The Decline of Child Sexual Abuse in Canada: Evidence From the 2014 General Social Survey.

OBJECTIVE: To provide evidence of trends in child sexual abuse (CSA) in Canada.
METHODS: Using data from 15,801 males and 18,669 females who responded to the 2014 General Social Survey (GSS), we compared the prevalence of CSA by age cohorts. Age cohort patterns were examined for several sub-populations including males, females, Indigenous peoples, and people living in low-income households.
RESULTS: After an increase in the post-World War II period, there has been a decline in CSA in Canada since the early 1990s. Findings indicate a decline for both sexes; although, the evidence is more compelling for females. There is also evidence of a decline for Indigenous peoples, for those living in low-income households, and regardless of the relationship to the perpetrator (i.e., family member, a teacher/professor/tutor, a babysitter, a nanny, other non-family member but known to the respondent, or a stranger).
CONCLUSIONS: In Canada, evidence from 3 retrospective population surveys suggests a decline in CSA since the early 1990s. However, given the associated harm, continued progress to the eradication of CSA is essential.

Public health surveillance of child sexual abuse (CSA) is important in our efforts to protect the safety of children. Such surveillance could use population-representative contemporaneous surveys of children and youth,[1] but in Canada, such data are non-existent at the national level. Like many other countries, Canada bases its incidence rates of CSA on data from official sources such as police[2] and child welfare.[3,4] However, CSA is often not reported to such agencies,[2,5,6] largely because CSA occurs in private, perpetrators do not confess, and children do not disclose it due to multiple barriers (e.g., need to self-protect, fragile social network, or a fear of social/relational consequences).[7]

Retrospective population surveys asking respondents about their childhood experiences of sexual abuse can instead be used to provide evidence of changes in prevalence over time.[8-15] However, there are problems with comparing the results from diverse cross-sectional retrospective surveys. Prevalence estimates of self-reporting CSA can vary substantially due to different definitions of CSA, data collection methods, and survey contexts. A different approach that has been used is to compare prevalence estimates of CSA for people in different age cohorts within the same survey.[9,10,12,16] This approach allows for the investigation of cohort differences in self-reporting CSA and reduces the possibility that differences are due to methodological discrepancies.

Studies from the United States (US) provide evidence of an increase in CSA following World War II (WWII).[10,13,15] Subsequently, data from numerous US victimization surveys provide evidence of a decline since the early 1990s[11] that parallels trends observed from child welfare agency data.[11] Less work has been done in Canada. Previously, CSA prevalence estimates were compared by birth cohorts using data from 2 representative Canadian surveys (the 2012 Canadian Community Health Survey-Mental Health [CCHS] and the 2004/2005 Canadian Gender, Alcohol, and Culture: An International Study [GENACIS]).[16] For the most part, results concurred with the US studies. For both sexes, there was evidence of an increase in CSA following WWII. The evidence for a decline since the early 1990s was compelling for females but weak for males.

It is important to recognise that trends in CSA may vary between population subgroups. CSA is a more common historical experience among certain subpopulations. For example, the results from a recent study based on longitudinal data found that CSA was associated with increased financial strain during adulthood.[17] Thus, adults living in low-income households may be more likely to have had CSA experiences. As well, reports of CSA in retrospective surveys are more common among Canadian Indigenous peoples than their non-Indigenous counterparts.[18,19] For the Indigenous population, there has been a call for action by the Truth and Reconciliation Committee to provide more information to close gaps in health outcomes between Indigenous and non-Indigenous communities.[20]

CSA can be intra- or extra-familial, and trends may differ by type of perpetrator. For example, there is some evidence that the prevalence of reporting CSA by a stranger was more common in the early 1900s.[13] As well, changes in family structure in the post-WWII period may have resulted in a general increase in all types of CSA.[21] In Canada, divorce became more common in the post-WWII period and peaked in the late 1980s, increasing the likelihood of children living with a single parent or a step parent,[21] which are both risk factors for CSA.[22-25]

Using the General Social Survey (GSS) 2014, we replicated earlier work by examining age cohort patterns by sex. We examined whether reporting CSA was relatively lower among those who were children before the end of WWII (1945 or earlier), relatively higher for those who were children after WWII to the early 1990s (1946–1992), and then relatively lower for those who were children after the early 1990s (1993 or after). We extended previous research by considering age cohort patterns for those living in lower-income households and Indigenous peoples, and examined age cohort patterns by perpetrator type. If evidence of a decline can be found in multiple surveys (with varying purposes and methodologies) and across sub-populations, then there can be enhanced confidence that the phenomenon is real.

Methods

Data Source

We used data from Statistics Canada’s GSS—Victimization-2014. The target population was household residents aged 15 y or older living in Canada’s 10 provinces and 3 territories. Two separate samples were selected for the GSS; one sample for the 10 provinces and one for the 3 territories. The survey frame was derived from the Census and various administrative sources from Statistics Canada’s dwelling frame, combining landline and cellular telephone numbers. To obtain better coverage, the frame consisted of groups of one or several telephone numbers associated with the same address. The sample was selected based on a stratified design employing probability sampling. The stratification was done at the province/census metropolitan area (CMA) level. More information about the sample design is available in the GSS Microdata User Guide.[26]

All interviews in the 10 provinces were conducted by telephone using computer-assisted telephone interviews (CATI). In the 3 territories, although most of the interviews were conducted by telephone using CATI, 30% of the interviews were conducted in person. The response rate was 52.9% for the 10 provinces (responding sample size of 33,127) and 58.7% for the 3 territories (responding sample size of 2,040). We pooled the 2 samples to produce estimates for all Canadians. From the combined sample (n = 35,167), 697 records were excluded due to non-response to the CSA questions resulting in a sample of 34,470 (15,801 males and 18,669 females).

Measures

The GSS questionnaire included numerous items on experiences of crime, fear of crime, social disorder, among others. Items on childhood maltreatment were asked close to the end of the interview and included physical abuse, sexual abuse, and exposure to intimate partner violence. GSS respondents were classified as having experienced CSA if they answered “one or more times” to either of the following questions (response categories were never, 1–2, 3–5, 6–10 or more than 10 times):

“Before age 15, how many times did an adult force you or attempt to force you into any unwanted sexual activity, by threatening you, holding you down, or hurting you in some way?”

“Before age 15, how many times did an adult touch you against your will in any sexual way? By this, I mean anything from unwanted touching or grabbing, to kissing or fondling.”

Among those reporting CSA, respondents were asked about their relationship to the perpetrator.

“Thinking about the most serious incident, what was the adult’s relationship to you?” (Multiple responses were not permitted.)

We grouped the responses to this item into the following categories:

Family member (excluding step-parent): Mother, father, grandmother, grandfather, brother, sister, other family member

Step-parent: Step-mother, step-father

Caregiver: Teacher/professor/tutor, baby-sitter/nanny

Other known person (not family): Boyfriend/girlfriend, ex-boyfriend/ex-girlfriend, neighbour, friend, acquaintance, classmate

Stranger: Known by sight only/stranger

Indigenous peoples were identified based on the following question: “Are you an Aboriginal person, that is, First Nations, Métis or Inuk (Inuit)? First Nations includes Status and Non-Status Indians.”

For the 10 provinces, the GSS income related variables were obtained from income tax files for calendar year 2013. For the 3 territories, respondents were asked “What is your best estimate of your total household income, received by all household members from all sources, before taxes and deductions, during the year ending December 31, 2013?”.

Analysis

We re-used the “childhood person years” (CPY) approach from a previous study to examine trends in CSA, summarised here for clarity.[16] CPYs are the individual units of time when the people in the study population were, by virtue of being aged 0 to 14 years, potentially exposed to CSA.

We hypothesized that prevalence of reporting CSA would be relatively higher in people who were children (ages 0 to 14 y) between 1946 and 1992, and relatively lower in those who were aged 0 to 14 y before the end of WWII (before 1946) or after the early 1990s (1993 and after). We calculated the percentage of their CPYs that those in each age group spent in the higher- and lower-risk periods. Table 1 illustrates the CPY calculations and indicates the number and percentage of CPYs spent by those in each age group (at the time of the survey) in the 3 time periods.

Table 1.

Number and Percentage of Childhood Person Years (Ages 0 to 14) Spent in Child Sexual Abuse Risk Periodsa Among Respondents to General Social Survey by Age Group Cohort.

			Number and % of CPYs Spent:
			Lower-risk		Higher-risk		Lower-risk
	Total CPYsb		1993 or after		1946–1992		1945 or earlier
Age groups (y)	(age 0–14 y)	Year of Birth	#	%	#	%	#	%
15–24	150	1990–1999	144	96	6	4	0	0
25–34	150	1980–1989	65	43	85	57	0	0
35–49	225	1965–1979	1	0	224	100	0	0
50–64	225	1950–1964	0	0	225	100	0	0
65–79	225	1935–1949	0	0	159	71	66	29
80–99	300	1915–1934	0	0	0	2	294	98

GSS, 2014 General Social Survey; CPYs, childhood person years; CSA, child sexual abuse.

aRisk periods: (1993 or After; From 1946 to 1992; and 1945 or Earlier).

bTotal CPYs = (# of birth years in cohort) × 15.

Consider the age group 15 to 24 y, shown in the first line of Table 1. These respondents had a total of 150 CPYs (the 15 y represented in the 0–14 y of childhood × the 10 possible ages included in the 15–24 age group, assuming that the 10 ages are equally represented). For GSS respondents in this age group, 6 CPYs were spent in the years 1946 to 1992, the higher-risk period, as follows: Those aged 15–21 years (born between 1993 and 1999) spent zero CPYs in the higher-risk period; those aged 22, 1 year; those aged 23, 2 years, and those aged 24, 3 years, leading to a combined total of 6 CPYs. Therefore, the youngest GSS age group spent 6 out of their 150 CPYs (4%) in the time-period we hypothesized to be higher-risk and the remainder (96%) in the time-period (1993 or after) we hypothesized to be lower-risk.

Figure 1 shows how we calculated the percentages of the CPYs for the higher-risk period. Each shaded box represents a CPY spent in the higher-risk period. For GSS respondents aged 15–24 y, there are 6 shaded boxes (higher-risk) out of a total of 150 CPYs. The remaining 144 boxes that would fall to the right in the lower-risk period (1993 or after) are not shown in the figure. Those aged 25–34 y spent 57% of their CPYs in the higher-risk period and the remaining 43% in the lower-risk period of 1993 or after. Those aged 50–64 y spent all 225 of their CPYs in the higher-risk period from 1946 to 1992, and respondents aged 35–49 y, nearly all of their CPYs in this period. Respondents aged 65–79 y spent 71% of their CPYs in the higher-risk period and the remaining 29% in the lower-risk period before the end of WWII. Those aged 80–99 y spent almost all of their CPYs (98%) in the lower-risk period before the end of WWII.

Figure 1.

Number and percentage of childhood person years (CPY)a (ages 0 to 14 y) spent in child sexual abuse (CSA)b higher risk period (1946 to 1992) among respondents to General Social Survey (GSS) by age cohort.

For the age groupings, we were somewhat constrained by sample sizes. It would have been advantageous to have more detailed age groupings, particularly for the youngest cohorts. This would have allowed us to elaborate more on the specific years when declines in CSA occurred. However, small sample sizes precluded the use of finer age breakdowns.

Initially we examined the prevalence of reporting CSA by age cohorts among respondents whose household income was less than $20,000; however, due to low sample sizes, it was necessary to expand this category and consider those with household income less than $40,000. When considering perpetrator type, prevalence estimates of CSA were stratified by gender because perpetrators against girls are more likely to be family members than are those for boys.[27,28]. Low sample sizes precluded an examination of CSA prevalence by specific types of perpetrators for males. Among females it was possible to examine age group in relation to 4 types of perpetrators: family members, caregivers, other people known (but not family), or strangers. It was necessary to include step-mother/step-father with family members due to low sample counts.

We conducted our analyses using SAS Enterprise Guide 5.1 (SAS Institute Inc; Cary, North Carolina, USA). All estimates are based on weighted data. Variance and 95% confidence intervals were calculated using the bootstrap technique to account for the complex survey design of the GSS.[26]

Results

An inverted U (∩) association was observed between age group and the prevalence of reporting CSA (Table 2). The highest prevalence was observed for those aged 35–49 y (10.0%) and 50–64 y (11.4%). Individuals in these age groups would have spent all or most of their childhood years in the higher-risk period of 1946 to 1992. The lowest prevalence estimates were observed for those in the youngest and oldest age groups. Among those in the oldest age group (80 y or older), who spent 98% of their childhood years in the lower-risk period of 1945 or earlier, prevalence was 4.0%. For those in the youngest age group (15–24 y), who spent 96% of their childhood years in the lower-risk period of 1993 or after, prevalence was 2.4%. The inverted U association was observed for both sexes (Table 2) but the pattern was more pronounced for females. For males, it was necessary to combine age groups 15–24 y and 25–34 y due to the small number reporting CSA among those aged 15–24 y.

Table 2.

Prevalence of Reporting Child Sexual Abuse (CSA) by Sex and Age Group, Population Aged 15 or Older (Canada, 2014).

					n	n
					Reporting	Missing
	Year of birth	% Reporting CSA (95% CI)			CSA	CSA
Total		8.1		(7.8–8.5)	3,302	697
15–24	1990–1999	2.4	*	(1.7–3.1)	117	26
25–34	1980–1989	5.7	*	(4.7–6.6)	286	50
35–49	1965–1979	10.0	*	(9.1–10.9)	841	150
50–64 (reference)	1950–1964	11.4		(10.6–12.2)	1,310	234
65–79	1935–1949	9.8	*	(8.8–10.8)	672	183
80–99	1915–1934	4.0	*	(2.9–5.0)	76	54
Malesa		4.0		(3.6–4.4)	746	328
15–34	1980–1999	1.3	*	(0.8–1.8)	61	32
35–49	1965–1979	4.9		(3.8–5.9)	179	73
50–64 (reference)	1950–1964	5.8		(4.9–6.6)	309	113
65–79	1935–1949	6.4		(5.3–7.6)	182	92
80–99	1915–1934	2.0	*	(0.8–3.3)	15	18
Females		12.1		(11.5–12.8)	2,556	369
15–24	1990–1999	4.5	*	(3.1–6.0)	103	12
25–34	1980–1989	9.2	*	(7.6–10.9)	239	32
35–49	1965–1979	15.1		(13.7–16.4)	662	77
50–64 (reference)	1950–1964	16.9		(15.6–18.1)	1,001	121
65–79	1935–1949	12.8	*	(11.2–14.4)	490	91
80–99	1915–1934	5.3	*	(3.6–6.9)	61	36

Source: Statistics Canada, GSS: 2014 General Social Survey: Victimization.

Note: CSA, Child sexual abuse.

aAmong males, age groups 15–24 and 25–34 were combined due to the small number reporting child sexual abuse. It is recommended that if the sample count contributing to the calculation of an estimate is less than 15, the estimate should not be released.[26]

* Significantly different from reference (P<0.05).

Likewise, an inverted U association between age group and the prevalence of CSA was observed for those living in households with income less than $40,000 (Table 3). Among Indigenous peoples, compared with those spending most of their childhood years in the higher-risk period, prevalence was lower for those aged 15–34 y (who spent more of their childhood years in the lower-risk period of 1993 or after). However, prevalence was high for those who spent most of their childhood years in the period before the end of WWII. The same inverted U association was observed between age group and CSA for females, regardless of the perpetrator identity.

Table 3.

Prevalence of Reporting Child Sexual Abuse by Age Group and Selected Variables and Subpopulations, Population Aged 15 or Older, Canada, 2014.

	Year of birth	% Reporting CSA (95% CI)
Prevalence among those living in
households with income less than $40,000		10.5		(9.6–11.5)
15–24 y	1990–1999	6.0	*	(2.7–9.4)
25–34 y	1980–1989	9.9	*	(6.6–13.2)
35–49 y	1965–1979	13.3		(10.5–16.2)
50–64 y (reference)	1950–1964	14.8		(12.7–16.9)
65–79 y	1935–1949	10.1		(8.5–11.6)
80 or older	1934 or earlier	4.1	*	(2.4–5.8)
Prevalence among Indigenous peoplesa		13.8		(11.4–16.2)
15–34 y	1980-1999	7.0	*	(4.4–9.6)
35–49 y	1965-1979	16.9		(11.2–22.6)
50–64 y (reference)	1950-1964	19.5		(14.2–24.8)
65 y or older	1949 or earlier	20.1		(12.3–27.8)
Prevalence of females reporting CSA by a family memberb		5.4		(5.0–5.8)
15–24 y	1990–1999	1.6	*	(0.8–2.4)
25–34 y	1980–1989	4.3	*	(3.1–5.5)
35–49 y	1965–1979	6.6		(5.6–7.6)
50–64 y (reference)	1950–1964	7.4		(6.5–8.3)
65–79 y	1935–1949	6.1		(5.0–7.2)
80 or older	1934 or earlier	2.3	*	(1.1–3.5)
Prevalence of females reporting CSA by a caregiverc		0.6		(0.5–0.8)
15–34 y	1980–1999	0.5		(0.2–0.7)
35–49 y	1965–1979	1.0		(0.5–1.4)
50–64 y (reference)	1950–1964	0.7		(0.5–1.9)
65 y or older	1949 or earlier	0.3	*	(0.1–0.5)
Prevalence of females reporting CSA by other
known person (not family)d		3.7		(3.3–4.1)
15–24 y	1990–1999	1.9	*	(0.8–3.0)
25–34 y	1980–1989	2.4	*	(1.7–3.2)
35–49 y	1965–1979	4.6		(3.8–5.4)
50–64 y (reference)	1950–1964	5.0		(4.2–5.7)
65–79 y	1935–1949	3.8		(2.8–4.8)
80 y or older	1934 or earlier	1.8	*	(0.9–2.7)
Prevalence of females reporting CSA by a strangere		2.2		(1.9–2.5)
15–24 y	1990–1999	0.7	*	(0.2–1.1)
25–34 y	1980–1989	1.2	*	(0.7–1.8)
35–49 y	1965–1979	2.6		(2.0–3.2)
50–64 y (reference)	1950–1964	3.5		(2.8–4.2)
65–79 y	1935–1949	2.6		(1.9–3.2)
80 y or older	1934 or earlier	0.8	*	(0.3–1.3)

Source: Statistics Canada, GSS: 2014 General Social Survey: Victimization.

Note: CSA, Child sexual abuse.

a Among Indigenous peoples age groups 15–24/25–34 and 65–79/80 or older were combined due to the small number reporting CSA.

b Mother, father, step-mother, step-father, grandmother, grandfather, brother, sister, or other family member.

c Teacher/professor/tutor, baby-sitter, nanny. Age groups 15–24/25–34 and 65–79/80 or older were combined due to the small number reporting CSA.

d Boyfriend/girlfriend, ex-boyfriend/ex-girlfriend, neighbour, friend, acquaintance, classmate.

e Stranger or known only by sight.

* Significantly different from reference (P < 0.05).

Discussion

This study, based on data from the GSS 2014, supports the findings from an earlier study (based on data collected in 2004/2005 and 2012), suggesting that, after an increase in the post-WWII period, there has been a decline in CSA in Canada since the early 1990s. The pattern was more pronounced among females. The inverted U (∩) association between age groups and the prevalence of reporting CSA underscores the importance of considering age cohorts when examining trends in CSA.

Evidence of an increase in the post-WWII period was found in previous Canadian[8] and US research.[10,13,15] A decline in more recent years has also been documented in the US,[11] Australia,[9] Finland,[12] New Zealand,[29] and Canada.[16] Potential reasons for the earlier increase and the more recent decline have been discussed elsewhere.[16,30,31]

Among female respondents, we found evidence for a decline in CSA since the early 1990s regardless of perpetrator identity. Sample sizes were insufficient to examine trends by perpetrator among males.

Even in the subpopulations most likely to have experienced high levels of CSA, such as Indigenous respondents and respondents living in lower income households, we found evidence of a decline in CSA since the early 1990s. Our finding that prevalence of reporting CSA is lowest for the youngest, compared to older, Indigenous respondents is encouraging. Those in the youngest cohort did not experience CSA in residential schools because they were born after the last residential school was closed in Canada. As well, programs developed to assist in healing processes may have been instrumental in breaking cycles of violence and thus protecting children.[19] Finally, the acknowledgement by the government of Canada of the harms caused by residential schools may have contributed to strengthening Indigenous communities.[19] However, among Indigenous peoples, there was no evidence that the period before the end of WWII was a period of lower risk of CSA; for all the Indigenous subgroups age 35 y and older, the prevalence reporting CSA was high. A possible explanation is that Indigenous children as a group were less likely to be safe before WWII, due to victimization in residential schools and colonization more generally.[19,32]

Differences by age cohorts in willingness to disclose CSA on a victimization survey may complicate the interpretation of our results. However, 2 studies partially addressed this issue. An Australian study asked respondents about their “openness” and “comfort” level during the interview and the authors concluded that these factors did not relate to the prevalence of reporting CSA by age cohort.[9] A social desirability scale was included in a study based on the Finnish population. It was concluded that the declining trend could not be explained by a higher likelihood of socially desirable responses by younger cohorts.[12]

Strengths and Limitations

The strengths of this study include confidence in the validity of the CSA assessment, given the use of multiple behaviour-based items.[33] Additionally, replication of the use of the CPY approach to compare age cohorts within this large Canadian representative survey and within subpopulations provides an advance in our understanding of trends in CSA in Canada. To date, findings from 3 Canadian national surveys, conducted over 10 years, with different methodologies and survey context (a victimization survey, a mental health survey and a survey on alcohol-use problems) provide concurrent validation that CSA has declined in Canada since the early 1990s.

There are some limitations to consider when interpreting the results of this study:1) The extent to which the low response rate to the GSS has an impact on prevalence estimates of CSA is unknown. The prevalence estimates based on GSS are somewhat lower than those observed in the CCHS and GENACIS (based on matched date of birth cohorts). One reason for the lower GSS CSA prevalence estimates is the reference period for reporting CSA experiences. GSS respondents were asked about incidents that occurred before the age of 15 whereas CCHS and GENACIS respondents were asked about incidents that occurred before age 16. 2) A second limitation is the non-response to the items on CSA. These respondents either refused to answer the questions on CSA or responded, “don’t know”. An analysis of the characteristics of these non-respondents indicated that they were more similar to those who reported CSA than those who did not (e.g., they were more likely to report poor health outcomes). As well, non-response to these items was more common among older respondents, which could impact our age cohort analyses. We conducted sensitivity analyses to address this issue. We assumed that anyone with missing values had experienced CSA (i.e., we imputed “yes” to missing values for the CSA items). When we reran the prevalence estimates with the imputed values, the same inverted U association persisted between age groups and CSA for both sexes and the significance testing among the age groups remained identical. 3) The coverage of the GSS excluded people who were homeless and those living in institutions—populations for which experiences of CSA are likely more prevalent. It is possible that trends in CSA may differ for these vulnerable sub-populations. 4) As the length of time since the CSA occurred increases, so too will downward biases on prevalence estimates due to premature mortality or residing in health care institutions; people who have died or who are living in institutions cannot be in the sample and report their CSA. Therefore, evidence of an increase between the pre- and post-WWII periods is likely to be erroneously exaggerated, and evidence of a decline since 1993 may be somewhat understated. 5) When examining trends by income, ideally, we would have considered household income at the time the CSA occurred but instead used household income at the time of the survey. 6) Our CPY analysis is limited by some small cell sizes. It would have been advantageous to have smaller age groupings particularly for the youngest cohorts. However, small sample sizes necessitated collapsing the youngest 2 cohorts for males and Indigenous peoples. 7) When examining trends by perpetrator type, it was necessary to include step-mother/step-father with family members due to low sample counts. It is possible that respondents may have interpreted situations where the perpetrator was a parent’s partner differently (e.g., as a step-parent, other family member or acquaintance). 8) Finally, it was assumed that the risk of CSA was constant within each of the identified risk periods, and that people of each age contributed equally to each cohort.

Conclusion and Implications

Although retrospective surveys are effective in providing evidence of trends in CSA, population-representative repeated surveys of school-aged children would give more timely information to assess potential further progress. A recent review suggests that youth self-report surveys conducted with careful attention to ethics and confidentiality are a viable way of collecting information on the occurrence of child maltreatment incidents.[34] Such surveys would be useful in assessing the need for support services to children and families, and for programs aimed at eradicating CSA.

In a 2012 position paper, the National Children’s Advocacy Centre concluded that the evidence of a decline in CSA in the US was convincing enough to merit a message of progress: “There are solutions that work, we are making incredible progress, and everyone has a role to play in efforts to end child abuse.”[35] In Canada, we now have evidence from 3 retrospective population surveys suggesting a corresponding decline. However, given its immediate and long-term harms, continued work toward the eradication of sexual abuse of children is essential and in keeping with the United Nations sustainable development goal to end all forms of violence against children.[36]

The universality of the decline suggests that public health prevention efforts also have an effect in disadvantaged communities. As such, prevention efforts could focus on targeting economically and otherwise disadvantaged groups where CSA is more common, to hasten the decline and promote health equity.