Objective
The objective of the study was to determine the rate of return for repeat abortion in relation to postabortion contraceptive method choice 24 months onward from an intervention study.
Study Design
This was a prospective cohort study involving a hospital note search for 510 women 24 months after an abortion.
Results
Women using long-acting reversible contraceptive (LARC) methods (intrauterine device [IUD] and depot medroxyprogesterone acetate) had significantly lower return rates for repeat abortion (6.45%; 95% confidence interval [CI], 4.0–9.8) than non-LARC users, of whom 14.5% returned (95% CI, 9.9–20.2). A Cox proportional hazard analysis showed that the postabortion method choice was significantly related to the likelihood of returning for a repeat abortion ( P = .002), controlling for major demographic factors and previous pregnancy history. Using the pill as a reference group for risk of repeat abortion, the IUD hazard ratio (HR) was 0.36 (95% CI, 0.17–0.77), the depot medroxyprogesterone acetate HR was 0.55 (95% CI, 0.21–1.45), and the HR for all other methods was 1.8 (95% CI, 0.83–3.92).
Conclusion
This study provides strong support for the promotion of immediate postabortion access to LARC methods (particularly intrauterine devices) to prevent repeat abortion.
The potential role of long-acting reversible contraceptive (LARC) methods in reducing unintended pregnancies and abortions has been frequently reported. These methods (injectable, implantable, and intrauterine devices [IUDs]) are safe and effective and associated with very low failure rates because of their low reliance on user compliance for efficacy. Furthermore, contraceptive implants and intrauterine devices are more cost effective than the oral contraceptive pill, are suitable for use by women of all reproductive ages (including nulliparous women), and can be usually be inserted immediately postpartum or after an abortion.
See Journal Club, page 93
Historically, the use of intrauterine methods and implants among contraceptive-aged women has been low because of a range of factors that include misinformation and lack of knowledge among providers and patients, high up-front method and insertion costs, and a lack of skills and training for IUD and implant insertion.
Guidelines and best-practice recommendations are now recommending that clinicians discuss and encourage the use of LARC methods by all reproductive-aged women. National strategies have also been implemented in some countries to increase the use of highly effective reversible contraception by providing methods provided free of charge to women (for example, in Scotland), and large research projects are underway to encourage the use of LARC methods by removing financial barriers (the Contraceptive CHOICE project).
To date, evidence for the link between LARC use and a reduction in repeat abortion has come from observational studies and case note reviews. Few intervention studies designed to promote the use of LARC methods by women presenting for abortion have reported on the impact of method choice on rate of return for repeat abortion.
In 2008, we undertook an intervention study in which the use of LARC methods increased significantly, from 45% at baseline to 61% during the intervention, with a 6-fold increase in the choice of the levonorgestrel intrauterine system (LNG-IUS, from 6% to 36%). The study ran over 10 weeks at a public hospital abortion clinic and was designed to promote use of LARC methods that included depot medroxyprogesterone acetate (DMPA), the LNG-IUS, and copper multiload Cu375 (Cu-IUD).
All methods were available to women free of charge (the LNG-IUS was usually $360 [New Zealand]), information updates were provided to clinic staff about patient suitability for LARC methods, and posters and counseling were used to raise awareness of methods among patients at the clinic. Contraceptive implants were not available at this clinic during the study period. Phone calls made to LARC users at 6 weeks and 6 months after the abortion revealed high rates of method continuation. Of those contacted at 6 months (78%), 81% of women with a LNG-IUS and 74% of those with a Cu-IUD had retained them, and 71% of DMPA users were continuing.
The present follow-up study analyzed data collected via a hospital note search to determine the rate of return for repeat abortion in relation to postabortion contraceptive method choice 24 months onward from the intervention study.
Materials and Methods
This prospective cohort study involved a note search at a public hospital abortion clinic that provides a free service to New Zealand residents. Ethics approval was obtained from the Central Regional Ethics Committee in June 2008 (reference CEN/08/04/015) and permission to conduct the research obtained from the Hospital Research Committee (May 2008). Data matching was undertaken by using unique patient National Health Index numbers to match events for the 510 intervention study participants to hospital abortion clinic discharge data in the 24 month period following the intervention study. Return for repeat abortion was recorded as a dichotomous variable (yes/no) and the date of the procedure used to calculate the time between initial and subsequent abortions.
The survival function for time to repeat abortion was plotted using a Kaplan-Meier survival curve, and a log-rank test was used to test for statistical significance between methods (unadjusted for covariate influence). A Cox proportional hazards model was used to examine the effect of postabortion contraceptive method choice on the likelihood of a return for repeat abortion within the follow-up period, adjusting for the potential confounding effects of covariates (age, parity, previous abortion, ethnicity, socioeconomic deprivation level).
Ethnicity refers to the ethnic group to which an individual belongs and was collected via self-report using the standardized New Zealand 2001 census question. Socioeconomic deprivation is a measure of socioeconomic status that is derived from an individual’s residential address and based on census data, in which scores range from 1 (least deprived) to 10 (most deprived). All covariates were defined at the time of recruitment into the intervention study (in 2008).
Hazard ratios, 95% confidence intervals (CIs), and P values were calculated using SAS 9.2 (SAS Institute, Cary, NC); the survival curve ( Figure ) was produced in R 2.11 (R Foundation, Vienna, Austria). All analyses were conducted on the basis of intention to treat; any reported change in method use at the telephone follow-up during the intervention study was not taken into account. That is, all LARC users at the time of the intervention were analyzed in the LARC group, even if they had since switched to a non-LARC method, and vice versa.
Results
The overall repeat abortion rate at 24 month follow-up was 9.61% (49 of 510; 95% CI, 7.2–12.5). LARC users had a lower return rate with 6.45% (20 of 310; 95% CI, 4.0–9.8) presenting for a repeat abortion during the follow-up period compared with 14.5% (29 of 200; 95% CI, 9.9–20.2) of non-LARC users. Return rates for repeat abortion are presented in Table 1 in relation to self-reported method status at the 6 month telephone follow-up.
