**3.3.8 PTSD and older adults**

The majority of stroke survivors are older adults, with almost 80% of first-ever strokes occurring in people aged 65 years or older (Stroke Association, 2011). Knowledge regarding the prevalence and determinants of PTSD as well as its phenomenology in older adults, more generally, is limited (Averill & Beck, 2000; Cook & O'Donnell, 2005). Moreover, the majority of previous research on PSTD in older adults has focused on holocaust survivors, combat veterans and survivors of natural disasters, rather than on survivors of lifethreatening illnesses such as stroke (Cook & O'Donnell, 2005). Some studies have highlighted differences between younger and older adults in the experience and/or reporting of PTSD symptoms (Acierno et al., 2002; Davidson et al., 1990; Fontana & Rosenheck, 1994), although other work has suggested that their PTSD reactions are quite

Posttraumatic Stress Disorder after Stroke: A Review of Quantitative Studies 261

In order to provide accurate prevalence rates of PTSD after stroke, it is essential that structured clinical interviews are routinely employed. The Clinician-Administered PTSD Scale (Weathers et al., 2001) is widely regarded as the measure of choice for PTSD assessment as it is standardized, and can be used to provide both a PTSD diagnosis and a continuous measure of symptom severity. The PTSD module of the SCID (First et al., 1995) is also recommended for PTSD diagnosis, although it fails to provide a measure of PTSD symptom severity. When administering a structured clinical interview the researcher should consider whether the symptoms of PTSD, such as disturbed sleep, difficulty concentrating and amnesia, are better accounted for by alternative explanations (e.g., physical effect of the stroke, medication use, hospital environment) before they are categorized as having a psychogenic origin (O'Donnell et al., 2003). Moreover, this type of more in-depth questioning will also help the researcher to distinguish PTSD from generalized anxiety

Future research on the prevalence of post-stroke PTSD should employ longitudinal designs with multiple assessments of PTSD at fixed time points after stroke in order to provide accurate point prevalence rates and to chart the natural course of PTSD after stroke. Similarly, research on the correlates of PTSD after stroke should utilize prospective designs in which potential risk factors are assessed shortly after stroke (e.g., within one month) and related to PTSD caseness and symptom severity at subsequent time points while controlling

Large samples are essential to establish the prevalence of psychiatric disorders in new populations (O'Donnell et al., 2003). Future research on the prevalence of PTSD after stroke should therefore aim to recruit larger sample sizes than have been recruited to date. The actual sample size required to accurately estimate different prevalence rates in a population can be calculated (Daniel, 1999). Naing et al. (2006) recommend that the precision of the estimate should be ±5% when the expected prevalence rate is greater than 10%, and half the expected prevalence rate when less than 10%. Current prevalence estimates for PTSD after stroke range from 3% to 31%. The sample size required to estimate a 3% prevalence rate with 95% confidence intervals at 1.5% precision (i.e., between 1.5% to 4.5%) is 497, whereas the sample size required to estimate a 31% prevalence rate with 95% confidence intervals at 5% precision (i.e., between 26% and 36%) is 329. The adoption of multi-site studies are likely to aid the recruitment of such sample sizes and also address possible population differences across sites. In addition, studies on the predictors of PTSD after stroke need to be sufficiently powered to assess the impact of a full range of independent variables. Tabachnick and Fidell (2007) recommend that, in order to have adequate statistical power, the sample size for a regression analysis should be at least 50 + 8*k* (where *k* = number of independent variables). For example, for a regression analysis with 25 independent variables the sample size should be at least 250.

Future studies should provide more information on the representativeness of their samples. Ideally, where there are common care pathways for stroke victims as in the UK, consecutive

**3.4.1 Measurement of PTSD** 

disorder and/or major depression.

for the effects of initial PTSD symptoms.

**3.4.4 Sample representativeness** 

**3.4.2 Study design** 

**3.4.3 Sample sizes** 

similar (Bleich et al., 2005; Chung et al., 2005; Kohn et al., 2005). In relation to stroke, a number of studies have reported negative correlations between age and the severity of PTSD symptoms (Field et al., 2008; Sampson et al., 2003; Sharkey, 2007), although other studies have reported non-significant correlations (Bruggimann et al., 2006; Merriman et al., 2007). Nonetheless, there are a number of specific factors that may need to be considered when assessing PTSD in older people (Cook & O'Donnell, 2005). First, older adults are likely to have experienced multiple lifetime traumas (Creamer & Parslow, 2008) which may compound the impact of the current trauma, and vice-versa (Bechtle-Higgins & Follette, 2002). Second, older adults are more likely to suffer from cognitive impairments and dementia which may impact on their ability to fully process trauma memories. Third, older adults may have a more accepting attitude to illness and its psychological consequences.

#### **3.3.9 PTSD and medical events**

The assessment of PTSD symptoms following medical events, such as stroke, is complicated by the possibility that such symptoms may be confounded with the effects of physical illness and/or its treatment. For example, some of the hyperarousal (e.g., disturbed sleep, irritability, difficulty concentrating) and avoidance (e.g., diminished interest, detachment) symptoms that are used in the diagnosis of PTSD are also common problems experienced by survivors as a consequence of their stroke. Similarly, psychogenic amnesia is also included in the diagnostic criteria for PTSD. However, stroke is often associated with periods of amnesia that may have an organic (i.e., physical), rather than psychogenic, origin. The difficulty in differentiating between organic versus psychogenic causes of specific symptoms has implications for the assessment of PTSD. If such symptoms are simply taken to be part of PTSD rather having an organic origin, this is likely to lead to inflated estimates of the prevalence of PTSD. This effect is likely to be amplified when self-report measures are used to assess PTSD as alternative explanations for such symptoms cannot be explored.

Further, Mundy and Baum (2004) have argued that the nature of PTSD intrusion symptoms for medical events might be qualitatively different to those for other traumatic events. Whereas the focus of intrusions for more traditional traumas, such combat injuries and assaults, is on the past events, intrusions for medical events may also be future-oriented focusing on concerns about treatment, disease recurrence and ongoing functional impairment. Thus, in addition to having flashbacks to the traumatic event (in the past), individuals who have survived a life-threatening illness such as stroke may also have intrusive negative thoughts about the future (e.g., "Will I live to see my grandchildren grow up?", "Will I be able to work again?"). To date, there has been no phenomenological studies on the experience of PTSD after stroke. If the intrusions experienced by stroke survivors are found to be predominantly future-oriented, this would raise serious questions as to whether such clinical presentations are best thought of as PTSD – which, by definition, is *post-*trauma and characterized by being haunted by past horror – rather than anxiety about the (future) consequences of the stroke.

#### **3.4 Recommendations for future research**

On the basis of the review of studies on the prevalence and correlates of PTSD after stroke, five main recommendations for future research are made focusing on (i) the measurement of PTSD, (ii) study design, (iii) sample sizes, (iv) sample representativeness, and (v) the assessment of risk factors.

#### **3.4.1 Measurement of PTSD**

260 Post Traumatic Stress Disorders in a Global Context

similar (Bleich et al., 2005; Chung et al., 2005; Kohn et al., 2005). In relation to stroke, a number of studies have reported negative correlations between age and the severity of PTSD symptoms (Field et al., 2008; Sampson et al., 2003; Sharkey, 2007), although other studies have reported non-significant correlations (Bruggimann et al., 2006; Merriman et al., 2007). Nonetheless, there are a number of specific factors that may need to be considered when assessing PTSD in older people (Cook & O'Donnell, 2005). First, older adults are likely to have experienced multiple lifetime traumas (Creamer & Parslow, 2008) which may compound the impact of the current trauma, and vice-versa (Bechtle-Higgins & Follette, 2002). Second, older adults are more likely to suffer from cognitive impairments and dementia which may impact on their ability to fully process trauma memories. Third, older adults may

The assessment of PTSD symptoms following medical events, such as stroke, is complicated by the possibility that such symptoms may be confounded with the effects of physical illness and/or its treatment. For example, some of the hyperarousal (e.g., disturbed sleep, irritability, difficulty concentrating) and avoidance (e.g., diminished interest, detachment) symptoms that are used in the diagnosis of PTSD are also common problems experienced by survivors as a consequence of their stroke. Similarly, psychogenic amnesia is also included in the diagnostic criteria for PTSD. However, stroke is often associated with periods of amnesia that may have an organic (i.e., physical), rather than psychogenic, origin. The difficulty in differentiating between organic versus psychogenic causes of specific symptoms has implications for the assessment of PTSD. If such symptoms are simply taken to be part of PTSD rather having an organic origin, this is likely to lead to inflated estimates of the prevalence of PTSD. This effect is likely to be amplified when self-report measures are used to assess PTSD as alternative explanations for such symptoms cannot be explored. Further, Mundy and Baum (2004) have argued that the nature of PTSD intrusion symptoms for medical events might be qualitatively different to those for other traumatic events. Whereas the focus of intrusions for more traditional traumas, such combat injuries and assaults, is on the past events, intrusions for medical events may also be future-oriented focusing on concerns about treatment, disease recurrence and ongoing functional impairment. Thus, in addition to having flashbacks to the traumatic event (in the past), individuals who have survived a life-threatening illness such as stroke may also have intrusive negative thoughts about the future (e.g., "Will I live to see my grandchildren grow up?", "Will I be able to work again?"). To date, there has been no phenomenological studies on the experience of PTSD after stroke. If the intrusions experienced by stroke survivors are found to be predominantly future-oriented, this would raise serious questions as to whether such clinical presentations are best thought of as PTSD – which, by definition, is *post-*trauma and characterized by being haunted by past horror – rather

On the basis of the review of studies on the prevalence and correlates of PTSD after stroke, five main recommendations for future research are made focusing on (i) the measurement of PTSD, (ii) study design, (iii) sample sizes, (iv) sample representativeness, and (v) the

have a more accepting attitude to illness and its psychological consequences.

than anxiety about the (future) consequences of the stroke.

**3.4 Recommendations for future research** 

assessment of risk factors.

**3.3.9 PTSD and medical events** 

In order to provide accurate prevalence rates of PTSD after stroke, it is essential that structured clinical interviews are routinely employed. The Clinician-Administered PTSD Scale (Weathers et al., 2001) is widely regarded as the measure of choice for PTSD assessment as it is standardized, and can be used to provide both a PTSD diagnosis and a continuous measure of symptom severity. The PTSD module of the SCID (First et al., 1995) is also recommended for PTSD diagnosis, although it fails to provide a measure of PTSD symptom severity. When administering a structured clinical interview the researcher should consider whether the symptoms of PTSD, such as disturbed sleep, difficulty concentrating and amnesia, are better accounted for by alternative explanations (e.g., physical effect of the stroke, medication use, hospital environment) before they are categorized as having a psychogenic origin (O'Donnell et al., 2003). Moreover, this type of more in-depth questioning will also help the researcher to distinguish PTSD from generalized anxiety disorder and/or major depression.

#### **3.4.2 Study design**

Future research on the prevalence of post-stroke PTSD should employ longitudinal designs with multiple assessments of PTSD at fixed time points after stroke in order to provide accurate point prevalence rates and to chart the natural course of PTSD after stroke. Similarly, research on the correlates of PTSD after stroke should utilize prospective designs in which potential risk factors are assessed shortly after stroke (e.g., within one month) and related to PTSD caseness and symptom severity at subsequent time points while controlling for the effects of initial PTSD symptoms.

#### **3.4.3 Sample sizes**

Large samples are essential to establish the prevalence of psychiatric disorders in new populations (O'Donnell et al., 2003). Future research on the prevalence of PTSD after stroke should therefore aim to recruit larger sample sizes than have been recruited to date. The actual sample size required to accurately estimate different prevalence rates in a population can be calculated (Daniel, 1999). Naing et al. (2006) recommend that the precision of the estimate should be ±5% when the expected prevalence rate is greater than 10%, and half the expected prevalence rate when less than 10%. Current prevalence estimates for PTSD after stroke range from 3% to 31%. The sample size required to estimate a 3% prevalence rate with 95% confidence intervals at 1.5% precision (i.e., between 1.5% to 4.5%) is 497, whereas the sample size required to estimate a 31% prevalence rate with 95% confidence intervals at 5% precision (i.e., between 26% and 36%) is 329. The adoption of multi-site studies are likely to aid the recruitment of such sample sizes and also address possible population differences across sites. In addition, studies on the predictors of PTSD after stroke need to be sufficiently powered to assess the impact of a full range of independent variables. Tabachnick and Fidell (2007) recommend that, in order to have adequate statistical power, the sample size for a regression analysis should be at least 50 + 8*k* (where *k* = number of independent variables). For example, for a regression analysis with 25 independent variables the sample size should be at least 250.

#### **3.4.4 Sample representativeness**

Future studies should provide more information on the representativeness of their samples. Ideally, where there are common care pathways for stroke victims as in the UK, consecutive

Posttraumatic Stress Disorder after Stroke: A Review of Quantitative Studies 263

Post-stroke traumatic stress is an important but relatively neglected psychological consequence of stroke. It would be valuable to have reliable and accurate prevalence data from clinical diagnostic interviews with large, representative samples of stroke survivors collected over several time points. In addition, further work is required on the assessment of potential risk factors for the development of PTSD. This should include assessment of a full range of risk factors, including variables from current models of PTSD, shortly after stroke that can be related to subsequent PTSD caseness and symptom severity at later time points. A better understanding of the risk factors for PTSD after stroke has important clinical implications for the management of stroke survivors. It may assist in better differentiating the organic effects of stroke from the behavioural and psychological symptoms of the psychiatric disorder. More importantly, there is now a large body of evidence to guide the effective treatment of PTSD (Ponniah & Hollon, 2009). Appropriate use of such interventions has the potential to improve the quality of life, and reduce the care costs, of

Acierno, R., Brady, K.L., Gray, M., Kilpatrick, D.G., Resnick, H., & Best, C.L. (2002).

older and younger adults. *Journal of Clinical Geropsychology,* 8, pp. 13-23. Adams, R.E., & Boscarino, J.A. (2006). Predictors of PTSD and delayed PTSD after disaster:

American Psychiatric Association. (1994). *Diagnostic and Statistical Manual of Mental Disorders*

American Stroke Association (2011). About stroke. Retrieved 10 June 2011. Available from <http://www.strokeassociation.org/STROKEORG/AboutStroke/About-

Astrom, M. (1996). Generalized anxiety disorder in stroke patients: A 3-year longitudinal

Averill, P. M., & Beck, J. (2000). Post-traumatic stress disorder in older adults: A conceptual

Barker-Collo, S.L. (2007). Depression and anxiety 3 months post stroke: Prevalence and

Bechtle-Higgins, A., & Follette, V.M. (2002). Frequency and impact of interpersonal trauma

Berry, E. (1998). Post-traumatic stress disorder after subarachnoid haemorrhage. *British* 

Blake, D., Weathers, F., Nagy, L., Kaloupek, D., Klauminzer, G., Charney, D., & Keane, T.

Bleich, A., Gelkopf, M., Melamed, Y., & Solomon, Z. (2005). Emotional impact of exposure to

(1992). *Clinician Administered PTSD Scale*. National Centre for Post Traumatic Stress

terrorism among young-old and old-old Israeli citizens. *American Journal of Geriatric* 

(4th ed.). American Psychiatric Association, Washington D.C..

Correlates. *Archives of Clinical Neuropsychology,* 22, pp. 519-531.

in older women. *Journal of Clinical Geropsychology*, 8, pp. 215-226.

Stroke\_UCM\_308529\_SubHomePage.jsp>

review. *Journal of Anxiety Disorders*, 14, pp. 133-156.

*Journal of Clinical Psychology*, 37, pp. 365-367.

study. *Stroke,* 27, pp. 270-275.

Disorder, Boston.

*Psychiatry,* 13, pp. 705-712.

Psychopathology following interpersonal violence: A comparison of risk factors in

The impact of disclosure and psychosocial resources. *Journal of Mental Disease*, 194,

**4. Conclusions** 

this population.

**5. References** 

pp. 485-493.

admissions to stroke units/wards should be recruited. The resultant sample should then be compared with the patient population from which it was drawn in order to assess its representativeness. Given the severity of stroke and the ensuing levels of disability, it is likely that many stroke survivors will be unable to give informed consent and/or complete self-report measures or clinical diagnostic interviews. This is likely to affect the representativeness of the sample and restrict the extent to which the findings can be generalized to all stroke survivors. Studies should therefore provide detailed information on their recruitment procedures and exclusion criteria. In addition, future work should attempt to amend recruitment and assessment procedures in order, as far as possible, to recruit stroke survivors with communication and cognitive impairments into studies on post-stroke PTSD. For example, research on depression and aphasia (Thomas & Lincoln, 2008) has used visual analogue scales (Brumfitt & Sheeran, 1999) to assess emotional distress in stroke survivors with communication difficulties. Future work should therefore also focus on developing measures of PTSD symptom severity that can be completed by stroke survivors with communication difficulties and/or cognitive impairments.

#### **3.4.5 Assessment of risk factors**

Future research should assess a comprehensive range of potential risk factors, including stroke details, when assessing the predictors of post-stroke PTSD. In particular, future research should draw upon current models of PTSD (Brewin & Holmes, 2003) to assess the impact of more proximal psychological variables that have been found to be the strongest correlates of PTSD symptomatology across a range of traumas in meta-analytic reviews (e.g., Brewin et al., 2000; Ozer et al., 2003). Future research should routinely employ prospective designs and conduct multivariate analyses in which proximal psychological variables (e.g., appraisals, memory processes) are assessed shortly after stroke (e.g., within one month) and are related to the subsequent development of PTSD and/or symptom severity at later time points, while controlling for the influence of more distal factors (e.g., stroke details, demographics) and initial PTSD symptoms. In this way, future studies may assess the extent to which the effects of distal variables are mediated by these more proximal variables, thereby increasing our understanding of the mechanisms, or processes, underlying the development of PTSD after stroke.

#### **3.5 Ethical considerations**

When studying psychological reactions to life-threatening illnesses, such as stroke, that are associated with severe levels of disability and instability in the patient's medical condition, researchers need to be cognisant of ethical as well as scientific considerations (Tedstone & Tarrier, 2003). Particular attention needs to be paid to issues of informed consent given the cognitive and communication impairments experienced by many stroke survivors. PTSD has become a popular diagnosis over recent years (Summerfield, 2001), as evidenced by the increasing range of events, including life-threatening illnesses, that have been the focus of PTSD research (Tedstone & Tarrier, 2003). While such research has increased our understanding of psychological reactions to life-threatening illnesses, researchers should be aware of the risk of pathologising normal reactions to a traumatic event that may naturally remit over time (Middleton & Shaw, 2000).
