TL;DR
A three-year study of nearly 10,000 pregnant women in Singapore found that approximately 10% sought referrals to terminate unintended pregnancies (UP). The research identifies high-risk demographics—specifically adolescents, women over 40, and those of Indian ethnicity—signaling an urgent need for targeted, community-based family planning and contraceptive counseling.
What Problem Does This Paper Address?
Unintended pregnancy (UP) remains a significant global public health challenge, even in highly developed nations with legalized abortion and accessible healthcare. UP is associated with complex socioeconomic burdens, emotional distress, and increased health risks for both mother and child.
In Singapore, while the overall birth rate is declining, the specific drivers behind women seeking UP referrals in primary care settings have been under-researched. There is a particular "data dearth" regarding how ethnicity, age, and socioeconomic status intersect to influence UP rates. This paper seeks to fill that gap by analyzing who is seeking these referrals at public polyclinics, providing the evidence base needed to design more effective, person-centric reproductive health services.
Methods
The researchers conducted a retrospective study using Electronic Medical Records (EMR) from eight public primary care clinics (polyclinics) in Singapore.
- Population: 9,794 pregnant women aged 15–54 who visited the clinics between July 2017 and June 2020.
- Outcome Measure: The primary outcome was whether a woman requested a gynecologist referral for the "termination of pregnancy."
- Analysis: The team used univariate and multivariable logistic regression to calculate odds ratios (OR), identifying which factors were most strongly associated with requesting a UP referral. They also performed sub-analyses across different age bands and ethnic groups to uncover nuanced trends.
Datasets
The study utilized a large-scale clinical dataset extracted from the SingHealth Polyclinics (SHP) electronic medical record system (Sunrise Clinical Manager). Data processing was handled via the Electronic Health Intelligence System (eHINTS) to ensure de-identification and data integrity. While not a "new" public benchmark in the machine learning sense, it represents a robust, real-world clinical dataset of a multi-ethnic Asian population.
Key Insights and Ideas
- The "U-Shaped" Age Risk: The risk of UP follows a U-shaped curve. Adolescents (<20 years) and older women (>40 years) were significantly more likely to seek referrals. The authors suggest that older women may misperceive a decline in fertility and prematurely stop using contraception.
- Ethnic Disparities: Women of Indian ethnicity had more than double the odds (OR 2.09) of seeking a UP referral compared to Chinese women. This suggests cultural or systemic barriers to effective family planning that require culturally sensitive interventions.
- The Contraception Gap: Alarmingly, only 2.9% of the women seeking UP referrals were recorded as being prescribed contraceptives. This highlights a massive "unmet need" and a failure of the current system to provide "upstream" preventive care.
- Social Vulnerability: Single women were over 7.5 times more likely to request a UP referral than married women. Additionally, lower socioeconomic status (proxied by the Community Health Assist Scheme, or CHAS) was associated with higher referral rates.
Key Results
- Prevalence: 9.94% of the gravid women in the study requested a referral for UP termination.
- Repeat Occurrences: 10.7% of these women had a history of more than one prior unintended pregnancy, indicating a cycle that current healthcare interventions are failing to break.
- Multivariable Analysis Highlights:
- Age >40: Adjusted OR of 5.08 (the highest risk age group).
- Single Status: Adjusted OR of 7.59.
- Indian Ethnicity: Adjusted OR of 2.09.
- Geography: Polyclinics in the eastern and northeastern regions saw higher referral rates, likely due to a higher density of younger families in newer housing estates.
Limitations and Future Work
The authors acknowledge several limitations:
- Selection Bias: The data only includes women who visited public polyclinics. It misses those who go directly to private GPs or specialist obstetricians.
- Lack of Qualitative Context: EMR data provides the "who" and "what," but not the "why." The psychosocial reasons behind the decision to terminate remain unknown.
- Causality: As a retrospective study, it can show associations but cannot definitively prove cause and effect.
Future Work: The authors call for the "SAFE" (Surveillance on Abortion and Family planning in primary carE) approach. They suggest that primary care professionals should use postnatal visits as a critical window to provide contraceptive counseling and advocate for the use of Long-Acting Reversible Contraceptives (LARCs) to reduce the incidence of UP.
Summary updated