Send letter to journal:
Re: Coverage of neonatal screening: Failure of coverage of failure of information system

Dear Editor,

We were very interested to read the paper by Dr. Ades and colleagues[1].

As the authors correctly pointed out, in 1985 Birmingham established a bi-directional electronic link between the neonatal laboratory database and the then five child health systems covering the City of Birmingham (approximately 16,000 births)[2]. Babies’ details, each with a unique system number, are transferred each night to the laboratory database. A pre-printed label with the baby’s details and a unique number is used for the screening request, thus enabling the laboratory to book in the screening request via the unique number. Test results for the screening tests (phenylketonuria, congenital hypothyroidism and haemoglobinopathies) are transferred electronically on a batch basis each night to the child health record.

This system, which covers the child health system for Birmingham (Northern and Southern) has demonstrated that it can achieve a >99.9% coverage rate and allows an assessment of screening performance.

In June 2000, we piloted an expansion of this system to extend to Sandwell Health Authority (births 4,000 per annum). Unfortunately the use of the pre-printed label was felt to be impractical in Sandwell. This was mainly due to the timescales and logistics involved with ensuring the label was available to the midwife when the specimen was collected. Instead, we decided to use a system of matching the specimen card against the transferred record on the laboratory database using a specimen matching criteria list. Seven demographics were included in the list:-

1. Baby’s date of birth
2. Baby’s surname
3. Address
4. Birth weight
5. Mother’s surname and forename
6. Mother’s date of birth
7. Rank and gender (multiple births)

A specimen card was only matched if either:-

a)Three of the demographics and the baby’s date of birth were the same (slight misspellings, obvious name transpositions (baby/mother) and obvious date transpositions were accepted).
b)For records where the date of birth was not the same (even if similar) all other demographics provided must have been the same as the specimen card.

For records which did not match a new record was created on the laboratory database.

The results from the pilot project have been very successful. During July 2000-June 2001, the average match rate percentage for matching specimen cards to records on the laboratory database using the matching criteria list was 97.45% (min-max range 95.4%-98.7%).

During the same period, the average number of days for which results were available on the Trust’s Child Health System following birth was 21.65 days (min-max range 19.2-29.6 days).

This has now enabled Sandwell’s Child Health Department (using the Child Health Computer System) to identify babies with outstanding screening results within the recommended national standard of 28 days of age.

As part of a Pathology Modernisation award, we are planning to expand further across the West Midlands Region.These data provide further support to the value of linkage between child health systems and the neonatal screening laboratory data base. As Dr. Ades and colleagues stated, the use of the NHS number as the unique identifier has the potential to provide a national system.

Yours sincerely,

Dr. Anne Green
Consultant Clinical Biochemist
Birmingham Children's Hospital NHS Trust

Mr Jayesh Patel
Information Systems Manager
Sandwell Healthcare NHS Trust

Dr. Helen Grindulis
Consultant Paediatrician
Sandwell Healthcare NHS Trust

Dr. Sue Ibbotson
Deputy Regional Director of Public Health
West Midlands Regional Office

References
(1) AE Ades, J Walker, R Jones, I Smith. Coverage of neonatal screening: failure of coverage or failure of information system. Arch.Dis.Child. 2001; 84: 476-479.
(2) Paul D Griffiths, John Morris, Janet Assheton, Anne Green. An online computerised system for neonatal screening. Elsevier Science Publishers B.V. (Biomedical Division) 1987. Advances in neonatal screening. B.L. Therrell, Jr., editor.