A physiologically unstable child is exposed to numerous procedures and technology in the Pediatric Intensive Care Unit (PICU) and a disturbed environment(1-3). This makes children vulnerable to stress experienced at a physical, psychological and social level(4,5). Thus in a temperamentally ‘at risk’ child, admission to PICU can be akin to psychological trauma(6). It is presumed that sicker the child, greater the interventions and more the psychological trauma(7). This trauma can have repercussions in the form of a post-traumatic stress disorder(5,8). Children may not verbalize such distress(3). Therefore their anguish may not be detected and the outcome of unresolved psychological trauma might be detrimental(3,7).

Attention is seldom directed towards the emotional well being of children in PICUs in India. If care of children in the PICU is to be holistic, there is a need to investigate presence of psychological morbidity in survivors so that remedial measures can be introduced. The objective of this study was to determine whether PICU hospitalization is associated with adverse psychological outcome; to identify factors contributing to psychological distress; and to determine whether the effects continue beyond the period of hospitalization.

A prospective cohort study was performed over 8 months from November 2003 through June 2004 after approval by the Institutional Ethics Committee. Almost all admissions to PICU are from the emergency ward and pediatric wards and are unscheduled. Neonates and surgical patients are not admitted. There is no formal counseling to prepare children for PICU admission. One guardian is permitted to stay in PICU. Soft toys and piped music are provided for recreation. For patient care, five senior staff members supervise 12 resident doctors working in shifts of eight hours. Two nurses trained in pediatrics work in shifts of eight hours each. Patient to doctor ratio is 2:1 and patient to nurse ratio is 4.5:1. The three-year average of annual number of admissions was 516. The bed occupancy through-out the year was 100% with average stay of 6.2 days. The average annual mortality varies from 25% to 29%.

Consecutive children admitted to the PICU and fulfilling inclusion criteria (age ³5 years, absence of neurological abnormalities precluding interview and duration of stay of at least 48 hours) were enrolled as Cases. Readmitted survivors were excluded. Controls were age and sex matched children from the pediatric ward without previous PICU admission and ward stay of at least 48 hours. Demographic data, duration of hospitalization and nature of disease was recorded. Socioeconomic status was determined using the modified Kuppuswamy scale(9). Pediatric Risk of Mortality-III (PRISM) score obtained at 24 hours of admission was used to grade severity of illness in patients to correlate with the psychological outcome(10). Therapeutic Intervention Scoring System (TISS-76)(11) was also recorded for each patient to objectively record intensity of critical care.

Patients were interviewed by SBB under the supervision of PS within 24 hours of discharge from PICU. Controls were interviewed on the day of discharge. A second interview was scheduled a month later. The Temperament Measurement Schedule is a standardized itemized questionnaire, which was administered to the parent/guardian to assess premorbid temperament(12). Five domains were judged: sociability, emotionality, energy, distractibility and rhythmicity. Impact of Event Scale (IES), a questionnaire of 15 items was administered to determine psychological outcome. It explores intrusive symptoms (nightmares, imagery, insomnia and emotional reactions) and avoidance behavior (numbing, avoidance and staying away from reminders) and together provides a subjective score. This has been validated for children over eight years of age(13). Impairment on IES was graded as follows: 0 to 8 – sub clinical, 9 to 25 – mild, 26 to 43 – moderately severe and >44 – severe. The Birleson depression Self-Rating Scale(14) is an 18-item questionnaire exploring symptoms like sleep, appetite, pain, interest and mood. A score of ³13 establishes presence of depression. The scale is validated for children 7-13 years old. Self-esteem was determined by the ten-item Self-Esteem Scale (SES) of Rosenberg(15). A score of <23 denotes low self-esteem(16). All questionnaires were administered in the vernacular and simplified by converting Likert scale to visual analogues. Children in either group demonstrating abnormalities were referred to the psychiatric service.

Amongst 337 PICU admissions, 34 children fulfilled inclusion criteria. Four patients could not be interviewed due to absence of recollection of PICU stay or language barrier. The final study population was 30 Cases (10 males, 20 females; M:F ratio 1:2) and an equal number of Controls (16 males, 14 females; M: F ratio 1:0.87). Their mean age was 8.5±5.2 years (range: 5-12) and 8.9±6.2 years (range: 6-12), respectively. Cases and controls were comparable with respect to demographic parameters and socioeconomic status. Duration of ward stay (11±11.8 days) exceeded that of PICU (6.6±4.9 days) but the difference was not significant (P=0.064). The pattern of disease differed signifi-cantly between the two study groups: infectious disease (including five patients with tetanus), cardiovascular disease like complications of rheumatic heart disease and neurological disorders were most frequent accounting for 33%, 23% and 20% of the cases respectively. Genitourinary (nephrotic syndrome or glomerulonephritis), hemato-logical and respiratory diseases were common amongst Controls (27%, 20% and 17% respectively).

The mean PRISM score was 3.27±3.44 (maximum score: 18). Means of the TISS score in Cases was 8.97±4.22 (range: 4-18) [highest score] and 5.85±2.80 (range: 3-15 [average score] while that in Controls was 1.4±1.59 (range: 0-8) [highest score] and 0.9±0.74 (range: 0-3) [average score]; both scores were significantly greater in Cases than in Controls (P<0.05).

The premorbid temperament of children in both groups was comparable in all domains except energy. One case and seven controls had high energy levels and five cases and one control scored low. However, this difference did not influence the psychological outcome.

The mean IES score in cases was 1.56±2.18, while controls had a significantly lower score of 0.10 ±0.40 (P=0.002, unpaired t test). When individually analyzed, the mean intrusion score in cases (1.1±1.83) was significantly higher than controls (0.10 ±0.40, P=0.005) but avoidance scores (0.15±0.61 and 0.0 respectively) did not differ significantly. Higher proportion of cases (13, 43%) developed intrusive thoughts as compared to controls (2, 6.7%) (P=0.001). The intrusive thoughts reported were images related to PICU hospitalization (12 cases, 1 control), bad dreams (9 cases, 2 controls), strong feelings (5 cases), thoughts (3 cases), difficulty falling asleep because of images or thoughts related to PICU (1 case). The score for each item was 1 (rarely experienced) in all except 2 patients who scored 3 (sometimes experienced). Presence of intrusive thoughts correlated with the TISS score (Spearman’s analysis; coefficient of correlation 0.507, P=0.004 for highest score and P<0.001 for average score). Avoidance and low self-esteem was present in two cases and none of the controls while depression was present in 4 cases (13.3%) and one control (3.3%). These differences were not significant. Cases and controls were comparable with respect to scores on the self-esteem and depression scales and the TISS scores did not correlate with these parameters.

Age, socioeconomic status, education, nature of illness, severity of disease, duration of PICU hospitalization and premorbid temperament did not influence the psychological outcome.

Seventeen cases and 22 controls were available for follow-up. Two cases had expired after discharge from PICU and 11 did not follow-up. In the control group, eight did not follow-up. Overall, the Cases and Controls were now comparable with respect to the psychological scores (mean IES-Intrusion score: Cases 0.28±0.57 versus Controls 0.05±0.22, P=0.097; mean depression scores: Cases 5.00±3.74 versus controls 5.29±1.98, P=0.766). Moreover, it was observed that in the cases, scores on all the scales were lower as compared to those obtained immediately following discharge from PICU (Table I).

Early studies on behavioral response of children to PICU hospitalization have identified abnormalities like anxiety, withdrawal, fear, anger and hostility(7). Such subjective studies may not be relevant to the current milieu of PICU where sicker children are surviving due to intensive and invasive monitoring and management supported by increasing reliance on equipments.

Our study convincingly documents psychological stress in children surviving PICU. We hypothesized that negative experiences in PICU could lead to psychological stress in a child ill equipped with mature coping mechanisms(3,7). Children evidently remember experiences in PICU. At least one bad experience was recorded in 50% of children in a PICU in Mumbai(2). Proportion of children expressing negative feelings about PICU was 24%(2). Such feelings may be underestimated or neglected, as children would express such feelings only if asked(3).

PICU survivors in the present study developed intrusive thoughts despite lower TISS scores. Rennick, et al.(7) demonstrated that children in PICU exposed to greater number of invasive procedures (cut-off score of ³121) constitute a high-risk group for adverse psychological sequelae. Our study suggests that even minimally invasive interventions adversely impact psychological well-being of PICU survivors. However, scores on follow-up were comparable with controls indicating that psychological abnormalities resolved within a month. Thus, children in our study probably had an acute stress reaction not amounting to post traumatic stress disorder(16). Early resolution of psychological stress could probably be due to low intensity of invasiveness. Higher scores on the invasiveness scale recorded by Rennick, et al. could have led to a greater impact persisting beyond six weeks(7,8).

Rennick, et al.(8) found no significant difference between PICU and ward hospitalization. This was attributed to the effect of sedation or analgesia administered to PICU patients in diminishing the impact. In India, sedation/ analgesia may not be used as liberally. This could have contributed to the difference in scores obtained between cases and controls in the present study.

The study does have some limitations. The IES scale is an adult scale also used in children(13,17). The depression scale is created for children older than seven years. Despite these limitations it is clear that even low intensity of interventions in PICU leads to acute albeit transient distress. It is desirable to prevent this by ensuring positive recollections of PICU stay and judicious use of sedation and analgesia. PICU staff must recognize manifestations of psychological distress(5). Services of a child psychologist or psychiatrist must be ensured. A warm and empathetic atmosphere, play therapy and permitting a parent especially the mother to stay with the child would be fitting(6,18).

The authors thank Dr Nilima Kshirsagar, Dean of Seth GS Medical College & KEM Hospital for granting permission to publish the paper using hospital data, Dr Anuradha Sovani, Professor of Psychology, University of Mumbai for guidance in scoring and interpretation of patient data and Dr Abhiram Kasbe, Associate Professor, Department of Preventive & Social Medicine, TN Medical College and BYL Nair Hospital, Mumbai for statistical analysis.

Contributors: MNM, SBB and HRS designed the study, interpreted data and results, performed literature review and drafted the manuscript. SBB collected the data and PS supervised data recording and interpretation. MST contributed to patient management and provided inputs for data interpretation and manuscript preparation.

Funding: None

Competing interests: The paper was awarded first prize at the 1st Asia Pacific Congress on Pediatric Critical Care and 7th National Congress on Pediatric Critical Care at Pune, India, October 2005.