Hostname: page-component-cd9895bd7-gbm5v Total loading time: 0 Render date: 2024-12-22T23:09:14.307Z Has data issue: false hasContentIssue false
  • English
  • Français

Psychological networks in clinical populations: investigating the consequences of Berkson's bias

Published online by Cambridge University Press:  04 December 2019

Jill de Ron Open the ORCID record for Jill de Ron [Opens in a new window] ,
Eiko I. Fried  and
Sacha Epskamp
Jill de Ron*
Affiliation:
Department of Psychological Methods, University of Amsterdam, Amsterdam, The Netherlands
Eiko I. Fried
Affiliation:
Department of Clinical Psychology, Leiden University, Leiden, The Netherlands
Sacha Epskamp
Affiliation:
Department of Psychological Methods, University of Amsterdam, Amsterdam, The Netherlands
*
Author for correspondence: Jill de Ron, E-mail: jillderon93@gmail.com
Get access Rights & Permissions [Opens in a new window]

Abstract

Background

In clinical research, populations are often selected on the sum-score of diagnostic criteria such as symptoms. Estimating statistical models where a subset of the data is selected based on a function of the analyzed variables introduces Berkson's bias, which presents a potential threat to the validity of findings in the clinical literature. The aim of the present paper is to investigate the effect of Berkson's bias on the performance of the two most commonly used psychological network models: the Gaussian Graphical Model (GGM) for continuous and ordinal data, and the Ising Model for binary data.

Methods

In two simulation studies, we test how well the two models recover a true network structure when estimation is based on a subset of the data typically seen in clinical studies. The network is based on a dataset of 2807 patients diagnosed with major depression, and nodes in the network are items from the Hamilton Rating Scale for Depression (HRSD). The simulation studies test different scenarios by varying (1) sample size and (2) the cut-off value of the sum-score which governs the selection of participants.

Results

The results of both studies indicate that higher cut-off values are associated with worse recovery of the network structure. As expected from the Berkson's bias literature, selection reduced recovery rates by inducing negative connections between the items.

Conclusion

Our findings provide evidence that Berkson's bias is a considerable and underappreciated problem in the clinical network literature. Furthermore, we discuss potential solutions to circumvent Berkson's bias and their pitfalls.

Keywords

Berkson's biasconditioning on a colliderpsychological networksselection biassimulation study
Type
Original Articles
Information
Psychological Medicine , Volume 51 , Issue 1 , January 2021 , pp. 168 - 176
Check for updates
Copyright
Copyright © Cambridge University Press 2019

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

American Psychiatric Association (2013) Diagnostic and statistical manual of mental disorders: DSM-5. Arlington, VA, American Psychiatric Association.CrossRefGoogle Scholar
Berkson, J (1946) Limitations of the application of fourfold table analysis to hospital data. Biometrics Bulletin 2, 47.CrossRefGoogle ScholarPubMed
Borsboom, D (2017) A network theory of mental disorders. World Psychiatry 16, 5–13.CrossRefGoogle ScholarPubMed
Borsboom, D, Rhemtulla, M, Cramer, AOJ, Van Der Maas, HLJ, Scheffer, M and Dolan, CV (2016) Kinds versus continua: a review of psychometric approaches to uncover the structure of psychiatric constructs. Psychological Medicine 46, 1567–1579.CrossRefGoogle ScholarPubMed
Bringmann, LF, Vissers, N, Wichers, M, Geschwind, N, Kuppens, P, Peeters, F, Borsboom, D and Tuerlinckx, F (2013) A network approach to psychopathology: new insights into clinical longitudinal data. PLoS ONE 8, e60188.CrossRefGoogle ScholarPubMed
Chen, J and Chen, Z (2008) Extended Bayesian information criteria for model selection with large model spaces. Biometrika Trust Biometrika 95, 759–771.CrossRefGoogle Scholar
Cole, SR, Platt, RW, Schisterman, EF, Chu, H, Westreich, D, Richardson, D and Poole, C (2009) Illustrating bias due to conditioning on a collider. International Journal of Epidemiology 39, 417–420.CrossRefGoogle ScholarPubMed
Cramer, AOJ, Waldorp, LJ, van der Maas, HLJ and Borsboom, D (2010) Comorbidity: a network perspective. Behavioral and Brain Sciences 33, 137–150.CrossRefGoogle ScholarPubMed
Cusin, C, Yang, H, Yeung, A and Fava, M (2009) Rating scales for depression. In Baer, L and Blais, MA (eds), Handbook of Clinical Rating Scales and Assessment in Psychiatry and Mental Health. Totowa, NJ: Humana Press, pp. 7–35.CrossRefGoogle Scholar
Elwert, F and Winship, C (2014) Endogenous selection bias: the problem of conditioning on a collider variable. Annual Review of Sociology 40, 31–53.CrossRefGoogle ScholarPubMed
Epskamp, S (2014) IsingSampler: Sampling methods and distribution functions for Ising model [Computer Software Manual]. (R package version 1.0).Google Scholar
Epskamp, S and Fried, EI (2018) A tutorial on regularized partial correlation networks. Psychological Methods 23, 617–634.CrossRefGoogle ScholarPubMed
Epskamp, S, Kruis, J and Marsman, M (2017 a) Estimating psychopathological networks: be careful what you wish for. PLoS ONE 12, e0179891.CrossRefGoogle Scholar
Epskamp, S, Rhemtulla, MT and Borsboom, D (2017 b) Generalized network psychometrics: combining network and latent variable models. Psychometrika 82, 904–927.CrossRefGoogle ScholarPubMed
Epskamp, S, Waldorp, LJ, Mõttus, R and Borsboom, D (2018) The Gaussian graphical model in cross-sectional and time-series data. Multivariate Behavioral Research 53, 453–480.CrossRefGoogle ScholarPubMed
Fava, M, Rush, AJ, Trivedi, MH, Nierenberg, AA, Thase, ME, Sackeim, HA, Quitkin, FM, Wisniewski, S, Lavori, PW, Rosenbaum, JF and Kupfer, DJ (2003) Background and rationale for the sequenced treatment alternatives to relieve depression (STAR*D) study. Psychiatric Clinics of North America 26, 457–494.CrossRefGoogle ScholarPubMed
Foygel, R and Drton, M (2010) Extended Bayesian Information Criteria for Gaussian Graphical Models. In John D. Lafferty, Christopher K. I. Williams, John Shawe-Taylor, Richard S. Zemel and Aron Culotta (eds), ‘NIPS’, Curran Associates, Inc., pp. 604–612.Google Scholar
Fried, EI and Nesse, RM (2015) Depression is not a consistent syndrome: an investigation of unique symptom patterns in the STAR∗D study. Journal of Affective Disorders 172, 96–102.CrossRefGoogle Scholar
Fried, EI, Epskamp, S, Nesse, RM, Tuerlinckx, F and Borsboom, D (2016) What are ‘good’ depression symptoms? Comparing the centrality of DSM and non-DSM symptoms of depression in a network analysis. Journal of Affective Disorders 189, 314–320.CrossRefGoogle Scholar
Fritz, J, Fried, E, Goodyer, I and Wilkinson, P (2018) A network model of resilience factors for adolescents with and without exposure to childhood adversity. Scientific Reports 8, 15774.CrossRefGoogle ScholarPubMed
Hamilton, M (1960) A rating scale for depression. Journal of Neurology, Neurosurgery, and Psychiatry 23, 56–62.CrossRefGoogle ScholarPubMed
Haslam, N, Holland, E and Kuppens, P (2012) Categories versus dimensions in personality and psychopathology: a quantitative review of taxometric research. Psychological Medicine 42, 903–920.CrossRefGoogle ScholarPubMed
Haslbeck, J, Borsboom, D and Waldorp, L (2018) Moderated Network Models. arXiv preprint arXiv:1807.02877.Google Scholar
Ising, E (1925) Report on the theory of ferromagnetism. Zeitschrift Für Physik 31, 253–258.CrossRefGoogle Scholar
Koller, D and Friedman, N (2009) Probabilistic Graphical Models: Principles and Techniques. Foundations vol 2009. Cambridge, MA, USA: The MIT press.Google Scholar
Kotov, R, Krueger, RF and Watson, D (2018) A paradigm shift in psychiatric classification: the Hierarchical Taxonomy Of Psychopathology (HiTOP). World Psychiatry 17, 24–25.CrossRefGoogle Scholar
Lauritzen, SL (1996) Graphical models (Vol. 17). Clarendon Press.Google Scholar
Marsman, M, Borsboom, D, Kruis, J, Epskamp, S, van Bork, R, Waldorp, LJ, Maas, HLJVD, Maris, G, Bork, V, Waldorp, LJ, Van Der Maas, HLJ, Maris, G and Marsman, M (2018) An Introduction to network psychometrics: relating Ising network models to item response theory models. Multivariate Behavioral Research 53, 15–35.CrossRefGoogle ScholarPubMed
Meredith, W (1964) Notes on factorial invariance. Psychometrika 29, 177–185.CrossRefGoogle Scholar
Molenaar, D, Dolan, CV, Wicherts, JM and van der Maas, HLJ (2010) Modeling differentiation of cognitive abilities within the higher-order factor model using moderated factor analysis. Intelligence 38, 611–624.CrossRefGoogle Scholar
Muthén, BO (1989) Latent variable modeling in heterogeneous populations. Psychometrika 54, 557–585.CrossRefGoogle Scholar
Nesselroade, JR and Thompson, WW (1995) Selection and related threats to group comparisons: an example comparing factorial structures of higher and lower ability groups of adult twins. Psychological Bulletin 117, 271.CrossRefGoogle ScholarPubMed
Pearl, J (2000) Causality: Models, Reasoning and Inference, vol 29. Cambridge, UK: Cambridge Univ Press.Google Scholar
Persons, JB (1986) The advantages of studying psychological phenomena rather than psychiatric diagnoses. American Psychologist 41, 1252–1260.CrossRefGoogle ScholarPubMed
Rosseel, Y (2012) Lavaan: an R package for structural equation modeling and more. Journal of Statistical Computing 48, 1–36.Google Scholar
R Core Team (2016) R: A language and environment for statistical computing [Computer software manual]. Vienna, Austria. Retrieved from www.R-project.org/.Google Scholar
Rush, AJ, Fava, M, Wisniewski, SR, Lavori, PW, Trivedi, MH, Sackeim, HA, Thase, ME, Nierenberg, AA, Quitkin, FM, Kashner, TM, Kupfer, DJ, Rosenbaum, JF, Alpert, J, Stewart, JW, McGrath, PJ, Biggs, MM, Shores-Wilson, K, Lebowitz, BD, Ritz, L and Niederehe, G (2004) Sequenced treatment alternatives to relieve depression (STAR*D): rationale and design. Controlled Clinical Trials 25, 119–142.CrossRefGoogle ScholarPubMed
Santor, DA, Gregus, M and Welch, A (2006) FOCUS ARTICLE: eight decades of measurement in depression. Measurement: Interdisciplinary Research & Perspective 4, 135–155.Google Scholar
van Borkulo, CD, Borsboom, D, Epskamp, S, Blanken, TF, Boschloo, L, Schoevers, RA and Waldorp, LJ (2014) A new method for constructing networks from binary data. Scientific Reports 4, 5918.CrossRefGoogle ScholarPubMed
Westreich, D (2012) Berksons bias, selection bias, and missing data. Epidemiology 23, 159–164.CrossRefGoogle Scholar
WHO (2016) International Classification of Diseases (ICD) 10. Available at http://apps.who.int/classifications/icd10/browse/2016/en#/XVI.Google Scholar
Supplementary material: File

de Ron et al. supplementary material

de Ron et al. supplementary material 1

Download de Ron et al. supplementary material(File)
File 2.8 KB
Supplementary material: File

de Ron et al. supplementary material

de Ron et al. supplementary material 2

Download de Ron et al. supplementary material(File)
File 97.1 KB
Supplementary material: File

de Ron et al. supplementary material

de Ron et al. supplementary material 3

Download de Ron et al. supplementary material(File)
File 24.6 KB
Supplementary material: File

de Ron et al. supplementary material

de Ron et al. supplementary material 4

Download de Ron et al. supplementary material(File)
File 51.3 KB
Supplementary material: PDF

de Ron et al. supplementary material

de Ron et al. supplementary material 5

Download de Ron et al. supplementary material(PDF)
PDF 209.8 KB
Supplementary material: File

de Ron et al. supplementary material

de Ron et al. supplementary material 6

Download de Ron et al. supplementary material(File)
File 48.9 KB