1
|
Rousson V, Trächsel B, Iglesias K, Baggio S. Evaluating the cost of simplicity in score building: An example from alcohol research. PLoS One 2023; 18:e0294671. [PMID: 38011173 PMCID: PMC10681198 DOI: 10.1371/journal.pone.0294671] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/21/2023] [Accepted: 10/20/2023] [Indexed: 11/29/2023] Open
Abstract
Building a score from a questionnaire to predict a binary gold standard is a common research question in psychology and health sciences. When building this score, researchers may have to choose between statistical performance and simplicity. A practical question is to what extent it is worth sacrificing the former to improve the latter. We investigated this research question using real data, in which the aim was to predict an alcohol use disorder (AUD) diagnosis from 20 self-reported binary questions in young Swiss men (n = 233, mean age = 26). We compared the statistical performance using the area under the ROC curve (AUC) of (a) a "refined score" obtained by logistic regression and several simplified versions of it ("simple scores"): with (b) 3, (c) 2, and (d) 1 digit(s), and (e) a "sum score" that did not allow negative coefficients. We used four estimation methods: (a) maximum likelihood, (b) backward selection, (c) LASSO, and (d) ridge penalty. We also used bootstrap procedures to correct for optimism. Simple scores, especially sum scores, performed almost identically or even slightly better than the refined score (respective ranges of corrected AUCs for refined and sum scores: 0.828-0.848, 0.835-0.850), with the best performance been achieved by LASSO. Our example data demonstrated that simplifying a score to predict a binary outcome does not necessarily imply a major loss in statistical performance, while it may improve its implementation, interpretation, and acceptability. Our study thus provides further empirical evidence of the potential benefits of using sum scores in psychology and health sciences.
Collapse
Affiliation(s)
- Valentin Rousson
- Division of Biostatistics, Center for Primary Care and Public Health (Unisanté), University of Lausanne, Lausanne, Switzerland
| | - Bastien Trächsel
- Division of Biostatistics, Center for Primary Care and Public Health (Unisanté), University of Lausanne, Lausanne, Switzerland
| | - Katia Iglesias
- School of Health Sciences Fribourg (HEdS-FR), HES-SO University of Applied Sciences and Arts of Western Switzerland, Fribourg, Switzerland
| | - Stéphanie Baggio
- Institute of Primary Health Care (BIHAM), University of Bern, Bern, Switzerland
- Laboratory of Population Health (#PopHealthLab), University of Fribourg, Fribourg, Switzerland
| |
Collapse
|
2
|
Iglesias K, Lannoy S, Sporkert F, Daeppen JB, Gmel G, Baggio S. Performance of self-reported measures of alcohol use and of harmful drinking patterns against ethyl glucuronide hair testing among young Swiss men. PLoS One 2020; 15:e0244336. [PMID: 33362226 PMCID: PMC7757898 DOI: 10.1371/journal.pone.0244336] [Citation(s) in RCA: 4] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/20/2020] [Accepted: 12/07/2020] [Indexed: 12/21/2022] Open
Abstract
Background There is a need for empirical studies assessing the psychometric properties of self-reported alcohol use as measures of excessive chronic drinking (ECD) compared to those of objective measures, such as ethyl glucuronide (EtG). Objectives To test the quality of self-reported measures of alcohol use and of risky single-occasion drinking (RSOD) to detect ECD assessed by EtG. Methods A total of 227 samples of hair from young Swiss men were used for the determination of EtG. Self-reported measures of alcohol use (previous twelve-month and previous-week alcohol use) and RSOD were assessed. Using EtG (<30 pg/mg) as the gold standard of ECD assessment, the sensitivity and specificity were computed, and the AUROC were compared for alcohol use measures and RSOD. Logistic regressions were used to test the contribution of RSOD to the understanding of ECD after controlling for alcohol use. Results A total of 23.3% of participants presented with ECD. Previous twelve-month alcohol use with a cut-off of >15 drinks per week (sensitivity = 75.5%, specificity = 78.7%) and weekly RSOD (sensitivity = 75.5%, specificity = 70.1%) yielded acceptable psychometric properties. No cut-off for previous-week alcohol use gave acceptable results. In the multivariate logistic regression, after controlling for the previous twelve months of alcohol use, RSOD was still significantly associated with EtG (p = .016). Conclusion Self-reported measures of the previous twelve months of alcohol use and RSOD were acceptable measures of ECD for population-based screening. Self-reported RSOD appeared to be an interesting screening measure, in addition to the previous twelve months of alcohol use, to understand ECD among young people.
Collapse
Affiliation(s)
- Katia Iglesias
- School of Health Sciences (HEdS-FR), HES-SO University of Applied Sciences and Arts of Western Switzerland, Fribourg, Fribourg, Switzerland
- * E-mail:
| | - Séverine Lannoy
- Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, California, United States of America
- Laboratory for Experimental Psychopathology, Psychological Sciences Research Institute, Université Catholique de Louvain, Louvain-la-Neuve, Brabant Wallon, Belgium
| | - Frank Sporkert
- Center of Legal Medicine, Forensic Toxicology and Chemistry Unit, Lausanne and Geneva Universities, Lausanne, Vaud, Switzerland
| | - Jean-Bernard Daeppen
- Addiction Medicine, Department of Psychiatry, Lausanne University Hospital, Lausanne, Vaud, Switzerland
| | - Gerhard Gmel
- Addiction Medicine, Department of Psychiatry, Lausanne University Hospital, Lausanne, Vaud, Switzerland
- Addiction Switzerland, Lausanne, Vaud, Switzerland
- Centre for Addiction and Mental Health, Toronto, Ontario, Canada
- University of the West of England, Bristol, United Kingdom
| | - Stéphanie Baggio
- Division of Prison Health, Geneva University Hospitals and University of Geneva, Thônex, Geneva, Switzerland
- Office of Corrections, Department of Justice and Home Affairs of the Canton of Zurich, Zurich, Switzerland
| |
Collapse
|
3
|
Baggio S, Baudat S, Daeppen JB, Gmel G, Heller P, Perroud N, Rothen S, Sporkert F, Studer J, Wolff H, Iglesias K. Screening for alcohol use disorder among individuals with comorbid psychiatric disorders: Diagnostic accuracy in a sample of young Swiss men. Addict Behav 2020; 106:106354. [PMID: 32135396 DOI: 10.1016/j.addbeh.2020.106354] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/17/2019] [Revised: 01/27/2020] [Accepted: 02/11/2020] [Indexed: 12/27/2022]
Abstract
Alcohol use disorder (AUD) is frequently comorbid with other psychiatric disorders. However, few studies investigated the psychometric properties of AUD screening tools in presence of co-occurring disorders. This study examined the diagnostic accuracy of a short AUD screening tool among young adults, in the presence of high vs. low or moderate symptomatology of other common psychiatric disorders. Data were collected among young Swiss men (n = 233) between 2016 and 2018. Measures included a diagnostic interview for AUD and screening tools for AUD and other psychiatric disorders (attention deficit hyperactivity disorder, antisocial personality disorder, bipolar disorder, borderline personality disorder, major depressive disorder, and social anxiety disorder). We computed receiver operating characteristic curves to test whether the AUD screening tool was an accurate indicator of AUD for groups with high vs. low or moderate symptomatology of each psychiatric disorder. The results showed that the optimal cut-off score was ≥3 (the original cut-off of the scale) for participants with a low or moderate symptomatology and ≥4 for participants with a high symptomatology. Our findings highlighted the urgent need for an integrated approach to screening. Psychiatric comorbidities should be included in the screen for AUD to obtain accurate results.
Collapse
|
4
|
Baggio S, Trächsel B, Rousson V, Rothen S, Studer J, Marmet S, Heller P, Sporkert F, Daeppen JB, Gmel G, Iglesias K. Identifying an accurate self-reported screening tool for alcohol use disorder: evidence from a Swiss, male population-based assessment. Addiction 2020; 115:426-436. [PMID: 31656049 DOI: 10.1111/add.14864] [Citation(s) in RCA: 6] [Impact Index Per Article: 1.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 04/11/2019] [Revised: 06/21/2019] [Accepted: 10/14/2019] [Indexed: 12/26/2022]
Abstract
BACKGROUND AND AIMS Short screenings for alcohol use disorder (AUD) are crucial for public health purposes, but current self-reported measures have several pitfalls and may be unreliable. The main aim of our study was to provide empirical evidence on the psychometric performance of self-reports currently used. Our research questions were: compared with a gold standard clinical interview, how accurate are (1) self-reported AUD, (2) self-reported alcohol use over time and (3) biomarkers of alcohol use among Swiss men? Finally, we aimed to identify an alternative screening tool. DESIGN A single-center study with a cross-sectional design and a stratified sample selection. SETTING Lausanne University Hospital (Switzerland) from October 2017 to June 2018. PARTICIPANTS We selected participants from the French-speaking participants of the ongoing Cohort Study on Substance Use and Risk Factors (n = 233). The sample included young men aged on average 27.0 years. MEASUREMENTS We used the Diagnostic Interview for Genetic Studies as the gold standard for DSM-5 AUD. The self-reported measures included 11 criteria for AUD, nine alcohol-related consequences, and previous 12 months' alcohol use. We also assessed biomarkers of chronic excessive drinking (ethyl glucuronide and phosphatidylethanol). FINDINGS None of the self-reported measures/biomarkers taken alone displayed both sensitivity and specificity close to 100% with respect to the gold standard (e.g. self-reported AUD: sensitivity = 92.3%, specificity = 45.8%). The best model combined eight self-reported criteria of AUD and four alcohol-related consequences. Using a cut-off of three, this screening tool yielded acceptable sensitivity (83.3%) and specificity (78.7%). CONCLUSIONS Neither self-reported alcohol use disorder nor heavy alcohol use appear to be adequate to screen for alcohol use disorder among young men from the Swiss population. The best screening alternative for alcohol use disorder among young Swiss men appears to be a combination of eight symptoms of alcohol use disorder and four alcohol-related consequences.
Collapse
Affiliation(s)
- Stéphanie Baggio
- Division of Prison Health, Geneva University Hospitals and University of Geneva, Thônex, Switzerland.,Department of Forensic Psychiatry, Institute of Forensic Medicine, University of Bern, Bern, Switzerland
| | - Bastien Trächsel
- Division of Biostatistics, Center for Primary Care and Public Health (Unisanté), University of Lausanne, Lausanne, Switzerland
| | - Valentin Rousson
- Division of Biostatistics, Center for Primary Care and Public Health (Unisanté), University of Lausanne, Lausanne, Switzerland
| | - Stéphane Rothen
- Addiction Division, Geneva University Hospitals and University of Geneva, Geneva, Switzerland
| | - Joseph Studer
- Addiction Medicine, Department of Psychiatry, Lausanne University Hospital, Lausanne, Switzerland
| | - Simon Marmet
- Addiction Medicine, Department of Psychiatry, Lausanne University Hospital, Lausanne, Switzerland
| | - Patrick Heller
- Division of Prison Health, Geneva University Hospitals and University of Geneva, Thônex, Switzerland.,Adult Psychiatry Division, Department of Mental Health and Psychiatry, Geneva University Hospitals, Thônex, Switzerland
| | - Frank Sporkert
- Centre of Legal Medicine, Forensic Toxicology and Chemistry Unit, Lausanne and Geneva Universities, Lausanne, Switzerland
| | - Jean-Bernard Daeppen
- Addiction Medicine, Department of Psychiatry, Lausanne University Hospital, Lausanne, Switzerland
| | - Gerhard Gmel
- Addiction Medicine, Department of Psychiatry, Lausanne University Hospital, Lausanne, Switzerland.,Addiction Switzerland, Lausanne, Switzerland.,Centre for Addiction and Mental Health, Toronto,, ON, Canada.,University of the West of England, Bristol, UK
| | - Katia Iglesias
- School of Health Sciences, HES-SO University of Applied Sciences and Arts of Western Switzerland, Fribourg, Switzerland
| |
Collapse
|
5
|
Baggio S, Iglesias K. Commentary on : Caution is needed when using self-reported alcohol use disorder screening tools. Addict Behav 2020; 100:106115. [PMID: 31494525 DOI: 10.1016/j.addbeh.2019.106115] [Citation(s) in RCA: 4] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/25/2019] [Revised: 08/30/2019] [Accepted: 08/30/2019] [Indexed: 11/29/2022]
Affiliation(s)
- Stéphanie Baggio
- Division of Prison Health, Geneva University Hospitals and University of Geneva, Chemin du Petit Bel Air 2, 1226 Thônex, Switzerland; Department of Forensic Psychiatry, Institute of Forensic Medicine, University of Bern, Falkenplatz 16, 3012 Bern, Switzerland.
| | - Katia Iglesias
- School of Health Sciences, HES-SO University of Applied Sciences and Arts of Western Switzerland, Route des Arsenaux 16a, 1700 Fribourg, Switzerland.
| |
Collapse
|
6
|
Sporkert F, Iglesias K, Jacot S, Deglon J, Baggio S. Assessment of heavy alcohol drinking among young Swiss men. TOXICOLOGIE ANALYTIQUE ET CLINIQUE 2019. [DOI: 10.1016/j.toxac.2019.03.017] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/26/2022]
|