Meltzer-Brody, Samantha MD, MPH*; Hartmann, Katherine MD, PhD†§; Miller, William C. MD, PhD, MPH‡§; Scott, Jewel*; Garrett, Joanne PhD‡; Davidson, Jonathan MD¶
Posttraumatic stress disorder (PTSD) is a debilitating form of an anxiety disorder triggered by personal experience of serious trauma (eg, sexual abuse or assault, victim of violent crime or severe motor vehicle accident). The course of untreated disease averages more than 5 years and may lead to additional psychiatric comorbidity, including a greatly increased risk of major depression and suicide.1 Early diagnosis of PTSD is critical, because prompt initiation of effective treatment may improve the long-term outcome of PTSD and prevent needless suffering.2
Population-based studies suggest that exposure to traumatic events is common and ranges from 30–69%.3–5 Approximately 1 in 4 individuals who experience trauma will develop PTSD,6 and women are twice as likely to develop PTSD after traumatic events compared with men.7 This finding is partially explained by sexual assault or rape having consistently been shown to cause the highest rates of PTSD.4,5,7–9 Most women will not receive psychiatric evaluation or treatment, although the vast majority will receive health care during the period in which they have symptomatic PTSD. Women who are victims of rape or sexual assault are twice as likely to seek medical services than nonvictims10,11 and often seek medical services much more often than psychological services in the year after being assaulted (72.6% compared with 19%).12 Thus, because of the high prevalence rates of trauma and PTSD in women, accurate screening and prompt diagnosis of women presenting for gynecologic primary care is important, so that treatment can be initiated.
Formal evaluation for PTSD requires a skilled assessment by a psychiatrist or psychologist. The standard diagnostic procedure is a structured clinical interview based on Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) diagnostic criteria.13 Diagnosis of PTSD requires history of a traumatic event and the individual's intense emotional reaction to it. The DSM-IV definition of trauma requires 2 elements: 1) experiencing, witnessing, or learning of an event that involved actual or threatened death, serious injury, or threat of harm to self or others and 2) having the trauma followed by intense feelings of fear, helplessness, or horror that are associated with persistent reexperiencing of the event (Box).
Diagnostic and Statistical Manual of Mental Disorders, 4th edition
Diagnostic Criteria for Posttraumatic Stress Disorder
* Reexperiencing symptoms: Nightmares and flashbacks; intrusive thoughts
* Avoidance symptoms: Avoiding thoughts or places associated with the trauma
* Emotional numbing symptoms: Feeling unable to have sad or loving feelings
* Hyperarousal symptoms: Sleep and concentration difficulties; hypervigilance
Despite the high prevalence of trauma and PTSD, recent studies have documented that posttraumatic stress disorder is often underdiagnosed or misdiagnosed in primary care settings,14,15 including those that treat pregnant women seeking routine prenatal care.16 Many valid reasons may contribute to a missed diagnosis of PTSD, including 1) lack of familiarity with PTSD leading to discomfort in screening for and diagnosing the disorder, 2) a high degree of overlap between PTSD and other anxiety disorders or depression, 3) initial presentation to primary care providers for specific physical complaints and inability to recognize the connection between past trauma and present somatic symptoms,17,18 and 4) time-consuming formal assessment of PTSD DSM-IV criteria.
For these reasons, it is not likely to be feasible or efficient for providers in primary care fields to make the diagnosis of PTSD, or alternatively, to simply refer all women with a history of trauma for evaluation. Thus, a briefer scale to assess PTSD symptoms could have great clinical usefulness as a triage tool. The SPAN, (acronym for symptoms of startle, physiological arousal, anger, and emotional numbness; Multi-Health Systems Inc., North Tonawanda, NY) is a brief 4-item, patient rated, written instrument for PTSD that we developed in an outpatient psychiatry clinic for case finding.19 Although the SPAN was developed for use as a screening instrument in high-risk psychiatric populations where the prevalence rate of PTSD was approximately 50%, we felt that it showed promise as an instrument to help triage patients with a history of trauma in a primary care setting, where the presentation of PTSD may be less severe and prevalence rates are lower. The estimated prevalence of PTSD in primary care settings is 12%.20 Additionally, in primary care settings PTSD can be a persistent illness, and many patients who have recovered from full-blown PTSD continue to suffer from subthreshold symptoms of PTSD.21
The prevalence of trauma and PTSD symptoms in women presenting to a gynecology clinic for routine care has been rarely studied. Therefore, we sought to evaluate the prevalence of trauma among women seeking primary care in a gynecology setting and to evaluate a potential tool to refine the process of referral for further psychiatric evaluation among women with trauma. Specifically, our aims were 1) to document the prevalence of a history of serious trauma among women seeking routine annual gynecologic care in an academic continuity care setting and 2) to evaluate a brief 4-item instrument assessing PTSD symptoms (the SPAN) as a potential tool for triaging women with a history of trauma for further evaluation.
SUBJECTS AND METHODS
Study participants (N = 292) were all women who presented to the University of North Carolina outpatient gynecology clinic for a routine annual examination between June 2001 and March 2002. This clinic is a gynecologic continuity clinic, located in an academic medical center that provides care to a diverse population of women, including a large uninsured and Medicaid patient population. Only patients presenting for a routine annual care examination were eligible for our study.
This project was reviewed and approved by the Institutional Review Board at the University of North Carolina at Chapel Hill, and study participants gave informed consent before taking part in this study. Participants completed a 1-time voluntary survey called the “UNC Women's Health Survey,” which contained questions about medical history, traumatic experiences, and overall health status. Participants were asked to indicate whether they had a history of a traumatic event and to describe the trauma. Specifically, patients were asked: “Have you had a severe trauma(s) or stress in which you thought you or another person might die or experience serious harm? Examples include: physical attack, mugging, rape, severe automobile accidents, natural or manmade disasters, being diagnosed with a life threatening illness or sexual abuse.” Patients who identified more than 1 trauma were asked to identify the type of trauma that was most disturbing to them. Those women who reported a traumatic event were also asked to complete the SPAN questions. Patients who endorsed a history of trauma on the survey were eligible for a psychiatric assessment. This project was open to English-speaking patients only, and there were no explicit exclusion criteria stated.
The SPAN is derived from the Davidson Trauma Scale,22 a validated, 17-item self-rating scale that reflects the symptoms diagnostic of PTSD as defined in DSM-IV.13 Each item on the SPAN is scored on a 0–4 scale from being “not at all distressing” to “extremely distressing” and can be administered in approximately 1 minute. The rating of each item is summed for the total SPAN score. From our previous work, a score of 5 or greater is considered a positive score. This cutoff score of 5 or more was recently validated in a study of Chinese earthquake survivors.23 We used a cutoff of 5 for our primary analyses, but also evaluated the performance across a range of possible cutoffs.
It was found that the SPAN closely corresponded to the diagnosis of PTSD by structured clinical interview. The outpatient psychiatry patient population had a prevalence rate of PTSD of 50%. In the psychiatry clinic in which the SPAN was developed, it performed with a sensitivity of 84% and a specificity of 91%. The SPAN includes a trauma screening question that states, “Have you had a severe trauma or stressor in which you thought you or another person might die or experience serious harm?” After the trauma screening question, the 4-item SPAN evaluates how much distress the patient has experienced during the past week in regard to symptoms of startle, physiological arousal at reminders of the trauma, anger, and numbness.19
Patients who reported a history of trauma on the survey were eligible for the psychiatric assessment. Of the 292 patients who completed surveys, 88 patients reported a history of a traumatic event and were eligible for further study participation. Of the 88 patients, 32 of 88 (36%) agreed and could be found to participate in the full psychiatric assessment.
This 60-minute, face-to-face assessment included The Mini-International Neuropsychiatric Interview24 as the standard structured clinical interview and was conducted by a psychiatrist who was blinded to the SPAN scores from the survey. The Mini-International Neuropsychiatric Interview is a short structured diagnostic interview, jointly developed by psychiatrists in the United States and in Europe for both DSM-IV and ICD-10 psychiatric disorders. It was designed to meet the need for a short but accurate structured clinical interview (it may be administered in approximately 15 minutes), while having very good agreement with the significantly much longer Structured Clinical Interview for DSM.24,25 The Mini-International Neuropsychiatric Interview includes a psychiatric assessment of current diagnoses of PTSD, panic disorder, generalized anxiety disorder, social phobia, major depression, psychotic disorders, and substance abuse or dependence. These comorbid psychiatric diagnoses were selected because of the frequent co-occurrence of PTSD with other anxiety disorders, major depression, and substance use.20,21 The assessment of psychotic symptoms was included for the sake of completeness.
All data were entered into a computerized database and analyzed using Stata 6.0 software (StataCorp., College Station, TX). Exploratory univariate analysis of demographic variables was assessed. We compared the characteristics of persons who completed structured interviews and persons who did not complete the structured interviews using t tests for continuous variables and the Fisher exact test for categorical variables. To assess the diagnostic performance of the SPAN compared with a structured clinical interview for a diagnosis of PTSD, we calculated the sensitivity, specificity, and positive and negative likelihood ratios. We also calculated 95% confidence intervals for these measures.26,27 Likelihood ratios reflect the change in odds of disease for a given test result. A likelihood ratio with a value of 1 provides no useful information. To assess the overall performance of the SPAN, we calculated the area under the receiver operating characteristic curve. To assess the potential usefulness of this instrument, we evaluated the posttest probability (positive predictive value or 1 minus negative predictive value) across a range of prevalences. In all evaluations, a diagnosis of PTSD on the structured clinical interview served as the reference standard.
Of the 292 participants who completed surveys in our study, the mean age was 34 (± 12) years, 43% were white, 49% were African American, and 46% were single. Eighty-eight patients who completed a survey reported a history of a traumatic event. Types of reported traumatic events included sexual abuse or rape, physical abuse, motor vehicle accidents, being diagnosed with a life-threatening illness, sudden unexpected death or diagnosis with a life-threatening illness of a family member, victim of natural disaster, and victim of war.
Among the 88 patients who reported a history of trauma, 32 completed a psychiatric assessment and 56 did not. These 2 groups did not differ significantly by age, marital status, race, or percentage with a positive SPAN score (Table 1). Of the 56 patients who did not complete a psychiatric assessment, the vast majority reported being unable to return to the clinic or lack of available time as the main reasons for refusing participation.
Of the 32 patients completing the psychiatric assessment, 25 (78%) patients met the criteria for a diagnosis of PTSD based on the structured clinical interview. Expressed in other terms, the single question regarding a history of trauma had a positive predictive value of 78% (95% confidence interval [CI] 0.60–0.91) in our patient population. Nearly all of these patients (n = 22 [88%]) with a diagnosis of PTSD were not receiving any kind of psychiatric treatment for PTSD or any other psychiatric disorder. Additionally, among the 32 patients interviewed, 31 met criteria for 1 or more DSM-IV psychiatric diagnoses, including major depression (62%), panic disorder (28%), generalized anxiety disorder (19%), social phobia (28%), and substance abuse disorders (15%).
Considering the full range of the SPAN, the instrument performed adequately (receiver operating characteristic curve area 0.75; 95% CI 0.57–0.94). Using the standard cutoff of 5 or greater, the SPAN performed with a sensitivity of 72% (95% CI 51–98) and specificity of 71% (95% CI 29–96), corresponding to a positive likelihood ratio of 2.52 (95% CI 0.76–8.34) and negative likelihood ratio of 0.39 (95% CI 0.18–0.86) (Table 2). Given our population prevalence of 78% among those women undergoing the structured clinical interview, a positive predictive value of 90%, and a negative predictive value of 42%, we were able to calculate posttest probabilities using a cutoff of 5 or greater (Fig. 1).
Of the alternative cutoffs considered, only a cutoff of 6 or greater provided acceptable performance characteristics. Using a cutoff of 6 or greater, the sensitivity was decreased, but the specificity increased, leading to an increase in the positive likelihood ratio and decrease in the negative likelihood ratio (Table 3).
In our population, the SPAN (cutoff 5 or greater) would have missed 7 patients who were diagnosed with PTSD by interview and would have resulted in the additional referral of 2 patients who did not have PTSD by interview. This 4-item scale evaluates startle, physiological arousal at reminders of the trauma, anger, and numbness. Three of these symptoms (startle, physiological arousal, and numbness) are necessary and specific for a diagnosis of PTSD. However, patients with other anxiety disorders may demonstrate significant anger and irritability or heightened physiological arousal symptoms and could be detected by a positive SPAN score. However, of note in our study, the 2 patients who had “false positive” results on the SPAN had another anxiety disorder (panic or generalized anxiety disorder) as well as comorbid major depression.
Approximately 1 in 3 women presenting for routine annual care in our academic continuity gynecology clinic have a history of a traumatic event. The prevalence of trauma we observed is comparable to national samples; however, the proportion of those with trauma who have PTSD is much higher than might have been anticipated, perhaps as a result of bias in those women who elected to participate in the study.
Our goal was to develop a quick, efficient, and low-cost tool to determine which patients in a primary care setting require further evaluation. Screening for PTSD is a 2-step process: 1) determining whether the patient has experienced a trauma and 2) assessment of PTSD symptoms. We believe that the SPAN may offer clinicians a useful way of triaging those patients who are in need of psychiatric referral. For example, after asking the relatively rough screening question of “have you ever had a traumatic experience,” this 4- item case-finding instrument may be a valuable tool for the busy primary care clinician. Additionally, we feel that the SPAN offers advantages over other screening instruments for PTSD. For example, Breslau et al28 developed a 7-item screening scale for PTSD that differs from the SPAN in the following ways: First, it is an interview measure and is not self-rated. Second, the Breslau scale uses a yes or no response to symptoms, whereas the SPAN uses a Likert scale that measures symptom severity on a 5-point scale. Finally, the Breslau scale measures lifetime prevalence of PTSD, whereas the SPAN measures current PTSD symptomatology.
In our study a relatively small number of women were willing to be contacted and completed the psychiatric interview. This small sample size limited the precision of our estimates, as reflected in the wide confidence intervals. Furthermore, there is likely sampling bias from nonresponse to the trauma items on the survey and from those who reported trauma but were unwilling or unavailable to return for a psychiatric interview. In addition, the survey used an open-ended question format to ask patients about a history of trauma. We can speculate that asking patients to specifically check off each individual type of trauma sustained may have yielded an increased number of women endorsing a trauma history.
Given the limitations, it is difficult to assert with confidence whether over or underestimating prevalence of trauma and proportion of women with PTSD occurred. Nonetheless, it is still possible to compare potential strategies for identifying women with PTSD. As an example, consider a practice that provides routine annual care for 1,000 women each year. In our population 30% of women had a history of a traumatic event. This is represented in the middle column in Table 3 that summarizes the potential usefulness of the SPAN for triage across a range of prevalence of trauma. If all women with a history of trauma were referred for formal psychiatric evaluation, 300 referrals would be made, compared with 165 per 1,000 by triaging those with a trauma history using the SPAN score to determine referral. Reduction in referrals is important because the current structure of the mental health system is not equipped to accommodate referrals for all women with trauma histories. However, most importantly, such triage would increase the number of women with PTSD who are diagnosed by 2.4- to 2.7-fold, a substantial improvement over the most likely strategy, which is either no or infrequent use of screening or referral.
Currently, routine evaluation for a history of trauma and PTSD often does not occur in primary care settings or in obstetric or gynecologic settings. Many women with histories of trauma and PTSD are not receiving appropriate care for their psychiatric symptoms. Busy primary care physicians need efficient and easy-to-use instruments to help them triage patients who require further evaluation. Furthermore, multiple studies have demonstrated that obstetrician–gynecologists provide more office-based general medical evaluations for women of reproductive age than either family practice physicians or internists29 and that increasingly, obstetrician–gynecologists have expanded training and expertise in overall women's primary health care.30
The SPAN is a readily useable tool, which can enable primary care providers to have a mechanism for triage, greatly improving the numbers of patients with PTSD detected compared with the status quo. Prompt initiation of treatment does improve the long-term outcome of PTSD and prevents needless and ongoing suffering. Current treatment strategies for PTSD include psychotherapy (cognitive-behavioral therapy, exposure therapies, dynamic psychotherapies) and psychopharmacology (including selective serotonin reuptake inhibitors, tricyclic antidepressants, anticonvulsants, and atypical antipsychotics).31–34 However, only the selective serotonin reuptake inhibitors have been proven effective and safe in long-term randomized controlled trials.31
Thus, the findings of this study suggest that asking women about a history of a traumatic event is extremely important, because this question alone is quite predictive in detecting psychiatric illness that is often untreated. The SPAN seems to be a valuable instrument for triaging patients in need of more formal psychiatric evaluation in primary care settings, although further refinement of the SPAN could help to improve its performance in a women's health setting. The next step is to clarify and further improve this tool, which is likely best accomplished by adding an additional question or questions about PTSD symptomatology that would capture a more gender-specific presentation of PTSD in women.
1. Davidson JR, Hughes D, Blazer DG, George LK. Posttraumatic stress disorder in the community: an epidemiological study. Psychol Med 1991;21:713–21.
2. Davidson JR. Biological therapies for posttraumatic stress disorder: an overview. J Clin Psychiatry 1997;58(suppl 9):29–32.
3. Kessler R, Sonnega A, Bromet E, Hughes M, Nelson CB. Posttraumatic stress disorder in the National Comorbidity Survey. Arch Gen Psychiatry 1995;52:1048–60.
4. Breslau N, Davis GC, Andreski P, Peterson E. Traumatic events and posttraumatic stress disorder in an urban population of young adults. Arch Gen Psychiatry 1991;48:216–22.
5. Resnick H, Kilpatrick DG, Dansky BS, Saunders BE, Best CL. Prevalence of civilian trauma and posttraumatic stress disorder in a representative sample of women. J Consult Clin Psychol 1993;61:984–91.
6. Green BL. Psychosocial research in traumatic stress: an update. J Trauma Stress 1994;7:341–62.
7. Breslau N, Davis GC, Andreski P, Peterson EL, Schultz LR. Sex differences in posttraumatic stress disorder. Arch Gen Psychiatry 1997;54:1044–8.
8. Kessler RC, Frank RG. The impact of psychiatric disorders on work loss days. Psychol Med 1997;27:861–73.
9. Norris FH. Epidemiology of trauma: frequency and impact of different potentially traumatic events on different demographic groups. J Consult Clin Psychol 1992;60:409–18.
10. Golding JM. Sexual assault history and physical health in randomly selected Los Angeles women. Health Psychol 1994;13:130–8.
11. Waigandt A, Wallace DL, Phelps L, Miller DA. The impact of sexual assault on physical health status. J Trauma Stress 1990;3:93–102.
12. Kimerling R, Calhoun KS. Somatic symptoms, social support, and treatment seeking among sexual assault victims. J Consult Clin Psychol 1994;62:333–40.
13. Diagnostic and statistical manual of mental disorders. (DSM-IV). 4th ed. Washington (DC): American Psychiatric Association; 1994.
14. Lecrubier Y. Posttraumatic stress disorder in primary care: a hidden diagnosis. J Clin Psychiatry 2004;65(suppl):49–54.
15. Rodriguez BF, Weisberg RB, Pagano ME, Machan JT, Culpepper L, Keller MB. Mental health treatment received by primary care patients with posttraumatic stress disorder. J Clin Psychiatry 2003;64:1230–6.
16. Loveland Cook CA, Flick LH, Homan SM, Campbell C, McSweeney M, Gallagher ME. Posttraumatic stress disorder in pregnancy: prevalence, risk factors and treatment. Obstet Gynecol 2004;103:710–7.
17. Samson AY, Bensen S, Beck A, Price D, Nimmer C. Post-traumatic stress disorder in primary care. J Fam Pract 1999;48:222–7.
18. McFarlane AC, Atchison M, Rafalowicz E, Papay P. Physical symptoms in post-traumatic stress disorder. J Psychosom Res 1994;38:715–26.
19. Meltzer-Brody S, Churchill E, Davidson JR. Derivation of the SPAN, a brief diagnostic screening test for post-traumatic stress disorder. Psychiatry Res 1999;88:63–70.
20. Stein MB, McQuaid JR, Pedrelli P, Lenox R, McCahill ME. Posttraumatic stress disorder in the primary care medical setting. Gen Hosp Psychiatry 2000;22:261–9.
21. Zlotnick C, Rodriguez BF, Weisberg RB, Bruce SE, Spencer MA, Culpepper L, et al. Chronicity in posttraumatic stress disorder and predictors of the course of posttraumatic stress disorder among primary care patients. J Nerv Ment Dis 2004;192:153–9.
22. Davidson JR, Book SW, Colket JT, Tupler LA, Roth S, David D, et al. Assessment of a new self-rating scale for post-traumatic stress disorder. Psychol Med 1997;27:153–60.
23. Chen CH, Shen WW, Tan HK, Chou JY, Lu ML. The validation study and application of stratum-specific likelihood ratios in the Chinese version of SPAN. Compr Psychiatry 2003;44:78–81.
24. Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller E, et al. The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry 1998;59(suppl):22–33.
25. Spitzer RL, Williams JB, Gibbon M, First MB. The Structured Clinical Interview for DSM-III-R (SCID). I: History, rationale, and description. Arch Gen Psychiatry 1992;49:624–9.
26. Simel DL, Samsa GP, Matchar DB. Likelihood ratios with confidence: sample size estimation for studies. J Clin Epidemiol 1991;44:763–70.
27. Delong ER, Delong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operator characteristic curves: a nonparametric approach. Biometrics 1988;44:837–45.
28. Breslau N, Peterson EL, Kessler RC, Schultz LR. Short screening scale for DSM-IV posttraumatic stress disorder. Am J Psychiatry 1999;156:908–11.
29. Leader S, Perales PJ. Provision of primary- preventive health care services by obstetrician-gynecologist. Obstet Gynecol 1995;85:391–5.
30. Brown CV. Primary care for women: the role of the obstetrician-gynecologist. Clin Obstet Gynecol 1999;42:306–13.
31. Stein DJ, Bandelow B, Hollander E, Nutt DJ, Okasha A, Pollack MH, et al. WCA Recommendations for the long-term treatment of posttraumatic stress disorder. CNS Spectr 2003;8(suppl):31–9.
32. Ahearn EP, Krohn A, Connor KM, Davidson JR. Pharmacologic treatment of posttraumatic stress disorder: a focus on antipsychotic use. Ann Clin Psychiatry 2003;15:193–201.
33. Schoenfeld FB, Marmar CR, Neylan TC. Current concepts in pharmacotherapy for posttraumatic stress disorder. Psychiatr Serv 2004;55:519–31.
34. Yehuda R. Post-traumatic stress disorder. N Engl J Med 2002;346:108–14.