Systematic media review: A novel method to assess mass-trauma epidemiology in absence of databases—A pilot-study in Rwanda

Objective Surge capacity refers to preparedness of health systems to face sudden patient inflows, such as mass-casualty incidents (MCI). To strengthen surge capacity, it is essential to understand MCI epidemiology, which is poorly studied in low- and middle-income countries lacking trauma databases. We propose a novel approach, the “systematic media review”, to analyze mass-trauma epidemiology; here piloted in Rwanda. Methods A systematic media review of non-academic publications of MCIs in Rwanda between January 1st, 2010, and September 1st, 2020 was conducted using NexisUni, an academic database for news, business, and legal sources previously used in sociolegal research. All articles identified by the search strategy were screened using eligibility criteria. Data were extracted in a RedCap form and analyzed using descriptive statistics. Findings Of 3187 articles identified, 247 met inclusion criteria. In total, 117 MCIs were described, of which 73 (62.4%) were road-traffic accidents, 23 (19.7%) natural hazards, 20 (17.1%) acts of violence/terrorism, and 1 (0.09%) boat collision. Of Rwanda’s 30 Districts, 29 were affected by mass-trauma, with the rural Western province most frequently affected. Road-traffic accidents was the leading MCI until 2017 when natural hazards became most common. The median number of injured persons per event was 11 (IQR 5–18), and median on-site deaths was 2 (IQR 1–6); with natural hazards having the highest median deaths (6 [IQR 2–18]). Conclusion In Rwanda, MCIs have decreased, although landslides/floods are increasing, preventing a decrease in trauma-related mortality. By training journalists in “mass-casualty reporting”, the potential of the “systematic media review” could be further enhanced, as a way to collect MCI data in settings without databases.

Surge capacity refers to preparedness of health systems to face sudden patient inflows, 29 such as mass-casualty incidents (MCI). To strengthen surge capacity, it is essential to 30 understand MCI epidemiology, which is poorly studied in low-and middle-income countries 31 lacking trauma databases. We propose a novel approach, the "systematic media review", to 32 analyze mass-trauma epidemiology; here piloted in Rwanda In Rwanda, MCIs have decreased, although landslides/floods are increasing, preventing a 51 decrease in trauma-related mortality. By training journalists in "mass-casualty reporting", the 52 solutions to triage and localize patients in real-time to specific MCIs and geographical areas 80 has been proposed, although, this has not yet been systematically implemented (7,8). In high-81 income countries (HICs), surge capacity is frequently monitored and MCI protocols are 82 developed using data collected in local databases (9-12). However, these protocols have 83 limited applicability in LMICs due to vast differences in preconditions, such as the disproportion 84 of the global physician density, which is 0.14/1000 people in Rwanda, as compared to 85 4.25/1000 in Germany or 2.61/1000 in the U.S (13). 86 87 Rwanda is a small but densely populated low-income nation in East Africa, with hilly 88 topography and high annual precipitations occurring primarily in two annual "rainy seasons". 89 The Ministry of Emergency Management compiles yearly reports on disasters (14) and 90 provides detailed national contingency plans (15-17). Since 2011, an injury registry has been 91 implemented at the two university teaching hospitals (18), however, a holistic understanding 92 of MCI epidemiology is hindered by data being limited to certain types of disasters and lack of 93 information on MCIs recorded at tertiary, provincial, or district hospitals. 94 95 These obstacles led us to look for new, feasible ways to estimate the burden of MCI in limited-96 resource settings. This study, therefore, aims to pilot the novel methodology "systematic media 97 review" in Rwanda, to identify mass-trauma events and assess their epidemiology using media 98 as the data source. 99 100

101
Defining mass-trauma / mass-casualty incidents 102 While "disaster" in the context of health care describes situations where available resources 103 are not enough to provide adequate patient care (19), "mass-trauma" and "mass-casualty 104 incidents" lack globally accepted definitions. Algorithms have been proposed to quantify the 105 maximum number of casualties that a hospital can manage at the same time, such as the 106 Hospital Acute Care Surge Capacity (HACSC) (20) but, have not been validated in more 107 resource-limited contexts. In this study, we defined MCIs as events causing three or more 108 injuries, as this was expected to be the lower threshold for when health facilities in Rwanda 109 may need to utilize "surge capacity" to meet the patient surge. Inclusion criteria were news articles, radio/news transcripts, and governmental/non-117 governmental reports published between January 1st, 2010 and September 1st, 2020, which 118 reported traumatic events that occurred in Rwanda and caused three or more injuries or, in 119 the case of missing injury data, where the reported number of on-site deaths ≥1 and the trauma 120 mechanism suggested the possibility of three or more injuries. 121

122
The exclusion criteria were languages other than English, French, or Kinyarwanda; 123 occurrences before January 1 st , 2010 or after September 1 st , 2020; explicit mention of the total 124 number of injuries or deaths being less than three; non-traumatic mass-casualty events; 125 location of MCI solely outside of Rwanda; no mention of a specific event (e.g. descriptions of 126 annual trends) or no mention of time and location. Articles were split by two study members 127 (LV and MD) who completed eligibility screening, and data were extracted through RedCap 128 (Appendix 2). 129 130 Data analysis 131 In cases where different articles described an MCI with the same trauma mechanism, 132 province, and date (+/-one day), this was considered as the same event, unless there was 133 information that indicated these were distinct MCIs. Articles describing the same MCI were 134 merged and data was cleaned as following: in cases of discrepancies regarding the number 135 of injuries, the largest number was chosen to avoid missing any data or on-site deaths; in case 136 of discrepancies in terms of which districts were affected, the largest number of districts was 137 chosen. Epidemiological patterns including the number of MCIs, number of injuries and on-138 site deaths, and temporal and geographical trends in MCI incidence, type, and number of 139 injuries and deaths were analyzed using descriptive statistics and Fisher's exact test using 140

184
In Kigali, acts of violence/terrorism were the most common MCI, although reported as rare, 185 isolated terrorist-related shootings and grenade blasts, where 75% occurred between 2011-186 2013, and only one act of violence/terrorism-related MCI occurred between 2017-2020. In the 187 other regions, road-traffic accidents were most common, except for the Western province 188 where road-traffic accidents and natural hazards were equally common ( Table 1) adjusted life years, is decreasing in Rwanda and globally (26,27). In this study, we similarly 218 found that mass-trauma events and the associated injuries have decreased in Rwanda during 219 the past 10 years. Yet, mortality in mass-trauma events did not improve during the study 220 period, which may be explained by the relative increase in natural hazards that tend to be 221 associated with high victim tolls (28,29). Increased frequency and intensity of natural hazards 222 is a known effect of climate change (43,44). Countries with weak health infrastructure are 223 especially vulnerable (34) and it is therefore imperative that developing trauma systems and 224 national surgical / disaster plans also take into consideration changing meteorologic patterns 225 (45,46). According to recent data by the Ministry in Charge of Emergency Management, 226 landslides and floods are the predominant types of natural hazards in Rwanda (14,15), when 227 excluding lightning which is not typically included in the four main categories of natural 228 disasters: hydrological, meteorological, climatological, and geophysical disasters (30). In our 229 study, natural hazards, including lightning, was also the trauma mechanism with the largest 230 number of on-site deaths, and the second highest median number of people injured, when 231 excluding the single boat-accident categorized as "other". Although there is limited data on the 232 direct impact of floods (31), floods lead to increased injury and mortality compared to 233 landslides (32,33), which aligns with our findings. In addition to the consequences on morbidity 234 and mortality, landslides and floods cause large socioeconomic costs, as houses and 235 workplaces may be destroyed, but these aspects were outside the aims of this study (34). 236 237 Road-traffic accidents accounted for the bulk of mass-trauma in Rwanda until 2017, which 238 resembles patterns reported by high-income countries (9,35) and the neighboring country 239 Tanzania (36). Globally, road-traffic accidents cause approximately 1.35 million deaths 240 annually and between 20-50 million non-fatal injuries, with 93% of fatalities coming from LMICs 241 (37). The African Region counts 40% more road deaths per 100,000 people compared to other 242 LMICs, and according to the World Health Organization, most injuries and deaths involve 243 vulnerable road users, such as pedestrians, cyclists, and motorcyclists (37). This contrasts 244 with our findings where car and bus passengers were most affected, and pedestrians and 245 bicyclists were involved in 15.1% and 4.1% of MCIs, respectively. This is lower than a study 246 from Kigali where pedestrians were involved in 35% of accidents in 2012-2016 (38) This study has multiple limitations. Firstly, the term "injuries" can range from very mild injuries 290 not needing healthcare to severe injuries leading to death, which makes comparisons between 291 events difficult. Similarly, the threshold for what constitutes mass-casualty incidents in Rwanda 292 has not been previously determined and therefore, our definition was arbitrary. A previous 293 study of surge capacity at provincial/tertiary hospitals in the neighboring country Tanzania 294 used a fixed number of 10 for a road traffic accident to be considered a MCI (36). Therefore, 295 the threshold of three may be appropriate in rural areas of Rwanda, although it should possibly 296 be adjusted upwards in urban areas. Thirdly, in cases where multiple articles described the same event, there were sometimes 307 discrepancies in the number of people injured / deaths reported. This was particularly common 308 for natural hazards, which, unlike road traffic accidents, are not always clearly limited in time 309 and exact location. Although some demographic data on MCI victims were provided, the 310 relative lack of details could make it difficult to use this method to track clinical outcomes. 311 312

313
The systematic media review can be used to assess mass-trauma epidemiology in contexts 314 where systematic data collection on MCIs is limited. In Rwanda, the number of MCIs has 315 decreased, although landslides/floods are increasing, hindering a decrease in mortality. MCI 316 protocols in Rwanda should put an emphasis on rural areas, where most natural hazards 317 occurred, and include modified referral protocols for critical patients. To improve the quality of 318 the systematic media review, there is potential in training journalists in "mass-casualty 319 reporting". Further studies will pair this novel method with clinical data collection to validate 320 the method and to give a more granular understanding of the epidemiology of MCIs in Rwanda (8)