Child maltreatment is a significant social and health problem in the United States. Data from state child protective service systems indicate that in 2007 there were over 3.2 million reports, involving an estimated 5.8 million children, to child protective service systems for suspected child maltreatment, resulting in over 794,000 cases of substantiated child maltreatment. Children under 7 years of age were at greatest risk for maltreatment, accounting for 55.6% of all substantiated reports in 2007.Child neglect is by far the most common form of reported maltreatment (59%), and children 7 years and under constituted 61.3% of all child neglect (DHHS 2009).

Direct and indirect costs of child neglect represent an extensive public health problem (Kelley et al. 1997; Wang and Holton 2007). Mitigating these costs through prevention is possible, particularly in the area of child safety. Preventable, unintentional injury occurs in homes. For instance, of all the drowning accidents that occurred in the U.S. in 1995, 78% happened in the home, and 71% occurred in bathtubs (Brenner et al. 2001). Nearly 90% of calls to poison control centers took place from homes in 2003: 52% of all calls involved children under age 6, and 61% of those cases represented unintentional exposure to a poison. Among the top products resulting in poisoning were cosmetics and personal care products (13.4%); cleaners (9.7%); and pain remedies (7.8%; Watson et al. 2004).

There is evidence that in-home, skills-based parent training can be effective in reducing hazards that lead to unintentional injury for young children. Gershater-Molko et al. (2003) found a 78% mean percent reduction in household hazards among maltreating families completing the home safety module of the SafeCare program. Other skills-based parent training programs, such as Triple P and Parent-Child Interaction Therapy (PCIT), also have an evidence base in preventing child maltreatment (Prinz et al. 2009; Shanley and Niec 2010). Each has theoretical foundations in learning and social learning theory. Triple P represents a public health/universal approach to preventing child maltreatment. PCIT is mostly conducted in clinics, whereas SafeCare is a home visiting program. Triple P and PCIT do not focus on skills deficits related to child neglect. Such interventions are typically quite intensive, and the effectiveness of such programs depends on parental engagement and completion of the program (Gershater-Molko et al. 2003). For example, the Safety module of the SafeCare intervention requires 6 in-home sessions that last 90–120 min.

New technologies may be useful in helping engage families, increase compliance, and increase program effectiveness (Self-Brown and Whitaker 2008). For example, Bigelow et al. (2008) are using cell phones as a means to improve engagement and effectiveness of a child maltreatment intervention in a randomized controlled trial. Specifically, cell phones are being used to increase communication between home visitors and mothers, and daily texts are used to improve compliance with the intervention. Early results suggest that use of cell phones has increased attendance, retention, and satisfaction relative to intervention without cell phones.

In this study, we examined whether the iPhone could be used to augment the home safety intervention of SafeCare. As with any new intervention, it is important to conduct small-scale testing before broad implementation, but this is particularly true for interventions involving new technology such as the iPhone, which may be expensive. Single-case research design allows examination of behavior change at the individual level, which is important to understand if the ultimate interest is changing behaviors of larger numbers of people (Kazdin 1982). These designs have shown high internal validity and limited external validity without multiple replications or ensuing larger group design research.

SafeCare is an evidence-based program that emanated from Project 12-Ways (Lutzker et al. 1998). It consists of three skill-based modules that address proximal risk factors for physical abuse and neglect: parent-child interactions, health care, and home safety. Depending on the needs and progress of the family, each module typically involves six sessions, with pre- and posttraining assessments conducted to ensure skill acquisition. As the program is behavioral in nature, all skills are modeled by a home visitor and practiced by the parent, with coaching and positive feedback; demonstration of mastery is required to advance to other modules.

The current study focuses only on the safety module of SafeCare. Home safety addresses safety hazards in the home and is based on 10 categories on the Home Accident Prevention Inventory-Revised (HAPI-R). Training begins in the room with the greatest number of hazards, and the home visitor and parents methodically identify and remove hazards (e.g., poisons, electrical hazards, choking, and suffocation hazards). Following the initial assessment, parents are supported by the home visitor as they gradually take responsibility for securing rooms (Lutzker and Bigelow 2002; Whitaker et al. 2008).

Use of technology to augment the SafeCare home safety module began when Barone et al. (1986) developed an audio-slide show to assist parents in identifying safety hazards and to make determinations as to what hazards were accessible to children. While this technology enhancement to the intervention did not supplant the home visitor, it did reduce the amount of time required for home visitor training and time in the home. Further, Mandel et al. (1998) developed a series of four video tapes, approximately 10 min each, to deliver instruction on identifying and securing hazards. As with the Barone et al. research, the tapes did not eliminate the home visitor, but reduced home visitor training and provided positive feedback concerning the medium.

While both the Barone and Mandel studies found positive results, neither was widely incorporated into ongoing implementation of SafeCare. In the current research, we employed the iPhone as a technological enhancement to the SafeCare safety module. We used the iPhone as both an intervention enrichment and a data collection tool. We used a multiple baseline design (Kazdin 1982) across household rooms replicated across families to examine whether the iPhone-enhanced safety module would reduce safety hazards in three homes. Families involved in the project were given iPhones during the study; the phones were used to record video of rooms following intervention and also to increase communication between families and the home visitor. Parent satisfaction with the iPhone technology was assessed to further understand the feasibility of this hybrid technology and home visiting approach.

Method

Participants

Three families living in metro Atlanta participated in the iSafety research. Each family had previously completed a relationship enhancement program targeting low-income parents living together. Nine families were selected by the relationship enhancement program director and referred to the National SafeCare Training and Research Center (NSTRC) following their verbal consent to do so. Each of the nine families was called. Four declined to participate and one was unreachable. Four families that met the following selection criteria agreed to the study: have at least one child age 5 or younger; consent to participate in the iSafety project; and state a desire for a safer home. One family agreed to participate, but then was not at home for the first meeting or reachable thereafter; consequently, this family was dropped. Thus three families participated.

Sample

The research took place in the homes of participants who lived in Atlanta metro-area neighborhoods. Family A lived in a rented duplex in a high-poverty, deteriorating neighborhood. The home had a severe roach infestation and was extremely crowded with boxes, excess furniture, and people. There were multiple family units within the household, two with infants, as well as daily daycare service for three toddlers. While the unmarried couple chose not to report income, the mother had shared that she qualified for Medicaid during her pregnancy. Family B lived in a new single-family structure, which they owned, in a moderate-income subdivision. The two-story home had been newly furnished when they purchased the home 2 years prior and was very well maintained. The mother stayed at home full-time and reported an annual household income of $50,000+. The father of this family had completed a technical degree and secured a professional career position since the family’s initial participation in the relationship enhancement program from which they were recruited for this study. Family C lived in a rented home that was well-kept, albeit in need of minor repairs. Minor roach infestation was apparent. They reported an income of less than $10,000 annually.

Observation System: the Home Accident Prevention Inventory-Revised (HAPI-R)

The iSafety intervention is based on a single assessment tool, the HAPI-R. The HAPI-R was used to record hazards in 10 categories: poisonous solids and liquids; fire and electrical; mechanical objects; small objects and choking; sharp objects; firearms; falling, tripping, and activity restriction; crush; drowning; and organic matter and allergens. Some categories were further subdivided into smaller units.

To be counted on the HAPI-R during an observation, hazards had to be accessible. Accessible hazards are defined as those within reach and unsecured for the referent child, that is, oldest child in the home up to age five. Reach was derived by measuring the distance from the floor to the referent child’s outstretched fingertips while standing on tiptoes. The child was assumed to have the ability to climb on any surface at or below eye level. Unsecured was defined as any item not in intact childproof containers or locked. A separate HAPI-R was used for each room.

Observer Training

Observers were graduate students enrolled in the Master of Public Health degree program at Georgia State University and graduate research assistants at the NSTRC. During training, observers were provided with written material about the SafeCare safety module consisting of copies of the slides used during the presentation, HAPI-R, and HAPI-R Definitions-Home Visitor Version. The training focused on use of the HAPI-R through discussion, video, and practice in a staged room. A quiz on hazard identification was administered and each trainee achieved a minimum of 85%, as determined by the NSTRC trainer conducting the sessions.

Reliability

Reliability Training

The home visitor and observers practiced face-to-face reliability during in situ and staged training over several weeks until each observer and the home visitor achieved increasingly accurate reliability ratings consistently. The home visitor and the observers each scored the same room using the HAPI-R independently of one another and without conversation. Following each observation, the home visitor led a discussion of HAPI-R identification and counts to refine common application of the tool. Reliability with iPhone video was similarly practiced with actual iPhone video footage from participant homes and practice reliability scores with video were consistent with face-to-face reliability scores.

Calculating Reliability

Interobserver reliability was measured for each family during initial baseline data collection and during two additional baseline or training visits, randomly determined. During reliability sessions, the home visitor and an observer assessed hazards via the HAPI-R in three rooms per household. If one of the randomly assigned visits consisted of video assessment, the home visitor and the observer used iPhone video, which was emailed to the home visitor by the family. Rules empirically established and listed below were applied to each observer’s HAPI-R and determined yes-no reliability within each of the 28 HAPI-R subcategories of hazards where either observer scored an item. This created very conservative reliability estimates as categories in which both observers scored “no hazards” were not included in the estimates.

  1. 1)

    Subcategories with 10 or more hazards required agreement between observers within 4 items to be scored a yes for reliability.

  2. 2)

    Subcategories with between 3 and 9 hazards required agreement between observers within 2 items to be scored a yes for reliability.

  3. 3)

    Subcategories with 1 or 2 hazards required perfect agreement between observers to be scored a yes for reliability.

Percentage agreement was then calculated. All valid subcategories were entered into the following formula to derive an interobserver reliability score:

$$ \frac{{Agreements}}{{Agreements + Disagreements}} * 100 $$

Three independent observations were made by the home visitor and an observer in all intervention rooms across families, with at least one observation occurring during the face-to-face baseline condition, and means across Families A–C were 60%, 94%, and 76%, respectively. Subsequent treatment-phase reliability observations were based on iPhone video and face-to-face data, and means for each family were as follows: 86%, 100%, and 88%. Accuracy checks were calculated between the home visitor and iPhone video with Family A only, and the following percentages were obtained: 75%, 86%, and 88%.

E-V accuracy

An accuracy check was assessed between home visitor observations and video observations (E-V) for one family. Because the iPhone was used at once as a tool for the intervention and assessment during the posttreatment phase, E-V accuracy was not assessed during baseline.

Materials

An iPhone 3GS was provided to each family. This model is the first in the smartphone class to offer video capture. This feature enabled participants to video rooms for data collection and send to the home visitor via email with relative ease and quality sufficient to maintain high E-V accuracy. The iPhone was also used to communicate specific feedback and praise related to safety content, as well as to facilitate logistics. Printed materials consisting of a list of abbreviated HAPI-R definitions, photocopies of pictures of poisonous house plants common in Georgia, and a handout for travel safety were provided to each family.

All families were provided with home safety devices and additional items needed to secure specific safety and cleanliness hazards. A local-area Target store awarded a $50.00 gift card for the supplies.

Experimental Procedure

Orientation and Baseline

During the first home visit, the home visitor discussed the iSafety project in broad terms. It was explained that the intervention generally required six visits and that the focus would be on securing hazards and creating a healthy environment in three selected rooms. The home visitor’s iPhone was introduced to briefly demonstrate video and email capacity and ease of use. Informed consent was obtained, and the Home Assessment Consent form was explained and signed. Institutional Review Board approval was granted by Georgia State University.

The referent child was measured for eye level and for reach. Reliability data were collected across rooms, across families during this visit. Before leaving, the home visitor summarized the session, invited questions, previewed the next session, and confirmed a date and meeting time.

The second visit began with a brief overview about the purpose of the session, and baseline data were collected by the home visitor in the three predetermined rooms. Data from each of the three rooms were examined to determine in which room the data showed the greatest stability. While all data were relatively stable across participant families throughout baseline, the room in which the number of hazards increased the most was generally targeted for intervention.

Parent Training

Once the target room was selected during the second visit, parent training on the categories of hazards and ways to secure or remove the hazards took place via conversation, demonstration, and practice. The parent was given and invited to use her printed materials as a reference throughout. The interactions focused on identifying specific hazards according to baseline data, identifying similar household hazards, and understanding why items were potentially dangerous. Supervision was stressed as key throughout.

Several of the hazards in the focal room were removed or secured systematically as a team, and several items were left as “homework” for the parent, based on readiness. If needed for the focal room, the parent was presented with a home safety device kit, and the devices were either installed by the home visitor and participant, or the home visitor showed the participant how to install the devices and verified that the required tools were available.

The parent was given the iPhone at the end of the second visit. Basic iPhone functions were shown, including the video feature and email. A Gmail™ account had been established for the study using the participant-chosen pseudonym. The home visitor and participant practiced filming and emailing video clips to the home visitor. The parent was asked to finish securing hazards in the room, video the room, and email the information to the home visitor by a mutually agreed-upon date prior to the third visit. A reminder alarm for the third visit was set on the iPhone.

In addition to the capturing homework video following visit two, the iPhone was used to communicate with the participant. Specific praise and coaching was communicated via texting or email, sometime resulting in further reduction of hazards. While the goal was to remove all hazards, significant reductions diminishing hazards in a room was deemed acceptable for continuation. A reminder about the third session was texted prior to the visit.

The third and fourth home visits followed the same pattern as the second visit, with different rooms, until hazards were reduced to near zero. Face-to-face data collection in each room continued at the beginning of the visit until the intervention took place, and then iPhone video data collection supplanted it, significantly shortening the duration of each successive visit. A fifth home visit was made to share information about topics such as travel safety, food storage and preparation safety, lead poisoning, and fire safety and smoke detectors; to review home safety information; and to stress the importance of supervision and evolving adjustments as children’s abilities changed. The iPhone was collected during this visit. No communication took place between this and the final visit.

Follow-up

The follow-up sixth and final visit, approximately one month following the fifth, included face-to-face data collection of all three rooms and an opportunity to clarify information and express concerns related to safety and supervision. A brief questionnaire was completed by all participants to gather socio-demographic data and preintervention experience levels with smartphones, as per an amendment granted by the Institutional Review Board.

Experimental Design

A multiple baseline design across household rooms replicated across families was used to evaluate the effect of the intervention. Data were collected face-to-face via the HAPI-R during each visit in each room until baseline stability or an ascending hazards trend was established, and one room was selected to initialize the intervention. The data collection then moved to video, each time via the HAPI-R, subsequent to the intervention. Follow-up data were collected face-to-face. The use of single-case research designs with a small number of participants allows intervention researchers to carefully examine an intervention and “tweak” any potential flaws before exposing a larger number of participants to a randomized control or comparison trial.

Consumer Evaluation

Upon completion of the intervention, families completed the Parent Satisfaction Survey designed to measure the parent’s perception of the program’s utility in terms of safety and ease of use in terms of the iPhone. The two categories of questions centered around satisfaction with the program (e.g., Since I finished the iSafety program, my home is much safer/safer/the same/ less safe) and with the iPhone (e.g., When the home visitor explained the information about how to use the iPhone, I understood well/understood/was a little confused/was very confused). The survey consisted of 10 sentences that the parent finished by selecting from a four-item, Likert-type scale ending. Additional space was provided for comments.

Results

The effects of the safety module intervention and iPhone enhancement on the total number of hazards per room per family are shown in Figs. 1, 2 and 3. Receipt of the safety intervention with iPhone enhancement dramatically reduced the number of hazards in each room for all three families, which was maintained through the follow-up visit. Baseline data show average hazards ranging from 43 to 81 per room for Family A, from 28 to 116 for Family B, and from 11 to 121 for Family C. During the training condition, a significant decrease in the range of average hazards per room was demonstrated across families: 10–17 for Family A, 1–5 for Family B, and 0–9 for Family C. This represents an average reduction in household hazards of 74%, 93%, and 97%, respectively and indicates the number of hazards diminished considerably as a result of the safety intervention with iPhone enhancement. These reductions were maintained or reduced further during the follow-up condition for Families A and B. The bathroom and kitchen for Family C showed slight increases in hazards from the training condition, but still large reductions from baseline; the children’s bedroom was maintained at zero hazards.

Fig. 1
figure 1

Total hazards per room for Family A over the course of the study and follow-up period

Full size image
Fig. 2
figure 2

Total hazards per room for Family B over the course of the study and follow-up period

Full size image
Fig. 3
figure 3

Total hazards per room for Family C over the course of the study and follow-up period

Full size image

Further, there are indications that reductions continued in rooms across the training condition. For Family A, hazards were immediately reduced in the living room, stabilized over the next two data collection points, dropped yet lower when there was no longer specific discussion of hazards in that room, and again showed a significant drop during follow-up data collection. As a bathroom closet was secured during the training phase, the father independently move hazardous items from the living room into that locked closet; that, among other efforts, contributed to ongoing reductions in the living room. Thus, generalization occurred.

Communication via iPhone was categorized into one of three categories: logistical, content question (initiated by the participant), or feedback (initiated by the home visitor). Logistical questions or reminders constituted the largest mode of communication across Families A–C (65%, 63%, and 53%, respectively). The mode communication by far was texting (86% average), though over the month that the intervention took place, this did not represent an onerous number of text messages. There were 33 text messages between the home visitor and Family A, and 38 and 29 with Families B and C, respectively. Fewer than 5 phone exchanges occurred, and the range of email message across families was 1–8.

A consumer satisfaction survey completed by all families showed positive results. Reactions to the program and the iPhone enhancement were wholly favorable. Parents considered their homes safer and expressed confidence in recognizing and securing hazards. The time required to make homes safer was not perceived as burdensome by any family, nor were the communications. No parents had difficulty adapting to use of the phone, as evidenced by no iPhone training session lasting longer than 5 min and by no subsequent communication related to technical aspects of the device. When asked about their experience with texting and emailing, all parents were reported to have either “a lot of experience” or “regular” experience.

Discussion

The SafeCare safety module enhanced by iPhone yielded substantial reductions in household hazards in all rooms across all participants, including a 1-month follow-up postintervention. The iPhone was used as both intervention tool and data collection tool, and the drop in the number of hazards in households in this research is consistent with earlier SafeCare studies specifically focused on the safety module in which the HAPI or HAPI-R was a pivotal component (Tertinger et al. 1984; see also Barone et al. 1986; Mandel et al. 1998; Metchikian et al. 1999). That is, previous studies showed a significant reduction in hazards during the training condition as compared to baseline and followed similar procedures. The face-to-face time was reduced in the current study with the iPhone serving as a data collection tool, compared to previous studies with all data collection taking place face-to-face. Thus, the iPhone enhancement appears to yield results as would be expected during standard, face-to-face data collection. Parents also believed the intervention to have been beneficial. All participants felt their home was safer having completed the program and more confident in identifying and securing hazards. Two families indicated on more than one occasion that they were not familiar with certain facts about household safety or with safety devices prior to the intervention.

Thus, this research suggests promise in achieving a yet more efficacious household hazard reduction intervention in several ways. Technology such as the iPhone may yield a more efficient intervention from a cost perspective. Implementation costs for interventions such as SafeCare are considerable. By using iPhone video for data collection, the number of sessions was reduce to five and can likely be reduced further. The current iPhone protocol also called for greater independence by the family for intervention. This may further reduce the costs of intervention. As further research using iPhones and other technology is conducted, costs-effectiveness analyses, such as the one being conducted by Bigelow et al. (2008) in their cell phone enhancement research, will become critical.

The use of the iPhone may make home safety interventions more acceptable and engaging, and help reduce attrition. Home visitation, and in particular, home visitation during which personal space is scrutinized for hazards, may be implicated in high attrition–low participation rates common to child maltreatment interventions (Sangvai et al. 2007). During the current study, Family A’s home was “cleaned up” immediately before the intervention visit because the matriarch/grandmother was uncomfortable with how “messy” the home had been during the first visit by “strangers.” Interestingly, the cleaning up resulted in removal of few hazards. The risk of this family dropping out due to the invasive nature of home visits may have been mitigated by the decreasing length of the visits over the course of the intervention.

The iPhone greatly enhanced communication between visits through texting, email, and phone or voicemail messages, and made the logistics of intervention considerably easier. This is critical for populations at risk for maltreatment, where missed appointments are frequent and attrition is common (Gershater-Molko et al. 2003; Sangvai et al. 2007). Interestingly, there were no missed appointments and all families completed the program. Use of such technology in reducing attrition and increasing compliance shows potential. To the extent that the participants, all between 20- and 30-years-old, are characteristic of their peers, mastery of this type of technology is firmly in place and is not an obstacle for widespread use in interventions.

Limitations

A number of limitations to this research must be noted. While the cost savings suggested by reduction in face-to-face visits is promising, these economies obviously depend on the ubiquity of handheld devices with high-quality video capacity and wireless connectivity. Conversely, the anecdotal comments by participants indicating excitement about having access to an iPhone would be negated by its very universality. How much this novelty contributed to the intervention effects merits further consideration.

While the iPhone is a state-of-the-art smartphone, it has limitations that became apparent during study implementation. The lack of back lighting for video recording made it extremely difficult to read the video at times. Shooting scorable video under beds and in closets, both places where children have access and where ingestible objects had been found, was challenging. The restrictions on video file size was also problematic, and on average, four individual clips per room were needed to allow for scoring. Thus, each family had to send up to 12 separate videos individually. While on the Parent Satisfaction Survey all parents responded that the iPhone was helpful or very helpful and easy to use and reported no dissatisfaction with upload speed or the number of videos required for data collection, this seems excessive. Technologies will certainly continue to evolve and improve, and it is likely that both of the problems mentioned above will be resolved.

Related to the limitations of the video is how critical it was for the home visitor to be familiar with the home and the layout of the rooms. Only because of the baseline condition over time followed by direct work alongside the participant in securing a room from hazards did the video footage afford a sense of dimension and allow remote assessment. Further, familiarity with the types of hazards that had been present over time alerted the home visitor to possible hazards that might be seen in the video. A related point is that some individuals may have difficulty following instructions to capture useable video. One family, despite repeated instructions, required ongoing redirection, with video footage either too fast, too dark, or incomplete. This became an issue for all families as the study neared an end. As noted earlier, by this point data collection was conducted exclusively by video, necessitating often 12 or more emails between visits. Clearly, iPhone video will not soon obviate the human element, but it does hold considerable promise for reducing the amount of time in situ per visit and the number of home visits overall.

Future Directions

The integration of an iPhone to enhance the SafeCare safety module offers a novel and innovative method for engaging families at risk of child neglect. Obviously, more studies are needed to examine the effectiveness of an iPhone enhanced intervention, and if those studies confirm the iPhones effectiveness, cost studies that compare an iPhone enhanced safety intervention to the standard face-to-face safety intervention. The relevance of such studies could change quickly as tools such as the iPhone may become ubiquitous, or may be replaced by some different technology.

In addition to the data collection and intervention enhancement represented by this study, the capacity of the iPhone in general is promising for other aspects of parent training (Self-Brown and Whitaker 2008). In terms of safety, the video capabilities could be used to produce short “how-to” segments for installing safety locks on cabinets that parents could access at will. Videos that provide brief, engaging information about preventable accidents by age could also be useful in helping parents remain vigilant as their children’s capabilities evolve. Other video could present parent-child interactions, health, or virtually any other topic of interest to parents. Additionally, there are myriad applications created for the iPhone (applets) and other smartphones that might also enhance the program. Applets with repetitive reminders for completing intervention activities and scheduling visits can be useful in increasing participation and reducing attrition. Other applets that produce lists that participants can check off are also promising in creating greater compliance, particularly ones that can be customized to perform novel “rewards,” such as playing favorite music or showing personal photos when an item on a “to do” list is completed. This study demonstrated that the iPhone may be able to reduce the number of face-to-face visits during the SafeCare home safety module. If future research also demonstrates that an application with a training video can also reduce the number of intervention visits, it would be useful to conduct a randomized control trial comparing iPhone to no iPhone with the SafeCare home safety module that also includes a cost analysis.

In summary, despite the limitations noted, this research adds to the nascent literature on the use of everyday technology to enhance evidence-based home visiting programs. The iPhone holds promise with other SafeCare modules and for other evidence-based practices.