SCID TRAINING SEQUENCE OF STEPS
Ideally, training should involve the following sequence:
1. Study the Basic Features, Conventions and Usage, as well as DO’s and DON'Ts sections in the SCID User’s Guide (included with your SCID-RV purchase).
2. Carefully read through every word of the SCID, making sure you understand all of the instructions, questions, and diagnostic criteria. As you are reading through each module, refer to the corresponding section in the SCID User’s Guide that discusses that part of the SCID. Also review the Diagnostic Features and Differential Diagnosis sections of the DSM-IV text for those disorders included in the SCID.
3. Now practice reading the SCID questions aloud so that eventually it sounds as if SCID is your mother tongue.
4. Try out the SCID with a colleague (or significant other) who can assume the role of a subject.
5. Watch the 11-hour, eight DVD training program: SCID 101.
6. Role-play the cases included in Appendix B of the User’s Guide with a colleague. These have been designed to take you through the SCID modules, not necessarily to demonstrate your dramatic talent.
7. Watch DVDs of SCID interviews and make your own ratings as the interview proceeds. Compare your ratings with the "expert" ratings of the interview that are included with each DVD. Click here for information on how to order the Recorded SCID Interviews.
8. Try out the SCID on actual subjects who are as representative as possible of those who will be included in your research study. If possible, these rehearsals should be joint interviews with all raters making independent ratings, followed by a discussion of the interviewing technique and all sources of rating disagreement.
9. Consider setting up an on-site SCID training workshop conducted by an associate of the Biometrics Research Department. The goal of these one- or two-day workshops is to provide live supervision of SCID interviews. Click here for more information.
10. Conduct a series of group SCID sessions where the SCID is administered by one interviewer, while the other SCID interviewers observe and contemporaneously make their own ratings. At the conclusion of the SCID interview, the interviewers go through the SCID and compare ratings, with discussion focusing on those ratings where there are disagreements. These exercises can help to insure that all of the SCID interviewers are in agreement with respect to their understanding of the DSM-IV diagnostic criteria and SCID methodology. It is recommended that these group SCID sessions be conducted periodically throughout the life of a study in order to minimize rater drift.
11. If multiple raters are going to be administering the SCID, it is important to implement procedures to ensure diagnostic agreement between the raters. Ideally you should try to do a "test-retest reliability study" in which the interview is repeated with the same subject within a short period of time by a second interviewer. You will learn the most from such a study if you record the interviews either on video or audiotape, then have each interviewer review and rate the other interviewer. This should be followed by a discussion of any sources of disagreement.
Alternatively, you can conduct a joint reliability study where SCID interviews conducted by one interviewer are observed by the other interviewers who independently make their own ratings. Although a joint reliability design is a less rigorous procedure for assessing reliability, the practical limitations of getting subjects to agree to repeated SCID interviews often necessitates this type of design. You will need to make a series of audio or videotapes conducted by the interviewers, suspending the skip-outs for all the diagnoses that are of interest, and have each tape rated by all the interviewers. In such a procedure, if the answer to the initial screening question (e.g., "Have you ever had a time when you were depressed or down for most of the day nearly every day?") is anything other than "No," the interviewer continues to ask all the subsequent questions. Raters will not know what judgments the interviewer is making, and the reliability of the diagnosis, as well as of the symptom ratings, can be assessed. In general, we would recommend a minimum of ten joint interviews, although the more the better. Another rule of thumb is to do enough interviews so as to have at least five cases of each type of diagnosis that you are interested in studying. For example, if you are doing a study in which you are identifying subjects with Panic Disorder and/or Generalized Anxiety Disorder, there should be enough interviews conducted for the purposes of determining reliability so that you end up with at least five cases of Panic Disorder and five cases of Generalized Anxiety Disorder. Note that SCIDs on subjects recruited for participation in the reliability study can potentially be used for the actual study you are conducting (i.e., you do not have to "waste" precious subjects for the reliability study) as long as you reach a group consensus about the SCID ratings.
12. Another way to insure that the SCID interviewers are administering the SCID properly is to record the interviewers and then send either a video or audiotape (plus the accompanying rated SCID) to Biometrics Research for review and critique. Click here if you want additional information.
13. The more of the above listed training steps you complete, the more skilled the SCID interviewers will become. Before letting a new trainee start doing SCID’s for the actual study, it is advisable to evaluate the trainee’s ability to administer the SCID properly. This is best done by someone at your site with SCID experience-- the last two pages of the SCID User’s Guide include a SCID evaluation form that indicates the various areas of proficiency that are required to be demonstrated by raters (e.g., "Obtained enough information to make judgments on each item" and "helped rambling subject to focus on the issue under consideration"). Alternatively, you can record the trainee’s interview on audiotape or videotape and sent the tape to us for review and critique (see step 12 above). When evaluating the trainee’s SCID skills, you should focus on those diagnoses that are most critical for the study, e.g., for an interviewer in a study of anxiety and depression, it may not be very important to deal with the nuances of the differential diagnosis between schizophrenia and schizoaffective disorder. The most important issue is how competent the interviewer is with subjects similar to those who will be assessed in the study. Unfortunately, there is no formal mechanism to "certify" individuals as SCID trained because we have not yet been able to develop an all-purpose assessment procedure that would insure the SCID interviewer’s ability to handle any situation that might arise with a subject.