Validation of a modified Chinese‐language version of the Davos Assessment of Cognitive Biases Scale (MCL‐DACOBS) in a sample of Chinese patients with schizophrenia

Abstract Introduction The Davos Assessment of Cognitive Biases Scale (DACOBS) is widely used to assess cognitive biases in patients who have schizophrenia. However, the lack of a modified Chinese‐language version of the DACOBS (MCL‐DACOBS) precludes Chinese schizophrenic patients from treatment aimed at normalizing cognitive biases, impacting their prognosis. Here, we aimed to produce a DACOBS for China and test the validity and reliability of the resultant MCL‐DACOBS. Methods Eighteen researchers collaborated to develop the MCL‐DACOBS: A total of 15 researchers modified and translated the English version of the DACOBS, 1 native‐English‐speaking researcher back‐translated the scale, and 2 Chinese sinologists localized and optimized the language of the MCL‐DACOBS. Forty‐two volunteers checked the scale items’ comprehensibility, and the two sinologists performed further localization and optimization based on their feedback. The final version of the MCL‐DACOBS used in this study was thus derived from the harmonized English‐language version of the scale. Confirmatory factor analyses (CFAs) were used to examine the best latent structure of the MCL‐DACOBS. Cronbach's α and intraclass correlation coefficients (ICCs) were used to check the reliability. The discriminative ability of the MCL‐DACOBS was assessed according to the area under the receiver operating characteristic curve. Results The CFA showed that all items loaded onto factors with loadings >0.400. A two‐factor structure showed a good model fit (root mean square error of approximation = .018, Tucker–Lewis index = .978, comparative fit index = .984). Promax rotation demonstrated that each item had a high factor load (0.432–0.774). Cronbach's α coefficient and ICC for the MCL‐DOCABS were .965 and .957, respectively, indicating that the scale has ideal reliability. Conclusion The MCL‐DACOBS has good validity and good reliability, and its psychometric properties indicate that it is a valid tool for measuring cognitive biases in Chinese patients with schizophrenia.

Cognitive biases affect several cognitive domains, including decisionmaking/reasoning, attention, motivation, memory recall, and style of attribution of meaning (Beck et al., 2011).Increasing numbers of studies have been reporting cognitive biases that contribute to cognitive impairments in patients with schizophrenia, and cognitive bias and cognitive impairments appear to produce reciprocal deterioration and worsen disability (Frydecka et al., 2022;Hu et al., 2022;Navalón et al., 2021;Tang et al., 2022).Accordingly, assessing cognitive bias in patients with schizophrenia may enable the cognitive performance of patients with schizophrenia to be improved, thus improving their prognosis.
An ideal cognitive bias assessment tool can provide effective information for us to understand schizophrenia-associated cognitive impairments.The Davos Assessment of Cognitive Biases Scale (DACOBS), which was developed by Professor van der Gaag et al.
It consists of 42 items (Table 1), which are divided into seven 6-item subscales.The DACOBS can be used to discriminate patients with psychosis from healthy respondents (Marchi et al., 2022;Müller et al., 2021;Shoham et al., 2022).It has been translated into Polish, Dutch, and Flemish versions, which have been validated in the respective associated populations (Livet et al., 2020;Whitman et al., 2002).Currently, there is not yet a DACOBS suitable for evaluating and measuring specific cognitive biases with a full-range Likert-type scale (ranging from 1 for strongly disagree to 7 for strongly agree) that refers to the last 2 weeks.The DACOBS aims to measure four forms of cognitive bias (jumping to conclusion; belief inflexibility bias; attention to threat bias; and external attribution bias) as well as two types of cognitive limitations (social cognition problems; and subjective cognitive problems), and avoidance behavior (safety behaviors).Our team agreed unanimously to revise the response scale to range from 0 (strongly disagree) to 7 (strongly agree) to identify a suitable cutoff point.Because the DACOBS is based on patient self-reporting, we provide clinical-use instructions indicating administrators should use it selectively in cases where the patient has stable symptoms, has demonstrated clear insight into their symptoms, and does not have any current long-term or shortterm memory impairments, especially related to memory (Gawęda et al., 2018;Pena-Garijo et al., 2022).
The aim of this work was to produce a DACOBS that can be used

Localization and optimization
A of the scale (Livet et al., 2022).

Constructive validity evaluation
The validity of the MCL-DACOBS was assessed using exploratory factor analysis (EFA) and confirmatory factor analysis (CFA).Factor loading of items into structural factors was used to assess the struc-ture of the MCL-DACOBS.The Tucker-Lewis index (TLI), comparative fit index (CFI), and root mean square error of approximation (RMSEA) values were calculated to test the fitness of a structural model of the MCL-DACOBS.

Reliability evaluation
Four hundred and twenty patients were assessed independently by 15 raters, who were blinded to the patients' cognitive bias statuses.
Internal consistency was tested using ICC, and split-half reliability was tested using Cronbach's α coefficient.

Definition of the cutoff point
The consistency of 15 professional doctors who have worked in cognition rescue intervention for more than 20 years was taken as the clinical standard.The area under the ROC curve (AUC) method was used (Furuno et al., 2023;Poznanovic et al., 2023;Kamizawa et, al., 2023;DeVellis., 2017;Bourgin et al., 2020) to set cutoff points for cognitive bias severity, and the sensitivity and specificity of putative cutoffs were calculated.

Participants
The

Construct validity
The EFA, performed with scores from a randomly selected subsam- Promax rotation demonstrated that each item had a high factor load (range, 0.432-0.774;Table 2).

Reliability
The scale had good adaptability, as demonstrated by the total ICC value of the interrater consistency (.957).Good split-half reliability was demonstrated by Cronbach's α coefficient of the total scale (.965) (Dunn et al., 2014;McGraw & Wong, 1996;Reise, 2012).

Cutoff points
Using the clinical evaluation standard of the severity of the cognitive bias features as a reference, ROC curve analysis demonstrated that cutoff points of ≥20 and ≥10 identified cases of severe and mild cognitive bias, respectively, and with very good sensitivity and good specificity.The AUC, 95% confidence interval, sensitivity, and specificity values are reported in Table 3. Accordingly, MCL-DACOBS scores of 11-19 were defined as indicating moderate cognitive bias.

DISCUSSION
The MCL-DACOBS was demonstrated to have ideal psychometric properties in Chinese patients with schizophrenia, with good con-structive validity and reliability for identifying cognitive bias severity (Horton et al., 1987).We defined MCL-DACOBS score ranges for three grades of cognitive bias severity.Our data demonstrated that our modified scoring did not affect the constructive validity of the MCL-DACOBS; rather, its reliability was better than the reliabilities that have been determined for other language versions of the DACOBS.
The present findings indicate that it is feasible to use the presently

Limitations
This study has two notable limitations.First, we used a self-report questionnaire.Such self-reporting, by its nature, can be affected by hiding, social desirability, and misunderstanding.Task-based tests are better for obtaining empirical evidence of a specific cognitive bias.Second, the protocol did not include test-retest reliability.Reproducibility over time (test-retest) is one of several ways to classify and measure reliability, as is internal consistency (Banks & Marotta, 2007;Carlson et al., 2021), which measures how well the individual item scores correlate with each other.The internal consistency of the MCL-DACOBS was shown to be satisfactory in the current study.

CONCLUSION
The MCL-DACOBS has good validity, good reliability, and its psychometric properties indicate that it is a valid tool for measuring cognitive biases in Chinese patients with schizophrenia.
to assess cognitive bias in Chinese with psychosis and to validate it in a Chinese population sample.We tested the validity and reliability of the produced modified Chinese-language DACOBS (MCL-DACOBS) in Chinese patients with schizophrenia.The following three hypotheses were examined: (1) The factor structure of the MCL-DACOBS satisfies the criteria of constructive validity; (2) the MCL-DACOBS has good reliability as demonstrated by Cronbach's α coefficient and intraclass correlation coefficient (ICC) values; (3) the receiver operating characteristic (ROC) curve analysis method can be used to define a clinically meaningful cutoff MCL-DACOBS score for cognitive bias severity.
vided written informed consent according to the ethics committees' guidelines before data collection.The patient inclusion criteria were (1) diagnosis of schizophrenia by at least two senior psychiatrists, (2) age ≥18 years, (3) ≥12 years education, (4) ability to understand the MCL-DACOBS, (5) full insight about the disease, and (6) Mandarin Chinese speaking fluency.The same inclusion criteria, with the exception of the schizophrenia diagnosis, were applied in volunteer selection.The exclusion criteria for patients were (1) intellectual disability, (2) neurodegenerative disease, (3) history of a personality disorder, (4) brain trauma, and (5) any other factor that might interfere with cognition.After receiving a full description of the study, all patients and their guardians provided written consent to study participation.2.1.1Criteria of the MCL-DACOBS using illustrations All participants were required to have demonstrated full insight and accurate memory at the time of the MCL-DACOBS assessment.Insight was assessed with the VAGUS insight into psychosis scale (www.
group of 15 researchers collaborated to develop the MCL-DACOBS; a total of 15 researchers modified and translated the English version of the DACOBS, 1 native-English-speaking specialist researcher back-translated the scale, and 2 Chinese sinology researchers localized and optimized the language of the MCL-DACOBS.A group of 42 senior psychiatrists evaluated the scale items' comprehensibility.The aforementioned sinologists and English-language specialists performed further localization and optimization based on the psychiatrists' feedback.The final version of the MCL-DACOBS used in this study was thus derived from the harmonized English-language version ple (n = 420), demonstrated that the MCL-DACOBS had a two-factor structure, measuring social cognitive bias (16 items) and neurocognitive bias (26 items).All factor loadings were >0.4.The Kaiser-Meyer-Olkin (.900) and Bartlett's test of sphericity (χ 2 = 2034.735)values verified the appropriateness of the sample for factor analysis.The cumulative variance contribution extent was 79.86%.The following goodness-of-fit indices for the two-factor structure were obtained: CFI = .948;TLI = .784;normed fit index = .896;RMSEA = .018;goodness of fit index = .891;and adjusted goodness of fit index = .771.
developed MCL-DACOBS to assess the cognitive bias of Chinese patients with schizophrenia.To avoid the impact of biases introduced by patient-reported content on our study, we required that the patients undergoing MCL-DACOBS assessment should have full insight and good memory.Hence, by this method, the MCL-DACOBS was designed to assure assessment accuracy.Regarding the discriminant power of the MCL-DACOBS, our data indicate that schizophrenia patients with mild, moderate, and severe cognitive biases can be identified based on total MCL-DACOBS scores of ≤10, 11-19, and ≥20, respectively, in agreement with the results of van der Gaag et al. (2013).These cutoff points were affirmed by the clinical experience of senior physicians who have worked in cognitive rescue for more than 20 years.The time and effort that we invested into translating the Englishlanguage version DACOBS into the MCL-DACOBS in the present work was fully consistent with the challenges of developing cross-cultural assessment tools, recognizing that cognitive bias is influenced more severely by culture than by other indices.Consequently, results should be explained carefully when using the current version of the MCL-DACOBS.Assessing cognitive bias in patients with schizophrenia may facilitate therapy to improve their cognitive performance, thus improving clinical outcomes.Further studies of the MCL-DACOBS should be conducted to improve its clinical utility in the Chinese medical context.
Factor loading of MCL-Davos Assessment of Cognitive Biases Scale (DACOBS) items.
420 participants (229 males/191 females) were interviewed by trained psychologists (doctoral level) with experience in cognitive measurement tool administration and who worked in the Tianjin Fourth TA B L E 2