We conducted a descriptive analysis of survey results obtained during the enrollment phase of a prospective, randomized longitudinal study, which evaluated the impact of patient-targeted educational brochures on patient-initiated contacts with local support resources and patient implementation of harm-reduction measures.
The study took place at an ED based in the only Level One trauma hospital of a large, southeastern U.S. city. The hospital is academically affiliated and staffed by faculty and residents from two local medical schools, and the ED sees approximately 105,000 patient visits each year. This study was reviewed and approved by the institutional review boards of our university and the hospital research oversight committee.
All African-American women seeking medical care in the ED who were in the waiting room between the ages of 18–55 were eligible for participation in this study, regardless of chief complaint. Women were excluded if they did not speak English, if they were acutely intoxicated, critically ill, currently taking anti-psychotic medication, or if they were otherwise unable to stand for 15 minutes.
Eligible patients were approached by research assistants (who were present in the ED Monday through Wednesday, 12pm to 8pm) informed about the nature of the study and asked if they were willing to participate. Women who agreed then read and completed an informed consent form and were t taken to a private booth in the ED to complete the survey on a touch-screen computer. The survey was designed with a skip pattern, with the first question in each section inquiring about relationship status or substance use in the prior 12 months. If participants answered “no” to this initial question, the survey advanced to the next section.
The survey included questions from several previously validated instruments, including the Index of Spousal Abuse (ISA),11
the Tolerance, Worried, Eye openers, Amnesia, K(Cut) down survey (TWEAK, an alcohol-abuse survey),12
Drug Abuse Screening Test (DAST),13
the Hooked on Nicotine Checklist (HONC),14
the Beck Depression Inventory (BDI),15
as well as a brief questionnaire assessing participants’ self-report of their general health, health behaviors, as well as economic and interpersonal resources.
The Index of Spousal Abuse is a 30-item scale designed to detect spousal abuse in women. There are two subscales, the ISA-P (measuring severity of physical abuse) and the ISA-NP (measuring severity of nonphysical abuse).16
Each question is answered on a Likert-type scale and scored on a scale of 1–5 points each. The questions are phrased in the present tense and focus on detection of abuse at the time of survey administration. For the purposes of this study, any woman with an ISA-P score ≥ 10, or an ISA-NP score of ≥ 25 was considered to have a positive IPV screen.17
We used the TWEAK (Tolerance, Worried, Eye openers, Amnesia, K(C)ut down) scale for detecting alcohol abuse. This instrument, which consists of five questions, was developed by combining elements from both the MAST and CAGE questionnaires.18
The questions are all answered yes/no, and the test is scored on a seven-point scale. In this study, a score of ≥ 2 was used to identify a positive screen.19,20
The Drug Abuse Screening Test (DAST) is a questionnaire consisting of yes/no answers. In this survey, we utilized the DAST-20, an abbreviated format shown to correlate nearly perfectly (r=0.99) with the longer 28-item survey.21
In this survey, we utilized a score of ≥ 6 to indicate a positive screening for drug dependence.
The Hooked on Nicotine Checklist (HONC) is a 10-item tool initially developed to assess adolescents’ loss of autonomy over tobacco, and has since been validated for use in adults.22
In this yes/no questionnaire, we utilized a score of ≥ 1 to indicate a positive screen.
The Beck Depression Inventory, II (BDI- II) was used to assess the presence and severity of depressive symptoms.23
In this study, we used a BDI-II score of 20 or greater, consistent with moderate to severe depression, as a positive depression screen. Prior validation studies have established an overall classification rate of 88% using this cut point (sensitivity 71%, specificity 88%).24
The survey also included a series of questions about general health and well-being, including questions about patients’ perceived state of health, family medical history, social and family situation, and health-related behaviors.
We analyzed the survey data utilizing t-test and chi square analysis to determine the associations between IPV status and presence of mental health symptoms, alcohol or substance abuse, and general health assessment and social/family support. All usable data were included for participants who were unable to finish the survey.