Impact of September 11, 2001 Events on Substance Abuse and Mental Health in the New York Area

Appendix D

Mental Health Variable Specifications for
September 11 Analysis

Listed below are the mental health problem variable specifications used for the September 11 analysis. These specifications are based on those defined in the 2001 CAI questionnaire, with a few minor adjustments. Note that these data are only available for 2001. Also, only persons aged 18 or older were asked these questions; thus, the 23,133 youths aged 12 to 17 interviewed in 2001 were excluded from the analyses. This document uses the CAI specification variable names for ease in referencing the specifications, but final edited data were used for generating all the estimates found in the tables.

D.1 Distress

Listed below are the questions and definition of the source variables that were used to define "distress."

Question: "Think of one month in the past 12 months when you were the most depressed, anxious, or emotionally stressed. If there was no month like this, think of a typical month. During that month, how often did you feel:

DSEFFORT - that everything was an effort?
DSDOWN - down on yourself, no good, or worthless?
DSTIRED - tired out for no good reason?
DSHOPE - hopeless?

DSNERV1 - nervous?
DSNERV2 - so nervous that nothing could calm you down?

DSFIDG - restless or fidgety?
DSSITSTL - so restless or fidgety that you could not sit still?

DSDEPR - sad or depressed?
DSNOCHR - so sad or depressed that nothing could cheer you up?"

For each of the distress criteria, we defined a value of 1 to indicate the respondent qualified for that distress criteria and a value of 0 to indicate he/she did not qualify. The first four source variables listed above are used independently to define one distress criterion each. Due to some connecting skip patterns for a few of the other source variables, we have paired those questions with connecting skip logic into one criterion each for distress. Therefore, the last six source variables listed above are grouped into three pairs since the first of each pair is a gateway question to the second. A combined rule over both questions was used for these to determine if respondents qualified for that measure.

The criteria variables were recoded as follows for the first four source variables listed above (DSEFFORT, DSDOWN, DSTIRED, DSHOPE):

1. Defined as positive (i.e., recode = 1) if respondents reported that they experienced that situation "all of the time, most of the time, or some of the time" (source var = 1,2,3).
2. Otherwise, defined as negative (i.e., recode = 0) if respondents reported that they experienced that situation "a little of the time" or "none of the time" (source var = 4,5).
3. Otherwise set to missing and excluded.

The criteria variables were recoded as follows for the three paired variables listed above (DSNERV1, DSNERV2; DSFIDG, DSSITSL; DSDEPR, DSNOCHR):

1. Defined as positive (i.e., recode = 1) if respondents reported that they experienced the first OR second situations "all of the time, most of the time, or some of the time" (either source var = 1,2,3). This includes those who indicated a positive for the first situation but did not respond to the second situation (don't know, refused, etc).
2. Defined as negative (i.e., recode = 0) if respondents reported that they experienced the first situation OR second situation "a little of the time" or "none of the time" (second source var = 4,5). This includes those who indicated a negative for the first situation but did not respond to the second situation (don't know, refused, etc).
3. Otherwise set to missing and excluded (both answers of don't know, refused, no answer, etc).

For reference, we have included a sample of one of the paired criteria to show the way they were created (note that all the 2000 cases [n = 71,764] are missing since data were not available in 2000). Missing data are indicated by a period in the age 18 column. Data are shown for persons aged 18 or older (age18 = 2). DSTNRVOS corresponds to CAI spec DSNERV1 and DSTNCALM corresponds to DSNERV2:

Note that even though we are combining these paired variables into one criterion over both variables, it still classifies a respondent as a "Yes" for distress if he or she said "Yes" to either question.

Therefore, our final distress analysis variable (DISTRESS) is defined as follows:

1. Defined as distressed (i.e., DISTRESS = 1) if respondents reported positive for any of the seven criteria.
2. Otherwise, defined as not being distressed (i.e., DISTRESS = 0) if respondents reported negative to four or more of the seven criteria.
3. Otherwise, set to missing and excluded.

Our final number of distress symptoms variable (DTSYMP) is defined as follows:

1. Defined only for respondents who are distressed (DISTRESS = 1).
2. Then number of symptoms is computed by summing up the number of criteria for which respondents reported positive, giving us a range of 1 to 7.
3. Otherwise, set to missing and excluded.

We have included below some frequencies to show the final analysis variables based on this criterion (note that all the 2000 cases [n = 71,764] are missing since data were not available in 2000).

age18	dtsymp	Frequency	Percent	Cumulative Frequency	Cumulative Percent
.	.	71764	51.01	71764	51.01
2	.	22832	16.23	117729	83.68
2	1	6394	4.54	124123	88.22
2	2	3539	2.52	127662	90.74
2	3	2790	1.98	130452	92.72
2	4	2360	1.68	132812	94.40
2	5	2310	1.64	135122	96.04
2	6	2331	1.66	137453	97.70
2	7	3240	2.30	140693	100.00

D.2 Disorder

These are 20 disorder questions coming from a variety of gateway questions, and we have included pertinent question information for each source variable.

DEFEELPR - Changes in sleep, energy, appetite, or ability to concentrate when sad/depressed?
DELOSTPR - Changes in sleep, energy, appetite.....when lost interest in things?
MASLEEP - Could hardly sleep, but still not tired when excited or hyper?
MAPWRS - Had special powers when excited or hyper?
GAPROB - Difficulties in sleep or concentration, feeling dizzy, easily tired, on edge, or irritable?
NPVOICE - Voices coming from inside your head or out of the air?
NPFORCE - Force taking over your mind?
NPXRAY - Force inserting thoughts into your head by x-rays or laser beams?
NPTHGHTS - Your own thoughts were being stolen out of your mind?
NPCOMM - Some force trying to communicate with you by sending signs or signals?
NPPLOT - Unfair plot going on to harm you?
NPVISION - In past 12 months have you seen a vision?
PNATKNUM - How many sudden attacks of fear in the past year?
PTWEEK1 - In how many weeks did you have adverse reactions to very stressful situations?

PHAVOID1 - How often do you try to avoid your most feared social situation?
PHAVD - How often do you try to avoid other feared public activities?
PHUPSET1 - How fearful of most feared social situation?
PHFEAR - How fearful of most feared public situation?
PHLIFE1 - How much does fear of social situation interfere with everyday activities?
PHINT - How much does fear of public situation interfere with everyday activities?

Due to the level of variation in the possible responses and the skip logic of the preceding source variables, we defined disorder criteria such that a value of 1 indicates the respondent qualified for that criterion and a value of 0 indicates he/she did not qualify. The first 14 source variables listed above are recoded independently. The last six source variables are interrelated due to skip logic and are grouped into three pairs. For the pairs, a combined rule over both questions was used to determine if respondents qualified for that criteria.

The first 14 criteria were recoded as either a positive, negative, or unknown as described below:

Positive responses to first 14 source variables:

• For each of the first 12 source variables listed above, defined as positive (i.e., recode = 1) if respondents reported a "Yes" (source = 1).
• For PNATKNUM, defined as positive (i.e., recode = 1) when respondents reported between 4 and 500 sudden fear attacks (source = 4 - 500).
• For PTWEEK1, defined as positive (i.e., recode = 1) when respondents reported between 4 and 52 weeks where they had adverse reaction to stress (source = 4 - 52).

Negative responses to first 14 source variables:

• For each of the first 12 source variables listed above, defined as negative (i.e., recode = 0) if respondents reported a "No" or legitimately skipped question (if applicable) (source = 2,99).
• For PNATKNUM, defined as negative (i.e., recode = 0) when respondents said they had less than four attacks of fear in past year or legitimately skipped question (source = 1,2,3,999).
• For PTWEEK1, defined as negative (i.e., recode = 0) when respondents said they had adverse reactions to stress in less than 4 weeks or legitimately skipped question (source = 1,2,3,99).

Otherwise for first 14 source variables, all other cases defined as missing and excluded.

The last six source variables have connecting skip logic, so they were grouped in pairs accordingly. PHUPSET1 is a gateway question for both PHAVOID1 and PHLIFE, and PHFEAR is a gateway question for both PHAVD and PHINT.

Definition for four paired criteria:

• Paired criteria 1 defined as positive when respondent reported "Fear is severe" for social situations (PHUPSET1 = 3) OR tries to avoid social situations "Often" (PHAVOID1 = 1).
• Paired criteria 1 defined as negative when respondent reported "Fear is mild" or "Fear is moderate" for social situations or legitimately skipped question (PHUPSET1 = 1,2,99) OR said "Sometimes, rarely, never" avoids social situations or legitimately skipped question (PHAVOID1 = 2,3,4,99).

• Otherwise, all other cases defined as missing and excluded.

• Paired criteria 2 defined as positive when respondent reported "Fear is severe" for social situations (PHUPSET1 = 3) OR fear interferes with everyday activities "A lot" (PHLIFE1 = 4).
• Paired criteria 2 defined as negative when respondent reported "Fear is mild" or "Fear is moderate" for social situations or legitimately skipped question (PHUPSET1 = 1,2,99) OR when fear interferes with everyday activities "Not at all, a little, or some" or legitimately skipped question (PHLIFE1 = 1,2,3,99).
• Otherwise, all other cases defined as missing and excluded.
• Paired criteria 3 defined as positive when respondent reported "Fear is severe" in situation (PHFEAR = 3) OR "Often" tries to avoid situation (PHAVD = 1).
• Paired criteria 3 defined as negative when respondent reported "Fear is mild" or "Fear is moderate" for situation or legitimately skipped question (PHFEAR = 1,2,99) OR said "Sometimes, rarely, never" avoids situation or legitimately skipped question (PHAVD = 2,3,4,99).
• Otherwise, all other cases defined as missing and excluded.

• Paired criteria 4 defined as positive when respondent reported "Fear is severe" in situation (PHFEAR = 3) OR fear interferes with everyday activities "A lot" (PHINT = 4).
• Paired criteria 4 defined as negative when respondent reported "Fear is mild" or "Fear is moderate" in situation or legitimately skipped question (PHFEAR = 1,2,99) OR when fear interferes with everyday activities "Not at all, a little, or some" or legitimately skipped question (PHINT = 1,2,3,99).
• Otherwise, all other cases defined as missing and excluded.

For reference, we have included a sample of one of the paired criteria (do13) to show the way they were created (note that all the 2000 cases [n = 71,764] are missing since data were not available in 2000). PHBLEVEL corresponds to CAI specification PHUPSET1 and PHBAVOID corresponds to PHAVOID1:

Using the above recoded (yes, no, missing) criteria variables, we then created two final disorder analysis variables that were used to generate the disorder estimates. We have included below some frequencies to show the final analysis variable based on these criteria.

Disorder (DISORDER):

1. Defined as having disorder (DISORDER = 1) if respondents were coded as positive for 1 or more of the 18 criteria (14 individual and 4 paired criteria).
2. Otherwise, defined as not having disorder (DISORDER = 0) if respondents were coded as negative for 9 or more source variables.
3. Otherwise, set to missing and excluded.

Mean Number of Disorder Symptoms (DOSYMP):

1. Defined only for respondents who are defined as having disorder (DISORDER = ).
2. Compute number of symptoms by summing up the number of criteria for which respondents reported positive, giving us a range of 1 to 18.
3. Otherwise, set to missing and excluded.

age18	dosymp	Frequency	Percent	Cumulative Frequency	Cumulative Percent
.	.	71764	51.01	71764	51.01
2	.	35147	24.98	130044	92.43
2	1	5199	3.70	135243	96.13
2	2	2582	1.84	137825	97.96
2	3	1164	0.83	138989	98.79
2	4	625	0.44	139614	99.23
2	5	377	0.27	139991	99.50
2	6	250	0.18	140241	99.68
2	7	157	0.11	140398	99.79
2	8	124	0.09	140522	99.88
2	9	59	0.04	140581	99.92
2	10	47	0.03	140628	99.95
2	11	28	0.02	140656	99.97
2	12	14	0.01	140670	99.98
2	13	11	0.01	140681	99.99
2	14	7	0.00	140688	100.00
2	15	3	0.00	140691	100.00
2	17	2	0.00	140693	100.00

This page was last updated on June 16, 2008.

SAMHSA, an agency in the Department of Health and Human Services, is the Federal Government's lead agency for improving the quality and availability of substance abuse prevention, addiction treatment, and mental health services in the United States.