ARCHIVED: Results: Congenital anomalies surveillance in Canada: results of a 2006-2007 survey on availability of selected data variables in Canadian provinces and territories
Results
Types of Programs in Canadian Provinces and Territories
The types of programs that either collect data or are custodians of databases containing information on congenital anomalies generally fell into three broad categories, summarized below. Figure 2 presents the type of program(s) in each jurisdiction.
Administrative Databases
All jurisdictions, through various provincial and territorial ministries, have administrative databases that contain data on congenital anomalies. Among others, these databases typically include vital statistics (births, deaths, stillbirths), the Discharge Abstract Database (DAD) for hospital admissions, and physician claims data. In principle, these databases can be linked together within jurisdictions to create anonymized individual level information that could be used for congenital anomalies surveillance purposes, such as is routinely done in some jurisdictions as part of the National Diabetes Surveillance System.32 However, it is unknown to what degree this type of analysis is routinely done or whether the capacity to undertake this type of record linkage currently exists in all jurisdictions. Hospital records originating from the birth admission and vital statistics could be used to identify cases with known congenital anomalies at the time of birth, and using the DAD and physicians claims, new diagnoses of congenital anomalies occurring after the birth could also be identified with potential for a long ascertainment period. The limitations of relying on these data sources for congenital anomalies surveillance are similar to those listed for the current Canadian Congenital Anomalies Surveillance System.
Perinatal Database Programs
In addition to administrative data, several jurisdictions also have population-based perinatal databases through their perinatal programs. Currently, these databases exist in Newfoundland and Labrador, Prince Edward Island, Nova Scotia, Ontario (currently in development under the auspices of the Ontario Perinatal Surveillance System), Alberta (not included in this survey), British Columbia and Yukon (not included in this survey). Perinatal databases generally contain data on live births and stillbirths, including detailed information about the pregnancy, such as maternal demographic characteristics and clinical information from the pregnancy and birth. These databases include information on infants and fetuses with congenital anomalies that are diagnosed at or before the time of birth; however, there is no information on pregnancy losses before 20 weeks’ gestation or on pregnancies ending in elective termination. These databases vary in the length of follow-up of infants after birth – the majority would not ascertain cases with congenital anomalies if they were detected after the birth admission, or in some cases, beyond the neonatal period. Some of the programs have ascertainment of infant vital status up to one year after birth.
Fetal Anomaly Databases and Congenital Anomalies Surveillance Programs
In addition to administrative databases and perinatal databases, four jurisdictions also have programs that specifically collect information on congenital anomalies. A unique feature of these four databases is the collection of information on pregnancies that are terminated as a result of a prenatal diagnosis of a congenital anomaly as well as information on spontaneous pregnancy losses that occur prior to 20 weeks due to a diagnosed congenital anomaly.
These four programs include two fetal anomaly databases (in Nova Scotia and Ontario) and two congenital anomalies surveillance programs (in Alberta and British Columbia). The fetal anomaly database programs were created for the purposes of coordinating prenatal care and follow-up for pregnant women undergoing prenatal diagnosis, as well as for conducting clinical research. Women whose developing baby receives a confirmed or suspected antenatal diagnosis of a congenital anomaly are ascertained by these programs and antenatal diagnostic information as well as information on pregnancy outcome (including verification of the anomaly after birth or termination of pregnancy) is collected. Although the fetal anomaly databases are not primarily used for surveillance activities, they are important potential data sources for congenital anomalies surveillance. The two congenital anomalies surveillance programs – the Alberta Congenital Anomalies Surveillance System and British Columbia’s Health Status Registry (HSR) – both conduct ongoing surveillance of congenital anomalies. Both programs have an ascertainment period of at least one-year (the HSR does not have an age limit for congenital anomalies).
Figure 2 Types of programs collecting congenital anomalies data in Canada*
* AD: administrative databases; PD: perinatal databases; FAD: fetal anomaly databases; CASP: congenital anomalies surveillance programs.
Figure 2 - Text Equivalent
ARCHIVED - Figure 2 - Types of programs collecting congenital anomalies data in Canada
Figure 2 illustrates a colourized map of Canada categorizing the types of programs collecting or holding data on congenital anomalies, by province and territory. The legend indicates that programs fall into four broad categories, including: congenital anomalies surveillance programs (CASP); fetal anomaly databases (FAD); perinatal databases (PD); and, administrative databases (AD). Provinces with a combination of dedicated congenital anomalies surveillance programs (CASP) in addition to perinatal databases (PD) and administrative databases (AD) are colour-coded in blue and include British Columbia and Alberta. Provinces with perinatal databases (PD), fetal anomaly databases (FAD) and administrative databases are represented by a white to blue gradient colour-coding, and include Ontario and Nova Scotia. Prince Edward Island and Newfoundland and Labrador are colour-coded white, signifying that these provinces have a perinatal database (PD) and administrative databases (AD). Finally, all other provinces and territories, including Yukon Territory, Northwest Territories, Nunavut, Saskatchewan, Manitoba, Quebec, and New Brunswick are colour-coded in orange, meaning that they rely on administrative databases (AD) for information on congenital anomalies.
Availability of Minimum Data Variables
Figure 3 presents the number of programs (out of 16 responding programs) with at least some type of information on each of the minimum data variables included in the survey. For certain data elements, more than one bar is shown in order to present different aspects of the data (e.g., date of death is represented by one bar for the number of programs with follow-up for infant vital status limited to the time of birth/birth admission, and a second bar to represent the number of programs with follow-up for infant vital status extending beyond the birth admission, from between 7 days to end of first year after birth). The number of programs with some information on these variables was high, particularly for date of birth, place of birth, hospital ID of birth, date of death (with followup of infant vital status at time of birth/birth admission), live/stillbirth flag, mother’s date of birth, mother’s postal code, diagnostic codes, birth weight, infant sex and plurality, for which all or almost all of the responding programs had information.
Figure 3 Availability of selected minimum data variables among the 16 responding programs*
† Follow-up for infant vital status at time of birth/birth admission.
‡ Follow-up for infant vital status between 7 days to end of first year after birth.
§ At delivery (note that one program only has this information at time of prenatal screening because they do not have pregnancy outcome information). SGC – Standard Geographic Code.
# RCPCH – Royal College of Paediatrics and Child Health.
Figure 3 - Text Equivalent
ARCHIVED - Figure 3 - Availability of selected minimum data variables among 16 responding programs
Figure 3 captures the level of availability, among 16 responding programs, of a select number of variables included in a survey aimed at identifying a minimum national dataset for congenital anomalies surveillance. The figure is depicted by a horizontal bar graph, in which the y-axis contains 18 selected variables, and the x-axis represents the level of availability (from 0-16) among all responding programs. Variables included: date of birth (15); place of birth (15); hospital ID of birth (15); date of death – from follow-up of infant vital status at time of birth/birth admission (15); data of death – from follow-up of infant vital status between 7 days to 1st year after birth (8); live/stillbirth (15); termination (9); infant sex (14); mother’s date of birth (16); mother’s postal code (16); mother’s standard geographic code (6); mother’s health region (12); diagnostic codes – any (16); diagnostic codes – International Classification of Diseases (ICD) (13); diagnostic codes – Royal College of Paediatrics and Child Health (RCPCH) (2); birth weight – in grams (15); gestational age (15); and, plurality (16).
Figure 3 shows that the number of programs with at least some type of information for each of the minimum data variables in the survey was quite high. Nevertheless, since the type of information collected by each of the programs was not necessarily uniform, Figure 3 presents the most optimistic level of availability. For example, although 9 of the 15 programs collecting pregnancy outcome information reported that they had information on terminations of pregnancy, there was a variety of data sources from which these programs obtained this information. In some cases, data were only available if the pregnancy termination occurred during a hospital admission and was coded in the DAD. In this instance, the information pertaining to the congenital anomaly would only be associated with the mother’s record and would only contain broad diagnostic codes indicating that the termination was due to a congenital anomaly. Another program had information on terminations of pregnancy (including those done in clinics), but was not able to distinguish between terminations done for an anomaly versus those done for other reasons. A further database did not distinguish between naturally occurring interruptions of pregnancy and elective interruptions of pregnancy.
The survey did not specifically ask programs whether they had a variable for date of death – rather they were asked about collection of data for pregnancy outcome (live/stillbirth) and for the length of follow-up for live births. Of the 15 programs that were collecting information on pregnancy outcome, all had information that allowed reporting on live births and stillbirths, including reporting on infant deaths that occurred during the birth admission. About half (eight programs) had follow-up for live births between 7 days and 1 year after birth, with only three of those programs having follow-up for vital status to the end of the first year after birth.
While all of the programs had information on diagnosis of congenital anomalies, there were differences in the coding of these anomalies. The majority of the programs (13 of 16 programs) use ICD-10 coding.10 This includes the program that does not currently have pregnancy outcome information, and therefore the codes are based on prenatal screening diagnoses. In addition to ICD-10 codes, two programs use Royal College of Paediatrics and Child Health codes,33 which are a more refined adaptation of ICD codes that are recommended in Canada for congenital anomaly surveillance.34 Of the three programs that do not use ICD-10, one uses a detailed list of database-specific codes that were developed by clinical experts in prenatal diagnosis. The other two programs use a drop-down list of anomalies – in one case, this is a list of nine conditions plus an optional ‘other’ field that can be used for text description.
Finally, all of the 15 programs with pregnancy outcome information had information on completed weeks of gestation at birth; however, the method for determining gestational age varied across programs. Five programs had three gestational age estimates – one based on last menstrual period, one based on ultrasound and a third based on clinical examination of the infant at birth. Four programs did not know the method of determination used for their gestational age variable and the remaining programs were based on either one or two of the three methods listed above.
Availability of Recommended Data Variables
Figure 4 presents the number of programs (out of 16 responding programs) with at least some type of information on each of the recommended data variables included in the survey. The number of programs with information on the recommended variables was much lower than for the minimum data variables, and ranged from a high of 10 out of 16 programs for information on diagnostic tests and procedures and nine out of 16 for father’s age and date of last menstrual period, to only one out of 16 programs for information on immigrant status of mother.
Figure 4 Availability of selected recommended data variables among the 16 responding programs
Figure 4 - Text Equivalent
ARCHIVED - Figure 4 - Availability of selected recommended data variables among 16 responding programs
Figure 4 captures the level of availability, among 16 responding programs, of a set of recommended variables (over and above the 18 minimum selected variables) included in a survey aimed at identifying a minimum national dataset for congenital anomalies surveillance. The figure is depicted by a horizontal bar graph, in which the y-axis contains 14 selected variables, and the x-axis represents the level of availability (from 0-16) among all responding programs. Variables included: mother’s immigrant status (1); mother’s ethnicity (8); mother’s aboriginal status (5); father’s date of birth (6); father’s age ( 9); father’s ethnicity (4); father’s aboriginal status (3); text description of anomaly (7); diagnostic tests (10); date of last menstrual period (LMP) (9); prenatal screening (8); mother’s education (6); mother’s occupation (2); and, father’s occupation (2).
Similar to the caveat for the availability of minimum data variables, Figure 4 presents the most optimistic level of availability for the recommended data variables included in the survey. For instance, 10 programs reported that they collected data on diagnostic tests and procedures; however, only six of those programs actually collect information on test results for all subjects. The other four programs only have the data available if they are captured in the DAD during a hospital admission. Given that many of these tests would only be done during the context of an ambulatory visit, the latter programs would not uniformly have the information available for all subjects in their database.
The availability of information on parental ethnicity and socioeconomic variables is presently quite low in Canadian jurisdictions. Eight programs reported some type of information on maternal ethnicity and four programs reported some type of information on paternal ethnicity. An even lower number of programs specifically indicated that they had information on mother’s (five out of 16 programs) and father’s (three out of 16 programs) Aboriginal identification (First Nations, Inuit or Métis). Six programs reported having information on mother’s level of education; however, several noted that this information tended to be poorly documented in medical records. Only two programs reported having information available on maternal and paternal occupation.
Page details
- Date modified: