In Massachusetts, motor vehicle occupant injury is a leading cause of mortality among residents between 1 and 44 years of age.1 It is also a major cause of cognitive and functional impairment, resulting in significant financial losses to both individuals and society.2 According to the National Highway Traffic Safety Administration (NHTSA), safety belt use (SBU) reduces motor vehicle fatalities by as much as 45%.3 Consistent SBU using 3-point shoulder and lap belt occupant restraints has been shown to reduce death or serious motor traffic–related injuries to the head, chest, or abdomen by 86%.4 In 2004, while the national average of SBU in the United States was more than 80%, the prevalence of SBU among motor vehicle occupants in Massachusetts was among the lowest in the nation at 63%.5,6 To date, Massachusetts law authorizes “secondary” safety belt enforcement, allowing police to write a traffic citation for not wearing a safety belt only if a motorist is stopped for another violation (e.g., speeding). Increasing SBU in the United States is an important public health goal enumerated in the Healthy People 2010 objectives.7 To emphasize this, the U.S. Preventive Services Task Force has suggested that health care providers use their clinical encounters as an opportunity to discuss topics such as injury prevention with their patients.8 A visit to an emergency department (ED) could represent an important opportunity for providing harm reduction counseling to patients to improve their SBU.9–11 Bernstein et al.11 suggested that “hospital EDs offer a concentrated opportunity not available elsewhere for brief counseling during a teachable moment for contemplating change in behavior and improving outcomes.” Researchers are only now starting to take advantage of the acute care setting to engage patients in health-promoting behavior change.11–13 Much like in the primary care setting, there is a need to develop, evaluate, and disseminate brief motivational interventions for increasing SBU, tailored to the acute care environment.14,15 We sought to adapt techniques that have been successful in changing other health-related behaviors to improve SBU among individuals in the acute care setting who have, in terms of adopting harm reduction behaviors, been historically difficult to reach. Therefore, we conducted a study to test a brief motivational intervention targeted to increase self-reported SBU among ED patients.
Objectives: Brief motivational interventions have shown promise in reducing harmful behaviors. The authors tested an intervention to increase safety belt use (SBU) among emergency department (ED) patients.
Methods: From February 2006 to May 2006, the authors conducted a randomized trial of adult ED patients at a teaching hospital in Boston. ED patients were systematically sampled for self-reported SBU. Those with SBU other than “always” were asked to participate. At baseline, participants answered a 9-item series of situational SBU questions, each scored on a 5-point Likert scale. SBU was defined as a continuous variable (9-item average) and as a dichotomous variable (response of “always” across all items). Participants were randomized to an intervention or a control group. The intervention group received a 5- to 7-minute intervention, adapted from classic motivational interviewing techniques, by a trained interventionist. Participants completed a 3-month follow-up phone survey to determine changes from baseline SBU. Continuous and dichotomous SBU were analyzed via analysis of covariance and chi-square testing.
Results: Of 432 eligible patients, 292 enrolled (mean age 35 years, standard deviation [SD] ±11 years; 61% male). At baseline, the intervention and control groups had similar mean (±SD) SBU scores (2.8 [±1.1] vs. 2.6 [±1.1], p = 0.31) and SBU prevalence (each 0%). At 3 months, 81% completed follow-up. The intervention group had significantly greater improvement in mean (±SD) SBU scores than controls (0.76 [±0.91] vs. 0.34 [±0.88], p < 0.001). Also, SBU prevalence of “always” was higher for the intervention group than controls (14.4% vs. 5.9%, p = 0.03).
Conclusions: Participants receiving a brief motivational intervention reported higher SBU at follow-up compared to controls. An ED-based intervention may be useful to increase SBU.
This study was a randomized controlled trial of a brief motivational intervention targeted to increase self-reported SBU among ED patients who were identified as not consistently wearing their safety belt. The institutional review board at our institution approved the study. All participants provided informed consent.
Study Setting and Population
The study was conducted at a university teaching hospital Level 1 trauma center in Boston that treats more than 90,000 ED patients annually.
Inclusion Criteria. ED patients age >21 years presenting for care for any medical reason were eligible, provided that they spoke English, were considered “stable” according to the triage nurse on duty (i.e., did not have serious or life-threatening condition and were not assigned an “emergent” priority level), had a normal mental status, and were able to give free and autonomous consent. Mental status was assessed by the patient’s ability to demonstrate alertness and orientation to person, time, and place. Inclusion was based on the results of the single-item query regarding overall SBU (i.e., patients that gave any answer other than “always” wear safety belts were considered eligible). Questions regarding a patient’s stability level were addressed with the attending ED physician.
Exclusion Criteria. Patients were excluded if they had no phone or were homeless (making the follow-up telephone contact or locating the participant after initial enrollment logistically challenging); had altered mental status secondary to pain, distress, medication, brain injury, or acute psychiatric illness; were too ill to be interviewed (i.e., major trauma or medical illnesses); were prisoners in custody; or had not ridden in a car in the past 30 days. Patients who had used ethanol or another psychotropic agent were not specifically excluded unless the degree of intoxication resulted in an alteration in mental status that interfered with the ability to provide informed consent. As needed, the treating physician was asked to confirm the patient’s level of orientation.
Between the hours of noon and 8 pm on weekdays from February to May 2006, research assistants (RAs) who had completed National Institutes of Health (NIH) certification on human subject protection approached every adult ED patient (age >21 years) by systematically going room-to-room to assess patients for study eligibility. The RAs were supervised and monitored by a full-time research nurse.
Upon providing verbal assent, patients were issued a 1-page Health and Safety Screening Form. This confidential, self-administered form included demographic questions (age, gender, race, ethnicity, marital status, education), questions on an array of health-related issues (e.g., alcohol use, height and weight, firearms in the home, SBU), and a question asking if they were seeking care as a result of a car crash (“Is your ED visit today related to a car crash?”). A single screening question regarding SBU was embedded within the form (“Think about the times you’ve ridden in a car in the past 30 days. Overall how often did you wear a seat belt? ‘Always,’‘More than half the time,’‘About half the time,’‘Less than half the time,’‘Never’”). The SBU assessment question was imbedded within a larger health screening questionnaire to minimize the possibility of overreporting of SBU. Patients who screened positive (i.e., gave any answer other than “always” wear safety belts) on the SBU screening question were asked to participate in the study.
Participants gave written informed consent and were then asked to complete a detailed intake questionnaire, which included contact information, and a series of questions regarding situational SBU, scored on a 5-point Likert scale (“always” = 5, “never” = 1). SBU was assessed by nine separate questions on scenario-specific aspects of transportation (highway, local, daytime, nighttime, driver, front-seat passenger, back-seat, short rides, and long rides; see Appendix 1, available as an online Data Supplement at <link removed as it refers to this location>). The nine-question survey was developed via a focus group of public health researchers and emergency physicians. We piloted the questions among a sample of ED patients for content and syntax prior to commencing a previous study that demonstrated low self-reported SBU among ED patients at our facility.16 Further description of the measure, as well as a comparison between a single-item versus the nine-item measure, has been presented elsewhere.16 We defined self-reported SBU in two ways: first, as a continuous variable (an average of the scores on the nine items), and second, as a dichotomous variable (response of “always” vs. less than “always” across all nine items). This nine-item dichotomous measure was used to define SBU prevalence (i.e., prevalence of self-reported belt use of “always”).
Randomization. After completing the baseline intake questionnaire, participants were randomized into two groups: a control group that received a brochure with general injury prevention information and an intervention group that received a brief motivational intervention adapted to increase SBU (see Appendix 2, available as an online Data Supplement at <link removed as it refers to this location>). The intervention as performed in this study generally required 5 to 7 minutes to complete.
Telephone Follow-up. At 3-months postenrollment, research staff blinded to group assignment contacted participants from both groups. A structured, scripted telephone survey was used to determine changes in self-reported SBU from baseline among groups.
Descriptive statistics were used to characterize the baseline characteristics of the randomized subjects; comparisons between the intervention and control groups were analyzed through chi-square tests and Student’s t-tests as appropriate. Further descriptive analyses were employed to compare the baseline characteristics of those subjects in each group with follow-up responses compared to those lost to follow-up. To assess the difference between the two groups in SBU prevalence at 3 months, we used a chi-square test; to assess and compare the changes from baseline to 3 months in the two groups, we conducted analysis of covariance on change using the baseline value as a covariate. The study was powered to detect a 20% change between groups. Of those subjects with outcome data, the comparison of results was done between the two randomized groups only. The data analysis for this paper was generated using a commercially available software package (SAS Version 9, SAS Institute, Cary, NC).
Characteristics of Study Subjects
A total of 1,659 individuals were approached, and 1,377 (83%) patients completed the screening form. Of these, 524 screened positive (i.e., 38% of those that completed the screening form had not “always” worn their safety belt in the past 30 days), 427 met inclusion criteria, and 292 participants (68% of those eligible) were enrolled in the study (Figure 1). The median age of study participants was 32 years, and 113 (39%) were female. There were 56% (n = 164) African Americans, 1% (n = 3) Asians, 1% (n = 2) Native American, and 24% (n = 71) whites. There were 10% (n = 30) persons of Hispanic ethnicity. Among respondents, 86% had at least a high school education. Those who identified themselves as non-Hispanic African Americans (61%) tended to be more likely to enroll than non-Hispanic whites (54%), Hispanics (46%), or other race/ethnicities (48%) (p = 0.048). Those born in the United States were more likely to enroll than those who were not (58% vs. 48%, p = 0.04). Those with at least a high school education were more likely to enroll than those without (59% vs. 41%, p = 0.002). Among participants, 145 were assigned to the control group, and 147 to the intervention group. There were no statistical demographic differences according to group assignment (Table 1).
|Demographic Variable||Controls (n = 145), %||Intervention (n = 147), %||p-Value|
|Median Age (yr)||31||33||0.579|
|Some high school||9||16|
|Valid driver’s license?||67||65||0.683|
|Own a car?||48||43||0.353|
|Drove after drinking (past 30 days)||15||11||0.314|
|Ridden with a driver who consumed alcohol prior to driving (past 30 days)?||22||28||0.209|
|In the ED today as a result of a crash?||10||7||0.287|
|Ever been in car crash?||84||78||0.197|
|SBU (past 30 days)|
|More than half the time||31||31|
|About half the time||20||27|
|Less than half the time||25||27|
At baseline, the intervention group (n = 147) and control group (n = 145) had similar mean (±standard deviation [SD]) SBU scores (2.8 [±1.1] vs. 2.6 [±1.1], p = 0.31), and SBU prevalence of SBU of “always” across the nine items (each 0%). At 3-months postvisit, 81% (n = 236, 118 in each group) completed the follow-up phone call, 80% in the intervention group and 81% in the control group. Participants were more likely to follow-up if they were African American (compared to whites) or if they had a valid driver’s license. Other comparisons including education, gender, age, car ownership, previous car accident, drinking while driving in the past 30 days, riding with a driver who drank and drove in past 30 days, or comparison of the general SBU question were not significantly different between subjects who completed the 3-month interview and those who did not complete the 3-month interview. In addition, there was no significant difference in mean SBU scores in each of the nine categories of seat belt use in those who completed the 3-month interview compared to those who did not complete the 3-month interview. The intervention group had significantly higher improvement in mean (±SD) SBU scores than the control group at 3-month follow-up (0.76 [±0.91] vs. 0.34 [±0.88], p < 0.001). Also, the intervention group had 2.4 times greater increase in self-reported prevalence of SBU of “always” across all nine-items than the control group (14.4% vs. 5.9%, p = 0.03; Figure 2). As shown in Table 2, compared to the control group, the intervention group had statistically greater improvement on eight of the nine items at the 3-month follow-up. Even after adjusting for multiple testing via a Bonferroni correction (i.e., 9-item measures in addition to the overall score, or 0.05/10 = 0.005), seven of nine outcomes remain statistically significant.
|n||Mean (SD)||n||Mean (SD)|
|Highway SBU||116||0.33 (1.26)||118||0.58 (1.26)||0.035|
|Local streets SBU||118||0.28 (1.27)||118||0.74 (1.15)||<0.001|
|Daytime SBU||118||0.30 (1.29)||118||0.80 (1.14)||<0.001|
|Nighttime SBU||117||0.38 (1.14)||114||0.88 (1.21)||<0.001|
|Driver SBU||90||0.52 (1.16)||90||0.71 (1.18)||0.135|
|Front seat passenger SBU||117||0.45 (1.30)||116||0.82 (1.20)||0.002|
|Back seat passenger SBU||112||0.20 (1.27)||104||0.61 (1.23)||0.004|
|Less than 10 minutes SBU||118||0.08 (1.17)||117||0.71 (1.33)||<0.001|
|More than 10 minutes SBU||118||0.55 (1.16)||118||0.86 (1.10)||0.005|
|Overall change, intake to 3-month follow-up||118||0.34 (0.88)||117||0.76 (0.91)||<0.001|
Among adult patients attending a Boston ED, a brief motivational intervention was adapted to increase SBU behavior. We found that this brief ED-based intervention increased the self-reported prevalence of wearing a safety belt “always” at 3 months compared to a control group. Additionally, self-reported SBU among individuals randomized to the intervention group showed a statistically significant improvement in eight of nine separate SBU scenarios compared to the control group, at 3-months poststudy.
Participants in this study, drawn from an urban population of difficult-to-reach ED patients, were screened for SBU with a multi-item measure. None of the participants had self-reported SBU of “always” across all nine situations at baseline. At the 3-month follow-up period, 14.4% of participants in the intervention group had reported SBU of “always.” This rate was nearly 2.5 times greater than that of the control group that received a general injury prevention brochure. Overall, compared to baseline scores, participants randomized to the intervention group showed statistically significant improvement on eight of nine scenarios. Although differences were statistically significant, future investigation will be required to determine if these improvements over the control group represent observable differences. Much like in other risky behaviors, modest improvement in health promotion behavior may represent clinically meaningful change if adopted at the population level.12,16
Our findings are in accord with those of other studies that have examined the effects of interventions to increase SBU. Injured adolescents identified in an ED setting receiving a targeted brief motivational intervention for injury-related risk behaviors were noted at 3- and 6-months postenrollment to have 34% increases in SBU of “always” at 3-month follow-up.17 While injury severity in that study was associated with a positive behavior change, the added benefit of the intervention itself was only detectable in the less severely injured. In a study conducted among pregnant women, an educational intervention resulted in a 12.2% increase in self-reported SBU after the intervention, as well as a significant increase in knowledge of belt effectiveness.18 In a rural neighborhood health center in Arizona, brief counseling raised self-reported SBU by 15.2% from baseline.19
Adopting stricter legislation (i.e., primary belt enforcement), conducting highly visible enforcement campaigns (e.g., “Click It Or Ticket”), strategic use or both paid and earned media (e.g., television, radio, Internet), or improvements in belt use reminder technologies are countermeasures that have been shown to increase SBU among broad elements of the population.20–25 However, a more intensive approach, such as a brief motivational intervention conducted in a nontraditional setting, may be required for individuals that are resistant to the standard countermeasures.
In this randomized controlled trial we were unable to perform an intent-to-treat analysis. As mentioned above, among assignment groups, loss to follow-up was not significantly related to any known characteristic. However, this does not fully guarantee an unbiased estimate of the true outcome. An intention-to-treat analysis could provide a conservative estimate of the treatment effect compared with what would be expected if there was 100% follow-up. The true outcome in this study is only estimated from the data at hand and may or may not be correct depending on all the outcome data. Additionally, at present the score for the multi-item measure of self-reported SBU does not have intuitive meaning in clinical practice. However, from a public health perspective, any degree of improvement in self-reported SBU score among ED patients reporting less than “always” wearing their belts represents meaningful change. Further evaluation of this 9-item measure compared with directly observed SBU is necessary to determine the degree to which this score corresponds to a specific level of SBU behavior.
We acknowledge that self-reported methods of assessing SBU often suffer from some degree of overestimation of actual SBU, compared with observational methods.26,27 However, we assert that the multi-item measure used in this study enhanced its level of rigor. These conservative measures may estimate actual belt use better than a single-item query.26 Based on our study design, we suspect that overreporting of SBU behavior by respondents was minimized. Additionally, the multi-item measure of SBU was used to identify various situation-specific motor vehicle occupant scenarios, which may aid in delivering situation-specific risk communication messages to part-time belt users. Although these situations were not mutually exclusive (i.e., highway travel often consists of trips longer than 10 minutes), based on an informal focus group of public health experts and emergency clinicians alike, these represent common scenarios in which motorists commonly engage. In this pilot study, these items were scored equally; however, specific items could be given increased weight according to vehicle miles traveled as a risk exposure proxy in the future.
A motor vehicle occupant injury-related ED visit itself has been shown in previous studies to have a positive effect on behavior change.17,28 However, our study was not a priori powered to examine this issue (i.e., an injury that sensitizes a patient to receiving messages regarding behavior change—the “teachable moment”). In future studies, oversampling of injured patients could be conducted to have sufficient power to adequately test this theory.
We provided an injury prevention brochure to participants assigned to the control group. This is not, strictly speaking, a true control group. With a true control group, the differences between groups in this study may have been even greater. To examine the effects of providing the brochure, a second control group that was provided neither the intervention nor the brochure would be required.
The generalizability of our findings to the general population may be limited, given that our sample was drawn from ED patients in Boston, Massachusetts, and may differ from other populations (i.e., race/ethnicity, educational level, percentage with a driver’s licenses, or who own their own vehicle). Apart from geographic differences, clinical populations can differ from population-based samples in terms of health and other factors, including SBU.29–31 However, it has been reported that ED patients have higher injury-prone behaviors than other clinical populations.32 This setting provided the opportunity for conducting a targeted intervention among an injury-prone population with low self-reported SBU.
In this pilot study, ED patients receiving a brief motivational intervention by a trained interventionist reported higher self-reported SBU at a 3-month follow-up compared to a control group. Although limited by self-report, an ED-based intervention may be useful to increase SBU among difficult-to-reach individuals.
The authors appreciate Ms. Kathy Shea for her administrative assistance in support of this project. They also thank Dr. Brian D. Johnston for his assistance in developing the measurement instrument. Finally, the authors thank Dr. Jonathan Olshaker, Dr. Holly Hackman, Ms. Carlene Pavlos, Ms. Heather Rothenberg, Ms. Sarah Hughes, Ms. Caroline Hymoff, and Ms. Cindy Rodgers for their helpful suggestions in the development and implementation of this project.