Population-based linked administrative databases were accessed at ICES (formerly known as the Institute for Clinical Evaluative Sciences) in Toronto, Ontario to conduct this retrospective study. ICES is an independent, non-profit research institute whose legal status under Ontario’s health information privacy law allows it to collect and analyze health care and demographic data, without consent, for health system evaluation and improvement. ICES houses data on individuals eligible for coverage under the Ontario Health Insurance Plan (OHIP). OHIP provides publicly funded health care coverage at no cost to those with Canadian citizenship, permanent resident status and some temporary residents with work permits.
Study design and inclusion/exclusion criteria
We took a retrospective longitudinal cohort approach and enumerated all induced abortions occurring since migration (i.e., since receiving permanent resident status) to Canada, measured per 100 person-years (PY) of follow-up time. Given our study design, we did not estimate annual abortion rates per 1000 women, as is typically seen in cross-sectional or repeated cross-sectional studies. Persons who indicated their sex to be female were included in the study if they immigrated to Canada between April 1, 1991 and December 31, 2012, were registered for OHIP and resided and were alive in Ontario for at least one year between 1 April 1991 and 31 March 2014 following arrival. For each year a female met the stated eligibility criteria and was between the ages of 15–44 years, a year of follow up time was assigned and summed in PY. Induced abortions were enumerated for females who contributed a PY of follow-up time in the year abortion(s) occurred. To examine induced abortion rates prior to a birth, the population included those who delivered up to three or more consecutive singleton births in Ontario hospitals between April 1993 and March 2014. Those who did not have all their births in Ontario were excluded from birth order analyses, along with all their children.
This study does not include asylum seekers awaiting a refugee determination hearing which decides their eligibility for permanent residency. Asylum seekers are eligible for federally funded health care services while they wait for their hearing (i.e., Interim Federal Health Program). This study also does not include migrants: with temporary (1 year) work permits who fill Canada’s short-term labour needs and are eligible for provincially funded health care; with temporary student permits who may be eligible for student health care plans; who are undocumented (estimated at ½ million, the majority initially entered Canada legally but have overstayed their permits [16]) who have limited or no access to publicly funded health care.
Data sources
Several population-based administrative databases were linked to perform this study. These datasets were linked using unique encoded identifiers and analyzed at ICES.
The Ontario portion of the Immigration, Refugee and Citizenship Canada Permanent Resident Database (IRCC-PRD) contains the legal immigration records for all individuals who obtained permanent residency in Canada between January 1985 and December 2012 and intended to reside in Ontario. About 86% of individuals in the Ontario portion of the IRCC-PRD were linked to Ontario’s healthcare registry consisting of Ontarians eligible for publicly funded healthcare insurance in Ontario between April 1, 1990 and March 31, 2014 with a valid health card number. There were small standardized differences (< 0.2) between linked and unlinked individuals across nearly all sociodemographic variables and regions of birth indicating that the linked individuals were largely representative of the original IRCC-PRD [17].
Reporting of induced surgical or pharmacological abortions in publicly funded clinics is mandatory and captured in the Canadian Institute for Health Information’s (CIHI) Discharge Abstract Database (DAD - hospital acute care visits), Same-Day Surgery and the National Ambulatory Care Reporting system (NACRS - emergency department visits). Out-of-hospital abortions are captured using OHIP billing codes for surgical abortions. The CIHI DAD was also used to identify hospital births to immigrant females and their birth characteristics including birth order.
The Office of the Registrar General’s Vital Statistics Death registry (1991–2013), supplemented by mortality recorded in the healthcare registry and other administrative databases was used to identify any deaths among immigrants included in the initial population. For the year in which a death occurred, follow-up time was calculated from January 1 to the date of death that year.
Variables
Induced abortion (henceforth “abortion”) was defined as any surgical or pharmacologically induced termination of pregnancy in the absence of a diagnosis of spontaneous abortion. Given the single-payer health care system in Ontario and the use of population-based databases (rather than self-reported surveys), we report on all induced abortions except a minority conducted in private clinics. If two abortion procedures were recorded within 40 days of each other for a given woman, they were considered the same event and first abortion date was used. CIHI diagnosis codes and OHIP billing codes for induced abortion are described elsewhere [18]. Late induced abortions were any abortions occurring at 15 weeks or more gestation identified using fee code S785. Late terminations are eligible for publicly funded payment if gestational age is confirmed by ultrasonography. In 2017 Mifesgymiso (i.e., Mifespristone) for pharmacologically induced abortion became available free of charge in Ontario, [19] however no abortions were carried out using this drug during the time period of this study. Medically indicated abortions (i.e., for health reasons) could not be excluded since Canadians do not have to provide a reason for abortion. Consequently, Canadian statistics on abortion reasons could not be found but 2018 data from Florida (USA) indicates ~ 1% of all abortions in that state were due to a life endangering physical condition or serious fetal genetic defect/deformity/abnormality. An additional 3% were performed to protect the physical/emotional/psychological health of the mother (not life endangering); while the remainder were due to rape/incest (< 0.1%), social/economic reasons (20%) or no reason provided was provided (75%) [20].
The exposure of primary interest was region of birth categorized using country of birth as reported in the IRCC-PRD and the United Nations geographical classification system which categorizes countries into 22 sub-regions [21]. Sub-region categories were explored for similarity in induced abortion rates to identify geographically adjacent sub-regions or sub-regions with similar levels of “development” for which aggregation would not mask important heterogeneity. Based on this exploration some sub-regions were aggregated into larger regions.
We also examined abortion rates by birth order which was defined as the complete sequence of up to three consecutive live births from the same mother since there were few females that had more than three births. At the time of each pregnancy the alive status of previous siblings was determined using the vital statistics registry and in-hospital death data. If prior siblings were not alive by the estimated conception date of the current live birth, the birth order was reduced for each subsequent live birth.
Covariates as reported in the IRCC-PRD included: country of birth, age at arrival (continuous and categorized as 0–14, 15–19, 20–24, 25–29, 30–34, 35–39, 40-44 years), year of arrival (continuous and categorized as 5-year intervals), education level at arrival (0–9 years, 10–12 years, 13+ years, trade certificate/non-university diploma, university degree), marital status at arrival (single, married/common-law, separated/divorced/widowed), official language ability at arrival (English and/or French, neither English or French), refugee status (refugee, non-refugee immigrants [economic and family class immigrants]) and years residing in Ontario (5-year intervals). Residential neighborhood income quintile (1 = lowest to 5 = highest) for the year after arrival was reported in Ontario’s healthcare registry and utilized for analysis. The year after arrival was used to capture residential income quintile for immigrants arriving later in the year.
Analyses
Population counts and proportions were calculated for year of arrival, age of arrival, education at arrival, income quintile the year after arrival, marital status at arrival, official languages at arrival, refugee status and years residing in Ontario stratified by region of birth.
For all regression analyses unadjusted and adjusted rate ratios (RR and ARR, respectively) were estimated using Poisson regression with the outcome specified as the count of induced abortions and the offset specified as the time in years residing in Ontario since arrival for each female. Crude abortion rates (per 100 PY) and associated 95% CIs were estimated for females from each region and each country if the country population was ≥500 females. Unadjusted and adjusted rate ratios were estimated for region of birth specified as the independent variable and confounders included year of arrival, age at arrival, education at arrival, neighborhood income at arrival and refugee status. Adjusting for year and age at arrival were related to our hypothesis that pre-migration exposures may impact induced abortion after arrival. Specifically, those arriving as adults may be more likely to adhere to the sexual and reproductive health norms related to their country of birth at the time they emigrated. For those arriving younger, sexual and reproductive health norms may be shaped by both the country of origin (through their parents) as well as by the Canadian context.
Persons born in the United States/ Northern & Western Europe/Australia & New Zealand were chosen as the comparator given similar levels of development. Study subjects with missing outcome and covariate data were excluded from regression analyses. We did not adjust for marital status at arrival since this is likely to become misclassified with increasing length of stay, particularly for those arriving younger and single. In secondary analyses, unadjusted and adjusted Poisson regression was conducted for each world region of birth to examine the association between abortion and each of year, age, education, income quintile (at arrival) and refugee status.
Birth order analyses were restricted to mothers who had three or more consecutive singleton births in Ontario between 1993 and 2012. Induced abortion rates (per 100 births) and 95% CIs were estimated for abortions occurring prior to the 1st birth, between the 1st and 2nd births (i.e., prior to the 2nd birth) and between the 2nd and 3rd or higher births (i.e., prior to the third or higher birth) for each region of birth. RRs and 95% CI were also estimated comparing each birth order to that prior to the 1st birth.