gastroschisis and mode of delivery: national trends and evaluation of outcomes, 1991 to 2005 t. mac...

14
Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird T. Mac Bird University of Arkansas for Medical University of Arkansas for Medical Sciences Sciences College of Medicine, Department of College of Medicine, Department of Pediatrics Pediatrics College of Public Health, Department of College of Public Health, Department of Health Policy and Management Health Policy and Management

Upload: gary-pearson

Post on 24-Dec-2015

214 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

Gastroschisis and Mode of Delivery: National Trends and

Evaluation of Outcomes, 1991 to 2005

T. Mac BirdT. Mac Bird

University of Arkansas for Medical SciencesUniversity of Arkansas for Medical Sciences

College of Medicine, Department of PediatricsCollege of Medicine, Department of Pediatrics

College of Public Health, Department of Health Policy College of Public Health, Department of Health Policy and Managementand Management

Page 2: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

BackgroundBackground

Gastroschisis is a severe birth defect where the intestines are herniated through the abdominal wall and exposed directly to amniotic fluid

Affects 1 in 2500 live births

Rates of Gastroschisis have been reported to be rising in both in the US and in other nations around the world

Page 3: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

BackgroundBackground

No consensus on preferred mode of No consensus on preferred mode of delivery for infants with gastroschisis. delivery for infants with gastroschisis.

Proponents of delivery by c-section argue– Reduced risk of infection– Reduces damage to eviscerated intestines– Allows scheduling of neonatal care team

However, no evidence of improved outcome over vaginal delivery

Page 4: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

ObjectivesObjectives

Use national hospital discharge data to determine trends in CS for infants with gastroschisis

Use instrumental variable methods to approximate the results of a randomized controlled trial using observational data

Determine if outcomes for infants with gastroschisis differ by mode of delivery

Page 5: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

DatabaseDatabase

Nationwide Inpatient Sample (NIS)Nationwide Inpatient Sample (NIS)– Created by the Agency for Healthcare Created by the Agency for Healthcare

Research and QualityResearch and Quality– Approximates a 20% stratified random sample Approximates a 20% stratified random sample

of all community hospital in the USof all community hospital in the US– ~ 8 million discharges per year~ 8 million discharges per year

Kid’s Inpatient Database (KID)Kid’s Inpatient Database (KID)– Pediatric specificPediatric specific– Larger pediatric sample for available yearsLarger pediatric sample for available years

Page 6: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

SubjectsSubjects

Live births in NIS/KID between 1991 and 2005

Diagnosis code for abdominal wall defect (ICD-9-CM 765.7)

Procedure code for gastroschisis repair (ICD-9-CM 54.71)

Singleton birthSingleton birth

Received surgical repair in birth hospitalReceived surgical repair in birth hospital

Page 7: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

AnalysisAnalysis

Bivariate probit models with instrumental Bivariate probit models with instrumental variablesvariables

Dependent Variables – Mortality, necrotizing Dependent Variables – Mortality, necrotizing enterocolitis, sepsis, respiratory distress, enterocolitis, sepsis, respiratory distress, transitory tachypnea, intestinal surgery, transitory tachypnea, intestinal surgery, medical complications, staged closuremedical complications, staged closure

Independent Variable – Mode of delivery, c-Independent Variable – Mode of delivery, c-section = 1, vaginal delivery = 0section = 1, vaginal delivery = 0

Page 8: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

AnalysisAnalysis

Instrumental Variables – Admission to hospital Instrumental Variables – Admission to hospital on a weekend, c-section rates for infants with on a weekend, c-section rates for infants with gastroschisis by US regional division per yeargastroschisis by US regional division per yearControl Variables – basic demographics, Control Variables – basic demographics, hospital characteristics, 10 indicators for hospital characteristics, 10 indicators for comorbid birth defect categories, mechanical comorbid birth defect categories, mechanical ventilation, fetal distress, maternal ventilation, fetal distress, maternal complications, small for gestational age, 4 birth complications, small for gestational age, 4 birth weight categoriesweight categories

Page 9: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

ResultsResults

2,785 (10,073 weighted) cases of 2,785 (10,073 weighted) cases of gastroschisisgastroschisis1,498 (5,332 weighted) or 52.9% born by 1,498 (5,332 weighted) or 52.9% born by c-sectionc-section1,497 (5,293 weighted) or 52.6% were 1,497 (5,293 weighted) or 52.6% were born pretermborn pretermRates increased from low of 1.7/10,000 Rates increased from low of 1.7/10,000 live births in 1996 to a high of 4.4/10,000 live births in 1996 to a high of 4.4/10,000 in 2005in 2005

Page 10: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

ResultsResults

0%

10%

20%

30%

40%

50%

60%

70%

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

Pe

rce

nt

of

De

live

rie

s b

y C

es

are

an

Se

cti

on

GastroschisisCS Rates

Overall CSRates

Page 11: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

ResultsResults

Table 4. Comparison of outcomes by mode of delivery for infants with gastroschisis.

  Bivariate Analysis Instrumental Variable

Outcome Variable Crude OR 95% CI AIRR 95% CI

Died in Hospital 1.49 (0.92, 2.41) 2.10 (0.89, 4.95)

Necrotizing Enterocolitis 1.29 (0.88, 1.89) 1.91 (1.24, 2.94)

Sepsis 0.90 (0.70, 1.17) 1.10 (0.69, 1.75)

Respiratory Distress 1.10 (0.82, 1.48) 1.15 (0.69, 1.90)

Transitory Tachypnea 1.12 (0.63, 2.00)

Intestinal Surgery 1.00 (0.81, 1.24) 0.82 (0.52, 1.29)

Medical Complication 0.91 (0.74, 1.12) 0.94 (0.64, 1.38)

Staged/Silo Closure 0.94 (0.77, 1.15) 0.92 (0.60, 1.43)

Bold indicates statistical significance.

Page 12: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

ConclusionsConclusions

Rates of gastroschisis have increased Rates of gastroschisis have increased steadily over the past 15 years in the steadily over the past 15 years in the United StatesUnited States

This study found no benefit to delivery of This study found no benefit to delivery of infants with gastroschisis by c-section infants with gastroschisis by c-section compared to vaginal birthcompared to vaginal birth

Delivery of infants with gastroschisis by c-Delivery of infants with gastroschisis by c-section could be harmfulsection could be harmful

Page 13: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

LimitationsLimitations

Discharge data created for reimbursementDischarge data created for reimbursement– Coding practices can vary by hospital, state, Coding practices can vary by hospital, state,

region, and over timeregion, and over time– The more severe/expensive a diagnosis or The more severe/expensive a diagnosis or

procedure the more likely to be codedprocedure the more likely to be coded

No code for gastroschisisNo code for gastroschisis– Algorithm likely under codes gastroschisis Algorithm likely under codes gastroschisis

slightly although positive predictive value slightly although positive predictive value should be highshould be high

Page 14: Gastroschisis and Mode of Delivery: National Trends and Evaluation of Outcomes, 1991 to 2005 T. Mac Bird University of Arkansas for Medical Sciences College

AcknowledgementsAcknowledgements

Evan KokoskaEvan Kokoska

Chenghui LiChenghui Li

Mario ClevesMario Cleves

Nirvana ManningNirvana Manning

Mick TilfordMick Tilford

James RobbinsJames Robbins

Charlotte HobbsCharlotte Hobbs

This study was funded in part by a grant from the Arkansas Biosciences Institute