INTRODUCTION
As medical technology has improved the treatment of conditions that take life suddenly, people are living longer, and are therefore suffering from and dying of chronic illnesses more commonly than ever before. More than 80% of those >65 years of age have at least one chronic illness (Wolff et al., Reference Wolff, Starfield and Anderson2002) and such conditions have become the leading cause of death in high-income nations (Strong et al., Reference Strong, Mathers and Leeder2005). In other words, most individuals die of an illness of which they are aware well in advance of dying. Depression, anxiety, and general emotional distress are now well documented in this population, and models of intervention have been proposed (Kaasa et al., Reference Kaasa, Malt and Hagen1993; Haley et al., Reference Haley, Larson and Kasl-Godley2003). Less well understood has been PTSD.
Although a number of studies report a heightened prevalence of PTSD in patients with potentially terminal conditions, healthcare providers may be neither sufficiently aware of nor adequately trained to address this issue (Samson et al., Reference Samson, Bensen and Beck1999). In this article, I review research pertaining to PTSD at the end of life, paying particular attention to the implications of this literature for the counselors and clinicians who work directly with people with this condition. In the first half or this article, I discuss how PTSD symptoms may detrimentally affect the dying process. In the second half, I use this information as a “jumping off point” to suggest a preliminary psychosocial treatment model. By gathering together the relevant research and clinical literatures, I hope to provide a structure within which continued investigation and treatment innovation can take place.
Although some work directly addresses PTSD in “terminally ill” patients for whom death is imminent, such patients are notoriously difficult to study because of subject attrition from sudden death (George, Reference George2002). Moreover, George (Reference George2002) observed that studies have suffered methodologically from the fact that it is rarely clear when a patient's “life-threatening” or “life-limiting” illness becomes “terminal.” A major challenge in writing about end-of-life issues is that none of these terms are operationalized consistently in the literature, and they often overlap. Furthermore, research indicates that physicians are poor at prognostication as patients approach the end of life, often dramatically overestimating time remaining (Christakis & Lamont, Reference Christakis and Lamont2000). Therefore, in order to offer the most complete and clinically relevant picture possible, I cite studies not only of individuals categorized as having terminal illnesses, but also studies concerning life-threatening and life-limiting illnesses, where appropriate.
PREVALENCE OF PTSD AT THE END OF LIFE
The prevalence of PTSD in cancer varies widely by study. From 3 to 35% of people with cancer experience PTSD (Alter et al., Reference Alter, Pelcovitz and Axelrod1996; Green et al., Reference Green, Krupnick and Rowland2000; Mundy et al., Reference Mundy, Blanchard and Cirenza2000; Amir & Ramati, Reference Amir and Ramati2002; Kangas et al., Reference Kangas, Henry and Bryant2005) compared to the 8% lifetime prevalence in the general population (Kessler et al., Reference Kessler, Sonnega and Bromet1995). The prevalence of intrusion symptoms in those who do not necessarily meet full PTSD criteria appears higher, for example, 68% in one study involving breast cancer (Matsuoka et al., Reference Matsuoka, Inagaki and Sugawara2005). Given that well over half of individuals with cancer will experience remission (depending considerably on cancer type and year diagnosed), these samples include many patients who are not necessarily facing the end of life (Jemal et al., Reference Jemal, Siegel and Xu2010). For this reason, the prevalence of PTSD specifically at the end of life is difficult to determine with accuracy. Kaasa and colleagues (Reference Kaasa, Malt and Hagen1993) offer a rare study of PTSD symptoms in patients receiving palliative radiation for terminal cancers. Although PTSD was not directly diagnosed in this study, the Impact of Events Scale (Horowitz et al., Reference Horowitz, Wilner and Alvarez1979) was used to measure PTSD-like intrusion and avoidance symptoms. The results indicate that 82% of individuals scored in the high range (≥20) for intrusion symptoms and 88% for avoidance symptoms. Horowitz (Reference Horowitz, Goldberger and Breznitz1982) suggests that such high scores indicate a significant stress response syndrome that requires intervention and likely warrants a PTSD diagnosis. Therefore, although the prevalence of syndromal PTSD in patients with terminal cancers remains unknown, the prevalence of PTSD symptoms appears high.
With regard to non-cancer diagnoses, PTSD is experienced by 30.2–53.8% of HIV+ patients (Kelly et al., Reference Kelly, Raphael and Judd1998; Smith et al., Reference Smith, Egert and Winkel2002), 27.5% of patients with adult respiratory distress syndromes (Schelling et al., Reference Schelling, Stoll and Haller1998), and 14–59% of patients treated in an Intensive Care Unit (ICU) for various life-threatening conditions (Tedstone & Tarrier, Reference Tedstone and Tarrier2003). These rates are similar to those found in other trauma populations, but substantially higher than in the general population.
Unfortunately, providers may misdiagnose PTSD as general anxiety, depression, delirium, or other conditions perceived to exist in seriously ill patients (Woods, Reference Woods2003; Feldman & Periyakoil, Reference Feldman and Periyakoil2006), and patients may not receive appropriate care for the disorder. In one study of HIV+ women treated as outpatients, 59% of those meeting full diagnostic criteria for PTSD and 78% meeting partial criteria were not being treated for their trauma symptoms (Martinez et al., Reference Martinez, Israelski and Walker2002). Woods (Reference Woods2003) observes that patients typically do not complain directly of PTSD, and physicians therefore may not be attuned to its assessment and diagnosis. Additionally, a much greater number of empirical studies address depression and non-PTSD anxiety disorders than address PTSD at the end of life. Woods (Reference Woods2003) suggests that this imbalance of information may result in a misclassification bias and an overestimation of those other disorders because of failure to assess PTSD. Another difficulty is that PTSD symptoms overlap somewhat with those of other disorders as well as with non-pathological reactions to terminal illness (Block, Reference Block2006). Obviously, the experience of depressed and anxious affects is very common at the end of life (Block, Reference Block2006), whereas syndromal PTSD is not as common. These facts render a clinician's careful differential diagnosis particularly important, because therapeutic approaches obviously differ substantially for PTSD versus other conditions.
CAUSES OF PTSD AT THE END OF LIFE
The literature suggests that PTSD may appear at the end of life primarily for two reasons. First, the physical illness may trigger pre-existing or currently existing trauma symptomatology (Macleod, Reference Macleod1994). Research indicates that prior trauma history predicts PTSD in serious medical illness (Green et al., Reference Green, Krupnick and Rowland2000; Ruzich et al., Reference Ruzich and Chee Leong Loot2005). A feature of PTSD is that stimuli and situations that remind individuals of the original trauma trigger symptoms. Both learning that one has a potentially terminal illness and witnessing the weakening of one's body could serve as such reminders.
Second, the physical illness may serve as the primary instigating event for PTSD. In this regard, numerous studies have documented PTSD symptoms with content specifically related to the life-threatening or terminal illness (e.g., Butler et al., Reference Butler, Koopman and Classen1999; Jacobsen et al., Reference Jacobsen, Sadler and Booth-Jones2002; DuHamel et al., Reference DuHamel, Ostroff and Ashman2004; Kangas et al., Reference Kangas, Henry and Bryant2005). It is worth mentioning a historical controversy in this regard. The Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR) defines a “trauma” (i.e., a Criterion A stressor) as any event that a person experiences, witnesses, or is confronted with that threatens death or serious injury, or is a threat to the physical integrity of oneself or others (American Psychiatric Association, 2000). This event must elicit a response involving fear, helplessness, or horror. Common traumas include natural disasters, combat exposure, domestic violence, childhood abuse, and sexual assault, among others (Foy, Reference Foy and Foy1992; Flett et al., Reference Flett, Kazantis and Long2004). Medical illness generally was not considered to be a traumatic stressor in the diagnostic system until the publication of DSM-IV in 1994 (American Psychiatric Association, 1994), principally because previous versions of the DSM (e.g., American Psychiatric Association, 1987) required that such events fall “outside the range of usual human experience.” Elimination of this clause now enables clinicians to diagnose PTSD in direct response to physical illness.
It is not known whether one of these explanations accounts for most cases of PTSD at the end of life. Regardless, the deleterious effects of PTSD at the end of life may be similar.
EFFECTS OF PTSD AT THE END OF LIFE
Before discussing the particular ways in which PTSD can hamper functioning at the end of life, it may be helpful to review the typical symptoms of the disorder. The DSM-IV-TR divides these symptoms into three clusters (American Psychiatric Association, 2000). The first cluster consists of re-experiencing symptoms such as recurrent and intrusive disturbing memories, nightmares, and hallucination-like flashbacks. The second cluster consists of avoidance symptoms. Attempts to evade activities, places, people, and conversations that remind the individual of the trauma can lead to significant social and occupational impairment. Other symptoms in this cluster include attempts to suppress trauma-related memories and emotions, amnesia for aspects of the trauma, affective “numbness,” and a sense of foreshortened future. The last cluster consists of hyperarousal symptoms such as hypervigilance, irritability, exaggerated startle response, and insomnia. Other notable psychosocial issues associated with PTSD include (but are not limited to) co-morbid mood and anxiety disorders, substance abuse (American Psychiatric Association, 2000), mistrust in authority (Glover et al., Reference Glover, Peleski and Bruno1990; Kubany et al., Reference Kubany, Gino and Denny2005), and suicide (Oquendo et al., Reference Oquendo, Brent and Birmaher2005).
Although these symptoms can appear at any stage of life, a number of effects may be unique to the end of life. Specifically, PTSD can encumber factors that surveys of patients, their families, and healthcare providers indicate are associated with a “good” death — emotional and practical support, doctor–patient communication, life review, and acceptance of dying (Low & Payne, Reference Low and Payne1996; Pierson et al., Reference Pierson, Curtis and Patrick2002; Steinhauser et al., Reference Steinhauser, Christakis and Clipp2007).
Reliance on Others for Support
Patients at the end of life often become reliant on others, particularly family members. Research indicates that patients with adequate emotional and practical support experience less depression (Gençóz & Aston, Reference Gençóz and Astan2006), lower pain levels (Steele et al., Reference Steele, Mills and Hardin2005), better overall quality of life (Tang et al., Reference Tang, Aarsonson and Forbes2004), and may live longer (Herbst-Damm & Kulik, Reference Herbst-Damm and Kulik2005) than those without such support. Unfortunately, PTSD can disrupt familial support networks. Research on physically healthy male veterans shows that PTSD is associated with severely elevated levels of marital and family dysfunction (Jordan et al., Reference Jordan, Marmar and Fairbank1992). Moreover, spouses can experience secondary traumatization, a state in which PTSD symptoms and other forms of distress result from being in relationship with the traumatized individual (Beckham et al., Reference Beckham, Lytle and Feldman1996; Bramsen et al., Reference Bramsen, van der Ploeg and Twisk2002). Such secondary traumatization is strongly associated with “caregiver burden,” the daily weight of caring for someone with PTSD (Dekel & Solomon, Reference Dekel, Solomon, Figley and Nash2007). Perhaps as a result of these factors, individuals with PTSD are more likely than the general public to be separated or divorced (Riggs et al., Reference Riggs, Byrne and Weathers1998; Afifi et al., Reference Afifi, Cox and Enns2006).
Although this research concerns physically healthy individuals, terminally ill patients obviously are influenced by their pre-morbid psychosocial histories, at least to a certain extent. One can therefore cautiously extrapolate these effects to such patients. Because of divorce or separation, patients with long-standing PTSD may be isolated and lack caregiver support. Despite 70% of people stating that they wish to be cared for at home prior to death (Foreman et al., Reference Foreman, Hunt and Luke2006), those with PTSD may face elevated risk of spending their remaining days in institutional settings. Last, lack of support may diminish the aforementioned benefits of adequate caregiving, leading to increased pain and depression, and decreased quality and quantity of life. Clearly, further research is needed to address these hypotheses directly.
Physician–Patient Communication
In a meta-analysis of 21 studies concerning patients with a variety of physical illnesses, including life-threatening and terminal conditions, Stewart (Reference Stewart1995) found that quality of physician–patient communication was associated with emotional health, functioning, physiological measures such as blood pressure and blood sugar, and pain control. Moreover, in surveys of dying patients and their families, communication with physicians is consistently cited as an important factor in quality of life (e.g., Teno et al., Reference Teno, Clarridge and Casey2004; Steinhauser et al., Reference Steinhauser, Christakis and Clipp2007).
Certain features of PTSD can hamper physician–patient communication. First, good medical communication at the end of life involves discussions of mortality, which can serve as trauma reminders. As discussed, individuals with PTSD often attempt to avoid stimuli that remind them of the original trauma. There is a clear empirical link between PTSD and such avoidance coping (Bryant & Harvey, Reference Bryant and Harvey1995; Tiet et al., Reference Tiet, Rosen and Cavella2006), and studies demonstrate a relationship between avoidance coping and poorer health status in patients with cancer (Epping-Jordan et al., Reference Epping-Jordan, Compas and Howell1994; Roesch et al., Reference Roesch, Adams and Hines2005).
Second, a common symptom of PTSD, particularly among men, is irritability and anger (Beckham et al., Reference Beckham, Moore and Reynolds2000; Jakupcak et al., Reference Jakupcak, Conybeare and Phelps2007). In those who have experienced combat trauma, a related feature is mistrust, particularly of those in authority (Glover et al., Reference Glover, Peleski and Bruno1990; Kubany et al., Reference Kubany, Gino and Denny2005). Such symptoms may lead patients to be suspicious or disregarding of medical advice. Indeed, research demonstrates elevated levels of nonadherance to medical advice in patients with PTSD (Boarts et al., Reference Boarts, Sledjeski and Bogart2006; Woods & Terhaar, Reference Woods and Terhaar2007).
Life Review
Decades ago, Butler (Reference Butler1963) observed that older adults engage in a meaning-making process known as life review. Often taking the form of reminiscence or story-telling, he theorized that, as people realize they have limited time remaining, they naturally examine their lives. Although evidence is mixed (Giambra, Reference Giambra1977), some studies show that older adults engage in more reminiscence than middle-age adults (Lieberman & Falk, Reference Lieberman and Falk1971; Revere & Tobin, Reference Revere and Tobin1980). Research also demonstrates a relationship between reminiscence and better psychosocial adjustment in older adults (Havighurst & Glasser, 1972; Boylin et al., Reference Boylin, Gordon and Nahrke1976). Furthermore, numerous studies demonstrate the efficacy of life-review therapies in reducing depression and negative affect in older adults (Arean et al., Reference Arean, Perri and Nezu1993; Scogin & McElreath, Reference Scogin and McElreath1994; Haight et al., Reference Haight, Michel and Hendrix1998; Serrano et al., Reference Serrano, Lotorre, Gatz and Montanes2004). Two studies also suggest that this therapy reduces depression in those with terminal cancers (Ando et al., Reference Ando, Tsuda and Morita2007) as well as less severe cancers (Ando et al., Reference Ando, Tsuda and Moorey2006).
The symptoms of PTSD may interfere with life review in two ways. First, to avoid trauma memories, patients may fail to engage in the process. Second, those who engage in life review without adequate coping resources may experience significant negative affect. Westwood and McLean (Reference Westwood, McLean, Kunz and Soltys2007) suggest that, when intense trauma memories are present, there is potential for re-traumatization. Consistent with this assertion, Haight (Reference Haight and Hendricks1995) reviewed nearly 100 published life reviews and found that 7% involved negative emotional outcomes.
Acceptance of Death
It is a classic perspective in the death-and-dying literature (almost a “truism” in the field) that acceptance of death's inevitability is a sign of psychological health in patients with terminal illness (Kubler-Ross, Reference Kubler-Ross1969). Although little research addresses this topic, one study of healthy older adults demonstrates a correlation between death acceptance and life satisfaction (Flint et al., Reference Flint, Gayton and Ozmon1983). Despite the pervasiveness of this assumption, further research is clearly needed before it is treated as “fact.”
Nonetheless, PTSD may interfere with acceptance of death through the use of avoidance coping. As discussed, research documents an association between avoidance coping and poorer health status in patients with cancer (Epping-Jordan et al., Reference Epping-Jordan, Compas and Howell1994; Roesch et al., Reference Roesch, Adams and Hines2005). Often labeled “denial” by end-of-life care professionals, Goetzke (Reference Goetzke1995) observed that such a coping strategy “can cause harm. Patients or families may incur astronomical costs in search of a cure or a better prognosis. They'll ignore pain because it may mean the disease is getting worse. They'll lose valuable time that could be spent on unfinished business” (p. 20).
EXISTING TREATMENTS FOR PTSD
Although numerous psychotherapies have been successfully applied to PTSD (e.g., Meichenbaum, Reference Meichenbaum1985; Resick & Schnicke, Reference Resick and Schnicke1992; Bleiberg & Markowitz, Reference Bleiberg and Markowitz2005), the preponderance of efficacy data support two treatments — exposure-based cognitive-behavioral therapy (CBT) and eye movement desensitization and reprocessing (EMDR). Research suggests that both function at least partially through exposure (Davidson & Parker, Reference Davidson and Parker2001), prompting patients to re-experience distressing memories or triggers of PTSD symptomatology in a gradual manner.
Exposure-based CBT, particularly prolonged exposure, boasts the largest body of empirical support (Foa et al., Reference Foa, Rothbaum and Riggs1991, Reference Foa, Dancu and Hembree1999; Taylor et al., Reference Taylor, Thordarson and Maxfield2003; see Nemeroff et al., Reference Nemeroff, Bremner and Foa2006). As detailed by Foa and Rothbaum (Reference Foa and Rothbaum1998), prolonged exposure consists of repeated, extended periods of exposure to traumatic memories and PTSD triggers (i.e., stimuli associated with the trauma). Thereby, patients undergo “emotional processing,” recognizing that such memories and triggers are not objectively dangerous, and experiencing a reduction in symptoms. Prolonged exposure is often combined with other CBT techniques, including psychoeducation, breathing and relaxation training, and cognitive restructuring.
The procedure of EMDR is described by Shapiro (Reference Shapiro1995), although it has been implemented somewhat differently across reports. Patients focus on a mental image from the trauma along with the associated beliefs and feelings, while following the therapist's finger with their eyes as it passes from side to side. Patients indicate any changes in the images, beliefs, or feelings that result, and this procedure continues with this new set of elements. Thus, the therapy follows chains of associations with the goals of desensitizing the original image and reprocessing the beliefs surrounding that image. EMDR's mechanisms of action are hotly debated, with particular controversy centering on the necessity of eye movements or other forms of bilateral stimulation (Davidson & Parker, Reference Davidson and Parker2001; Rogers & Silver, Reference Rogers and Silver2002). Despite this controversy, ample evidence supports its efficacy, even in comparison to other treatments (Davidson & Parker, Reference Davidson and Parker2001; Ironson et al., Reference Ironson, Freund and Williams2002; Wilson et al., Reference Wilson, Becker and Tinker1995).
Despite the notable efficacy of exposure-based CBT and EMDR, neither treatment (as presently performed) is optimally suited to address PTSD in individuals at the end of life. There are several reasons for this assertion, all of which are related to the unique characteristics of the population. First, because these interventions most often are delivered in 8–16 sessions (Davidson & Parker, Reference Davidson and Parker2001; Taylor et al., Reference Taylor, Thordarson and Maxfield2003), some patients may die before treatment is completed. Consider patients in hospice care. In the United States, patients generally qualify for hospice if a physician notes that they likely will not survive for more than 6 months, though most hospice referrals do not occur until much later. According to the National Hospice and Palliative Care Organization (2008), the mean length of time patients spent on hospice in the United States was 67.4 days in 2007. The median length of stay was only 20 days. Although numerous sessions may be needed to produce long-lasting reductions in PTSD symptoms using these therapies, a focus on long-term outcome may not be relevant for patients with such limited prognoses. Of course, therapists may encounter patients earlier in the disease trajectory, rendering traditional PTSD treatments very appropriate. Close communication with patients' physicians can provide the necessary information for clinical decision making in this regard. It is important to note, however, that research demonstrates that physicians tend to overestimate life expectancy in terminally ill patients by a factor of five (Christakis & Lamont, Reference Christakis and Lamont2000).
Second, typical 50–90 minute sessions are too lengthy for many patients at the end of life, for whom fatigue is common (Jackson & Block, Reference Jackson and Block2002). For example, Teunissen and colleagues (Reference Teunissen, Wesker and Kruitwagen2007) report an 88% prevalence of fatigue in the last 2 weeks of life in patients with incurable cancers. In fact, among patients with cancer in general, fatigue has been described as an almost universal symptom (Wagner & Cella, Reference Wagner and Cella2004). For this reason, clinicians may need to limit session duration substantially.
Third, both exposure-based CBT and EMDR require patients to recollect vivid trauma memories, which may lead to a temporary increase in PTSD symptoms. This especially may be the case for exposure-based CBT (Ironson et al., Reference Ironson, Freund and Williams2002), which functions partially through the elicitation of emotional arousal (Foa et al., Reference Foa, Riggs and Massie1995). Nishith et al. (Reference Nishith, Resick and Griffin2002) randomly assigned 171 female victims of rape to 9–12 sessions of one of two CBT therapies involving exposure — prolonged exposure or cognitive processing therapy — or to a “minimal attention” control condition. In both CBT conditions, trauma re-experiencing symptoms got worse before they got better, peaking in session 6 before declining. Emotional arousal symptoms remained relatively level until session 6, at which point these also began to diminish. As such, it is worth considering on a case-by-case basis whether a patient who may die soon after initiation of treatment could die with greater PTSD symptoms than if he or she had not received therapy. By no means should this imply that exposure-based therapies are necessarily inappropriate, but caution should be taken to carefully evaluate patient physical health and prognosis before offering this type of treatment. Interestingly, EMDR has shown some promise in courses of three or fewer sessions (Wilson et al., Reference Wilson, Becker and Tinker1995), but sessions generally are 90 minutes in length, necessitating relatively high energy levels. In addition, one case study documented the effective use of trauma-focused CBT — an exposure-based therapy — after bone marrow transplant for leukemia (Duhamel et al., 2000). It is important to note, however, that this patient was healthy enough to live nearly 3 years beyond the bone marrow transplantation as well as attend 10 therapy sessions and a 6-month follow-up assessment.
Fourth, PTSD at the end of life often involves an ongoing trauma rather than a traumatic event or set of events that occurred in the past and have since discontinued. In their grounded theory analysis of the trauma-related themes of six patients who had undergone peripheral blood stem cell transplantation for multiple myeloma, Dahan and Auerbech (Reference Dahan and Auerbach2006) conclude that “myeloma differs from other traumas by virtue of its ongoing nature. Without the distance of time to process the event, patients' ability to grasp and integrate what they have been through may be hindered” (p. 385). This also could be said of nearly any terminal illness. There has been much debate in the clinical literature regarding the wisdom of treating ongoing or very recent trauma with techniques that require in-depth discussion of the traumatic events, with some arguing in favor (e.g., Robinson, Reference Robinson2004) and some against (e.g., Devilly & Cotton, Reference Devilly and Cotton2003). In particular, critical incident stress debriefing (CISD) (Mitchell, Reference Mitchell1983) and its variants have been criticized for their potentially re-traumatizing or further traumatizing effects. After reviewing the literature, Devilly et al. (Reference Devilly, Gist and Cotton2006) conclude that the best treatment evidence supports the application of exposure-based CBT beginning 2–6 weeks after a traumatic event has occurred. Unfortunately, such a waiting period is not possible when the trauma involves an ongoing set of circumstances surrounding terminal illness that will not likely cease until death.
Last, typical interventions for PTSD do not address issues that are unique to the end of life. Problems with doctor–patient communication and worries about medical matters may be interwoven with PTSD symptoms. As already discussed at length, the most damaging effects of PTSD may not be the symptoms at all, but the ways in which the symptoms interfere with a “good” dying process.
Pharmacologic treatments suffer from similar drawbacks. Although a full review of the pharmacologic treatment of PTSD is beyond the scope of this article, the most commonly used medications for PTSD are antidepressants — primarily selective serotonin reuptake inhibitors and, to a lesser extent, tricyclic antidepressants (Pearlstein, Reference Pearlstein2000). These medications may require ≥6 weeks to alleviate symptoms, rendering them ineffective for patients with limited prognoses (see Feldman & Periyakoil, Reference Feldman and Periyakoil2006). In addition, medications alone do not address the particular ways in which PTSD interferes with a “good” dying process. As such, although medication can be an important tool in the treatment of PTSD at the end of life, it offers an incomplete solution to this complex psychosocial problem.
As effective as these treatments certainly are for treating PTSD in physically healthy patients, the fundamental difficulty with applying them to individuals at the end of life is that they fail to follow a palliative care philosophy. Medical professionals, of course, often distinguish between curative and palliative care (Feldman & Lasher, Reference Feldman and Lasher2008). When a person is diagnosed with cancer, for example, care may initially focus on cure. Therapies such as chemotherapy are frequently effective at achieving this end but can lead to unpleasant side effects. Such side effects are acceptable, however, because the treatments likely will lead to long-term cure or remission of the cancer. When a patient enters the “terminal” stage of an illness (and sometimes before), physicians may suggest transitioning to a palliative approach. Given that cure is less likely, physicians may emphasize interventions that increase physical comfort and quality of life, possibly discontinuing potentially unpleasant curative treatments.
The World Health Organization (2009) defines palliative care as, “An approach that improves the quality of life of patients and their families facing the problems associated with life-threatening illness, through the prevention and relief of suffering.” Hallenbeck (Reference Hallenbeck2003) adds that “the emphasis is on symptom management . . . with no effort to prolong or hasten death” (p. 5). Curative care may eliminate symptoms as a result of cure, of course, but this is not its primary intent. Palliative approaches specifically target symptom management and quality of life.
Traditional treatments for PTSD follow a curative model. Although they may require considerable time and possibly lead to short-term discomfort, they will likely bring about long-term remission of the disorder. This model makes sense when applied to physically healthy individuals. When patients are dying, however, a psychotherapeutic approach paralleling medical palliative care may be called for. Such an approach would aggressively target immediate emotional suffering, but deemphasize long-term cure of PTSD.
A PALLIATIVE APPROACH TO PTSD AT THE END OF LIFE
The proposed psychosocial treatment model consists of cognitive and behavioral techniques, nearly all of which are drawn from evidence-based psychotherapies for PTSD. In this sense, it is nothing new. The value of the present model, tentatively named stepwise psychosocial palliative care (SPPC), is primarily a heuristic one. Specifically, given the unique characteristics of the population, the stepwise nature of the model allows therapists to determine which types of interventions to use and when to use them. Similar to an approach for geriatric populations by Hyer and Woods (Reference Hyer, Woods, Follette, Ruzek and Follette1998), this model consists of three stages: I) palliate immediate discomfort and provide social supports, II) provide psychoeducation and enhance coping skills, and III) treat specific trauma issues.
These stages are ordered to maximize the short-term effectiveness of each therapeutic interaction, such that quality of life may be improved even when very little time is available. At each stage, therapists assess the effectiveness of interventions and gather information regarding the patient's prognosis. Therapists move to later stages only under two conditions: (1) earlier stages were not effective, and (2) the patient's prognosis allows sufficient time to implement later stages. This approach is thus applicable to patients for whom death is imminent and for those earlier in the disease trajectory.
Stage I: Palliate Immediate Discomfort and Provide Social Supports
At Stage I, the therapist focuses on making concrete changes in the patient's environment that can rapidly reduce distress and physical discomfort. Because patients may have only weeks or months to live, emphasis is placed on short-term maximization of quality of life.
As in all forms of therapy, the development of a trusting therapeutic relationship is essential. Use of active listening and validation is important at this stage, and these skills are widely recognized as necessary for providers in all healthcare professions when working with patients at the end of life (Larson, Reference Larson1993; Schofield et al., Reference Schofield, Carey and Love2006). In the context of this developing relationship, the therapist is likely to discover practical concerns that are distressing to the patient. Aiding in making funeral arrangements, re-establishing contact with distant loved ones, or simply finding the right television channel can dramatically improve quality of life in the short term. The therapist may work with family members or other caregivers, helping them to make similar environmental changes to palliate patient discomfort. Additional interventions may include mediating family conflicts, educating other healthcare providers and family about how to avoid triggering PTSD symptoms, finding answers to health questions, and mediating conversations with medical providers. The therapist also can convey his or her diagnosis of PTSD to the physician and inquire about the appropriateness of pharmacologic treatment.
As can be seen in these suggested interventions, at this stage the therapist should offer to solve problems for the patient, if possible and desired by the patient. This recommendation can make therapists uncomfortable who are not accustomed to providing end-of-life care. Clinical training typically emphasizes teaching patients skills to help themselves solve their own problems. Even the practice of giving advice is viewed by some as fostering dependence, and therefore should be prohibited as a breach of professional ethics (Couture & Sutherland, Reference Couture and Sutherland2006). This prohibition, however, follows from a curative philosophy in which patients have the requisite time, energy, and health to solve their own problems. Dying patients, almost by definition, do not meet these requirements. Of course, such problem-solving efforts should be undertaken only with patients' permission.
The length of Stage I, and indeed of all three stages of the SPPC model, can be highly variable. In some cases, the therapist may choose never to move to later stages, having determined that Stage I interventions sufficiently alleviate distress. In contrast, the therapist may determine within the first clinical session that Stage I interventions will not achieve this end, thus proceeding rapidly to Stage II. In most cases, progress through the model will occur at a pace between these extremes. Such decisions are determined by multiple factors beyond the scope of this article, including the intensity of symptoms, specifics of the patient's difficulties, quality of the therapeutic relationship, patient's physical condition, and patient's goals.
Stage II: Provide Psychoeducation and Enhance Coping Skills
At Stage II, the therapist focuses primarily on helping the patient to develop skills to better cope with PTSD symptoms. In the previous stage, the therapist addressed immediate problems for the patient. In the present stage, the therapist teaches the patient behavioral strategies to improve his or her own quality of life. Although these strategies may provide more durable gains, they also require greater time to instill. For this reason, therapists move to Stage II only if earlier interventions were ineffective and patients' prognoses provide sufficient time. Of note, moving to Stage II does not mean abandoning Stage I interventions. The stages of this model do not represent discontinuous phases of therapy but sets of techniques that build on one another. Speaking at length about trauma memories is still discouraged at this stage, however, as this could increase distress.
At Stage II, the provider should offer psychoeducation to the patient and his or her family, if appropriate, regarding the symptoms of PTSD. Most evidence-based psychotherapy treatment protocols recommend providing such information (e.g., Shapiro, Reference Shapiro1995; Foa & Rothbaum, Reference Foa and Rothbaum1998), which helps render the symptoms predictable and manageable, and helps to secure cooperation in treatment. The fact that this component appears at Stage II does not mean that brief psychoeducation cannot be provided earlier. However, detailed psychoeducation in Stage I is not the most effective use of time if life is severely limited by prognosis.
Stage II interventions may also include traditional CBT techniques such as relaxation training (Goldfried & Davison, Reference Goldfried and Davison1994) and breathing retraining (Guell et al., Reference Guell, Resqueti and Sangenis2006), both of which target the somatic symptoms of anxiety. Thought-stopping (Wolpe, Reference Wolpe1973) and cognitive restructuring (Beck, Reference Beck1995) can be used to target distorted thinking about past events and current health conditions. Acceptance-based techniques drawn from acceptance and commitment therapy (Hayes et al., Reference Hayes, Strosahl and Wilson2003), mindfulness-based stress reduction (Kabat-Zinn, Reference Kabat-Zinn1990), and dialectical behavior therapy (Linehan, Reference Linehan1993) also may be helpful in addressing negative thoughts, emotions, and somatic symptoms. Last, communication training (McKay et al., Reference McKay, Davis and Fanning1995) can be used to improve the patient's relationship with healthcare providers and others.
These techniques are, of course, quite different from one another and likely offer differing levels of effectiveness. It is beyond the scope of this article to review the efficacy of all possible techniques. For heuristic purposes, it is important to note that what unites Stage II interventions is their emphasis on skills training, which generally requires greater energy and time on the part of the patient than do Stage I interventions but will likely lead to longer-lasting gains.
Stage III: Treat Specific Trauma Issues
In neither of the preceding stages was the content of the trauma directly addressed. This strategy prevents unnecessary distress, given that many patients may die prior to completion of Stage II. In the present stage, however, patient and therapist focus on the trauma memories. Therapists only use Stage III interventions under three conditions: (1) earlier interventions were ineffective at adequately addressing the patient's suffering, (2) the patient's prognosis permits the use of longer-term strategies, and (3) the patient indicates an openness to speaking about the trauma.
Any exposure-based approach potentially may be used at this stage. Especially, careful consideration should be given to patient prognosis and level of energy, however, before a standard course of prolonged exposure is undertaken, because of its potential for temporarily increased distress (Ironson et al., Reference Ironson, Freund and Williams2002; Nishith et al., Reference Nishith, Resick and Griffin2002). Prolonged exposure involves vividly re-experiencing trauma memories without escaping the emotions they trigger. During a session, emotional distress generally increases, levels off, and eventually decreases. Prolonged exposure treatment guidelines suggest that much of an hour session may be necessary to produce such clinically significant reductions in distress, particularly concerning especially distressing aspects of the trauma (Foa & Rothbaum, Reference Foa and Rothbaum1998). As discussed previously, because many patients at the end of life are too fatigued to engage in exposures of this length, sessions may end while distress is still high. Of course, should the therapist believe that lengthy sessions are possible based on clinical observation and information provided by the physician, prolonged exposure may be a viable option at this stage.
Therapists may wish to alter the procedures of existing therapies (e.g., exposure-based CBT or EMDR) to accommodate clients' current levels of functioning and to help avoid the potential for heightened distress. If session duration is shortened because of patient fatigue, therapists may consider offering to meet more frequently. Therapists also should consider exposing patients very gradually to trauma memories, increasing the intensity and detail of exposure more slowly from session to session than they would with healthy individuals. This precaution may help ensure that shortened sessions are still long enough for distress ultimately to diminish (i.e., for extinction to occur).
Therapists also may consider alternative approaches for focusing on the trauma. One approach that has shown promise in end-of-life care is life-review therapy. Maercker (Reference Maercker2002) defines this approach as, “a directed process of reminiscence in which a therapist helps the patient to organize and evaluate his or her memories” (p. 240). Although there is considerable empirical support for life-review therapy for late-life depression (Arean et al., Reference Arean, Perri and Nezu1993; Haight et al., Reference Haight, Michel and Hendrix1998; Serrano et al., Reference Serrano, Lotorre, Gatz and Montanes2004; see Scogin & McElreath, Reference Scogin and McElreath1994 for a review), only one study addresses its effectiveness with regard to PTSD (i.e., Maercker, Reference Maercker2002). Though this study is not methodologically rigorous, it quantitatively documents reductions in both intrusion and avoidance symptoms in four case reports.
Life-review therapy proceeds in a semi-structured manner. Although the patient determines the particular content and level of detail of memories discussed, the therapist uses a set of planned questions to encourage sharing of relevant information. The focus is on helping the patient re-experience and make meaning of memories. Foa and Riggs (Reference Foa, Riggs, Oldham, Riba and Tasman1993) suggest that PTSD may be the result of “the victim's failure to process the trauma successfully” using his or her pre-existing schemas. Life-review therapy provides the patient an opportunity for such processing (Butler et al., Reference Butler, Lewis and Sunderland1998; Romanuik & Romanuik, Reference Romanuik and Romanuik1981). This sort of meaning making is similar to the “emotional processing” found in exposure-based CBT and “reprocessing” found in EMDR. In contrast to these approaches, however, life- review therapy involves discussing aspects of the entire lifespan, rather than focusing exclusively on the trauma (Maercker, Reference Maercker2002). This feature of treatment may be especially relevant to older patients given the aforementioned research indicating an association between lifespan-oriented life review and better general psychosocial adjustment in later life (Revere & Tobin, Reference Revere and Tobin1980).
It is important to note, however, that no controlled studies regarding the efficacy of life-review therapy for PTSD have been performed. Because it does not contain the extended periods of exposure present in both prolonged exposure-based CBT and EMDR, it may not be as effective. A life-review approach is mentioned here merely as one possibility that may be more appropriate to patients' limited prognoses and energy levels.
CONCLUSIONS
Although a significant number of patients will suffer from PTSD as they face the end of life, existing models of PTSD treatment do not adequately address the particular needs of these patients. Such treatments are focused on long-term cure of PTSD rather than short-term maximization of quality of life. The SPPC approach suggested in this article was designed to address these concerns and provide a viable treatment alternative. With continued work and increased clinical awareness, perhaps the impact of PTSD on the precious final months of life can be substantially reduced.
ACKNOWLEDGMENTS
The author thanks Julia Kasl-Godley for her thorough and helpful comments on an earlier draft of this article.
