaccinated individuals, provided additional motives underlying polarization. further revealed information overload, needles, previous negative experiences, not getting healthcare follow-up after if needed, aversion due (psychiatric) illness (e.g., eating disorders) hesitance. Conclusions influential consistent two European samples, highlighting their generalizability populations. These offer insights that could be adapted by public health campaigns mitigating misconceptions related toward increasing Moreover, results highlight responsibility, mediators perception vaccines, minimize polarization provide accurate portrayals vaccine-related incidents, reducing aggravating reactance

Language: Английский

Citations

0
The Structure of Benevolent Childhood Experiences: A Latent Class Analysis and Association with Mental Health Outcomes and Psychological Factors in a Large Adult UK Sample DOI Creative Commons
Andrea Zagaria, Thanos Karatzias, Philip Hyland

et al.

Adversity and Resilience Science, Journal Year: 2025, Volume and Issue: unknown

Published: March 28, 2025

Language: Английский

Citations

0
Perceived manageability of debt and mental health during the COVID-19 pandemic: A UK population analysis DOI Creative Commons
Mark Shevlin, Enya Redican, Philip Hyland

et al.

PLoS ONE, Journal Year: 2022, Volume and Issue: 17(9), P. e0274052 - e0274052

Published: Sept. 21, 2022

Objectives This study examined the association between perceived manageability of debt and risk depression, anxiety, mental health help-seeking among a nationally representative sample adults living in United Kingdom (UK). Methods Data was derived from COVID-19 Psychological Research Consortium (C19PRC) Study Wave 6 (August/September 2021) which psychological, social, economic effects pandemic on UK adult population. Bivariate logistic regression analyses were conducted to determine different levels (i.e., “easily manageable”, “some problems”, “quite serious “very “cannot manage at all”) related outcomes. Results Almost quarter (24%, n = 494) reported management problems, associated with higher help-seeking. After adjusting for demographic variables (e.g. income, receipt benefits), analysis demonstrated dose-response increasing problems Specifically, adjusted odds ratios anxiety ranged 2.28 (‘some problems’) 11.18 (‘very problems’), depression 2.80 16.21 (‘cannot all’), 1.69 3.18 (‘quite problems’, ‘very problems’). Conclusion highlights that represent robust predictor mental-health help seeking.

Language: Английский

Citations

8
An 18‐month follow‐up of the Covid‐19 psychology research consortium study panel: Survey design and fieldwork procedures for Wave 6 DOI Creative Commons
Orla McBride, Sarah Butter, Antón P. Martínez

et al.

International Journal of Methods in Psychiatric Research, Journal Year: 2022, Volume and Issue: 32(2)

Published: Oct. 10, 2022

Abstract Objectives Established in March 2020, the C19PRC Study monitors psychological and socio‐economic impact of pandemic UK other countries. This paper describes protocol for Wave 6 (August–September 2021). Methods The survey assessed: COVID‐19 related experiences; experiences common mental health disorders; characteristics; social political attitudes. Adult participants from any previous wave ( N = 3170) were re‐invited, sample replenishment procedures helped manage attrition. Weights calculated using a raking algorithm to ensure on‐going original panel (from baseline) was nationally representative terms gender, age, household income, amongst factors. Results 1643 adults re‐interviewed at (51.8% retention rate). Non‐participation higher younger adults, those born outside UK, living cities. Of recruited baseline, 54.3% 1100) participated 6. New respondent 415) entered this wave, resulting cross‐sectional 2058 adults. procedure re‐balanced longitudinal within 1.3% population estimates selected socio‐demographic characteristics. Conclusions outlines growing strength publicly available data COVID‐19‐related interdisciplinary research.

Language: Английский

Citations

8
The emotional consequences of novel political identities: Brexit and mental health in the United Kingdom DOI Open Access
Richard P. Bentall, Azzam Alsuhibani, Kate Bennett

et al.

Published: May 25, 2023

Following the 2016 EU Referendum on Britain’s membership of European Union, many people described themselves as ‘Leavers’ or ‘Remainers’. Here, we examine emotional responses associated with Brexit identities using survey data collected from two nationally representative samples British public in 2019 (N = 638) and 2021 2,058). Confirmatory factor analysis indicated that both had coherent Leave Remain identities. and, to a lesser extent, (regardless how actually voted referendum) predicted distress about Brexit-related events clinical symptoms depression anxiety at time points. Structural equation models suggested effect was largely mediated by events. We demonstrate lasting impact mental health UK citizens formation novel political has been more important this process than voting behaviour.

Language: Английский

Citations

0
The emotional consequences of novel political identities: Brexit and mental health in the United Kingdom DOI Creative Commons
Richard P. Bentall, Azzam Alsuhibani, Kate Bennett

et al.

Political Psychology, Journal Year: 2023, Volume and Issue: unknown

Published: Aug. 25, 2023

Abstract Following the 2016 EU referendum on Britain's membership in European Union, many people described themselves as “Leavers” or “Remainers.” Here, we examine emotional responses associated with Brexit identities using survey data collected from two nationally representative samples of British public 2019 ( N = 638) and 2021 2,058). Confirmatory factor analysis indicated that both had coherent Leave Remain identities. and, to a lesser extent, (regardless how actually voted referendum) predicted distress about Brexit‐related events clinical symptoms depression anxiety at time points. Structural equation models suggested effect was largely mediated by events. We demonstrate lasting impact mental health UK citizens show formation novel political has been more important this process than voting behavior.

Language: Английский

Citations

0
Using natural language processing to facilitate the harmonization of mental health questionnaires: a validation study using real-world data DOI Open Access
Eoin McElroy, Thomas Andrew Wood, Raymond Bond

et al.

Published: Sept. 1, 2023

Pooling data from different sources may help further mental health research by providing larger sample sizes and allowing cross-study comparisons; however, the heterogeneity in how variables are measured across studies poses a significant challenge to this process. This study explores potential of natural language processing (NLP) harmonize questionnaires matching similar items based on their semantic content. Using Sentence-BERT model, content 39 questions 5 scales was converted into numeric vectors representing These enabled calculation cosine similarity scores, which served as measure between (N=741 item pairs). representative UK adults (N=2,058), Spearman rank correlations were also calculated for same pairs items. We then tested Pearson these two indices, found moderate overall correlation (r = .48, p <.001) scores coefficients. In holdout sample, exhibited ability predict real-world with mean error +/- 0.05, suggesting utility NLP identifying pooling. indicates that score can actual participants would answer However, struggled replicate more complex correlational structures (i.e. latent factors) data. contributes burgeoning field retrospective harmonization highlighting facilitate pooling research. Nevertheless, researchers cautioned verify psychometric equivalence matched items, not fully capture intricate structures.

Language: Английский

Citations

0