The World Mental Health Composite International Diagnostic Interview
The World Mental Health Survey Initiative
Health and Work Performance Questionaire
Appendix to:
Kessler, R.C., Barker, P.R., Colpe, L.J., Epstein, J.F., Gfroerer, J.C., Hiripi, E., Howes, M.J, Normand, S-L.T., Manderscheid, R.W., Walters, E.E., Zaslavsky, A.M. (2003). Screening for serious mental illness in the general population. Archives of General Psychiatry. 60(2), 184-189.

Mental Health Module

Appendix for Screening for Serious Mental Illness in the General Population:

1. Introduction

This appendix presents the text of the ACASI module used in the in-person interviews carried out as the second phase of the Boston pilot study. The nine sections of this module are as follows:

I. The ACASI Tutorial: Questions INTROACASI through ANYQUES (pp. 1-3)

II. Demographics: Questions DEMINTRO through SCHOOL (pp. 3-4)

III. Work Limitations: Questions WORKSCHL through PROBTYPE (p. 4)

IV. The truncated CIDI-SF: Questions PNATTACK through NPVISION (pp. 5-13) OCD questions are OCBEH1 through OCWEEK2 (pp. 9-10)

V. Treatment: Questions SEEDR through MEDWEEK (pp. 13-14)

VI. The K/10 scales: Questions DSNERV1 through DSDOWN (pp. 14-16)

VII. The K/6 scales: Questions DSNERV1, DSHOPE, DSFIDG, DSNOCHR, DSEFFORT and DSDOWN (pp.14-16)

VIII. The modified WHO-DAS: Questions LITHINK through IMDAYS (pp. 17-21)

IX. Perceived effectiveness of treatment: Question IMHELP (p. 21)

2. Scoring the truncated CIDI-SF diagnostic classifications

a. Likely panic attack was defined dichotomously as PNREACT = 1. Number of attacks (PNATKNUM) was also included as a predictor of SMI in some of the exploratory prediction equations.

b. Number of weeks with MD (the larger of the two entries in responses to DEWEEK1 and DEWEEK2) was defined continuously in the range 0-52 and was included as a predictor of SMI in some of the exploratory prediction equations. Missing values were coded 0. Likely major depression (MD) was defined dichotomously as (DEFEELPR = 1 or DELOSTPR = 1) and number of weeks with MD >= 6 and DEDAYSAD=1-2.

c. Likely mania was defined dichotomously as either (MASLEEP = 1 and MAPWRS = 1) or MAMED = 1.

d. Likely social anxiety disorder was defined dichotomously as PHUPSET1=3 and PHAVOID1 = 1 and PHLIFE1 = 4.

e. Likely agoraphobia was defined dichotomously as PHFEAR=3 and PHAVD = 1 and PHINT = 3-4.

f. Likely obsessive-compulsive disorder (OCD) was defined dichotomously as either ( 26<= OCWEEK1 =<52) or ( 26<=OCWEEK2 =<52). Because all respondents with likely OCD also met criteria for a number of other CIDI-SF disorders, OCD was deleted from the version of the SMI Screening module included in the NHSDA.

g. Likely generalized anxiety disorder (GAD) was defined dichotomously as GAPROB = 1. Number of weeks with GAD (GAWEEK1) was defined continuously in the range 0-52 and was also included as a predictor of SMI in some of the exploratory prediction equations. Missing values were coded 0.

h. Likely post-traumatic stress disorder (PTSD) was defined dichotomously as PTWEEK1 >= 4.

i. Possible non-affective psychosis (NAP) was defined dichotomously as the sum of yes responses in the question series NPVOICE through NPVISION being in the range 2-7. Number of NAP symptoms was defined continuously in the range 0-7, but could not be included in the prediction equation because of sparse data.

3. Scoring the K10/K6

Optimal continuous scaling of the K10 and K6 scales based on the two-parameter IRT model used to select the items for these scales is described by Kessler et al. (2002). Simple scaling involves transforming response categories of the ten K10 questions (DSNERV1 through DSDOWN) so that "all of the time" is coded 4, "Most of the time" 3, "some of the time" 2, "a little of the time" 1, "none of the time" 0, and "Don't know" and "Refuse" also coded 0, and summing across the transformed responses to obtain a scale with a 0-24 range for the K6 (DSNERV1, DSHOPE, DSFIDG, DSNOCHR, DSEFFORT, and DSDOWN) and 0-40 for the K10. The rank-order correlation between scores based on optimal scaling and simple scaling were greater than .90 for both the K10 and the K6 in the pilot sample.

4. Scoring the modified WHO-DAS

The modified WHO-DAS was scored continuously in the range 0-48 by recoding the 1-4 response categories of the four-category WHO-DAS variables (LITHINK through LIWKQUIC) to 0-3 and summing after three series of recodes. The first series of recodes classified "Don't know" and "Refused" responses as 0 on the transformed response scale. The second series of recodes classified original responses coded 5. Five codes appeared in only eight of the WHO-SAD questions. These questions deal with functions that are not performed by all people, such as cleaning house. The 5 code allowed the respondent to tell us that they do not perform this kind of activity. A follow-up question asked people who responded in this way if the reason for their nonperformance was that they were unable to perform these activities because of health problems. Five codes were recoded as 3 if the response to this follow-up question was yes. The 5 codes were recoded to 0 otherwise. The third series of recodes assigned all respondents who skipped out of the WHO-DAS questions a score of 0 on the 0-48 scale.

Frequency of impairment (IMDAYS) was defined continuously in the range 2-7, with "Every day" coded 7, "Most days" coded 5, and "Only one or two days a week" coded 2. "Don't know" and "Refused" were coded 2.

Number of weeks with impairment (IMWEEK) was defined continuously in the range 0-52. Missing values were coded 0. "Don't know" and "refused" were coded 4.

Two multiplicative measures were also created and used as predictors of SMI in some exploratory equations. One was the product of WHO-DAS scores times weeks of symptoms. The second was the product of WHO-DAS times weeks of symptoms times frequency of impairments.

5. Scoring other predictors

a. Disability due to emotional problems was defined dichotomously as PROBTYPE = 2-3 versus all others.

b. Professional help-seeking for emotional problems was defined dichotomously as SEEDR = 1.

c. Number of outpatient visits for emotional problems was defined continuously in the range 0-365 as the larger of the entries in either DRVISITS or DRVISDK. Missing values were coded 0.

d. Hospitalization for emotional problems was defined dichotomously as HOSPTL = 1.

e. Medication for emotional problems was defined dichotomously as MEDS = 1.

f. Weeks on medication for emotional problems (MEDWEEK) was defined continuously in the range 0-52. Missing values were coded 0.

6. Revised module used in the NHSDA

The instrument presented here was revised based on the results of the Boston pilot study for use in the NHSDA. The revised version of the module, as it appears in the NHSDA, can be found at http://www.samhsa.gov/oas/nhsda/methods.cfm

7. Further refinement of scoring

The simple scoring procedure presented in the paper will be reviewed and, if needed, revised based on analysis of the National Comorbidity Survey Replication (NCS-R).
----------
Kessler, R.C., Andrews, G., Colpe, L.J., Hiripi, E., Mroczek, D.K., Normand, S.-L.T., Walters, E.E., & Zaslavsky, A. (2002). Short screening scales to monitor population prevalances and trends in nonspecific psychological distress. Psychological Medicine. 32(6), 959-976.