Details of HTA project in progress
Last updated: 31 August 2010 - Next update due: 7 September 2010
Research type: |
Primary Research (e.g. trial) |
Project title: |
Pulse oximetry as a screening test for congenital heart disease in newborn babies |
Project ref: |
06/06/03 |
Cost: |
£940,787 |
Chief Investigator : |
Dr Andrew K Ewer, Senior Lecturer/Consultant Neonatologist, Neonatology, University of Birmingham |
Start Date: |
July 2007 |
Estimated date of publication in HTA journal series: |
Late 2011. This date takes account of time for report preparation and printing based on current average times for these activities. |
Plain English Summary |
Congenital Heart Disease (CHD) is the commonest group of congenital malformations and affects 7-8/1000 live born newborns. It contributes to 3% of all infant mortality and 46% of deaths from congenital malformations with most deaths occurring in the first year of life. Currently in the UK, all newborn babies undergo a routine screening examination, usually in the first 24 hours after birth, during which, among other things, a careful assessment of the cardiovascular system is undertaken. However, it is estimated that over 50% of babies with undiagnosed CHD failed to be picked up by routine neonatal examination. The need for an accurate, simple, non-invasive test for CHD in the early neonatal period has led a number of investigators to examine the use of pulse oximetry and although results are encouraging, as we demonstrated in our systematic review, there is a clear need for a larger, robust, well conducted study to confirm the value, acceptability and cost effectiveness of such a screening programme. This large multi-centre study will determine the accuracy of Pulse Oximetry (PO) screening for CHD in newborns. It will evaluate the acceptability of PO to both parents and health professionals. The study will also assess the costs and cost effectiveness of utilising such screening in combination with clinical examination in the early detection of potentially life-threatening CHD. All women booking into seven large obstetric units, in the West Midlands area, will be invited to participate. The study aims to recruit 20,000 women over 12 months for the assessment of test accuracy and acceptability. The main analysis for diagnostic accuracy will estimate sensitivity, specificity, predictive values, likelihood ratios and their confidence intervals. Using multivariable logistic regression analysis, predictive probabilities for various combinations of history, antenatal tests and oximetry results will be generated. |
Project Abstract: |
Design: Multicentre prospective study to assess diagnostic accuracy and cost-effectiveness of routine PO to screen for CHD in newborns. Study design will meet the STARD criteria for methodological quality of test accuracy studies (3). We will examine acceptability of PO for parents and health professionals using a questionnaire to a sample of participants and focus groups of parents of children with CHD (screen positive or negative). Primary data on test accuracy, cost and resource use will be collected and used in a cost effectiveness analysis based on the structure of a pre-existing decision analytic model (4). Follow up data to 1 year retrieved as part of the study will be used and additional data required will be extracted from the literature where appropriate. Setting: 7 maternity units in West Midlands (WM) (total annual deliveries - 35 000); Cardiology referral centres, WM Perinatal Institute and GP surgeries to identify children with CHD. Target population: Asymptomatic newborns born after 35 weeks gestation. Exclusions: symptoms of CHD at birth, extracardiac defects associated with CHD or mother unable to give written informed consent Health technologies being assessed: PO to measure oxygen saturation as a screening for CHD. PO will be performed on one hand and one foot of all asymptomatic newborns to measure functional oxygen saturation, before 24 hours or discharge from hospital and ideally at 3-6 hours. Those with saturation <95% will be examined, and if normal have a repeat PO after 1-2 hours. Wherever possible, screening will be performed before 24 hours of age in order to minimise those presenting with cardiac symptoms prior to screening. Screen positive newborns with persistently low oxygen saturation or with signs suggesting CHD will undergo a hyperoxia test and echocardiography, the gold standard to diagnose clinically significant, non significant CHD or no CHD. The condition of screen negative newborns will be followed up through databases maintained by the WM congenital anomaly register (CAR), regional cardiology referral unit, regional perinatal mortality survey and Read codes of primary care trusts to identify any CHD diagnosed by 1 year of life. These are the different gold standards used to ascertain the outcome in those screened negative for the tests. Cost and Outcome measurement: Primary outcome is timely diagnosis of life-threatening CHD. Secondary outcome is significant CHD and non-cardiac serious illness. Economic evaluation based on the primary and secondary endpoints will be extrapolated in the decision model to principal economic outcome of cost per infant death avoided at one year. Sample size: Assuming a 'true' sensitivity of 75% and specificity of 99.5%, sample sizes for CHD prevalence between 2 and 5 per 1,000, using one-tailed significance level of 2.5% have been computed. A sample size of 20,000 will have 80% power to demonstrate that sensitivity is above 61% at a prevalence of 5 per 1000 in the overall sample. For the subgroup of women who are not selectively screened antenatally, we anticipate the prevalence to be around 2 per 1000, a level at which sensitivity of 52% can be detected with 80% power. Our assumption about disease prevalence is quite conservative. If prevalence in our study was higher than 0.5%, we will have more than the projected power to undertake subgroup analyses confidently. On this basis, considering the likely prevalence, we plan to recruit a sample size of 20,000. Consent and data collection: Written informed consent will be obtained from every woman prior to delivery. Informed consent is currently obtained for other newborn screening (including hearing and blood spot) and in the WM current uptake for this is 99%. Data would be collected and input into a web-based study database. Project Timetable: Start-up: 3 month period of training provision of information to pregnant women. Recruitment: 8-9 months, assuming data is adequately collected from 95% of deliveries from the 7 centres. Follow-up: 12 months after last delivery. Analysis and Reporting: 6 months. Total project: 30 months |
Project Protocol: |
Project protocol (pdf format, 584 kbytes) |
URL of this page: |
http://www.hta.ac.uk/1624 |
News feeds