Solution-oriented research for paediatric burn prevention in Mongolia: An assessment of prevention tools.

Child burn injuries in Mongolia are often caused by electric cooking appliances used on the floor or low table in traditional tent-like dwellings (called a ger) which have no separate kitchen. To prevent these injuries, we developed a context-specific kitchen rack to make electric appliances inaccessible to children, and the rack was provided to 50 families with children aged 0-3 years living in gers for a pilot test. In the present study, we investigated their opinions about the rack after they used it for about 10 months through semi-structured interviews, their willingness-to-pay (WTP) for the rack using a contingent valuation method, and their preference for potential modifications of the rack using best-worst scaling. The estimated median WTP was about USD 40 (which was higher than USD 37 at the baseline when they started to use the rack). The highest priority of modifications of the rack was to enclose the lower section of the rack with doors (which was originally open without doors to reduce the production cost). A few families did not use the rack in winter because they used heating stoves instead of electric appliances for cooking, but we found a unanimous view that the rack reduces burn injuries to children, which may be reflected in their increased WTP for the rack. These findings would guide us to make our burn prevention efforts more relevant to real-life situations and socially acceptable in Mongolia.

Introduction

Globally, a large number of child burn injuries occur disproportionately across countries with different income levels. Child burn mortality rate in low-, middle-, and high-income countries is estimated to be 4.5, 1.4, and 0.6 per 100,000 children aged 1–14 years, respectively [1]. The rate in Mongolia (8.1) is much higher than these rates [1]. Every year, about 1200 children (or 3 per 1000 children) aged <5 years are admitted for burn injuries to the National Trauma and Orthopedic Research Center, the only hospital providing tertiary care for burn injuries in Mongolia [2]. Our previous hospital-based survey revealed that many infants and toddlers were scalded severely due to electric pots and kettles that are commonly used on the floor or low tables in the traditional tent-like dwelling, called a ger [3].

To prevent child burn injuries inflicted by electric pots and kettles, we previously developed a context-specific steel kitchen rack to make these electric appliances inaccessible to children (Fig. 1). The rack is 75 cm high with the top section enclosed with a grid where electric appliances are placed. The lower section is open space for storage. In the pilot survey, we distributed the rack to 50 households with an infant or toddler to examine their acceptance and willingness-to-pay for the rack and get their feedback on the rack after distribution [4]. Generally, the survey participants appreciated enhanced safety and reduced anxiety about burn risk to children by using the rack, but they also reported that children entered the lower section of the rack to play, the corners/edge of the rack might hurt children if they collide with the rack, and the plug of some electric appliances cannot pass through the grid of the top section, which they described as inconvenient.

Fig. 1

Kitchen rack used in a Monglian tent-like dwelling.

Based on the feedback, we modified the rack visually (Fig. 2), though we might not be able to modify all elements (attributes) of modifications in real production of the rack. So, we decided to investigate the relative importance of the attributes of the rack to identify priority of modifications. We were also interested to examine whether their willingness-to-pay (WTP) had changed after they used the rack for about 10 months. The aim of this follow-up study was therefore to estimate their WTP for the original kitchen rack and to determine the relative importance of the attributes of the rack they considered for further modifications. We also explored their views and opinions about the rack through qualitative data analysis. In this way, this paper described the process of solution-oriented research for child burn prevention.

Fig. 2

A sample choice set of original and modified kitchen racks for best–worst scaling.

Methods

Participants

Participants were 50 families living in gers with a child aged 0–3 years to whom we provided the kitchen rack between May and June 2020. The detail procedure of recruiting the participants was described in our previous study [4], but briefly, the families were randomly recruited from the list of eligible families in five sub-districts within the district called Songinokhairkhan in Ulaanbaatar, the capital of Mongolia. The study district and sub-districts were purposively selected based on the size of population living in ger areas. We obtained the list of eligible families from the administrative offices of the selected sub-districts and contacted each family by phone. About 30% of the families we tried to contact were unavailable after we tried to contact them 3 times, and 14 families we contacted appeared to be non-users of an electric pot. They were replaced with additional families that were further randomly selected from the list. We recruited a total of 50 families made up of 10 families in each of five sub-districts. In May and June 2020, trained research assistants visited the households to install the kitchen rack and conducted structured interviews with mothers as the lead person for cooking. The mothers were invited to participate in group discussions about the use of the rack after using it for 4–6 weeks. Trained research assistants conducted five sets of group discussions using a semi-structured interview guide, and a total of 36 mothers participated.

In the follow-up survey conducted in March 2021, we found that 4 families temporarily moved to rural areas because of the COVID-19 pandemic, while 1 family could not be contacted by phone and was found to have moved out, and 1 family refused to participate in the follow-up survey. Consequently, 44 families were included in the follow-up survey. Table 1 shows the baseline characteristics of the 44 families and other 6 families lost to follow-up. Though caution should be exercised when comparing these two groups given the small number of families involved, the families lost to follow-up tended to be single parent (mother only) family with a lower income and older children.

Table 1

Baseline characteristics of 44 families included in the follow-up survey and 6 families lost to follow-up.

	44 families	6 families
Number of children aged <5 years in the family
1	13	3
2	26	3
3	5
Sex of children aged <5 years (n = 89)
Male	42	4
Female	38	5
Age of children aged <5 years (n = 89)
<1 year	19
1 year	23	1
2 years	23	2
3 years	6	3
4 years	9	3
Age of parents, median (range)
Father	30 (23−44)	32 (28−39)
Mother	29 (23−44)	25 (18−35)
Number of cohabitants in the family
3 to 4	15	5
5 to 6	22	1
>6	7
Median (range)	5 (3−10)	4 (3−6)
Single parent (mother only) family	6	2
Monthly income (in MNT)a
500,000 or lower	4	3
500,001–900,000	24	2
900,001 –1,600,000	15	1
1,600,001 or above	1

MNT is a Mongolian currency (Tugrik). MNT 100,000 is equivalent to USD 35.

Data collection

Trained research assistants visited the participants’ households in March 2021 for data collection. To elicit their opinions about the kitchen rack, semi-structured interviews were conducted and audio-recorded. To measure their willingness-to-pay and preference for potential modifications of the rack, structured questionnaires were used with measures as explained below.

Measures

We used a contingent valuation method with a double-bounded dichotomous choice format to estimate participants’ willingness-to-pay for the original kitchen rack [5,6]. This method and format were used in the baseline survey [4], where participants were presented with one of predetermined four potential prices for the rack (MNT 30,000, 50,000, 70,000, and 90,000; MNT 100,000 is about USD 35) and asked whether they are willing to pay that price. If yes, they were asked whether they are willing to pay double of the price; if no, half of the price. Their responses formed contingent valuation data.

We used best–worst scaling to determine the relative importance of the attributes or potential modifications of the kitchen rack [5,7,8]. In best–worst scaling, respondents are presented with a set of attributes to select the best and the worst attributes, and this is repeated a certain number of times with different sets of attributes. In our previous survey, the participants suggested the need for modifications to the lower section of the rack, the corners/edge of the rack, and the grid size of the top section of the rack. So, we created several images of the modified rack, in which the lower section was enclosed with/without doors, the corners were made rounded, and opening was made in the top section for the flex of electric appliances (Fig. 2).

Considering the cost reduction of the modified rack, we created six patterns of the lower section (three sides solidly enclosed, with/without doors on the 4th side; enclosed with a grid, with/without doors; enclosed with bars, with/without doors), supposing that the cost will be cheaper if the lower section is enclosed with a grid or bars instead of being solidly enclosed. Consequently, nine versions of the rack were formulated: 1) the original rack,2–7) the original rack with six patterns of the lower section, namely “solid enclosure with/without doors”, “grid enclosure with/without doors”, and “bar enclosure with/without doors”, 8) the original rack with “rounded corners”, and 9) the original rack with “flex opening”.

Based on nine versions of the rack, we created 12 sets of six versions in different combination. Through the 12 sets, participants were presented to each version with equal frequency and combination: the number of times each version appeared in the 12 sets was eight times, and each version appeared with each other in the same set five times (e.g., versions 1 and 2 appeared together in 5 of the 12 sets). Such sets were created using a balanced incomplete block design [5]. For each set, participants were requested to select the best version among six versions first, and then the worst version. So, 24 data points per respondent (12 best and 12 worst from the 12 sets). Fig. 2 were collected per sample set.

Besides their willingness-to-pay and preference for potential modifications of the rack, we asked participants about seasonal variations in usage, whether their children had burn injuries during the follow-up period, potential of reducing the risk of injury when using the rack, and the views of visitors whilst the rack was in use. We also asked the current household monthly income because the income level might have changed due to the COVID-19 pandemic.

Analysis

Using contingent valuation data of 44 participants, we estimated a median WTP and its 95% confidence interval (CI), controlling for the household income level at present, previous child burn experiences including experiences during the follow-up, and the number of children in the household, as in the baseline survey. For this estimation, we used the DCchoice package from the Comprehensive R Archive Network [5]. The details of contingent valuation data analysis in our study can be found in our previous study [4].

With the best–worst scaling data from 42 participants with complete responses, we counted the total number of times each of the nine versions of the modified rack was chosen as the best or the worst to find which version was most frequently chosen as the best or the worst. The maximum number of times each version could be chosen (or voted) as the best or the worst was 336 because each version appeared 8 times in 12 sets from which 42 participants chose the best and the worst (i.e., 8 × 42 = 336 votes). To facilitate the interpretation of the results, we calculated a standardized score for each version ranging from −1.0 (worst) to +1.0 (best). The standardized score of the version 1, for example, was the number of times the version 1 was chosen as the best subtracted by the number of times the version 1 was chosen as the worst, that was divided by the maximum number of times the version 1 could be chosen as the best or the worst (i.e., 8 × 42 = 336 votes). For this analysis, we used the support.BWS, crossdes, and dfidx packages from the Comprehensive R Archive Network [5].

For qualitative data analysis, two researchers reviewed the text content of the interviews, one using manual content analysis and the other using NVivo software; both identified key points raised which then compared. We found little difference in the agreed content as the majority of the responses were very short and focused on functional uses of the rack or factual elements of their experiences. Quotations included in the results section have been taken verbatim from the translations into English.

Results

A median willingness-to-pay (WTP) for the original kitchen rack was MNT 115,000 (95% CI: 98,000–134,000) or about USD 40 at the follow-up survey about 10 months after they were provided with the rack; whereas the median WTP was MNT 106,000 (95% CI: 84,000–130,000) or about USD 37 at the baseline survey when they were provided with the rack. Regarding their preference for potential modifications of the kitchen rack, Table 2 shows the total number of times each version of the modified rack was chosen as the best or the worst, with the standardized score of each version. Many participants preferred solid enclosure of the lower section of the rack with doors (261 out of a total of 336 votes for the best), that was followed by flex opening (85 votes for the best). On the other hand, they considered bar enclosure with and without doors least preferable (169 and 109 votes out of a total of 336 votes for the worst, respectively). According to the standardized score or relative importance of 9 versions of the modified rack, solid enclosure with doors was highly valued than other attributes, and flex opening was valued next, while bar and grid enclosure with/without doors appeared to be unpopular.

Table 2

Total number of times each version of the modified kitchen rack was chosen as the best and the worst among 42 respondents, with the standardized score of each version.

	No. of times chosen
Versions	Best	Worst	Standardized scorea
Original rack with solid enclosure with doors	261	3	0.77
Original rack with flex opening	85	3	0.24
Original rack with solid enclosure	54	11	0.13
Original rack with rounded corners	20	19	0.00
Original rack	10	21	−0.03
Original rack with grid enclosure	12	40	−0.08
Original rack with grid enclosure with doors	10	87	−0.23
Original rack with bar enclosure	5	109	−0.31
Original rack with bar enclosure with doors	5	169	−0.49

The standardized score ranges from −1.0 (worst) to +1.0 (best). The score for the version 1, for example, was the number of times the version 1 was chosen as the best subtracted by the number of times the version 1 was chosen as the worst (261−3 = 258), that was divided by the maximum number of times (8 times) the version 1 could be chosen as the best or the worst among 42 participants (8 × 42 = 336), thus 258/336 = 0.77.

In the interview, only two of the families reported a burn incident while the rack was in the ger, but in both cases the rack was not in use. Of these, one was a serious burn with hospitalisation for 20 days with 2 sets of surgery: in this incident the electric pot had been taken off the rack and used on the floor. The second burn occurred when the child ran past the stove (used in winter to heat the ger) and burned their arm; the injury was not serious and was not related to the use of the rack.

Six families had moved the kitchen rack outside of the ger during the winter as they returned to traditional use of the heating stove for cooking; 20 of the families used the rack less frequently in winter due to the convenience of the heating stove, as summarized by this family:

Family 14: “I think it is okay to use in summer. But the frequency of the use will decrease in winter. We use it in summer because we use the electric pot regularly. Our children play and run a lot so that we need to protect them by putting the kitchen rack.”

The families continued to voice their desire for changes, particularly in relation to the lower space in the rack:

Family 12: “It would be better if the lower part of the kitchen rack is covered and has door. So that, children cannot open it when we store things there.”

After continued use for many months the family consistently reported that the rack was easy to use, very convenient and stable. We found unanimous agreement that the rack had the potential to substantially reduce burn injuries in the home:

Family 7: “I think it has reduced. For me, it was convenient as children cannot touch, reach or pull the electric pot. I liked it a lot.”

Family 44: “In the past, the child used to touch and burn arms when cooking with an electric pot. The rack clearly prevents from such episodes. Now there are no such cases.”

Comments from visitors to the ger were consistently positive, with the most frequent comment being that is was ‘convenient (especially for toddlers)’; other comments included: ‘want one /where can I get one’; ‘safe’; ‘beautiful’; ‘easy’; ‘very nice’.

Family 12: “When we had visitor from countryside, they asked us to have it. We explained and refused to give it to them”

Family 41: “There was no negative comments. But visitors who are moms are interested in it. They said it would be easy to handle things and the child would not be burnt”.

We found one point of confusion expressed in many households, which was that they thought that the rack would heat up when placed next to the heating stove. In fact, this will not happen, and the families reported no cases when this did actually happen – only that they feared it would:

Family 43: “Of course, it will be heated if I put it right next to the stove.”

Discussion

We found that the participants’ willingness-to-pay (WTP) for the kitchen rack had increased about 10 months after they received the rack. Though a few participants did not use the rack at the follow-up survey, most of the participants continued to use it, and their understanding of its usefulness should have resulted in the increased WTP for the rack.

Discontinued use of the kitchen rack was anticipated especially in winter (when the follow-up survey was conducted) because the stove is used all day long for heating and can also be used for cooking. It is understandable that those not using electric pots in winter did not use the rack for cooking. Yet, most of the participants continued using electric pots in the rack as well as the stove for cooking in winter, even if it was only to boil a kettle whilst using the heating stove as their main way to cook meals. No one specifically reported that limited space within their dwelling was a reason for not using the rack, but a few participants reported moving the rack out of the house to make space for the heating stove suggesting a desire avoid overcrowding the living space.

What was not anticipated was the occurrence of a child burn injury caused by an electric pot which was used on the floor in one of the participating families, even though they were instructed to use the pot in the rack. This single incident poses a difficult question about how to ensure behavioral change alongside using the pot in the rack. The mother felt she was “still getting used to the rack” at the time of burn injury, caused by 4 children running around with one getting pushed over; she reported that she had been using the rack ever since. The incident is a reminder of the strength of influence of traditional cooking practices and emphasizes the the benefit of a comprehensive, culturally sensitive, implementation programme when introducing such cultural changes.

Regarding the modification of the kitchen rack, solid enclosure of the lower section with doors was the highest priority, while enclosing the lower section with bars or grids was unpopular with less priority than flex opening and rounded corners. This implies that appearance of the lower section is also important. These findings give us a clue to modify the rack to be further accepted.

It is clear that while all participants could clearly see the benefits of the kitchen rack, twe propose that more information be made available (e.g., leaflets, Question and Answer Responses, a wider public health campaign) to support the implementation of the use of the rack should the rack become available on a wider scale. Support in terms of how to use the rack around the year, ideal locations for use within the ger, more information on the fact that the rack is made of materials that do not get hot, and clearer instructions on the general use of the rack could all be helpful. The consistently reported interest from others (visitors) suggests a general interest in the rack and availability for purchase, which supports the unanimous view that the rack reduces burn injuries to children.

Finally, our study limitations should be acknowledged. First, we recruited the participants randomly from a group of eligible participants, but the sample size was small, resulting in less precise estimates (as seen in a relatively wider confidence interval of the estimated WTP). Second, best–worst scaling may be better than traditional rating scales at discriminating the relative importance of the attributes (potential modifications of the kitchen rack) [8], while we could not produce modified kitchen racks to show to the participants to get their full understanding of the modifications. We believe that showing the images of modified racks helped them understand the modifications better than simple descriptions of the modifications without such visual aid. Third, we had methodological challenges in using a contingent valuation method to estimate the WTP in this follow-up survey as in the baseline survey, which was fully explained in our previous report [4]. In the present study, we used the same method to assure the comparability of the estimated WTPs in both surveys.

In conclusion, we found that the participants’ willingness-to-pay for the original kitchen rack had increased after a while and solidly enclosing the lower section of the rack with doors was the top priority among potential modifications of the rack. These findings would guide us in our efforts for child burn prevention in Mongolia. As such, a solution-oriented research is potentially useful to make burn prevention efforts more relevant to real-life situations.