A growing body of literature suggests that child maltreatment and traumatic stressors have long-term consequences for adult health behavior and adult health outcomes. Adults with a history of child maltreatment have been shown to be at increased risk of multiple health risk behaviors and risk factors for chronic disease including overweight or obesity [1,2], smoking [3-5], and physical inactivity [6], and some have reported associations with higher prevalences of heart [7-9], lung [10], and liver[11] disease, diabetes [9], and depression or depressive disorders [12-17] – chronic conditions associated with higher health care service utilization.

Health care resources are scarce, and therefore research focused on explaining and/or predicting utilization of health care services is both of interest and of importance to policymakers. Use of prescription drugs, one of the most rapidly growing sectors of health care in the US, has increased substantially over the past 10 to 15 years, although use of some prescription drugs, such as hormone replacement therapy [18] and alpha- (α-) blockers [19], has decreased. In 2000, three in five US adults filled at least one prescription. [20] Nationally, the number of purchased prescription drugs increased 75% between 1993 and 2005 from 2.0 to 3.5 billion per year. [21,22] In 2004 an estimated 45% of American adults were taking prescription drugs on a regular basis and 27% were taking them occasionally [23].

Increases in prescribed drug use are reflected in expenditures for these medications. Spending on prescription drugs in the United States in 2003 was nearly $180 billion, or 11% of total national health expenditures, and was more than four times the amount spent on prescription drugs in 1990.[24] Future spending on prescription drugs is projected to increase 10.5% between 2005 and 2013, but this may turn out to be an underestimate as the projection was made before the passage of the Medicare Prescription Drug, Improvement, and Modernization Act of 2003 [25].

Few studies have assessed the relationship of child maltreatment and related traumatic stressors to the use of prescription drugs in adulthood. [26-29] Our earlier analyses of data from the Adverse Childhood Experiences (ACE) Study have demonstrated that an integer count of the number of categories of abuse, exposure to domestic violence, and other forms of serious household dysfunction (the ACE Score) [30] experienced during childhood has a strong graded relationship to a wide variety of health and social problems from adolescence to adulthood (Table 1). [2,3,6,7,11,12,31-43] Therefore, we hypothesized that the ACE Score would be associated with increased rates and number of classes of prescriptions drugs used during an average of 6.1 years of follow-up of adults in the ACE Study cohort.

In this large study of HMO patients we found that rates of prescriptions increased in a graded fashion as the ACE Score increased. This pattern was particularly evident in the younger age groups – whose rates were increased by as much as 60% for persons with ACE Scores of 5 or more. Previous research has shown ACEs to be associated with earlier onsets of health risks such as smoking [3], alcohol [34] and illicit drug use [36], and sexual intercourse. [31,35] Thus, ACEs may "accelerate" the onset of health risks and illnesses, in the process increasing the use of prescription drugs among younger persons. If this is the case, the greatest relative effects on prescription rates would be expected among younger persons.

Among older persons, the graded relationship between the ACE Score and prescription rates was attenuated. This could be due to differential morbidity and mortality because ACEs influence a multitude of health and social problems (Table 1). Older persons affected by ACEs might expectedly have higher levels of multimorbidity [31] or severe health problems (such as ischemic heart disease [7] or liver disease [11]), requiring ongoing specialty care. In this scenario, older persons would be less likely to visit a clinic for wellness care and hence, less likely to have enrolled in the study. Moreover, ACE Scores tend to be lower among older persons, [6] possibly as a result of increased mortality over time leading to a decreased likelihood that persons with high levels of ACEs would survive to be in the older age groups included in the study.

Our finding that the risk of being prescribed multiple classes of drugs during follow-up increased in a strong graded fashion as the ACE Score increased lends further support for the idea that ACEs matter long after they occur. We found a 2-fold increase for young and middle-aged persons and 1.7-fold for the older persons with Scores ≥ 5. Thus, there was little evidence of attenuation among older persons as was observed for rates of prescription use. This is likely due to the choice of the upper decile of number of drug classes – an extreme measure – which may have selected persons with comorbid conditions that resulted from exposure to ACEs. We have previously shown that the mean number of a variety of health-related problems [31] and the number of risk factors for the leading causes of death increases as the ACE Score increases [6].

When we controlled for documented ACE-related health and social problems (Table 1), the apparent effects of ACEs on rates of prescription drug use were reduced by 67–76% (median, (69%); similarly, the risk of using a high number of classes of drugs during follow-up was reduced by 26–67% (median, 61%). Thus, as would be expected, the documented ACE-related conditions among participants appear to account for the majority, although not all, of the increase use of prescription medications we observed. Because screening for childhood traumatic stressors is not yet a routine part of adult medical care, some clinicians are likely identifying and treating these conditions without a full understanding of their origins in the long-term neurobiologic effects [31,48,49] of childhood stressors.

Relationships between child maltreatment and prescription drug use in adulthood among adult survivors of child maltreatment have been examined previously, but the relationships remain unclear as studies are often limited by study design, use of clinical populations versus community-based samples, self-reported health care service utilization measures, examination of only one or two types of maltreatment, and suboptimal statistical analyses as in the case of no multivariable adjustment of comparisons. In our study, the multivariable-adjusted relative rate of prescriptions as well as the relative risk of use of a high number of classes of drugs increased with a higher ACE Score. In a study of 3333 women aged 18–64 years who were members of a large health maintenance organization in the northwestern United States, Bonomi and collegues [26] observed that women with a history of both physical and sexual childhood abuse had more pharmacy fills (adjusted incidence rate ratio = 1.57; 95% CI: 1.33–1.86) than women without a history of child physical or sexual abuse. In a study of 150 women aged 17–49 years seen consecutively for non-emergency medical care by a family practitioner in a health maintenance organization, Sansone and colleagues [29] found a significant association between sexual abuse and the number of prescribed medications, obtained from a physician review of patient medical records, during the 12-months following completion of a clinic survey; lifetime physical or emotional abuse were not associated with use. Participants were not queried for abuse that occurred only during childhood; rather, for each type of abuse participants provided an age range during which the event(s) occurred.

Using data from women members of a health maintenance organization, Farley and Patsalides [27] report significantly more prescription medications obtained from medical record review among women with a history of childhood physical and sexual abuse (n = 27) compared to women without a history of either form of abuse (n = 26). The study is limited by a low response rate (14%) for the mailed survey, the absence of any quantitative data on the prescription drug use, and the absence of any multivariable adjustment in the statistical comparisons of abuse groups. In a clinical sample of 75 women with fibromyalgia, Alexander and associates [28] observed an increased use of pain medications and greater outpatient service use among women with a history of sexual or physical abuse compared to those without such a history. The study did not stipulate when the abuse occurred (i.e., childhood or adolescence vs adulthood) and did not include multivariable-adjusted statistical analyses.

The analyses herein have several strengths. Prescriptions were obtained prospectively from electronic, administrative pharmacy claims data and therefore are not subject to differential misreporting. The relationship of ACEs to prescription drug use is not limited to any specific class of drugs (data not shown). Future analyses will detail the relationship of ACEs to increased use of individual classes of drugs.

Our results should be interpreted keeping the following limitations in mind. Because of the sensitive nature of questions about ACEs and affective problems, the responses probably represent an underreporting of their actual occurrence. However, our estimates of the prevalence of childhood exposures are similar to estimates from nationally representative surveys [50,51] indicating that the experiences of our participants are comparable to those of the larger population of adults. For example, in our study we found that 16% of the men and 25% of the women met the case definition for contact sexual abuse; a national telephone survey of adults conducted by Finkelhor and colleagues [52] using similar criteria for sexual abuse estimated that 16% of men and 27% of women had been sexually abused. Of the men in our study, 28% had been physically abused as boys, which closely parallels the percentage (31%) found in a population-based study of men in Ontario that used questions from the same scales. [53] The similarity in estimates of the prevalence of these childhood exposures between the ACE Study and other population-based studies suggests that our findings are likely to be applicable in other settings. Also, when utilizing retrospective reports of adverse childhood experiences, several factors may inevitably lead to variability in the responses over time. These include difficulty recalling the experiences the due to the time lapse between the events in question and the research survey, variability in responses may occur due to the sensitive nature of the questions and the subjects' knowledge of the "social taboos" of responding to such questions, and incomplete or total inability to recall the experiences due to memory impairments as a result of stressful childhood experiences. Dube and associates [54] observed that retrospective responses to childhood abuse and related forms of serious household dysfunction are generally stable over time concluding that there is good to excellent reliability in the reports of adverse childhood experiences during adulthood.

The mind-body dichotomy that persists in Western medical training may lead clinicians away from understanding the role that childhood trauma and stress has on the health of their adult patients. Childhood stressors are known to produce changes in the developing brain that affect emotions, behavior, and cognition [55] which in turn can impair health and quality of life via numerous pathways (Table 1). These traumatic pathophysiological insults may be "silent" until much later in life, [55,56] leading clinicians to prescribe medications to treat symptoms and illnesses without knowledge of their potential origins in the disruptive effects of ACEs on neurodevelopment. As improvements in the treatment of persons affected by traumatic stress evolve, understanding the role of these childhood experiences on adult health will become increasingly important in making decisions about prognosis, diagnosis, and treatment.

The authors declare that they have no competing interests.

Study conception and design: RFA, DWB. Acquisition of data: RFA, VJF, SRD. Analysis and interpretation of data: RFA, DWB, WHG. Drafting of manuscript: RFA, DWB, VJF, SRD, WHG. Critical revision: RFA, DWB, VJF, SRD, WHG. All authors read and approved the final manuscript.

The pre-publication history for this paper can be accessed here: