Development and community-based validation of eight item banks to assess mental health

Author’s Accepted Manuscript Development and community-based validation of eight item banks to assess mental health Philip J. Batterham, Matthew Sunderland, Natacha Carragher, Alison L. Calear www.elsevier.com/locate/psychres

PII: DOI: Reference:

S0165-1781(15)30469-8 http://dx.doi.org/10.1016/j.psychres.2016.07.011 PSY9819

To appear in: Psychiatry Research Received date: 6 October 2015 Revised date: 6 May 2016 Accepted date: 4 July 2016 Cite this article as: Philip J. Batterham, Matthew Sunderland, Natacha Carragher and Alison L. Calear, Development and community-based validation of eight item banks to assess mental health, Psychiatry Research, http://dx.doi.org/10.1016/j.psychres.2016.07.011 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Development and community-based validation of eight item banks to assess mental health Philip J. Batterhama*, Matthew Sunderlandb, Natacha Carragherb, Alison L. Caleara a

National Institute for Mental Health Research, Research School of Population Health, The

Australian National University, Canberra, Australia b

NHMRC Centre of Research Excellence in Mental Health and Substance Use, University of

New South Wales, Sydney, Australia *

Corresponding author. National Institute for Mental Health Research, Research School of Population Health, 63 Eggleston Road, The Australian National University, Acton ACT 2601 AUSTRALIA. Tel.: +61 2 61251031; fax: +61 2 61250733. [email protected] Abstract There is a need for precise but brief screening of mental health problems in a range of settings. The development of item banks to assess depression and anxiety has resulted in new adaptive and static screeners that accurately assess severity of symptoms. However, expansion to a wider array of mental health problems is required. The current study developed item banks for eight mental health problems: social anxiety disorder, panic disorder, post-traumatic stress disorder, obsessive-compulsive disorder, adult attention-deficit hyperactivity disorder, drug use, psychosis and suicidality. The item banks were calibrated in a population-based Australian adult sample (N=3,175) by administering large item pools (4575 items) and excluding items on the basis of local dependence or measurement noninvariance. Item Response Theory parameters were estimated for each item bank using a twoparameter graded response model. Each bank consisted of 19-47 items, demonstrating excellent fit and precision across a range of -1 to 3 standard deviations from the mean. No previous study has developed such a broad range of mental health item banks. The calibrated item banks will form the basis of a new system of static and adaptive measures to screen for a broad array of mental health problems in the community. 1

Keywords: screening, assessment, anxiety disorders, obsessive-compulsive disorder, adult attention-deficit hyperactivity disorder, substance use, psychosis, suicidality 1. Introduction Mental disorders are currently under-diagnosed (Kamerow et al., 1986; Goldman et al., 1999; Kessler et al., 2002; Mann et al., 2005), and treatment rates are inadequate (Kamerow et al., 1986; Ormel et al., 1991; Hirschfeld et al., 1997; Goldman et al., 1999). If prevention and early intervention efforts are to be successful, there must be better systems available to rapidly assess individual need for treatment (Gilbody et al., 2006; Insel, 2009). The identification and treatment of mental health problems has traditionally been difficult as a result of poor symptom recognition (Bower et al., 2000; Greer et al., 2004; Zachrisson et al., 2006), and a general lack of help-seeking behaviour or reporting of symptoms (Kamerow et al., 1986; Ormel et al., 1991; Simon and VonKorff, 1995; Goldman et al., 1999; Rickwood et al., 2005). Additional challenges in assessing mental health include the high comorbidity between disorders and the need for clinicians to distinguish between a wide range of mental health problems with different but overlapping symptom presentations (Mineka et al., 1998). New screening tools take advantage of statistical algorithms to adapt assessment to the individual, resulting in measures that are more brief, precise, flexible and efficient than traditional measures (Revicki and Cella, 1997; Cella et al., 2010; Pilkonis et al., 2011). Specifically, Item Response Theory (IRT) and Computerized Adaptive Testing (CAT) have made it possible to maximise the level of precision exhibited by a large pool of items while only administering a fraction of the total pool to respondents. IRT provides a statistical framework that calibrates the observed responses from different items to an individual’s location on a latent construct of interest (Embretson and Reise, 2013), such as severity of anxiety symptoms. Having prior information regarding the performance of each item allows developers to select a reduced set of items that optimally discriminate between respondents at 2

varying severity on the trait continuum. By combining the information from IRT analyses with a CAT approach to assessment, measures can be individually tailored based on item response parameters in order to efficiently triangulate underlying trait scores until a predetermined level of precision has been met (Bjorner et al., 2007). An example of such measures comes from the Patient Reported Outcomes Measurement Information System (PROMIS), a system of highly reliable, valid, flexible, precise, and responsive assessment tools that measure patient-reported health status, including measures for depression, anxiety, alcohol use, physical function, pain, fatigue, social function, and sleep disturbance (Cella et al., 2010; Pilkonis et al., 2013). The PROMIS tools have been shown to greatly reduce the reporting burden on patients. Findings indicate that five-item adaptive screeners and eight-item static screeners for depression discriminate a broader range of depression severity than the commonly used 20-item CES-D (Choi et al., 2010), representing up to 75% greater efficiency in assessment. Although the PROMIS depression and anxiety tools appear to be highly useful as screeners for use in a number of clinical and research settings (Choi et al., 2010; Pilkonis et al., 2011), expansion of available item banks to a broader array of mental health problems and further validation is required. While this study was conducted independently of the PROMIS initiative, we largely adopted the analytical approach used by the PROMIS group. In addition, there is a need to develop screening tools for a range of general and specific mental health problems using a consistent measurement approach. The development of efficient screening tools involves calibration of a bank of items using community-based data, from which appropriate items can be selected for use in screening. Therefore, the aim of this project was to extend the instruments available for mental health screening by developing new item banks to assess eight specific mental health problems. The mental health problems were selected on the basis of high prevalence and/or a paucity of existing brief measures.

3

This study adopted a common item and response framework to develop item banks that could be used alongside PROMIS item banks for depression and anxiety. Initial item pools were developed using a systematic process of item selection reported elsewhere (Batterham et al., 2015). In the present study, these item pools were tested in a large community-based Australian sample to form calibrated item banks assessing social anxiety disorder (SAD), panic disorder (PD), post-traumatic stress disorder (PTSD), obsessive-compulsive disorder (OCD), adult attention-deficit hyperactivity disorder (ADHD), drug use, psychosis and suicidality. The first six item banks were designed to measure specific disorders, while the psychosis item bank was designed to cover a range of psychotic disorders as there are currently few self-report measures covering this area. Given their high prevalence and burden among people with mental disorders, an item bank for suicidal thoughts and behaviours was also developed. It is anticipated that the availability of such a broad range of item banks may be used to develop measures that overcome difficulties in assessing multiple, overlapping mental health problems. This paper reports on the calibration of the new item banks in the community-based sample, which involved removing highly inter-correlated items, removing items that displayed measurement non-invariance on the basis of gender, age group or education, and estimating population-based Item Response Theory parameters for the items ultimately selected into the item banks. DSM-5 criteria for the selected disorders were also evaluated in the present study, which will facilitate subsequent development of disorder-based screeners, along with the development of static and adaptive dimensional measures for the disorders of interest. 2. Method 2.1 Participants and procedure

4

Respondents were recruited from the online social media website Facebook using a series of advertisements targeted to all Australian adults aged 18 year or older during AugustDecember 2014. The target population of Facebook users aged 18 years or older was 8.8 million, representing approximately 45% of the total Australian population in that age bracket. The advertisements resulted in 39,945 users clicking on the advertisement. A total of 10,082 adults consented to participate in the survey with 5,011 (49.7%) completing the survey. As the survey covered a broad range of disorders with lengthy scales for each, consenting participants were given the option of completing a brief form of the survey (lasting approximately 30 min), which comprised one of three versions covering a different combination of mental disorders, or one of three full forms (lasting approximately 60 min) that covered all disorders of interest but were presented in a different order. The survey was conducted online using Limesurvey software, with data stored on a secure server at the Australian National University, Canberra. The study received ethics approval from the Australian National University Human Research Ethics Committee (protocol #2013/509). To facilitate generation of normative data, a weighting scheme was developed to match the Australian general population. The weighting scheme was based on the presence of anxiety, affective, and substance use disorders, accounting for comorbidity between these disorder categories in each age and gender group using representative population-based data on disorder prevalence from the 2007 Australian National Survey of Mental Health and Wellbeing (Slade et al., 2009). The data were also weighted to account for the age and gender distribution of the Australian population using recent census data. Specifically, the weighting scheme assigned participants to one of 96 subgroups (based on presence of depression, anxiety disorders and substance use disorders, across gender and six age groups), with weights in each subgroup ranging from 0.03 to 10.29 reflecting the proportion of each subgroup in the general population. The weights accounted for considerably higher rates of

5

psychopathology and overrepresentation of females and middle-aged individuals in the sample relative to the general population. Application of weights necessitated that the respondents completed the full version of the survey to ensure that complete DSM-5 diagnoses for anxiety, affective, or substance use disorders were available. Therefore, the current study only included participants who opted to complete the full form of the survey, comprising 3,175 participants (63.4% of completers). 2.2 Construction of item pools Construction of the item pools, which formed the candidate items included in this study, has been described previously (Batterham et al., 2015). Briefly, a four stage process was used to select items, which included: (1) systematic literature searches for existing scales of each mental health problem, (2) item selection, standardization, and refinement, (3) item feedback by consumers and expert researchers/clinicians, and (4) reduction of the item pool based on consumer/expert feedback and item content overlap. This process resulted in a total of 23,721 references identified with 310 scales extracted across the eight mental health domains. A total of 6,900 items were standardized using a consistent tense (past-test), timeframe (past 30 days), and response scale (1=Never, 2=Rarely, 3=Sometimes, 4=Often, 5=Always) before undergoing further refinement. Two reviewers eliminated items that were duplicates (or highly similar), unrelated to the disorder, ambiguous, did not fit the 30 day timeframe, or were not specific to the disorder of interest. The remaining 2,002 items were selected to be reviewed by consumers and experts for a secondary round of refinement resulting in a total of 463 items (45-75 items per pool) selected for inclusion in the survey (Batterham et al., 2015) and analysed in the current study. 2.3 Calibration of item banks Calibrating the item banks using IRT requires several assumptions of the item pool to be met. Primarily, the item banks must ideally demonstrate a unidimensional structure. To examine

6

this feature, confirmatory factor analysis (CFA) was conducted using limited information weighted least squares with mean and variance adjustment. This estimator is suitable for the analysis of categorical data using polychoric correlations. Items with a factor loading on a single factor solution <0.4 were excluded and the analysis was repeated to ensure the remaining items were strongly reflective of a single factor.Model fit was determined using the Comparative fit index (CFI), Tucker-Lewis fit index (TLI), and root mean square error of approximation (RMSEA) with good model fit determined using established cut-off values of ≥ 0.90 for CFI and TLI and values ≤ 0.08 for RMSEA (Hu and Bentler, 1998). A second assumption requires that the items exhibit local independence (i.e., the items correlate with one another only through the shared relationship with the latent variable). To identify locally dependent items, modification indices of residual correlations were inspected. Large modification indices associated with significant residual correlations ≥0.3 between item pairs were identified and the content across the item pairs were inspected. One of the items in each pair with a significantly high residual correlation was excluded from the final item bank if it displayed similarity in content and poorer readability in comparison with the partner item. The model was iteratively re-estimated (up to four times) after the exclusion of locally dependent items, to determine whether additional locally dependent items could be identified using these criteria. A final assumption requires that the underlying factor structure and the item response parameters estimated by the model are invariant. Non-invariance or differential item functioning (DIF) implies that the items possess different item properties depending on an external or seemingly unrelated variable (e.g., socio-demographic features). For items with non-invariance, respondents with different socio-demographic backgrounds endorse an item in a different manner despite being matched in terms of latent severity. Evidence of invariance is a necessary prerequisite for group comparisons. DIF was examined in the

7

current study using a multigroup CFA approach, across age [18-55 years vs. 56+ years, with the split chosen to balance subsamples and reflect the observation that symptom patterns are often different in older adults than younger adults (Sunderland et al., 2014)], gender (males vs. females), and education level (post-high school education vs. high school or lower). The multigroup approach proceeded in a series of steps by fitting nested models that estimated increasing levels of measurement invariance separately across age, gender, and education level. First, a baseline or configural model was estimated that assumed both groups of interest possessed the same factor structure but the factor loadings and thresholds were allowed to vary across groups. To identify the model, the factor means for each group were fixed to zero and the factor variances were fixed to one. Second, a comparison or scalar model was estimated that constrained the loadings and thresholds to equality across the groups of interest (note: for categorical items the loadings and thresholds are constrained to equality in tandem as both parameters influence the response curves). Factor means and variances were freely estimated in the comparison group, but were fixed to zero and one respectively in the reference group. The two nested models were then compared using a chi-square difference test implemented by the DIFFTEST function in Mplus. If the chi-square difference test resulted in nonsignificance (p > 0.05) then this would imply that the scalar model does not provide a significant decrement in model fit and all the items can be assumed invariant. If the chisquare difference test demonstrated a significant decrement in model fit (p < 0.05) associated with the scalar model then the modification indices were inspected to determine the specific item(s) that demonstrated significant levels of invariance. Items with the highest modification indices pertaining to the loadings or thresholds were identified and the parameters for those specific items were allowed to vary between the groups whilst the other items were constrained to equality in a partial scalar model. The baseline and partial scalar models were

8

then compared to ensure invariance of the remaining items could be assumed. The items that were identified as non-invariant were subsequently removed from the final item banks. Consistent with large-scale assessment instruments, the PTSD and drug use item pools included a gating question that excluded respondents from completing the full item pool if they had not experienced a traumatic event or if they had never used drugs. This substantially reduced the number of respondents included in the PTSD and drug use analyses. To accurately determine non-invariance across the chosen sub-groups, respondents who completed the brief version of the questionnaire and endorsed the gating questions were also included in these analyses. This resulted in a sample size of 2,978 for PTSD and 968 for drug use. However, the final IRT parameters for the PTSD and drug use item banks for the weighted sample were estimated using only respondents who completed the full version of the questionnaire and endorsed the gating questions. The final item banks were calibrated using the two-parameter graded response model suitable for ordinal data by estimating a unidimensional CFA with a full information robust maximum likelihood estimator and a logit link function. This method takes into account the statistical similarities between IRT and factor analysis for ordered categorical items when converting item factor loadings and thresholds to the respective IRT discrimination and difficulty parameters (Takane and De Leeuw, 1987). The two-parameter model provides a discrimination parameter, which is highly useful for distinguishing between items when generating short-form scales, and is consistent with other item banks developed independently (Pilkonis et al., 2011). The analyses and item calibration were conducted using Mplus version 7.2 (Muthén and Muthén, 2013). 3. Results The characteristics of the calibration sample are detailed in Table 1. Population weights were applied to account for overrepresentation of specific subgroups. Both weighted and

9

unweighted characteristics are displayed in Table 1. The development of the item banks were conducted on the unweighted sample, but the final IRT parameters for the weighted sample are reported below to more accurately reflect likely performance in the general population. The full list of items tested in the initial item pools has been previously published (Batterham et al., 2015). Indices of model fit for the initial models are provided in Table 2. Each item pool was found to display a unidimensional structure with good fit, with the exception of the OCD item pool (RMSEA = 0.07, CFI = 0.88, TLI = 0.88) and the ADHD item pool (RMSEA = 0.10, CFI = 0.91, TLI = 0.91). For the OCD and ADHD item pools, bifactor models were fit to determine the presence of essential unidimensionality using exploratory and confirmatory analysis (Reise et al., 2015). A bifactor model is a special type of multidimensional model that fits a single general factor that accounts for the relationship amongst all items. Additionally, specific factors are also estimated that account for the residual relationship between subgroups of common items. For OCD, a bifactor model was examined that described a single general OCD factor as well as three specific factors that represented items associated with perfectionism, fears of contamination, and unwanted intrusive thoughts. This model provided excellent fit (RMSEA = 0.04, CFI = 0.96, TLI = 0.96). For ADHD, a bifactor model with two specific factors representing attention deficit and hyperactivity/impulsivity, along with a single general ADHD factor, provided excellent fit (RMSEA = 0.07, CFI = 0.96, TLI = 0.95). In all other item pools, each item loaded on a single dimension with a loading of ≥0.6. A summary of the item selection process is presented in Table 3, including the number of items tested, the number of items excluded due to local dependence (large modification index indicating strong inter-item correlations), the number of items excluded due to measurement non-invariance on the basis of age/gender/education/multiple factors, and the final number of items included in the calibrated item banks. The final item banks ranged from 18-47 items.

10

The fit indices for the final item banks reflected excellent model fit, with CFI ≥ 0.96, TLI ≥ 0.95, and RMSEA ≤ 0.07 in all cases. Fit indices for the final item bank models are also reported in Table 2. To examine whether the excellent fit of the models may have been attributable to overfitting, we re-estimated the models in two random halves of the sample. Excellent model fit was maintained in these re-estimations. The IRT parameters, including discrimination (slopes) and difficulty (thresholds), for the resulting item banks are provided in Table 4, based on the weighted data. In unweighted data, discrimination estimates tended to be similar to those for the weighted data but thresholds were lower, reflecting the higher prevalence of psychopathology in the sample relative to the general population. For the OCD and ADHD item banks, previous studies have indicated that an item bank can be assumed as unidimensional (and safely ignore potential levels of multidimensionality) if the factor loadings associated with the general factor of the bifactor model are similar to the factor loadings estimated using a unidimensional model (Reise et al., 2011; Reise et al., 2015). Inspection of the factor loadings for the general OCD and ADHD factors of the bifactor models indicated that all items continued to load highly on their respective general factors, with average differences of 0.05 in comparison to the factor loadings of the unidimensional models for both item banks. The general factor of the OCD bifactor model accounted for 54% of the variance whereas the specific factors explained 6%, 5%, and 3% (14% combined), respectively. Similarly, the general ADHD factor accounted for 55% of variance compared to 12% and 9% (21% combined) for the specific factors. Given the dominance of the general factors, levels of multidimensionality could safely be ignored and the final OCD and ADHD item banks were estimated using unidimensional IRT methods. The high difficulty parameters for most items reflect the skewed response pattern, with few of the participants reporting each of the symptoms. Figure 1 shows information curves for all

11

eight item banks (based again on weighted data) and indicates that the item banks were accurate across a broad range of severity levels, although tending to provide most information between -1 and 3 standard deviations from the population mean. There was some variation in the regions of greatest accuracy, with use, PD, OCD, drug use and psychosis centred around theta values of approximately 2, PTSD and suicidality around 1, and SAD and ADHD around 0. As would be expected, longer item banks tended to provide more information than shorter banks. 4. Discussion The current study developed and calibrated eight new item banks for a range of mental health problems. Item banks had 18-47 items each, with all items displaying local independence and measurement invariance within its bank. All of the item banks were unidimensional or essentially unidimensional, with good fit. The banks accurately measured the dimensions of interest between -1 and 3 SD from the mean. The information curves shown in Figure 1 indicate that the item banks were most accurate among those with elevated symptoms, which suggest the banks may be more appropriate for identifying individuals with high levels of psychopathology than discriminating among healthy individuals. This outcome is similar to the findings from the PROMIS emotional distress item banks (Pilkonis et al., 2011). It has been suggested that bipolar IRT models may not best represent psychopathological data, indicating that testing of unipolar IRT models (Lucke, 2015) may be warranted. However, unipolar IRT models are relatively new and have not been tested widely or implemented in statistical software. Alternatively, construction of a separate item bank for wellbeing may be required to distinguish individuals with few symptoms from those who are healthier. The development of these item banks provides a useful resource for researchers and clinicians aiming to more efficiently assess mental health. The calibrated item banks may now be used to develop short static and adaptive screeners for each of the domains of mental

12

health, as well as developing multidimensional screeners to more rapidly assess whether an individual is experiencing a mental health problem. Items from this study have already been used to develop a brief screen for psychological distress that is more accurate than existing screeners (Batterham et al., in press). Additional screeners for specific disorders may be developed directly from the item banks by identifying subsets of items that best distinguish diagnostic criteria. Similarly, brief dimensional scales may be developed by identifying the most informative items across a broad range of severity levels, as has been done with the PROMIS emotional distress item banks (Choi et al., 2010; Pilkonis et al., 2011). In addition, data on DSM-5 criteria for each of the disorders was collected in the current study, enabling the development of separate screeners based on accuracy in meeting criteria for each disorder (Batterham et al., in press). This study had several strengths, including the use of a large, community-based sample weighted to the general population, the inclusion of a broad range of mental health problems, and the extraction of items demonstrating local independence and measurement invariance. There were also limitations to the current research. First, the process of item selection could have been more data-driven, such that the original item pools extracted from the literature (ranging 159-405 items per pool) were tested. However, given the broad scope of the project, it was not feasible to administer such lengthy item banks. Therefore, a systematic initial process of item selection was conducted (Batterham et al., 2015). Second, online social networks were used to recruit for this study, which may not be representative of the broader community. This problem with community-based research is also observed in traditional recruitment settings such as postal surveys (Batterham, 2014). Population weights were used to address this issue of representativeness for the final IRT estimates, with threshold estimates tending to be reduced by weighting due to the overrepresentation of individuals with psychopathology in the sample. Nevertheless, in developing item banks, it is more

13

important to obtain a sample that is sufficiently large and diverse to obtain stable item parameter estimates, than a sample that is representative of population demographic characteristics (Bjorner et al., 2007). By obtaining a sample with over-representation of participants with elevated symptoms of psychopathology, we are confident that assessment of people across the entire severity range was conducted. Further validation studies in other community-based and clinical settings may strengthen the evidence for the present calibration outcomes, comparing item bank responses to those on legacy scales and clinical interviews. Further validation of the item banks may also be required to rule out overfitting of the models, with an independent validation sample currently being collected. Finally, the study was constrained in the number of specific item banks that could be developed. Future development of item banks for other mental health problems such as personality disorders, eating disorders and sleep-wake disorders may also be beneficial. In conclusion, the item banks that were developed and calibrated in the current study will form the basis of a new system of static and adaptive measures to screen for a broad array of mental health problems in the community. It is anticipated that this approach will provide more efficient screening, leading to higher recognition and more appropriate levels of treatment for mental health problems. Acknowledgements PJB, MS and ALC are supported by NHMRC Fellowships 1083311, 1052327 and 1013199. The study was funded by NHMRC Project Grant 1043952. We thank Jacqueline Brewer, Ph.D., trial manager, for providing assistance with the data collection for the study. The authors declare no conflict of interest.

14

References Batterham, P.J., 2014. Recruitment of mental health survey participants using internet advertising. Int J Meth Psych Res 23, 184-191. Batterham, P.J., Brewer, J.L., Tjhin, A., Sunderland, M., Carragher, N., Calear, A.L., 2015. Systematic item selection process applied to developing item pools for assessing multiple mental health problems. J Clin Epidemiol 68, 913-919. Batterham, P.J., Sunderland, M., Carragher, N., Calear, A.L., Mackinnon, A.J., Slade, T., in press. The Distress Questionnaire-5: Validation of a population screener for psychological distress more accurate than the K6/K10. Journal of Clinical Epidemiology. Bjorner, J.B., Chang, C.H., Thissen, D., Reeve, B.B., 2007. Developing tailored instruments: item banking and computerized adaptive assessment. Qual Life Res 16 Suppl 1, 95108. Bower, P., West, R., Tylee, A., Hann, M., 2000. Symptom attribution and the recognition of psychiatric morbidity. J Psychosom Res 48, 157-160. Cella, D., Riley, W., Stone, A., Rothrock, N., Reeve, B., Yount, S., Amtmann, D., Bode, R., Buysse, D., Choi, S., Cook, K., Devellis, R., DeWalt, D., Fries, J.F., Gershon, R., Hahn, E.A., Lai, J.S., Pilkonis, P., Revicki, D., Rose, M., Weinfurt, K., Hays, R., 2010. The Patient-Reported Outcomes Measurement Information System (PROMIS) developed and tested its first wave of adult self-reported health outcome item banks: 2005-2008. J Clin Epidemiol 63, 1179-1194. Choi, S.W., Reise, S.P., Pilkonis, P.A., Hays, R.D., Cella, D., 2010. Efficiency of static and computer adaptive short forms compared to full-length measures of depressive symptoms. Qual Life Res 19, 125-136.

15

Embretson, S.E., Reise, S.P., 2013. Item response theory for psychologists. Psychology Press, Mahwah, NJ. Gilbody, S., Sheldon, T., Wessely, S., 2006. Should we screen for depression? BMJ 332, 1027-1030. Goldman, L.S., Nielsen, N.H., Champion, H.C., 1999. Awareness, diagnosis, and treatment of depression. J Gen Intern Med 14, 569-580. Greer, J., Halgin, R., Harvey, E., 2004. Global versus specific symptom attributions: predicting the recognition and treatment of psychological distress in primary care. J Psychosom Res 57, 521-527. Hirschfeld, R.M., Keller, M.B., Panico, S., Arons, B.S., Barlow, D., Davidoff, F., Endicott, J., Froom, J., Goldstein, M., Gorman, J.M., Marek, R.G., Maurer, T.A., Meyer, R., Phillips, K., Ross, J., Schwenk, T.L., Sharfstein, S.S., Thase, M.E., Wyatt, R.J., 1997. The National Depressive and Manic-Depressive Association consensus statement on the undertreatment of depression. JAMA 277, 333-340. Hu, L.T., Bentler, P.M., 1998. Fit indices in covariance structure modeling: Sensitivity to underparameterized model misspecification. Psychological methods 3, 424-453. Insel, T.R., 2009. Translating scientific opportunity into public health impact: a strategic plan for research on mental illness. Arch Gen Psychiatry 66, 128-133. Kamerow, D.B., Pincus, H.A., Macdonald, D.I., 1986. Alcohol abuse, other drug abuse, and mental disorders in medical practice. Prevalence, costs, recognition, and treatment. JAMA 255, 2054-2057. Kessler, D., Bennewith, O., Lewis, G., Sharp, D., 2002. Detection of depression and anxiety in primary care: follow up study. BMJ 325, 1016-1017. Lucke, J.F., 2015. Unipolar Item Response Models, in: Reise, S.P., Revicki, D.A. (Eds.), Handbook of item response theory modeling. Taylor and Francis, New York, NY.

16

Mann, J.J., Apter, A., Bertolote, J., Beautrais, A., Currier, D., Haas, A., Hegerl, U., Lonnqvist, J., Malone, K., Marusic, A., Mehlum, L., Patton, G., Phillips, M., Rutz, W., Rihmer, Z., Schmidtke, A., Shaffer, D., Silverman, M., Takahashi, Y., Varnik, A., Wasserman, D., Yip, P., Hendin, H., 2005. Suicide prevention strategies: a systematic review. JAMA 294, 2064-2074. Mineka, S., Watson, D., Clark, L.A., 1998. Comorbidity of anxiety and unipolar mood disorders. Annu Rev Psychol 49, 377-412. Ormel, J., Koeter, M.W., van den Brink, W., van de Willige, G., 1991. Recognition, management, and course of anxiety and depression in general practice. Arch Gen Psychiatry 48, 700-706. Pilkonis, P.A., Choi, S.W., Reise, S.P., Stover, A.M., Riley, W.T., Cella, D., 2011. Item banks for measuring emotional distress from the Patient-Reported Outcomes Measurement Information System (PROMIS(R)): depression, anxiety, and anger. Assessment 18, 263-283. Pilkonis, P.A., Yu, L., Colditz, J., Dodds, N., Johnston, K.L., Maihoefer, C., Stover, A.M., Daley, D.C., McCarty, D., 2013. Item banks for alcohol use from the Patient-Reported Outcomes Measurement Information System (PROMIS): use, consequences, and expectancies. Drug Alcohol Depend 130, 167-177. Reise, S.P., Cook, K.F., Moore, T.M., 2015. Evaluating the impact of multidimensionality on unidimensional item response theory model parameters, in: Reise, S.P., Revicki, D.A. (Eds.), Handbook of item response theory modeling. Taylor and Francis, New York, NY. Reise, S.P., Ventura, J., Keefe, R.S., Baade, L.E., Gold, J.M., Green, M.F., Kern, R.S., Mesholam-Gately, R., Nuechterlein, K.H., Seidman, L.J., Bilder, R., 2011. Bifactor

17

and item response theory analyses of interviewer report scales of cognitive impairment in schizophrenia. Psychol Assess 23, 245-261. Revicki, D.A., Cella, D.F., 1997. Health status assessment for the twenty-first century: item response theory, item banking and computer adaptive testing. Qual Life Res 6, 595600. Rickwood, D., Deane, F.P., Wilson, C.J., Ciarrochi, J., 2005. Young people's help-seeking for mental health problems. AeJAMH (Australian e-Journal for the Advancement of Mental Health) 4. Simon, G.E., VonKorff, M., 1995. Recognition, management, and outcomes of depression in primary care. Arch Fam Med 4, 99-105. Slade, T., Johnston, A., Oakley Browne, M.A., Andrews, G., Whiteford, H., 2009. 2007 National Survey of Mental Health and Wellbeing: methods and key findings. Aust N Z J Psychiatry 43, 594-605. Sunderland, M., Carragher, N., Buchan, H., Batterham, P.J., Slade, T., 2014. Comparing profiles of mental disorder across birth cohorts: results from the 2007 Australian National Survey of Mental Health and Wellbeing. Aust N Z J Psychiatry 48, 452-463. Takane, Y., De Leeuw, J., 1987. On the relationship between item response theory and factor analysis of discretized variables. Psychometrika 52, 393-408. Zachrisson, H.D., Rodje, K., Mykletun, A., 2006. Utilization of health services in relation to mental health problems in adolescents: a population based survey. BMC Public Health 6, 34.

18

Table 1: Sample characteristics (n = 3,175)

Characteristic Age 18-25 26-35 36-45 46-55 56-65 >=66 Gender Male Female Highest level of education Less than high school High school Certificate/Diploma Bachelor Degree Higher degree Prefer not to answer Employment status Full-time Part-time/casual Unemployed Not working (study, maternity, retirement) Prefer not to answer Location of residence Metropolitan Regional Rural/remote Language spoken at home English only English and other Other language only

Unweighted Count Percent 371 277 506 830 842 349

442.3 582.7 558.6 549.6 522.6 519.2

13.9% 18.4% 17.6% 17.3% 16.5% 16.4%

20.4% 1556.9 79.6% 1618.1

49.0% 51.0%

12.0% 13.9% 30.3% 19.8% 23.5% 0.4%

335.7 507.1 835.3 703.6 777.0 16.3

10.6% 16.0% 26.3% 22.2% 24.5% 0.5%

856 819 383 1070 47

27.0% 1001.5 25.8% 712.9 12.1% 358.4 33.7% 1052.8 1.5% 49.4

31.5% 22.5% 11.3% 33.2% 1.6%

1421 1279 475

44.8% 1552.1 40.3% 1204.4 15.0% 418.5

48.9% 37.9% 13.2%

2962 199 14

93.3% 2888.7 6.3% 268.0 0.4% 18.3

91.0% 8.4% 0.6%

648 2527 382 440 963 630 747 13

11.7% 8.7% 15.9% 26.1% 26.5% 11.0%

Weighted Count Percent

19

Table 2: Model fit statistics for initial item pools and final item banks Initial item pool Domain

Final item bank

RMSEA

CFI

TLI

RMSEA

CFI

TLI

Social anxiety disorder

0.07

0.97

0.97

0.07

0.99

0.99

Panic disorder

0.07

0.97

0.97

0.06

0.99

0.99

Post-traumatic stress disorder

0.06

0.97

0.97

0.06

0.98

0.98

Obsessive-compulsive disorder*

0.04

0.96

0.96

0.03

0.98

0.98

Adult ADHD*

0.07

0.96

0.95

0.04

0.98

0.98

Drug use

0.05

0.97

0.97

0.05

0.99

0.99

Psychosis

0.04

0.93

0.93

0.05

0.96

0.95

Suicidality

0.06

0.99

0.99

0.06

0.99

0.99

Notes: * bifactor model fit statistics (otherwise unidimensional models are reported); ADHD: Attention-deficit hyperactivity disorder; CFI: Comparative fit index; TLI: TuckerLewis fit index; RMSEA: root mean square error of approximation

20

Table 3: Construction of calibrated item banks from item pools: item selection process Items removed

Domain

Initial item Local pool dependence size

Age NI only

Gender Education NI only NI only

Multiple Calibrated source item bank NI size

Social anxiety disorder

63

28

2

2

5

0

26

Panic disorder

46

20

3

0

0

4

19

75

22

3

2

0

1

47

61

11

5

0

0

1

44

Adult ADHD

45

13

6

0

1

2

26

Drug use

51

19

5

0

0

0

27

Psychosis

69

43

1

0

3

2

20

Suicidality

53

22

3

2

1

2

23

Post-traumatic stress disorder Obsessivecompulsive disorder

Notes: ADHD: Attention-deficit hyperactivity disorder; NI: measurement non-invariance

21

Table 4: IRT parameters based on population weighted data for the final item banks

Item

Slope (discrimin ation)

Difficulty parameters 1

2

3

4

0.3 9 1.0 2 0.4 3 0.5 0 0.5 0 0.9 4 0.4 8 0.5 5 0.7 2 0.5 5 0.7 5 0.6 1 0.8 7 0.4 4 0.2 6 0.6 8 0.4 7 0.7 0 0.6 4 0.7 0 0.8 0 0.1 3

1. 13 1. 52 1. 13 1. 29 1. 14 1. 50 1. 10 1. 11 1. 37 1. 17 1. 38 1. 22 1. 53 1. 06 1. 00 1. 36 1. 06 1. 38 1. 29 1. 29 1. 43

1.6 6 2.0 2 1.7 5 2.0 1 1.8 2 2.0 8 1.6 8 1.6 6 2.0 4 1.7 5 2.0 3 1.8 6 2.1 1 1.6 5 1.6 1 1.9 9 1.7 8 2.0 1 1.9 0 1.9 4 2.2 0

0. 55

1.2 7

SOCIAL ANXIETY ITEM BANK I felt nervous during social situations

4.45

0.11

My fear of social situations was distressing

4.29

0.52

I felt tense about mixing in a group

4.22

I felt uncomfortable when I was around others

4.11

I was unable to relax during social situations

3.92

I was scared in social situations

3.79

0.46

I felt awkward and tense if I knew people were watching me

3.72

0.05

I feared attending parties or social events

3.59

0.09

I had difficulty concentrating in social situations

3.53

0.22

I was fearful of being the focus of attention in social situations

3.50

0.09

Fear of embarrassment caused me to avoid doing things or speaking to people

3.45

0.26

I feared feeling self-conscious in front of others

3.43

0.09

I found social settings upsetting

3.43

0.31

I was nervous with people unless I knew them well

3.37

I felt self-conscious around others

3.36

I felt self-conscious or embarrassed in crowded places

3.32

0.20

I attempted to hide my reaction to social situations

3.13

0.05

I felt that people could easily see when I was nervous

3.00

0.20

I felt that I would say something embarrassing when talking

2.94

0.07

2.88

0.23

2.83

0.33

2.83

0.55

I feared looking unnatural and artificial during social situations I tried to avoid working on tasks when other people were around Other people were more comfortable in social situations than I was

0.12 0.07 0.07

0.11 0.33

22

Item

Slope (discrimin ation)

Difficulty parameters 1

2

3

4

0.2 7 0.4 6 0.6 9 0.5 2

1. 05 1. 10 1. 34 1. 34

1.7 1 1.7 8 2.1 1 2.2 7

1.5 0 1.4 5 1.3 6 1.7 9 1.5 1 1.7 2 1.5 6 1.2 7 1.5 4 1.3 0

2. 04 1. 98 1. 93 2. 20 2. 11 2. 26 2. 22 1. 86 2. 16 2. 04

2.4 6 2.4 3 2.3 6 2.6 4 2.6 1 2.6 8 2.8 0 2.4 7 2.7 6 2.7 3

1.4 8 1.4 4 1.4 5 2.0 0 2.0 3 1.5 6 2.0 4 1.3 9 2.2 2

2. 12 2. 12 2. 06 2. 58 2. 67 2. 24 2. 67 1. 71 3. 02

2.6 1 2.7 0 2.5 9 3.4 7 3.3 6 2.8 1 3.0 6 2.3 0 3.9 6

I worried about what others thought of me

2.40

0.37

It was important for me not to appear nervous

2.37

0.01

I avoided crowded places

2.35

0.21

I avoided disagreeing with or expressing disapproval to others

1.95

0.12

PANIC DISORDER BANK I felt afraid of certain activities because I feared having a panic attack

4.86

1.05

I worried a lot about having any panic symptoms I worried about having an unexpected anxiety spell or panic attack I feared going out of the house alone in case I had a panic attack I changed my behaviour or did something different because of a panic attack

4.74

1.00

4.63

0.94

4.53

1.38

3.96

1.13

I anticipated that I would have a panic attack

3.76

1.14

I became really frightened for no obvious reason I experienced panic symptoms very easily in stressful situations I worried that I would suddenly get a scared feeling when there was nothing to be afraid of I had a sudden unexpected period of intense fear, anxiety or discomfort I was anxious about going to or being in some places or situations because I feared that help would not be there if I needed it

3.59

1.01

3.59

0.76

3.40

1.04

3.12

0.73

3.11

1.03

It scared me when I felt shaky

2.90

0.90

I feared fear

2.85

0.96

I felt terrified, my heart raced and I thought I might be dying

2.77

1.37

I was frightened by nausea

2.43

1.40

I feared being in a strange place I sought help from my parents, spouse, friends or neighbours because of panic symptoms I had a safety zone (i.e. usually around the home) where I could travel around freely

2.43

0.89

2.20

1.49

2.13

1.11

I feared fainting

1.91

1.47

23

Item

Slope (discrimin ation)

Difficulty parameters 1

2

3

4

0.9 9 0.9 7 1.0 0 1.1 1 1.0 4 0.8 2 1.0 6 1.0 7 1.0 2 0.8 0

1. 35 1. 32 1. 56 1. 57 1. 53 1. 37 1. 48 1. 47 1. 48 1. 32

1.8 8 1.8 7 2.0 6 2.0 4 2.1 0 1.8 7 1.9 1 1.9 4 2.0 6 1.9 2

1.0 2 0.8 5 1.0 6 0.7 7 0.9 9 0.8 1 0.9 5 0.7 4 1.2 7 1.0 4 0.7 3 1.2 4 1.3 5

1. 45 1. 25 1. 65 1. 29 1. 47 1. 23 1. 50 1. 30 1. 80 1. 40 1. 40 1. 68 1. 78

1.8 8 1.7 1 2.2 7 1.9 7 2.1 5 1.6 6 1.9 9 1.9 2 2.4 5 1.9 1 1.9 4 2.0 4 2.2 5

PTSD ITEM BANK I could not feel close to others because of the trauma I felt cut off and isolated from other people because of the trauma

4.82

0.67

4.54

0.68

I felt like I could not relax because of the trauma

4.48

0.61

I did not enjoy the company of others because of the trauma

4.29

0.75

I was easily agitated because of the trauma

4.27

0.69

I felt on edge and distressed when thinking about the trauma

4.25

0.39

I felt hopeless about the future because of the trauma

4.15

0.72

I felt dead inside because of the trauma I found that certain activities or things became pointless, meaningless or insignificant because of the trauma Thoughts I had related to the trauma bothered me emotionally I avoided activities or things that evoked feelings of loneliness, crying or other distressing emotions related to the trauma The trauma caused significant distress or impairment to my social, work or home life I had difficulty concentrating on what I was doing because of the trauma

4.14

0.78

3.93

0.69

3.92

0.30

3.83

0.72

3.81

0.53

3.74

0.66

I felt vulnerable because of the trauma

3.72

0.39

I experienced intrusive, unwanted thoughts about the trauma I thought that no one understood how I felt, not even my family, because of the trauma

3.71

0.58

3.67

0.48

I felt helpless because of the trauma

3.65

0.59

I experienced anxiety because of the trauma

3.62

0.27

I had daydreams of the trauma that felt very real

3.62

0.89

I felt inadequate because of the trauma

3.60

0.65

I had painful images, memories or thoughts of the trauma

3.59

0.22

I experienced intense fear because of the trauma I felt like an object, rather than a person, because of the trauma

3.58

0.86

3.57

1.05

24

Item

Slope (discrimin ation)

Difficulty parameters 1

I felt depressed because of the trauma I felt numb or detached from my surroundings because of the trauma

3.56

0.36

3.53

0.69

I felt lonely because of the trauma

3.51

0.47

I had trouble trusting others because of the trauma I startled easily when someone touched me, spoke to me or approached me unexpectedly because of the trauma Things around me felt unreal or dreamlike because of the trauma

3.45

0.68

3.43

0.83

3.39

0.85

I expected something bad to happen because of the trauma I felt like I was different from other people because of the trauma

3.36

0.88

3.27

0.49

I noticed changes in my appetite because of the trauma

3.14

0.87

I felt as though the trauma was happening again I felt as if my future would somehow be cut short because of the trauma I had physical reactions, such as sweating, shaking and a pounding heart, because of the trauma I felt that I did not laugh or cry at the same things as other people because of the trauma I experienced reminders that brought back feelings about the trauma I was bothered, more than I expected, by feelings of grief because of the trauma

3.12

1.08

3.11

0.96

3.07

0.79

3.04

0.73

3.03

0.16

3.02

0.45

I had bursts of anger because of the trauma

2.97

0.75

I had bad dreams or nightmares about the trauma

2.93

0.71

I was overly alert because of the trauma

2.77

0.56

I felt ashamed of my emotional reactions to the trauma I felt that the world was a dangerous place because of the trauma

2.76

0.94

2.45

0.75

I tried not to talk about what happened during the trauma

2.42

0.55

I tried not to think about what happened during the trauma

2.30

0.46

I was afraid that the trauma would happen again

2.26

0.51

I had difficulty remembering important aspects of the trauma

2.01

0.97

2

3

4

0.7 1 1.0 8 0.8 1 1.0 0 1.2 7 1.2 5 1.2 2 0.8 6 1.2 1 1.5 5 1.2 4 1.1 9 1.0 8 0.6 5 0.8 4 1.3 0 1.1 2 0.9 5 1.2 6 1.1 6 0.9 5 0.8 3 1.0 4 1.4 1

1. 26 1. 60 1. 32 1. 36 1. 65 1. 80 1. 73 1. 40 1. 76 2. 04 1. 75 1. 72 1. 65 1. 42 1. 36 1. 79 1. 74 1. 56 1. 72 1. 73 1. 45 1. 42 1. 73 1. 99

1.9 2 2.1 5 1.8 0 1.9 2 2.0 3 2.3 8 2.2 6 1.8 3 2.3 0 2.6 3 2.1 5 2.3 5 2.1 9 2.2 3 1.9 9 2.3 3 2.3 2 2.1 9 2.1 3 2.3 1 1.9 1 1.9 2 2.3 8 2.5 5

25

Item OCD ITEM BANK I needed to repeat routine activities to prevent terrible consequences I worried a lot if I did not do something exactly the way I liked it I repeatedly checked to see if I had made a mistake or done something terrible I felt driven to repeat routine actions a certain number of times or until it felt just right I felt driven to repeatedly check something even after it had been checked by others I felt the urgent need to know whether some job or task that I had completed had been done correctly I felt compelled to do certain things, even though my reason told me it was not necessary I felt preoccupied with unwanted thoughts about being contaminated I felt unsettled or guilty if I had not been able to do something exactly as I would like No matter how many times I checked something, I could not help wondering whether I had done everything correctly After doing something carefully, I still had the impression I had not finished it I had difficulty concentrating unless things were in the right place

Slope (discrimin ation)

Difficulty parameters 1

2.93

1.54

2.83

0.61

2.69

0.76

2.58

1.41

2.49

0.92

2.48

0.54

2.44

1.04

2.33

2.01

2.31

0.39

2.31

0.68

2.30

0.65

2.29

0.69

I felt overwhelmed by a need for precision or perfection

2.28

0.57

I got stuck doing routine behaviours and it slowed me down

2.24

1.03

I spent far too much time washing my hands I constantly evaluated whether my thoughts and actions were appropriate I felt compelled to follow a very strict routine when doing ordinary things I had thoughts that I might be responsible for something terrible happening I felt that having bad thoughts or urges meant that I was likely to act on them I spent far too long getting ready to leave home each day because I had to do everything exactly right I felt that, even if harm was very unlikely, I had to try to prevent it at any cost I felt very disturbed when I got some contaminated substance on me

2.18

1.67

2.15

0.19

2.15

1.00

2.14

1.11

2.11

1.36

2.11

1.45

2.10

1.18

2.08

1.44

I checked things so that nothing terrible would happen I felt driven to repeatedly ask for reassurance that I had said

2.07 2.04

0.92 0.79

2

3

4

2.1 2 1.2 4 1.4 7 1.9 9 1.5 9 1.1 1 1.6 7 2.6 4 1.0 6 1.3 9 1.3 8 1.3 5 1.1 9 1.6 1 2.3 8 0.8 4 1.6 2 1.7 1 2.0 9 2.0 9 1.8 5 2.0 8 1.4 5 1.4

2. 66 1. 91 2. 19 2. 60 2. 43 1. 81 2. 43 3. 37 1. 88 2. 09 2. 13 2. 13 1. 95 2. 41 2. 98 1. 67 2. 37 2. 58 3. 02 2. 84 2. 52 2. 72 2. 14 2.

3.2 8 2.6 6 2.8 1 3.4 4 3.2 0 2.6 5 2.9 3 3.7 0 2.9 0 2.8 3 3.0 6 3.0 3 2.7 9 3.2 7 3.5 9 2.5 5 3.1 2 3.5 2 3.8 9 3.7 1 2.9 9 3.0 5 2.9 5 3.0

26

Item or done something correctly I avoided doing certain things or going certain places because of contamination concerns I had to keep working at something until it was done exactly right I experienced upsetting and unwanted thoughts about losing control

Slope (discrimin ation)

Difficulty parameters 1

2.03

1.70

2.03

0.40

2.02

0.87

I felt that there was only one right way to do things Whenever I lost control of my thoughts, I had to struggle to regain control I felt that having an upsetting thought made it more likely for that to happen When I started thinking of certain things, I became obsessed with them I felt that having nasty thoughts meant that I was a terrible person I felt that it was ultimately my responsibility to ensure that everything was in order

2.01

0.55

2.01

0.89

1.99

1.08

1.93

0.77

1.88

0.87

1.87

0.25

I feared I would act on an unwanted impulse

1.85

1.17

I did certain things thoroughly in order to get them just right

1.79

0.40

Things had to be perfect according to my own standards I had thoughts that I would harm other people without meaning to hurt them I felt that I had to repeat certain words or phrases in my mind in order to erase bad thoughts, feelings or actions I felt preoccupied with unwanted and intrusive thoughts about harming myself

1.75

0.19

1.68

1.50

1.66

1.43

1.64

1.61

I was concerned I would become ill because of contamination

1.63

1.67

I had difficulty avoiding having upsetting thoughts

1.58

0.37

I had to clean myself to get rid of contamination

1.56

1.75

I found it difficult to touch rubbish or dirty things

1.53

1.27

I did things to prevent or remove contact with contaminants PSYCHOSIS ITEM BANK

1.51

1.39

I heard voices saying bad things about me

3.37

2.25

I heard voices that came from inside my head

2.77

1.86

I felt as if someone or something was playing games with my mind

2.71

1.73

2

3

4

9 2.4 1 1.1 5 1.5 5 1.3 2 1.5 7 1.6 9 1.4 6 1.4 8 0.9 0 1.7 9 1.0 7 0.8 2 2.2 7 2.0 9 2.2 2 2.5 5 1.1 3 2.3 5 2.0 0 2.1 0

27 3. 29 1. 99 2. 40 2. 03 2. 37 2. 54 2. 27 2. 35 1. 62 2. 78 1. 97 1. 64 3. 13 2. 87 3. 06 3. 54 2. 12 3. 49 2. 81 3. 09

8 3.7 2 2.9 0 3.1 9 2.9 7 3.2 7 3.4 9 3.3 7 3.2 4 2.3 9 3.4 6 3.1 7 2.6 4 3.9 5 3.8 0 4.1 8 4.5 3 3.3 4 4.0 3 3.7 8 4.0 9

2.5 4 2.2 8 2.1 4

2. 85 2. 77 3. 00

3.0 8 3.4 1 3.6 5

27

Item

Slope (discrimin ation)

Difficulty parameters 1

I felt that there were odd or unusual things going on around me that I could not explain

2.70

1.54

I felt as if my actions might have been controlled by others

2.56

2.25

I was confused about whether something I experienced was real or imaginary

2.56

1.41

I saw things that other people could not see

2.46

1.98

Other people's thoughts came into my mind

2.28

2.04

I heard voices that sounded like ordinary conversation

2.28

2.08

I doubted that my dreams were the products of my own mind

2.24

1.98

I thought that people would harm me if given an opportunity

2.11

1.61

I heard my own thoughts being repeated

2.10

1.75

I felt that things on television, on radio, in magazines or in newspapers had a special meaning for me

2.03

1.91

I was convinced that people singled me out

2.01

1.34

I thought that I had special powers that other people lacked

1.92

2.14

I felt that there was a conspiracy against me

1.86

1.66

I had days where lights or colours seemed brighter or more intense than usual

1.85

1.53

I believed that I had been punished without cause

1.74

1.51

I experienced smells that other people could not smell

1.39

2.14

I thought there was something seriously wrong with my body

1.30

1.27

2

3

4

2.0 3 2.7 5 2.0 1 2.4 2 2.6 3 2.4 7 2.3 9 2.1 3 2.1 6 2.5 8 1.9 5 2.7 2 2.3 0 2.0 5 2.1 6 2.7 6 1.9 2

2. 86 3. 31 2. 83 3. 11 3. 17 3. 21 3. 33 2. 87 2. 80 3. 54 2. 86 3. 30 3. 31 3. 19 2. 89 3. 80 2. 98

3.6 8 3.8 9 3.6 7 3.5 5 4.2 1 3.9 6 3.8 9 3.4 9 3.6 4 4.0 2 3.4 8 3.8 2 4.0 1 4.3 8 3.6 7 5.1 3 4.0 8

0.3 3 0.1 1 0.1 7 0.4 8 0.3 8 0.1 9

1. 31 1. 09 1. 01 1. 32 1. 28

2.0 4 2.0 9 2.0 2 2.3 3 2.2 2

0. 86

1.9 9

ADHD ITEM BANK I had difficulty sustaining attention on tasks

3.57

I had trouble concentrating

3.30

I was easily distracted

3.27

I had difficulty organising my thoughts

3.24

I had difficulty completing tasks

2.85

0.42 0.69 0.61 0.24 0.41

My mind wandered

2.70

1.05

28

Item

Slope (discrimin ation)

Unrelated thoughts seemed to pop into my head I had difficulty keeping my attention when doing boring or repetitive tasks While listening to others, my attention drifted to unrelated thoughts When I was spoken to directly, it took me longer than other people to understand what was being said I was reluctant to do tasks that required sustained mental effort

2.61

I jumped from task to task without finishing the first

2.30

I failed to give close attention to details

2.26

I lost my train of thought when conversing with others

2.21

I forgot to do things

2.20

I had racing thoughts

2.04

I did not listen when spoken to directly

2.01

2.52 2.47 2.39 2.31

Difficulty parameters 1

2

3

4

0.31 0.55 0.78

0.3 3 0.1 6 0.1 2 1.0 3 0.4 5 0.3 8 0.5 6 0.3 4 0.0 4 0.6 4 1.0 0 0.1 2 0.7 6 0.6 7 1.2 5 0.3 0 1.3 2 2.0 3 0.8 3 1.1 8

1. 22 1. 08 1. 16 1. 94 1. 33 1. 31 1. 67 1. 54 1. 30 1. 58 2. 25

2.2 0 2.0 6 2.3 6 2.9 4 2.3 0 2.5 1 2.9 3 2.8 3 2.5 3 2.6 0 3.3 9

0. 90 1. 94 1. 77 2. 43 1. 42 2. 49 3. 28 2. 52 2. 57

2.0 4 3.0 4 2.9 2 3.6 9 2.6 8 3.5 9 4.4 5 4.4 6 3.9 1

1.5 4 1.3 0 1.0 4

1. 86 1. 82 1. 61

2.3 8 2.3 0 2.0 8

0.24 0.26 0.45 0.36 0.57 1.03 0.03 0.02 0.97 0.37 0.27

I put off difficult tasks

2.01

I made careless mistakes

1.90

I misjudged time

1.79

I blurted out things

1.70

I was bored easily

1.65

0.37 0.48

I had difficulty waiting my turn

1.55

0.34

I engaged in reckless behaviour

1.36

I interrupted others

1.27

I was active, restless or always on the go SUICIDALITY ITEM BANK

1.20

1.07 0.57 0.02

I seriously considered attempting suicide

6.51

1.16

I wanted to end my life

6.47

0.95

I thought I would be better off dead

6.19

0.69

29

Item

Slope (discrimin ation)

Difficulty parameters 1

My reasons for dying outweighed my reasons for living

6.07

0.97

I thought that killing myself was the best solution to my problems

6.06

0.93

I thought that the only way out for me was to die

5.77

0.88

I was ready to kill myself and was only waiting for the right opportunity

5.54

1.48

I made plans to take my own life

5.37

1.44

I wished I were dead

5.35

0.74

I wanted to make an active suicide attempt

5.34

1.38

I thought that dying would be a relief

4.91

0.52

I thought of when I would kill myself

4.69

1.21

I felt that I had no reason to live

4.53

0.70

I seriously wanted to harm myself

4.21

1.20

I thought that my family and friends would be better off if I was dead

4.11

0.84

I felt that I deserved to die

3.60

1.15

I felt capable of carrying out a suicide attempt

3.39

1.14

I concealed from others that I was thinking about suicide

3.20

0.88

I thought that things would not get better

3.01

0.39

I was in the deepest despair and distress

2.93

0.86

I mentioned to someone that I might end my life

2.83

1.69

I heard voices telling me to kill myself

2.25

2.68

I was unafraid of dying

1.26

0.35

5.68

1.56

4.76

1.29

4.52 4.47

1.41 1.32

2

3

4

1.3 4 1.2 6 1.2 9 1.8 2 1.7 6 1.1 6 1.7 5 1.0 0 1.5 3 1.1 1 1.5 7 1.2 7 1.5 6 1.5 6 1.1 1 0.8 6 1.3 7 2.1 8 2.8 3 0.9 3

1. 78 1. 72 1. 77 2. 14 2. 07 1. 65 2. 16 1. 53 1. 99 1. 63 1. 89 1. 73 1. 99 1. 97 1. 33 1. 46 1. 97 2. 66 3. 16 1. 61

2.2 2 2.2 0 2.1 2 2.4 8 2.4 5 2.0 6 2.5 4 2.0 1 2.5 1 2.1 7 2.4 5 2.1 7 2.4 2 2.3 0 1.5 2 2.1 5 2.6 7 3.0 4 3.6 5 2.1 8

2.0 9 1.8 2 1.7 6 1.9

2. 55 2. 21 2. 23 2.

2.8 1 2.7 2 2.5 0 3.0

DRUG USE ITEM BANK I spent a lot of my time getting drugs I made sure that I had drugs or money for drugs before concentrating on other things If something or someone stopped me from using drugs when I wanted to get high, I would become anxious or upset I lost interest in activities and hobbies because of my drug use

30

Item

Slope (discrimin ation)

Difficulty parameters 1

I needed to take a drug the morning after heavy drug use

4.46

1.67

I felt I had to carry on taking drugs once I had started I felt powerless to prevent myself taking drugs when I was anxious, uptight or unhappy

4.22

1.53

4.20

1.30

I neglected my family because of my drug use I continued to use drugs even though I thought it was causing problems with my emotions, nerves or mental health

3.88

1.69

3.79

1.40

I spent a great deal of time thinking about drugs I failed to do what was normally expected of me because of my drug use

3.70

1.27

3.66

1.45

I felt tense or anxious as my supply of drugs began to run out

3.60

1.15

I planned my days around taking drugs I deliberately used drugs before going out if I felt there might not be the opportunity to use them later

3.56

1.41

3.35

1.26

I spent a lot of my time using drugs

3.34

1.24

I found myself taking more drugs than I intended to

3.27

1.25

I used drugs to deal with my frustration

3.16

0.98

I felt psychologically dependent on drugs

3.13

1.27

I wanted or needed to reduce my drug use, but was unable to My family or friends worried or complained about my drug use I felt that I needed more drugs than I used to in order to get the same effect

3.09

1.54

3.06

1.60

2.93

1.37

I regretted something I did as a result of my drug use

2.57

1.59

I lost weight or did not eat properly because of my drug use I used drugs in situations that could be unsafe, such as driving, operating machinery or caring for children

2.53

1.55

2.49

1.56

I felt physically dependent on drugs

2.44

1.44

I particularly enjoyed getting a really strong effect from drugs

1.95

0.72

I experienced withdrawal symptoms after using drugs

1.83

1.45

2

3

4

1 2.0 1 1.9 8 1.6 2 2.0 7 1.6 4 1.9 3 2.0 4 1.6 4 1.9 2 1.5 4 1.7 5 1.7 0 1.3 6 1.6 5 1.9 0 2.2 4 1.8 1 2.3 8 2.1 8 2.0 4 1.8 8 1.1 8 2.1 9

35 2. 69 2. 33 2. 07 2. 55 2. 06 2. 41 2. 49 2. 22 2. 49 2. 01 2. 28 2. 30 1. 99 2. 23 2. 21 2. 71 2. 33 2. 83 2. 59 2. 35 2. 66 1. 70 2. 55

6 3.0 5 2.6 8 2.6 9 2.9 7 2.4 8 2.9 7 3.1 3 2.5 7 2.7 6 2.6 2 2.9 9 2.6 9 2.6 5 2.5 6 2.6 5 3.0 2 2.9 6 3.1 6 3.1 3 3.1 6 2.9 9 2.1 8 2.8 1

31

Figure 1: Information curves for the eight item banks 180

SAD

160

PD PTSD

140

Test information

OCD 120

ADHD Drug use

100

Psychosis Suicidality

80 60 40 20

0

-3

-2

-1

0

1

2

3

4

5

6

Severity / item difficulty (θ)

Highlights 

Eight new item banks of 19-47 items were developed to assess a range of mental health problems



Items were selected on the basis of local independence and measurement invariance



Item response theory parameters weighted to the general population are provided



The item banks will form a new system of static and adaptive mental health screeners

32