|What is the ICPE?|
The International Consortium in Psychiatric Epidemiology (ICPE) is a consortium funded by the US National Institute of Health. The main goal of the ICPE is to facilitate cross-national comparative epidemiologic studies of psychiatric disorders. This will be done by bringing together researchers from around the world who are carrying out general population epidemiologic surveys of psychiatric disorders based on the Composite International Diagnostic Interview (CIDI) and providing various resources for coordination. The Principal Investigator of the ICPE is Ronald Kessler, from Harvard Medical School. The CO-PI is Bedirhan Ustun, from the World Health Organization. Both Kessler and Ustun have expertise in psychiatric epidemiology. The administrative home of the ICPE is the Survey Research Center (SRC) in the Institute for Social Research (ISR) at the University of Michigan. The ISR is one of the leading academic survey research training centers in the world. The resources of ISR, including its state-of-the-art survey methodology programs in study design, sampling, instrument development, field and data processing, and its extensive summer training programs in data collection and data analysis, are at the disposal of ICPE members.
Resources of the Consortium
The main ICPE resources are as follows:
1. The ICPE hosts an annual meeting of members each year.
2. The Michigan World Health Organization CIDI Training and Reference Center (TRC) sponsors a series of Summer Institute courses in psychiatric epidemiology in conjunction with the meeting of the ICPE that are open to students from around the world. The course this year is a one-week CIDI training course taught by the staff of the Survey Operations Division of SRC.
Information about the CIDI TRC Summer Training Program can be obtained by clicking on the following link: Summer Training Program Information.
3. The ICPE supports a large UNIX server housed administratively within the Inter-university Consortium of Political and Social Research (ICPSR) at the University of Michigan for the life of the consortium. The ICPSR is the world's largest academic behavioral science data archive. ICPSR staff are highly experienced in transporting archived survey data using a variety of media throughout the world.
4. The ICPE supports a senior data manager to build and maintain files on the server. The files contain data from the surveys of the ICPE members.
5. The ICPE will maintain a library of SAS macros and other program files that can be used by ICPE members to analyze their data. SAS is the core statistical package of the ICPE.
6. The ICPE maintains an e-mail network for communication among members.
7. The ICPE has a webpage for purposes of disseminating products of the Consortium to the scientific community, including preprints and reprints of Consortium publications.
8. The ICPE UNIX server can be used by any member, with the use of a special unique password, for data analysis of large cross-national files. The server can also be used to partition and then download data and program files from the server by FTP to carry out analyses locally.
9. The ICPE makes available the services of a senior analyst to carry out a series of common analyses and generate a series of common tables and figures of basic results for each survey in the pooled data file (e.g., prevalence tables, age-of-onset curves, analyses of cohort effects, basic survival models studying the effects of sociodemographic variables on onset and course of disorders, etc.). The set-up files for these basic runs are saved for use in generating parallel results for new surveys as they are added to the pooled data file.
10. The ICPE provides biostatistical consultation to ICPE members.
11. The ICPE provides coordination of several sorts for collaborations that go beyond simple comparative data analyses, including the following:
(a) A work group was established at the 1996 annual meeting under the direction of Dr. Ustun for collaboration in CIDI Advisory Committee test-retest and validity studies of the CIDI.
(b) Work groups have also been established for collaboration in the development of expanded survey modules that will be available for use by ICPE members to increase the opportunities for cross-survey comparisons. Interest has been expressed already in modules on exposure to ethnic violence and other types of interpersonal trauma, services, broadly defined risk factors, and cultural sensitivity in assessing psychopathology. Other modules will be added as interest emerges.
(c) A number of groups have developed CIDI computer programs capable of being used in personal interviews (CAPI: Computer-Assisted Personal Interviews). These programs were made available for inspection at the 1996 annual meeting of the ICPE.
12. The ICPE coordinates cross-national data analyses and preparation of reports using several approaches:
(a) Concerning data analysis, we have three approaches. Our main approach, as noted above, is to develop a common cross-national data file and have a senior analyst located in Ann Arbor carry out standard data analyses in parallel across all the surveys. Both survey-specific and pooled results are produced.
(b) The second data analysis approach, again, noted above, is to facilitate analyses carried out by collaborators on the extremely large ICPE data files by giving them access to the ICPE UNIX server for remote batch computing. A library of SAS macros and biostatistical consultation is available to collaborators to facilitate this type of work.
(c) The third data analysis approach will be to have collaborators use the Michigan server to partition the large ICPE data files for downloading to their own computers for local analysis. Again, the SAS macro library and access to ICPE biostatistical support will be available to facilitate this type of work.
13. Concerning coordination of report preparation, the following approaches are used in addition to any others the ICPE members devise:
(a) We use a two-tiered system of analysis. In the first tier, common analyses of each separate Consortium survey are carried out by the ICPE senior data analyst. The results of the analyses for a single survey are distributed to the collaborators involved in that survey in order to facilitate analysis of their own data. Access to the ICPE server for computationally intensive analyses, access to the SAS macro library, and consultation with the ICPE biostatistical support staff are additional ways to facilitate local analyses. Local collaborators are responsible for writing reports of the results of their own studies and should publish these reports as they see fit in professional journals.
(b) In addition, the senior analyst carries out pooled between-country analyses. The PI and CO-PI take responsibility for preparing reports based on these pooled analyses. Regarding publication of these analyses, two approaches are used. First, it is sometimes useful to assemble the results of the parallel within-country analyses in addition to the pooled between-country analyses in edited volumes that have one chapter for each country and a separate overview chapter for pooled analyses. Second, there are occasions when the significance of the results does not warrant the publication of an edited book, in which case the pooled analyses are published separately as a journal article jointly by all ICPE members whose data are included in the report.
For more information on the ICPE, contact:
Tel: 617 432 3587
All content © 2005 Harvard School of Medicine