A systematic review of interactive multimedia interventions to promote children’s communication with health professionals: implications for communicating with overweight children

While there was a wide age variation amongst study participants, several papers [7, 65, 69] analysed data within narrower age groups. Data relating to the age category closest to, and incorporating 7 to 11-year-olds, were selected for inclusion. Only two studies focussed on 7 to 11, or 7 to 12-year-olds; either as the total study population [72], or analysed as a subgroup [69]. Differing age ranges in turn affected the mean ages of recruited children.

The ability to compare mean ages was compromised by a lack of data. With the exception of five studies [6, 9, 66‐68], papers failed to provide the mean age of recruited children. One paper reported the median age of the comparison and intervention group [8] instead of the mean. This study reported a significant age difference (p = 0.012) between the two groups. Five studies [7, 65, 68‐70] offered no indication as to group similarity or differences in participant’s mean ages. Where necessary, and feasible, the mean age of participants was calculated from study data [8, 71‐74] (Figure 4). Within the ten studies, the mean age of children was 10.11 ± 2.13 with the mean range from 7.40 to 13.05-years.

The heterogeneity evident in the age ranges of child-participants was further apparent in the MIs themselves (Table 5). Diversity existed on a number levels; including platform delivery (CD-ROM, web-based, etc.) and the setting in which child-participants accessed the MI (medical settings, school, home, etc.).

Perhaps the most obvious difference between the selected studies was what the MIs were designed to achieve. Most of the interventions had either stated or implied aims of educating children about their health-related condition; and improving skills and confidence to manage symptoms, triggers and medication. Some MIs [65, 69, 70, 73] had an additional dimension; to improve social support. Kidz with Leukemia: A Space Adventure provided a short video clip on “siblings views on leukemia”[69], p. 299. Packy and Marlon[70] had a two-player option, in which children needed to help one another to succeed in the game, encouraging children to communicate with others about their diabetes. The interventions used by Hazzard, et al.[65] and Runge, et al.[73] both offered chatroom facilities; Hazzard et al. featured videoconferencing, and Runge et al. enabled email contact with other children. Social networking and peer-to-peer communication opportunities may have affected study outcomes, such as self-efficacy and motivation to change, to a greater or lesser degree, thereby confounding the effect of the MIs that relied solely on animation design elements.

Outcomes may also have been influenced by the extent to which health professionals were engaged in the MI. Clinicians were more involved in the MI delivery in three studies [6, 9, 73], although clinicians were only consistently involved with each child (and their family) in the Eckler [9] study; orthodontists actively using the MI to deliver information within the consultation. In contrast, Redsell, et al.[6] mentions a ‘passport’ given to children to indicate the MI modules they had worked through and record comments for later discussion with the school nurse, but omits to report on whether this resulted in children and school nurses discussing the MI. Similarly, asthma experts (not the children’s health professionals) in Runge, et al.[73] were available online for scheduled chatroom chats and via email but it is unclear the extent to which these facilities were used. In an attempt to prevent performance bias, two studies [7, 72] aimed to avert child-clinician communication about the intervention through blinding. Bartholomew, et al.[67] and Bartholomew, et al.[68] actively encouraged physicians to become more involved in the care of children using the MI. Researchers encountered a general resistance, these health professionals cited lack of time [68] and the MI content not suited to their institution’s clinical procedures [69] as barriers to engaging with the MI.

In addition to the MI, some studies provided additional information to intervention group children. Children in nine of the included studies [8, 9, 66‐71, 74] were given the MI only. Others were given the MI as well as education sessions [73] or more frequent face-to-face consultations [6, 65, 72]. One study [7] provided written information as well as periodic contact with a health professional. Other papers failed to clarify whether or not health information was given to children in the comparison group [68] or what form this may have taken [9]. Inequality in study intervention approaches hinders comparison of results. One might argue that these anomalies could be accounted for by the choice of comparison treatment.

A number of MIs [6, 9, 66, 67, 74, 75] were designed for use at outpatient clinic settings. There was variation within these studies; one researcher demonstrated the MI [9], others offered researcher assistance, as required [66, 74], and another did not state the level of support available [6]. Five studies required children to work independently, at home [8, 69, 70, 72, 73]. Again, not all had a similar experience; Connelly, et al.[72] posted the MI to the child’s home, while Mcpherson, et al.[8] and Ritterband, et al.[71] visited children in their homes to offer initial support. Some children used the MI in the nurse’s office at school [68] and others during a period of hospitalisation [65]. The setting and level of researcher involvement will have affected children’s experiences of the MI, particularly for younger or less confident children, thereby creating discrepancies between study findings.

There were significant differences of MI use, both within and between studies. Some studies [6, 9, 72] controlled the order in which multimedia modules were viewed, or indeed offered a limited time in which children were able to use the MI [74]. Other studies [65‐67, 69‐71, 73] allowed children freedom over how they chose to use the MI, if at all. Consequently, reported MI usage also varied greatly. Some children [9, 74] had a total of 30-minutes or less, and others [70] were reported to have used the interventions for up to 34-hours.

The study period for half of the interventions were for less than six-months; one [9] was used within a half-hour consultation, another [65] spanned three-days, while the majority of shorter studies were for three [69, 74] or four-months [6]. Of those lasting 6-months or more, 3 [8, 70, 73] ended at 6-months, 2 [7, 68] were for 12-months and one study [67] had a mean duration of 7.6-months, with participant involvement ranging from 4 to 15.6-months. Most (10 of the 14) studies had no follow-up period (Table 4). Connelly, et al.[72] tracked outcomes for three-months post-intervention and two studies [6, 73] had a six-month follow-up period. One school-based study [68] continued to collect data for two years post-intervention.

Variations in intervention usage, study duration and follow-up period all would have influenced the opportunities children had to familiarise themselves with the multimedia content, and subsequently make behavioural changes.

Prior to exploring the potential health outcomes, the measures used to determine these outcomes should be examined. Of the 14 included studies, all but 2 [67, 68] use the same measures for knowledge acquisition, changes in self-efficacy and health outcomes. Dragone, et al.[69] and Mcpherson, et al.[8] derive their self-efficacy ‘Health Locus of Control’ measure from the same source (Table 6). Measures used to evaluate knowledge gain and differences in self-efficacy were validated prior to their use in most studies reporting these outcomes. There was less certainty over the reliability of instruments used to measure changes in specified health conditions, or specific health symptoms such as HbA1c (glycated haemoglobin, an indicator of long term blood glucose control in diabetes) or FEV₁ (forced expiratory volume, used as a measure of lung function in asthma management). With the exception of Bartholomew, et al.[67], Bartholomew, et al.[68], and Connelly, et al.[72], the included papers neglected to provide adequate information about the instruments used to measure health outcomes. In one study [70], children were recruited from two paediatric diabetic clinics that sent blood samples to different pathology laboratories, each using different HbA1c normative reference ranges.

A number of health outcomes depended on parent recall. These included episodes of emergency medical care and hospitalisation for the chronic condition under investigation. Timescales for recollecting these episodes of emergency care varied from one-month [8] to three-months [70], six-months [73], nine-months [7] and even 12-months [67]. Memory is inherently unreliable. Only Runge, et al.[73] attempted to validate the reliability of the data obtained through parent recall, by using GP electronic records.

Despite the diversity of the research presented in these papers, and their associated methodological differences and limitations, there is value in reflecting on individual study outcomes.

The stated aim of this review was to evaluate the effectiveness of MIs in promoting the quality of face-to-face communication between health professionals and children, within a health context. One paper [9] aimed to facilitate child-clinician communication. In this study, orthodontists used a MI to clarify treatment options with children and parents. Children who viewed the MI felt more involved in the consultation (p < 0.05) but did not report to feel more understood by the orthodontist (p = ns). Another study considered the effect of a MI on a different type of face-to-face interaction; Brown, et al. [70] considered its effect on children talking to parents about diabetes. The study found the effect to be statistically significant (p = 0.0025).

Of further interest, was the extent to which MIs affected the child’s (and family’s) involvement in their healthcare or treatment. Outcomes connected to this aim are attendance rates and relative satisfaction with the MI. However, for both outcomes, minimal comparison data were available.

Three studies reported on appointment attendance. Bartholomew, et al.[67] cited significantly greater attendance rates among intervention group children. In contrast, Homer, et al.[66] found no significant difference between the intervention and comparison groups in this regard. However, data from Redsell, et al.[6] are less clear. The paper reports no difference in attendance rates. Isolating the data of children who failed to attend two consecutive appointments; 26 (24.1%) of the intervention group versus 12 (13.8%) of the comparison group; implied better appointment attendance amongst children who did not use the MI [6]. This leads to an inconclusive assessment of the affect of MIs on attendance rates.

Another proxy for evaluating the quality of communication should be satisfaction with care. However, data were extracted from only three papers [9, 69, 70]. Outcomes from several of the included studies [7, 8, 65, 67, 71] were discounted because of a lack of comparison group data. Davis, et al.[74] designed a cross-over trial, having no comparison group for this outcome measure. Homer, et al.[66] reported no significant group differences for parental satisfaction on physician communication, physician involvement or overall care. The study reported that healthcare professionals were not engaged in the MI and neither were parents; with some parents choosing to sit in a separate room, away from where their child was using the MI, potentially explaining this finding. The ‘not significant’ outcome of Brown, et al.[70] could be interpreted as the MI was no more engaging than the alternative treatment. In this case however, the comparison group received a non-health-related MI. These data could therefore equally be interpreted; that the health-related MI was as enjoyable as the entertainment game. In fact, children in this study used the MI for an average of 18-hours in the first 3-months and 16-hours in the second 3-months. Likewise, Eckler [9] found no significant difference in preference for either the MI or traditional visual aids–dental models, radiographs and photographs–all of which could conceivably be interesting and informative to children within a consultation-setting. Data from Dragone, et al.[69] implied that the MI was equally as useable as the book, but potentially more engaging (as more children used the MI more often and for longer).

Of the three studies, Dragone, et al.[69] was rated ‘high’ quality and at lower risk of bias. However, it would be unreasonable to judge the extent to which MIs encourage 7 to 11-year-old child engagement with health-related treatment, based on 17 children.

The findings for the effect of MIs on knowledge gained are mixed. Six MIs [7, 65, 67, 70, 71] were found to be no more helpful than comparative interventions. The other six MIs were found to be effective [9, 69] or very effective [8, 66, 68, 74] education devices. Three [8, 69, 74] of these studies had a comparatively more robust research strategy, lending more weight to the suggestion that multimedia may be a worthwhile health education tool. Moreover, equipped with greater knowledge about a health condition, and it’s treatment, may enable children to understand health-related conversation in a more meaningful way, thereby facilitating their future involvement in consultations-settings with health professionals [2].

Apart from Hazzard, et al.[65] and Homer, et al.[66], all studies reporting on self-efficacy, presented data pointing towards a significant link between MIs and enhanced self-belief [6, 8, 67, 70], a number of which were highly significant [68, 69, 74]. Although intervention children in the Eckler [9] study were no more confident about their treatment than their counterparts in the comparison group, they were sure of the problems associated with the proposed orthodontic treatment. These data echo those reported for knowledge acquisition.

Self-belief is affected by the time spent learning and practising new behaviours, as confirmed by Bartholomew, et al.[67] who found a correlation between time using the MI, and self-efficacy. Children using the STARBRIGHT: Fitting Cystic Fibrosis into your Everyday Life MI, only used it for 30 minutes [74], probably not enough time to influence self-belief in a meaningful way. Children in the Mcpherson, et al.[8] study typically used the Asthma Files once, for 60 to 90 minutes.

Care must be taken when quantifying the variable and subjective nature of self-efficacy. However, children who feel more able to manage their health condition have a degree of self-control that may afford them greater self-confidence when communicating with clinicians. Parental involvement in the education activity will likely also play a role in a child’s ability to improve self-management [66].

The litmus test of someone’s ability to cope with, or manage, a health condition is how this translates into physical health and wellbeing. Likewise, the ultimate goal of many of these MIs is to change health behaviour and improve disease-related symptoms.

Of the four studies [7, 67, 71, 72] that described MIs as having made a significant improvement to symptoms; three [7, 67, 72] demonstrated inconsistencies. Krishna, et al.[7] found a correlation between use of the IMPACT MI and fewer days of asthma symptoms. However, this outcome was not confirmed by an improvement in lung function (FEV₁). Bartholomew, et al.[67] reported fewer asthma symptoms, but only in children with mild asthma. Connelly, et al.[72] too had mixed results. The Headache Index Composite score (derived from data collected through headache diaries) was lower in children who used the MI. However, the Pediatric Migraine Disability Assessment questionnaire showed no significant difference between the intervention and comparison group children. The authors pointed out that, following the one-month intervention, children were tracked only for a further three-months but their data points towards a gradual decline in Migraine Disability scores amongst children who used the MI.

One of the major limitations for many of the identified studies [7, 67, 71] was the brief length of post-intervention follow-up. Despite these, some MIs have played some role in improving health outcomes for some children.

The evidence analysed thus far suggest that MIs are less beneficial for severe medical conditions. Based on this assumption, emergency medical care and hospitalisation rates would not benefit from MIs. Data presented in Table 6 confirms that only one [7] out of the handful studies reporting on this area, found MIs to be effective to this end.