Delirium is a frequent, underrecognized, complex neuropsychiatric syndrome. It is an acute state that is believed to be organically based. According to the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), characteristics of delirium include rapid onset; fluctuating course; and disturbances of attention, memory, thought, perception, and behavior that "do not occur within the context of a severely reduced level of arousal such as coma."1, 2 While delirium in hospitalized adults, and particularly older adults, is well documented, affecting between 42% and 80% of this population,3 its occurrence among hospitalized children is less clear. A recent literature review found the reported prevalence of delirium among hospitalized children to range from 13% to 28%; one study, which used "clinical suspicion" as its index, reported just 5%.4 Yet one retrospective study of children ages seven to 17 years who had been hospitalized found that more than 30% recalled having hallucinatory experiences during their stay.5 In almost all cases (94%), these delusional memories were "highly disturbing," with persistent and threatening content. Except in two cases, the location associated with these memories was the pediatric ICU (PICU).
Delirium in adults has been associated with longer hospital stays, poorer functional outcomes, and high mortality rates.6 Although there have been fewer studies among children, one study of pediatric psychiatric patients found that delirium was associated with prolonged hospital stays and high mortality rates.7 Factors that increase the likelihood of delirium in hospitalized patients include disturbed sleep, discomfort, dehydration, limited access to food and fluids, isolation, and immobility.8 In other words, the acute care hospital environment itself may be a precipitating factor. The specific pathophysiology of delirium remains largely unknown. That said, research in traumatic brain injury indicates that the immature, developing brain has different responses to oxygen deprivation and cytokine release than the adult brain, making children potentially more susceptible to occurrences of delirium.9, 10
In children, acute delirium can present in one of three subtypes: hypoactive, hyperactive, or mixed (also known as emerging or veiled) delirium.11 Hypoactive delirium, thought to be related to the poorest outcomes, is characterized by inattention and lethargy. Hyperactive delirium is characterized by heightened psychomotor activity (such as agitation and pulling out catheters) and hallucinations. In mixed delirium, the patient fluctuates between the hyperactive and hypoactive subtypes. The patient may present with disturbed cognition or attention (or both) and severe anxiety "often accompanied with moaning and restlessness," yet without clear agitation or lethargy.11
Estimates of the economic burden of delirium vary. One U.S. study conducted among elderly patients found that delirium accounted for added health care costs of between $16,303 and $64,421 per patient annually, resulting in an overall annual burden of $38 billion to $152 billion.12 A study in the Netherlands by Smeets and colleagues found that pediatric delirium increased the duration of PICU stay by an average of 2.39 days, which corresponded to an increase of 1.5% in direct medical costs.13 And in a study of costs associated with pediatric delirium, Traube and colleagues found that median total PICU costs were almost four times higher for patients with delirium than for those without ($18,832 versus $4,803, respectively).14 After controlling for age, sex, illness severity, and PICU length of stay, delirium was associated with an 85% increase in PICU costs.
It's likely that a better understanding of the risk factors and characteristics of pediatric delirium would lead to earlier identification, referrals to psychiatric consultation when appropriate, and the development of delirium reduction and prevention programs. These approaches have been shown to improve outcomes,12 and thus might also reduce costs.
Nurses, particularly pediatric nurses, are well positioned to recognize the risk factors for and characteristics of pediatric delirium and to address ways to reduce such risk. Thus, we framed the research question as, "How do nurses recognize the risk factors and characteristics of delirium in children?" The purpose of this study was to examine the evidence regarding the risk factors for and characteristics of acute pediatric delirium in hospitalized children.
METHODS
Design. A descriptive epidemiological systematic review method was used. Descriptive epidemiological studies examine the distribution of a disease or condition for the purposes of establishing prevention or management programs (or both) and informing their planning and evaluation.15 Such studies do so by describing the characteristics of people (race, age, and sex, for example), place (geographic location), and time (a specific year or span of time), rather than by examining the effects of an intervention. For this review, epidemiological study designs-including prospective and retrospective cohort studies with or without control groups, and cross-sectional and case-control studies that addressed nonemergent delirium (delirium unrelated to the administration of anesthesia, for example)-were sought.
The target population was hospitalized children, from birth through 21 years, who had been diagnosed with delirium. Studies were selected for inclusion if they involved this target population, used a delirium screening method, described risk factors for or characteristics of delirium episodes (or both), and were published in English. Single case studies, intervention studies, and qualitative studies, as well as non-English publications, were excluded.
Selection of studies for inclusion. A four-step search strategy for articles published in English from 1990 through 2015 was used. First, we searched MEDLINE and CINAHL using an initial list of keywords generated from a concept map based on the research question; this method also allowed us to identify additional keywords. Second, in consultation with a health sciences librarian, we used the expanded list of keywords to search 20 databases (see Electronic Databases Searched). Third, the reference lists of all identified articles were searched to identify additional studies. Lastly, we searched the selected Web of Science articles by author names to determine if any other relevant publications existed.
Two reviewers (CH and SP) screened all articles that were identified by title and abstract, using the inclusion criteria. The resulting list was then reviewed by the entire review team. Each article selected for retrieval was assessed independently by teams of two reviewers (all authors participated) for methodological quality, using the appropriate critical appraisal tool from the Joanna Briggs Institute Reviewers' Manual.16 Each team of reviewers compared the results of this assessment and jointly decided whether to include or exclude each study. Any disagreement was discussed with the entire team until consensus was achieved. Studies were included if they met at least 80% of the methodological criteria in the critical appraisal tool. The appraisal criteria included questions regarding sample selection, sufficiency of sample descriptions, inclusion criteria, assessment of outcomes, reliability of outcome measures, and appropriateness of statistical measures used. (For details regarding the full search strategy, please contact the corresponding author.)
Data were extracted from articles using a pro forma data extraction tool. For each study, the following data were extracted: information on the sample (age, sex, diagnosis, preexisting conditions, sample size), study design, study location (PICU, non-PICU), delirium incidence, delirium screening method used, identified symptoms of delirium, and time of occurrence of such symptoms.
RESULTS
Of the 301 articles originally identified, 52 appeared to meet the inclusion criteria based on the abstract or title. The review team subsequently excluded 31 of these. Reasons for exclusion were as follows: six were single case studies; five used a duplicate sample; five were literature reviews; five were not full text (two were letters to the editor, three were abstracts only); four studied the wrong population (postdischarge patients); three were non-English publications; two were systematic reviews (one a review of interventions, the other a review of diagnostic tools); and one was about the validity and reliability of tool development. Our review examined the remaining 21 quantitative studies, of which 11 were cohort studies and 10 were case series or case reports. Of these 21 studies, 10 were conducted in the United States, seven in the Netherlands, three in India, and one in South Africa. (For a flow diagram of the search strategy, see Figure 1.) Per the Joanna Briggs Institute's levels of evidence,17 all 21 studies were graded as level 3 or 4.
Overall, the studies involved 2,616 children. Because of differences in how studies reported delirium diagnosis, it was not possible to determine an overall prevalence rate. Nor did all studies report sex. In the 10 studies that did report both sex and a diagnosis of delirium, 54% of the 633 subjects were boys and 46% were girls, although this difference was not significant. The majority of subjects were PICU patients and had been diagnosed with delirium using a delirium rating scale (22%), by a psychiatric consultation liaison team (14%), or both (64%).
Findings from the 21 studies that met the inclusion criteria are presented according to the purposes of this review: to identify delirium risk factors in children from birth through 21 years of age and to identify characteristics of acute pediatric delirium. (See Table 1 for full details.7, 11, 13, 18-35) We defined risk factors as those conditions and circumstances that predispose a child to the development of acute delirium. We defined characteristics as the individual symptoms and patterns of symptoms found in hospitalized children with acute delirium.
Risk factors for acute pediatric delirium.Age. Using the Cornell Assessment of Pediatric Delirium (CAPD) scale to assess PICU patients, Traube and colleagues found that delirium prevalence was lowest in children older than 13 years (3.6%).18 But Schieveld and colleagues, using the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR) criteria for delirium with evaluation by a child neuropsychiatrist, found delirium prevalence ranging from 3% in children ages three years or younger to 19% in those ages 16 to 18 years.11
Sex. As noted above, 10 studies reported both sex and a diagnosis of delirium, with delirium occurring overall in more boys than girls.11, 18-26 For example, Traube and colleagues reported that, of 111 PICU patients, the prevalence of delirium was 18.8% in girls and 21.7% in boys.18 And Smith and colleagues found that, of nine patients who developed delirium, 22.2% were girls and 77.8% were boys.24 Similarly, in the other eight study populations, more boys than girls were diagnosed with delirium. This suggests that male sex may be a risk factor.
Anxiety. Jones and colleagues reported that children with preexisting anxiety were more likely than others "to exhibit apprehension toward staff, inattentiveness and distractibility, motor agitation when awake, and misbehavior and disobedience."27 They found that such children were at higher risk for developing psychological trauma and behavior problems that would require intervention, although they didn't specifically cite delirium. Schieveld and colleagues noted that anxiety was associated with a presentation of mixed delirium.11
Developmental delay. Traube and colleagues found that children with developmental delay were diagnosed with delirium almost three times as often as children without such delay.18 In a pilot study of 50 PICU patients by Silver and colleagues, the prevalence of delirium was 29%, and the researchers noted that 24% of the overall sample had a prior diagnosis of developmental delay.28
Mechanical ventilation. Van Dijk and colleagues investigated pediatric delirium in 29 PICU patients admitted over an eight-year period.26 In their sample, 79% were mechanically ventilated. In a larger study of 877 PICU patients, Schieveld and colleagues found that 85% of those with delirium were on mechanical ventilation.29 Similarly, in a study of 147 PICU patients, Smeets and colleagues found that 84% of those with delirium were on mechanical ventilation, compared with 42% of those without delirium.13 Janssen and colleagues, studying 154 PICU patients, reported that 53.8% of those with delirium were mechanically ventilated, compared with just 11.7% of those without delirium.21 And in the study by Traube and colleagues, the researchers found that delirium prevalence was lowest (5.2%) among children who were not on respiratory support.18
A note on other potential risk factors. Our focus in this review was on noniatrogenic, nonpharmacological risk factors and characteristics that fit within the epidemiological framework of person, place, and time. Thus, we did not consider factors that may be secondary to the patient's acute illness or iatrogenic in nature, including metabolic disturbances, acute infection, hypoxemia, anemia, acidosis, and hypotension. Some commonly used medications that are considered precipitating risk factors for delirium, such as opioids, anxiolytics, tricyclic antidepressants, and corticosteroids, were also excluded.
Characteristics of acute pediatric delirium identified in the 21 studies included agitation, disorientation, hallucinations (primarily visual), inattention, and sleep-wake cycle disturbances. Aggression, irritability, restlessness or fidgeting, and impaired memory were among others noted. Jones and colleagues reported that such characteristics were significantly more common in children in pediatric or cardiovascular ICUs than in children in general units.27 In a case series of very young children with neuroblastoma, Traube and colleagues described a lack of sustained sleep, agitation (such as thrashing in bed), and a lack of eye contact as characteristics of postoperative delirium.30
Patterns of delirium symptoms in children are similar to those seen in adults, especially older adults. In seven studies involving a total of 447 pediatric patients and use of the Delirium Rating Scale Revised-98 (DRS-R-98)-which incorporates the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, criteria and is often used to assess delirium in adults-all of the children exhibited characteristics of delirium according to the scale.7, 21, 23, 31-34 This suggests that the presentation of delirium episodes in children and adolescents is similar to that in adults. In one study of 38 patients by Grover and colleagues, all of the subjects had sleep-wake cycle disturbances and disorientation, and a majority had impaired attention (89.5%), impaired short-term memory (84.2%), agitation (68.4%), and lability of affect (rapid changes in emotion unrelated to external events; 60.5%).20 Karnik and colleagues reported that delirium symptoms observed in two hospitalized adolescents were comparable and included impaired attention, agitation, periods of confusion, memory loss, and sleep-wake cycle disturbances.32 In a study of 110 children by Turkel and colleagues, delirium symptoms included sleep-wake disturbances, motor agitation, altered thought process, disorientation, inattention, short- and long-term memory problems, and visual-spatial difficulties.34
There was some evidence of differences in the presentation of pediatric and adult delirium. Another study by Grover and colleagues, conducted among 30 children ages eight to 18 years, concluded that although the pattern of delirium in this population was similar to that seen in adults, children had a higher frequency of lability of affect.19 Leentjens and colleagues, comparing delirium symptoms in 46 children and 119 adults, reported that children tended to have more severe perceptual disturbances, hallucinations, agitation, and lability of affect.23 But children were less likely to exhibit sleep-wake cycle disturbances.
DISCUSSION
While the precise mechanism in the development of acute delirium is unknown, the results of this review indicate that children who are more severely ill or developmentally delayed are at higher risk. Children who are in a PICU, mechanically ventilated, and male may also be at greater risk, as are those with a high Pediatric Risk of Mortality II (PRISM II) score (at or above the 60th percentile).
Overall, acute pediatric delirium presents with disturbances in behavior, circadian rhythm, cognition, consciousness, and perception (see Characteristics of Pediatric Delirium). The characteristics of pediatric delirium were found to be wide ranging. The five most common characteristics overall were agitation, disorientation, hallucinations (primarily visual), inattention, and sleep-wake cycle disturbances (see Table 220, 25, 26, 31, 35). Symptoms reflective of the hyperactive type of delirium were the most prevalent. These include aggression, agitation, hallucinations, and restlessness or fidgeting.
Early recognition of acute pediatric delirium can be enhanced by the routine use of valid screening tools in PICUs. Such assessment is an important basic element in timely recognition, management, and prevention strategies. Since the clinical course of delirium fluctuates, screening may be most useful if done at least once per shift. It stands to reason that frequent screening, leading to earlier detection and appropriate treatment, could also shorten lengths of stay. The CAPD, a recently developed eight-question rapid assessment tool, has the potential to facilitate early recognition.18 This tool has been found to be comparable to the use of DSM-IV-TR criteria and psychiatric evaluation in recognizing acute pediatric delirium.28 The CAPD uses age-based developmental anchor points (newborn; four, six, eight, and 28 weeks; and one and two years), and can be completed by nurses in less than five minutes.18 It has shown excellent reliability and validity, and provides a structured, developmentally informed approach to the screening and assessment of delirium in children.18, 35
Given that delirium has a rapid onset and fluctuating course, it's essential to use delirium screening tools first at the time of admission to establish a baseline for comparison with later status. As Hatherill and colleagues stated, the "temporal aspect" of cognitive changes "underscor[es] the need for a longitudinal assessment and serial interviews."36 This is especially important for children with developmental delays or preexisting anxiety, as these conditions have been found to be risk factors for the development of delirium.18, 27 The PRISM II score has also been found to be predictive of delirium for children scoring in the 60th percentile or above.13, 29
It's useful to consider the similarities between the characteristics of delirium in children and the elderly in the hospital setting. These include rapid onset, aggression, agitation, impaired attention, irritability, and memory disturbances.7, 19, 23, 25 Two studies found that in children, symptoms were more often seen at night.19, 26 In a study comparing symptoms in children and older adults, Grover and colleagues noted that children had more severe lability of affect, but less severe language, short-term memory, and visual-spatial disturbances.19 Furthermore, it's important to note that Traube and colleagues found that parents were often aware of their child's cognitive and behavioral changes before the hospital staff were.18
Implications. The clinical signs of delirium vary and are often dependent on a patient's age and, among infants and children, developmental stage. Of all health care professionals, nurses interact most closely and often with patients, and are thus vital to both early recognition and prevention of delirium in hospitalized children.
Delirium as a potential cause of agitation and sleeplessness must be considered. Indeed, Schieveld and Leentjens found that in severely ill young children, psychiatric symptoms and behavior changes were more likely due to delirium than to a psychiatric disorder.33 The early recognition and management of pediatric delirium may help prevent unwarranted laboratory testing and imaging studies, which can cause children and families unnecessary distress, as well as increasing hospital costs.
Measures to support and maintain a child's usual sleep-wake cycles include fostering a routine sleep schedule, reducing ambient noise, and lowering light levels at night.37 Nurses and other hospital staff need to ensure that loud discussions don't take place near a sleeping child and that any unnecessary noise-producing equipment is shut off. Usual age-appropriate comfort measures are also important. These may include providing a pacifier or other favorite toy and singing songs to lessen fear and anxiety during both routine and nonroutine procedures, and encouraging parents to physically and verbally support and comfort their child.38
Facilitating parental presence at the bedside and listening to parents' concerns about changes in their child's behavior are also important for the early recognition of delirium. Parents may be the first to observe changes in their child's cognition or behavior that are symptomatic of delirium. Moreover, parental presence at the bedside may enhance the child's comfort and lessen anxiety, and might prevent delirium from developing. Parents need education regarding delirium prevention, treatment, and follow-up care, including the fact that delirium symptoms may persist after discharge.5, 37 Emotional support for parents who witness the often upsetting behaviors of a child with delirium is essential. Because hallucinations are a commonly reported characteristic, staff education on the potential for delusional activity may be needed. Lastly, increased psychological monitoring and support for children with delirium and their families should be considered, as well as routine psychiatric consultation for children at risk for delirium.
Limitations. Per the Joanna Briggs Institute's levels of evidence,17 all 21 studies were graded as level 3 or 4, meaning that our findings must be viewed in light of their limitations. For example, several studies used a retrospective design, inferring data from recorded facts or narratives in the medical record. Because of the studies' low levels of evidence grades and varying methods of assessment, a meta-analysis was not possible. This review is further limited by its observational nature, which does not allow for causal attribution. The severity of delirium episodes could not be considered as an outcome measure because only a few of the studies assessed delirium using a standardized rating scale.
CONCLUSIONS
Acute delirium is reportedly among the top six preventable conditions in hospitalized older adults,39 and may rank similarly high in the pediatric population. Children in acute care are especially susceptible to developing delirium. This systematic review examined the evidence regarding risk factors for and characteristics of pediatric delirium. It is our hope that the findings will lead to more timely recognition and enhance prevention.
In accordance with the epidemiological framework of person, place, and time, we found the following: risk factors may include male sex, preexisting anxiety, developmental delay, and being on mechanical ventilation (person); the five most common characteristics of pediatric delirium include agitation, disorientation, hallucinations (primarily visual), inattention, and sleep-wake cycle disturbances; patterns of delirium symptoms in children are similar to, but not exactly the same as, those found in adults and the elderly; among hospitalized children, the most susceptible were those admitted to a PICU (place); and, although the symptoms of pediatric delirium fluctuate, many episodes occurred at night (time). (See Table 3.) A key finding of this review is that delirium is multifactorial, related to treatment (specifically, mechanical ventilation) and to a hospital environment (such as the PICU) that deprives patients of normal sleep-wake cycles and familiar routines.
REFERENCES