Non-consensual condom removal among adults: prevalence, correlates, and outcomes in quantitative studies – A systematic review

1.2 Young adults, gender, and risk context

Building on the conceptual framing of NCCR as a violation of consent with significant health and psychosocial consequences, existing research suggests that young adults aged 18 years and above are particularly vulnerable to this form of sexual harm. This life stage is often characterised by increased sexual exploration, a higher likelihood of casual or non-committed sexual relationships, and variable condom use, especially in social contexts where power imbalances, substance use, and peer norms complicate discussions of consent and sexual boundaries.^12,13 Although awareness of condom use and sexual health risks is generally high among young people, evidence consistently demonstrates that knowledge alone does not reliably translate into protective sexual behaviour.¹⁴

Within these contexts, gendered patterns of victimisation and perpetration are consistently observed. Women and girls are more frequently reported as victims of NCCR, particularly those with a prior history of sexual victimisation, suggesting cumulative vulnerability over time.¹⁵ In contrast, perpetrators are more likely to exhibit hostile or antisocial traits, sexual entitlement, and broader patterns of coercive or aggressive behaviour.^16,17 Importantly, existing evidence indicates that NCCR often co-occurs with other forms of sexual coercion rather than occurring as an isolated behaviour, reinforcing its classification as part of a broader continuum of sexual violence.

1.3 Legal ambiguity and global variation

Despite growing recognition of NCCR as a violation of consent, legal responses remain inconsistent across jurisdictions. In the United Kingdom, NCCR is not explicitly defined as a criminal offence, creating uncertainty regarding prosecution under existing sexual offences legislation.^18,19 Similar ambiguity exists across much of the United States and Canada, where NCCR is often prosecutable only if it results in demonstrable bodily harm, such as pregnancy or STI transmission.^2,3

In contrast, a small number of jurisdictions have taken clearer legal positions. California classifies NCCR as sexual battery, Singapore criminalised the act under its Criminal Law Reform Act, and Australian courts have recognised NCCR as a violation of consent.^6,20 These disparities shape public perceptions of seriousness and accountability. Quantitative evidence suggests that even where individuals recognise NCCR as wrong, not all view it as criminal, and support for prosecution often increases only when physical consequences such as pregnancy or STIs occur.²¹

1.4 Prevalence and population-level uncertainty

Quantitative studies indicate that NCCR is not a rare occurrence, although reported prevalence varies widely across settings and populations. Researchers⁵ found that rates of NCCR victimisation ranged from 7.9% to 43.0% for women and 5.0% to 19.0% for MSMs; rates of NCCR perpetration ranged from 5.1% to 9.8% for men and 0% for women. Evidence also suggests an increase in reported prevalence over the past decade, alongside a rise in scholarly publications addressing NCCR.⁹ However, the true population burden of NCCR remains difficult to determine. Prevalence estimates are shaped by substantial heterogeneity in study design, sampling strategies, recall periods, and definitions of NCCR. Many studies rely on convenience samples, particularly university students, and few provide sufficiently detailed demographic stratification to allow comparison across age groups, genders, geographic regions, or sexual orientations. This lack of comparability limits the usefulness of existing data for public health surveillance and prevention planning.

Existing reviews and remaining gaps

A few review papers have contributed important insights into NCCR, but have not addressed key epidemiological questions. A previous study conducted a rapid review aimed at resolving conceptual and definitional inconsistencies, proposing a broader framework of “nonconsensual condom-use deception.”⁶ While valuable for theoretical clarity, their review was not designed to synthesise prevalence estimates or examine demographic or regional variation.

A scoping review mapped the extent and nature of empirical research on NCCR.⁵ As intended in scoping methodology, the review was broad and inclusive, synthesising diverse study designs and populations to identify themes and gaps. However, it did not focus on a systematic comparison of prevalence estimates or on the methodological drivers of variation.

A further narrative synthesis examined NCCR primarily in relation to alcohol-involved rape and abortion policy barriers following the Dobbs decision.²² While highlighting the reproductive consequences of NCCR, this work did not aim to quantify prevalence or assess population-level patterns.

Collectively, these reviews establish NCCR as a legitimate concern but leave a critical gap in the literature: the absence of a systematic, prevalence-focused synthesis of quantitative evidence that examines how NCCR varies by age, gender, region, and study design.

Aim of the current systematic review

Guided by a pre-specified study protocol (Appendix 1- Study Protocol), the present systematic review aims to synthesise quantitative evidence on the prevalence of non-consensual condom removal among adults aged 18 years and above. Specifically, the review seeks to:

By consolidating prevalence data and critically examining methodological heterogeneity, this review aims to strengthen the epidemiological evidence base on NCCR and support more informed public health responses, legal discourse, and prevention strategies for this under-recognised form of sexual violence.

Study design and reporting framework

This review was conducted as a systematic review of quantitative studies reporting the prevalence of non-consensual condom removal (NCCR) among adults aged 18 years and above. The review followed established methodological guidance for prevalence reviews and was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A review protocol was developed prior to commencing the review to guide methodology and ensure transparency, in line with updated systematic review guidance.^23,24 The PRISMA flow diagram detailing the study selection process is provided in Appendix 3.

Databases

Three electronic databases were searched: CINAHL, MEDLINE, and APA PsycINFO.

Search strategy

A two-step search strategy was employed to identify relevant studies. The first step involved an initial exploratory (wild) search, a broad, unstructured scan of literature using basic keywords related to condom sabotage, NCCR and young people. This phase aimed to familiarise the researchers with commonly used terminology, refine inclusion concepts, and identify alternative spellings or phrasing across different disciplines and databases. Based on this exploratory phase, the second step involved a more systematic, advanced search using Boolean operators (AND, OR), truncation (*), and phrase searching (“”).

The final search terms included in the above 3 databases are as follows:

Across all databases, searches were limited to peer-reviewed journal articles published in English. The final search yielded 348 records from MEDLINE, 23 from CINAHL, and 62 from APA PsycINFO. One additional study was identified through citation searching, bringing the total to 11 studies. Detailed database-specific search strings are available in Appendix 2.

To enhance completeness, backwards and forward citation searching was conducted for all included studies. In addition, Google Scholar was searched systematically, with screening limited to the first 20 pages of results, consistent with best practice recommendations for supplementary searching in systematic reviews.

Study selection

All records identified through database searches were imported into EndNote reference management software. Duplicate records were identified and removed using EndNote’s automated duplicate detection function, followed by manual verification to ensure that all duplicates were appropriately identified and removed.

Study selection was conducted in a staged process in accordance with PRISMA 2020 guidance. In the first stage, titles and abstracts were screened to identify potentially relevant studies addressing non-consensual condom removal (NCCR), including studies that did not explicitly use the term stealthing but examined related concepts. Records that did not indicate stealthing or NCCR in the title or abstract were excluded at this stage.

Full texts were sought for all records considered potentially eligible. One report could not be retrieved. During initial full-text inspection, a small number of reports were excluded because, despite appearing potentially relevant at the screening stage, they did not address stealthing or NCCR upon closer examination. The remaining full-text articles were then assessed in detail for eligibility against the predefined inclusion and exclusion criteria. Full-text assessment was conducted independently by two reviewers, with any disagreements resolved through discussion and consensus. Reasons for exclusion at the full-text stage were documented and are presented in the PRISMA flow diagram (Appendix 3).

Studies were eligible for inclusion if they reported quantitative prevalence estimates of NCCR victimisation and/or perpetration among adults aged 18 years and above, were published in peer-reviewed journals, and were available in full text in English. Studies were excluded if they were qualitative in design, opinion pieces, editorials, commentaries, or conference abstracts without full text. Studies that did not report prevalence data or included participants under the age of 18 were also excluded.

Quality appraisal

Methodological quality was assessed using the Joanna Briggs Institute (JBI) Critical Appraisal Checklist for Prevalence Studies. This tool evaluates domains, including sampling adequacy, measurement reliability, statistical analysis, and response rates. Quality appraisal was conducted to inform the interpretation of findings rather than to exclude studies based on quality alone. Results of the quality assessment are reported narratively below and summarised in tabular form in Appendix 4.

Critical appraisal of included studies (n = 11)

All 11 included studies were appraised using the JBI Critical Appraisal Checklist for Analytical Cross-Sectional Studies (for mixed-methods papers, appraisal focused on the quantitative component only). Overall, the methodological quality was generally strong: eight studies were rated high quality, while three were rated moderate quality. Across studies, inclusion criteria and study settings were usually well described, and most papers provided clear operational definitions of NCCR, supporting consistency in case identification. Several studies also demonstrated stronger internal validity through appropriate inferential analyses, including multivariable regression to account for potential confounding (e.g., clinic-based and some community/online samples). Two reviewers independently appraised study quality, with any disagreements resolved through discussion; full item-level appraisal tables for each included study are provided in Appendix 4.

Common limitations were largely shared across the evidence base. First, sampling approaches were predominantly non-probability (e.g., university students, MTurk participants, clinic attendees, and convenience/community samples), reducing generalisability to broader adult populations. Second, all studies relied on self-reports, introducing potential recall and social desirability bias, especially given the sensitive nature of sexual violence and condom practices. Third, measurement of NCCR was frequently based on single items or study-developed questions, with limited evidence of psychometric validation; this was a key reason for moderate ratings in some studies. Finally, while several studies tested correlates of NCCR, not all employed multivariable adjustment, meaning residual confounding could not be ruled out in bivariate-only analyses. As expected for this literature, all included studies were cross-sectional; therefore, causal inference was not possible across the evidence base.

In summary, the included studies provided usable prevalence and correlational evidence, with the strongest studies characterised by clear definitions, robust analytic strategies (including adjustment for confounders), and transparent reporting. However, interpretation should remain cautious due to consistent risks related to selection bias, self-report measurement, heterogeneity in NCCR measurement, and limited representativeness of samples.

Data extraction

Data were extracted using a standardised and pilot-tested data extraction form to ensure consistency and accuracy. Extracted variables included study characteristics (author, year, country, and setting), methodological details (study design, sampling strategy, recruitment method, and analytical approach), and participant characteristics (sample size, age range or mean age, and gender). Prevalence estimates of NCCR victimisation and/or perpetration were recorded, along with the time frame of measurement (e.g. lifetime or past-year prevalence). Where available, subgroup analyses by age, gender, sexual orientation, or region were extracted. Secondary outcomes, such as disclosure behaviour and reported sexual, reproductive, or mental health impacts, were also captured when reported.

The full data extraction tool and extracted variables are provided in Appendix 5.

Data synthesis

Study characteristics and prevalence estimates were synthesised using descriptive and narrative synthesis. Meta-analysis was not conducted due to substantial heterogeneity across studies, including differences in populations, definitions of NCCR, study designs, and outcome measures. As a result, a narrative synthesis approach was adopted to describe patterns and variability in reported prevalence estimates.

Study selection

Database searches identified records from MEDLINE, CINAHL, and APA PsycINFO, supplemented by Google Scholar screening and citation chasing. After de-duplication and screening, 11 studies met the inclusion criteria and were included in the synthesis.

Study characteristics

Eleven quantitative studies (n = 11) were included, with most using cross-sectional survey designs.^{2,10,12,15,16,25–30} Of the 11 studies, two were mixed-methods papers from which quantitative findings were extracted.^2,25

Sample sizes varied considerably across studies, with the smallest study including 96 participants and the largest including 2,550 participants. The detailed data extraction table is in Appendix-5.

Studies were conducted across multiple regions, including the United States (n = 6),^{12,16,25,26,29,31} Australia/New Zealand, (n = 1),¹⁵ Australia,(n = 1),¹⁰ Canada, (n = 1),² Brazil, (n = 1),,²⁷ and India (n = 1).³⁰

Sampling frames included university student populations (n = 5),,^{2,15,25,27,30} local community samples (n = 5),^{12,15,16,26,31} local community and college students sample (n = 1),¹⁵ and (n = 1) from Amazon Mechanical Turk²⁹ and a sexual health clinic sample (n = 1).¹⁰

Recall periods differed substantially (e.g., lifetime/ever vs “since age 14/16” vs not specified), contributing to heterogeneity in prevalence estimates. Four studies (n = 4) reported NCCR prevalence since age fourteen.^12,16,25,31 Five studies (n = 5) reported lifetime or ever experienced NCCR.^{2,10,15,27,30} One study (n = 1) reported the prevalence since age sixteen.²⁶ and one study (n = 1) did not specify the recall period.²⁹

Participant populations varied across studies. Two studies (n = 2) reported recruiting the MSM population alongside women.,^10,29 while one study (n = 1) included men, women and MSM within the sample.²⁹ Four studies (n = 4) recruited both men and women.^2,25–27 Two studies (n = 2) included only male participants.^12,16 And three studies (n = 3) recruited women participants.^15,30,31

Five studies (n = 5) assessed alcohol use and/or other substance use as potential correlates of NCCR victimisation and/or perpetration, using a range of analytic approaches, including descriptive, event-level, and multivariable analyses.^{10,12,25,26,30}

Objective 1: Prevalence of non-consensual condom removal

Prevalence of NCCR victimisation varied by population group, setting, and recall period. Among women, victimisation prevalence ranged from 9.3% in a young women sample in Australia/New Zealand.¹⁵ to 32% in an Australian sexual health clinic sample.¹⁰

In U.S. samples, women’s victimisation prevalence included 12.2% in a community sample aged 21–30²⁸ and 24% in an MTurk adult sample.²⁹ In Canada, a university sample reported 18.7% victimisation.² 11.44% among young women in Brazil²⁷ And in India, female university students reported 14.4% victimization.³⁰

Among MSMs, victimisation prevalence was 19% in the Australian sexual health clinic sample¹⁰and 35% in the MTurk sample.²⁹ Where men’s victimisation was reported, estimates were lower; e.g., 2.4% in a U.S. university sample,²⁵ and 5% among emerging adults in a U.S. community/online sample.²⁶

Perpetration prevalence was reported less consistently. Where measured among men, perpetration ranged from 1.33% in a Brazilian university student sample²⁷ to approximately 19–19.8% in U.S. samples of men,^12,29 with an additional estimate of 6.1% among U.S. emerging adults²⁶ and 9.8% in a U.S. sample of men aged 21–30.³² Two studies also reported repeated victimisation/perpetration events rather than single episodes.^16,26

The prevalence estimates reported across the included studies are summarised in Appendix 6.

The findings from the included studies indicate that NCCR is a prevalent phenomenon across diverse populations and settings, with substantial variation by gender, sexual orientation, and recruitment context. Overall, victimisation prevalence was consistently higher among women than men, with estimates among women ranging from 9.3% to 32%. The highest prevalence was observed among women attending sexual health clinics, suggesting that clinical populations may capture individuals at heightened sexual risk or with greater exposure to coercive sexual practices. In contrast, lower but still notable prevalence was reported among women in university and community-based samples, with victimisation estimates ranging from 9.3% to 24%, indicating that NCCR is not confined to high-risk or clinical contexts.

Victimisation prevalence among men was reported less frequently and, where available, tended to be lower than among women in mixed-gender samples. However, studies focusing on MSM revealed comparatively high levels of victimisation, with prevalence ranging from 19% to 35%. These findings suggest that MSM may experience distinct vulnerabilities related to sexual negotiation, power dynamics, or contextual factors within same-sex partnerships that are not adequately captured in heteronormative research frameworks.

Data on NCCR perpetration were comparatively sparse but revealed important patterns. Male perpetration prevalence varied widely, ranging from 1.33% to 19.8%, depending on population characteristics and study design. Higher perpetration prevalence was reported in studies focusing exclusively on men, particularly among younger, single heterosexual men, whereas substantially lower prevalence was observed in mixed-gender university samples. This variability may reflect differences in disclosure willingness, social desirability bias, or how perpetration was operationalised across studies. The absence of quantitative estimates of female perpetration across the included studies highlights a significant gap in the literature and limits understanding of gendered patterns of perpetration.

Taken together, the findings suggest that NCCR is neither rare nor limited to a specific demographic group, but rather represents a widespread sexual health and consent issue. The wide range of prevalence estimates underscores the influence of study population, setting, and measurement approaches, and highlights the need for more consistent definitions and standardised outcome measures. Improved methodological.

Objective 2: Variation in prevalence by demographics and correlates

3.2.1 Alcohol and substance use as correlates of NCCR victimisation and perpetration

Across studies, correlates of NCCR victimisation and perpetration were heterogeneous, but several consistent quantitative patterns emerged. Substance use, particularly alcohol, was the most repeatedly supported correlate of victimisation. Victimisation among men was associated with binge drinking (OR = 1.49, p < .05) and use of other illegal drugs (OR = 1.79, p < .05), while among women victimisation increased with older age (OR = 1.13, p < .05) and was also associated with binge drinking (OR = 1.28, p < .05), cannabis use (OR = 1.29, p < .05) and other drug use (OR = 1.22, p < .05).²⁶ A further study reported large differences by substance exposure: NCCR prevalence was 28.4% among binge drinkers vs 3.2% among non-binge drinkers, and 38.7% among substance users vs 3.3% among non-users (both p < .01, as reported).³⁰

Incident-level evidence from a sexual health clinic sample reported that alcohol and other drug use commonly co-occurred with NCCR events. At the time of the NCCR incident, 57% of women and 41% of MSM reported having consumed alcohol, while 12% of women and 13% of MSM reported use of other drugs (either with or without alcohol). Participants also frequently reported partner substance use during the incident: among women, 68% reported the partner had consumed alcohol, and 19% reported partner use of other drugs. Among MSM, 40% reported partner alcohol use and 12% reported partner use of other drugs.¹⁰

Event-level and multivariable evidence linking NCCR perpetration to alcohol-related factors among men.¹² NCCR perpetration was positively correlated with the number of drinks consumed during the sexual event (rpb = 0.26, p = .01) and with men’s perceptions of their partners’ intoxication (rpb = 0.34, p < .01). NCCR was also significantly more likely to occur when a partner had consumed alcohol. Of the perpetrators, 72.2% were with a partner who had consumed alcohol as compared to 27.8% reported having a sober partner r (χ² = 6.13, p = .02).

In multivariable analyses, alcohol-related expectancies specific to condom use and coercive sexual behaviour remained strong predictors of NCCR perpetration. Men who reported drinking to enhance sex had more than three times higher odds of engaging in NCCR (OR = 3.12, 95% CI 1.49–6.56), while stronger sexual coercion-related alcohol expectancies were associated with nearly a fivefold increase in odds of NCCR (OR = 4.74, 95% CI 1.60–21.64).

Alcohol use was independently associated with increased risk of NCCR victimisation.²⁵ In the multivariable logistic regression model, alcohol use was associated with a 23% increase in the odds of NCCR victimisation (OR = 1.23, 95% CI 1.02–1.49).

A detailed overview of studies assessing alcohol and substance use variables and their associations with NCCR victimisation and perpetration is presented in Appendix 7.

Across the above studies, alcohol and substance use were the most consistently identified correlates of NCCR. Regression-based analyses showed that alcohol use, binge drinking, and other drug use were independently associated with increased odds of NCCR victimisation among both men and women, with markedly higher prevalence among substance users compared with non-users. Incident-level findings further indicated that NCCR events frequently occurred in alcohol-involved sexual encounters, including contexts in which both victims and partners had consumed alcohol. Event-level and multivariable analyses among men also demonstrated that NCCR perpetration was more likely during alcohol-involved encounters and was strongly associated with alcohol-related expectancies specific to condom use and coercive sexual behaviour. Overall, the evidence indicates that alcohol-related behaviours and intoxication contexts are closely linked to both NCCR victimisation and perpetration.

3.2.2 Variation in NCCR by sexual orientation and gender identity

Across the included studies, sexual orientation and gender identity were consistently associated with variation in non-consensual condom removal (NCCR) victimisation. Sexual orientation emerged as a significant correlate where examined. Heterosexual orientation was associated with significantly lower odds of NCCR victimisation among men (OR = 0.24, p < .05), indicating elevated risk among non-heterosexual men.²⁶ Women identifying as non-heterosexual were significantly more likely to experience NCCR than expected based on sample distribution (z = 2.35), corresponding to more than twice the odds of victimisation compared with heterosexual women.¹⁵ Together, these findings indicate that sexual minority status was associated with increased likelihood of NCCR victimisation across different populations and study designs.

Gender identity and sex also showed strong associations with NCCR victimisation. Identifying as a woman was associated with substantially higher odds of NCCR victimisation compared with cisgender men (OR = 13.41, 95% CI 1.05–172.10), while identifying as a gender minority was associated with even greater odds (OR = 30.95, 95% CI 1.54–622.78) in adjusted analyses.²⁵ Significant associations have also been reported between NCCR victimisation and both female biological sex and identifying as a woman (p = 0.000 for both).²⁷ These findings indicate that gender identity and sex function as independent correlates of NCCR victimisation, contributing to observed variation beyond other demographic and behavioural factors.

Beyond sexual orientation and gender identity, several studies identified additional correlates related to relational and behavioural context. Women who had experienced NCCR were significantly more likely to be current sex workers (AOR = 2.87, 95% CI 2.01–4.11, p < .001), highlighting the role of occupational and relationship context in shaping NCCR risk.¹⁰ NCCR perpetration among men was associated with attitudinal and behavioural factors, with higher odds observed among men reporting greater hostility toward women (OR = 1.47) and more severe histories of sexual aggression (OR = 1.06) in adjusted analyses.¹⁶

Objective 3: Reported outcomes and impacts associated with NCCR

3.3.1. Psychological and mental health impacts

Across studies, NCCR was consistently associated with adverse psychological and mental health outcomes. Victims of NCCR were significantly more likely to report elevated post-traumatic stress disorder (PTSD) symptoms compared with non-victims, even after accounting for other forms of sexual violence.³⁰ The authors explicitly noted that a causal relationship could not be established, but that NCCR and PTSD were significantly and independently associated, underscoring the psychological burden linked to the experience.

High levels of emotional distress were also reported across multiple studies. The majority of participants who experienced NCCR reported significant psychological distress, with 67.2% stating they were “bothered a lot” by the incident.,² Similarly, emotional stress was identified as the most common outcome following NCCR, reported by 56% of women and 52% of MSM attending sexual health clinics.¹⁰

Evidence further indicated that NCCR was associated with diminished sexual self-efficacy and perceived sexual agency. Women who experienced NCCR reported significantly lower perceived control over sexual situations than non-victims (p = .05, r = .11), as well as a reduced ability to refuse unwanted sex, reflected in lower sexual self-efficacy scores (p = .04, r = .11),¹⁵ Similarly, adverse impacts on sexual wellbeing, including reduced confidence, assertiveness, and increased shame-related feelings, have also been reported.³¹ Together, these findings suggest that NCCR may undermine sexual agency and confidence beyond the immediate incident.

Mental health impacts were also evident in specific populations. MSMs who had experienced NCCR were significantly more likely to report symptoms of anxiety or depression (AOR = 2.13, 95% CI 1.25–3.60, p = .005).¹⁰ Male victims of NCCR reported higher levels of depression and anxiety than any other victim or non-victim group, indicating that psychological distress may vary by gender and victimisation context.²⁹

Notably, substantial heterogeneity in perceived impact has been observed. Approximately 25% of women and over one-third of MSM reported no consequences following NCCR, suggesting variation in interpretation, coping, or the contextual meaning of the experience.¹⁰

3.3.2. Physical and reproductive health consequences

Several studies documented physical and reproductive health outcomes associated with NCCR. Participants reported concrete health consequences following NCCR, including sexually transmitted infections (STIs; n = 4), pregnancy (n = 2), and both STI and pregnancy (n = 1).² Although absolute numbers were small, these findings highlight the tangible sexual and reproductive risks associated with non-consensual condom removal.

Chi-square analyses demonstrated that men with a history of NCCR perpetration were significantly more likely to report having been diagnosed with an STI (29.5% vs 15.1%) and to report that a partner had experienced an unplanned pregnancy (46.7% vs 25.8%) compared with men without such a history.¹⁶ In a related analysis,²⁸ further, it was reported that STI diagnoses were positively associated with receipt of condom use resistance (CUR), but not with engagement in CUR, suggesting differential health risks for victims versus perpetrators.

STIs were reported by 8% of women and 5% of MSM, while HIV infection was reported by 1% of women and 2% of MSM following experiences of NCCR among clinic based populations.¹⁰ Collectively, these findings indicate that NCCR is associated with increased exposure to sexually transmitted infections and unplanned pregnancy across diverse populations.

3.3.3 Experiential, relational, and perceptual impacts

Beyond individual health outcomes, NCCR was associated with relational and interpersonal harms. Participants reported relationship strain following NCCR, including interpersonal conflict or fights (14% of women; 9% of MSM) and relationship breakdowns (9% of women; 4% of MSM), indicating that NCCR may disrupt trust and relationship stability.¹⁰

NCCR often occurred within broader contexts of relationship dysfunction.²⁹ Male perpetrators reported higher levels of partner violence and control, insecure attachment styles (avoidant and anxious), and elevated antisocial and borderline personality traits. Victims, particularly male victims, reported greater levels of violence and control within romantic relationships and endorsed more symptoms of depression and anxiety than other groups. These findings suggest that NCCR may be embedded within wider patterns of interpersonal harm and psychological vulnerability.

Experiential impacts were also evident in how participants perceived and conceptualised NCCR. Most individuals who experienced NCCR described it as negative and emphasised themes of violation, lack of consent, and betrayal of trust.² However, individuals who had experienced NCCR were significantly less likely to label the experience as sexual assault compared with those who had not experienced it (women: OR = 0.29, 95% CI 0.22–0.40; men: OR = 0.31, 95% CI 0.21–0.45),¹⁰ highlighting a disconnect between lived harm and formal recognition of NCCR as sexual violence.

Summary of findings

Eleven studies were included. Reported prevalence of non-consensual condom removal (NCCR) victimisation ranged from 9.3% to 32% among women, 19% to 35% among men who have sex with men (MSM), and 2.4% to 5% among men in general population samples. Reported perpetration ranged from 1.33% to 19.8% among men.

Variation in NCCR victimisation was associated with sexual orientation and gender identity. Non-heterosexual men had lower odds of protection from victimisation compared with heterosexual men (OR = 0.24, p < .05), whereas non-heterosexual women had a higher likelihood of victimisation (z = 2.35, p < .05). Identifying as a woman (OR = 13.41, 95% CI 1.05–172.10) or as a gender minority (OR = 30.95, 95% CI 1.54–622.78) was associated with higher odds of NCCR victimisation compared with cisgender men.

Relationship and contextual correlates included current sex work among women (AOR = 2.87, 95% CI 2.01–4.11), perpetrator hostility toward women (OR = 1.47), and sexual aggression history (OR = 1.06).

Alcohol and substance use were commonly reported behavioural correlates. Victimisation prevalence was higher among binge drinkers (28.4% vs 3.2%) and substance users (38.7% vs 3.3%) compared with non-users, while perpetration was associated with alcohol-related expectancies (OR range = 3.12–4.74).

Across studies, NCCR was associated with adverse impacts including elevated post-traumatic stress disorder symptoms, anxiety or depression (AOR = 2.13, 95% CI 1.25–3.60), emotional distress (52–67%), reduced sexual self-efficacy (p = .04–.05), sexually transmitted infections (5–8%), unplanned pregnancy (25–47%), and relationship conflict or breakdown (4–14%). Although causal relationships could not be established, studies reported consistent associations between NCCR and negative psychological, physical, and relational outcomes.

Comparison with the broader literature on prevalence and measurement

The wide variation in prevalence observed in this review is consistent with prior syntheses of the NCCR literature. A recent scoping review concluded that most empirical studies on stealthing are cross-sectional and report highly variable victimisation prevalence across populations, with higher estimates commonly observed among women and sexual minority samples, and perpetration estimates reported less consistently.⁵ These findings mirror the pattern observed in the present review and reinforce the challenges associated with estimating population-level prevalence.

Furthermore, prevalence estimates observed in this review align with the wider literature on NCCR and condom use resistance, which indicates that NCCR is reported by a substantial minority of sexually active individuals, particularly young adults and clinic-attending populations.³³ variability in prevalence across studies has been attributed to differences in study populations, recall periods, and behavioural definitions, a pattern that has also been documented more broadly in research on sexual violence and coercive sexual behaviours.^34,35

Additionally, NCCR has been situated within the wider literature on coercive condom use resistance (CUR), which conceptualises NCCR as one tactic within a broader range of behaviours aimed at undermining condom use.^33,36 This framing helps explain why prevalence estimates are sensitive to measurement choices. Studies using narrow, single-item measures capture explicit condom removal without consent, whereas broader CUR measures may capture multiple forms of condom-related coercion, including sabotage or deception, which may share overlapping but non-identical correlates and prevalence profiles.⁶ Given these definitional and methodological differences, the use of narrative synthesis in the present review was appropriate.

Comparison with external evidence on correlates of NCCR

This review identified recurring correlates of NCCR related to alcohol or substance use, sexual minority status (where measured), relationship and sexual history factors, and prior experiences of sexual victimisation. These findings are consistent with broader evidence indicating that coercive condom-related practices often cluster with prior sexual victimisation and occur within sexual risk contexts characterised by impaired consent negotiation.^33,36 Research on coercive condom use resistance (which includes condom sabotage and other condom-related coercion) has linked perpetration to sexual aggression history, hostility toward women, and related coercive dynamics, particularly in male samples.^37–39

The review also found that victimisation prevalence among MSM was elevated in certain samples. External evidence supports this observation and highlights that condom-related coercion and deception occur in both same-sex and opposite-sex encounters, reinforcing the need for inclusive measurement approaches that do not assume exclusively heterosexual contexts.^5,40

Comparison with external evidence on outcomes and consequences

Outcomes associated with NCCR were inconsistently reported across the included studies but included sexual and reproductive health concerns, such as risk of sexually transmitted infections and pregnancy, as well as psychological impacts. The wider literature supports these concerns, indicating that coerced condomless sex practices are associated with adverse physical, emotional, and mental health outcomes.^33,36 However, systematic assessment of outcomes remains limited, with few studies examining longer-term consequences or integrating validated mental health measures.

Relationship context and reproductive coercion

Findings linking NCCR to relationship dynamics, sex work, and controlling behaviours are consistent with the reproductive coercion and intimate partner violence literature, which documents associations between birth control sabotage, unintended pregnancy, and sexual health harms.⁴¹ Conceptualising NCCR within this broader continuum of sexual and reproductive control may help clarify mechanisms and inform prevention strategies.

Alcohol, substance use, and sexual violence

The strong and consistent associations between alcohol use and NCCR identified in this review align with extensive evidence linking intoxication to sexual aggression, impaired consent processes, and increased sexual risk-taking.⁴² Event-level and expectancy-based findings observed in NCCR-specific studies extend established models of alcohol-related sexual risk to this specific form of sexual boundary violation.

Psychological, physical, and relational impacts

Associations between NCCR and PTSD symptoms, anxiety, depression, and emotional distress observed in this review mirror well-established links between sexual victimisation and adverse mental health outcomes reported in population-level studies.³⁴ Importantly, while cross-sectional designs preclude causal inference, the persistence of these associations after accounting for other forms of sexual violence suggests that NCCR may independently contribute to psychological harm, consistent with broader sexual violence research.³³

Physical and reproductive health consequences identified in this review, including sexually transmitted infections and unplanned pregnancy, align with findings from studies on reproductive coercion and sexual risk behaviours.^41,42 Together, these findings support framing NCCR as a public health concern with implications beyond consent alone.

Implications for practice, policy, and research

Clinically, these findings support routine, trauma-informed sexual health screening that explicitly includes questions about condom removal and related coercive condom use resistance, particularly in sexual health services and populations with elevated risk. Prevention efforts may benefit from integrating NCCR content into sexual consent education and alcohol-harm reduction strategies, given the repeated association between alcohol/substance use and both victimisation and perpetration. At the policy level, the evidence strengthens calls to treat NCCR as a form of sexual violence in legal and institutional responses, consistent with existing scholarship on consent and “rape-adjacent” practices.³

Future research priorities include: (1) standardised, validated measures of NCCR and consistent recall periods to improve comparability; (2) more probability-based and culturally diverse sampling to address generalisability gaps; and (3) longitudinal or event-level designs (e.g., EMA) to reduce recall bias and clarify temporal pathways linking alcohol, coercion, and harm.

Strengths and limitations of the evidence base

This review provides a structured synthesis focused on quantitative prevalence, demographic patterning, and outcomes. However, interpretation is constrained by consistent limitations in the primary studies: cross-sectional designs precluding causal inference, heavy reliance on self-report for sensitive experiences, heterogeneous measurement (single-item vs multi-item proxies), and frequent use of convenience samples (students, MTurk, clinic attendees). These issues likely contribute to variability in prevalence estimates and underscore the appropriateness of narrative synthesis rather than meta-analysis.

This review adds new knowledge beyond existing reviews

Firstly, existing evidence syntheses on NCCR/“stealthing” have primarily taken a scoping or conceptual approach, mapping how the phenomenon has been defined, measured, and discussed across diverse study types and disciplines (e.g., qualitative studies, commentaries, legal scholarship), rather than producing a focused synthesis of quantitative prevalence estimates and objective-led outcomes. However, this review contributes to new and more policy-actionable knowledge in four ways. A) it is a systematic review of quantitative studies that explicitly report NCCR prevalence in adults, allowing a clearer description of prevalence ranges by population group (women, MSM, men perpetration) and setting, rather than a general mapping of the topic. B) it applies an objective-led synthesis (prevalence → demographic variation/correlates → outcomes), which strengthens transparency about what the available evidence can and cannot support for public health and prevention decision-making. C) it incorporates and synthesises recent international prevalence studies (including non-US settings) and explicitly identifies where evidence remains geographically concentrated, an issue repeatedly noted as a gap in prior reviews.

Finally, by emphasising the consequences that are actually measured (e.g., sexual/reproductive outcomes, mental health impacts, disclosure/reporting where available), this review directly addresses the outcomes gap highlighted in the wider conceptual literature on NCCR measurement and consent, including calls for clearer operationalisation and consistent measurement of harms. This strengthens the review’s contribution to both research priorities (standardised measures, consistent recall periods, subgroup reporting) and policy/clinical practice (screening in sexual health settings, consent education, and targeted prevention for higher-risk groups).