Published ahead of print on January 31, 2008, doi:10.1164/rccm.200706-857OC
© 2008 American Thoracic Society doi: 10.1164/rccm.200706-857OC
Children's Factual and Delusional Memories of Intensive Care1 Pediatric Psychology Service, St. George's Hospital, London, United Kingdom; 2 Community Health Sciences, St. George's University of London, London, United Kingdom; and 3 Pediatric Intensive Care Unit, Great Ormond Street Hospital for Children, London, United Kingdom Correspondence and requests for reprints should be addressed to Gilian Colville, B.Sc., M.Phil., Consultant Clinical Psychologist, Pediatric Psychology Service, St. George's Hospital, London SW17 0QT, UK. E-mail: gcolvill{at}sgul.ac.uk
Rationale: Delusional memories are significantly associated with post-traumatic stress in adult patients after intensive care. Objectives: In this study, we attempted to establish whether this relationship was found in children. We also examined the association between factual memory and distress. Methods: One hundred two consecutive children, aged between 7 and 17 years, were interviewed about their pediatric intensive care unit (PICU) experience 3 months after discharge from a PICU. Principal measures were the ICU Memory Tool (a checklist of intensive care memories) and an abbreviated version of the Impact of Event Scale (a screen for post-traumatic stress disorder). Measurements and Main Results: In total, 64 of 102 (63%) children reported at least one factual memory of their admission and 33 of 102 (32%) reported delusional memories, including disturbing hallucinations. Traumatic brain injury was negatively associated with factual memory (odds ratio, 0.23; 95% confidence interval [CI], 0.09–0.58; P = 0.002). Longer duration of opiates/benzodiazepines was associated with delusional memory (odds ratio, 4.98; 95% CI, 1.3–20.0; P = 0.023). Post-traumatic stress scores were higher in children reporting delusional memories (adjusted difference, 3.0; 95% CI, 0.06–5.9; P = 0.045) when illness severity and emergency status were controlled for. Factual memory was not significantly associated with post-traumatic stress. Conclusions: This study indicates that delusional memories are reported by almost one-third of children and are associated both with the duration of opiates/benzodiazepines and risk of post-traumatic stress. More research is needed on the presence of delusional memories and associated risk factors in children receiving intensive care treatment.
Key Words: post-traumatic stress • hallucinations • memory • opiates • benzodiazepines
Adverse psychological reactions are increasingly being recognized in a significant proportion of both adults and children after treatment in critical care settings (1–3). In particular, symptoms of post-traumatic stress have been noted. To meet the criteria for post-traumatic stress disorder, a person needs to have experienced a traumatic event, usually involving a threat to life or physical integrity, and in addition be displaying three types of symptom: reexperiencing (e.g., flashbacks, intrusive images), avoiding reminders of the traumatic event (e.g., not talking about it), and hyperarousal (e.g., exaggerated startle response or irritability) (4). Although, intuitively, it would be expected that those patients who remember their time in intensive care would be at risk of higher levels of post-traumatic stress symptoms, this has not turned out to be the case in adults. An association has been found, however, between memory for delusional events at the time of admission (defined as nightmares, hallucinations, dreams, or the sense that someone was trying to harm the patient) and subsequent post-traumatic stress (5). This may explain the apparent anomaly that some patients who remember very little factual information about their intensive care hospitalization continue to be very distressed when reminded of it. The fact that pediatric intensive care unit (PICU) survivors are at higher risk of post-traumatic stress than other pediatric patients (3) and that the number of invasive procedures in the PICU is associated with distress at follow-up (6) has led to an assumption that traumatic experiences in the PICU cause post-traumatic stress. However, a study of recall for events in 38 children during admission found that only two-thirds remembered anything and the majority of their memories were neutral or positive (7). In summary, there is no information on rates of delusional memory in children in critical care settings. Furthermore, although a number of studies have reported post-traumatic symptoms in children, links with particular types of memory and subsequent distress have not been examined to date. The aims of this study were as follows: (1) to establish the nature and extent of factual and delusional memories, in a representative sample of children after PICU discharge; (2) to examine the relationships between both types of memory and demographic and medical variables; and (3) to investigate whether there was a significant relationship between the presence of either type of memory and subsequent rate of post-traumatic stress symptoms. It was hypothesized that a significant number of children would report delusional memories and that the presence of delusional memories would be associated with higher rates of post-traumatic stress symptoms. Some of the results of this study have been reported previously in the form of an abstract (8).
Over the course of 18 months, the families of consecutive surviving children older than 7 years were approached, 6 weeks after discharge, to take part in the research project. Children were excluded if the reason for admission was self-harm. Other reasons for exclusion were significant learning difficulties (defined as requiring special educational placement), readmission between admission and contact, discharge on long-term ventilation, palliative care, or address abroad. The setting was a 21-bed ICU in a tertiary children's hospital in an inner city area. Ethical permission for the study was granted by the hospital's ethics committee. Parents and children were provided with separate, age-appropriate information sheets about the study, and written, informed consent was obtained from the parents of all participating children. Families were given the option to be interviewed at home or in hospital and, when requested, interview dates were changed to fit around existing outpatient appointments. During the interviews, which were conducted by an experienced clinical psychologist (G.C.), children were asked about the aspects of the PICU admission they were able to recall, using a checklist of factual memories, delusional memories, and memories of feelings provided in the ICU Memory Tool (9); and their responses were audiotaped and transcribed verbatim. They also completed the Children's Revised Impact of Event Scale (10), which measures post-traumatic stress symptomatology. Demographic and medical data were obtained from the child's medical record. Illness severity was measured using the Pediatric Index of Mortality (11). Social deprivation was measured using the Townsend Deprivation Index, which is derived from U.K. Census data for a given geographical district and relates to four variables: car ownership, employment status, home ownership, and level of accommodation overcrowding (12). The ICU Memory Tool was developed for use with adult patients after intensive care admission and has good test–retest reliability and internal consistency (9). It has been adapted for use in a number of countries (13–15), but its use with children has not been previously reported. Patients were coded as having factual memory for admission if they recalled any of the factual memory subscale items and were coded as having delusional memory if they recalled any item on the delusional memory subscale. Post-traumatic stress symptoms were measured using the Children's Revised Impact of Event Scale. This is an eight-item screen for post-traumatic stress symptoms in children aged between 7 and 18 years, with established reliability and validity (10), which has been used with children who have experienced many types of trauma, including war, disaster, and road accidents (16–18). A cutoff score of 17 or greater has been found to classify correctly over 80% of children with a diagnosis of post-traumatic stress disorder (19).
Statistical Analyses Additional detail on methods is provided in the online supplement.
Sample Characteristics All children aged 7 years or over who were discharged between February 1, 2004, and July 31, 2005, were considered for inclusion in the study. The following exclusions were made: readmission (n = 29), death (n = 33), palliative care (n = 4), learning difficulties (n = 79), overseas address (n = 20), self-harm (n = 3), or long-term ventilation (n = 4). Of the 132 remaining eligible cases, 102 (77%) agreed to take part, of whom only 3 had previously been admitted to the PICU. The median time to interview was 3 months (range, 1.8–5.7 mo). Sample characteristics are given in Table 1.
Reasons for not taking part, which were provided by the families of 20 of 30 children, were as follows: child too ill (n = 9), child reported to remember nothing (n = 6), or child unwilling to be interviewed (n = 5). The children in nonrecruited families were older, and had shorter stays. Otherwise, the two groups were comparable in terms of demographic and medical variables (see Table E1 of the online supplement). In both groups, socioeconomically deprived families were overrepresented as compared with the general population. This distribution is consistent with national data for the United Kingdom (20).
Children's Memories
In total, 33 of 102 children (32%) reported having experienced at least one delusional memory. Two children described positive feelings associated with what they experienced, but in all other cases the delusional memories were experienced as highly disturbing, particularly in relation to their persistence and threatening content (see Table 3 for children's descriptions of their hallucinations). In all but two cases, the delusional memories originated in the PICU. These hallucinations were most frequently visual only (n = 16), but two children reported auditory hallucinations only; two reported visual and tactile hallucinations; three reported visual and auditory hallucinations; and one reported visual, auditory, and tactile hallucinations. In addition, two children reported a strong sensation that their parents had been replaced by impostors. Eleven children described multiple hallucinations, beginning near the end of their PICU stay and continuing for several days after discharge, interfering with their sleep.
Cronbach  was calculated for the three subscales used from the ICU Memory Tool, to check for internal consistency. Values were 0.94 for factual memories, 0.81 for memories of feelings, and 0.71 for delusional memories.
Data Transformation In addition, for the purposes of the proposed statistical analyses, length of stay and length of sedation were cut respectively at greater than and greater than or equal to 2 days (a median split in both cases to ensure roughly equal numbers in each group) to transform these data, which were not normally distributed, into categorical variables, because we could not make the assumption of a linear relationship with post-traumatic stress or presence of factual or delusional memory.
Associations with Factual Memory
When these three variables were entered simultaneously into a logistic regression, only traumatic brain injury remained as a significant predictor, reducing the likelihood of a child remembering anything factual about the admission by 77%. There was no independent effect of duration of opiates/benzodiazepines or of emergency admission.
Associations with Delusional Memory
In a logistic regression, children were nearly five times more likely to report a delusional memory if they were receiving opiates and/or benzodiazepines for 2 or more days, as compared with the remainder of the sample, after adjusting for length of stay. Length of stay was not significantly related to the presence of delusional memory, after controlling for duration of opiates/benzodiazepines. Furthermore, although report of delusional memory was highest in children receiving opiates/benzodiazepines for 2 days or more (23/48, 48%), a number of children who had been prescribed these medications for less than 2 days also reported delusional memories (10/54, 19%). All children reporting hallucinations in this study had received opiates, but this finding needs to be interpreted with caution, because there were very few children who were not prescribed opiates (8/102, 8%).
Post-traumatic Stress Symptoms The univariate associations between post-traumatic stress score and demographic and medical variables are given in Table 6. Two variables, namely illness severity and emergency status, were significantly associated with post-traumatic stress score.
Separate linear regression models were examined to assess the strength of the association with post-traumatic stress for (1) factual memory and (2) delusional memory, controlling for the two confounding variables given above (Table 7). No significant relationship was found between factual memory and post-traumatic stress. In contrast, however, children who reported any delusional memory had significantly higher post-traumatic stress scores than those who did not.
Finally, nonparametric statistics were used to compare the post-traumatic stress scores of children with both types of memory with those of children who reported delusional memory alone. No significant difference was found between groups (median [range], 12 [3–24] for those with both types of memory, vs. 7.5 [0–23], for those with delusional memory only; Mann-Whitney P = 0.71).
Nearly one in three children in this sample, who were all verbal and cognitively normal, reported delusional memories associated with their PICU admission. Furthermore, the hypothesis that delusional memories would be positively associated with higher rates of post-traumatic stress symptoms was confirmed. However, the suggestion that the coexistence of factual memory might be protective psychologically (5) was not upheld in this sample. It may be that children, as a result of lack of experience, are more disturbed by hallucinatory experiences and therefore less reassured by coexisting factual memories. Alternatively, it might just be that this finding was not replicated in a larger sample. Overall rates of post-traumatic stress were similar to those found in another study of 19 PICU survivors (3). A recent review of the prevalence of post-traumatic stress disorder in ICU survivors has recommended that researchers in this area keep exclusion criteria to a minimum in order to determine the scale of this problem for the population as a whole (22). Given that children with traumatic brain injury make up a significant proportion of admissions to the PICU studied, it was decided to include them. The finding that they were as likely as the other children admitted to suffer significant levels of post-traumatic stress, even though their factual memory for admission was poorer, is consistent with research in adults (23). This is an important finding, which calls into question the assumption that children who appear to remember nothing will not be distressed by what has happened to them, because it demonstrates that psychological distress is not simply a function of the degree to which patients remember their PICU hospitalization. The strength of the association between duration of sedation/analgesia and presence of delusional memory is consistent with the adult literature (15, 24) and with a number of reports of children suffering withdrawal symptoms, such as hallucinations, in association with the types of medication most commonly used in the PICU (e.g., benzodiazepines and opiates) (25–27). The finding that some patients reported delusional memories after being sedated for less than 2 days has also been reported in adults in intensive care (28), and is worthy of further investigation. The variation in sensitivity to medication may reflect genetic differences, such as those recently reported in adults found to be at higher risk of experiencing delirium in intensive care settings (29). This study has a number of strengths. The sample is believed to be the largest to date, both in terms of self-reported psychological symptoms and of reported memories in children after PICU admission. The paucity of information on children's direct experience in this situation is indicative of just how difficult these studies are to undertake. The majority of children admitted to PICU are nonverbal by virtue of their age, and this fact, together with parents' understandable reticence about overburdening them (30), makes recruitment difficult. It is likely that the flexibility regarding timing and site of interview (31) in this study enhanced the recruitment rate, which was high for a study of this type, and thereby minimized differences between recruited and nonrecruited children, which were minimal. In particular, the sample was representative of the population from which it was drawn with regard to groups often underrepresented in research in this area, such as the socially deprived (32) and children with chronic illness (33), increasing the validity of the results. Finally, this study is one of a very small number of studies of children's experience in this situation to make use of established standardized psychological measures, facilitating replication and comparison with groups of other traumatized children in future research. A number of limitations, however, should be acknowledged. First, there was no control group. Further research on the prevalence of delusional memories in other groups of hospitalized children is clearly needed. Also, there was no formal assessment of either withdrawal symptoms or delirium, reflecting the lack of validated measures for use with children of this age (21, 34), and, because this study reports only on the experience of children older than 7, the findings may not be applicable to younger children. Furthermore, in the interests of brevity, a screening measure was used to assess post-traumatic stress rather than a full diagnostic interview. Estimates of probable post-traumatic stress disorder should therefore be scaled down accordingly. Finally, given that delusional memory was retrospectively assessed at the same time as post-traumatic stress, causality cannot strictly be inferred. This association needs to be further explored in a prospective study.
Implications for Further Research The content of the memories described in this study was very similar to that reported by adults after critical care treatment (5, 35, 36). The children's hallucinations were more like those described by adults with organic delirium than those reported by patients with psychosis, in that they were predominately visual and frequently incorporated cues from the immediate environment. These phenomena cannot be said to constitute definitive evidence of delirium in children, because the interviews did not include assessment of inattention, which is regarded as essential for formal diagnosis of this condition (38). However, they are strongly suggestive of it and indicate that more research is required into the assessment of delirium across the age range, given its association with morbidity and mortality (39). A recent study of 40 pediatric cases has queried the applicability of the adult psychiatric diagnostic criteria in delirium in the PICU setting (34). To examine whether the associations between delirium and mortality and morbidity found in adults also apply to children, it will be necessary first to develop a measure of delirium appropriate for use with this population. It may be possible to adapt the Delirium Rating Scale (40), which has been used with children in a psychiatric setting (41), or either the Confusion Assessment Method for the Intensive Care Unit (42) or the Intensive Care Delirium Screening Checklist (43), both of which have been designed specifically for use with adults in critical care settings. However, the diagnosis of this condition in young (and particularly preverbal) children, who make up the majority of admissions to the PICU, will be a particular challenge. Once a validated measure of delirium is available, it will be possible to determine whether symptoms can be prevented or reduced by manipulating variables that have been shown to affect delirium in adult and which are amenable to experimental manipulation, such as sedation and the ward environment. Promising avenues for future research include the following: changing the type of sedation (44); changing the mode of delivery either by altering rates of weaning (45) or by interrupting sedation, a strategy that has been associated with a reduction in delirium and subsequent post-traumatic stress in adults (46); and altering the child's immediate environment by simplifying physical surroundings, demarcating day and night and providing him or her with regular information designed to reassure and reorientate (47, 48). Further evaluation of sedation regimes in terms of psychological as well as physical aftereffects, in adults and children, is also warranted, given the dearth of research in this area (49).
Implications for Clinical Practice
The authors thank Helen Tighe for help with data collection, Jill Rolfe for transcribing the interviews, and the families who took part for sharing their experiences without financial compensation.
Supported by a Leading Practice Through Research Award from The Health Foundation, UK. This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200706-857OC on January 31, 2008 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form June 12, 2007; accepted in final form January 28, 2008
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
