According to the Field Representative's Manual for 1999, this set of questions on panic attacks was included "to determine if the Sample Adult has attacks when they suddenly feel frightened, anxious, or very uneasy; whether they occurred in a life-threatening situation; how many attacks in the past year and when was the most recent one; did they happen when the Sample Adult was not in danger or the center of attention; and what other physical reactions occurred when they were having these attacks."

The 1999 NHIS CIDI-SF includes screening scales for major depression and anxiety (generalized anxiety disorder and panic attack). More information on the major depression questions--for dyphoric mood and for anhedonia, respectively--can be found in DEP2WKEV and DEPLIFUN. For more information on the generalized anxiety disorder questions, see WOR1MO.

The first of these questions about the circumstances of the attacks asked, "Did any of these attacks occur when you were in a life-threatening situation?" (PANICTHREAT).

Persons with an affirmative response in PANICTHREAT were asked, "Did any of these attacks occur when you were not in a life-threatening situation?" (PANICNOTHRET).

Respondents who indicated that their attack(s) occurred in a non-life-threatening situation (A "No" in PANICTHREAT or a "Yes" in PANICNOTHRET) received an additional screening question that asked, "Did [this attack/all of these attacks] happen in a situation when you were not in danger or not the center of attention?" (PANICNODANGR). Persons whose answer to PANICTHREAT was "unknown" and, in the NHIS instrument, those who volunteered that their attack "wasn't in a life-threatening situation" in PANICTHREAT also received PANICNODANGR.

In the NHIS instrument, persons who passed the first three filter questions (PANICYR, PANICTHREAT, PANICNOTHRET) and who answered "Yes" in PANICNODANGR then received a series of questions on panic attack symptoms.

These additional filter questions determined whether the attacks occurred in situations that usually caused the respondent unreasonably strong fear, and only respondents who indicated that their attacks did not occur in such situations then received the panic attack symptom questions in the WHO CIDI-SF.

Respondents who passed the sets of filter questions in the WHO and NHIS instruments (namely, PANICYR, PANICTHREAT, PANICNOTHRET, and PANICNODANGR for the NHIS version) were asked about the following panic attack symptoms:

• Heart pounds in panic attacks (PANICHEART)

• Chest/stomach discomfort in panic attacks (PANICCHEST)

• Sweat in panic attacks (PANICSWEAT)

• Tremble in panic attacks (PANICTREMBL)

• Hot flashes or chills in panic attacks (PANICCHILL)

• Surroundings seem unreal in panic attacks (PANICUNREAL)

In both the NHIS and WHO CIDI-SF instruments, persons who passed the first three filters (PANICYR, PANICTHREAT, PANICNOTHRET) were also asked, "About how many attacks did you have in the past 12 months?" (PANICYRNO).

Respondents who received PANICYRNO were also asked, "In what month and year did you have the most recent attack?" Responses to this question were recorded in two variables, PANICWHENMO and PANICWHENYR, that respectively report the month and year of the respondents' most recent attack.

According to the Appendix on Adult Mental Health Documentation included in the 1999 NHIS Survey Description, "NCHS suggests that individual CIDI-SF items not be used for purposes beyond those recommended by the World Health Organization (WHO)." The Survey Description then directs users to the WHO's website for instructions on how to calculate the probability of psychiatric diagnosis based on responses to individual CIDI-SF items.

As previously noted, respondents must pass all filter questions to receive further questions on panic attack symptoms (with four filter questions included in the NHIS instrument). Respondents who do not pass these filters receive a score of zero (zero probability of psychiatric caseness for panic attack).

According to the scoring system for the CIDI-SF screening scale for panic attack, respondents who pass the filter questions and go on to the panic attack symptom questions receive one point for each of the symptoms they endorse. Summing the number of endorsed symptoms results in a panic attack score ranging from 0 to 6. Respondents with a score of 3 or more are classified as a probable case for panic attack.

Users are cautioned, however, that the WHO CIDI-SF survey instrument uses more stringent criteria in its skip pattern than does the NHIS survey instrument. As mentioned above, according to the scoring guide published by the World Health Organization, only respondents who report that their attacks do not usually occur in situations that cause them unreasonably strong fear receive further symptom questions. The NHIS survey instrument, by contrast, does not include such exclusion criteria. More people will therefore be classified as probable cases for panic attack with the NHIS survey instrument than with the WHO CIDI-SF instrument.

Respondents who failed to pass the full series of filter questions (i.e., persons who did not have at least one panic attack in the past twelve months that occurred in a non-life-threatening situation when they were neither in danger nor the center of attention) receive a score of 0 in PANICCDC. For persons who did pass the series of filter questions, the value of PANICCDC equals the number of physical symptoms of panic attack acknowledged, with a maximum value of 6. PANICCDC thus has a value ranging from 0 to 6, with a score of 3 or more indicating a probable case for panic attack.