Predicting Protective Factors of Physical and Mental Health for Survivors of Residential Fire
Symptoms of posttraumatic stress disorder (PTSD), depression, and physical health outcomes are three of the most common health outcomes evaluated for trauma survivors and several lines of empirical and meta-analytic research have demonstrated many risk factors for PTSD. Further, examining trauma survivors’ responses through a resilience orientation has grown increasingly popular over the past decade. However, the resilience orientation has little support among adult trauma survivor populations and none when evaluating physical health as part of an integrated health index (combining PTSD, depression, and physical health outcomes). Through examination of residential fire survivors, the current project evaluates the predictive validity of protective factors of PTSD as they relate to this integrated health index. Participants were assessed via self-report and semi-structured interviews approximately 4 months post-fire. Through evaluation of the integrated health index, peritraumatic emotionality and resource loss were found to significantly predict a resilient group of residential fire survivors 4 months post-fire. The present study suggests lower sustained resource loss and lower peritraumatic emotionality are significant protective factors for resiliency from residential fire.
KeywordsPTSD Depression Somatic health Peritraumatic emotionality Social support Resource loss
Current knowledge regarding the negative influence of traumatic events on physical and mental health is represented by a growing body of literature. Empirical and meta-analytic research supports significant effects of traumatic experiences and exposure severity on psychological disorders [1, 2] and physical health outcomes [3, 4, 5, 6, 7]. Further, high levels of comorbidity between posttraumatic stress disorder (PTSD), depression, and physical health problems are evident [8, 9]. Altogether, empirical findings suggest PTSD is an important predictor of health outcomes among trauma survivors. Less is known about mechanisms of such relationships, particularly protective factors for adult trauma survivors (i.e. whether protective factors for PTSD predict overall health following a residential fire). Given the nascent state of the resiliency literature for both mental and physical health outcomes of adults following a traumatic event, further investigation is warranted.
1 The Risk/Resiliency Paradigm
In the psychosocial trauma literature, a variety of theoretical models of PTSD have examined risk factors to further enhance the understanding of the disorder. Theories of PTSD utilizing risk factors include (but are not limited to) emotional processing theory , Elhers and Clark’s  cognitive model of PTSD, the diathesis-stress model , and the dose–response model . The understanding of risk factors through the lens of such theories has contributed to an emphasis placed on protective factors as necessary for prevention and treatment of psychopathology.
Protective factors have important implications for understanding health related vulnerability and resilience (both psychological and physical) in post-trauma recovery contexts. Bonanno  defines resiliency as “the ability of adults in otherwise normal circumstances who are exposed to an isolated and potentially highly disruptive event such as the death of a close relation or a violent life-threatening situation to maintain relatively stable, healthy levels of psychological and physical functioning (p. 20).” Therefore, it is important to continue to identify factors that protect one from post-trauma problems or disorders, as well as to extend the research to identify potential preventative factors pertaining to self-reported health. Such empirical identification and extension of the literature will yield a more accurate and encompassing understanding of resiliency.
2 Protective Factors are Influential in Fire Survivors’ Health
Posttraumatic Stress Disorder is a psychopathology characterized by a strong emotional reaction to a traumatic event along with symptoms consisting of three symptom clusters (re-experiencing, avoidance, and increased arousal; APA 2000). Following trauma exposure, survivors often overreact to subsequent stressors, which facilitates a cycle of vulnerability to hyperarousal .
Posttraumatic stress disorder has been shown to significantly and negatively impact trauma survivors’ health [16, 17] and is the most commonly endorsed disorder following traumatic events . Even when severity of exposure to traumatic events is controlled for, the effects of PTSD on somatic symptoms remain pronounced . Posttraumatic stress disorder has been found to mediate the relationship between exposure to trauma and self-reported health outcomes in combat veterans [20, 21] and to partially mediate the same relationship in a group of individuals exposed to toxic gases . Additionally, depressive symptoms may have a significant impact on trauma survivors’ health: Given the high rates of comorbidity between PTSD and depression, reported as high as 50% , depressive symptomatology likely plays a significant role in trauma survivors’ health .
Given the robust literature of PTSD, the impact of PTSD on physical health, and the relationship between PTSD and depression, it may be useful to identify protective factors of PTSD and test their effectiveness in predicting resiliency in residential fire survivors. By comparing and contrasting these variables, researchers may gain a clearer understanding of mechanistic and facilitating protective factors and variables that cause the link between PTSD, depression, and self-reported health outcomes.
3 Protective Factors Against Psychological and Physical Health Problems
A growing body of research provides preliminary evidence for factors associated with protection from PTSD development. A variety of factors have been identified as intervening variables of resiliency in child abuse, including social support and a positive family environment, financial resources, and access to higher education . Alim et al.  evaluated protective factors in adults exposed to a variety of traumas and conceptualized resiliency as an individual having no current psychological diagnosis (i.e., absence of symptoms). Results of the project found members of the resilient group were largely male, experienced lower levels of trauma exposure severity, and reported having a purpose in life. In another context (characterized by ongoing threat of terrorist attacks in the Middle East), Hobfoll et al.  evaluated factors associated with PTSD, depression, and resiliency. Results indicated significantly fewer individuals were considered resilient (22%) as compared to those considered chronically distressed (54%). Additionally, the authors reported psychosocial resource loss, socioeconomic status, perceived social support, majority ethnic status (Jewish) as predictors of resiliency.
Resiliency research conducted in the wake of the terrorist attacks that occurred in the United States on September 11, 2001, examined psychological outcomes of survivors of the attacks. Specifically, this research assessed New York City residents 6 months after the September 11th, 2001 terrorist attacks , and highlighted results indicating that 65% of the sample of residents met criteria for being resilient (as defined by having one or no symptoms of PTSD). Lower levels of resiliency were found among those who experienced high trauma exposure severity. The current study was conceptualized from this foundational research. Further research done in the same context extended these findings to assess inverse risk factors of PTSD as protective factors (e.g., if females are more at risk for developing symptoms, protective factor ). The study found ethnicity (Caucasian, African American, and Hispanic), lower level of trauma exposure, no reduction in income, high social support, and a lack of chronic disease, being of the male gender, and being of an age below 65 years were all protective factors. Given the novelty of assessing protective factors of health outcomes posttrauma, similar logic (identifying risk factors and assessing the inverse for resiliency) is utilized in the current study.
The resistance and resiliency literature on self-reported health outcomes of trauma survivors who have sustained traumatic exposure(s) is minimal. One study examined hospitalized patients during the severe acute respiratory syndrome (SARS) outbreak in Hong Kong . The latent class analysis used in this study identified four groups of participants, based on psychological distress: chronic dysfunctional, delayed dysfunction, recovery, and resilience.
Results identified three groups had better self-reported health than the chronic dysfunctional group. Social support, less health-related worry, and male gender were factors of the resilient and recovered groups. Finally, the resilient group had more social support than the delayed dysfunctional group and significantly better self-reported health than the recovery group. A review of the literature yielded no results for resiliency involving resource loss or peritraumatic emotionality’s influence on self-reported health for residential fire survivors.
4 Rationale for the Project and Hypotheses
The literature regarding risk and protective factors for PTSD is well established. A number of pre-existing attributes, characteristics of the trauma, and posttrauma factors have been identified as either risk or protective attributes for trauma survivors. Additionally, a sound theoretical model  has been established regarding trauma survivors’ health. However, no research to date has investigated characteristics of the trauma and posttrauma protective factors for adult trauma survivors’ health in a sample of residential fire survivors.
The current project investigated protective factors of residential fire survivor’s mental and physical health following the fire. In order to assess a comprehensive definition of adult resiliency , peritraumatic emotionality, resource loss, and social support were used to predict a resilient (or minimally symptomatic) group of trauma survivors (using a comprehensive health construct) among residential fire survivors. No study to date has examined these factors in the context of fire survival.
5.1.1 Brief Symptom Inventory (BSI)
The BSI  is a commonly used self-report measure of psychological constructs that allows participants to self-report distress levels on nine distinct factors. For the current study, the somatization scale was utilized as a somatic health complaint (i.e. stomach ache, chest pains, etc.) assessment, and the depression subscales was used as a proxy for depressive symptomatology. For the logistic regression analysis, symptoms were rated as either absent (0 to 1 symptom) or present (if between 2 and 4 symptoms). Adequate reliability for these subscales (somatization α = .83; depression α = .88) is demonstrated with the current sample.
5.1.2 Anxiety Disorders Interview Schedule-IV: Lifetime Version (ADIS)
The ADIS  is a semi-structured clinical interview designed to assess a variety of psychological disorders. For the current study, both PTSD and health history modules of the ADIS were used to assess PTSD symptomatology and prior physical health problems. The continuous distress rating of PTSD symptomatology of the ADIS PTSD module was used to assess PTSD symptomatology. For the logistic regression analysis, posttraumatic stress symptoms were considered quantitatively as either absent (rating from 0 to 3) or present (rating 4+). Internal consistency for the PTSD item was good (α = .92).
5.1.3 Fire Questionnaire (FQ)
The FQ  is a semi-structured interview designed for the residential fire project. The FQ assesses 16 factors and provides an assessment for retrospectively reported peritraumatic emotionality for the current study . Participant ratings of feelings of helplessness, hopelessness, and fear were averaged to create the peritraumatic emotionality variable. The FQ demonstrated adequate reliability (α = .73)
5.1.4 Social Support Questionnaire (SSQ)
The SSQ  is a self-report instrument that assesses subjective level of social support. The SSQ asks participants to report both the number of individuals who provide social support and perceived quality of support; for the purposes of this study a total social support coefficient is utilized in analyses. The SSQ demonstrated excellent reliability (α = .97).
5.1.5 Resource Loss Scale (RLS)
The RLS  is a 53 item self-report measure of loss sustained during and following a traumatic event (i.e. objects, energy, personal characteristics, etc.). For the current study, all items were summed to comprise the total loss factor utilized in analyses. Excellent reliability was found with this scale (α = .95; ).
5.2 Creating an Integrated Health Index and Resiliency Dichotomy
Utilizing self-reported PTSD, depressive, and somatic symptoms, a dichotomous whole health index was created in line with Bonanno’s  formulation of resiliency. Consistent with grouping systems from previous work , participants were grouped into either a resilient group (having one or fewer endorsements in each of the three pathology categories; n = 9, 20.5% of the sample) or symptomatic (having greater than one endorsement in one or more of the pathology categories; n = 35, 79.54% of the sample). This ratio of division of groups was found to be similar to some previous results of resilient groups when considering PTSD and depression as the primary outcome , but lower than others .
5.3 Participants and Procedure
Data for the project came from the National Institute of Mental Health sponsored residential fire project NIMH 5RO1 MH49147-03 (for further details, please see ) that assessed survivors of residential fire for years following their sustained incident. A sample of 44 adults, ages 24 to 79 (mean = 38.55, SD = 10.35) met inclusion criteria for this study. Thirty-five of the participants were women, 23 were Caucasian, 19 African-American, 1 Hispanic. Participants were recruited from five locations in the southeast region of the United States, and were largely accessed via live interviews. To be included in the original study, families must have sustained loss of 15% or greater of their home and/or personal belongings during said house fire. Data from the 4 month post-fire assessment point was used for the current project. Missing data was addressed via a mean imputation procedure where necessary for the social support, resource loss, and PTSD variables.
6.1 Control Variables and Functioning Level for the Current Sample
In addition to exploration of demographic control variables (i.e., age, gender, education, and ethnicity), participant health histories were examined as a way to further understand the comprehensive posttrauma health of fire survivors. Assessment of health history required participants to either confirm or deny current and previous health problems, including: diabetes, heart problems, high/low blood pressure, cancer, thyroid disease, other hormonal problems, asthma, respiratory problems, migraines, stroke, gastrointestinal problems, and blood diseases. Notably, participants with a history of respiratory problems reported significantly higher depression symptomology (t = −2.31, p < .05) and somatic symptomology (t = 2.23, p < .05). Further, a significant difference between groups was found for age (t (42) = 2.71, p < .01) amongst the resilient group. No other health problems or demographic variables were found to significantly differ between the two groups.
Measures Range, Means, and Standard Deviations
Fire questionnaire: peritraumatic emotionality, home during fire, and perceived control
N = 44
0 to 3.0
N = 44
0 to 3.0
N = 44
0 to 3.0
Peritraumatic emotionality (mean)
N = 44
0 to 3.0
Resource loss scale: total resource loss
N = 44
7 to 149
Social support questionnaire: social support
N = 44
0 to 53.00
Anxiety disorder interview schedule: posttraumatic stress disorder (averaged across symptoms)
N = 44
0 to 6.47
Brief symptom inventory: somatic and depression symptoms (averaged across symptoms)
N = 44
0 to 3.0
N = 44
0 to 3.17
6.2 Primary Analyses
The above reported outcome groups (resilient vs. symptomatic) were analyzed in each set of the following logistic regressions to understand how the suggested protective factors both individually, and simultaneously, predicted the integrated health index. In the first three sets of individual logistic regressions the integrated health index was regressed onto peritraumatic emotionality, social support (i.e., satisfaction and quantity), and resource loss, respectively. In the first logistic regression, results yielded peritraumatic emotionality as a significant influencer of the integrated health index (Wald statistic (1) = 5.143, p < .05), which predicted 33.3% of resilient participants and 91.4% of symptomatic cases. In the second logistic regression, social support was not significantly related to the integrated health index (Wald statistic (1) = 2.48, p > .05). In the third regression, results showed resource loss significantly predicted the integrated health index (Wald statistic (1) = 5.92, p < .05), predicting 91.4% of symptomatic participants and 44.4% of resilient participants.
In the final set of logistic regressions comprehensive health was regressed onto the significant predictors simultaneously (i.e., resource loss and peritraumatic emotionality). In this regression, resource loss (Wald statistic (1) = 5.49, p < .05) significantly predicted the integrated health index. The influence of peritraumatic emotionality, however, was non-significant, though approaching an alpha level of .05 (Wald statistic (1) = 3.85, p = .05). In this final regression, a total of 84.1% of participants were predicted by the model; 91.4% of symptomatic cases were predicted, and 55.6% of resilient participants were predicted.
The study made a first attempt at evaluating if low peritraumatic emotionality, social support, and low resource loss would serve as protective factors for fire survivors via exploration of an integrated health index. Whereas social support was not a significant predictor, resource loss and peritraumatic emotionality each individually predicted the mental and somatic health construct for both groups (i.e., resilient vs. symptomatic responders). Further, when the independent variables were entered into a model simultaneously, resource loss maintained a significant influence on the integrated health index, predicting 55.1% of membership in the resiliency group and 91.4% of symptomatic group membership (peritraumatic emotionality did not predict outcomes in the simultaneous regression model). These results further support low resource loss as a protective factor for residential fire survivors.
The non-significant relationship between social support and psychological/somatic health may have resulted from two issues. First, social support may not have had sufficient time to influence the outcome; Ozer et al.  suggest that strongest findings related to social support and PTSD occur when assessment is conducted 3 years post-trauma. Second, social support was operationalized as both quantity and perceived satisfaction with one’s social network. Some lines of research suggest in post-trauma contexts, distinctions between the dynamic relationships among received social support, perceived social support, and social networks [38, 39] should be made. Thus, future research may focus on examining dynamic social support processes conducted via longitudinal design.
This study provides further empirical evidence to the field of adult trauma resiliency research. The first advance lies in the first attempt at identification and operationalization of an integrative whole health construct via merging PTSD, depression, and somatic health symptomatology. Additionally, with proper health history screening (something not previously done in the trauma literature), this study represents a more comprehensive way to understand trauma survivor health outcomes. Further, despite the relatively low frequency of pre-trauma health problems, controlling for pre-trauma health provides increased confidence in the notion that somatic health complaints reported are associated with trauma exposure. Finally, an additional advance posed by this study relates to the novelty of the trauma population in which this model was examined (i.e. identification of protective factors for adult survivors of residential fire).
The results of this study have notable implications for early intervention and assessment. Mental health clinicians may be able to use assessment of peritraumatic emotionality and resource loss as early screening tools to detect those who are perhaps more likely to be resilient following a trauma event. These protective factors can be evaluated prior to the recommended wait times for assessment of acute stress disorder .
The cross-sectional design of this study limits the ability to make causal inference of the risk factors’ impact on both groups (given data of the sample was from one assessment point post-fire). Studies designed longitudinally that employ more sophisticated analytical techniques (e.g., hierarchical linear modeling) may increase understanding of dynamic post-trauma processes. Additionally, the relatively small sample size limits generalizability and reduces power. Such power reduction may have influenced the fact that social support was not significantly related to outcomes, as has been suggested by previous findings [41, 42]. However, considering the limited sample size and finding of peritraumatic emotionality’s near significant prediction of the groups, during the simultaneous regression a larger and more powerful sample may allow for detection of peritraumatic emotionality’s effect on residential fire survivors’ outcomes. Finally, the current project did not assess pre-trauma functioning as a holistic construct, as suggested by Bonanno  (only physical health). Future research may benefit from larger samples that integrate more comprehensive pre-trauma circumstances and functioning assessments (e.g., pre-trauma resources, depression, anxiety, etc.).
In conclusion, the current study adds further support to the growing body of literature evaluating protective factors for adult residential fire survivors. By evaluating residential fire survivors 4 months posttrauma, the study is able to evaluate protective factors of PTSD, depression, and somatic health (as an integrated health construct). Additionally, this study extends previous findings indicating peritraumatic emotionality and resource losses are important predictors of both psychological disorders and physical health complaints posttrauma. The findings also extend the notion that physical and mental health can be evaluated as a unified construct. Future research can build on findings and further evaluate additional protective factors of survivors’ physical and mental health following residential fire.
- 3.Green BL, Kimerling R (2004) Trauma, posttraumatic stress disorder, and health status. In: Schnurr P, Green B (eds) Trauma and health: physical health consequences of exposure to extreme stress. American Psychological Association, Washington, DC, p 13–42Google Scholar
- 6.Schnurr P, Green B (2004) Trauma and health: physical health consequences of exposure to extreme stress. American Psychological Association, Washington, DCGoogle Scholar
- 16.Schnurr P, Jankowski MK (1999) Physical health and post-traumatic stress disorder: review and synthesis. Semin Clin Neuropsychiatry 4(4):295–304Google Scholar
- 18.Shalev AY, Freedman S, Peri T, Brandes D, Sahar T, Orr SP, Pitman RK (1998) Prospective study in posttraumatic stress disorder and depression following trauma. Am J Psychiatry 155:630–637Google Scholar
- 30.Schnurr P, Green B (2004) Understanding relationships among trauma, posttraumatic stress disorder, and health outcomes. In: Schnurr P, Green B (eds) Trauma and health: physical health consequences of exposure to extreme stress. American Psychological Association, Washington, DC, p 247–275Google Scholar
- 31.Derogatis LR (1993) Brief symptom inventory (BSI) administration, scoring, and procedures manual, 3rd edn. National Computer Systems, MinneapolisGoogle Scholar
- 32.Di Nardo PA, Brown TA, Barlow DH (1994) Arvciety disorders interview schedule for DSM-IV: lifetime version (ADIS-1V-L). Psychological Corporation, San AntonioGoogle Scholar
- 40.Litz BT, Gray MJ (2004) Early intervention for trauma in adults: a framework for first aid and secondary prevention. In: Litz B (ed) Early intervention for trauma and traumatic loss. Guildford Press, New York, p 87–111Google Scholar
- 42.Vogt D, King D, King L (2007) Risk pathways for PTSD. In: Friedman MJ, Keane MT, Resick PA (eds) Handbook of PTSD: science and practice. The Guilford Press, New York, p 99–115Google Scholar