Running head: TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Tailored Messaging Feedback to Improve Parent Knowledge and Behavior Practices on Pediatric Drowning Prevention Jodi Riggs BSN, RN Arizona State University DNP 712 1 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 2 Tailored Messaging Feedback to Improve Parent Knowledge and Behavior Practices on Pediatric Drowning Prevention Abstract Introduction and Background: Drowning is the leading cause of preventable injury death in Arizona for children under five years old. Tailored education has demonstrated efficacy in behavior change and knowledge retention. The purpose of this evidence-based project was to evaluate if tailored education improved knowledge and self-reported behaviors related to pediatric drowning. The Elaboration Likelihood Model provided the framework for this project. Methods/Experimental Approach: The prospective pilot project was conducted using the Iowa Model of Evidence Based Practice. Parents with children under five years, presenting with low acuity complaints in a pediatric emergency department were approached. A baseline assessment identified high-risk behaviors and a custom education plan was delivered to parents. Outcome variables were measured at baseline and three weeks after initial assessment. Results: The average parent age was 29 (M = 28.5; SD = 6.35) years. Participant (n=29) responses were analyzed using descriptive statistics. Participants (n = 27, 93%) reported likelihood to change behaviors and 29 (100%) perceived the tailored intervention as relevant. Secondary outcome variables were not measured at three weeks due to a lack of survey response. Conclusions: Parents reported a high likelihood of behavior change when water safety education was tailored and relevant to their child. The tailored intervention evoked positive interaction and receptivity from parents and suggested a high motivation to make a behavior change. The effect of the intervention could not be tested due to the lack of follow-up and post data collection. The design of this evidence-based project is quantifiable and replicable in a low-acuity setting, which TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT allows for future evaluations of self-reported behavior change and knowledge improvement. Funding: No sponsorship or financial conflict of interest. Keywords: pediatric drowning, drowning, water safety, tailored education, supervision, tailored messaging, mobile-based, web- based, technology-based, education, parent knowledge, safety behavior 3 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 4 Drowning is a devastating and complex global public health issue. In Arizona, fatal drowning is the leading causes of preventable injury death for children ages one to four years old. With appropriate guidance, parental knowledge and behaviors, which increase water safety, can be improved. Tailored interventions have been shown to be an effective means of parent education. Using an assessment-based approach with tailored educational feedback is a promising intervention to improve parent knowledge and promote behavior change. Background & Significance Fatal drowning is defined as death from water submersion suffocation. Nonfatal drowning is defined as the recovery from a water submersion. Both are devastating and complex global public health issues (Al-Qurashi et al., 2017). The per capita drowning rate takes into account the mass population, which has found that children under five years old are among the highest risk age group for drowning (Arizona Child Fatality Review [ACFR], 2018; Centers for Disease Control [CDC], 2017; U. S. Department of Health and Human Services [USDHHS], 2012). In the state of Arizona, fatal drowning is considered a preventable injury death. Lack of supervision accounted for the majority of those preventable deaths in children (ACFR, 2018; HHS, 2012; Safe Kids, 2016). Supervision is defined as the act of watching over someone or something with attentiveness (Morronigello, Sandomierski, Zdzieborski, & McCollam, 2012). In 62% of drowning cases, adult supervision was compromised by distractions such as taking a phone call or responding to a text message (Safe Kids, 2016). Various elements affect adult supervision, including parent experience, style of parenting, and perceptions of child injury (Huynh, Demeter, Burke, & Upperman, 2017). Distracted behavior and inadequate supervision increase risk of preventable injuries, including drowning among children (Huynh et al., 2017). TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 5 Lack of understanding of developmental norms may increase drowning risk for children in this age group. Permissive parenting styles, attributing injuries as a normal variation of childhood and limited enforced safety rules results in an increase of childhood injuries (Huynh et al., 2017). Further, injury risk may be increased when parents are unable to predict a child’s behavior. Limited parental knowledge of child development results in unrealistic behavior expectations of the child (Kendrick et al., 2013). Education is needed to increase developmental knowledge among parents. Educational programming should account for a variety of components such as education level, maternal age, child age, and socioeconomic status. Parent education is also needed to increase knowledge about the use of life jackets to reduce drowning risk. When incidents involving boating deaths were examined, people who survived were more than two times more likely to have been wearing a life jacket (O’Connor & O’Connor, 2005). One of the main reasons life jackets are necessary for inexperienced swimmers is due to the lack of understanding of how even the slightest lapse in supervision increases the risk for drowning (Safe Kids, 2016). Finally, parent education about appropriate emergency response in response to a drowning is needed. When cardiac arrest related to drowning occurs, cardiopulmonary resuscitation provides the best chance of neurological recovery (Topjian et al., 2012). Emergency departments treat nearly nine million children for injuries annually. Fatal drowning and near drowning places a substantial burden on the health care system (USDHHS, 2012). All childhood injuries, including drowning, requiring hospitalization incur approximately $87 billion in medical and societal costs (USDHHS, 2012). An average of 4,700 children per year, younger than five years old, were treated in the emergency department for a non-fatal drowning between 2015 and 2017 (Consumer Product Safety Commission, 2018). A reported TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 6 96% of the non-fatal drowning victims treated in the emergency department took place in a swimming pool. Children younger than five years old comprised approximately 76% of the victims of a pool- related submersion fatality. These fatalities generally occurred during the summer months of May, June, July, and August (CPSC, 2018). In addition, children from ethnic minorities, such as Hispanic or African-American, experience a non-fatal drowning more than four times that of Caucasian children (Felton, Myers, Liu, & Winders-Davis, 2015). Lifetime medical and work loss costs associated with a fatal drowning are estimated at nearly $5.7 million and $1.2 billion respectively (HHS, 2012). Children requiring inpatient hospital services incurred costs of approximately $6,372 per episode compared to an ED visit at approximately $4,528 per episode (Felton et al., 2015). In the United States, the annual cost of caring for a drowning victim equates to $173 million (World Health Organization [WHO], 2018). This data demonstrate how financially devastating water incidents are for families and the healthcare system. Unfortunately, fatal drownings increased 30% from 2016 to 2017 and continue to occur despite a parent or family member supervising at swimming pools (ACFR, 2018; Nansel et al., 2002; O’Connor & O’Connor, 2005). Encouraging an adult to be close by while a child is swimming or near a pool is wise, but is an incomplete strategy to prevent drowning (Hossain, Mani, Sidik, Hayati, & Rahman, 2015; Morrongiello et al., 2012). The National Action Plan for Child Injury Prevention by the Centers for Disease Control and Prevention is committed to preventing childhood injury and drowning. Their goal is to raise awareness, highlight prevention solutions, and mobilize action (HHS, 2012). An Arizona pediatric hospital organization also has taken action by examining studies about drowning causation, directing conversations with parents, and participating in child fatality review teams (T. Isaacson, personal communication, October 2017). The Injury Prevention TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 7 Center is able to evaluate adverse childhood experiences (ACES), trauma, substance abuse, lack of understanding of developmental norms, and behavioral health issues as contributors to drowning risk. In addition, the team examines parental behaviors and parents view supervision or non-supervision as contributing to increased risk (T. Isaacson, personal communication, October 2017). With so many factors to consider, delivering concise and clear educational messaging is critical. Improved communication and education may better inform families to implement evidence-based injury prevention strategies into their daily routines (McDonald et al., 2005; Morrongiello et al., 2012; Nansel et al., 2002). The use of innovative media and educational technologies to examine parent knowledge and behaviors surrounding water safety is promising (Dijkstra, & De Vries, 1999). Using an assessment-based approach to target specific parental needs will help guide tailored education so that parents are more likely to engage in behavior change and continue that improvement long term (Morrongiello et al., 2012; Nansel et al., 2002). Purpose Statement The purpose of this evidence-based project was to evaluate if tailored education improves parent knowledge and self-reported behaviors related to drowning after identifying high-risk behaviors from a risk assessment. Problem Statement and PICO question This review evaluates how tailored education improves parent knowledge and selfreported behaviors related to drowning. The evidence search PICO question is as follows: For caregivers with children age one to four years old (P), how does a parent education program (I), compared to a community education program (C), improve caregiver knowledge (O). TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 8 Search Strategy The main aspects of the PICO question were searched in several databases including PubMed, the Cochrane Library, CINAHL, PsycInfo, and the Turning Research Into Practice (TRIP) databases (See Appendix G). The initial search strategy included the key words: parent or adult or caregiver (population) and community education or parent education or parent based or community based (interventions). The outcome of interest included parent knowledge or drowning prevention or child safety. Search limits were applied to include all literature from 2013 to 2018 in order to narrow to a more manageable search. The PubMed search initially produced 1,192 results, and this modified search produced 2,128 results in the Cochrane Library; 27,056 results in CINAHL; 23,004 results in PsycInfo; and 627 results in TRIP database. No search limits were placed in PubMed due to the moderately narrow search results generated by the key words used. Next, a combination of the key words were used in each database using the Boolean connector “and” and “or” with the year search limits of 2013 to 2018. In addition, the literature was filtered to obtain reviewed articles only. This moderately refined search yielded 7,054 results in PsycInfo; an insignificant reduction to 26,472 results in CINAHL; and 2,128 results in the Cochrane Library. The TRIP database was unable to filter the literature with a peer-reviewed option during the search. Specific key words were utilized in the five databases including: parent program and drowning or water safety or supervision. This detailed search yielded 13 results in PubMed; 79 results in PsycInfo; 64 results in CINAHL; and 11 results in the Cochrane Library. The TRIP database search limitations included the ability to select articles based on specific types of methodologies, which produced varying results. A secondary hand search was pursued in PubMed to identify specific types of education TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 9 interventions. A combination of search terms used were parent-delivered or parenting education intervention or mobile or parent-based intervention or parent training program or web-based or injury prevention program and parent knowledge or child safety or injury prevention. Search limits of 2013 to 2018 were applied to help narrow the search. These parameters led to four outlier articles that were identified through this hand search. Throughout the initial and secondary search process, non-English studies were excluded. After the completion of the exhaustive literature search, a total of 25 articles addressed at least two elements of the PICO question. Twenty articles resulted with the search terms web-based, mobile health technology and parent knowledge, and five articles were revealed with the search term parent-training programs. Fifteen studies were excluded from the critical appraisal due to insignificant findings, erroneous populations groups, lack of generalizability, and interventions unrelated to the desired outcome. The final yield resulted in 10 studies (See Appendix A). These studies included seven randomized controlled trials (RCTs), one non-random controlled trial, one quasi- experimental study and one systematic review (SR). The flow chart detailing the process of article search strategy and selection can be found in Appendix G. Review of Literature and Evidence Synthesis A study by Carlson-Gielen and colleagues (2007) was the first to use the Precaution Adoption Process Model (PAPM) to address parental safety behaviors, parental thoughts and beliefs and how these elements affected the persuasiveness of messaging. One additional study used the PAPM approach by utilizing staged algorithms to identify parents’ level of safety behaviors (Shields et al., 2013). Five of the studies reported a reliable and valid tool (McDonald et al., 2005; Nansel et al., 2007; Morrongiello et al., 2012; Gittelman et al., 2014; Glassman et al., 2017). These instruments included a 9-item scale to examine persuasiveness of an TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 10 intervention, survey questions using a Likert-scale, and the American for Society Quality (ASQ) scale, which examines six constructs of the Health Belief Model (HBM) (See Appendix B). Homogeneity was noted in the type of intervention used to deliver tailored messaging to parents. Technological-based interventions were used across all ten studies. Three of the studies used computerized tailored messaging (Carlson-Gielen et al., 2007, Nansel et al., 2002, & Shields et al., 2013) and two used kiosk-based messaging (Gittelman et al., 2014 & McDonald et al., 2005) but both offered parent feedback reports. Other delivery methods included web-based videos, social marketing campaigns, and video-based tailored messaging (VanBeelen et al., 2014, Glassman et al., 2017, & Morrongiello et al., 2012). The systematic review by Omaki and colleagues (2017) identified seven RCTs, which examined the impact of kiosk, computer and mobile technology-based tailored systems. The findings of this review were consistent with previous research and solidified the positive impact technological approaches offer to busy clinical practices. Each technological approach had the ability to tailor safety education to parents and provide immediate feedback in a time-efficient manner. Homogeneity was noted in the examination of safety knowledge scores and behavioral impact of the intervention on parents. This was evaluated in half of the studies (all citations). Two studies examined a combination of self-reported and observed behaviors (Omaki et al., 2017 & Shields et al., 2013). Other dependent variables (DV) included parental perception of safety behaviors, barriers to supervision, risk score assessment, and parent-provider communication (Glassman et al., 2017, Morrongiello et al., 2012, Nansel et al., 2002, & Nansel et al., 2007). (See Appendix B). Homogeneity was noted in the findings of the significant impact of improved parental knowledge and safety behaviors across all 10 studies that measured these variables. The interventions with a tailored approach resulted in significant findings (Nansel et al., 2002, TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 11 McDonald et al., 2005, Carlson-Gielen et al., 2007, Nansel et al., 2007, Morrongiello et al., 2012, Shields et al., 2013, & VanBeelen et al., 2014). Furthermore, parents offered tailored education, based on their needs and experiences, were more likely to engage in behavior change, retain improved safety knowledge, and be able to continue safer practices over a longer period of time. In seven of the studies, homogeneity was noted in the setting type (Carlson-Gielsen et al., 2007, Gittelman et al., 2014, McDonald et al., 2005, Nansel et al., 2002, Nansel et al., 2007, & Shields et al., 2013). These locations included either in an emergency department or pediatric clinic waiting room. Injury prevention education was well received by parents when their child or children presented to these settings with an injury complaint. All 10 studies used a combination of self-reported surveys, true/false questions, and multiple-choice questions to gain insight to parental safety behaviors and knowledge (See Appendix B). Evidence Conclusions Understanding risk factors associated with water safety is crucial for the improvement of behavior change and knowledge among parents with young children. Using an assessment-based approach, such as intervention tailoring, parents are able to provide personal behavioral information to receive the most appropriate type of messaging feedback specific to their needs. Parents are more likely to engage in behavior change and continue that improvement long term. Technology-based delivery systems can reach a large population, provide a single platform for providers to tailor water safety information for parents, and compliment verbal instructions given during healthcare visits. A tailored approach can be used in an outpatient setting to address the elements of parental water safety knowledge and behavior, and provide immediate feedback. These elements were consistently recommended across all studies. Conceptual & Theoretical Model The Elaboration Likelihood Model (ELM), which was developed by Richard Petty and TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 12 John Cacioppo (Petty & Cacioppo, 1986). This theory states that a person is more likely to process and consider information that is relevant to their needs (Petty & Cacioppo, 1986) (See Appendix H). Elaboration is defined as the cognitive act of examining a potentially persuasive argument or opinion. The model proposes that people are active thinkers and processors (Petty & Cacioppo, 1986). The model was designed to explain how communication persuades behavior and attitude changes. The effect of persuasion plays a key role in how people view various issues and objects. Two types of persuasion processes are dependent upon the degree of elaboration. Both routes involve systematic thinking and other cognitive shortcuts. The central route, which comes from information received that is relevant to personal experiences, occurs when elaboration is high. Parents will use this route to critically analyze the merit of the information received about their drowning risk behaviors and will likely respond in a positive way, regardless of previous knowledge and beliefs (O’Keefe, 2008). The peripheral route relies on outlier sources to guide attitude and belief instead of information processing. Receiving information from an informal source negates the need to critically analyze what is being communicated. Parents who accept advice from family and friends about drowning risk behaviors employ a superficial method of behavior change, without any merit to the information received (O’Keefe, 2008). Personalized knowledge and beliefs from parents can be customized into tailored messaging to promote behavioral change based on individual needs and experiences (McDonald et al., 2005). In this project, persuasive, relevant messaging can impact how parents view their child’s risk of drowning and lead to a high likelihood to engage in behavior change. Using the ELM approach has been shown to result in a permanent attitude change (McDonald et al., 2005; O’Keefe, 2008). Therefore, stimulated cognitive activity can lead to overall improved safety TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 13 knowledge concerning drowning and a decrease in high-risk behaviors. The primary outcome of interest in this evidence-based project was the perceived relevance of tailored education and the likelihood to make a behavior change after the tailored intervention. The plan was to survey participants with two questions about the outcomes following the tailored discussion with the handouts. Additional outcomes of interest, including knowledge retention, usefulness of tailored education, and self-reported behavior change demonstrated by a decrease in risk score from baseline to follow-up also were planned Evidence Based Practice Model The Iowa Model of Evidence-Based Practice to Promote Quality Care (Titler, 1994) was chosen to guide this project planning and implementation (Appendix I). The Iowa Model is widely used and focuses on problem-solving steps to help clinicians identify a clinical problem or knowledge gap (Titler, Steelman, Budreau, Buckwalter, & Goode, 2001). The step-wise approach includes identifying triggers, clinical application, organizational priorities, forming a team, piloting a practice change, evaluating the pilot, and evaluating practice changes and dissemination of the results (Melnyk & Fineout-Overholt, 2011; Titler et al., 2001). This model emphasizes scientific knowledge about the problem, and formulates a multidisciplinary approach to identify and address a high priority issue among the organization. Robust evidence suggests lack of supervision is among the top causes of pediatric drowning. Children younger than five years of age are most vulnerable. Parental lack of knowledge and perceptions of water safety and dangers factor into undesired outcomes. The Iowa Model helped identify the problem and offered a realistic approach to design and implement an evidence-based project focused on tailored drowning education. The Iowa Model for this project was used in the implementation process in the pediatric TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 14 emergency department setting. The use of the model allowed key stakeholders to evaluate a pilot practice change in order to determine its usefulness in a setting with families presenting to the emergency department with low-acuity complaints. Project team members were carefully selected and included the project team leader, principal investigators, and clinical manager. Methods Privacy Institutional Review Board approval was obtained after an expedited review process through the children’s hospital system and Arizona State University. Private health information, including personal emails, was collected for this evidence-based project. All participant information was de-identified and assigned a participant identification number (ID). The survey responses remained anonymous. Names or any other specific identifiers were not be required, nor obtained. Data was exported from the Research Electronic Data Capture (REDCap) system to the SPSS® software, version 25 from the tablet by the project leader. The de-identified data obtained from the surveys was securely stored in an SPSS file on the project leader’s personal computer which is password protected. The REDCap system is a secure, web-based application designed to support data capture for research studies providing 1) an intuitive interface for validated data entry, 2) audit trails for tracking data manipulation and export procedures, 3) automated export procedures for seamless data downloads to common statistical packages, and 4) procedures for importing data from external sources (Harris, Taylor, Thielke, Payne, Gonzalez, & Conde, 2009). The baseline survey link was generated from the REDCap system and contained a singular URL to the survey, which did not track IP addresses, e-mail addresses, or any other identifying information. The follow-up survey link that was e-mailed to the participant contained a unique URL that was only TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 15 valid for one submission via the REDCap system. After discussion, consent forms were signed prior to beginning the survey to protect private information due to the collection of personal email addresses. Consent forms were stored in a file locked cabinet in the Center for Family Health & Safety office at Phoenix Children’s Hospital and will be shredded after project results are disseminated. Project Design This evidence based project was conducted to examine the effectiveness of tailored education for promoting knowledge retention and self-reported behavior change related to drowning prevention by parents with children less than five years old. Outcome measures of interest included likelihood of behavior change and parental views of relevance of tailored education immediately following the tailored intervention. Additional planned outcome measures of interest included usefulness of the tailored education, knowledge retention, and self-reported behavior change. Parent knowledge and self-reported behaviors related to water safety are identified based on a calculated score from the parent assessment survey (Mangione, Imre, Chow, Linsiski, & Heitz, 2015; McDonald et al., 2005; Morrongiello e al., 2012). A summary page with risk scores in five areas related to drowning prevention was provided, as well as two tailored education handouts to parents. Participants were approached in the patient exam room during their visit. Participants who satisfied the inclusion criteria were invited to participate in the pilot project after a description of the project was read to the participant. The participants were asked to take a baseline parent assessment risk survey and engage in a brief discussion with provided tailored education handouts. A follow-up parent assessment risk survey was emailed to participants three weeks after the baseline survey was complete the REDCap automated survey reminder system, TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 16 with two attempts/reminders sent one day apart to remind parents to complete the follow-up survey. Inclusion and Exclusion Criteria Participants included parents or legal guardians of children under five years old presenting to the Pediatric Emergency Department (PED) with a low acuity medical or injury complaint, at local children’s hospital in Phoenix, Arizona, Parents were English speaking. Participants with a child presenting with a high-acuity medical or injury complaint, including children requiring urgent or immediate medical care, as well as healthcare needs requiring lifesaving treatment by emergency healthcare providers and staff were excluded. In addition, participants with children who were five years of age and older and non-English speaking were also excluded. Data Collection Participants, who satisfied the inclusion criteria, were invited to participate in evidencebased project after the project leader read a description of the project to the participant. Participants recruited completed the following forms: (a) consent form and (b) Health Insurance Portability and Accountability Act of 1996 (HIPAA) authorization form. After the senior injury prevention specialist obtained consent prior to the participants beginning the survey on the electronic tablet. Demographic questions included questions about gender, relationship to child, age, level of education, employment status, race/ethnicity, and age of the child. The participants entered a working email address in a free-text box and were required to supply this information before they could continue with the survey. Baseline self-reported behaviors and knowledge assessment were measured using a 20item risk assessment survey. The risk assessment survey was developed by the project leader TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 17 and reviewed by three field experts for content validity (see Appendix J). The survey was based on the Appraisal Statements Questionnaire (ASQ). Previous Cronbach alphas for the 3- and 4item statements ranged from .82 to .85. This survey evaluated parents’ self-reported behaviors and knowledge in five categories related to drowning risks, with four questions per category (Morrongiello et al., 2012). A Likert scale was used to evaluate survey responses and generate a risk score (Vagias, 2006). The baseline risk assessment survey addressed five categories related to drowning risk behaviors including supervision practices, emergency preparedness, developmental milestones, life jackets, and pool parties. The survey took approximately 10-15 minutes for the participants to finish. Each response had an assigned point value from one to five. A maximum value of 20 was set for each of the five drowning risk categories, with a total possible risk score of 100. The highest risk was identified by the highest numerical value (McDonald et al., 2005; Morrongiello et al., 2012; Nansel et al., 2002). The participant was given a summary page, which included a score for each category of drowning risk and a total risk score (Appendix K). The top two categories of drowning risk were identified and discussed with each participant using two tailored educational handouts based on self-reported behaviors. The perceived relevance of the tailored intervention and likelihood of behavior change were measured following the tailored discussion by a post intervention survey. A link to the follow-up risk assessment survey was emailed to the participants three weeks later. REDCap sent this survey link through the automated survey reminder system. A total of three reminders one day apart, were made to collect follow-up data and measure the additional outcome variables. At the completion of each interaction with the participants, the participants were given a neon-colored index card with information about the follow-up survey including the TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 18 email address they would receive the email from and the date they could expect to receive the email. Data Analysis Descriptive statistics were used to describe the sample and outcome variables. Inferential statistics was used to analyze the data. A two-tailed test was used and the critical value was set at p < 0.05. Further analyses were conducted using the Spearman’s Correlation to examine the relationships between the risk scores in each of the five categories, the total risk score, parent age, education, employment, and the total number of children. In addition, an independent t-test was used to determine if the mean total risk score was significantly different between parent age, education, employment, and the total number of children. Results Demographics The foundational studies for the development of this evidence-based project had a mean attrition rate of 18% (McDonald et al., 2005; Morrongiello et al., 2012; Nansel et al., 2002). The number of participants was considered based on this attrition rate. We anticipated recruiting 36 participants. A total of 29 participants agreed to complete the baseline risk assessment survey and participate in the tailored intervention. The average age of the sample was 29 years (M = 28.5; SD = 6.35) with the age range from 17 to 44 years of age. Twenty-six (88.7%) were the parent of a child under five years old; 25 (86.2%) of the participants were female; and 4 (13.8%) were male. Fifteen (51.7%) had a high school diploma or less education, 12 (41.4%) had some college or tech school training, and 2 (6.9%) were college graduates. Sixteen (55.2%) were employed 40 or more hours per week, 5 (17.2%) worked 39 hours or less, and 8 (27.6%) were unemployed. Nineteen (65.5%) of the participants were Hispanic, and 10 (34.5%) were non- TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 19 Hispanic. Fifteen (51.7%) reported having two children or less and 14 (48.2%) had three or more children. Demographic information is displayed in Appendix C. Outcome variables The project protocol and outcome measures guided the project team to collect baseline data on the relevance of the tailored intervention and the likelihood to make a behavior change. Data analysis was used to determine how tailored messaging was perceived and appropriate in this setting but due to a lack of follow-up response the effectiveness of the tailored intervention was not measured. Twenty-nine (100%) of the participants reported that the tailored intervention was relevant to their child. Twenty-seven (93%) reported a likelihood of behavior change when water safety education was tailored to their child (Appendix L). Knowledge retention and selfreported behavior change were not measured due to a lack of follow-up participant response. A moderate positive correlation exists between total number of children and milestones risk score, (r s (29) = 0.46 , p = 0.01). No other significant correlations exist between the risk scores and parent demographics. Correlation results are displayed in Table 1 (Appendix D). Life jacket use (M = 9.3, SD = 2.7) and emergency preparedness (M = 9.6, SD = 2.9) were identified as the top two categories with the highest risk score amongst the participants. Categories of drowning risk results are demonstrated in Figure 1 (Appendix E). An independent t-test was conducted to compare the total risk score between parents less than 25 years old (M = 43.6, SD = 11.6) and parents 25 years and older (M = 41.8, SD = 6.1), t(7), p = 0.70. An independent t-test was conducted to compare the total risk score and parents with two children or less (M = 41.5, SD = 8.8) and parents with three or more children (M = 42.9 , SD = 6.0 ), t(27), p = 0.63. An independent t-test was conducted to compare the total risk score and parents with less than a college degree (M = 42.6 , SD = 7.6) and parents who graduated TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 20 college with at least a bachelor’s degree (M = 37.5, SD = 3.5), t(27), p = 0.37. An independent ttest was conducted to compare the total risk score and unemployed parents (M = 41, SD = 9.2) and employed parents (M = 42.7 , SD = 6.9), t(27), p = 0.60. The t-test results are displayed in Appendix F. Discussion The purpose of the pilot project was to examine how parents can be persuaded to engage in behavior changes to reduce their child’s drowning risk by using an assessment-based approach and providing tailored education. This evidence-based project gave parents an opportunity to self-report knowledge and behavior practices surrounding water safety. By using an assessmentbased approach and identifying areas that increase a child’s risk for drowning, the tailored intervention evoked a positive interaction and receptivity from parents. The results suggest parents are more likely to make a behavior change when water safety education is tailored and perceived as relevant to their child. Clinical populations vary considerably in terms of knowledge, behavior practices, and learning needs, and tailored education targets specific needs and exhibits effectiveness by offering a high level of perceived relevance to the recipient (Nansel et al, 2002). The customized education was highly focused. Parents received information on the top two categories of drowning risk based on their responses. They engaged in an open discussion with the author about simple ways to keep their child safer at the pool. However, the effect of the intervention could not be fully tested due to the lack of follow-up and post data collection. Therefore, behavior change and knowledge improvement was not measured. The email approach for follow-up was a poor communication method for this convenience sample. It is undetermined why participants did not complete the post survey when they received it by email. TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 21 Life jacket use and emergency preparedness were identified as the top two categories with the highest risk core. These results suggest that there is a lack of understanding about the use of a lifejacket and what do in a drowning emergency. Certain life jackets are designed to prevent a drowning by keeping a child’s head and face out of the water and in a vertical position. Life jackets should be properly fitted and continually sized as the child grows (USCG, 2019). Bystander cardiopulmonary resuscitation (CPR) can prevent a fatal drowning with proper technique and ventilation. Topjian and colleagues (2012) discuss drowning survival and how crucial effective CPR can be to prevent neurological deficits and ultimately, death. No significant difference was found between parent demographics and total risk score. However, parents with an education level less than a college degree had a total risk score of 5.1 points higher than parents with at least a college degree. Low socioeconomic status has been found to be associated with parental unsafe injury-preventable behaviors, as well as undesired safety knowledge and beliefs (McDonald et al., 2005). The results in this project may be attributable to the parental knowledge gap concerning life jacket use and emergency preparedness. Practice Implications and Future Research Findings from this evidence-based project suggest using an assessment-based approach to tailor water safety information for parents with young children. However, due to the lack of post intervention data, it is unclear if a behavior change or knowledge retention took place. The design of this evidence-based project is quantifiable and replicable in a low-acuity setting. Phone calls and text messaging may be a more effective option to collect follow-up data. In the future, self-reported behavior change could be measured by a decrease in the total risk score from baseline to follow-up. Knowledge improvement could be measured by the self-reporting of TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 22 sharing the tailored education with others. Replication of this evidence-based project and intervention is realistic to further evaluate self-reported behavior change and knowledge improvement. Safety team members may consider implementing this intervention during the summer months, with a larger sample size, and utilizing an alternate method to collect follow-up data. Further, offering the risk assessment survey and tailored education in languages other than English may have a positive impact on the sample size, ability to collect post intervention data, and contribute to the generalizability of the results. Limitations There were several limitations to this evidence-based project, which are important to discuss for future research. Participants were asked to self-report behaviors and knowledge about water safety, as opposed to recording direct observations of their behaviors around a swimming pool. However, a majority of the behaviors were assessed by a level of appraisal of certain behavior practices and perceived danger of a given situation which only be measured by self-reporting. This project included a small convenience sample from a pediatric emergency room setting. While participating families presented with low acuity complaints and had no impingement of the care they received, it is unclear how their current situation factored into selfreporting behaviors. The exclusion criteria, which eliminated non-English speaking parents from recruitment, was also a limitation to this project because a large part of the population seen in this pediatric emergency room were Spanish-speaking. Lastly, a large limitation was the lack of follow-up response by use of emailing surveys. The effect of the intervention was not determined due to lack of post intervention survey data collection. Conclusion This evidence-based project demonstrates that tailored water safety education is a TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 23 promising way to educate and communicate with families about water safety. Parents reported a high likelihood of behavior change when water safety education was tailored and relevant to their child. The tailored intervention evoked positive interaction and receptivity from parents and suggested a high motivation to make a behavior change. However, the effect of the intervention could not be tested due to the lack of participant response to the follow up email survey. The design of this evidence-based project is quantifiable and replicable in a low-acuity setting. Future delivery of the intervention should use an alternative method to measure selfreported behavior change and knowledge improvement. TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 24 References Al-Qurashi, F. O., Yousef, A. A., Aljoudi, A., Alzahrani, S. M., Al-Jawder, N. Y., Al-Ahmar, A. K.,Al-Majed, M. S., & Abouollo, H. M. (2017). A review of nonfatal drowning in the pediatric-age group: A 10-year experience at a university hospital in Saudi Arabia. Pediatric Emergency Care, 00(00), 1-5. doi:10.1097/PEC.0000000000001232 Arizona Child Fatality Review Program (2018). Twenty-fifth annual report. Retrieved from: https://www.azdhs.gov/documents/prevention/womens-childrens-health/reports-factsheets/child-fatality-review-annual-reports/cfr-annual-report-2018.pdf Carlson-Gielen, A., McKenzie, L. B., McDonald, E. M., Shields, W. C., Wang, M-C., Cheng, Y-J., Weaver, N. L. & Walker, A. R. (2007). Using a computer kiosk to promote child safety: Results of a randomized controlled trial in an urban pediatric emergency department. Pediatrics, 120(2), 330-339. doi: 10.1542/peds.2006-2703 Centers for Disease Control and Prevention. (2017). Web based injury statistics query and reporting system (WISQRS) [Online]. Retrieved from: https://www.cdc.gov/ncipc/wisqrs Consumer Product Safety Commission (2018). Pool or spa submersion: Estimated non-fatal drowning injuries and reported drownings, 2018 report. Retrieved from: https://www.cpsc.gov/s3fs-public/Pool_Spa_Submersion_Estimated_nonfatal_drowning _injuries_and_reported_drownings_2018_report_2018H002.pdf?hak4TSK.2PVleeVQeM kFJkJ0BSnoSP4Z Dijkstra, A. & De Vries, H. (1999). The development of computer-generated tailored interventions. Patient Education and Counseling, 36, 193-203. doi: 10.1016/S0783 -399(98)00135-9 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 25 Felton, H., Myers, J., Liu, G., & Winders Davis, D. (2015). Unintentional, non-fatal drowning of children: US trends and racial/ethnic disparities. BMJ Open Journal, 5, e008444. doi:10.1136/bmjopen-2015-008444 Gittelman, M. A., Pomerantz, W. J., McClanahan, N., Damon, A. & Ho, M. (2014). A computerized kiosk to teach injury prevention: Is it as effective as human interaction?. Journal of Trauma Acute Care Surgery, 77(3), S2-S7. doi: 10.1097/TA .0000000000000317 Glassman, T.J., Castor, T., Karmakar, M., Blavos, A., Dagenhard, P., Domigan, J…., Kucharewski, R. (2017). A social marketing intervention to prevent drowning among inner-city youth. Health Promotion Practice, 19(2), 175-183. doi: 10.1177 /152483991773255 Harris, P. A. Taylor, R., Thielke, R., Payne, J., Gonzalez, N. & Conde, J. G. (2009). Research electronic data capture (REDCap): A metadata-driven methodology and workflow process for providing translational research informatics support, Journal of Biomedical Informatics, 42(2), 377-81. doi: 10.1016/j.jbi.2008.08.010 Hossain, M., Mani, K. K. C., Sidik, S. M., Hayati, K. S., & Rahman, A. K. M. (2015). Randomizedcontrolled trial on drowning prevention for parents with children aged below five years in Bangladesh: A study protocol. BMC Public Health, 15, 1-7. doi:10.1186/s12889-015-1823-1 Huynh, H. T., Demeter, N. E., Burke, R. V., & Upperman, J.S. (2017). The role of adult perceptions and supervision behavior in preventing child injury. Journal of Community Health, 42, 649-655. doi:10.1007/s10900-016-0300-9 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 26 Kendrick, D., Mulvaney, C. A., Ye, L., Stevens, T., Mytton, J. A., & Stewart-Brown, S. (2013). Parenting interventions for the prevention of unintentional injuries in childhood. Cochrane Database of Systematic Reviews, 3, 1-73. doi: 10.1002/14651858.CD006020.pub3 Mangione, T. W., Imre, M., Chow, W., Linsinski, H. E., & Heitz, E. (2015). 2014 Life jacket wear rate observation study. Retrieved from: https://www.uscgboating.org/library /national-live-jacket-wear-study/2014%20CORE%20Report.pdf McDonald, E. M., Solomon, B., Shields, W., Serwint, J. R., Jacobsen, H., Weaver,N. L.,…Gielen, A. C. (2005). Evaluation of kiosk-based tailoring to promote household safety behaviors in an urban pediatric primary care practice. Patient Education and Counseling, 58, 168181. doi: 10. 1016/j.pec.2004.08.015 Melnyk, B. M., & Fineout-Overholt, E. (2015). Evidence-based practice in nursing and healthcare: A guide to best practice (3rd ed.). Lippincott, Williams & Wilkins. ISBN-13: 978-1451190946, ISBN-10: 1451190948 Morrongiello, B. A., Sandomierski, M., Zdieborski, D., & McCollam, H. (2012). A randomized controlled trial evaluating the impact of the supervising for home safety program on parent appraisals of injury risk and need to actively supervise. Health Psychology, 31(5), 601611. doi:10.1037/aoo28214 Nansel, T. R., Weaver, N., Donlin, M., Jacobsen, H., Kreuter, M. W. & Simons-Morton, B. (2002). Baby, Be Safe: The effect of tailored communications for pediatric injury prevention provided in a primary care setting. Patient Education and Counseling, 46(3), 175-190. doi: 10.1016/S0738-399(01)00211-7 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 27 Nansel, T. R., Weaver, N. L., Jacobsen, H. A., Glasheen, C., & Kreuter, M. W. (2007). Preventing unintentional pediatric injuries: A tailored intervention for parents and providers. Health Education Research, 23(4), 656-669. doi: 10.1093/her/cym041 O’Connor, P., & O’Connor, N. (2005). Causes and prevention of boating fatalities. Accident analysis and prevention, 37(4), 689-698. doi: 10.1016/j.aap.2005.03.011 O’Keefe, D. J. (2008). Elaboration likelihood model. In W. Donsbach (Ed.), The International Encyclopedia of Communication. (pp. 1475-1480). Malden, MA: Blackwell Publishing Ltd. Retrieved from: http://www.communicationcache.com/uploads/1/0/8/8/10887248 /elaboration_likelihood_model_-_okeefe.pdf Omaki, E., Rizzutti, N., Shields W. Zhu, J., McDonald, E., Stevens, M. W., & Gielen, A. (2017). A systematic review of technology-based interventions for unintentional injury prevention education and behavior change. Injury Prevention, 23, 138-146. doi: 10.1136/injuryprev- 2015041740 Petty, R. E. & Cacioppo, J. T. (1986). Elaboration likelihood model of persuasion. Advances in Experimental Social Psychology, 19, 124-181. doi: 10.1016/S0065-2601(08)60214-2 Reavy, K. (2016). Inquiry and leadership: A resource for the DNP project. Philadelphia, PA: F.A. Davis. Safe Kids World Wide (2016). Keeping kids safe in and around water: Exploring misconceptions that lead to drowning. Retrieved from: https://www.safekids.org/sites/default/files /water_safety_study_2016.pdf TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 28 Shields, W., McDonald, E., McKenzie, L., Wang, M-C., Walker, A.R. & Gielen, A. (2013). Utilizing the pediatric emergency department to deliver tailored safety messages: Results of a randomized controlled trial. Pediatric Emergency Care, 29(5), 628-634. doi: 10.1097/PEC.0b013e31828e9cd2 Titler, M. G., Steelman, V. J., Budreau, G., Buckwalter, K. C., & Goode, C. J. (2001). The Iowa model of evidence-based practice to promote quality care. Critical Care Nursing Clinics of North America, 13(4), 497-509. doi: 10.1016/S0899-5885(18)30017-0 Topjian, A. A., Berg, R. A., Bierens, J. J., Branche, C. M., Clark, R. S., Friberg, H., … Warner, D. S. (2012). Brain resuscitation in the drowning victim. Neurocritical care, 17(3), 441– 467. doi:10.1007/s12028-012-9747-4 United States Coast Guard Boating. (2019). Life Jacket Wear: Wearing your life jacket. Retrieved from: http://www.uscgboating.org/recreational-boaters/life-jacket-wearwearing-your-life-jacket.php U.S. Department of Health and Human Services, Centers for Disease Control and Prevention & National Center for Injury Prevention and Control. (2012). National action plan for child injury prevention: An agenda to prevent injuries and promote the safety of children and adolescents in the United States. Retrieved from: https://www.cdc.gov/safechild/nap/ Vagias, W. M. (2006). Likert-type scale response anchors. Clemson International Institute for Tourism & Research Development, Department of Parks, Recreation and Tourism Management. Clemson University. TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 29 Van Beelen, M. E. J., Beirens, T. M. J., Strujik, M. K., den Hertog, P., van Beeck, E. F., & Raat, H. (2014). Effectiveness of web-based tailored advice on parents’ child safety behaviors: Randomized controlled trial. Journal of Medical Internet Research, 16(1), e17. doi: 10.2196/jmir.2521 World Health Organization (2018). Drowning. Retrieved from: https://www.who.int/newsroom/fact-sheets/detail/drowning TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Citation Nansel, T.R. (2002). Baby, Be Safe: The effect of tailored communications for pediatric injury prevention provided in a primary care setting. Country: United States (Washington, DC) Funding: not reported Bias: sample bias Theory/ Conceptual Framework Health Belief Model Design/ Method/ Purpose Design: twogroup RCT Purpose: to determine effectiveness of tailored communication s for promoting adoption of child injury prevention measures by parents Sample/Setting Major Variables Studied N = 174 IG: n = 85 (tailored) CG: n = 89 (generic) IV: computerized tailored messaging Demographics: Respondents M, Ethnicity AA: 84.7% M % Mothers: 76.3% DV: FU injury risk score Setting: pediatric clinic Inclusion: parents of children 6-20 months of age at time of well visit 30 Measurement of data Overall risk score based on 5-itempsychosocial construct for injury-related locus of control α = 0.58 ↓ valid/reliable Time frame– Assessment questions 10-15 minutes FU phone assessment 3 weeks post intervention Attrition: 18.3% (lost to FU) Appendix A Evaluation Table Table 1 Final yield from literature search. Data analysis ANCOVA (↑ sensitivity of main effects) Findings DV IG: M ↓ in score: 4.68 SD = 6.44 T = -3.45 df = 1 p = 0.01 (↓risk score for IG) Decision to use in practice LOE: II Strength: RCT design; time-efficient; feasible; ease of development into practice; moderate attrition rate Weakness: self-reporting; convenience sample; primarily African-American parents; sampling limitation; many disconnected phone numbers; limited validity/reliability for measurement tool Conclusion: tailored messaging is an effective way to provide injury prevention education to parents of young children Clinical significance: cost efficient, tailored education program for parents to discuss injury prevention behaviors with their provider, which is easily integrated within regular patient flow in the primary care setting TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Citation McDonald, E.M. (2005). Evaluation of kiosk-based tailoring to promote household safety behaviors in an urban pediatric primary care practice Country: United States Funding: National Center for Injury Prevention and Control Bias: possible social desirability bias at FU Theory/ Conceptual Framework Elaboration Likelihood Model 31 Design/ Method/ Purpose Sample/Setting Major Variables Studied Measurement of data Data analysis Findings Design: 2-group RCT N = 144 IG: n = 70 (tailored feedback report) CG: n = 74 (no feedback) IV: kiosk-based tailoring intervention Risk index calculation (r/v tool for risk behavior) X2 or Fisher’s exact DV1 IG, CG Front seat Car safety: 5%, 16% p < 0.05 No stair gate needed: 3%, 14% p < 0.05 Ipecac for poisons: 57%, 39% p < 0.05 (↑ knowledge on safety items in IG) DV2 IG, CG Believe injuries r/t supervision: 73%, 93% p < 0.05 child belief of own safety: 25%, 45% p < 0.05 teach to obey parent for IP: 64%, 86% p < 0.05 (↑ + IP beliefs in IG) DV3 IG: 51 (20-90%) CG: 44 (11-70%) p = 0.01 (↑ % correct in IG) Purpose: describe development and feasibility of implementing computertailored injury prevention intervention in busy urban primary care practice; report results of program’s impact on parent’ home and child passenger safety knowledge, beliefs, and behavior Demographics: MA, mo. (child) 9.38 MA, y (mother) 26 Mother education HS: 74% E, AA children: 94% Setting: urban hospital-based academic primary care practice Inclusion: parents of children between ages 6wks and 24 mo. Under care of participating MD, English-speaking, living in same home as child Attrition: 16% (lost to FU) DV1: safety knowledge DV2: prevention beliefs DV3: safety behaviors (overall risk score) Time frame of the intervention14- minutes to complete assessment FU phone assessment 4 weeks post intervention t-test Decision to use in practice LOE: II Strength: RCT design; provides evidence to support future efforts of computer tailoring; moderate attrition rate Weakness: small pilot study; self reporting; “one-shot” intervention (one visit); baseline assessment may have influenced subsequent behavior; validity/reliability of tools not reported Conclusion: the kiosk intervention resulted in more safety knowledge, positive prevention beliefs, and selected safety behaviors. Overall risk score in IG was significantly better. Clinical significance: tailored communication is widely accepted by parents and providers and can be incorporated into a busy clinical setting. TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 32 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Citation Carlson-Gielen, A. (2007). Using a computer kiosk to promote child safety: Results of a randomized, controlled trial in an urban pediatric emergency department Country: United States Funding: National Institute of Child Health and Human Development; subcontract to the Health Communication Research Laboratory at St. Louis University Bias: possible overestimation of safety behaviors Theory/ Conceptual Framework Elaboration Likelihood Model & Precaution Adoption Process Model Design/ Method/ Purpose Design: RCT Purpose: to evaluate theorybased, computertailored intervention designed to promote parents’ car seat, smoke alarm, and poison storage safety knowledge and behaviors Sample/Setting N = 901 IG: n = 448 CG: n = 453 Demographics: Child’s age, y 1-2: 42% Gender m/f = 50% Respondents Mother’s: 90% 20-29: 55.3% AA: 93% HS degree: 75% PCI: < $5K: 63% Setting: ED of level 1 pediatric trauma center in Baltimore City Inclusion: parents of children between 4 and 66 months c/o injury or medical issue, English-speaking, living in Baltimore City, living with child at least halftime. parent/guardian Exclusion: parent of child whose visit was suspicious of child abuse or neglect, critically ill or injured Attrition: 16% (lost to FU) Major Variables Studied 33 Measurement of data Data analysis Findings IV1: computer kiosk tailored report IV2: per capita income Multiple choice; T/F; interview questions Sample t- test DV1: knowledge outcomes DV2: behavioral outcomes for LE & HE 1 CSS use 2 SA use 3 PS DV3: Parent Anxiety Staging algorithms (r/v not tested but recommended use in predicting behavior stages) IV1 & DV1: answered correct, M ± SD IG: 72.6 ± 13.9 CG: 66.4 ± 14.8 t = 5.87 p = .000 Time frame of the intervention12-minute assessment FU phone assessment one month post intervention State- Trait Anxiety Inventory (α = 0.86) Ordinal and logistic regression analyses- for IG to compare behavioral outcomes & per capita income as IV IV1 & DV2: OR (95% CI),p EA for IG LE 1 1.15 (0.85-1.54 p > 0.05 2 0.95 (0.63-1.42) p > .05 3 0.77 (0.54-1.11) p > .05 HE 1.70 (1.20-2.41) p = .003 2.07 (1.16-3.69) p = .01 2.01 (1.27-3.16) p = .003 IV2 & DV2: (IG) LE,LI 1 0.96 (0.66-1.39) p > .05 2 0.72 (0.43-1.21) p > .05 3 0.91 (0.57-1.44) p > .05 LE,HI 1 2.09 (1.28-3.40) p = .003 2 1.27 (0.63-2.59) p > .05 3 0.72 (0.41-1.28) p > .05 HE,LI Decision to use in practice LOE: II Strength: focus on preschool-aged population; use of stage-based behavior change theory; application of PAPM to child safety and computer-tailored interventions; RCT design; reasonable cross- section of families; large sample of families; same time spent on computer kiosk in both groups Weakness: self-reporting; majority of mother respondents Conclusion: improvements in safety knowledge, short term behavior outcomes, and correct safety reporting behaviors; injury prevention programs in ED setting can be offered to wide spectrum of families; financial barriers for low-income families are a priority for correct safety behaviors Clinical significance: results showed minimal intrusion to patient flow or time with provider and computer technology for patient education can be applied to busy EDs and other clinical settings. TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 34 1 1.08 (0.68-1.72) p > .05 2 1.80 (0.83-3.91) p > .05 3 2.70 (1.42-5.11) p = .002 HE,HI 1 3.28 (1.94-5.54) p = .000 2 1.90 (0.81-4.48) p > .05 3 1.81 (0.94-3.45) p > .05 DV3: M ± SD IG: 34.94 ± 10.25 CG: 34.96 ± 11.35 p > .05 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Citation Nansel, T.R. (2007). Preventing unintentional pediatric injuries: A tailored intervention for parents and provider Country: United States Funding: Intramural Research Program of the National Institutes of Health, National Institute of Child Health and Human Development Bias: potential social desirability reporting bias among providers or parents Theory/ Conceptual Framework Health Belief Model Design/ Method/ Purpose Design: controlled trial w/o random Purpose: to determine efficacy of delivering T-IPI to parents and concurrent T-IPI to parents and physicians on adoption of safety practice Sample/Setting N = 594 IG1: n = 192 IG2: n = 221 CG: n = 188 (G-IPI) Demographics: M % age range, y (parents) 21-25: 35.6% M % Home owners: 29.8% M % education College: 41.9% HS: 44.5% M % Ethnicity Non-AA: 10% AA: 58.7% M % Income < $10k: 41.2% Setting: Midwestern pediatric clinic Inclusion: parents of children age 4 and younger attending wellchild visit Attrition: 49% (lost to FU) Major Variables Studied IV1: T-IPI only IV2: T-IPI + P IV3: education level DV1: provider-parent communication DV2: parental adoption of injury prevention behaviors DV3: participant reactions to information Time frame of the interventionFU phone assessment one month post intervention 35 Measurement of data Data analysis Computer kiosk assessment questions Descriptive statistics, multinomial logistic regression, chi-square analyses, Nine-item scale- assess persuasiveness of intervention (α = 0.97) ANOVA Findings Decision to use in practice OR (95% CI) LOE: III IV1 & DV1 n = 34.6% 1.2 (0.6-2.1) p > 0.05 IV2 & DV1 n = 28% 0.9 (0.5-1.6) p > 0.05 IV1 & DV2 n = 48.6% 2.0 (1.1-3.6) p < 0.05 IV2 & DV2 n = 45% 1.9 (1.0-3.4) p < 0.05 Strength: lower socioeconomic status sample IV1,IV2 & DV3 (persuasiveness rating) M = 8.4, 8.5; F = 1.31 df = 2 p = 0.27 IV3 & DV2 < HS & HS = T-IPI: p = 0.02 ( ↑ effectiveness c/w CG) Weakness: low FU response rate (overall decrease in response to phone surveys); reduced statistical power; large attrition rate Conclusion: tailored approach with targeted information may overcome impediments of generic-printed injury prevention information; addition of provider-directed tailored feedback did not enhance intervention effectiveness Clinical significance: potential enhancement of injury prevention anticipatory guidance to increase safety practices by parents with young children TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Citation Morrongiello, B.A. (2012). A randomized controlled trial evaluating the impact of the supervising for home safety program on parent appraisals of injury risk and need to actively supervise. Country: United States Funding: grants from the Social Science & Humanities Research Council of Canada and the United States Department of Health and Human Services, Centers for Disease Control, National Center fro Injury Prevention and Control Bias: sampling bias; selfreporting scores-post intervention Theory/ Conceptual Framework Health Belief Model Design/ Method/ Purpose Design: RCT Purpose: To evaluate the efficacy of the Supervising for Home Safety program for increasing parental injuryrisk appraisals and need to actively supervise 2through 5- year old children Sample/Setting N =163 IG – n = 81 CG- n = 82 Demographics: MA, mo. - child m: 44.79 f: 44.21 Mother education % college: 63.80% Father education % college: 66.25% Family PCI >$80k: 64.50% Setting: laboratory Inclusion: child 25years old, child/parent fluent in English, older sibling at least 2 years older than target child, mother agreed to be participating parent Exclusion: reported abnormal child development, twins, participation in pilot projects related to study Attrition: 6.4% (lost to FU) Major Variables Studied IV: tailored safety video DV1: ASQ scores for six subscales 1 Vulnerability 2 Severity 3 Preventability 4 Supervision 5 Self-efficacy 6 knowledge DV2: barriers to supervision Time frame of the intervention – 20-minute video FU assessment 4 months post intervention 36 Measurement of data ASQ subscales (M alpha = 0.83)reliable/valid tool Data analysis Findings ANOVAsample demographics p > .05 (randomization effective) Split-plot ANOVAintervention impact IV1 & DV1 3-way interaction of Group x Appraisal x Time, F(10, 1610) = 2.20, p <.05, partial n2 = .19 Split-plot ANOVA- types of barriers to supervision & FU w/ Bonferroni Pre & Post Intervention (Intermediate/1-year) Appraisals x Time, F(5, 800) = 3.58 and 6.35, p < .01 ↑ appraisal scoresimmediate and 1-year (persistent effects) IV & DV2 Barrier type reported F(3,258) = 30.65 p <. 01 Environment/location (↑ reporting) M = .52 SD = .35 p < .05 Decision to use in practice LOE: II Strength: RCT design; parental inquire on identifying barriers; minimal bias; low attrition rate Weakness: mostly Caucasian sample, majority of mothers participating; multiple intervention components Conclusion: tailored video messaging improves supervision practices by identifying different risk appraisals that motivate different types of safety behaviors, depending on parental supervision needs/experiences Clinical significance: safety behavior changes using this approach is sustainable and better retained by parents TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Citation Shields, W. (2013). Utilizing the pediatric emergency department to deliver tailored safety messages: Results of a randomized controlled trial Country: United States (Baltimore, MD) Funding: grant to Johns Hopkins Bloomberg School of Public Health from the National Institute of Child Health and Human Development Bias: none reported Theory/ Conceptual Framework Precaution Adoption Process Model Design/ Method/ Purpose Design: RCT Purpose: to evaluate impact of computer kiosk intervention on parents’ selfreported safety knowledge and observed behaviors; compare selfreported versus observed behaviors Sample/Setting N = 901 IG: n = 448 (tailored report) CG: n = 453 (generic report) Demographics: Age range, y (child) 1-2 y: 42% m: 51% Respondents AA: 93% Mothers: 90% Not married: 69% 20-29 y: 55% HS: 74% annual PCI< $5K: 63% Setting: pediatric trauma ED Inclusion: ESP, child between 4-6 months c/o treatment for injury or medical complaint, live in Baltimore Exclusion: parents of critical patients, parents of children with suspected child abuse/neglect Attrition: 20% (lost to FU) Major Variables Studied IV: tailored safety computer program DV1: safety knowledge DV2: self-reported safety behaviors DV3: observed safety behaviors Time frame of intervention – 10-12 minute assessment FU completed 2-4 weeks and 4-6 months post intervention 37 Measurement of data Data analysis Multiple choice and T/F questions t-tests (total M % correct scores) PAPM-stage based questions (r/v not tested but recommended use in predicting behavior stages) ordinal regression & logistic regression PPV and NPV & Fisher’s exact test Findings DV1 Total M % correct IG: 73.08 (13.6) CG: 69.41 (14.08) t = 3.54 p = 0.000 ES = 0.26 (IG ↑ % correct) DV2 (IG,CG) OR=1.36; 95% CI=1.05, 1.77 p = .02 (IG in ↑ PAPM stage) DV3 IG, CG (smoke alarms) PPV: 33%, 54% p = .09 NPV: 23%, 66% p = .06 (over-reported unsafe behaviors) IG, CG (poison storage) PPV: 16%, 14% p = 1.0 (under-reported unsafe behaviors) NPV: 100%, 100% p = 1.0 Decision to use in practice LOE: II Strength: home observation of safety behavior; RCT design; high FU rate; direction observation; reasonable cross sample of conditions; strong sample size Weakness: self-reporting; car seat observations not possible Conclusion: improved safety knowledge scores and increasing self-reported behaviors due to intervention; consistent knowledge gains evident after four months on computer tailoring home safety information Clinical significance: results are encouraging for clinical settings as they were achieved without burdening providers TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Citation Gittelman, M.A. (2014). A computerized kiosk to teach injury prevention: Is it as effective as human interaction? Country: United States (Cincinnati, OH) Funding: Ohio Department of Public Service-EMS Injury Prevention Research Grant Bias: low- convenience sampling Theory/ Conceptual Framework Elaboration Likelihood Model Design/ Method/ Purpose Sample/Setting Design: Prospective, randomized trial N = 317 KB-G: n = 172 IPS-G: n = 145 Purpose: determine if computerized kiosk in Pediatric ED can screen families for injury risk, encourage to make more safety changes at FU compared to IPS Demographics P/G MA: < 35y: 64% E (child) white: 62.8% f (child): 49.8% mother taking survey: 91.2% HS grad or lower: 34.4% Age range, y (child enrolled) 1-4y: 36.3% Setting: CCHPC ED Inclusion: parent/legal guardian w/ child between ages birth through 14y with c/o urgent/nonurgent Exclusion: parents of critically ill or emergent child, non-English speaking, no parent/guardian present at time of visit Attrition: 31% (lost to FU) Major Variables Studied IV1: KB injury instructions IV2: IPS instructions DV: change in % of M correct behavior responses per age Time frame of the intervention given: KB-G completed in 6 minutes IPS-G completed in 8 minutes FU completed 3 weeks after intervention 38 Measurement of data Survey questions based on AAP TIPP program (α = 0.869) reliable and valid survey Data analysis X2 Analysis (categorical variables) Student’s t test (continuous variables) Findings Decision to use in practice IV1, IV2 & DV <1 y 2.95 (7.83), 7.65 (10.16) p = 0.144 LOE: II 1-4 y -0.45 (9.92), 7.59 (8.74) p = 0.0002 Weakness: convenience sampling, parents only approached when CRCs available; IPS employed at ED safety resource center; KG practiced safer behaviors at initial kiosk screen; self-reporting; modest attrition rate 5-9 y 2.58 (8.13), 9.82 (9.04) p = 0.002 10-14 y 0.37 (9.36), 8.44 (6.14) p = 0.003 Overall change in % for DV IV1: 1.05 (9.04) IV2: 8.31 (8.62) p < 0.0001 Strength: kiosk screening time was short; families satisfied with service Conclusion: computerized kiosk in ED waiting room appropriate place to screen for child safety risk, but would be more effective if printed information was reviewed and given by provider (human being) Clinical significance: kiosk screening is quick and provides useful information for decrease injury risk and would be more beneficial if information was discussed with provider TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Citation Theory/ Conceptual Framework Van Beelen, M.E.J. (2014). Effectiveness of web-based tailored advice on parents’ child safety behaviors: Randomized controlled trial Health Belief Model Country: Netherlands Funding: grant from ZonMw, the Netherlands Organization for Health Research and Development Bias: none reported Design/ Method/ Purpose Design: RCT Purpose: to evaluate effectiveness of E-Health4Uth home safety on parents’ safety behaviors with regard to prevention of falls, poisoning, drowning and burns Sample/Setting N = 1292 IG – n = 643 (WB w/ personal counseling) CG- n = 649 (generic) Demographics: Mother 93.58% MA, y: 32.06 Low education: 15.19% Employed: 83.44% E, Dutch: 88.46% Father MA, y: 34.51 Low education: 22.4% Employed: 95.67% E, Dutch: 87.94% Children M: 51.32% MA, mo.: 7.21 Setting: well-baby clinics Inclusion: parents with child between 5-8 months old and eligible for routine well-baby visit Attrition: 6.6% (lost to follow-up) 39 Major Variables Studied Measurement of data IV: web-based tailored safety advice w/ personal counseling total risk score (max score 53 points- r/v tool for risk behavior) DV: unsafe behavior (total risk score at FU) Time frame of the intervention given – FU conducted 10 months post intervention Data analysis Findings Logistic (for specific safety behaviors & Linear regression (total risk score) analyses DV IG: M = 13.63 SD = 6.12 (1.00-3.00) CG: M = 15.34 SD = 6.07 (0.0037.00) OR -1.59, 95% CI 2.26 to -0.93 (IG with significant ↓ in total risk score) Decision to use in practice LOE: II Strength: focus on effect of tailored intervention on parents’ child safety behaviors and overall safety risk score; RCT design; large number of participants; few lost to FU Weakness: intervention not tested on mobile and tablet devices; potential recall bias from gift vouchers; self-report behavior may lead to underestimation Conclusion: use of tailored parent information combined with counseling from provider is effective in promoting parents’ child safety behavior Clinical significance: supports the use of web-based, tailored, safety advice for injury prevention in pediatric health care setting TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Citation Theory/ Conceptual Framework Glassman, T.J., (2017). A social marketing intervention to prevent drowning among inner-city youth Health Belief Model Country: U.S. Midwest Funding: grant from the Center for Injury Research and Policy funded by CDC Bias: Design/ Method/ Purpose Design: quasiexperimentalmatched pairs Purpose: determine if social marketing campaign guided by HBM could improve parent’s knowledge and perceptions concerning water safety, overall improve adult supervision Sample/Setting N = 65 IG: n = 39 ( swim lessons/messages) CG: n = 26 (no swim lessons or messages) Demographics: MA- mother: 43y MA- child: 9y E (AA) – 78% Setting: inner-city swimming course Inclusion: parent/guardian of inner-city youth, enrolled in 6-week session swimming course Attrition: not reported Major Variables Studied IV: social marketing intervention DV1: perceived threat DV2: perceived benefit DV3: perceived barrier DV4: self-efficacy DV5: behavioral intention DV6: knowledge Time frame of the intervention – IG: 6-weeks CG: 3 weeks 40 Measurement of data Structural survey questionnaires- 35 items using Likerttype scales (α = .70 to .83) (r/v tool) Data analysis Chi-square analysis (parent perception of child swimming ability); Two-way analysis of covariance Multivariate linear regression Findings X2 = 10.985; p = .012 ANCOVA: n2; F ME/post (DVs) 1 0.004; 0.202 2 0.001; 0.09 (p<.05) 3 0.000; 0.019 4 0.078; 4.543 (p<.05) 5 0.049; 0.964 6 0.272; 20.207 (p<.001) IE (DVs) 1 0.007; 0.365 2 0.001; 0.173 3 0.000; 0.012 4 0.071; 4.133 (p<.05) 5 1.250; 0.644 6 0.123; 7.594 (p<.05) Model R2 (DVs) 1 .504 2 .266 3 .189 4 .364 5 .100 6 .561 DV5 R2 = 0.336, F(1, 60) = 9.09, p < .01 DV1 β = .525, p = .00 Decision to use in practice LOE: III Strength: study consistent with extant literature Weakness: majority of female participants; low participation; social desirability Conclusion: increase in parent/guardians’ knowledge and self-efficacy on water safety, correctly identified risk factors associated with drowning Clinical significance: providers can use these results to develop feasible strategies (prevention messages to supplement swim lessons and decreasing drowning rates among at-risk youth. TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Citation Omaki, E. (2017). A systematic review of technology-based interventions for unintentional injury prevention education and behavior change Country: United States Funding: National Institutes of Health, National Institute of Child Health and Human Development Bias: low-risk due to possible social desirability or recall bias Theory/ Conceptual Framework Elaboration Likelihood Model Design/ Method/ Purpose Design: Systematic review Purpose: summarize evaluation of computer and mobiletechnology-based health behavior change applications in unintentional injury prevention; describe how successes can be applied to injuryprevention programs; identify research gaps 41 Sample/Setting Major Variables Studied Measurement of data N = 44 technologybased IP studies (RCTs n = 30) IV: computer-andmobiletechnology-based interventions DVs based on Target population DV1: knowledge impact DV2: behavior impact 1 Observed 2 Reported Bias risk assessment Demographics: Pediatric IP programs: 24 Setting: n/a Inclusion: CB interventions, program evaluationsdelivered by computer processor; studies that addressed unintentional injury prevention Exclusion: educational videos, computer-only data collection Attrition: 0 Time frame of electronic searches- article search published since 1990; eligible studies published in 2002 or later Categorized study types Data analysis Data abstraction form Findings 24 studies designed for Pediatric IP 5 study types identified 1 LHSP – 16 2 KB programs – 4 3 RHIP – 11 4 MT/PD – 2 VR – 11 BRA: 15 studies scored ≥ 20 DV1 Y – 20 No – 1 N/A – 23 DV2 Total N/A – 7 Decision to use in practice LOE: I Strength: several strong RCTs commonly used Weakness: relevant articles missed in search or excluded erroneously during review; English-speaking only articles used, no translator resources for foreign articles Conclusion: review provides evidence on computer-based programs and effectiveness in conveying safety information and positively influencing injury prevention and safety behaviors Clinical significance: providers have increased potential for use of computer-based programs for IP education 1 Observed Y – 19 No – 3 2 Reported Y – 11 No – 3 AA – African American, AAP – American Academy of Pediatrics, ANCOVA – analysis of covariance, ANOVA – analysis of variance, ASQ – American Society for Quality, BCT – Behavior Change Theory, BRA – bias risk assessment, CBA – controlled before and after studies, CCHPC ED – Cincinnati Children’s Hospital Medical Center Emergency Department, c/w – compared with, CB – cognitive behavioral, CB/MT – computer-and-mobile based technology intervention, CDC – Centers for Disease Control and Prevention, CG – control group, CI – confidence interval, c/o – complaint of, CSS – child safety seat use, df – degrees of freedom, DV – Dependent variable, E – ethnicity, EA – exposure analysis, ED – emergency department, ELM – Elaboration Likelihood Model, ES – effect size, ESP – English-speaking parents, f – female, FU – follow up, G-IPI – generic injury prevention information, HB – hospital-based, HBM – Health Belief Model, HE – high exposure, HI – high income, HS – high school degree, HS = - high school degree/equivalent, IE – interaction effect, IG - Intervention group, IG1 Intervention group1, IG2 - Intervention group2, IP – injury prevention, IPS – injury prevention specialist, ITT – intent-to-treat, IV - Independent variable, KB – kiosk-based, KG – kiosk group, LE - Low exposure, LI – low income, LHSP – locally hosted software program, LOE – level of evidence, m – Male, M –Mean, MA – mean age, mo – months, ME – main effect, MT/PD – mobile technology and portable device, N – Sample (population), NPV – negative predictive value, OR – odds ratio, n – Sample size (studies), N/A – not assessed/available, NR – not reported, PAPM – Precaution Adoption Process Model, P/G – parent/guardian, PCI – per capita income, PPV – positive predictive value, PS – poison storage, QE – quasi experimental, r/v – reliability and validity, Race – race, RCT – Randomized Controlled Trial, RHIP – remotely hosted internet program, S – swimming, SA – smoke alarm use, SD – standard deviation, SR – systematic review, T/F – true/false, T-IPI – tailored injury prevention information, T-IPI + P- tailored injury prevention information plus supplementary tailored provider information, TIPP – The Injury Prevention Program, TM – tailored messaging, VR – virtual reality program, w/o – without, WB - web-based tailored safety advice, y – years, Y - yes Running head: TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Appendix B Synthesis Table Table 1. Review of literature and evidence synthesis table. Author Year Design Level of evidence Number of subjects/studies Theoretical Framework Setting Pediatric clinic Swimming course ED Laboratory Demographics ↑ % Mother respondents ↑ % Low income (<$10k) ↑ % African American parents ↑ % HS degree or lower ↑ % 20-29 M age range, y- P/G Pediatric IP programs Child age, y (under 5) Independent Variables Computerized TM Kiosk-based TM T-IPI only T-IPI + P Video-based TM IPS instructions CarlsonGielen Gittelman Glassman McDonald Morrongiello Nansel 2007 RCT II 901 ELM & PAPM 2014 RCT II 317 ELM 2017 QE III 65 HBM 2005 RCT II 144 ELM 2012 RCT II 163 HBM 2002 RCT II 174 HBM 2007 RCT II 594 HBM X X X Nansel Omaki Shields Van Beelen 2017 SR I 44 studies BCT 2013 RCT II 901 PAPM 2014 RCT II 1292 HBM X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Web-based TM Social marketing Computer/Mobile technology Dependent Variables Knowledge impact Behavior impact- Observed Behavior impact- Self- report Provider-parent communication Provider-parent communication Parent perception Participant reactions to information ASQ scores Barriers to supervision Change in % of M correct behavior responses per age FU injury risk score Anxiety scores Findings Tailored > Generic Tailored w/ provider > Tailored w/o provider 43 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X AA – African American, AAP – American Academy of Pediatrics, ANCOVA – analysis of covariance, ANOVA – analysis of variance, ASQ – American Society for Quality, BCT – Behavior Change Theory, BRA – bias risk assessment, CBA – controlled before and after studies, CCHPC ED – Cincinnati Children’s Hospital Medical Center Emergency Department, c/w – compared with, CB – cognitive behavioral, CB/MT – computer-and-mobile based technology intervention, CDC – Centers for Disease Control and Prevention, CG – control group, CI – confidence interval, c/o – complaint of, CSS – child safety seat use, df – degrees of freedom, DV – Dependent variable, E – ethnicity, EA – exposure analysis, ED – emergency department, ELM – Elaboration Likelihood Model, ES – effect size, ESP – English-speaking parents, f – female, FU – follow up, G-IPI – generic injury prevention information, HB – hospital-based, HBM – Health Belief Model, HE – high exposure, HI – high income, HS – high school degree, HS = - high school degree/equivalent, IE – interaction effect, IG - Intervention group, IG1 Intervention group1, IG2 - Intervention group2, IP – injury prevention, IPS – injury prevention specialist, ITT – intent-to-treat, IV - Independent variable, KB – kiosk-based, KG – kiosk group, LE - Low exposure, LI – low income, LHSP – locally hosted software program, LOE – level of evidence, m – Male, M –Mean, MA – mean age, mo – months, ME – main effect, MT/PD – mobile technology and portable device, N – Sample (population), NPV – negative predictive value, OR – odds ratio, n – Sample size (studies), N/A – not assessed/available, NR – not reported, PAPM – Precaution Adoption Process Model, P/G – parent/guardian, PCI – per capita income, PPV – positive predictive value, PS – poison storage, QE – quasi experimental, r/v – reliability and validity, Race – race, RCT – Randomized Controlled Trial, RHIP – remotely hosted internet program, S – swimming, SA – smoke alarm use, SD – standard deviation, SR – systematic review, T/F – true/false, T-IPI – tailored injury prevention information, T-IPI + P- tailored injury prevention information plus supplementary tailored provider information, TIPP – The Injury Prevention Program, TM – tailored messaging, VR – virtual reality program, w/o – without, WB - web-based tailored safety advice, y – years, Y - yes TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Appendix C Demographics Table 1. Parent demographics Parent Demographics Relationship to child Parent Legal Guardian Gender Female Male Age <25 years 25 years and older Age not provided Education < High School High School or GED Some college/tech school College graduate Employment 40+ hours per week 1-39 hours per week Not employed Ethnicity White Hispanic African American Other Total # of children 2 or less 3 or more N Percent 26 3 89.7 10.3 25 4 86.2 13.8 7 20 2 25.9 74.1 ** 6 9 12 2 20.7 31 41.4 6.9 16 5 8 55.2 17.2 27.6 2 19 2 6 6.9 65.5 6.9 20.6 15 14 51.7 48.2 44 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 45 Appendix D Correlation Table Table 1. Correlation results between the risk scores and parent demographics. Risk Scores Milestones Parent age Education Employment Total # of children .10 -.28 -.18 .42* Pool Party -.07 -.24 -.11 .20 Supervision Life Jacket .13 .12 .16 -.05 Correlation is significant at the 0.01 level (2-tailed)** Correlation is significant at the 0.05 level (2-tailed)* .11 -.02 .06 -.15 Emergency Total Preparedness Risk Score -.01 .10 -.13 -.20 .03 -.03 -.08 .14 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 46 Appendix E Categories of Drowning Risk Table 1. Categories of drowning risk results. Milestones Mean SD 9.0 3.5 Pool Party 7.0 1.4 Supervision 7.2 2.1 Life Jacket 9.3 2.7 Emergency Preparedness 9.6 2.9 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 47 Appendix F Independent t-test for Total Risk Score and Parent Demographics Table 1. Independent t –test: Total Risk Score and Parent Age Total risk score Parent age <25 years old 25 years old + N 7 20 Mean 43.57 41.75 Levene’s test for equality of variances Total risk score Equal variances not assumed Std. deviation 11.57 6.09 t-test for equality of means Sig. t df Sig. (2-tailed) 95% CI Lower Upper .008 .534 7.196 .702 -8.952 12.595 p < .05 Table 2. Independent t-test: Total Risk Score and Total Number of Children Total risk score Total # of Children 2 children or less 3 children or more N Mean Std. deviation 15 14 41.53 42.93 8.806 6.044 Levene’s test for equality of variances Total risk score p < .05 Equal variances assumed t-test for equality of means Sig. t df Sig. (2-tailed) 95% CI Lower .150 -.494 27 .625 -7.192 Upper 4.402 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 48 Table 3. Independent t-test: Total Risk Score and Education Level Total risk score Education level < college graduate college graduate N 27 Mean 42.56 Std. deviation 7.633 2 37.50 3.536 Levene’s test for equality of variances Total risk score Equal variances assumed t-test for equality of means Sig. t df Sig. (2-tailed) 95% CI Lower .391 .917 27 .367 -6.253 Upper 16.364 p < .05 Table 4. Independent t-test: Total Risk Score and Employment Total risk score Employment Not employed employed N 8 21 Mean 41.00 42.67 Levene’s test for equality of variances Total risk score p < .05 Equal variances assumed Std. deviation 9.243 6.931 t-test for equality of means Sig. t df Sig. (2-tailed) 95% CI Lower Upper .345 -.528 27 .602 -8.144 4.810 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 49 Appendix G key words Search Specific terms used: main used: parent drowning, water safety, aspectsprograms, of PICO question with search supervision limits: 20132018, peer-reviewed Psych Info: n = 79 Psych Info: CINAHL: n = 7,054 n = 64 CINAHL: TRIP: n = n26, = varied 472 results TRIP: nCochrane = 627 Library: n = 11 Cochrane Library:n n= =132,128 PubMed: PubMed n = 1,192 Articles addressed at least TWO elements of PICO question n = 25 Secondary search in PubMed with specific education type search terms Filters applied: 2013-2018 Excluded non-English Articles with search terms: webbased, mobile health technology, parent knowledge n = 20 Articles with search terms: parent training programs n = 5 Outlier articles from hand search n=4 Articles excluded n = 15 Insignificant findings Erroneous population groups Lack of generalizability Interventions unrelated to desired outcomes Final yield n = 10 RCT: n = 7 Non-random control trial: n = 1 Quasi-experimental n = 1 Systematic Review: n = 1 Figure 1. Flow chart detailing process of article search strategy and selection. TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 50 Appendix H Figure 1. Elaboration Likelihood Model. (Petty & Cacioppo,1986) TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 51 Appendix I Figure 1.. The Iowa Model of Evidence-Based Practice to Promote Quality Care (1994) (Titler, 1994) TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Appendix J 52 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 53 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 54 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 55 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 56 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT 57 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Figure 1. 20-item tablet risk assessment survey. 58 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Appendix K Figure 1. Summary page: Poolside plan. 59 TAILORED MESSAGING FEEDBACK TO IMPROVE PARENT Appendix L Figure 1. Primary outcome variables: Likelihood of behavior change and relevance of tailored intervention. 60