Limits...
Using two on-going HIV studies to obtain clinical data from before, during and after pregnancy for HIV-positive women.

Huntington SE, Bansi LK, Thorne C, Anderson J, Newell ML, Taylor GP, Pillay D, Hill T, Tookey PA, Sabin CA, UK Collaborative HIV Cohort (UK CHIC) Study and National Study of HIV in Pregnancy and Childhood (NSHP - BMC Med Res Methodol (2012)

Bottom Line: In total, 2063 women were found in both datasets, representing 23.1% of HIV-positive women with a pregnancy in the UK (n = 8932).There was bias towards matching women with repeat pregnancies (35.9% (741/2063) of women found in both datasets had a repeat pregnancy compared to 21.9% (1502/6869) of women in NSHPC only) and matching women HIV diagnosed before their first reported pregnancy (54.8% (1131/2063) compared to 47.7% (3278/6869), respectively).Through the use of demographic data and clinical dates, records from two independent studies were successfully matched, providing data not available from either study alone.

View Article: PubMed Central - HTML - PubMed

Affiliation: Research Department of Infection & Population Health, University College London, Royal Free Campus, Rowland Hill Street, London, NW3 2PF, UK. susan.huntington.09@ucl.ac.uk

ABSTRACT

Background: The UK Collaborative HIV Cohort (UK CHIC) is an observational study that collates data on HIV-positive adults accessing HIV clinical care at (currently) 13 large clinics in the UK but does not collect pregnancy specific data. The National Study of HIV in Pregnancy and Childhood (NSHPC) collates data on HIV-positive women receiving antenatal care from every maternity unit in the UK and Ireland. Both studies collate pseudonymised data and neither dataset contains unique patient identifiers. A methodology was developed to find and match records for women reported to both studies thereby obtaining clinical and treatment data on pregnant HIV-positive women not available from either dataset alone.

Results: Women in UK CHIC receiving HIV-clinical care in 1996-2009, were found in the NSHPC dataset by initially 'linking' records with identical date-of-birth, linked records were then accepted as a genuine 'match', if they had further matching fields including CD4 test date. In total, 2063 women were found in both datasets, representing 23.1% of HIV-positive women with a pregnancy in the UK (n = 8932). Clinical data was available in UK CHIC following most pregnancies (92.0%, 2471/2685 pregnancies starting before 2009). There was bias towards matching women with repeat pregnancies (35.9% (741/2063) of women found in both datasets had a repeat pregnancy compared to 21.9% (1502/6869) of women in NSHPC only) and matching women HIV diagnosed before their first reported pregnancy (54.8% (1131/2063) compared to 47.7% (3278/6869), respectively).

Conclusions: Through the use of demographic data and clinical dates, records from two independent studies were successfully matched, providing data not available from either study alone.

Show MeSH
Criteria used to match records for women reported to NSHPC and UK CHIC. *Pairs were removed as one or both records in the pair were confirmed as a match in a different pair.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3475121&req=5

Figure 1: Criteria used to match records for women reported to NSHPC and UK CHIC. *Pairs were removed as one or both records in the pair were confirmed as a match in a different pair.

Mentions: Initially, records in NSHPC with a DOB identical to a record in UK CHIC were ‘linked’ and included in a temporary dataset referred to as the ‘linked DOB dataset’. Some women appeared multiple times in this dataset, as they were linked to multiple records with the same DOB. A series of criteria were then used (as described below) to assess which pairs were a genuine match. If records were confirmed as a ‘match’ (i.e. the NSHPC record referred to the same woman as the UK CHIC record) they were merged and moved to a ‘combined dataset’. All remaining occurrences of these women (i.e. as part of other linked pairs) were removed from the linked DOB dataset. The next stage of matching was then undertaken for pairs of records remaining in the linked DOB dataset (Figure 1). If at any stage a record was matched to multiple records, these were reviewed to identify the strongest match.


Using two on-going HIV studies to obtain clinical data from before, during and after pregnancy for HIV-positive women.

Huntington SE, Bansi LK, Thorne C, Anderson J, Newell ML, Taylor GP, Pillay D, Hill T, Tookey PA, Sabin CA, UK Collaborative HIV Cohort (UK CHIC) Study and National Study of HIV in Pregnancy and Childhood (NSHP - BMC Med Res Methodol (2012)

Criteria used to match records for women reported to NSHPC and UK CHIC. *Pairs were removed as one or both records in the pair were confirmed as a match in a different pair.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3475121&req=5

Figure 1: Criteria used to match records for women reported to NSHPC and UK CHIC. *Pairs were removed as one or both records in the pair were confirmed as a match in a different pair.
Mentions: Initially, records in NSHPC with a DOB identical to a record in UK CHIC were ‘linked’ and included in a temporary dataset referred to as the ‘linked DOB dataset’. Some women appeared multiple times in this dataset, as they were linked to multiple records with the same DOB. A series of criteria were then used (as described below) to assess which pairs were a genuine match. If records were confirmed as a ‘match’ (i.e. the NSHPC record referred to the same woman as the UK CHIC record) they were merged and moved to a ‘combined dataset’. All remaining occurrences of these women (i.e. as part of other linked pairs) were removed from the linked DOB dataset. The next stage of matching was then undertaken for pairs of records remaining in the linked DOB dataset (Figure 1). If at any stage a record was matched to multiple records, these were reviewed to identify the strongest match.

Bottom Line: In total, 2063 women were found in both datasets, representing 23.1% of HIV-positive women with a pregnancy in the UK (n = 8932).There was bias towards matching women with repeat pregnancies (35.9% (741/2063) of women found in both datasets had a repeat pregnancy compared to 21.9% (1502/6869) of women in NSHPC only) and matching women HIV diagnosed before their first reported pregnancy (54.8% (1131/2063) compared to 47.7% (3278/6869), respectively).Through the use of demographic data and clinical dates, records from two independent studies were successfully matched, providing data not available from either study alone.

View Article: PubMed Central - HTML - PubMed

Affiliation: Research Department of Infection & Population Health, University College London, Royal Free Campus, Rowland Hill Street, London, NW3 2PF, UK. susan.huntington.09@ucl.ac.uk

ABSTRACT

Background: The UK Collaborative HIV Cohort (UK CHIC) is an observational study that collates data on HIV-positive adults accessing HIV clinical care at (currently) 13 large clinics in the UK but does not collect pregnancy specific data. The National Study of HIV in Pregnancy and Childhood (NSHPC) collates data on HIV-positive women receiving antenatal care from every maternity unit in the UK and Ireland. Both studies collate pseudonymised data and neither dataset contains unique patient identifiers. A methodology was developed to find and match records for women reported to both studies thereby obtaining clinical and treatment data on pregnant HIV-positive women not available from either dataset alone.

Results: Women in UK CHIC receiving HIV-clinical care in 1996-2009, were found in the NSHPC dataset by initially 'linking' records with identical date-of-birth, linked records were then accepted as a genuine 'match', if they had further matching fields including CD4 test date. In total, 2063 women were found in both datasets, representing 23.1% of HIV-positive women with a pregnancy in the UK (n = 8932). Clinical data was available in UK CHIC following most pregnancies (92.0%, 2471/2685 pregnancies starting before 2009). There was bias towards matching women with repeat pregnancies (35.9% (741/2063) of women found in both datasets had a repeat pregnancy compared to 21.9% (1502/6869) of women in NSHPC only) and matching women HIV diagnosed before their first reported pregnancy (54.8% (1131/2063) compared to 47.7% (3278/6869), respectively).

Conclusions: Through the use of demographic data and clinical dates, records from two independent studies were successfully matched, providing data not available from either study alone.

Show MeSH