Hostname: page-component-848d4c4894-4rdrl Total loading time: 0 Render date: 2024-07-03T01:43:57.658Z Has data issue: false hasContentIssue false

Early Adversity Scale for Schizophrenia (Short Form) (EAS-Sz (SF))

Published online by Cambridge University Press:  23 November 2023

Patsy Di Prinzio*
Affiliation:
School of Population and Global Health, University of Western Australia, Perth, WA, Australia
Vera Anne Morgan
Affiliation:
School of Population and Global Health, University of Western Australia, Perth, WA, Australia
*
Corresponding author: Patsy Di Prinzio; Email: patsy.diprinzio@uwa.edu.au
Rights & Permissions [Opens in a new window]

Abstract

Type
Correspondence
Copyright
Copyright © The Author(s), 2023. Published by Cambridge University Press

To the Editor

We have developed the Early Adversity Scale for Schizophrenia (EAS-Sz) (Di Prinzio et al., Reference Di Prinzio, Bjork, Valuri, Ambrosi, Croft and Morgan2023) which is a weighted combination of 42 childhood adversity exposures up to age 10 that can be measured from cross-linked electronic health and social services registers. Five subscales corresponding to (1) discontinuity in parenting (Morgan et al., Reference Morgan, Kirkbride, Leff, Craig, Hutchinson, McKenzie and Fearon2007), (2) family functioning (Read, van Os, Morrison, & Ross, Reference Read, van Os, Morrison and Ross2005), (3) family structure (Haukka, Suvisaari, & Lonnqvist, Reference Haukka, Suvisaari and Lonnqvist2004), (4) area-level socioeconomic/demographic environment (Haukka et al., Reference Haukka, Suvisaari and Lonnqvist2004; Veling, Susser, Selten, & Hoek, Reference Veling, Susser, Selten and Hoek2015), and (5) family-level sociodemographic status (Wicks, Hjern, Gunnell, Lewis, & Dalman, Reference Wicks, Hjern, Gunnell, Lewis and Dalman2005) were also developed. See Table 1 for the variables included in each domain. Using routinely collected register data, we believe EAS-Sz has potential to help untangle contributions of genetic and environmental liability to schizophrenia risk. However, to increase the accessibility of EAS-Sz, we have now derived a shortened version of it and its subscales using simplified calculations, while maintaining useful levels of discrimination.

Table 1. Calculation of EAS-Sz (SF) and simplified domain-level subscales

a In area of mother’s residence at child’s birth, at 2001 census.

The calculation of the original EAS-Sz is a weighted combination of its five subscale scores. Each of these is in turn a weighted combination of its domain-specific adversity exposures. To simplify EAS-Sz, we used the original development dataset from which it was developed and identified cut-points in each distribution of subscale scores, some of which arose naturally because of the tendency for adversity exposures to cluster. We then explained and distinguished these groups by efficient checklists of exposures. For example, the distinct group of children with the highest scores in discontinuity in parenting had all experienced at least three adverse exposures, regardless of which three. The resulting five categorical scales (each a simplification of its original full domain continuous scale) were then combined through multivariable cox regression of time to schizophrenia diagnosis. Weights derived from regression coefficients were then assigned to each categorical level of each domain. A careful study of the regression coefficients showed that some category weightings within a domain could be assumed equal. This allowed for further simplification such that, when all domain scales were used to contribute to the overall scale, they could be explained by either two or three categories each. Refined regression-derived weightings for 13 categories were rounded and the summed rounded weightings within domains mapped to a domain-contribution to total score. See Table 1 for the steps involved in calculating this simplified version of EAS-Sz: Early Adversity Scale for Schizophrenia (Simplified Form) – EAS-Sz (SF).

This modified scale has a potential range from 0 to 10. Harrell's Concordance (Harrell, Califf, Pryor, Lee, & Rosati, Reference Harrell, Califf, Pryor, Lee and Rosati1982) values for predicting time to schizophrenia outcome were determined from the two original validation samples (C = 0.63 (95% CI 0.60–0.66) and C = 0.65 (95% CI 0.62–0.68)) which were consistent with those calculated for the original scale. Four category approximations for individual domains displayed concordance values which varied between validation samples and ranged between a minimum of C = 0.55 (95% CI 0.52–0.58) for Domain 4 to a maximum of C = 0.60 (95% CI 0.58–0.63) for Domain 3. Category level concordances were equivalent to those for their respective full scales for all domains. Calibration graphs (Royston, Reference Royston2015) for EAS-Sz (SF) displayed good agreement between predicted time to schizophrenia diagnosis curves and 95% CI bands of observed Kaplan–Meier (Kaplan & Meier, Reference Kaplan and Meier1958) curves.

For those 5% of children assigned an EAS-Sz (SF) score of five or greater, we estimated an average hazard rate of schizophrenia diagnosis at least four times higher than the rate estimated for the 63% of children assigned a score of zero or one.

EAS-Sz has been developed to help untangle contributions of genetic and environmental liability to schizophrenia risk. Initial validations show discrimination capacity well above chance. To increase the accessibility of EAS-Sz, we developed a simplified version, EAS-Sz (SF), derived through simplified calculations applied to a checklist. Concordance values for EAS-Sz (SF) from validation samples were equivalent to those for EAS-Sz. Results showed the concordance values for five domain sub-scales, each simplified to a four-category measure, were equivalent to the concordance values for the respective full sub-scales.

We have presented EAS-Sz (SF) and five domain-based four-category subscales. These are simplified versions of our original EAS-Sz and corresponding five domain scales and they display equivalent discriminative capacity. These new tools provide a computationally simplified method for researchers to simultaneously account for many early childhood adversity exposures when modeling the role of cumulative adversity in schizophrenia risk, including its interplay with genetic liability. While the greatest discrimination capacity is achieved when contributions across five domains are included, individual domain scales also provide some discrimination capacity above chance.

References

Di Prinzio, P., Bjork, J., Valuri, G., Ambrosi, T., Croft, M., & Morgan, V. A. (2023). Development and initial validation of a multivariable predictive early adversity scale for schizophrenia (EAS-Sz) using register data to quantify environmental risk for adult schizophrenia diagnosis after childhood exposure to adversity. Psychological Medicine, 53(11), 49905000.CrossRefGoogle ScholarPubMed
Harrell, F. E., Califf, R. M., Pryor, D. B., Lee, K. L., & Rosati, R. A. (1982). Evaluating the yield of medical tests. Jama-Journal of the American Medical Association, 247(18), 25432546.10.1001/jama.1982.03320430047030CrossRefGoogle ScholarPubMed
Haukka, J. K., Suvisaari, J., & Lonnqvist, J. (2004). Family structure and risk factors for schizophrenia: Case-sibling study. BMC Psychiatry, 4, 41.CrossRefGoogle ScholarPubMed
Kaplan, E. L., & Meier, P. (1958). Nonparametric-estimation from incomplete observations. Journal of the American Statistical Association, 53(282), 457481.10.1080/01621459.1958.10501452CrossRefGoogle Scholar
Morgan, C., Kirkbride, J., Leff, J., Craig, T., Hutchinson, G., McKenzie, K., … Fearon, P. (2007). Parental separation, loss and psychosis in different ethnic groups: A case-control study. Psychological Medicine, 37(4), 495503.CrossRefGoogle ScholarPubMed
Read, J., van Os, J., Morrison, A. P., & Ross, C. A. (2005). Childhood trauma, psychosis and schizophrenia: A literature review with theoretical and clinical implications. Acta Psychiatrica Scandinavica, 112(5), 330350.CrossRefGoogle ScholarPubMed
Royston, P. (2015). Tools for checking calibration of a Cox model in external validation: Prediction of population-averaged survival curves based on risk groups. Stata Journal, 15(1), 275291.CrossRefGoogle Scholar
Veling, W., Susser, E., Selten, J. P., & Hoek, H. W. (2015). Social disorganization of neighborhoods and incidence of psychotic disorders: A 7-year first-contact incidence study. Psychological Medicine, 45(9), 17891798.CrossRefGoogle ScholarPubMed
Wicks, S., Hjern, A., Gunnell, D., Lewis, G., & Dalman, C. (2005). Social adversity in childhood and the risk of developing psychosis: A national cohort study. American Journal of Psychiatry, 162(9), 16521657.10.1176/appi.ajp.162.9.1652CrossRefGoogle ScholarPubMed
Figure 0

Table 1. Calculation of EAS-Sz (SF) and simplified domain-level subscales