Brief Report: A Web-Based Pilot Study of Childhood Sexual Abuse, Recent Stressful Events, and Alcohol Use in Women

Abstract

Background and Objectives

Childhood sexual abuse (CSA) is a well-documented risk factor for problem drinking, but the possibility that recent stressful events exacerbate risk conferred by CSA has rarely been examined, and the few studies to do so have limited their investigations to negative events. The aim of the current study was to examine the associations between CSA, recent positive and negative stressful events, and women's alcohol consumption.

Methods

Eighty-nine women (42.7% reporting CSA) completed weekly web-based alcohol use surveys and a stressful event inventory assessing negative and positive events over a 12 week period.

Results

Independent of CSA history, experiencing negative events was associated with elevated likelihood of drinking to intoxication.

Discussion and Conclusions

Proximal distressing events are more closely linked than CSA history to short-term drinking patterns.

Scientific Significance

Tracking stressful negative events may be informative for identifying precipitants of heavy drinking.

INTRODUCTION

One in five women has experienced childhood sexual abuse (CSA),¹ a well-documented risk factor for problem drinking.² Children who have experienced abuse are at elevated risk for a range of stressful events later in life.³ Sexually abused girls are two to three times more likely to be sexually assaulted as adults than those without a history of CSA,⁴ and sexual assault – as well as less severe stressors—have been linked to heavy alcohol use.^5,6

The increased vulnerability to negative outcomes following stressful events in individuals with a history of child maltreatment, (i.e., “stress sensitization”), first demonstrated for depression, has only recently been examined in alcohol studies, but findings are consistent. Among individuals exposed to stressful events, heavy drinking⁶ and severe alcohol craving³ are more common in those with versus without a history of child maltreatment.

The aim of the current study was to examine the relationship between CSA, stressful events, and alcohol consumption in women, using a prospective web-based design. It expands on prior studies by: (i) querying exposure to stressors over a limited period of time (versus past year); and (ii) including positive stressful events – events typically viewed as positive but involve adapting to new circumstances and drawing on emotional or other resources (e.g., completing school) – in addition to negative ones (e.g., physically attacked). This allowed us to assess the relevance of the type of event to its association with alcohol use. We hypothesized that alcohol use would be elevated in women with a history of CSA who experienced either positive or negative stressful events.

METHODS

Participants

The sample was recruited from participants in the Missouri Adolescent Female Twin Study (MOAFTS), a longitudinal study of alcohol and related psychosocial outcomes in women (described in prior publications⁷). A total of 101 women were ascertained based on their CSA history. The final sample was composed of 89 women who had data on CSA status (42.7% with a history of CSA), completed the Stressful Life Event Inventory, and provided data for at least 9 weeks that the weekly survey was administered. The 12 women whose data were not used did not differ from the rest of the sample with respect to demographics or CSA history. Mean age was 31.6 (SD=2.7); 92.4% self-identified as White, 7.6% as African-American.

Procedures

Eligible MOAFTS participants were mailed recruitment letters, then contacted by telephone. Women who agreed to participate (88.9% of those sent letters) were scheduled for an orientation to the web-based survey system and sent a packet containing a consent form and study payment information. A research assistant conducted the orientation over the telephone, first reviewing the consent form and obtaining verbal consent, then e-mailing the participant a link to a blank survey that she opened and went through with the research assistant.

Alcohol use data were collected over a period of 12 weeks (a sufficient timeframe for capturing relatively infrequent severe stressors) using weekly rather than daily assessments to maximize participant retention. On Sunday mornings, participants received an e-mail containing a link to the Diary of Substance Use and Related Behaviors, with the survey's unique code and the participant's unique identifier. In week 12, participants received a link via e-mail to the Stressful Life Event Inventory.

Predictors

CSA status was derived using items from three sections of structured psychiatric interviews (the Semi-Structured Assessment for the Genetics of Alcoholism⁸) conducted with MOAFTS participants: (i) traumatic events (rape, sexual molestation); (ii) childhood experiences (forced sexual contact before age 16 with a family member or with a non family member 5 or more years older); and (iii) health problems (ever forced to have sexual intercourse). Consistent with prior MOAFTS publications using CSA data, criteria for CSA were met if any sexual abuse question in any wave of data collection was endorsed, with first occurrence before age 16.

Stressful events were assessed using the Stressful Life Event Inventory, a 34-item measure composed of all 17 items referring to stressful events during adulthood in the Life Stressor Checklist-Revised (LSC-R),⁹ 15 items from the Psychiatric Epidemiology Research Interview Life Events Scale,¹⁰ and two additional items (partner/spouse lost job, involved in serious accident). Participants reported the number of times and the dates on which each event occurred during the 12-week period. Events were categorized (by the authors based on the conceptual content) as negative (25 events, e.g., sexually assaulted, lost job) or positive (9 events, e.g., got married, made major purchase such as a house).

Outcomes

Mean weekly frequency of alcohol use, mean drinks consumed on weekday (Sunday through Thursday) and weekend (Friday and Saturday) drinking days, any drinking to intoxication, and maximum drinks consumed in one day (max drinks) during the 12 week period were derived from the Diary of Substance Use Behaviors and Environments. This 13 item measure was created for the current study to collect day by day reports of alcohol use (i.e., number of drinks consumed and drinking to intoxication), cigarette smoking, marijuana use, exposure to environments where these substances are commonly used, and context of use (i.e., places and people present).

Data Analysis

Frequency, quantity consumed on weekdays and weekend days, and max drinks variables were log transformed (due to skewness) and analyzed using multiple regression analyses. Due to the low frequency of drinking to intoxication, drinking to intoxication was analyzed as a dichotomous variable using logistic regression analyses. Interaction effects were tested by entering interaction terms representing CSA history by event exposure (dichotomously coded) into models. Analyses were adjusted for age, marital status, level of education, and household income. The clustered sandwich estimator was applied to adjust for the non-independence of observations in twins.

RESULTS

Stressful Events and CSA History

Over the 12 week period, 48.3% of participants reported experiencing one or more stressful events (27.2% negative, 35.9% positive). As seen in Table 1, neither the proportion of women endorsing events nor the number of separate events reported differed by CSA status for either event type.

TABLE 1.

Exposure to negative and positive stressful events during the 12 week study period by childhood sexual abuse (CSA) history and drinking behaviors by CSA history and stressful event exposure

Negative and positive stressful event exposure by CSA history
	Any negative event		Number of negative events (among those reporting any)		Any positive event		Number of positive events (among those reporting any)
	%	OR (RSE) [95%CI]	Mean (SD)	t(23)=.54 p=0.59	%	OR (RSE) [95%CI]	Mean (SD)	t(30)=.11 p=0.92
CSA	26.32	.86 (.39) [.35, 2.10]	2.10 (1.37)		36.84	1.06 (.52) [.41, 2.79]	1.36 (.50)
No CSA	29.41		1.73 (1.79)		35.29		1.39 (1.04)

CSA history, stressful event exposure, and drinking behaviorsa,b
	β (RSE) [95%CI for β]				OR (RSE) [95%CI]
	Mean drinking days per week (n=83)	Mean drinks on weekday drinking daysc (n=73)	Mean drinks on weekend drinking daysd (n=80)	Maximum drinks in one day (n=83)	Became intoxicated (n=83)
CSA	−.07 (0.37) [−.82, 0.68]	.22 (.16) [−.30, .35]	.18 (.25) [−.33, .68]	.16 (.23) [−.29, .61]	2.00 (1.41) [.50, 7.99]
Negative event	.17 (.33) [−.50, .84]	.14 (.13) [−.13, .41]	.10 (.15) [−.21, .41]	.00 (.23) [−.47, .47]	7.64* (6.55) [1.42, 41.01]
CSA × Negative event	.32 (.60) [−.89,1.54]	.10 (.24) [−.38, .57]	−.24 (0.38) [−1.01, .53]	.41 (.39) [−.37, 1.20]	.24 (.31) [.02, 3.06]
CSA	−.19 (.36) [−.93, .54]	−.32 (.16) [−.36, .30]	.05 (.22) [−.39, .49]	.10 (.24) [−.37, .58]	1.20 (.81) [.32, 4.50]
Positive event	−.46 (.37) [−1.21, .29]	−.12 (.16) [−.45, .21]	−.12 (.18) [−.47, .24]	−.22 (.21) [−.65, .20]	.94 (.64) [.25, 3.58]
CSA × Positive event	.53 (.60) [−.68, 1.74]	.19 (.26) [−.33, .71]	.13 (.27) [−.41, .67]	.40 (.33) [−.25, 1.06]	1.06 (1.02) [.16, 6.93]

OR, odds ratio; RSE, robust standard error; CI, confidence intervals.

^aAdjusted for age, marital status, level of education, and household income.;

^bAmong participants who endorsed alcohol use during the study period.;

^cparticipants who endorsed alcohol use on weekdays during the study period.;

^dAmong participants who endorsed alcohol use on weekends during the study period;

^*p<.05.

Patterns of Alcohol Use by CSA History and Stressful Event Exposure

No differences in likelihood of using alcohol (any, weekdays, or weekends) by exposure to negative events, positive events, or CSA history were observed, with the exception of CSA and any alcohol use. Women with a history of CSA had a reduced likelihood of consuming any alcohol during the study period (OR=.11, 95%CI:.02–.96 in the unadjusted model). As seen in Table 1, in adjusted models (based on drinkers), neither CSA nor exposure to positive events were associated with any of the alcohol outcomes, but exposure to negative events was associated with drinking to intoxication (OR=4.28, 95%CI:1.54–11.87). No interactions between CSA and positive events or CSA and negative events were significant.

DISCUSSION

The current web-based study characterized short-term patterns of alcohol use in women in relation to CSA history and current stressful events. Capturing stressful events over a limited period of time allowed for a more nuanced examination of the joint effects of CSA and proximal stressors on alcohol use than prior studies in this area. The incorporation of positive events into the stressful event measure created the opportunity to assess the specificity of changes in drinking behaviors to negative events versus stressful events more generally.

We did not observe either heavier alcohol use or higher exposure to stressful events in women with a history of CSA, as have prior studies of child maltreatment.^2,3 This may be due to a diminishing impact of CSA associated risk over the approximately 20 years since first CSA exposure (which may not have been observed if the sample had been ascertained on severe abuse history). Furthermore, we did not find evidence for stress sensitization. Rather, results indicated that negative events are associated with potentially risky alcohol use independent of CSA history. Taken together, our findings support two primary conclusions: first, proximal stressful events are more closely linked than distal risk factors to patterns of alcohol consumption and second, this association is specific to distressing events – not just any event that involves significant adjustment to new circumstances.

Limitations

First, as this was a pilot study, we conducted a substantial number of analyses and did not adjust for multiple testing, so findings should be viewed as suggestive. Second, statistical power was limited due to the relatively small sample size. Additional associations may be detectable with larger samples. Third, although participants were cued with start and end dates of the 12 week study period, more frequent administration of the Stressful Life Event Inventory would increase confidence in the accuracy of participant recall of stressful events. Fourth, as the primary measures were created for the purposes of this study, we do not have test-retest reliability or validity data on them.

CONCLUSIONS SCIENTIFIC SIGNIFICANCE, AND FUTURE DIRECTIONS

Distressing events are common and are associated with short-term drinking patterns, even in the absence of a childhood trauma consistently associated with problem drinking. Tracking alcohol use and concurrent stressful negative events may be informative in clinical settings for identifying situations that precipitate problem drinking.