Women’s Knowledge, Attitude, and Practice Regarding Oral Short-Acting Contraceptives After Induced Abortion: A Cross-Sectional Study in Beijing, China Authors: Shan He, Wei Ma, Jie Gao, Juan Zhao Published: 2026 ABSTRACT BackgroundAbout half of all patients seeking abortion report having a history of prior abortions and are at high risk for repeat abortion. Data about postabortion contraception in China are limited. This study investigated the knowledge, attitude, and practice (KAP) of women undergoing induced abortion regarding postoperative oral short-acting contraceptives.Study DesignThis cross-sectional study enrolled women at Beijing Luhe Hospital, Capital Medical University, from December 2023 to February 2024. The minimal calculated sample size was 385 participants. A self-administered questionnaire was developed to collect demographic information and assess knowledge, attitudes, and practices regarding postabortion oral short-acting contraceptives. Scores >70% of the possible maximum score of each dimension were considered as good knowledge, positive attitudes, and proactive practice.ResultsAmong 420 women with questionnaires passing quality control, mean knowledge, attitude, and practice scores were 5.60 ± 2.96/12 (46.67%), 31.27 ± 4.54/45 (69.49%), and 21.70 ± 2.94/25 (86.80%). Lower education (junior high school or below) was associated with lower odds of sufficient knowledge and positive attitudes. Higher knowledge scores were associated with higher odds of positive attitude. Higher knowledge and attitude scores and parity >1 were independently associated with higher odds of adequate practice, while age 36–39 was associated with lower odds.ConclusionIn Beijing, women after induced abortions had poor knowledge and attitudes scores, but a proactive intended practice toward short-acting contraception. Specific knowledge items were identified as requiring improvement. Educational interventions should particularly target women with a lower socioeconomic status to reduce repeated induced abortions. Background Induced abortion is the intentional medical or surgical termination of pregnancy for any reason.1 Unintended pregnancies account for about 50% of the 208 million pregnancies reported worldwide yearly.1,2 About 40%–60% of unintended pregnancies and about 30% of all pregnancies are reported to end in abortion.1,2 Both medical and surgical options for termination of pregnancy may be offered to patients seeking abortion in the first and second trimesters.1–3 Patient preference will vary. Mortality and other major complications are rare at all gestational ages in countries with access to safe abortion care.3 Postabortion care includes any or all of the following, as needed or desired: follow-up visits, management of regular adverse effects or complications, and contraception services.4 About half of all patients seeking abortion reported having a history of prior abortions and are at high risk for repeat abortion.3 The patients should initiate contraception as soon as possible, as ovulation may resume as early as 8–21 days after abortion.2–5 For patients having surgical abortions, initiating contraception on the day of the procedure should be recommended; all methods, excluding hysteroscopic sterilization, diaphragm, and cervical cap, can be considered.3,4 For patients having medical abortions with mifepristone and misoprostol or with misoprostol alone, hormonal contraception can be started immediately after the first pill of the medical abortion regimen.1,4,6 The use of short-term contraceptives following an abortion can effectively prevent pregnancy and promote endometrial repair. It can help reduce the risk of uterine adhesions and offer enhanced fertility protection, particularly for women desiring future pregnancies.6,7 Self-management after induced abortion is central to reducing the risk of recurrent unintended pregnancy, but some women who undergo induced abortion for medical reasons also need to plan for a desired future pregnancy and may therefore have different contraceptive priorities.1–4 In clinical practice, short-acting oral contraceptives comprise both combined oral contraceptives and progestogen-only pills, which differ in contraindications, safety profile, and counseling requirements, especially in the immediate postabortion period.1–4 Proper knowledge of contraception and pregnancy is necessary to consult the right healthcare professionals promptly regarding postabortion contraception, adopt appropriate health behaviors, and achieve optimal outcomes. Because oral contraceptives are taken at home, without direct supervision, adequate knowledge and favorable attitudes are essential to support proactive and consistent use. Knowledge, attitude, and practice (KAP) is one of several behavior‑change frameworks that propose different causal architectures (eg, Theory of Planned Behavior, Health Belief Model, and COM‑B) and use structured surveys to generate quantitative and qualitative information about gaps, misconceptions, and misunderstandings on a specific subject in a defined population.8,9 A study in Ethiopia showed that the practice of postabortion contraception was high, at 78%.10 A study in India showed that women admitted for self-induced abortion had a poor KAP of postabortion contraception.11 Another study from India showed that women seeking induced abortion had good knowledge and favorable attitudes toward contraception but poor practice.12 China has undergone rapid demographic and policy changes over recent decades, including the transition from the one‑child policy to a two‑child and subsequently three‑child policy, together with increasing government emphasis on fertility support and reproductive health protection. Within this context, induced abortion remains legally available in accredited medical institutions, but there is growing policy attention to the prevention of unintended pregnancy, reduction of repeated abortions, and provision of standardized postabortion counseling and contraception services. National and expert consensus documents in China have highlighted the importance of postabortion care, including timely initiation of appropriate contraceptive methods to protect endometrial health and preserve future fertility.13,14 No study has characterized Chinese women’s KAP toward short-acting oral contraceptives at the time of induced abortion, which is the most relevant decision window. Access to high‑quality counseling and contraceptive services, as well as women’s understanding and acceptance of postabortion contraception, may vary according to socioeconomic and regional factors. Clarifying women’s knowledge, attitudes, and practices regarding short‑acting oral contraceptives in this policy environment is therefore essential to inform targeted educational and service interventions. One study pointed out poor access to and utilization of postabortion care services in Guangzhou (China).15 Therefore, this study aimed to examine the KAP of postabortion women regarding the use of short-acting oral contraceptives. The results could help identify gaps in knowledge that could be targeted by future educational interventions to improve attitudes and practice. Methods Study Design and Participants This cross-sectional study enrolled women admitted to the hospital because they were seeking an induced abortion at Beijing Luhe Hospital, Capital Medical University, from December 2023 to February 2024. This study was approved by the ethics committee of Beijing Luhe Hospital, Capital Medical University (approval 2023-LHKY-125-02). All participants signed the informed consent form before completing the questionnaire. The inclusion criteria were 1) ≥18 years of age and 2) consulting for induced abortion. The exclusion criteria were 1) having already undergone medical abortion, or women with contraindications to oral short-acting contraceptives. Sample Size The minimal sample size was estimated using Cochran’s sample size formula for survey studies:16 \documentclass[12pt]{minimal} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \DeclareFontFamily{T1}{linotext}{} \DeclareFontShape{T1}{linotext}{m}{n} {linotext }{} \DeclareSymbolFont{linotext}{T1}{linotext}{m}{n} \DeclareSymbolFontAlphabet{\mathLINOTEXT}{linotext} \begin{document}$$n = {{{Z^2} \times p\left({1 - p} \right)} \over {{d^2}}}$$\end{document} where Z2 is the confidence coefficient, p is the proportion, and d is the margin of error. The sample size is maximized when p = 0.5. A 95% confidence interval involves a Z-value of 1.96. Precision was assumed at 5%. Hence, a minimum of 385 participants was needed. Questionnaire and Quality Control The questionnaire design was based on the “Expert Consensus on Promoting Endometrial Repair after Induced Abortion”, “Chinese Expert Consensus on Clinical Application of Compound Oral Contraceptives”, and “Contraceptive Service Standards after Induced Abortion (2018)”. After the initial draft of the questionnaire was designed, it was modified based on the opinions of two experts. The questionnaire was pilot-tested on a small scale (79 participants) using the same inclusion and exclusion criteria as the formal survey. The overall, knowledge, attitude, and practice Cronbach’s α (Feldt’s 95% confidence intervals) for the pilot survey were 0.820 (0.759–0.868), 0.833 (0.776–0.849), 0.767 (0.689–0.825), and 0.767 (0.686–0.823), indicating good questionnaire reliability. The Cronbach’s α and item-total correlations are shown in Supplement Tables 1–4. As two items (K11 and K12) were added after pilot testing, the Cronbach’s α and item-total correlations were also analyzed for the questionnaires collected in formal survey. The overall, knowledge, attitude, and practice Cronbach’s α (Feldt’s 95% confidence intervals) for the formal survey were 0.705 (0.665–0.745), 0.813 (0.788–0.838), 0.610 (0.558–0.663), and 0.810 (0.785–0.836), indicating acceptable questionnaire reliability. The Cronbach’s α and item-total correlations are shown in Supplement Tables 5–8. The final questionnaire was in Chinese and collected information in four dimensions, comprising 41 items (Supplementary materials-Questionnaire). The basic information section included 12 items, the knowledge dimension included 12 items, the attitude dimension included nine items, and the practice dimension included eight items. For the knowledge dimension, 1 point was given for a correct answer and 0 points for an incorrect or unclear answer, with a score range of 0–12 points. The attitude dimension used a 5-point Likert scale, with scores assigned based on the degree of positivity. A2, A3, and A9 were scored from strongly agree (5 points) to strongly disagree (1 point), while the remaining items were reverse-scored, with a score range of 9–45 points. In the practice section, P3 investigated the main reasons for not accepting oral short-acting contraceptives after induced abortion, P4 investigated whether long-acting contraceptives were considered after induced abortion if there were no plans for further pregnancy, and P8 investigated how knowledge about contraception and induced abortion surgery was obtained. The remaining five questions used a 5-point Likert scale, with scores ranging from strongly agree (5 points) to strongly disagree (1 point), with a score range of 5–25 points. The participants were recruited through convenience sampling through the outpatient clinics when they were admitted for an induced abortion in the gynecology and obstetrics clinic. The researchers invited the patients to participate in the study. If they accepted, they filled out the questionnaire before the abortion by scanning a QR code. The first page of the questionnaire was the consent form, which had to be completed. The data were collected using the Questionnaire Star platform. The researcher informed the participants that the enrollment would not affect clinical treatment and that all questionnaires were completed anonymously. The participants did not receive any education or guidance before completing the questionnaire. The Questionnaire Star platform was configured to require completion of all items, there were no missing data within collected questionnaires. If any questions arose during the answering process, the research team members were responsible for promptly answering the participants’ questions. After all data collection was completed, the research team members conducted a quality check of the questionnaire. Questionnaires with obvious logical errors or answered with the same option for all items were considered invalid. Statistical Analysis Stata 17.0 (Stata Corporation, College Station, TX, USA) was used for statistical analysis. Continuous variables were expressed as means ± standard deviations and analyzed using Student’s t-test (comparison of two groups) or ANOVA (comparison of more than two groups). Categorical variables were expressed as n (%) and analyzed using the chi-square test. Spearman correlation analysis was used to analyze the correlations between the KAP dimensions. The KAP dimension scores were defined as good knowledge, positive attitudes, and proactive practice when >70% of the maximum possible score (ie, >8.4 for knowledge, >31.5 for attitude, and >17.5 for practice).8,17–19 Multiple logistic regression was used with knowledge, attitude, and practice scores as dependent variables to analyze the independent risk factors for KAP scores. For the practice dimension, the participants were grouped as “Strongly agree” and “Agree” to the question “After an induced abortion, I will follow the doctor’s advice to take short-acting contraceptives orally.” in one group, and the others in another group. The characteristics of the two groups were analyzed. Variables with P <0.05 in the univariable logistic regression analyses were included in the multiple regression analyses. Age, parity, education, and income were included a priori in all multivariable models on substantive grounds, irrespective of their univariable P values, while other candidate variables continue to follow the previous screening approach. Variance inflation factors (VIFs) were calculated for all independent variables to assess multicollinearity, with VIF values close to 1 indicating negligible collinearity and values >5 suggesting potentially problematic multicollinearity that may inflate standard errors and destabilize regression estimates. Two-tailed P-values <0.05 were considered statistically significant. Results Characteristics of the Participants In this study, 442 participants completed questionnaires; 22 questionnaires were answered with the same option for all items and were excluded. Therefore, 420 participants who provided a questionnaire passing quality control (95.02%) were included for analysis. Most participants were 26–35 years old (58.57%), married (67.86%), with one or two previous pregnancies (56.90%), with one child (44.15%), normal body mass index (63.81%), urban residency (64.52%), college education or above (69.29%), employed (69.52%), income 5000–10,000 (43.57%), with medical insurance (88.10%), using contraception (94.05%), and without comorbidities (89.29%) (Table 1).Table 1Characteristics of the ParticipantsVariablesn (%)Knowledge ScoreAttitude ScorePractice ScoreMean ± SDPMean ± SDPMean ± SDPTotal score4205.60 ± 2.9631.27 ± 4.5421.70 ± 2.94Age (years)0.573a0.017a0.211a 18–2574 (17.62)5.59 ± 2.8931.91 ± 4.4522.14 ± 2.57 26–35246 (58.57)5.71 ± 2.9731.52 ± 4.6521.71 ± 3.01 36–39100 (23.81)5.34 ± 3.0130.17 ± 4.1821.34 ± 3.01Marital status0.409a0.284a0.173a Single115 (27.38)5.59 ± 3.0831.38 ± 4.6022.06 ± 2.65 Married285 (67.86)5.67 ± 2.8531.33 ± 4.5421.60 ± 3.01 Divorced/Widowed20 (4.76)4.75 ± 3.7429.70 ± 4.1220.90 ± 3.48Number of pregnancies0.313a0.461a0.350a 060 (14.29)5.20 ± 2.9631.65 ± 4.6622.18 ± 2.96 1–2239 (56.90)5.56 ± 2.9731.38 ± 4.5421.66 ± 2.88 ≥3121 (28.81)5.89 ± 2.9330.86 ± 4.4721.52 ± 3.05Number of childbirths0.441a0.130a0.537a 0127 (30.31)5.35 ± 3.1331.97 ± 4.3021.58 ± 3.12 1185 (44.15)5.62 ± 2.8631.01 ± 4.5721.86 ± 2.73 ≥2107 (25.54)5.85 ± 2.9330.98 ± 4.6321.50 ± 3.09Body mass index0.215a0.240a0.825a Underweight (<18.5 kg/m2)43 (10.24)5.93 ± 3.1331.79 ± 3.9722.07 ± 2.57 Normal (18.5–23.9 kg/m2)268 (63.81)5.62 ± 2.9731.39 ± 4.6121.62 ± 3.01 Overweight (24.0–27.9 kg/m2)88 (20.95)5.69 ± 2.9431.07 ± 4.6021.75 ± 3.05 Obesity (≥28 kg/m2)21 (5.00)4.33 ± 2.5029.48 ± 4.3221.67 ± 2.44Place of residence0.475b0.874b0.306b Non-urban149 (35.48)5.46 ± 2.9131.22 ± 4.5521.50 ± 2.77 Urban271 (64.52)5.68 ± 2.9931.30 ± 4.5421.80 ± 3.03Education<0.001a0.002a<0.001a Junior high school and below44 (10.48)3.77 ± 2.8529.45 ± 4.2719.43 ± 3.55 High school/vocational school85 (20.24)5.22 ± 2.6530.51 ± 4.4121.85 ± 2.58 College and above291 (69.29)5.99 ± 2.9631.77 ± 4.5321.99 ± 2.80Current employment status0.233a0.267a0.030a Employed292 (69.52)5.75 ± 3.0431.13 ± 4.4221.67 ± 3.01 Unemployed21 (5.00)5.33 ± 2.3931.67 ± 3.6021.81 ± 2.99 Self-employed25 (5.95)4.80 ± 2.7730.32 ± 5.2620.56 ± 2.71 Full-time housewife36 (8.57)5.94 ± 2.6032.72 ± 4.8222.94 ± 1.82 Other46 (10.95)4.93 ± 2.9831.32 ± 4.9321.41 ± 3.08Average monthly per capita income<0.001a0.126a0.013a <200012 (2.86)2.50 ± 2.1130.00 ± 3.4419.17 ± 2.79 2000–500075 (17.86)5.16 ± 3.1030.24 ± 4.3621.25 ± 3.47 5000–10,000183 (43.57)5.55 ± 2.7431.35 ± 4.3521.78 ± 2.69 10,000–20,000112 (26.67)5.98 ± 3.0731.92 ± 4.9022.05 ± 2.90 >20,00038 (9.05)6.61 ± 2.9231.39 ± 4.7521.89 ± 2.84Type of medical insurance0.261b0.191b0.067b Insured370 (88.10)5.66 ± 2.9131.38 ± 4.5121.79 ± 2.89 Uninsured50 (11.90)5.16 ± 3.2830.48 ± 4.7220.98 ± 3.24Contraceptive measures0.295b0.358b0.501b Used (multiple choices)395 (94.05)5.64 ± 2.9831.22 ± 4.4821.67 ± 2.96 Condoms348 (82.86)--- Short-acting contraceptives27 (6.43)--- Emergency contraceptives (morning-after pills)104 (24.76)--- Long-acting contraceptives10 (2.38)--- Other37 (8.81)--- Intrauterine device16 (3.81)--- Have not used any contraceptive measures25 (5.95)5.00 ± 2.6132.08 ± 5.3822.08 ± 2.69Comorbidities0.996b0.443b0.658b Yes (multiple choices)45 (10.71)5.60 ± 2.8930.78 ± 4.8321.51 ± 4.09 Diabetes2 (0.48)--- Hypertension7 (1.67)--- Viral hepatitis3 (0.71)--- Heart disease2 (0.48)--- Other12 (2.86)--- Thyroid disease21 (5.00)--- No375 (89.29)5.60 ± 2.9731.33 ± 4.5121.72 ± 2.78Note: aANOVA; bt-test. Table 1Characteristics of the ParticipantsVariablesn (%)Knowledge ScoreAttitude ScorePractice ScoreMean ± SDPMean ± SDPMean ± SDPTotal score4205.60 ± 2.9631.27 ± 4.5421.70 ± 2.94Age (years)0.573a0.017a0.211a 18–2574 (17.62)5.59 ± 2.8931.91 ± 4.4522.14 ± 2.57 26–35246 (58.57)5.71 ± 2.9731.52 ± 4.6521.71 ± 3.01 36–39100 (23.81)5.34 ± 3.0130.17 ± 4.1821.34 ± 3.01Marital status0.409a0.284a0.173a Single115 (27.38)5.59 ± 3.0831.38 ± 4.6022.06 ± 2.65 Married285 (67.86)5.67 ± 2.8531.33 ± 4.5421.60 ± 3.01 Divorced/Widowed20 (4.76)4.75 ± 3.7429.70 ± 4.1220.90 ± 3.48Number of pregnancies0.313a0.461a0.350a 060 (14.29)5.20 ± 2.9631.65 ± 4.6622.18 ± 2.96 1–2239 (56.90)5.56 ± 2.9731.38 ± 4.5421.66 ± 2.88 ≥3121 (28.81)5.89 ± 2.9330.86 ± 4.4721.52 ± 3.05Number of childbirths0.441a0.130a0.537a 0127 (30.31)5.35 ± 3.1331.97 ± 4.3021.58 ± 3.12 1185 (44.15)5.62 ± 2.8631.01 ± 4.5721.86 ± 2.73 ≥2107 (25.54)5.85 ± 2.9330.98 ± 4.6321.50 ± 3.09Body mass index0.215a0.240a0.825a Underweight (<18.5 kg/m2)43 (10.24)5.93 ± 3.1331.79 ± 3.9722.07 ± 2.57 Normal (18.5–23.9 kg/m2)268 (63.81)5.62 ± 2.9731.39 ± 4.6121.62 ± 3.01 Overweight (24.0–27.9 kg/m2)88 (20.95)5.69 ± 2.9431.07 ± 4.6021.75 ± 3.05 Obesity (≥28 kg/m2)21 (5.00)4.33 ± 2.5029.48 ± 4.3221.67 ± 2.44Place of residence0.475b0.874b0.306b Non-urban149 (35.48)5.46 ± 2.9131.22 ± 4.5521.50 ± 2.77 Urban271 (64.52)5.68 ± 2.9931.30 ± 4.5421.80 ± 3.03Education<0.001a0.002a<0.001a Junior high school and below44 (10.48)3.77 ± 2.8529.45 ± 4.2719.43 ± 3.55 High school/vocational school85 (20.24)5.22 ± 2.6530.51 ± 4.4121.85 ± 2.58 College and above291 (69.29)5.99 ± 2.9631.77 ± 4.5321.99 ± 2.80Current employment status0.233a0.267a0.030a Employed292 (69.52)5.75 ± 3.0431.13 ± 4.4221.67 ± 3.01 Unemployed21 (5.00)5.33 ± 2.3931.67 ± 3.6021.81 ± 2.99 Self-employed25 (5.95)4.80 ± 2.7730.32 ± 5.2620.56 ± 2.71 Full-time housewife36 (8.57)5.94 ± 2.6032.72 ± 4.8222.94 ± 1.82 Other46 (10.95)4.93 ± 2.9831.32 ± 4.9321.41 ± 3.08Average monthly per capita income<0.001a0.126a0.013a <200012 (2.86)2.50 ± 2.1130.00 ± 3.4419.17 ± 2.79 2000–500075 (17.86)5.16 ± 3.1030.24 ± 4.3621.25 ± 3.47 5000–10,000183 (43.57)5.55 ± 2.7431.35 ± 4.3521.78 ± 2.69 10,000–20,000112 (26.67)5.98 ± 3.0731.92 ± 4.9022.05 ± 2.90 >20,00038 (9.05)6.61 ± 2.9231.39 ± 4.7521.89 ± 2.84Type of medical insurance0.261b0.191b0.067b Insured370 (88.10)5.66 ± 2.9131.38 ± 4.5121.79 ± 2.89 Uninsured50 (11.90)5.16 ± 3.2830.48 ± 4.7220.98 ± 3.24Contraceptive measures0.295b0.358b0.501b Used (multiple choices)395 (94.05)5.64 ± 2.9831.22 ± 4.4821.67 ± 2.96 Condoms348 (82.86)--- Short-acting contraceptives27 (6.43)--- Emergency contraceptives (morning-after pills)104 (24.76)--- Long-acting contraceptives10 (2.38)--- Other37 (8.81)--- Intrauterine device16 (3.81)--- Have not used any contraceptive measures25 (5.95)5.00 ± 2.6132.08 ± 5.3822.08 ± 2.69Comorbidities0.996b0.443b0.658b Yes (multiple choices)45 (10.71)5.60 ± 2.8930.78 ± 4.8321.51 ± 4.09 Diabetes2 (0.48)--- Hypertension7 (1.67)--- Viral hepatitis3 (0.71)--- Heart disease2 (0.48)--- Other12 (2.86)--- Thyroid disease21 (5.00)--- No375 (89.29)5.60 ± 2.9731.33 ± 4.5121.72 ± 2.78Note: aANOVA; bt-test. Characteristics of the Participants Note: aANOVA; bt-test. Knowledge The mean knowledge score was 5.60 ± 2.96 on a maximum of 12 (46.67%), indicating poor knowledge. Higher knowledge scores were observed with higher education (P < 0.001) and higher income (P < 0.001) (Table 1). The knowledge item with the highest rate of correct answers was K2 (85.95%; “The more times induced abortion is performed, the closer the intervals, and the larger the gestational age, the more likely the endometrium is to be damaged”). The knowledge items with very poor correct answer rates were K12 (5.24%; “Short-acting contraceptives should not be used for a long time, as long-term use can cause great harm to the body”) and K8 (8.81%; “Oral short-acting contraceptives may increase the risk of fetal congenital malformations”) (Figure 1 and Supplement Table 9, including the scores for the individual items).Figure 1Rates of correct responses to the knowledge items.Horizontal bar graph showing rates of correct responses for knowledge items K1 to K12.Horizontal bar graph of rates of correct responses to the knowledge items. X-axis label: percent, range 0.00 percent to 100.00 percent, with tick labels at 0.00 percent, 10.00 percent, 20.00 percent, 30.00 percent, 40.00 percent, 50.00 percent, 60.00 percent, 70.00 percent, 80.00 percent, 90.00 percent, 100.00 percent. Y-axis label: k1, k2, k3, k4, k5, k6, k7, k8, k9, k10, k11, k12. Bar values: k1 equals 70.95 percent; k2 equals 85.95 percent; k3 equals 47.86 percent; k4 equals 72.14 percent; k5 equals 76.67 percent; k6 equals 22.14 percent; k7 equals 50.24 percent; k8 equals 8.81 percent; k9 equals 38.10 percent; k10 equals 40.48 percent; k11 equals 41.67 percent; k12 equals 5.24 percent. Rates of correct responses to the knowledge items. Attitude The mean attitude score was 31.27 ± 4.54 on a maximum of 45 (69.49%), indicating an unfavorable attitude. Higher attitude scores were observed with younger age (P = 0.017) and higher education (P = 0.002) (Table 1). Supplement Table 10 presents the distribution of the answers to the attitude items and the scores for the individual items. Practice The mean practice score was 21.70 ± 2.94 on a maximum of 25 (86.80%), indicating proactive practice. Higher practice scores were observed with higher education (P < 0.001), full-time housewife (P = 0.030), and higher income (P = 0.013) (Table 1). Supplement Table 11 presents the distribution of the answers to the practice items. The main reasons for not accepting short-acting oral contraceptives after surgery were worries about fertility (50.71%) and economic reasons (23.18%) (Figure 2). If there is no plan for future pregnancy, 33.81% would take long-acting contraception after abortion, 20.71% would not, 28.33% were uncertain, and 17.14% would follow the physician’s advice (Figure 3). The most important means of getting knowledge about postabortion contraception included the Internet (55.24%), healthcare providers (44.76%), community lectures (22.62%), and social circles (21.19%) (Figure 4). When comparing P1 (“After an induced abortion, I will follow the doctor’s advice to take antibiotics”) and P2 (“After an induced abortion, I will follow the doctor’s advice to take short-acting contraceptives orally”), the practice of taking antibiotics after surgery was better than that of taking short-acting contraceptives orally after surgery (P = 0.031) (Figure 5).Figure 2P3. The main reason for not accepting short-acting oral contraceptives after surgery. The x axis label: a. Do not think it is necessary; b. Economic reasons; c. Worried about affecting fertility; d. Worried about side effects; e. Worried about being discovered by others; f. Other reasons. The y axis label: number (n) of participants.Bar graph showing reasons for not accepting short acting oral contraceptives after surgery.Figure 3P4. If there is no plan for a future pregnancy, would you consider choosing long-acting contraception after an induced abortion? The x axis label: a. Yes; b. No; c. Uncertain; d. Follow the doctor’s advice. The y axis label: number (n) of participants.Bar graph showing participant responses about choosing long acting contraception after an induced abortion.Figure 4P8. Ways of acquiring knowledge about contraception and induced abortion. The x axis label: a. Community lectures; b. Internet; c. Magazines and newspapers; d. TV and radio; e. Exchange with friends; f. Introduction by medical staff; g. Other ways; h. Never interested in this information. The y axis label: percentage (%) of participants.Bar graph showing ways of acquiring knowledge about contraception and induced abortion.Figure 5Practice differences in taking antibiotics and short-acting contraceptives. Comparing P1 (Q1) “After an induced abortion, I will follow the doctor’s advice to take antibiotics” and P2 (Q2) “After an induced abortion, I will follow the doctor’s advice to take short-acting contraceptives orally” (paired t-test), the practice of taking antibiotics after surgery is better than that of taking short-acting contraceptives orally after surgery. 1 = strongly disagree; 2 = disagree; 3 = uncertain; 4 = agree; 5 = strongly agree.Violin plots comparing practice ratings for Q1 and Q2, with similar central values and a noted difference. P3. The main reason for not accepting short-acting oral contraceptives after surgery. The x axis label: a. Do not think it is necessary; b. Economic reasons; c. Worried about affecting fertility; d. Worried about side effects; e. Worried about being discovered by others; f. Other reasons. The y axis label: number (n) of participants. P4. If there is no plan for a future pregnancy, would you consider choosing long-acting contraception after an induced abortion? The x axis label: a. Yes; b. No; c. Uncertain; d. Follow the doctor’s advice. The y axis label: number (n) of participants. P8. Ways of acquiring knowledge about contraception and induced abortion. The x axis label: a. Community lectures; b. Internet; c. Magazines and newspapers; d. TV and radio; e. Exchange with friends; f. Introduction by medical staff; g. Other ways; h. Never interested in this information. The y axis label: percentage (%) of participants. Practice differences in taking antibiotics and short-acting contraceptives. Comparing P1 (Q1) “After an induced abortion, I will follow the doctor’s advice to take antibiotics” and P2 (Q2) “After an induced abortion, I will follow the doctor’s advice to take short-acting contraceptives orally” (paired t-test), the practice of taking antibiotics after surgery is better than that of taking short-acting contraceptives orally after surgery. 1 = strongly disagree; 2 = disagree; 3 = uncertain; 4 = agree; 5 = strongly agree. The participants were grouped based on their response to P2 as “Strongly agree” and “Agree” vs the others (Supplement Table 12). Compared with the unwilling group, the participants in the willing group had higher knowledge (P < 0.001) and attitude (P < 0.001) scores, and the proportion of college education was higher (P < 0.001) (Supplement Table 12, including the scores for the individual items). Correlations As shown in Table 2, the knowledge scores were correlated to the attitude (r = 0.174, P < 0.001) and practice (r = 0.394, P < 0.001) scores. The attitude scores were correlated to the practice scores (r = 0.216, P < 0.001).Table 2Spearman CorrelationKnowledgeAttitudePracticeKnowledge1Attitude0.174 (P < 0.001)1Practice0.394 (P < 0.001)0.216 (P < 0.001)1 Table 2Spearman CorrelationKnowledgeAttitudePracticeKnowledge1Attitude0.174 (P < 0.001)1Practice0.394 (P < 0.001)0.216 (P < 0.001)1 Spearman Correlation Multivariable Analysis All VIFs were <2, suggesting the absence of significant multicollinearity. Junior high school or below (OR = 0.08, 95% CI: 0.01, 0.61, P = 0.015) and not having used contraceptive measures (OR = 0.19, 95% CI: 0.01, 0.94, P = 0.021) were independently associated with lower odds of sufficient knowledge, while a monthly income >10,000 (OR = 1.89, 95% CI: 1.09–3.30, P = 0.025) was more strongly associated with higher odds of sufficient knowledge) (Table 3 and Figure 6). The knowledge scores (OR = 1.13, 95% CI: 1.05, 1.22, P < 0.001) were independently associated with higher odds of positive attitude, while junior high school or below (OR = 0.44, 95% CI: 0.20, 0.96, P = 0.038) was independently associated with lower odds of positive attitude (Table 4 and Figure 7). The knowledge scores (OR = 1.17, 95% CI: 1.02, 1.33, P = 0.027), the attitude scores (OR = 1.11, 95% CI: 1.00, 1.22, P = 0.047), one childbirth (OR = 3.60, 95% CI: 1.38, 9.37, P = 0.009), and >2 childbirths (OR = 4.55, 95% CI: 1.36, 15.22, P = 0.014) were independently associated with higher odds of adequate practice, while age 36–39 (OR = 0.22, 95% CI: 0.05, 0.94, P = 0.041) was independently associated with lower odds of adequate practice (Table 5 and Figure 8).Table 3Multivariable Logistic Regression Analysis for KnowledgeVariablesnUnivariateMultivariableORPORPAge18–2517/74ReferenceReference26–3553/2460.921 (0.502,1.751)0.79440.787 (0.375,1.652)0.527036–3918/1000.736 (0.349,1.557)0.41940.664 (0.265,1.664)0.3824Marital statusMarried58/285ReferenceSingle/Divorced/Widowed30/1351.118 (0.674,1.829)0.6600Number of pregnancies 010/60ReferenceReference 1–248/2391.257 (0.615,2.789)0.55020.933 (0.483,1.800)0.8355 ≥330/1211.648 (0.765,3.802)0.21761.676 (0.788,3.566)0.1800Number of childbirths 028/127Reference 136/1860.854 (0.491,1.497)0.5783 ≥224/1071.022 (0.548,1.897)0.9441Body mass indexNormal (18.5–23.9 kg/m2)57/268ReferenceUnderweight (<18.5 kg/m2)11/431.273 (0.582,2.614)0.5261Overweight (≥24 kg/m2)20/1090.832 (0.463,1.446)0.5241Place of residenceUrban58/271ReferenceNon-urban30/1490.926 (0.559,1.509)0.7600EducationCollege and above75/291ReferenceReferenceJunior high school and below1/440.067 (0.004,0.316)0.00810.078 (0.010,0.605)0.0146High school/vocational school12/850.473 (0.233,0.891)0.02740.590 (0.289,1.206)0.1478Current employment statusEmployed71/292ReferenceReferenceUnemployed17/1280.477 (0.261,0.831)0.01170.592 (0.319,1.097)0.0955Average monthly per capita income5000–10,00030/183ReferenceReference<500015/871.063 (0.527,2.071)0.86131.258 (0.614,2.579)0.5300>10,00043/1502.050 (1.214,3.498)0.00771.893 (1.086,3.301)0.0245Type of medical insuranceInsured78/370ReferenceUninsured10/500.936 (0.426,1.887)0.8601Contraceptive measuresUsed87/395ReferenceReferenceNot used1/250.148 (0.008,0.713)0.01210.185 (0.010,0.936)0.0208ComorbiditiesNo82/375ReferenceYes6/450.550 (0.203,1.253)0.1894Note: Goodness-of-fit: Nagelkerke R2= 0.1282, sensitivity = 0.7159, specificity = 0.5076, F1-score = 0.4013, and AUC = 0.6837. The variance inflation factors varied from 1.005 to 1.531.Abbreviations: OR, odds ratio; CI, confidence interval. Table 4Multivariable Logistic Regression Analysis for AttitudeVariablesnUnivariateMultivariableORPORPKnowledge1.156 (1.080,1.237)<0.00011.133 (1.054,1.217)0.0007Age18–2538/74ReferenceReference26–35126/2460.995 (0.590,1.674)0.98411.056 (0.568,1.965)0.862936–3939/1000.606 (0.328,1.110)0.10590.748 (0.357,1.566)0.4408Marital statusMarried142/285ReferenceSingle/Divorced/Widowed61/1350.830 (0.550,1.251)0.3745Number of pregnancies 032/60Reference 1–2120/2390.882 (0.498,1.555)0.6653 ≥351/1210.638 (0.341,1.186)0.1563Number of childbirths 068/127ReferenceReference 189/1860.804 (0.511,1.264)0.34580.785 (0.458,1.346)0.3796 ≥246/1070.654 (0.389,1.096)0.10830.807 (0.428,1.519)0.5054Body mass indexNormal (18.5–23.9 kg/m2)134/268ReferenceUnderweight (<18.5 kg/m2)20/430.870 (0.453,1.658)0.6712Overweight (≥24 kg/m2)49/1090.817 (0.521,1.276)0.3745Place of residenceUrban131/271ReferenceNon-urban72/1490.999 (0.669,1.491)0.9973EducationCollege and above158/291ReferenceReferenceJunior high school and below11/440.281 (0.131,0.560)0.00050.438 (0.200,0.957)0.0384High school/vocational school34/850.561 (0.341,0.914)0.02120.689 (0.410,1.158)0.1598Current employment statusEmployed140/292ReferenceUnemployed63/1281.052 (0.694,1.595)0.8100Average monthly per capita income5000–10,00092/183ReferenceReference<500030/870.521 (0.304,0.878)0.01550.636 (0.364,1.112)0.1123>10,00081/1501.161 (0.754,1.792)0.49840.976 (0.619,1.539)0.9173Type of medical insuranceInsured180/370ReferenceUninsured23/500.899 (0.494,1.625)0.7251Contraceptive measuresUsed191/395ReferenceNot used12/250.986 (0.433,2.226)0.9726ComorbiditiesNo182/375ReferenceYes21/450.928 (0.496,1.725)0.8129Note: Goodness-of-fit: Nagelkerke R2= 0.1075, sensitivity = 0.5911, specificity = 0.6389, F1-score = 0.5985, and AUC = 0.6634. The variance inflation factors varied from 1.056 to 1.555.Abbreviations: OR, odds ratio; CI, confidence interval. Table 5Multivariable Logistic Regression Analysis for PracticeVariablesnUnivariateMultivariableORPORPKnowledge1.271 (1.121,1.440)0.00021.165 (1.018,1.333)0.0266Attitude1.126 (1.036,1.224)0.00511.105 (1.001,1.218)0.0468Age18–2570/74ReferenceReference26–35228/2460.724 (0.204,2.017)0.57030.355 (0.098,1.282)0.114036–3989/1000.462 (0.124,1.417)0.20250.215 (0.049,0.939)0.0409Marital statusMarried262/285ReferenceSingle/Divorced/Widowed125/1351.097 (0.520,2.478)0.8137Number of pregnancies 055/60Reference 1–2219/2390.996 (0.320,2.586)0.9930 ≥3113/1211.284 (0.373,4.032)0.6734Number of childbirths 0112/127ReferenceReference 1174/1862.119 (0.945,4.892)0.07053.597 (1.380,9.372)0.0088 ≥2100/1071.913 (0.774,5.185)0.17464.554 (1.362,15.223)0.0138Body mass indexNormal (18.5–23.9 kg/m2)246/268ReferenceUnderweight (<18.5 kg/m2)41/431.833 (0.513,11.713)0.4236Overweight (≥24 kg/m2)100/1090.994 (0.456,2.344)0.9877Place of residenceUrban249/271ReferenceNon-urban138/1491.108 (0.532,2.437)0.7887EducationCollege and above272/291ReferenceReferenceJunior high school and below34/440.238 (0.104,0.570)0.00090.480 (0.160,1.443)0.1913High school/vocational school81/851.415 (0.514,4.983)0.53902.222 (0.638,7.740)0.2099Current employment statusEmployed269/292ReferenceUnemployed118/1281.009 (0.478,2.280)0.9820Average monthly per capita income5000–10,000171/183ReferenceReference<500075/870.439 (0.187,1.030)0.05590.639 (0.239,1.711)0.3727>10,000141/1501.099 (0.452,2.764)0.83511.122 (0.423,2.975)0.8178Type of medical insuranceInsured345/370ReferenceReferenceUninsured42/500.380 (0.167,0.950)0.02730.415 (0.148,1.169)0.0961Contraceptive measuresUsed364/395ReferenceNot used23/250.979 (0.272,6.280)0.9782ComorbiditiesNo349/375ReferenceReferenceYes38/450.404 (0.172,1.066)0.06560.413 (0.153,1.114)0.0807Note: Goodness-of-fit: Nagelkerke R2= 0.2268, sensitivity = 0.9378, specificity = 0.3939, F1-score = 0.9427, and AUC = 0.8238. The variance inflation factors varied from 1.044 to 1.573.Abbreviations: OR, odds ratio; CI, confidence interval. Figure 6Forest plot of the multivariable analysis of knowledge.Forest plot of odds ratios across age, childbirths, education, employment, income and contraceptive measures.The forest plot displays odds ratios with 95% confidence intervals for various variables. The x-axis ranges from 0.2 to 5.0, with a reference line at 1.0. Variables include age groups (26-35, 36-39), childbirths (1, ≥2), education levels (junior high and below, high school/vocational), employment status (unemployed), income levels (<5000, >10000) and contraceptive use (not used). Each variable has a square marker indicating the odds ratio, with confidence intervals spanning from about 0.1 to 4.0. Education at junior high and below and not using contraceptives, show the widest intervals, extending left of 0.2 and ending near 1.0. Higher income (>10000) and multiple childbirths (≥2) have estimates right of 1.0, with intervals above 2.0. Other variables have estimates near 1.0, with intervals crossing the reference line.Figure 7Forest plot of the multivariable analysis of attitude.Forest plot of attitude multivariable analysis showing odds ratios with confidence intervals around 1.0.Forest plot analysis of attitude shows odds ratios (95% CI) on the x-axis from 0.2 to 2.0, with a reference line at 1.0. The y-axis includes: Knowledge, Age groups (26-35, 36-39), Number of childbirths (1, ≥2), Education levels (Junior high and below, High school/vocational) and Average monthly per capita income (<5000, >10000). Each category has a point estimate with a 95% confidence interval. Knowledge is around 1.1 (CI: 1.05-1.2). Age 26-35 is about 1.05 (CI: 0.55-1.95), Age 36-39 is 0.75 (CI: 0.35-1.35). Childbirths 1 is 0.8 (CI: 0.45-1.3), ≥2 is 0.8 (CI: 0.4-1.35). Junior high education is 0.45 (CI: 0.2-0.95), High school/vocational is 0.7 (CI: 0.4-1.1). Income <5000 is 0.65 (CI: 0.35-1.1), >10000 is 1.0 (CI: 0.6-1.4).Figure 8Forest plot of the multivariable analysis of practice.Odds ratios: knowledge, attitude, age, births, education, income, insurance, comorbidities.Forest plot with odds ratios and confidence intervals for predictors. X-axis: Odds Ratio (95% CI) with ticks at 0.2, 0.5, 1.0, 2.0, 5.0, 10.0. Vertical line at odds ratio 1.0. Y-axis: Knowledge, Attitude, Age26-35, Age36-39, Number of childbirths 1, Number of childbirths ≥2, Education (Junior high and below, High school/vocational), Average monthly income (<5000, >10000), Medical insurance (Uninsured), Comorbidities (Yes). Estimates and 95% CIs: Knowledge ~1.2 (1.1-1.4); Attitude ~1.1 (1.0-1.3); Age26-35 ~0.4 (0.1-1.2); Age36-39 ~0.2 (0.08-0.9); Childbirths 1 ~3.6 (1.4-9.4); Childbirths ≥2 ~4.6 (1.4-15.2); Education (Junior high and below) ~0.5 (0.2-1.6); Education (High school/vocational) ~2.0 (0.8-9.0); Income <5000 ~0.7 (0.25-1.6); Income >10000 ~1.1 (0.4-3.0); Uninsured ~0.4 (0.15-1.2); Comorbidities Yes ~0.4 (0.15-1.2). Table 3Multivariable Logistic Regression Analysis for KnowledgeVariablesnUnivariateMultivariableORPORPAge18–2517/74ReferenceReference26–3553/2460.921 (0.502,1.751)0.79440.787 (0.375,1.652)0.527036–3918/1000.736 (0.349,1.557)0.41940.664 (0.265,1.664)0.3824Marital statusMarried58/285ReferenceSingle/Divorced/Widowed30/1351.118 (0.674,1.829)0.6600Number of pregnancies 010/60ReferenceReference 1–248/2391.257 (0.615,2.789)0.55020.933 (0.483,1.800)0.8355 ≥330/1211.648 (0.765,3.802)0.21761.676 (0.788,3.566)0.1800Number of childbirths 028/127Reference 136/1860.854 (0.491,1.497)0.5783 ≥224/1071.022 (0.548,1.897)0.9441Body mass indexNormal (18.5–23.9 kg/m2)57/268ReferenceUnderweight (<18.5 kg/m2)11/431.273 (0.582,2.614)0.5261Overweight (≥24 kg/m2)20/1090.832 (0.463,1.446)0.5241Place of residenceUrban58/271ReferenceNon-urban30/1490.926 (0.559,1.509)0.7600EducationCollege and above75/291ReferenceReferenceJunior high school and below1/440.067 (0.004,0.316)0.00810.078 (0.010,0.605)0.0146High school/vocational school12/850.473 (0.233,0.891)0.02740.590 (0.289,1.206)0.1478Current employment statusEmployed71/292ReferenceReferenceUnemployed17/1280.477 (0.261,0.831)0.01170.592 (0.319,1.097)0.0955Average monthly per capita income5000–10,00030/183ReferenceReference<500015/871.063 (0.527,2.071)0.86131.258 (0.614,2.579)0.5300>10,00043/1502.050 (1.214,3.498)0.00771.893 (1.086,3.301)0.0245Type of medical insuranceInsured78/370ReferenceUninsured10/500.936 (0.426,1.887)0.8601Contraceptive measuresUsed87/395ReferenceReferenceNot used1/250.148 (0.008,0.713)0.01210.185 (0.010,0.936)0.0208ComorbiditiesNo82/375ReferenceYes6/450.550 (0.203,1.253)0.1894Note: Goodness-of-fit: Nagelkerke R2= 0.1282, sensitivity = 0.7159, specificity = 0.5076, F1-score = 0.4013, and AUC = 0.6837. The variance inflation factors varied from 1.005 to 1.531.Abbreviations: OR, odds ratio; CI, confidence interval. Multivariable Logistic Regression Analysis for Knowledge Note: Goodness-of-fit: Nagelkerke R2= 0.1282, sensitivity = 0.7159, specificity = 0.5076, F1-score = 0.4013, and AUC = 0.6837. The variance inflation factors varied from 1.005 to 1.531. Abbreviations: OR, odds ratio; CI, confidence interval. Table 4Multivariable Logistic Regression Analysis for AttitudeVariablesnUnivariateMultivariableORPORPKnowledge1.156 (1.080,1.237)<0.00011.133 (1.054,1.217)0.0007Age18–2538/74ReferenceReference26–35126/2460.995 (0.590,1.674)0.98411.056 (0.568,1.965)0.862936–3939/1000.606 (0.328,1.110)0.10590.748 (0.357,1.566)0.4408Marital statusMarried142/285ReferenceSingle/Divorced/Widowed61/1350.830 (0.550,1.251)0.3745Number of pregnancies 032/60Reference 1–2120/2390.882 (0.498,1.555)0.6653 ≥351/1210.638 (0.341,1.186)0.1563Number of childbirths 068/127ReferenceReference 189/1860.804 (0.511,1.264)0.34580.785 (0.458,1.346)0.3796 ≥246/1070.654 (0.389,1.096)0.10830.807 (0.428,1.519)0.5054Body mass indexNormal (18.5–23.9 kg/m2)134/268ReferenceUnderweight (<18.5 kg/m2)20/430.870 (0.453,1.658)0.6712Overweight (≥24 kg/m2)49/1090.817 (0.521,1.276)0.3745Place of residenceUrban131/271ReferenceNon-urban72/1490.999 (0.669,1.491)0.9973EducationCollege and above158/291ReferenceReferenceJunior high school and below11/440.281 (0.131,0.560)0.00050.438 (0.200,0.957)0.0384High school/vocational school34/850.561 (0.341,0.914)0.02120.689 (0.410,1.158)0.1598Current employment statusEmployed140/292ReferenceUnemployed63/1281.052 (0.694,1.595)0.8100Average monthly per capita income5000–10,00092/183ReferenceReference<500030/870.521 (0.304,0.878)0.01550.636 (0.364,1.112)0.1123>10,00081/1501.161 (0.754,1.792)0.49840.976 (0.619,1.539)0.9173Type of medical insuranceInsured180/370ReferenceUninsured23/500.899 (0.494,1.625)0.7251Contraceptive measuresUsed191/395ReferenceNot used12/250.986 (0.433,2.226)0.9726ComorbiditiesNo182/375ReferenceYes21/450.928 (0.496,1.725)0.8129Note: Goodness-of-fit: Nagelkerke R2= 0.1075, sensitivity = 0.5911, specificity = 0.6389, F1-score = 0.5985, and AUC = 0.6634. The variance inflation factors varied from 1.056 to 1.555.Abbreviations: OR, odds ratio; CI, confidence interval. Multivariable Logistic Regression Analysis for Attitude Note: Goodness-of-fit: Nagelkerke R2= 0.1075, sensitivity = 0.5911, specificity = 0.6389, F1-score = 0.5985, and AUC = 0.6634. The variance inflation factors varied from 1.056 to 1.555. Abbreviations: OR, odds ratio; CI, confidence interval. Table 5Multivariable Logistic Regression Analysis for PracticeVariablesnUnivariateMultivariableORPORPKnowledge1.271 (1.121,1.440)0.00021.165 (1.018,1.333)0.0266Attitude1.126 (1.036,1.224)0.00511.105 (1.001,1.218)0.0468Age18–2570/74ReferenceReference26–35228/2460.724 (0.204,2.017)0.57030.355 (0.098,1.282)0.114036–3989/1000.462 (0.124,1.417)0.20250.215 (0.049,0.939)0.0409Marital statusMarried262/285ReferenceSingle/Divorced/Widowed125/1351.097 (0.520,2.478)0.8137Number of pregnancies 055/60Reference 1–2219/2390.996 (0.320,2.586)0.9930 ≥3113/1211.284 (0.373,4.032)0.6734Number of childbirths 0112/127ReferenceReference 1174/1862.119 (0.945,4.892)0.07053.597 (1.380,9.372)0.0088 ≥2100/1071.913 (0.774,5.185)0.17464.554 (1.362,15.223)0.0138Body mass indexNormal (18.5–23.9 kg/m2)246/268ReferenceUnderweight (<18.5 kg/m2)41/431.833 (0.513,11.713)0.4236Overweight (≥24 kg/m2)100/1090.994 (0.456,2.344)0.9877Place of residenceUrban249/271ReferenceNon-urban138/1491.108 (0.532,2.437)0.7887EducationCollege and above272/291ReferenceReferenceJunior high school and below34/440.238 (0.104,0.570)0.00090.480 (0.160,1.443)0.1913High school/vocational school81/851.415 (0.514,4.983)0.53902.222 (0.638,7.740)0.2099Current employment statusEmployed269/292ReferenceUnemployed118/1281.009 (0.478,2.280)0.9820Average monthly per capita income5000–10,000171/183ReferenceReference<500075/870.439 (0.187,1.030)0.05590.639 (0.239,1.711)0.3727>10,000141/1501.099 (0.452,2.764)0.83511.122 (0.423,2.975)0.8178Type of medical insuranceInsured345/370ReferenceReferenceUninsured42/500.380 (0.167,0.950)0.02730.415 (0.148,1.169)0.0961Contraceptive measuresUsed364/395ReferenceNot used23/250.979 (0.272,6.280)0.9782ComorbiditiesNo349/375ReferenceReferenceYes38/450.404 (0.172,1.066)0.06560.413 (0.153,1.114)0.0807Note: Goodness-of-fit: Nagelkerke R2= 0.2268, sensitivity = 0.9378, specificity = 0.3939, F1-score = 0.9427, and AUC = 0.8238. The variance inflation factors varied from 1.044 to 1.573.Abbreviations: OR, odds ratio; CI, confidence interval. Multivariable Logistic Regression Analysis for Practice Note: Goodness-of-fit: Nagelkerke R2= 0.2268, sensitivity = 0.9378, specificity = 0.3939, F1-score = 0.9427, and AUC = 0.8238. The variance inflation factors varied from 1.044 to 1.573. Abbreviations: OR, odds ratio; CI, confidence interval. Forest plot of the multivariable analysis of knowledge. Forest plot of the multivariable analysis of attitude. Forest plot of the multivariable analysis of practice. The sensitivity analyses based on the continuous KAP scores are shown in Supplement Tables 13–15. Two or more childbirths (β = 1.25, 95% CI: 0.39, 2.10, P = 0.005) and junior high school or higher (β = −2.19, 95% CI: −3.16, −1.21, P < 0.001) were independently associated with the knowledge scores. The knowledge scores (β = 0.22, 95% CI: 0.08, 0.37, P = 0.003) were independently associated with the attitude scores. The knowledge scores (β = 0.31, 95% CI: 0.21, 0.40, P < 0.001), the attitude scores (β = 0.07, 95% CI: 0.01, 0.13, P = 0.017), and junior high school or higher (β = −1.54, 95% CI: −2.49, −0.58, P = 0.002) were independently associated with the practice scores. Discussion Data about postabortion contraception are lacking in China. This study investigated the KAP of women who underwent induced abortion regarding postoperative oral short-acting contraceptives. The results suggest that women after an induced abortion had poor knowledge and unfavorable attitudes, but proactive practice toward short-acting contraception, and specific knowledge items were identified as requiring improvement. Educational interventions should particularly target women with a lower socioeconomic status. The decision on contraception and the type of contraception ultimately belongs to the women undergoing an induced abortion. Still, making the most appropriate decision requires proper knowledge of contraception about the advantages, side effects, the risk of pregnancy recurrence, and plans for fertility.1–4 The advantages of short-acting contraception include relatively mild and well-known side effects and the possibility of stopping them at any time to restore fertility. On the other hand, this method requires close adherence to the dosing schedule to minimize the risk of pregnancy or the combined use of a barrier contraceptive method. In this study, over half of the participants identified the Internet as their main source of information on postabortion contraception, followed closely by healthcare providers, with smaller but non‑negligible contributions from community lectures and social circles. Given the known risk of misinformation in online reproductive health content and the documented inconsistencies in contraceptive knowledge among providers,20–24 these channels represent critical targets for quality improvement. Clinician‑curated online resources on widely used Chinese health platforms (eg, large commercial health portals and mainstream social media health channels), structured provider training on up‑to‑date contraceptive evidence, and standardized peri‑abortion counseling protocols could help improve the accuracy and consistency of information delivered through both digital and clinical pathways. The second most important source of information was physicians. While international studies have documented important gaps in contraceptive knowledge among physicians,22–24 comparable large‑scale surveys among OB‑GYNs in mainland China remain limited, and this represents an important evidence gap for future research. Of course, physicians with inaccurate knowledge will pass on that inaccurate knowledge. According to the KAP framework, knowledge and attitudes are often conceptualized as important antecedents of health‑related behaviors.8,9 In the present study, knowledge was independently and positively associated with both attitude and practice, and attitude was independently and positively associated with practice. This pattern suggests that strengthening women’s understanding of short‑acting oral contraceptives could, in principle, support more favorable attitudes and better contraceptive behaviors. Basic and accurate information on short‑acting contraceptives (including potential harms and side effects, contraindications, implications for fertility and congenital malformations, and their role in endometrial repair after abortion) should therefore be provided to women undergoing induced abortion. At the same time, the marked discrepancy between poor knowledge and relatively high practice scores may have at least two explanations: women may genuinely be willing to follow physicians’ advice even when they do not fully understand the rationale, but the practice items were framed as “I will follow the doctor’s advice…” and were completed in the clinic before the abortion, so social desirability bias is also a plausible contributor. Similar patterns of poor knowledge but high reported practice have been observed in previous studies,10,12 although the available data remain limited and warrant cautious interpretation. Notably, the two worst‑answered knowledge items (long‑term safety and congenital malformation risk) corresponded closely to the most frequently cited reason for declining short‑acting contraception (worry about future fertility), highlighting a direct and policy‑relevant link between specific misconceptions and contraceptive hesitancy. Given that the poorest‑answered knowledge items concerned long‑term safety (K12) and congenital malformation risk (K8), and that worry about future fertility was the dominant stated reason for declining short‑acting oral contraceptives, postabortion counseling should explicitly and systematically address these specific misconceptions. Counseling protocols could include simple, evidence‑based explanations of the long‑term safety profile of short‑acting oral contraceptives and their lack of association with fetal congenital malformations, supported by written materials designed for a junior‑high reading level. In addition, clinician‑curated online content delivered through commonly used health platforms and social media, aligned with the messages provided in the clinic, may help correct misinformation encountered on the Internet. Such structured peri‑abortion counseling and tailored patient education are likely to be more effective than generic information in improving understanding and acceptance of short‑acting oral contraceptives, particularly among women with lower education and health literacy. In the present study, a low socioeconomic status was generally associated with poorer KAP, and participants with a high willingness to take postabortion contraception based on their physician’s recommendations had a higher level of education than participants with a low willingness. This finding is consistent with evidence that general health literacy is associated with socioeconomic status.25 It also highlights that women of a lower socioeconomic status and undergoing induced abortion should be particularly targeted by future educational interventions to improve KAP. Still, a study from Guangzhou highlighted that access to and utilization of postabortion care were limited, especially in disadvantaged women,15 suggesting that postabortion care services themselves should also be improved. The association of older age with lower odds of positive practice could be due to several reasons, including different fertility intentions at higher parity, perceived decline in fertility, method preference shift, cohort and information effects, and socioeconomic and structural confounding. The study design and data collected do not allow for determining the exact reason. The present sample was drawn from an urban hospital in Beijing, where most participants held urban residence and relatively high educational attainment. In China, structural factors such as the hukou system, urban–rural disparities, and the large migrant population in metropolitan areas can substantially influence access to family planning services, continuity of care, and eligibility for publicly funded counseling and contraception. The Guangzhou study by Wang et al15 highlighted that disadvantaged and migrant women face particular barriers to postabortion care. The relatively high use of the Internet and formal healthcare providers as information sources in our cohort may therefore not be generalizable to women in rural settings or those with unstable hukou status, underscoring the need to tailor interventions to different subgroups within China. Strengths and Limitations A strength of this study is that the participants filled out the questionnaire before the abortion; at that time, they had not yet received postoperative counseling about postabortion care, which likely reduced the influence of clinicians’ explanations and immediate postprocedure emotions on their answers. Nevertheless, several limitations should be acknowledged. First, the cross-sectional design precludes any causal inference between knowledge, attitudes, and practice; associations observed between KAP dimensions and sociodemographic factors may reflect underlying, unmeasured constructs rather than directional effects. In addition, the study provides only a single “snapshot” of women’s KAP at the time of abortion, and does not capture how their knowledge, attitudes, and practices may evolve after exposure to postabortion counseling, recovery experiences, or subsequent pregnancies; however, these data can serve as a historical baseline for future longitudinal or interventional studies. Second, this was a single-center study in an urban hospital in Beijing, with a relatively modest sample size compared with the broader population of women seeking induced abortion in China. Consequently, some 95% CIs were wide due to the small number of events in certain combined categories. Thus, the findings may not be generalizable to other regions, rural areas, or different levels of medical institutions with distinct service models and patient profiles. Third, the study did not differentiate between women undergoing induced abortion for unplanned pregnancy and those undergoing abortion for medical indications (eg, missed abortion), even though their fertility intentions, perceived risks, and contraceptive needs may differ substantially, which could dilute or mask important subgroup-specific patterns in KAP. Still, women with a previous history of abortion were excluded because they would have received education about contraception, introducing bias. Counseling pathways, timing of contraceptive discussion, and perceived side‑effect profiles may differ between medical and surgical abortion, and women who complete medical abortion before presentation may have a distinct KAP profile that is not captured here. As a result, the findings should be interpreted as most applicable to women managed in surgical settings or at the pre‑initiation stage of medical abortion, and may not generalize to all women undergoing medical abortion in other care pathways. Fourth, the questionnaire was developed by local investigators based on Chinese expert consensus documents and national standards, and although pilot testing showed acceptable reliability, its content and emphasis may reflect local practices, guidelines, and policies; certain items may not fully capture broader cultural, religious, or regional determinants of postabortion contraceptive decision-making. Furthermore, in routine practice, short‑acting oral contraceptives include both combined oral contraceptives and progestogen‑only pills, which have different contraindications, safety profiles, and counseling needs, particularly in the immediate postabortion period,1–4 and the present questionnaire focused on women’s overall KAP toward the use of such short-acting oral contraceptives after abortion, without distinguishing between specific formulations. Fifth, as with all self-administered KAP surveys, responses are vulnerable to social desirability bias, whereby participants provide answers that they perceive as socially acceptable or consistent with medical recommendations rather than their true beliefs or behaviors.26,27 Considering the relatively high practice scores compared with the knowledge and attitude scores, it is possible that some women overreported their willingness to follow physicians’ advice on short-acting contraception, which would lead to an overestimation of “adequate” practice. Furthermore, although the questionnaire underwent a pilot test, the absence of factor analysis and test-retest analysis limited its psychometric evaluation. Sixth, the use of a self-completed online questionnaire requires a minimal level of literacy and familiarity with smartphones or digital tools, so women with very low literacy or limited digital access may be underrepresented, potentially underestimating the true extent of poor knowledge and unfavorable attitudes in more disadvantaged groups. Seventh, because convenience sampling was used, the exact number of women who were invited to participate is unknown; hence, the response rate cannot be calculated. Conclusions Women seeking induced abortion at an urban hospital in Beijing had poor knowledge (mean 46.7% of the maximum score) and unfavorable attitudes (69.5%), but reported a high intention to follow doctors’ recommendations regarding short‑acting oral contraceptives after abortion (practice score 86.8%), with this knowledge-practice gap emerging as the key signal of the study. Two specific misconceptions, long‑term safety (K12) and congenital malformation risk (K8), which had the lowest correct response rates, directly mirrored the most frequently cited reason for declining short‑acting contraception, namely worry about future fertility, highlighting concrete targets for educational messaging. Women with junior‑high education or below and those in the lowest income strata consistently showed poorer KAP scores, indicating that they are the subgroup most likely to benefit from tailored educational and counseling interventions. It is also important to note that the practice dimension captured intended behavior (“I will follow the doctor’s advice”) rather than observed adherence, and the findings should be interpreted accordingly.