Professional Documents
Culture Documents
Halliday 1997
Halliday 1997
fects present at birth. In addition, a 12-year summary this study was begun to determine use of genetic coun-
report of birth defects in Victoria [Riley and Halliday, selling services and whether the type of birth defect
1996] showed that there is an equal or higher birth was related to rate of utilization. Was there greater use
prevalence of many of the sentinel defects compared to when there was a risk of recurrence, or when the ex-
other reports [EUROCAT, 1995; International Clear- pertise of a medical geneticist would be important in
inghouse, 1993], almost certainly reflecting a high level the management of the infant? A comparison to cases
of ascertainment. less likely to warrant referral was to be made.
Many birth defects have a recurrence risk or are We also wanted to look at a possible source of refer-
recognisable genetic disorders, making it desirable for ral, that is, at the hospitals at which the babies were
the parents to have genetic counselling. Such counsel- born. In addition, we examined variables that might
ling aims to provide realistic estimates of recurrence have influenced rates, such as mother’s country of birth
risk, independent of whether the parents plan another and maternal age. These had been shown to be impor-
pregnancy or not. A genetic counselling service should tant predictors of use of prenatal diagnostic testing in
give information on options pertaining to future repro- Victoria [Halliday et al., 1995].
duction and facilitate autonomous reproductive choice
[Chadwick, 1993; Clarke, 1993]. In case of other birth
defects, an attempted early, accurate diagnosis of a ge- MATERIALS AND METHODS
netic disorder or syndrome by a medical geneticist is Study Population
important for management and to alleviate parental
anxiety associated with lack of understanding and This study was undertaken with the use of the Birth
knowledge. Defects/Congenital Malformations Register (BDCMR)
The Victorian Clinical Genetics Service (VCGS) pro- at the Perinatal Data Collection Unit (PDCU), and the
vides a statewide genetic counselling service at the computerised Genetics File (GF) at the VCGS, located
Royal Children’s Hospital, at two major obstetric at the Murdoch Institute, Royal Children’s Hospital.
teaching hospitals, and, more recently, at several outer All babies born with a defect in 1991 and 1993, and who
metropolitan and country hospitals. Records on those were registered at the BDCMR, were subjects of this
seen by one of the six medical geneticists or six genetic study. Conditions are classified using the British Pae-
counsellors are held in a central register at the Mur- diatric Association Classification of Diseases, compat-
doch Institute. Therefore, it has been possible to link ible with the ICD9. The ICD9 coding of birth defects is
records on babies (or fetuses) with birth defects re- a routine task performed independently of this study
corded at the PDCU with those at the VCGS. by skilled medical record administrators and, when
This study was set up to determine how often there necessary, a consultant medical officer.
had been use of genetic counselling services by parents The BDCMR collects data from multiple sources on
after the birth of a baby with a defect. The birth defect all defects found in live births, still births, and termi-
was categorised as carrying a high, moderate, or low nations of pregnancy. Three isolated ‘‘minor birth de-
risk of recurrence or genetic cause and, therefore, war- fects’’ were deleted from the database, i.e., inguinal
ranting referral to a genetics service at the same level hernia, hydrocoele, and undescended testes. This ex-
of importance, i.e., high, moderate, or low need. clusion was done to prevent searching for referrals
A study of the use of genetic services for neural tube when referral was not appropriate.
defects in British Columbia [Sadovnick et al., 1987] The GF has been in existence since the 1970s and
showed that utilization of services may be influenced contains approximately 21,000 files with information
by factors that do not necessarily influence recurrence, on families or individuals. Anyone who has been seen
e.g., whether the infant was stillborn or live born. The by a member of the genetics team, at any one of the
authors found that, when the infant was live born, ser- clinics around the state, is entered into the GF system.
vices were used more frequently than when the infant All GF data were included in the search, which allowed
was stillborn. Similar findings were made in The Neth- for checking for utilization of genetic services that had
erlands [Cornel et al., 1992], where the overall use of eventuated prior to the birth date of the baby under
genetic services after the birth of a baby with a defect review, because of a family history of a genetic disor-
was 16.8%. der, advanced maternal age, or abnormality detected in
A small interview-based study [Hobus et al., 1995] on the current pregnancy.
why parents eligible for genetic counselling did or did
not seek it reported that knowledge of availability of Variables Selected
genetic counselling and recurrence risks were strong
indicators for uptake. However, there was evidence From the BDCMR and the linked perinatal data col-
that some couples did not want counselling or to deal lection at the PDCU, we selected variables available on
with the ‘‘burdensome’’ decision-making process in fu- all of the 1991 and 1993 cases. The variables required
ture pregnancies. Some parents were seen to resist for the analysis and linkage with the data at the VCGS
counselling because of irrational fears and uncertain- were mother’s surname, mother’s first name, baby’s
ties. surname, suburb, postcode, mother’s date of birth, ba-
Other than these studies, there are few publications by’s date of birth, hospital of baby’s birth, sex of baby,
to indicate, at the population level, how many parents and birth defect codes and descriptions. Maternal age
are seeking, or being given access to genetic counsel- at delivery could be calculated from knowledge of the
ling services after the birth of a malformed baby. Thus, mother’s date of birth and the baby’s date of birth.
Genetic Counselling for Birth Defects 5
Record Linkage patients also. We could not determine from these hos-
pitals the number of public and private patients, be-
A three-tiered approach was used to match records cause these data are not collected at the PDCU; there-
from the BDCMR database with records in the VCGS; fore, public hospitals with private facilities were cat-
1) surname enquiry, enabling user to search by sur- egorised as ‘‘metropolitan public.’’
name; this screen also provides the first name and
postcode; 2) patient enquiry, providing further infor- Statistical Analysis
mation on address, date of birth, and information on
the relationship between family members; and 3) diag- SPSS for Windows, Version 6.1, was used for the
nosis screen, providing the diagnosis on the patient. production of descriptive statistics, which included fre-
There was an attempted match with each record of a quencies and cross tabulations. EpiInfo Version 5.0
birth defect to a record at the VCGS up to mid-1995. In was used for relative risk calculations, x2 tests, and
about 90 cases, the ‘‘hard copy’’ of the genetics file was linear trend analysis.
checked for information as whether the family had ac-
tually been seen for counselling, or if the baby had been RESULTS
seen as a ward consult only, or if only a tissue sample In 1991, 2,253 babies and in 1993, 2,154 babies were
had been collected. registered in the BDCMR. Using the detailed record-
Coding linkage process between the BDCMR and the VCGS,
we attempted to match all but 72 cases with a birth
Every birth defect was assigned a code according to defect with cases in the VCGS genetics file system. The
whether or not the family needed genetic counselling. 72 unmatched cases were recorded in the BDCMR but
These categories were 1) high need, because of a risk of had insufficient identifying information, most being
recurrence or for diagnosis of a genetically caused terminations of pregnancy before 20 weeks of gesta-
anomaly, e.g., chromosomal abnormalities, single gene tional age and without identified names.
defects, dwarfing conditions, and many of those with Table I shows that most birth defects fell into the
multiple birth defects; 2) moderate need, e.g., cleft lip/ moderate need category, 44.5% of 1991 and 47.2% of
palate, neural tube defects; 3) low need, e.g., congenital 1993 birth defects. In 1991, 622 of 2,253 (27.6%) cases
dislocated hip, obstructive defects of renal pelvis. Ap- were in the high need category, and 25.6% of the 1993
pendix A shows the details of the coding groups used. cases were similarly categorised.
The decision to assign a multiple birth defect to cat- Table I also shows that the category of birth defect(s)
egory 1 or 2 depended on the pattern of birth defects influenced the number of times the genetic service was
and the probability that this combination represents a accessed. If an infant was born in 1991 with a defect(s)
diagnosable syndrome. The coding system for this categorised with a high need for referral, there was a
study was developed at the PDCU by a medical officer 39.7% chance that the parents would be seen (within 5
(C.R.) and medical geneticist (A.B.), who are both ex- years) by the VCGS. If the baby was born in 1993, there
pert in recognising patterns of birth defects [POSSUM was a 42.5% chance (within 3 years). This is not a sig-
Version 4, 1996] and genetic abnormalities. The re- nificant difference (relative risk 1.07; 95% confidence
searcher who performed the record linkage (O.G.) was interval 0.93–1.23). As the apparent need for referral
blinded to this coding system in order to eliminate bias declines, the actual utilization rate declines corre-
in searching more extensively for cases assigned the spondingly to less than 5%.
first code. Not included are 37 cases born in 1991 and 28 cases
The 130 hospitals at which babies were born were born in 1993 that came to the attention of the VCGS,
grouped into the following categories: 1) teaching hos- either through ward consultation or through labora-
pitals, 2) metropolitan private hospitals, 3) metropoli- tory tests. These parents did not necessarily receive
tan public hospitals, 4) country private hospitals, 5) genetic counselling.
country public hospitals, 6) other (home births, inter- It is important to note that most of the genetic ser-
state and overseas). In many cases, the baby would vice use was within 3 years of birth; it was possible to
have been transferred out of the hospital of birth to one do only a 3 year follow-up for the 1993 babies (1993 to
of the two major hospitals in which there was a paedi- 1995, inclusive). Table II shows the proportion found
atric service. each year for the 1991 births.
It was difficult to categorise some metropolitan pub- Use of the genetic service prior to the birth of the
lic hospitals; many of these facilities cater to private baby in the study year was counted in the total utili-
TABLE I. Need for Referral and Utilization Rates, 1991 and 1993
1991 1993
Need for referral Need for referral
High High Mod Low Total High Mod Low Total
Number in BDCMR 622 1,003 629 2,253 552 1,016 586 2,154
(row %) (27.6) (44.5) (27.9) (100) (25.6) (47.2) (27.2) (100)
Use of VCGS 247 106 27 380 235 86 28 349
(column %) (39.7) (10.6) (4.3) (16.9) (42.5) (8.5) (4.8) (16.2)
6 Halliday et al.
TABLE II. Timing of Use of Genetic Services than 3 years after the birth of the 1991 baby, i.e., in
Use of the VCGS
1994 or 1995. If we assume that this figure also holds
by those in true for the 1993 babies in 1996 and 1997, there would
Year of first appearance 1991 BDCMR be an increase in the total referral rate from 16.2% to
in VCGS register (%) 17.1% (Table I).
Prior to birth of 1991 baby, The hospital of baby’s birth and the genetic service
i.e., previous family history referral rate are shown in Table III. It was important to
or abnormality detected separate out the two teaching hospitals with genetic
in pregnancy 18 counselling services (B and C) for this analysis.
1991 48 Teaching hospital A has no genetic clinic and there-
1992 18
1993 7 fore no consultant medical geneticist on site at any
1994 3 time, unlike the other two teaching hospitals. There
1995 2 was a slight increase (nonsignificant) in uptake rate in
Unknown 4 1993 for the high need category. There was a reduction
in use of genetics service at teaching hospital B by
1993. At teaching hospital C, there was increased uti-
zation rate. The cases seen prior to the birth of the baby lization by high need cases and maintenance of the
in the study period could be divided into three catego- relatively substantial rate of counselling in the moder-
ries: mother seen in the current pregnancy because of a ate need category. Utilization rates in country private
prenatally diagnosed abnormality (25% of cases), hospitals were lower than elsewhere in 1991 but had
mother seen before or during early pregnancy for ad- increased by 1993.
vanced maternal age (12%), and those seen because of It was thought that the mother’s country of birth
a family history of a birth defect (63%). This last cat- might influence whether parents were referred for ge-
egory included cases in which the previous abnormality netic counselling. Therefore, the data were stratified
was the same as the current one, whereas others had a according to country of birth, as an indicator of lan-
dissimilar birth defect. guage and perhaps cultural barriers to service access;
For those with a previous dissimilar birth defect, and 1991 and 1992 results were pooled owing to small num-
for those seen because of advanced maternal age bers (Table IV). There was no significant difference be-
(AMA), a sample of genetics files was reviewed to de- tween utilization rates for women from different coun-
termine whether these women were seen again after tries of birth. (x2 4 8.28; P 4 0.14; df 4 5).
the birth of the current baby. Among the women who Utilization rates for women of different age groups
were seen before or during early pregnancy for AMA, were examined, again using pooled data for 1991 and
10 of 15 women were seen again after the birth of the 1993 (Table V). Many of those in the unknown age cat-
malformed baby. Twelve of the fifteen files checked for egory are those who have had a termination of preg-
women who gave birth to a baby with a dissimilar birth nancy for a birth defect and for whom there is no cal-
defect, were seen again after the birth of the current culated age at delivery. The highest rate of genetic ser-
baby. vice utilization is in this group (35%).
Five percent or 19 service referrals occurred more Women aged 40 years and over were at a higher risk
TABLE III. Hospital of Baby’s Birth, Malformation Category, and Referral Rates: Number in
CMR (% Referred)
1991 1993
Need for referral Need for referral
High Mod Low Total High Mod Low Total
Teaching hospital
A 48 95 55 198 30 90 41 161
(35.4) (7.4) (9.1) (14.6) (40.0) (6.7) (4.9) (12.4)
Teaching hospital
B 81 97 35 213 68 71 25 164
(42.0) (12.4) (0) (21.6) (32.4) (9.9) (0) (17.7)
Teaching hospital
C 147 140 75 362 131 138 62 331
(55.8) (22.9) (9.3) (33.4) (67.2) (21.7) (9.7) (37.5)
Metro-private 69 154 111 334 89 182 133 404
(34.8) (8.4) (3.6) (12.3) (41.6) (7.7) (6.0) (14.6)
Metro-public 121 238 144 503 99 268 143 510
(47.1) (7.6) (2.1) (15.5) (45.5) (6.7) (4.2) (13.5)
Country private 33 60 42 135 19 44 35 98
(27.3) (10.0) (0) (11.1) (42.1) (6.8) (0) (11.2)
Country public 90 203 162 455 89 215 145 449
(34.4) (9.4) (6.2) (13.2) (32.6) (8.4) (4.8) (12.0)
Other, e.g., home
births, interstate 33 16 4 53 27 8 2 37
(51.5) (6.3) (0) (34.0) (35.7) (12.5) (50) (32.4)
Genetic Counselling for Birth Defects 7
TABLE IV. Birth Defects and Use of Genetics Services for Mothers With Different
Countries of Birth, 1991 and 1993 Pooled
Need for referral
Country High Mod Low Total
Australia
No. in BDCMR 633 1444 968 3,045
No. (%) referred 251 (39.6) 129 (8.9) 39 (4.0) 419 (13.8)
Oceania, incl. New Zealand
No. in BDCMR 20 45 23 88
No. (%) referred 6 (33.3) 5 (11.1) 1 (4.3) 12 (13.6)
United Kingdom
No. in BDCMR 54 86 64 204
No. (%) referred 24 (44.4) 9 (10.5) 7 (10.9) 40 (16.7)
Europe
No. in BDCMR 43 101 51 195
No. (%) referred 18 (41.9) 5 (4.9) 3 (5.9) 26 (13.3)
Asia, incl. Middle East
No. in BDCMR 108 183 83 374
No. (%) referred 41 (38.0) 19 (10.4) 3 (3.6) 63 (16.8)
Othera
No. in BDCMR 30 45 16 91
No. (%) referred 13 (43.3) 3 (6.7) 0 (0) 16 (17.6)
a
Other includes North and South America, Africa, and unknown.
of giving birth to a malformed baby (4.7%) than the categorisation of appropriateness or not for refer-
younger women; this relates predominantly to aneu- ral. However, we ensured that the researcher was
ploidy. If chromosome anomalies are removed from the blinded to the high, moderate, and low need categori-
calculations, the percentage malformed drops to 3.3% sation in order to prevent bias introduced by a more
for women over age 40 years. detailed search for those in the ‘‘high need for referral’’
Women aged 40 years and over had the second high- category. Our results in fact showed an appropriate
est total utilization rate (24.7%). The relative risk of decline in utilization rates across the three categories.
utilization for this group, compared to women in the The slight, nonsignificant increase in use of services
20–29 year age group, was 1.88 (95% CI 1.31–2.68; P < by those in the high need category, from 40% to 42.5%,
0.001). The total rate is inflated by the higher propor- might represent an improving situation. To establish
tion of utilization in the moderate and low need catego- whether this is a real increase, it will be important to
ries. In fact, the highest utilization rate in the high reassess the situation in the next year or two.
need category is in the 35–39 year age group (48.8%). The rates quoted include cases in which the genetic
There was a significant linear increase in total utili- service was used by the family prior to the actual birth
zation rates with advancing maternal age (x2 for linear date of the baby in the BDCMR. From a sample of
trend 4 12.14; P < 0.001). The biggest difference in use genetics files accessed directly, we know that, for ap-
was in the low need category. proximately 70–80% of these families, there is another
counselling session. Knowledge of the service from a
DISCUSSION previous encounter would have encouraged couples to
seek access as they wished.
This is a population-based, record-linkage study, al- In a previous study in the same population [Halliday
lowing for a view of the whole state, in which there are et al., 1995], it was shown that hospital of baby’s birth,
approximately 65,000 births per year, more than 2,000 mother’s country of birth, and mother’s age at delivery
of which have a birth defect. The VCGS provides all the were significant predictors of uptake of prenatal diag-
genetic counselling in the state and makes up a genet- nosis. However, in this study, the only significant vari-
ics file on everyone seen in an outpatient setting at one able influencing uptake of genetic counselling was ma-
of their clinics. Record-linkage between the BDCMR ternal age. These three variables are discussed below.
and the genetic filing system at the VCGS allowed us to
measure uptake rates of genetic counselling after a Hospital of Baby’s Birth
baby had a birth defect.
The overall utilization rates for genetic counselling Table III shows that there was a slight increase in
when a baby was born with a defect were 16.9% in 1991 utilization of genetic counselling from teaching hospi-
and 16.2% in 1993 (Table I). The study follow-up time tal A, where, in fact, there was no genetic consultant.
was less for the 1993 babies, but, if we extrapolate from This may reflect a growing awareness of genetic ser-
the 1991 data, another 5% would be referred in the vices in the community.
poststudy follow-up time, resulting in use by 17.1% Meanwhile there had been a slight decrease in use of
rather than 16.2%. With or without these 5%, we have the genetics outpatient clinic at teaching hospital B.
not seen a significant increase in utilization of genetic This was probably due to the desired handling of some
services in Victoria. cases by a counsellor (nongenetic) and obstetrician at a
A limitation of this study was the difficulty arising in concurrent fetal diagnostic unit. This had been neces-
8 Halliday et al.
TABLE V. Maternal Age and Use of Genetic Services, 1991 and 1993
Need for referral
Number of births
Maternal age (years) (% malformed) High Mod Low Total
12–19
No. in BDCMR 4,885 34 75 54 163
No. (%) referred [3.3% in BDCMR] 15 (44.1) 5 (6.7) 1 (1.8) 21 (12.9)
20–29
No. in BDCMR 69,159 426 1023 651 2100
No. (%) referred [3.0% in BDCMR] 162 (38.6) 88 (8.6) 27 (4.1) 277 (13.2)
30–34
No. in BDCMR 39,957 246 552 370 1168
No. (%) referred [2.9% in BDCMR] 96 (39.0) 54 (9.8) 16 (4.3) 166 (14.2)
35–39
No. in BDCMR 13,821 127 221 116 464
No. (%) referred [3.3% in BDCMR] 62 (48.8) 18 (8.1) 7 (6.0) 87 (18.7)
ù40
No. in BDCMR 2,161 55 32 14 101
No. (%) referred [4.7% in BDCMR] 19 (34.5) 4 (12.5) 2 (14.3) 25 (24.7)
Unknown age
No. in BDCMR 286 117 8 411
No. (%) referred 120 (42.0) 23 (19.6) 2 (25.0) 145 (34.3)
sary because of the burden on the limited genetic ser- ploidy) was found, and, correspondingly, the mother’s
vice available. Importantly, there is also a paediatric age was a strong predictor of whether she had genetic
arm of this hospital, where there are experts familiar counselling or not (Table V). There was a twofold in-
with common birth defects and their causes. crease in overall chance of uptake by a woman over age
In the third large teaching hospital (C), there was an 40 years compared to women 20–29 years old. Utiliza-
increase in the genetic counselling service. This would tion, by older women, of genetic services was occurring
have accounted for the increased identification and ap- for the moderate and low need categories of birth de-
parent referral of high risk cases and maintenance of fects at a greater rate than in other age groups,
the relatively substantial rate of counselling in the whereas uptake by those in the high need category was
moderate risk category. The finding of a 67% utiliza- no higher. This may be evidence for older women gen-
tion rate in the high need category at this hospital for erally asking more questions because of their increased
the 1993 birth defects, compared with only 32–45% in knowledge and anxieties about having a baby with a
other hospitals, indicates that there is considerable birth defect. Given the availability in Victoria of rou-
room for improvement generally. tine prenatal diagnosis to women 37 years and older,
Overall, it was pleasing to see an improvement in they would more likely be aware of a genetic counsel-
utilization rates for the high need category in most hos- ling service and seek advice, even when the chance of
pital groupings. The lowest overall utilization rates recurrence was not particularly high. It may also re-
were from the country private hospitals, although flect the fact that older women tend to give birth in the
there was uptake by over 42% of the 1993 cases with a teaching hospitals, and this might account for some of
high need. the higher use observed, especially in the moderate
need category at teaching hospital C, where there is a
Mother’s Country of Birth prenatal diagnostic and genetics service on site.
The analysis of mother’s country of birth (Table IV)
showed that there was no particular pattern of use of CONCLUSIONS
genetic services. Those with a European or Asian back- At least 26% of birth defects are categorised as war-
ground, and therefore potentially non-English- ranting a high need for referral of parents for genetic
speaking, did not appear selectively disadvantaged in counselling, approximately 46% a moderate need, and
this situation. This was rather surprising; women of another 27% a low need. For the study years, only 16–
non-English-speaking background were significantly 17% of parents who had a baby with any birth defect
less likely than women of English-speaking back- were seen by a medical geneticist or genetic counsellor.
ground to have amniocentesis or chorion villus sam- However, where resident genetic counsellors are em-
pling for advanced maternal age [Halliday et al., 1995]. ployed fulltime, as at teaching hospital C, the utiliza-
Perhaps when there is an actual, rather than possible, tion rate for those in the high need category was 67%,
birth of a malformed baby, cultural, language or other and even those in moderate and lower need categories
barriers to access are removed, allowing for greater use were receiving counselling at relatively higher rates.
of genetic services. This would suggest that provision of such a service on
Mother’s Age at Delivery a wider scale would be beneficial to the community and
allow for appropriate management of more families
The expected association of increasing maternal age among whom a baby has a birth defect.
with increasing birth defect rate (primarily from aneu- It was not possible in this study to determine wheth-
Genetic Counselling for Birth Defects 9