doi:10.1111/j.1440-1754.2008.01323.x
ORIGINAL ARTICLE
Measuring sensorineural disability in preterm children using a public health screening strategy: A randomised controlled trial Margo A Pritchard,1 Paul B Colditz,1 David I Tudehope,2 Peter H Gray,3 David Cartwright,4 Neil R Wigg5 and Elaine M Beller6 1 Perinatal Research Centre, Royal Brisbane and Women’s Hospital, The University of Queensland, 2Division of Neonatology, Mater Mother’s Hospital, The University of Queensland, 3Growth and Development Unit, Mater Health Services, The University of Queensland, Brisbane, 4Grantley Stable Neonatal Unit, Royal Brisbane and Women’s Hospital, 5Community Child Health Service Royal Children’s Hospital and Health Service District, 6Queensland Clinical Trials Centre, Princess Alexandra Hospital, School of Population Health, The University of Queensland. Australia
Aim: To assess the efficacy of a preterm-targeted screening programme against the routine Australian National Health Medical Research Council (NHMRC) universal child health screening programme to detect disability in a general practice setting in children born ⱕ31 weeks gestation at 12-months of age. Methods: Multi-centred trial involving 202 preterm children randomised to receive the preterm-targeted or NHMRC programme. Primary outcome, correct identification of neurosensory disability by general practitioners assessed against gold standard paediatric assessments. Sensitivity analysis estimated interrater agreement and screening accuracy. Secondary outcomes, post natal depression (PND), parental stress, health service use, screening programme helpfulness and correct identification of levels of disability severity. Results: Of the 195 infants with data on the primary outcome in the preterm-targeted group, their general practitioners correctly identified the disability status of 61/93 (65.6%) children, as compared with 69/102 (67.6%) in the NHMRC group (odds ratios (OR) 0.91 95% confidence interval (CI) 0.50, 1.65). Responses where general practitioners were unsure of a child’s disability status were coded as incorrect and not paired for sensitivity analysis. Sensitivity analysis for 180 diagnostic pairs showed fair interrater agreement for both groups (preterm-targeted k = 0.30 vs. NHMRC k = 0.29) with screening test results favouring the preterm-targeted group with greater sensitivity (73% vs. 33%) but lower specificity (70% vs. 92%) resulting in more over referrals (30% vs. 8%); however, these had a significantly lower mean Developmental Quotient (DQ) score compared with non-disabled children. PND scores were higher in preterm-targeted group (OR 1.33 95% CI 0.01, 2.66). Conclusion: The preterm-targeted programme used by general practitioners: (i) did not improve overall identification of disability status compared to the NHMRC universal programme (Australian New Zealand Clinical Trails Registry number, ACTRN 12606000472572); however (ii) it did demonstrate greater efficacy as a screening tool in accurately identifying disabled children. Key words:
neurosensory; preterm; primary-health; randomised; targeted-screening.
Key Points 1 Accurate identification of the developmental and disability status in children born very preterm who rely on primary health universal childhood screening programs is poor. 2 A preterm-targeted child health screening program improves the precision of the screening and surveillance process for development and disability in primary health settings with test characteristics >70%. 3 Establishing an effective screening partnership between NICU and primary health care professionals may enhance developmental follow up in children born very preterm. Correspondence: Dr Margo Anne Pritchard, Perinatal Research Centre, Royal Brisbane and Women’s Hospital, The University of Queensland, Butterfield Street, Herston, 4029, Australia. Fax: +61 7 3636 1769; email:
[email protected] Accepted for publication 24 January 2008.
424
The estimated prevalence of disability from neurosensory impairment (cerebral palsy, cognitive delay, bilateral blindness and hearing loss requiring aids) in children born ⱕ31 weeks gestation is important information for resource planners, families, clinicians and researchers Established gold standard assessments conducted by neonatal intensive care unit (NICU) follow-up programmes have historically assessed early childhood disability but are not feasible at all the appropriate times for all children because of cost, lack of specialist clinicians and geographical constraints. Reported follow-up studies are more often hospital-based and show considerable variation in definitions of disability making them difficult to combine and compare for regions. Recently, randomised and interrater agreement studies using standardised questionnaires based on functional assessments have been shown to be a pragmatic alternative to NICU follow-up for the collection of populationbased disability estimates. Although these results are promising, there are several limitations which affect their reliability including use of untrained personnel, incomplete forms, limited
Journal of Paediatrics and Child Health 44 (2008) 424–431 © 2008 The Authors Journal compilation © 2008 Paediatrics and Child Health Division (Royal Australasian College of Physicians)
MA Pritchard et al.
Primary health preterm infant screening
Infants with a gestation ≤31 weeks gestation were assessed for eligibility (n = 423)
Excluded (n = 221) Did not meet inclusion criteria (n = 76) • Did not want to attend follow-up (n = 66) •
Had no general practitioners (n = 10)
Refused to participate (n = 18) Eligible but not approached (n = 127)
202 infants were randomised Allocated to the preterm-targeted group & received allocated intervention 98 (49%) Targeted screening & surveillance program • Singletons (n = 58) •
17 sets of twins (n = 34)
•
16 sets of twins (n = 32)
•
2 sets of triplets (n = 6)
•
1 set of triplets (n = 3)
Children not followed-up –
Children not followed-up•
moved overseas (n =1, 1%)
•
mother of twins not wanting further contact (n
•
child taken into foster care (n = 1, 1%)
•
mother with significant depression refused to attend (n = 1, 1%)
= 2, 2%) •
Allocated to the NHMRC group & received allocated intervention 104 (51%) NHMRC screening & surveillance program • Singletons (n = 69)
death (Sudden Infant Death Syndrome & undiagnosed Tetralogy of Fallot) (n = 2, 2%)
Primary outcome data available & analysed for 93 (95%) No cases excluded from analysis
Primary outcome data available & analysed for 102 (98%) No cases excluded from analysis
Fig. 1 Number of infants screened, enrolled and for whom follow-up data is available.
recognition of mild impairment and reliance on functional assessment alone, which is not specifically confined to measuring neurosensory impairment. Children not attending a NICU programme rely on the Australian National Health Medical Research Centre (NHMRC) universal child health screening schedule conducted by medical and nursing practitioners in the primary health sector to identify developmental deviations.1 We used the NHMRC recommended framework and developed an alternative programme of child health screens targeted to detect neurosensory impairment in children born ⱕ31 weeks gestation. It was anticipated that this approach would provide families of children with complex needs an ongoing platform to develop the multi-agency links required to provide the necessary range of services throughout childhood. The programme was not designed to replace tertiary follow-up programmes but as a sustainable dual level screening programme between tertiary and primary health services. This model ensured a broader resource and geographical capacity to optimise child-tracking, early identification and longitudinal collection of neurosensory disability. The primary study hypothesis is that a preterm-targeted screening programme used by general practitioners in a primary health care setting will be an effective method of identifying the
disability status of children born ⱕ31 weeks gestation. The aim of this study was to assess the efficacy of collecting disability information from the primary health sector. We compared the preterm-targeted programme with the NHMRC programme in detecting neurosensory impairment in children born ⱕ31 weeks gestation at 12 months corrected age for (i) identification of disability; and (ii) as a screening tool. A secondary goal was to assess the effects of the programme on the incidence of parental stress, health service use and programme helpfulness in managing the child’s health care.
Methods Participants Infants born ⱕ31 weeks gestation, whose family could nominate a general practitioner and who would return for a 12-month NICU paediatric assessment, were eligible for enrolment 2 weeks prior to NICU discharge. (Fig. 1) Infants were enrolled from November 2002 until December 2003 from Queensland’s three NICUs (Royal Brisbane Women’s Hospital n = 108, Mater Mothers’ Hospital n = 66 and The Townsville Hospital, n = 28). Baseline infant and maternal characteristics
Journal of Paediatrics and Child Health 44 (2008) 424–431 © 2008 The Authors Journal compilation © 2008 Paediatrics and Child Health Division (Royal Australasian College of Physicians)
425
Primary health preterm infant screening
MA Pritchard et al.
Table 1 Baseline infant and maternal characteristics of enrolled infants Child characteristics Variable
Male gender Singleton Gestation (weeks) mean (SD) Birthweight (g) mean (SD)
