Predictors of Help Seeking Among Connecticut Adults After September 11, 2001 I Mary L. Adams, MS, MPH, Julian D. Ford, PhD, and Wayne F.

Transcription

1 I RESEARCH AND PRACTICE I Predictors of Hel Seeking Among Connecticut Adults After Setember 11, 21 I Mary L. Adams, MS, MPH, Julian D. Ford, PhD, and Wayne F. Dailey, PhD The unrecedented events of Setember 11, 21, had widesread sychological and health effects in New York City as well as regionally and nationally. A study of 18 Manhattan adults conducted a few weeks after Setember 11 showed that 7.5% of resondents had symtoms consistent with osttraumatic stress disorder (PTSD) and that 9.7% aeared to have major deression.o Aroximately 75% of resondents who took art in a telehone survey of adult residents of New York State, New Jersey, and Connecticut reorted 1 or more sychological or health roblems as a result of Setember 11, with comarable results for all 3 states. 2 According to Schuster et al., 44% of a national samle of adults reorted exeriencing 1 or more of 5 sychological stress symtoms after Setember Physical effects and behavioral changes were also reorted in New York City and surrounding areas. Increases in symtom severity were found among resondents with asthma in lower Manhattan who reorted difficulty breathing owing to smoke and debris. 4 Vlahov et al. reorted that 28.8% of all resondents in that same survey oulation increased their use of cigarettes, alcohol, or marijuana after Setember 11.5 More than 1 in 5 smokers (21%) reorted increased smoking, and 3% of all resondents reorted increased alcohol consumtion, in a ost- Setember 11 survey of Connecticut, New Jersey, and New York State adults. 2 A key factor in recovery from the traumatic stress of disasters is timely receit of aroriate suort services. 6 Wang et al. found that eole in communities less affected by a major earthquake were more likely to develo PTSD and exhibited oorer recovery 3 to 6 months later than eole in more severely affected communities who received sustained hel, including health and mental health services. 7 Receit of hel also was more strongly associated with ostdisas- Objectives. We conducted a oulation-based telehone survey in an attemt to determine correlates of formal and informal hel seeking after Setember 11, 21. Methods. Between October 15 and December 31, 21, 1774 Connecticut Behavioral Risk Factor Surveillance System resondents were asked questions directly related to their exeriences of Setember 11. Results. Multivariate logistic regression analyses showed that receit of formal hel was redicted by slee roblems, close association with a victim, reorts of increased smoking or drinking, and receit of informal hel. Age, gender, reorts of 1 or more roblems, and formal hel seeking redicted receit of informal hel. Conclusions. Public health lanning and bioterrorism rearedness should include rograms addressing increased smoking and drinking, slee roblems, and bereavement in the wake of disasters. (Am J Public Health. 24;94: ) ter readjustment than extent of exosure to the earthquake. 8 Desite the benefits of such services, Caldera et al. 9 and Wang et al." found that most eole in affected communities do not seek hel foi stress or sychological roblems in the aftermath of disasters and that individuals with revious mental health roblems are most likely to seek hel. Factors thought to lay a role in seeking and receiving hel for mental and behavioral health roblems can be groued into 3 categories, according to Andersen's model. 1 "Predisosing characteristics," such as age, gender, cultural factors, and degree of exosure to trauma or roximity to victims," 1 have been shown to affect reorting of traumatic stress and inclination to seek hel in the wake of disaster. 1-3 "Enabling resources" such as health insurance coverage influence receit of care by affecting social suort or availability of services. 14 "Perceived need" leads a erson to actually decide to seek hel and to follow through with health care. Stress-related roblems, such as slee disturbances, mental and hysical health roblems, and increased smoking or alcohol consumtion, are associated with erceived need for hel.11" 5 Our study was designed to investigate redictors of obtaining hel after Setember 11, 21, among Connecticut adults who were regionally roximate to New York City. In articular, we sought to learn more about the arameters that distinguished hel seekers from those who did not seek hel, examining all 3 categories of factors that have been shown to influence hel seeking. We also saw this as an oortunity for ublic health ractitioners, mental health ractitioners, and members of academia to collaborate in an attemt to imrove ublic health rograms designed to resond to terrorism or other large-scale traumatic events or disasters. METHODS Data Data were collected through the Behavioral Risk Factor Surveillance System (BRFSS), a state-based telehone survey of randomly selected noninstitutionalized adults coordinated by the Centers for Disease Control and Prevention. 16 The samle of telehone numbers for each BRFSS state is udated quarterly to include newly connected hones, and each month a stratified subsamle is drawn to ensure that results will accurately reresent the full adult oulation. Connecticut decided to add questions in the fourth quarter of 21 to address the effects of Setember 11, with a articular focus on the World Trade Center attacks. The 17 questions added to measure sychological and emotional effects of the attacks were modifications of questions used after an earlier disaster Research and Practice I Peer Reviewed I Adams et al. American Joumal of Public Health I Setember 24, Vol 94,. 9

2 In Connecticut, the monthly BRFSS samle was stratified by county (n=8) for 21, and an annual total of 6 to 15 interviews were conducted in each county, deending on oulation. A total of 7752 surveys were conducted in 21; data derived from all 1774 interviews conducted between October 15 and December 31, 21, were used in the resent study. Twelve resondents who did not answer the question regarding receit of hel were removed from the analysis, resulting in a final samle size of Data were adjusted for number of adults and number of telehone lines in the household and for the different robabilities of selection in each stratum; data were further adjusted to be reresentative, in terms of age and gender, of the adult oulation of Connecticut Missing values were excluded from our analyses. Measures In our analyses, we used 2 deendent variables based on tye of hel resondents reorted receiving after Setember 11. Receit of hel was ascertained from a single question: "Did you get hel with roblems you have exerienced since the attacks?" Tye of hel received was assessed with a single question focusing on sources of hel. Those who received any formal suort services (e.g., medical assistance, services rovided by a sychologist, sychiatrist, social worker, other mental health rofessional, or religious counselor) were laced in the "formal hel" grou. The "informal hel" grou included those who reorted obtaining hel from family members, friends, neighbors, or any other source not identified as roviding formal hel. Twelve resondents who reorted receiving both tyes of hel were included in both grous. Resondents were asked whether they, or anyone they knew ersonally (limited to family members, friends, coworkers, acquaintances, and members of their community), had been a victim of the attacks. A "victim" was defned as someone who witnessed the attacks, who was injured or kidled in the attacks, or whose home or worklace was disruted or damaged. In our analyses, we groued resonses into 3 categories: victims or family members of a victim, those with a ersonal relationshi (other than family membershi) to a victim, and those with no relationshi to a victim. Resondents were asked how dose they were in roximity to the World Trade Center attack, and resonses were coded dichotomously as "in New York City" or "elsewhere." To assess access to suort relevant to loss and grieving, we asked resondents whether they had attended funerals or memorial services for friends, acquaintances, or community members killed in the attacks. Presence of health insurance coverage was used as a general measure of health care accessibility. As a means of assessing erceived roblems subsequent to Setember 11, resondents were asked "Since the attacks, have you exerienced any of the following feelings or roblems?" (This question, the 12th in a series of items focusing on Setember 11, followed the question asking whether hel was sought; however, it did not limit resonses to roblems directly resulting from the attacks.) Nervousness, worry, hoelessness, loss of control over external events, and worthlessness were groued, and resondents were coded dichotomously as having reorted none or I or more of these roblems. Anger, reorted by half of adl resondents, was not included owing to its ervasiveness 2 and, otentially, short duration. Slee roblems were considered searately. Information on only a few measures of interest was available before Setember 11, 21. Physical and mental health status were obtained from questions asked over the entire year as art of the BRFSS core.1 8 Poor hysical health days were measured via the following question: "w thinking about your hysical health, which includes hysical illness and injury, for how many days during the ast 3 days was your hysical health not good?" Poor mental health (measured as oor mental health days) was ascertained from the question "w thiriking about your mental health, which includes stress, deression, and roblems with emotions, for how many days during the ast 3 days was your mental health not good?" In most analyses, we were rimarily interested in any indication of roblems, so continuous variables were dichotomized ( days vs 1 or more days). Tobacco and alcohol revalence rates were also measured throughout 21. Smokers were defined as those who had smoked at least 1 cigarettes in their lifetime and currently smoked on some days or every day. Resondents who reorted that they had consumed at least 1 drink in the ast 3 days were considered drinkers. Changes in smoking and drinking were measured only in October through December. Current smokers who increased their smoking, those who began smoking after Setember 11, and those reorting increased alcohol consumtion since the attacks were considered to have increased their substance use. Data on ethnicity were derived from several questions that ermitted resondents to indicate more than one race/ethnicity. The result was the creation of 2 grous: White, limited to non-hisanic Whites, and non-white, which included individuals of all other races and ethnicities, including multiracial resondents. Age was entered into the multivariate model as a continuous variable (although it is resented categorically in Table 1 to show reresentations of distinct age cohorts). Statistical Analysis Stata Version 8. (Stata Cor; College Station, Tex) was used in all statistical analyses to account for the comlex samle design of the BRFSS. Characteristics of resondents receiving formal suort services or informal hel were examined and, via tests of indeendence based on the Pearson x 2 statistic, comared with characteristics of resondents who did not receive such hel. Multivariate logistic regression analyses were used to model receit of (1) formal hel and (2) informal hel as a function of the redictor variables suggested by the Andersen model,' along with otential confounders of current smoking and drinking. Odds ratios, 95% confidence intervals, and Wald F ratios were comuted. RESULTS The resonse rate for the study eriod (comleted interviews divided by sum of comleted interviews, terminated interviews, and refusals) was 69/o. Desite relatively low resonse rates, the quality of BRFSS data is high, and revalence data have been shown to be valid and reliable Table I resents characteristics of the study resondents according to tye of hel received. The figures Setember 24, Vol 94,. 9 American Journal of Public Health Acfams et al. I Peer Reviewed I Research and Practice I 1 597

3 TABLE I-Resondent Characteristics, bytye of Hel Received After Setember I1: Behavioral Risk Factor Surveillance System, 21 Gender Male Female Age, y Ž65 Race/ethnicity White n-white Location on Setember 11 New York Elsewhere Victim status Self or family member victim Other relation to victim association with victim Smoking status Current smoker t current smoker Alcohol use status Current drinker t current drinker Health insurance coverage P Attended funeral. of oor hysical health days > P Formal Hel Informal Hel Overall, no. (%), %, %, %,% Predisosing factors 727 (48.8) (51.2) (37.3) 813 (44.1) 328 (18.6) (83.6) (16.4) (4.2) (95.8) (7.7) 411 (22.3) 1217 (7.) (2.7) (79.3) (65.3) (34.7) Enabling factors 1638 (92.1) (7.9) (8.3) (91.7) Perceived need for services 571 (32.5) (67.5) Continued in the first column reflect actual numbers of survey resondents; ercentages are weighted and adjusted to be reresentative of all adults residing in the state. Results are groued for convenience according to the 3 categories of Andersen's model,' although some variables may reflect more than 1 category. Most resondents (84%) were non-hisanic White, and about half were male. More than 9% had health insurance coverage, and about 81 % were aged younger than 65 years. Nearly two thirds had consumed alcohol in the receding 3 days, and one fifth smoked. One third reorted that there were 1 or more days in the ast 3 days on which they would categorize their mental health as oor, the same ercentage reorting oor hysical health days in that time eriod. Aroximately half reorted 1 or more roblems, 5% reorted increased substance use, and 3 % either were victims or ersonally knew someone who was a victim. About 8% had attended a funeral, and 4% were in New York City on Setember 11. Overall, 6.4% (n= 117) of the survey resondents indicated that they had received hel for roblems exerienced subsequent to Setember 11; 3.3% reorted receiving formal suort (unweighted n=55), and 3.7% / reorted informal hel (unweighted n= 74). Twelve resondents reorted receiving both formal and informal hel, 43 resondents had received only formal hel, and 62 had received only informal hel. ne of the resondents reorted receit of hel from a social worker or emergency worker reresenting an organization such as the Red Cross. Table 2 shows the results of the multile logistic regression analysis that included all of the variables listed in Table 1. The following variables were significantly associated with receit of formal hel after adjustment for other variables: being a victim, being a family member of a victim or being involved in another relationshi with a victim, reorting slee roblems, reorting increased substance use, and reorting receit of informal hel. Predictors of receit of informal hel included being female, being younger, reorting 1 or more roblems, and receiving formal hel. Comarisons of resondents surveyed (in 21) before and after Setember 11 did not reveal changes in tobacco or alcohol use 1598 Research and Practice I Peer Reviewed I Adams et al. American Joumal of Public Health j Setember 24, Vol 94,. 9

4 TABLE 1-Continued. of oor mental health days.. of roblems' reorted 1 or more Slee roblems Increased substance uset. of resondents 63 (33.9) (66.1) (51.2) (48.8) (15.5) 43.7 < (84.5) (5.) (95.) '.1 < < te. Resondent numbers are unweighted; ercentages were adjusted to be reresentative of the state adult oulation by age and gender. P values were derived from Pearson x 2 tests of association. 'Including worry, nervousness, worthlessness, hoelessness, and lack of control over extemal events. btobacco or alcohol use, or both. revalence rates or in average number of drinks consumed er day. In addition, the ercentage of resondents reorting any oor hysical or oor mental health days in the receding 3 days did not change after Setember 11. Finally, resondents interviewed before Setember 11 and those surveyed after Setember 11 did not exhibit differences in regard to mean numbers of oor mental or hysical.health days. DISCUSSION Desite the otential benefits of receiving assistance in the wake of a traumatic event, 1 our findings are consistent with revious reorts 8 9 indicating that most eole affected by disaster do not receive hel in the aftermath. Our results are also consistent with Andersen'sl behavioral model of health care utilization, showing that redisosing factors and erceived need for services influence receit of hel. Three redisosing factors for seeking some tye of hel after Setember 11 were being a victim or a family member of a victim, being female, and being a younger adult These results mirror fndings indicating 1688 the otential effects of secondary traumatization21 or traumatic grief 2223 and suggest that not only direct exosure, but also secondary stress owing to a close relationshi to a victim, can be a otential indicator of need for hel in the wake of a disaster. Although we did not directly assess erceived need for services, 3 variables likely to be associated with such erceived need were assodated with' receit of some form of hel. Reorting 1 or more roblems, such as worry, nervousness, worthlessness, hoelessness, or lack of control over external events, was associated with receit of informal but not fornal hel. Increased rates of smoking or drinking and slee roblems were associated only with receit of formal suort services. Our results extend findings indicating that tobacco and alcohol use increased in New York City after Setember 115 by demonstrating that increases in use of these substances were also associated with receit of hel in a geograhically roximate oulation. Exosure to trauma and exosure to osttraumatic stress have been found to be associated with increased tobacco and alcohol use roblems among adolescents 2425 and adults However, it was the increase in use rather than smoking or alcohol consumtion er se that distinguished resondents who received hel: current smokers and drinkers were no more or less likely to receive hel than nonsmokers and nondrinkers. It is encouraging to learn that these otentially stress-related changes in substance use were associated with receit of formal hel. Screening eole for increases in tobacco use, alcohol consumtion, and slee roblems may be a better method of identifying individuals in need of formal hel in the wake of disaster than asking general questions regarding whether eole are exeriencing mental and behavioral roblems (reorts of which were unrelated to receit of formal hel in the resent multivariate analysis). We were not able to systematically assess enabling factors, and those measured (i.e., attending a funeral and ossessing health insurance coverage) had no effect on receit of hel. Other enabling factors may have been more relevant, or the magnitude of the event and relief efforts may have increased the relative effects of redisosing factors and erceived need or resulted in enabling resources being more widely available than usual. We found that rates of receit of formal and informal hel were similar but that redictors of these tyes of hel were different. We believe that our finrdings related to receit of formal hel have significant imlications for disaster relief systems. Providers sought for formal hel (rimarily medical or mental health rofessionals or religious counselors) included individuals who may have little or no training in disaster resonse yet often lay key roles in disaster relief. This finding underscores the imortance of rearing medical and mental health rofessionals and religious counselors to assist disaster victims and their families. In articular, our fimdings suggest that such roviders may need training to identify and assist affected individuals in obtaining services that will address roblems related to increased tobacco and alcohol use. Informal hel received from family and friends may be a critical source of social suort mitigating the imact of ostdisaster stress'4; however, this form of hel is difficult to define, and resondents who in fact received such hel may not have reorted it Setember 24, Vol 94,. 9 1 American Journal of Public Health Adams et at. I Peer Reviewed I Research and Practice

5 TABLE 2-Results of Multivariate Logistic Regression Analysis: Predictors of Receit of Hel, by Tye of Hel Received: Behavioral Risk Factor Surveillance System, 21. Gender Male Female Age Race/ethnicity White n-white Location on Setember 11, 21 New York Elsewhere Victim status association with victim Some relation to victim Self or family member victim Smoking status Current smoker t current smoker Alcohol use status Current drinker t current drinker Health insurance coverage Attended funeral. of oor hysical health days >. of oor mental health days >. of roblems reorteda 1 or more Slee roblems Increased substance useb Received alternate form of hel Adjusted Odds Ratio Formal Hel 95% Confidence Interval Predisosing factors.58, 2.78,1.3.42, , , 9.72** 1.88, 15.48**.51,3.2.87, 3.61 Enabling factors.32, , 5.39 Perceived need for services ,4.25,4.4.51, , 5.91* 1.11, 1.74* 2.55, 16.72*** Adjusted Odds Ratio Informal Hel 95% Confidence Interval 1.6,4.19*.96,.99**.39, , , , , , ,2.74.2, , , , 3.9*.32,1.8.61, , 16.74*** alncluding worry, nervousness, worthlessness, hoelessness, and lack of control over external events. btobacco or alcohol use, or both. *P<.5; **P<c.O1; ***P<.1. P values reresent the individual variables included in the model. In the case of each model, the overall P value for the Wald F test was <.1. For examle, resondents may have viewed informal hel not as "hel" but simly as art of the ordinary suort eole rovide to one another. Further research is needed to enhance our understanding of when, how, and froim whom eole seek and receive informal tyes of hel following disasters. That tobacco use, alcohol use, and oor mental and hysical health days did not increase in the 3 months after Setember 11 rovides some context for these results. Increases in substance use redicted receit of hel, but the increases observed were not sufficient, on a oulation basis, to affect survey results for smoking revalence or alcohol consumtion. Because recent changes in smoking and drinling are not nonnally measured on the BRFSS, we h,ave no way of knowing whether, on a regular basis, 5% of adults exhibit increases in smoking or drinking (as shown by the resent results) as art of fluctiating substance use atterns. This issue warrants flrther investigation. More than half of the resondents reorted nervousness, worry, feelings of worthlessness, or other stress-related roblems, yet there was no increase in the number of oor mental (or hysical) health days in the months immediately following Setember 11. We do not know the extent to which these reorted oor mental health days might reresent reexisting serious mental illness (which affects 5.4% of US adults 2 8 ) or how much of the hel received was art of ongoing treatment. It has been estimated that 15% of the US oulation receives mental health services each year, 2 2 while we found that 3.3% of our samle received formal hel in a 3-month eriod. Given the limits of our data, we were not able to determmie whether our figure reresents an increase in use of services above that exected or whether this figure rimarily reresents treatment of individuals with reexisting mental health roblems. The lack of an increase in oor mental health days in the months subsequent to Setember 11 suggests that many of the resondents who reorted mental health roblems (and receit of hel) may have had ongoing (erhas subclinical) difficulties. This finding is consistent with the results of other studies 8 9 indicating that individuals with reexisting mental health roblems are more likely 16 Research and Practice I Peer Reviewed I Adams et al. American Joumal of Public Health I Setember 24, Vol 94,. 9

6 than individuals without such roblems to seek hel after disasters. The findings that stress-related roblems such as worry, nervousness, and hoelessness were associated with receit of informal but not formal hel and that 2% of resondents who reorted no such roblems obtained formal hel suggest that other factors should be considered in ostdisaster screening. Exeriences of oor mental or hysical health days in the receding month were not significant redictors of receiving hel but may still warrant consideration as factors in early identification of individuals likely to need hel in the wake of a disaster. Although gender and ethnocultural background were not related to receit of formal hel in the first 2 to 3 months ostdisaster, these factors may lay a greater role in the develoment of osttraumatic stress and the need for hel at more distal time oints (e.g., 1 or more years ostdisaster). More detailed examination of the tyes of roblems associated with receit of different tyes of formal hel (e.g., medical, sychological, or crisis debriefing) over extended time eriods is warranted. Our fndings are subject to a number of limitations. The BRFSS excludes individuals without telehones and those who are unable or unwilling to articiate in a telehone survey, thus otentially excluding those most affected by the imact of a disaster. In addition, the survey questions did not address severity or duration of symtoms, so actual need for hel was not assessed. We did not distinguish between roblems associated with the attacks and other roblems that could have been ongoing or unrelated to Setember 11. Also, we did not address quality of hel received, and we had no way of measuring whether sufficient resources were available to meet serviceneeds. We also do not know whether or not some of the resondents needed hel but never sought it, whether others sought hel unsuccessfully, or whether those who received hel benefited from it. Moreover, it was imossible to isolate the effects of Setember 11 on the study oulation because of subsequent events that occurred during the time the survey was being conducted (e.g., the war on terrorism, deloyment of troos abroad, anthrax threats and deaths, and a lunging stock market). Patterns of hel seeking and changes in substance and tobacco use among eole affected by disasters have imortant imlications for the design and imlementation of disaster assistance resonse services. Findings such as those from this study should be considered in the develoment of service models and resonse systems for assisting communities in the aftermath of disasters. Public health rofessionals, hysicians, mental health roviders, and religious leaders should all be involved. Such lans also need to address ways to strengthen individual and community resilience and to resond to ostdisaster behavioral health needs, which change over time. Finally, results of rogram evaluations should be used to develo interventions that rovide the highest robability of success with the most efficient use of resources. 1 About the Authors At the time of this study, Mary L Adams was with the Connecticut Deartment of Public Health, Hartford. Julian D. Ford is with the Deartment of Psychiatry, University of Connecticut Health Center, Farmington. Wayne E Dailey is with the Connecticut Deartment ofmental Health and Addiction Services, Hartford, and the Deartment ofpsychiatry, Yale University School of Medicine. New Haven, Conn. Requestsfor rerints should be sent tojulian Ford, PhD, Deartment ofnpychiatry MC141, University of Connecticut Health Center, 263 Farmington Ave, Farmington, CT 63 ( ford@sychiatry.uchc.edu). This article was acceted vember 12, 23. Contibutors All of the authors were involved in the concetion of tlhe study, interretation of data, and dmfting and editing of the article. M.L. Adams was resonsible for data collection and analyzed the data. Acknowledgmnents This research was suorted in art by cooerative agreement from the Centers for Disease Control and Prevention (grant U58/CCU ) and by contracts under the Terrorism Related Disaster Relief Gmnts administered by the federal Centers for Substance Abuse Treatment (grant T1974), Mental Health Services (gmnt Tl 13 65), and Substance Abuse and Mental Health Services Administmrion (gmnt SM146). We thank Kevin Grifrith for comments on the article. Hluman Partciation Protection rotocol aroval was needed for this study. References 1. Galea S, Ahem J, Resnick H, et al. Psyclological sequelae of the Setember 11 terrorist attacks in New York City. NEnglJMed. 22;346: Melnick TA, Baker CT, Adams ML, et al. Psychological and emotional effects of the Setember 11 attacks on the World Tmde Center-Connecticut, New Jersey, and New York, 21. MMWR Morb Mortal Whly Re. 22;51: Schuster MA, Stein BD, Jaycox L, et al. A national survey of stress reactions after the Setember 11, 21 terrorist attacks. NEnglJ Med. 21;345: Fagan J, Galea S, Ahern J, et al. Self-reorted increase in asthma severity after the Setember 11 attacks on the World Trade Center-Manhattan, New York, 21. MMWRMorbMortal WilyRe. 22;51: Vlahov D, Galea S, Resnick H, et al. Increased use of cigarettes, alcohol, and marijuana among Manhattan, New York, residents after the Setember 11th terrorist attacks. AmJ Eidemiol. 22;155: Eustace K, MacDonald C, Long N. Cyclone Bola: a study of the sychological after-effects. Anxiety Stress Coing. 1999;12: Wang X, Gao L, Shinfuiku N, Zhang H, Zhao C, Shen Y. Longitudinal study of earthquake-related PTSD in a mndomly selected community samle in north China. AmJPsychiatry. 2;157: Wang X, Gao L, Zliang H, Zhao C, Shen Y, Shinfuku N. Post-earthquake quality of life and sychological well-being: longitudinal evaluation in a rural community samle in northem Ciina. Psychiatry Clin Nenrosci. 2;54: Caldera T, Palma L, Penayo U, Kullgren G. Psydcological imact of Hurricane Mitch in Nicaragua in a one-year ersective. Soc Psychiatry Psychiatr EidemioL 21;36: Andersen R. Revisiting the behavioral model and access to medical care: does it matter?j Health Soc Behav. 1995;36: Green B. Crossnational and ethnocultural issues in disaster. In: Marsella A, Friedman M, Gerrity E, Scur- Geld R, eds. Ethnocnltural Asects of Posttraumatic Stress Disorder: Issues, Research, and Clinical Alications. Washington, DC: American Psychological Assodation; 1996: rris F, Perrilla J, lbanez G, Murhy A. Sex differences in symtoms of osttmumatic stress disorder does culture lay a role?j Trauma Stress. 21;14: Perilla JL, rris FH, Lavizzo EA. Ethnicity, culture and disaster resonse: identifying and exlaining ethnic differences in PTSD six months after Hurricane Andrew. J Soc Clin Psychol. 22;21: Kaniasty K, rris F. In search of altruistic community: attems of sodal suort mobilization following Hurricane Hugo. AmJ Community Psychol. 1995; 23: Sattler DN, Preston A], Kaiser CF, Olivem VE, Valdez J, Scduster S. Hunicane Geoxges: a cross-national study examining' rearedness, resource loss, and sychological distress in the US Vurgin Islands, Puerto Rico, Dominican ReubEc, and the United States. J Trauma Stress. 21;15: Remington PL, Soith MY, Williamson DF, Anda RF, Gentry EM, HogeGn GC. Design, chamcteristics, and usefuiness of state-based behavioral risk factor surveillance: PublicHealth Re. 1988;13: Setember 24, Vol 94,. 9 American Journal of, Public Health Adams et al. I Peer Reviewed I Research and Practice 1 :161

7 RESEARCH AND PRACTICE I 17. Smith DW, Christiansen EH, Vincent R, Hann NE. Poulation effects of the bombing of Oklahoma City. J Okla State MledAssoc. 1999;92: Measriing Hlealtldy Days. Atlanta, Ga: Centers for Disease Control and Prevention; Arday DR, Tomar SL, Nelson DE, Merritt RI(l Schooley MW, Mowery P. State smoking revalence estimates: a comarison of the Behavioral Risk Factor Surveillance System and current oulation surveys. AmJPubliciHealth. 1997;87: Nelson DE, Holtzman D, Bolen J, Stanwick C. Reliability and validity of measures from the Behavioral Risk Factor Surveillance System (BRFSS). IntJPutblic Health. 21;46(sul 1): Bramsen I, van der Ploeg HM, Twisk JW. Secondary traumatization in Dutch coules of World War II survivors. J Consult Clin Psychol. 22;7: Dechant E, Jellinek M, Goodwin J, Prince JB. Processing acute traumatic grief: exacerbation of osttraumatic stress disorder after Setember 11 in a 9-year-old boy. Harv Rev Psychiatry. 22;1: Shear MY, Zuckoff A, Frank E. The syndrome of traumatic grief. CNS Sectrunos. 21;6: Acierno R, Kilatrick DG, Resnick HS, et al. Assault PTSD, farnily substance use, and deression as risk factors for cigarette use in youth: findings from the National Survey of Adolescents. J Trauma Stress. 21; 13: Kilatrick DG, Aciemo R, Saunders B, Resnick HS, Best CL, Schnurr PP. Risk factors for adolescent substance abuse and deendence: data from a national samle. J Consult Clin Psyciol. 2;68: Beckham JC. Smoking and anxiety in combat veterans with chronic osttraumatic stress disorder: a review.jpsychoactivedrugs. 1999;31: O den Velde W, Aarts PGH, Falger PRJ, et al. Alcohol use, cigarette consumtion and chronic osttraumatic stress disorder. Alcohol Alcohol. 22;37: Kessler RC, Berglund PA, Zhao S, et al. The 12- month revalence and correlates of serious mental illness (SMI). In: Manderscheid RW, Sonnenschein MA, eds. Mental Health, United States. Washington, DC: US Govemment Printing Office; 1996: Regier DA, Narrow WE, Rae DS, Manderscheid RW, Locke BZ, Goodwin FK. The de facto US mental and addictive disorders service system: Eidemiologic Catchment Area rosective 1-year revalence rates of disorders and services. Arch Gen Psychiatry. 1993;5: Second Edition ISBN Chronic Disease Eidemiology and Control Edited by Ross C. Brownson, PhD, Patrick Remington, MD, MPH, and James R. Davis With this book, you'll leam to: I Locate critical background information for develoing aroriate interventions I Enhance your technical caacity for delivering effective rograms I hmrove your knowledge about the methods used in chronic disease eidemiology I Identify diseases and risk factors I Examine the underlying biological or hysiological ages I softcover rocesses of disease $32. $32. APHA APo-members Members I Learn about high risk oulations, geograhlic var- $45. n-members Plus shiing and handling iatons, and trends I Plan, organize, and address revention and control methods American Public Health Association Publication Sales Web: APHA@TASCO1.com Tel: (31) FAX: (31) CHRN4JS Research and Practice I Peer Reviewed I Adams et ai. American Joumal of Public Health I Setember 24, Vol 94,. 9

8 COPYRIGHT INFORMATION TITLE: Predictors of Hel Seeking Among Connecticut Adults After Setember 11, 21 SOURCE: Am J Public Health 94 no9 S 24 WN: The magazine ublisher is the coyright holder of this article and it is reroduced with ermission. Further reroduction of this article in violation of the coyright is rohibited. To contact the ublisher: htt:// Coyright The H.W. Wilson Comany. All rights reserved.