1) To identify categories of burn injuries in paediatric patients.

2) What are the etiological sub-categories for burn injuries in paediatric patients with an emphasis on food or cooking associated items?

3) What patient outcomes are reported for burn injuries in paediatric patients?

A systematic search for relevant articles was performed by two independent researchers and a librarian. Databases searched were Cochrane, PubMed, CINAHL, Google Scholar and ProQuest. Medical subject headings and free-text terms of child* or paediatric/pediatric* and burn*or scald*, and noodles or ramen were combined with the search filter all types of study methodologies. Titles and abstracts were screened by two of three assigned reviewers with studies excluded if they did not meet the inclusion criteria. The full manuscripts of potentially eligible studies were then reviewed by two reviewers for inclusion. Any disagreement was resolved through discussion and by a third reviewer if necessary. Searches were conducted over a ten-year period from January 2009 to December 2019, to assist with identifying any trends in line with the reported popularity of instant foods ([2]).

Inclusion criteria included empirical research related to burns in children aged 0 - 18 years (Table 1), although records with a paediatric and adult population were included if paediatric data were able to be extracted. Exclusion Criteria included adult studies, intervention and outcomes that did not relate to burn injuries in paediatric patients. To ensure recency, publications where data were collected prior to 2005 were excluded, as were grey literature or print media, electronic data, theses, or government records where publishing is non-commercial and not the primary activity ([2]; [36]; [48]).

Figure 1.PRISMA flow diagram.

To test the reliability of search terms, a preliminary search was undertaken by a health librarian trained in database and literature searching. The search was repeated by the authors and extended across the five databases. The initial search resulted in 8035 records including one obtained from hand-searching as shown in Figure 1. Following the removal of duplicates (n = 35) and initial screening of titles and abstracts 73 reports were retained for full-text review by both authors. Eleven publications were not available in full-text or the English language, therefore 62 reports were assessed for eligibility and 16 publications retained in this review.

A data extraction tool was purposely developed by the project team using Microsoft Excel. A process of data extraction of all included studies was conducted with author, country, methods, intervention, and outcomes incorporated. Data to be extracted were entered into the quality assessment tool and reviewed by two reviewers independently, with each entry cross-checked and inconsistencies discussed until resolved. In the case of multiple publications of a study, these were highlighted and grouped. Authors of the articles were contacted if needed for clarifications regarding additional unpublished data. Extracted data included study and participant characteristics; burn injury incidence, aetiology, and descriptors; and hospital and patient-related outcomes. No assumptions were made about missing data.

Burn injuries were categorised as either a burn or a scald depending on a dry or wet causal agent ([50]) and divided into 12 aetiological sub-categories, with an emphasis on food or cooking-associated items. For example: hot liquids, soup or stew, noodles, and fire.

All the studies included were then reviewed for quality assessment as there were a variety of methods reported. The 43-item Epidemiology Appraisal Instrument (EAI) ([22]; [23]) was used to assess the validity and quality of included publications. Each item was scored from 0 - 2 with 0 for not described; 1 for unclear, and 2 if the item was clearly addressed in the publication. Items considered not applicable were not scored. Following the recommendation by Genaidy et al. ([27]), the mean score was calculated for each study. Publications with mean scores of 1.4 or greater were described as ‘good’; a mean score of at least 1.1 was considered ‘average’ and studies with a mean score of 1.0 or less ‘poor’. The instrument was independently applied to all articles included in the review by the researchers and any differences in scoring or reasoning were discussed until consensus was achieved. The mean score of studies was 1.6 (SD 0.2) therefore all articles were retained in the review (Table 1).

Statistical Database IBM SPSS v27 with alpha set at 0.05 was used for analyses. Descriptive statistics were used to calculate the incidence of burn injury by category and sub-categories, and student’s t-test (2-tailed) was used to compare the incidence of burns and scalds.

Sixteen publications were included in this review (Table 2). The publications described data originating in the US (n = 4), Europe and the UK (n = 4), Middle East (n = 3) Australasia (n = 3) and Africa (n = 2). Most studies (n = 15) aimed to explore the epidemiology, incidence, demographics, and characteristics of children who had sustained a burn or scald. Study designs were primarily retrospective and epidemiological (n = 14) with a focus on burn (n = 12) or scald (n = 4) incidence ([28]; [29]; [30]), demographics ([32]; [33]; [35]; [36]; [37]), aetiology and risk factors ([30]; [32]; [33]; [35]; [36]; [38]; [39]; [40]; [41]), injuries and outcomes ([32]; [33]; [35]; [37]; [39]; [40]; [41]; [42]; [43]; [44]; [45]). Two studies addressed burns or scalds secondary to cooking ([43]) with one focusing on instant noodles ([44]).

Table 2. Study designs and sample characteristics.

EAI: Epidemiological Appraisal Instrument.

Most studies (n = 8, 50%) included children aged 0 to 18 years, four studies focused on children to 5 years ([30]; [32]; [33]; [36]; [43]) and one study included children to 12 years ([28]). The calculated mean age was 8.8 years (n = 10, SD 4.16; median 3.05). A slightly higher incidence (56%) of burns was reported in male children (n = 14) although one small study reported a higher ratio of females in children aged under one year ([43]).

Data were collected over 3.6 years on average (SD 2.88, range 1 - 12), had a mean sample size of 47,504 (median 339) and focused on hospital emergency presentations (n = 11) or hospital admissions (n = 14) (Table 2). Admission rates ranged from 2.9% (Finland) ([43]); 4.3% - 14.6% (US) ([32]; [40]); 12.7% (Australia) ([5]; [39]); 26.8% (Switzerland) ([35]), to 86.1% (Iran) ([37]). Two studies addressed paediatric burn injuries treated in the outpatient setting ([36]; [39]).

Hospital admission rates implied that many burn injuries are severe (10 studies with an aggregate sample population of 759,996, 10 of the 15 articles described admission numbers with a total admitted population of 36,881). The severity of burn injuries in terms of size and depth was variously assessed as a percentage of total body surface area (TBSA) (n = 10) in 4 studies ([28]; [31]; [38]; [42]); mean TBSA (n = 8); by the Lund and Browder chart ([37]) and by mean centimetres ([28]). Burn injury depth was described by degrees ([28]; [38]; [41]) or by descriptors such as “deeper than superficial” ([39]), “deep partial thickness” ([30]), “mid depth” ([31]) and “full thickness” ([44]).

Numerous categories were reported in 11 studies and used to describe agents responsible for paediatric burns (n = 43) in line with research aims and designs. The most common being scald (n = 11), flame (n = 6) and contact burns (n = 4). The incidence of burns to scalds is conflicting. Most studies (n = 10) reported a higher prevalence of scalds, three studies reported a higher prevalence of burns ([29]; [39]; [40]) and yet the overall incidence of injuries from wet agents were fewer (n = 244,127) than that arising from dry agents (n = 368,112) (95% CI −15,941.6, 46,457.3 and −32559.7, 89192.2 respectively). No correlation was found between burn and scald incidence and high or low-income countries.

Multiple agents (n = 64) were identified as the cause of paediatric burns with hot water calculated to be implicated in 71.5% (n = 55,410) of scald injuries. Hot water was identified as a sole entity (n = 11); probable constituent of hot liquid, beverages and/or steam ([31]; [37]; [41]) in cooking and boiled food ([28]; [35]; [41]) bathing and hair braiding ([41]).

Food-related burn injuries were categorised as cooking or non-cooking ([32]), by the cooking apparatus e.g. oven door, barbeque or fireplace ([35]; [36]), microwave or stove top ([42]) or an associated action such as grabbing, pulling, touching or spilling food items ([32]; [36]; [41]; [43]).

Table 3. Burn and scald reported aetiologies and outcomes.

NR: not reported; SD: Standard Deviation; IQR: Inter-quartile range.

Identified food items were fats, oils or grease ([33]; [34]; [38]; [41]), soup and stew ([33]; [34]; [35]), porridge ([35]; [38]), rice ([35]) and noodles ([39]; [41]; [42]). Foods were also grouped together such as noodles and broth ([41]), or were not identified, limiting comparisons. Koltz et al. reported that the cooking method for noodles is a risk, with microwave cooking posing a higher risk for older children whereas younger children were at risk irrespective of whether the microwave or stove top were used ([42]) (Table 3).

Most burns occurred in the kitchen (79.7%) but could also occur at outside recreational venues with the latter associated with barbeques, fire pits and fireworks ([35]; [37]). No specific foods were identified for burns associated with outside venues.

The extensive search for empirical data was a strength of this study despite the low representation of articles for some countries. Grey literature was excluded meaning that some published data may not have been captured. It is recognised that findings related to burn incidence have likely been affected by studies with small sample sizes or missing data and therefore should be interpreted with caution. However, the global approach has ensured a comprehensive review that has demonstrated the need for definitional consistency with the coding and description of burns and burn injuries, to enhance research rigor and for health promotion purposes.

Considering the increasing popularity of instant soups in some Asian countries and the US ([20]), and anecdotal evidence from a public hospital in a culturally diverse region of Sydney, Australia, implying that paediatric burns associated with instant soups and noodles were rising, a scoping review of published literatures was undertaken to identify if specific foods posed a greater risk of paediatric burns.

In order to enhance research rigor, there is an urgent need for a universally recognised burn categorisation system.Early implications are that the cooking method and resultant change in foods in terms of viscosity, liquidity and temperature are key risk factors for burn injuries in children.These research findings can be valuable for health professionals who want to improve the paediatric burns associated with food and cooking methods.