ABSTRACT
BACKGROUND: Cardiorespiratory arrests are rare in paediatric intensive care units, yet intensive care nurses must be able to initiate resuscitation before medical assistance is available. For resuscitation to be successful, instant decision-making, team communication, and the coordinating role of the first responsible nurse are crucial. In-house resuscitation training for nurses includes technical and nontechnical skills. OBJECTIVES: The aim of this study was to develop a valid, reliable, and feasible assessment instrument, called the Professional Assessment Tool for Team Improvement, for the first responsible nurse's technical and nontechnical skills. METHODS: Instrument development followed the COnsensus-based Standards for the selection of health Measurement Instruments guidelines and professionals' expertise. To establish content validity, experts reached consensus via group discussions about the content and the operationalisation of this team role. The instrument was tested using two resuscitation assessment scenarios. Inter-rater reliability was established by assessing 71 nurses in live scenario sessions and videotaped sessions, using intraclass correlation coefficients and Cohen's kappa. Internal consistency for the total instrument was established using Cronbach's alpha. Construct validity was assessed by examining the associations between raters' assessments and nurses' self-assessment scores. RESULTS: The final instrument included 12 items, divided into four categories: Team role, Teamwork and communication, Technical skills, and Reporting. Intraclass correlation coefficients were good in both live and videotaped sessions (0.78-0.87). Cronbach's alpha was stable around 0.84. Feasibility was approved (assessment time reduced by >30%). CONCLUSIONS: The Professional Assessment Tool for Team Improvement appears to be a promising valid and reliable instrument to assess both technical and nontechnical skills of the first responsible paediatric intensive care unit nurse. The ability of the instrument to detect change over time (i.e., improvement of skills after training) needs to be established.
Subject(s)
Nurses , Resuscitation , Child , Clinical Competence , Communication , Humans , Intensive Care Units, Pediatric , Patient Care Team , Reproducibility of Results , Resuscitation/educationABSTRACT
PURPOSE: The quality of resuscitation and effective leadership are decisive for the outcome of a resuscitation. Nurses are usually the first responders upon cardiac arrest. Therefore, we started the "proficiency check" project, which aims to improve nurses' resuscitation and teamwork skills. This article describes the effectiveness of the proficiency check and nurses' experiences with it. DESIGN AND METHODS: This study was done among intensive care nurses working on a pediatric ICU (PICU) in the Netherlands. It was designed as a mixed-methods study combining a quantitative and a qualitative approach. Quantitative data were obtained through a pre-posttest comparison of nurses' resuscitation and teamwork skills, in a simulation setting. Qualitative data on nurses' experiences were collected through semi-structured individual interviews. RESULTS: Both resuscitation and teamwork skills improved significantly. In 39 nurses (32%), the improvement of both resuscitation and teamwork skills after the intervention was large (effect size >0.8). The experiences of nurses regarding the proficiency check were diverse: on the positive side, increased knowledge and confidence were reported, whereas negative experiences related, among other things, to stress and anxiety. CONCLUSIONS: Resuscitation and teamwork skills of PICU nurses can be enhanced by the 'proficiency check' studied here. This simulation-based training can be further improved by incorporating the nurses' experiences. PRACTICE IMPLICATIONS: A simulation-based assessment for resuscitation may play an important role in a PICU, and possibly for other skills and in other settings as well. Particular attention should be paid to the stress that many nurses experience due to skills assessment.
Subject(s)
Nurses , Simulation Training , Child , Clinical Competence , Humans , Intensive Care Units, Pediatric , Netherlands , ResuscitationABSTRACT
It has been identified that one of the dosages listed within the above article was incorrectly abbreviated during the production process. This is now presented correctly in this article.
ABSTRACT
We studied the occurrence of adverse events (AEs) in low-risk non-survivors (LNs), compared to low-risk survivors (LSs), high-risk non-survivors (HNs), and high-risk survivors (HSs) in two pediatric intensive care units (PICUs). The study was performed as a retrospective patient record review study, using a PICU-trigger tool. A random sample of 48 PICU patients (0-18 years) was chosen, stratified into four subgroups of 12 patients: LNs, LSs, HNs, and HSs. Primary outcome was the occurrence of AEs. The severity, preventability, and nature of the indentified AEs were determined. In total, 45 AEs were found in 20 patients. The occurrence of AEs in the LN group was significantly higher compared to that in the LS group and HN group (AE occurrence: LN 10/12 patients, LS 1/12 patients; HN 2/12 patients; HS 7/12 patients; LN-LS difference, p < 0.001; LN-HN difference, p < 0.01). The AE rate in the LN group was significantly higher compared to that in the LS and HN groups (median [IQR]: LN 0.12 [0.07-0.29], LS 0 [0-0], HN 0 [0-0], and HS 0.03 [0.0-0.17] AE/PICU day; LN-LS difference, p < 0.001; LN-HN difference, p < 0.01). The distribution of the AEs among the four groups was as follows: 25 AEs (LN), 2 AEs (LS), 8 AEs (HN), and 10 AEs (HS). Fifteen of forty-five AEs were preventable. In 2/12 LN patients, death occurred after a preventable AE. CONCLUSION: The occurrence of AEs in LNs was higher compared to that in LSs and HNs. Some AEs were severe and preventable and contributed to mortality. What is Known: ⢠59-76% of all PICU patients encounter at least one adverse event during their PICU stay. ⢠It is unknown if adverse events play a role in death of low-risk PICU patients. What is New: ⢠In low-risk PICU non-survivors, occurrence of adverse events is higher compared to low-risk PICU survivors and to high-risk PICU non-survivors. ⢠Severe and preventable adverse events occur in low-risk PICU non-survivors, some contributing to mortality.
Subject(s)
Intensive Care Units, Pediatric/statistics & numerical data , Medical Errors/statistics & numerical data , Survivors/statistics & numerical data , Adolescent , Child , Child, Preschool , Female , Hospital Mortality , Humans , Infant , Male , Patient Safety/statistics & numerical data , Retrospective Studies , Risk FactorsABSTRACT
OBJECTIVE: To determine differences between survivors and nonsurvivors and factors associated with mortality in pediatric intensive care patients with low risk of mortality. DESIGN: Retrospective cohort study. SETTING: Patients were selected from a national database including all admissions to the PICUs in The Netherlands between 2006 and 2012. PATIENTS: Patients less than 18 years old admitted to the PICU with a predicted mortality risk lower than 1% according to either the recalibrated Pediatric Risk of Mortality or the Pediatric Index of Mortality 2 were included. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: In total, 16,874 low-risk admissions were included of which 86 patients (0.5%) died. Nonsurvivors had more unplanned admissions (74.4% vs 38.5%; p < 0.001), had more complex chronic conditions (76.7% vs 58.8%; p = 0.001), were more often mechanically ventilated (88.1% vs 34.9%; p < 0.001), and had a longer length of stay (median, 11 [interquartile range, 5-32] d vs median, 3 [interquartile range, 2-5] d; p < 0.001) when compared with survivors. Factors significantly associated with mortality were complex chronic conditions (odds ratio, 3.29; 95% CI, 1.97-5.50), unplanned admissions (odds ratio, 5.78; 95% CI, 3.40-9.81), and admissions in spring/summer (odds ratio, 1.67; 95% CI, 1.08-2.58). CONCLUSIONS: Nonsurvivors in the PICU with a low predicted mortality risk have recognizable risk factors including complex chronic condition and unplanned admissions.
Subject(s)
Critical Care , Critical Illness/mortality , Hospital Mortality , Intensive Care Units, Pediatric , Adolescent , Child , Child, Preschool , Chronic Disease , Female , Humans , Infant , Infant, Newborn , Logistic Models , Male , Netherlands/epidemiology , Registries , Retrospective Studies , Risk Assessment , Risk FactorsABSTRACT
BACKGROUND: Central-line-associated bloodstream infections (CLABSIs) are a major problem in intensive care units (ICUs) worldwide. We aimed to quantify the effectiveness of central-line bundles (insertion or maintenance or both) to prevent these infections. METHODS: We searched Embase, MEDLINE OvidSP, Web-of-Science, and Cochrane Library to identify studies reporting the implementation of central-line bundles in adult ICU, paediatric ICU (PICU), or neonatal ICU (NICU) patients. We searched for studies published between Jan 1, 1990, and June 30, 2015. For the meta-analysis, crude estimates of infections were pooled by use of a DerSimonian and Laird random effect model. The primary outcome was the number of CLABSIs per 1000 catheter-days before and after implementation. Incidence risk ratios (IRRs) were obtained by use of random-effects models. FINDINGS: We initially identified 4337 records, and after excluding duplicates and those ineligible, 96 studies met the eligibility criteria, 79 of which contained sufficient information for a meta-analysis. Median CLABSIs incidence were 5·7 per 1000 catheter-days (range 1·2-46·3; IQR 3·1-9·5) on adult ICUs; 5·9 per 1000 catheter-days (range 2·6-31·1; 4·8-9·4) on PICUs; and 8·4 per 1000 catheter-days (range 2·6-24·1; 3·7-16·0) on NICUs. After implementation of central-line bundles the CLABSI incidence ranged from 0 to 19·5 per 1000 catheter-days (median 2·6, IQR 1·2-4·4) in all types of ICUs. In our meta-analysis the incidence of infections decreased significantly from median 6·4 per 1000 catheter-days (IQR 3·8-10·9) to 2·5 per 1000 catheter-days (1·4-4·8) after implementation of bundles (IRR 0·44, 95% CI 0·39-0·50, p<0·0001; I(2)=89%). INTERPRETATION: Implementation of central-line bundles has the potential to reduce the incidence of CLABSIs. FUNDING: None.
Subject(s)
Bacteremia/prevention & control , Catheter-Related Infections/drug therapy , Catheterization, Central Venous/adverse effects , Infection Control , Bacteremia/epidemiology , Catheter-Related Infections/prevention & control , Chlorhexidine/chemistry , Critical Illness , Cross Infection/prevention & control , Hand Hygiene , Humans , Incidence , Intensive Care Units , Intensive Care Units, NeonatalABSTRACT
UNLABELLED: The objectives of this study were to identify causal and contributing factors of serious patient safety incidents in a paediatric university hospital, to report on ensuing recommendations and to assess the extent of implementation of the recommendations. The possible causal and contributing factors identified in 17 incidents were classified by a system devised by Vincent et al. Proposed recommendations were classified by the same system, and degrees of implementation were established. A median of 5 causal and contributing factors per incident were identified. Twenty-two percent of all factors were related to teamwork and 22 % to task factors. A median of 5 recommendations per analysis were formulated. Most recommendations were related to task factors (36 %). The time load of each analysis was a mean of 27 h. One third of the recommendations have been acted upon, mostly those related to task and team factors. CONCLUSION: Incident analysis is time-consuming but yields indispensable information on causal and contributing factors, presenting numerous opportunities for quality improvement. The value of these analyses could be improved by appointing responsibilities and setting up time frames for implementation. A bottom-up approach with managerial support appears to be a key to turning incident analysis and quality improvement into an ongoing process.
Subject(s)
Critical Care/organization & administration , Documentation , Intensive Care Units/organization & administration , Patient Safety , Safety Management/organization & administration , Data Collection , Hospitals, University , Humans , Intensive Care Units, Pediatric/organization & administration , Medical Errors/prevention & control , Netherlands , Quality Assurance, Health Care/organization & administration , Retrospective StudiesABSTRACT
OBJECTIVE: Adverse event registration is a means to improve patient safety in a PICU. So far it has been used in European and North American countries mainly. We studied adverse events in a South African setting with the aims to 1) assess rates and types of adverse events with two different registration methods and 2) describe characteristics of patients experiencing adverse events. DESIGN: This study consisted of 1) a retrospective audit of randomly selected patient records and 2) a prospective observational study of real-time registration of AEs during ward rounds. Adverse events were identified using the Child Health Corporation of America - Pediatric Pharmacy Advocacy Group PICU trigger tool. SETTING: A multidisciplinary 20 bed PICU at the Red Cross War Memorial Children's Hospital in Cape Town. PATIENTS: The retrospective section of the study involved 80 randomly selected patients who had been discharged from the PICU, and the prospective study involved patients who were present in the PICU between March and June 2012. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The retrospective audit identified 260 adverse events in 61 patients (50.8 per 100 patient days). Nineteen patients (24%) did not have any adverse events. Catheter complications, hypoglycemia, and endotracheal tube malpositioning requiring repositioning were most frequent. Prospective registration during 58 ward rounds revealed 272 adverse events in 236 patients (27.2 per 100 patient days), particularly catheter complications, nosocomial infection, and surgical complications. Hundred thirty-two patients of the total 236 patients (56%) did not experience an adverse event. Patients experiencing adverse events underwent mechanical ventilation significantly more frequently. Length of stay was significantly associated with number of adverse events. CONCLUSIONS: The trigger tool method identifies a higher adverse event rate compared with real-time registration. Each method has a unique contribution to yield types of adverse events.
Subject(s)
Intensive Care Units, Pediatric/statistics & numerical data , Intensive Care Units, Pediatric/standards , Patient Safety/statistics & numerical data , Catheters/adverse effects , Child, Preschool , Cross Infection/epidemiology , Female , Humans , Hypoglycemia/epidemiology , Infant , Intubation, Intratracheal/adverse effects , Length of Stay/statistics & numerical data , Male , Medical Audit , Postoperative Complications/epidemiology , Prospective Studies , Registries , Respiration, Artificial/statistics & numerical data , Retrospective Studies , South Africa/epidemiologyABSTRACT
BACKGROUND: Central venous catheter-associated bloodstream infections in children are an increasingly recognized serious safety problem worldwide, but are often preventable. Central venous catheter bundles have proved effective to prevent such infections. Successful implementation requires changes in the hospital system as well as in healthcare professionals' behaviour. The aim of the study is to evaluate process and outcome of implementation of a state-of-the-art central venous catheter insertion and maintenance bundle in a large university children's hospital. METHODS/DESIGN: An interrupted time series design will be used; the study will encompass all children who need a central venous catheter. New state-of-the-art central venous catheter bundles will be developed. The Pronovost-model will guide the implementation process. We developed a tailored multifaceted implementation strategy consisting of reminders, feedback, management support, local opinion leaders, and education. Primary outcome measure is the number of catheter-associated infections per 1000 line-days. The process outcome is degree of adherence to use of these central venous catheter bundles is the secondary outcome. A cost-effectiveness analysis is part of the study. Outcomes will be monitored during three periods: baseline, pre-intervention, and post-intervention for over 48 months. DISCUSSION: This model-based implementation strategy will reveal the challenges of implementing a hospital-wide safety program. This work will add to the body of knowledge in the field of implementation. We postulate that healthcare workers' willingness to shift from providing habitual care to state-of-the-art care may reflect the need for consistent care improvement. Trial registration: Dutch trials registry, trial # 3635. TRIAL REGISTRATION: Dutch trials registry (http://www.trialregister.nl), trial # 3635.
Subject(s)
Catheterization, Central Venous/methods , Hospitals, Pediatric/organization & administration , Patient Care Bundles/methods , Catheter-Related Infections/prevention & control , Catheterization, Central Venous/economics , Catheterization, Central Venous/standards , Child , Child, Preschool , Cost-Benefit Analysis , Female , Hospitals, Pediatric/standards , Humans , Infant , Infant, Newborn , Male , Outcome and Process Assessment, Health Care , Patient Care Bundles/economics , Patient Care Bundles/standards , Program DevelopmentABSTRACT
AIMS AND OBJECTIVES: The Critical Nursing Situation Index (CNSI) is a checklist to detect nursing protocol violations. The objectives of this study were to determine incidences and severities of nursing protocol violations and to check whether corrective actions were taken. DESIGN: Prospective observational audit. METHODS: This study was performed in the Intensive Care Unit of the Erasmus MC-Sophia Children's Hospital, Rotterdam, The Netherlands. The CNSI was applied in the period February 2009 to February 2010 by 14 purpose-trained nurses whose interrater reliability had proved sufficient. The checklist addressed nine domains of nursing care: Basic care, Circulation, Respiration, Digestive tract, Infection, Invasive catheters, Medication, ECMO and Central nervous system. The trained nurses also recorded whether violations were discussed with the bedside nurse, whether they could be corrected; and whether they were justifiable. Protocol violations are justifiable when protocol adherence carries greater risk of harm to the patient. RESULTS: Protocol violations were identified for 987 of 8107 items (12·2%) checked in 238 observations in 126 patients. The percentage of protocol violations varied from 5% in the Medication domain to 26% in the Digestive tract domain. More than fifty percent (53·4%) of all protocol violations were corrected in the same shift; 22·3% of all protocol violations proved justifiable, however, these were rarely documented (6·4% of cases). Nurses' classification of the severity of the protocol violations was not reliable because linearly weighted kappa varied from 0 to 0·33. CONCLUSIONS: The CNSI is a useful tool to monitor and correct nursing protocol violations. RELEVANCE TO CLINICAL PRACTICE: Timely identification and correction of protocol violations will reduce possible adverse events resulting from these violations. Furthermore, this study made us aware that protocol violations may be justifiable in clinical practice provided they are well documented.
Subject(s)
Clinical Protocols/standards , Critical Care Nursing/standards , Quality of Health Care/statistics & numerical data , Checklist , Communication , Female , Humans , Male , Netherlands , Prospective StudiesABSTRACT
BACKGROUND: Cerebral MRI performed on preterm infants at term-equivalent 30 weeks' gestational age (GA) is increasingly performed as part of standard clinical care. OBJECTIVE: We evaluated safety of these early MRI procedures. MATERIALS AND METHODS: We retrospectively collected data on patient safety of preterm infants who underwent early MRI scans. Data were collected at fixed times before and after the MRI scan. MRI procedures were carried out according to a comprehensive guideline. RESULTS: A total of 52 infants underwent an MRI scan at 30 weeks' GA. Although no serious adverse events occurred and vital parameters remained stable during the procedure, minor adverse events were encountered in 26 infants (50%). The MRI was terminated in three infants (5.8%) because of respiratory instability. Increased respiratory support within 24 h after the MRI was necessary for 12 infants (23.1%) and was significantly associated with GA, birth weight and the mode of respiratory support. Hypothermia (core temperature < 36°C) occurred in nine infants (17.3%). Temperature dropped significantly after the MRI scan. CONCLUSION: Minor adverse events after MRI procedures at 30 weeks GA were common and should not be underestimated. A dedicated and comprehensive guideline for MRI procedures in preterm infants is essential.
Subject(s)
Hypothermia/epidemiology , Infant, Premature , Infant, Very Low Birth Weight , Magnetic Resonance Imaging/statistics & numerical data , Respiration Disorders/epidemiology , Female , Humans , Incidence , Infant, Newborn , Netherlands/epidemiology , Risk FactorsABSTRACT
UNLABELLED: The objectives of this study are to describe the number and nature of adverse events occurring in general pediatric practice, to describe factors contributing to the occurrence of these adverse events, and to report on the experience of pediatricians with reporting adverse events. It is a prospective study on 11 pediatric units in a 3-month period; adverse events were registered for all newly admitted patients. Ninety-four adverse events were registered in 88 of 5,669 patients, amounting to a 1.6 per 100 admissions rate and a 0.4 per 100 patient days rate. Ninety percent of the adverse events did not cause serious harm. Failed diagnostic procedures were most common. CONCLUSION: Adverse event registration in general pediatric practice is a first step in assessing quality and safety of care. It yields a considerable number of adverse events. Compliance to adverse event registration in daily practice is difficult but also key to optimal monitoring of quality of care.
Subject(s)
Hospitalization , Medical Errors/statistics & numerical data , Patient Safety/statistics & numerical data , Pediatrics/standards , Quality Indicators, Health Care/statistics & numerical data , Child , Humans , Netherlands , Outcome Assessment, Health Care , Patient Safety/standards , Prospective Studies , Quality Assurance, Health Care , RegistriesSubject(s)
Eosinophilia/complications , Eosinophilia/epidemiology , Intestinal Pseudo-Obstruction/etiology , Myenteric Plexus , Neuritis/complications , Neuritis/epidemiology , Anti-Inflammatory Agents , Beclomethasone , Chronic Disease , Colectomy , Comorbidity , Enteral Nutrition , Eosinophilia/drug therapy , Follow-Up Studies , Humans , Infant , Intestinal Pseudo-Obstruction/surgery , Male , Neuritis/drug therapy , Prednisone/toxicity , Treatment OutcomeABSTRACT
OBJECTIVE: Nowadays, the belief is widespread that a safety culture is crucial to achieving patient safety, yet there has been virtually no analysis of the safety culture in pediatric hospital settings so far. Our aim was to measure the safety climate in our unit, compare it with benchmarking data, and identify potential deficiencies. DESIGN: Prospective longitudinal survey study at two points in time. SETTING: Pediatric surgical intensive care unit at a Dutch university hospital. SUBJECTS: All unit personnel. INTERVENTIONS: To measure the safety climate, the Safety Attitudes Questionnaire was administered to physicians, nurses, nursing assistants, pharmacists, technicians, and ward clerks in both May 2006 and May 2007. This questionnaire assesses caregiver attitudes through use of the six following scales: teamwork climate, job satisfaction, perceptions of management, safety climate, working conditions, and stress recognition. Earlier research showed that the Safety Attitudes Questionnaire has good psychometric properties and produced benchmarking data that can be used to evaluate strengths and weaknesses in a given clinical unit against peers. MEASUREMENTS AND MAIN RESULTS: The response rates for the Safety Attitudes Questionnaire were 85% (May 2006) and 74% (May 2007). There were mixed findings regarding the difference between physicians and nurses: on three scales (i.e., teamwork climate, safety climate, and stress recognition), physicians scored better than nurses at both points in time. On another two scales (i.e., perceptions of management and working conditions), nurses consistently had higher mean scale scores. Probably due to the small number of physicians, only some of these differences between physicians and nurses reached the level of statistical significance. Compared to benchmarking data, scores on perceptions of management were higher than expected (p < .01), whereas scores on stress recognition were low (p < .001). The scores on the other scales were somewhat above (job satisfaction), close to (teamwork climate, safety climate), or somewhat below (working conditions) what was expected on the basis of benchmarking data, but no persistent significant differences were observed on these scales. CONCLUSIONS: Although on most domains the safety culture in our unit was good when compared to benchmark data, there is still room for improvement. This requires us to continue working on interventions intended to improve the safety culture, including crew resource management training, safety briefings, and senior executive walk rounds. More research is needed into the impact of creating a safety culture on patient outcomes.
Subject(s)
Attitude of Health Personnel , Intensive Care Units, Pediatric , Organizational Culture , Patient Safety , Surveys and Questionnaires/statistics & numerical data , Adolescent , Adult , Child , Child, Preschool , Female , Humans , Infant , Infant, Newborn , Longitudinal Studies , Male , Medical Errors/prevention & control , Middle Aged , Netherlands , Prospective Studies , Surgery Department, HospitalABSTRACT
OBJECTIVES: To evaluate availability and reliability of pediatric drug dosing guidelines in selected formularies for intensive care patients. Most drugs used in the pediatric intensive care unit are prescribed off-label, often on the guidance of limited information from commonly used drug formularies. DESIGN: Availability of dosing information on prescribed drugs in a Dutch intensive care unit from January 1, 2005 to December 31, 2006 was compared among four selected formularies (Micromedex, Lexi-Comp, Drug Formulary for Children, Drug Doses). Reliability of dosing guidelines was assessed by evaluating labeling status and literature data for the three most (midazolam, acetaminophen, and amoxicillin/clavulanic acid) and the three least (bosentan, ketanserin, and iloprost) prescribed drugs. MEASUREMENTS AND MAIN RESULTS: The selected formularies covered 68% to 86% of all 257 prescribed drugs. Guidelines differ widely on daily doses per kilogram, dose description, dosing regimen, and age ranges. For the three most prescribed and one of the least prescribed drugs (bosentan), dosing guidelines adequately reflected labeling status and existing (but scarce) literature. No dosing guidelines were available for iloprost, and only one dosing guideline was available for ketanserin. CONCLUSIONS: This study shows that four commonly used drug formularies give few and widely differing dosing guidelines for drugs prescribed in the intensive care unit. If guidelines exist, they seem to reflect labeling status (if present) and limited literature available. Findings from this study likely reflect the scarcity of drug studies in this population. Physicians should be aware of the limitations of these formularies for daily practice in this group of vulnerable patients.
Subject(s)
Drug Prescriptions/statistics & numerical data , Formularies, Hospital as Topic , Intensive Care Units, Pediatric , Humans , Netherlands , Reproducibility of Results , Statistics, NonparametricABSTRACT
BACKGROUND: The mortality for children with prolonged stay in pediatric intensive care units (PICU) is much higher than overall mortality. The incidence of withdrawal or limitation of therapy in this group is unknown. PURPOSE: To assess mortality and characteristics of children admitted for >or=28 days to our ICU, and to describe the extent to which limitations of care were involved in the terminal phase preceding death. METHODS: For the period 2003 to 2005 clinical data were collected retrospectively for children with prolonged stay (defined as >or=28 days) in a medical/surgical PICU of a university children's hospital. RESULTS: In the PICU, 4.4% of the children (116/2,607, equal gender, mean age 29 days) had a prolonged stay. Median (range) stay was 56 (28-546) days. These children accounted for 3% of total admissions and occupied 63% of total admission days. Mortality during admission for this group was five times higher (22%) than the average PICU mortality rate of 4.6%. Withdrawal or limitation of therapy preceded 70% of deaths. CONCLUSIONS: Children with prolonged stay in the PICU have a significantly high risk of mortality. Death is typically preceded by limitation of care.