Additional funding was provided by the B.C.'s Children's Hospital ... ences Corporation; Mount Sinai Hospital; North York General Hospital Foundation; Saint.
Transport risk index of physiologic stability: A practical system for assessing infant transport care Shoo K. Lee, MBBS, FRCPC, PhD, John A. F. Zupancic, MD, FRCPC, MS, Margaret Pendray, MBBS, FRCPC, Paul Thiessen, MD, FRCPC, Barbara Schmidt, MD, FRCPC, MSc, Robin Whyte, MBBS, FRCPC, David Shorten, RN, MN, Shawn Stewart, BA, and The Canadian Neonatal Network Objectives: To develop and validate a practical, physiology-based system for assessment of infant transport care. Study design: Transport teams prospectively collected data, before and after transport, from 1723 infants at 8 neonatal intensive care units (NICUs) from 1996 to 1997. We used logistic regression to derive a prediction model for mortality within 7 days of NICU admission and develop the Transport Risk Index of Physiologic Stability (TRIPS). We validated TRIPS for prediction of 7-day mortality, total NICU mortality (until discharge), and severe (≥grade 3) intraventricular hemorrhage. Results: TRIPS comprises 4 empirically weighted items (temperature, blood pressure, respiratory status, and response to noxious stimuli). TRIPS discriminated 7-day NICU mortality and total NICU mortality from survival with receiver operating characteristic areas of 0.83 and 0.76, respectively. There was good calibration across the full range of TRIPS scores and gestational age groups. Increase and decrease in TRIPS scores after transport were associated with increased and decreased mortality, respectively. The receiver operating characteristic area for TRIPS prediction of severe intraventricular hemorrhage was 0.74. Addition of TRIPS improved performance of prediction models in which gestational age and baseline population risk variables were used. Conclusions: TRIPS is validated for infant transport assessment. (J Pediatr 2001;139:220-6) From the Department of Pediatrics, University of British Columbia, Vancouver, British Columbia; Centre for Community Health and Health Evaluation Research, Vancouver, British Columbia; Department of Pediatrics, McMaster University, Hamilton, Ontario; Department of Pediatrics, Dalhousie University, Halifax, Nova Scotia; and Clinical Evaluation Services, Calgary Regional Health Authority, Calgary, Alberta, Canada.
Supported by Grant 40503 and Grant 00152 from the Medical Research Council of Canada. Additional funding was provided by the B.C.’s Children’s Hospital Foundation; Calgary Regional Health Authority; Division of Neonatology, Children’s Hospital of Eastern Ontario; Child Health Program, Dalhousie University Neonatal Perinatal Research Fund; Health Care Corporation of St John’s; The Neonatology Program, Hospital for Sick Children; Lawson Research Institute; Midland Walwyn Capital Inc; Division of Neonatology, Hamilton Health Sciences Corporation; Mount Sinai Hospital; North York General Hospital Foundation; Saint Joseph’s Health Centre; University of Saskatchewan Neonatal Research Fund; University of Western Ontario; and Women’s College Hospital. Submitted for publication Nov 14, 2000; revision received Jan 24, 2001; accepted Feb 22, 2001. Reprint requests: Shoo K. Lee, MBBS, FRCPC, PhD, Canadian Neonatal Network Coordinating Center, 4480 Oak St, Room E-414, Vancouver, British Columbia, Canada V6H 3V4. Copyright © 2001 by Mosby, Inc. 0022-3476/2001/$35.00 + 0 9/21/115576 doi:10.1067/mpd.2001.115576
220
Although infant transport is a key component of neonatal-perinatal care,1-5 there is a lack of suitable instruments with which to assess transport care. The Neonatal Stabilization ANTSS
Alberta Neonatal Transport Stabilization Score GA Gestational age NICU Neonatal intensive care unit ROC Receiver operating characteristic SGA Small for gestational age SNAP-II Score for Neonatal Acute Physiology, Version II TRIPS Transport Risk Index of Physiologic Stability
Score6 mainly assesses the transport process rather than infant stability. The transport score of Hermansen et al7 is only validated for very low birth weight infants, uses arbitrarily assigned scores, and relies on laboratory-based tests. The Neonatal Status Score8 and Alberta Neonatal Transport Stabilization Score9 use arbitrarily assigned scores rather than empiric weighting of physiologic derangement, and neither instrument has been validated. Illness severity scores, such as the Clinical Risk Index for Babies10 and Score for Neonatal Acute Physiology Version II,11 are not suitable for transport because they require prolonged data collection over 12 hours. Systems for assessing transport care are inherently more demanding and difficult to develop than admission illness severity scores because measurement is limited by the brief duration of the transport, paucity of laboratory support, and need for brief and rapid measurements that are sufficiently
LEE ET AL
THE JOURNAL OF PEDIATRICS VOLUME 139, NUMBER 2 sensitive to capture changes in patient condition resulting from the transport process. The purpose of this study was to create a practical, validated, empirically weighted, physiology-based system for assessment of key infant transport outcomes.
METHODS Sample Selection We examined all “outborn” infants who were transported to 8 tertiary level Canadian neonatal intensive care units by the hospital infant transport teams from January 1996 to October 1997 (n = 2437 infants).12 We included only inbound transports from community hospitals to tertiary NICUs and transports between tertiary NICUs (n = 12) because retrotransfers to community hospitals usually involve infants in stable condition and are less relevant to the objectives of this study. Complete transport data were available for 1723 (71%) infants, who were included in the study. We excluded 488 (20%) infants who were not recruited into the study and 226 (9%) infants with incomplete data. Non-recruitment and incomplete information were predominantly due to inadvertent omission of data collection immediately on arrival at the referring hospital, and not to difficulty with obtaining the data items. There were no significant (P < .05) differences in the characteristics (birth weight, gestational age, sex, multiple births, cesarean delivery, Apgar score at 5 minutes, small for gestational age) of transported infants who were included and those who were excluded from the study. Feedback from transport teams and examination of the distribution of missing data items (response to noxious stimuli 69%, blood glucose value 65%, perfusion 64%, blood pressure 61%, temperature 43%, ventilation 42%, and pulse oximetry 42%) did not indicate that certain items were more likely to be omitted because they were especially difficult to obtain.
Population In 1996, Canada had a population of 30 million13 with about 357,000 annual births.14 The 8 NICUs include more than 30% of all tertiary NICU beds in Canada and serve a population of approximately 10 million. The 8 NICUs ranged in size from 26 to 57 beds, 133 to 1129 admissions, and 0 to 7500 births annually. Two NICUs admitted only outborn infants. At the other 6 NICUs, outborn infants comprised 3% to 40% of admissions. Each NICU served a distinct geographic region and coordinated infant transfer within the region. Infant characteristics were similar to those at other Canadian NICUs.12 Our study cohort was part of a larger Canadian cohort from which SNAP-II was derived. Because SNAP-II was subsequently validated in 2 separate US cohorts,11 it is likely that TRIPS will also be valid for use in US NICUs.
Measurement of TRIPS Before and After Transport We prospectively collected data on modified items of the ANTSS.9 Variables included 7 physiologic measurements (temperature, blood pressure, respiratory distress, capillary filling time over the sternum, pulse oximetry, response to noxious stimuli, and blood glucose level measured with a glucose oxidase strip), representing 5 major physiologic systems—thermoregulatory, cardiovascular, respiratory, neurologic, and metabolic/endocrine. We combined respiratory distress and pulse oximetry into one variable (respiratory status) because they were clinically related, and we used the resultant 6 variables in model derivation. We excluded perinatal risk variables (GA, SGA, and Apgar score at 5 minutes) in order to create a pure physiology-based instrument that is not affected by changes in the relative contributions of birth weight and related perinatal risks to mortality over time. We excluded other nonphysiology-based risk factors that may
affect mortality (eg, antenatal steroid treatment, diagnostic classification, and congenital anomalies) to avoid measuring factors that are not attributable to the transport process. One member of the transport team collected the data immediately (within 15 minutes) on arrival at the referring hospital and immediately after arrival at the destination hospital, and other members of the transport team assessed the patient or engaged in care. Consequently, change in TRIPS score measured before and after transport includes the effect of care provided to the infant by the transport team.
Other Patient Information In the NICU, trained research assistants prospectively abstracted patient information from the mothers’ and infants’ charts daily, as part of a larger study of outcomes and practices in the Canadian Neonatal Network.12 Data were directly entered into laptop computers by using a customized data entry program with built-in error checking and a standard manual of operations and definitions. Data were electronically transmitted to the Canadian Neonatal Network Coordinating Centre for verification. Potential data errors were re-checked by site research assistants. Data management was conducted by the Canadian Neonatal Network Coordinating Centre, in concert with a steering committee comprising experienced researchers and neonatologists representing each of the 5 geographic regions (British Columbia, Prairie provinces, Ontario, Quebec, and Atlantic provinces) in Canada and with site investigators representing each of the participating hospitals. Patient information included demographic information, antenatal history, obstetric information, and selected outcomes.
Variable Definitions Study variables were defined according to the Canadian Neonatal Network SNAP Project Abstractor Manual. GA was defined as the best 221
LEE ET AL
THE JOURNAL OF PEDIATRICS AUGUST 2001
Table I. Characteristics of infants in derivation and validation cohorts
Characteristics
Derivation cohort
Validation cohort
P value
Infants (n) Transports (n) Birth weight (g) (mean ± SD) GA (wk) (mean ± SD) GAT (wk) (mean ± SD) Female sex (%) Cesarean delivery (%) Multiple births (%) 5-Minute Apgar score 32
PREDICTION OF IVH. A transport score should assess risk for both mortality and morbidity. We assessed TRIPS for prediction of severe (≥grade 3) IVH because IVH is well defined, usually occurs during the first few days of life, and is more likely to be affected by transport than other major NICU morbidity (eg, chronic
lung disease, infection). For this analysis, we only included infants ≤32 weeks’ GA who were transported to an NICU within 72 hours of birth because that is the most relevant period for development of IVH. COMPARISON OF TRIPS WITH ANTSS, GA, AND SNAP-II AS PREDICTION MODELS. Because ANTSS9 is an existing transport score, we compared performance of TRIPS with that of ANTSS for prediction of 7-day mortality, total NICU mortality, and severe IVH to determine whether risk weighting has added value. Because SNAP-II is a well-validated measure of physiologic stability, we compared TRIPS with SNAP-II. For comparison, we also compared TRIPS with GA alone. We used the method of Hanley and McNeil22 to compare ROC areas derived from the same cases. A P value 32 wk GAT Total NICU mortality All GAT ≤32 wk GAT >32 wk GAT Severe (≥grade 3) IVH ≤32 wk GAT
TRIPS
ANTSS
GA
SNAP-II
GA + variables*
TRIPS + GA + variables†
0.83 (0.47) 0.88 (0.58) 0.78 (0.97)
0.77‡ (0.82) 0.77 (0.79) 0.75 (0.58)
0.59‡ (0.13) 0.73‡ (0.30) 0.59‡ (0.04)
0.88 (0.44) 0.85 (0.75) 0.88 (0.47)
0.85‡ (0.61) 0.80 (0.87) 0.85‡ (0.88)
0.91‡ (0.67) 0.90 (0.98) 0.90‡ (0.23)
0.76 (0.22) 0.72 (0.23) 0.74 (0.58)
0.70‡ (0.97) 0.65‡ (0.88) 0.73 (0.89)
0.65‡ (0.03) 0.75 (0.66) 0.51‡ (0.05)
0.75 (0.52) 0.72 (0.71) 0.75 (0.80)
0.80‡ (0.05) 0.76 (0.11) 0.79‡ (0.55)
0.85‡ (0.70) 0.83‡ (0.88) 0.83‡ (0.83)
0.74 (0.14)
0.68‡ (0.13)
0.72 (0.74)
0.74 (0.60)
0.76 (0.21)
0.80‡ (0.21)
ROC areas for TRIPS scores are based on pre-transport data. Parentheses contain P values for Hosmer-Lemeshow goodness-of-fit statistic (values
