The impact of digital technology on health of populations affected by humanitarian crises: Recent innovations and current gaps

Digital technology is increasingly used in humanitarian action and promises to improve the health and social well-being of populations affected by both acute and protracted crises. We set out to (1) review the current landscape of digital technologies used by humanitarian actors and affected populations, (2) examine their impact on health and well-being of affected populations, and (3) consider the opportunities for and challenges faced by users of these technologies. Through a systematic search of academic databases and reports, we identified 50 digital technologies used by humanitarian actors, and/or populations affected by crises. We organized them according to the stage of the humanitarian cycle that they were used in, and the health outcomes or determinants of health they affected. Digital technologies were found to facilitate communication, coordination, and collection and analysis of data, enabling timely responses in humanitarian contexts. A lack of evaluation of these technologies, a paternalistic approach to their development, and issues of privacy and equity constituted major challenges. We highlight the need to create a space for dialogue between technology designers and populations affected by humanitarian crises.


Introduction
In 2016, 37 countries are facing humanitarian crises. The UN Office for the Coordination of Humanitarian Affairs (UNOCHA) estimates that 88 million people are in need of humanitarian assistance globally. 1 The recent influx of refugees into Europe highlights the political, social, and economic complexities that result in displacement both within and across borders. Traditional and social media coverage of these events has increased awareness of the plight of refugees, and highlighted how digital technologies have become essential tools for humanitarian workers and affected populations alike. 2 Digital technology allows for storage and analysis of large amounts of data using electronic devices. 3 Since the response to the Haiti earthquake in 2010, in which social media, text messages, and satellite imagery played an important role, 4 the use of digital technology in humanitarian action is now viewed as both ''essential and inescapable.'' 5 Such technologies underpin the functions of many humanitarian actors, including data collection, data analysis, information dissemination, coordination, advocacy, and fundraising. Networks of technicians have also been established to provide data analysis and geospatial technical assistance to humanitarian organizations. 6 The potential of digital technologies to support the health and social well-being of populations affected by humanitarian crises lies in their use to better understand complex situations and the needs of the affected communities, facilitating humanitarian response efforts, and engaging the affected populations in the response. 5 Our objectives were to (1) review the current landscape of digital technologies used by humanitarian actors and affected populations, (2) examine their impact on the health and well-being of affected populations, and (3) consider the opportunities for and the challenges faced by users of these technologies.

Methods
We conducted a systematic literature search to identify reports of the use of digital technologies to improve health assistance, health, or its determinants in populations affected by humanitarian crises. These populations were defined as people exposed to natural disasters, war, armed conflicts, or rescue missions, including displaced populations, refugees, forced migrants, or evacuated populations. Digital technologies refer to both hardware and software applications such as (but not limited to) mobile and desktop computers, telemedicine, telehealth, eHealth, and mHealth. We included studies or reports dating from 2005 to 2015 that aimed to improve the health of displaced populations or to facilitate the health work of humanitarian organizations. We excluded news articles, blogs, media reports, theoretical models of technology, and disaster planning documents, monthly and weekly updates from humanitarian agencies, and opinion pieces as well as reports about economic migrants. We also excluded initiatives that focused on shelter, water, sanitation and hygiene improvements, as they generally focused on the engineering of infrastructure, and were more than one degree removed from health.
We searched Medline (our strategy is included in Supplementary Material), Embase, Global Health, PyschInfo, Popline, and the Association for Computing Machinery (ACM) Library for published literature, and used Google to locate reports of projects for displaced populations, digital technology, disaster context, and health. We included peer-reviewed articles, books, book chapters, periodic reports, one-time reports, and websites published between 2005 and 2015. No language restrictions were placed on the search. We imported the articles and reports retrieved into Endnote X7 and removed duplicates. Two researchers conducted title-abstract and full-text screening independently according to the inclusion criteria outlined. Figure 1 presents the resulting review flow chart.

Results
The findings of this landscape review are structured according to an organizational framework illustrated in Figure 2. They are divided into technologies used by humanitarian actors and those used by populations affected by crises. Digital technologies are used in humanitarian programming to facilitate various functions of the preparedness for, response to, and recovery from emergencies, supporting health assistance and health and its determinants. 7 We identified a set of digital technology initiatives used largely by those affected by humanitarian crises and that can impact the various determinants of health including income, education, social support, access to healthcare services and others. 8 Table 1 summarizes the number of articles or reports we found through the systematic search for different types of digital technologies, defines these technologies, and categorizes them by user (humanitarian versus affected population). Table 2 summarizes the technologies used by humanitarian actors and aid workers, the context in which they have been deployed, their purpose, impact, and at what point in the humanitarian program cycle they have been used (preparedness, response, or recovery).

Humanitarian crisis preparedness
Five initiatives were found that use technologies in preparedness, of which geographical information systems (GIS) has been a major game changer. GIS now underpins risk assessments, vulnerability models, and spatial decision support systems (SDSS) that predict outbreak and spread of diseases. 9,10,63,64 Examples include disease surveillance systems that assess disease risk during natural disasters, and emergency preparedness of households with specific health needs. 63,64 These systems have been found to enable continuity of care throughout a disaster. 63,64 Another important element of preparedness is training. Game-based simulation exercises have been used to improve humanitarian responses and reduce evacuation time [See Table 2; mixed reality triage and evacuation (MiRTE)]. 11 The impact of digital technology on health in humanitarian crises SMS (and radio talk, radio and television talk shows) Text messaging service information is given on early warning alerts and local assistance numbers

47
The impact of digital technology on health in humanitarian crises

Oxfam/Post Typhoon in Philippines and in Thailand
Responses can be recorded during interviews on handheld devices with the aim to achieve food security and sustainable income Response Timely and accurate data entry, which allowed for efficient data monitoring as well as tracing accountability while maintaining data security.
Potential for coding errors, heightened security risks for data collectors, weather challenges, and inequalities in technological literacy The impact of digital technology on health in humanitarian crises Ó 2016 The Authors. 0197-5897 Journal of Public Health Policy Vol. 37, S2, S167-S200 S177

Humanitarian crisis preparedness and response
Five technologies linked preparedness to response and were used mostly in the context of natural disasters and conflicts. Social media and geolocation technologies inform decision-making and are used by humanitarian personnel to compile and analyze location data to create evacuation and emergency response models. [12][13][14] Surveillance systems based on social media were used to assess emerging disease risk, such as cholera-related tweets (Twitter) as an indicator of cholera outbreak in refugee camps in Kenya. 13 This innovative data compilation method allows a temporal analysis of such data and, when combined with GIS and SDSS, can enable the prediction of disease outbreaks and their mortality quite accurately. 9,13 Surveillance of web-based fora allows clinics run by NGOs in camps to submit reports of numbers of cases of disease observed, among vulnerable populations. 15 This aids risk assessment that can improve communication between agencies and support decision-making. 16,17 Finally, use of cloud-based electronic health records can successfully recover health records in case of disasters and damage to physically held records and local databases (See Table 2; Japan's Health Management System). 5

Humanitarian crisis response
We identified thirty technological initiatives that primarily serve humanitarian responses. Data entry with mobile devices is now widely used to facilitate the registration of displaced individuals, to conduct surveys, identify those in need of assistance, and to capture data on issues such as food security, vaccination rates, and mortality. (See Table 2; Last Mobile Solutions). [18][19][20] Accurate and timely data entry supported efficient data monitoring and provenance while maintaining data security. Practical challenges, however, include potential for coding errors, heightened security risks for data collection teams, weather challenges, and inequalities in technological literacy. 21 Geographic information systems (GIS) are used to portray areas of conflict through the creation of live crisis maps that document the intensity of conflict and identify the number of people in need of assistance. They are used for demand-based responsive supply for displaced populations and to create warnings (See Table 2; Ushahidi). 5,22-24 GIS can also be used to track population movements, to identify those in high-risk zones in cases of disease outbreaks, and to send text messages containing health information that can be sent to those proximal to an outbreak area. 10,25 Technologies that combine GIS and social media have made it possible for humanitarian organizations to aggregate posts and broadcast information to public officials or to rescue and relief organizations (See Table 2; Artificial Intelligence for Disaster Response Crisis Lex). 18,[26][27][28][29] When humanitarian response is initiated, GIS can evaluate the security, road safety, and effectiveness of vehicles, to inform logistics, coordination, and evacuation. In addition, unmanned aerial vehicles have been used to deliver humanitarian aid in hard to access areas and to map survivors (See Table 2; UN Aid necessities transporter). 30,31 Digital technologies such as the internet, SMS, and social media are also used to deliver training for healthcare providers during crises. [32][33][34] Portable medical devices enable physicians to run tests 32,33 (See Table 2; Cellophone) including eye examinations, tests for tuberculosis, and the identification of biomarkers for HIV positive patients.

Preparedness, response, and recovery
Surveillance systems enable information gathering about displaced populations or evacuees. 65 Data aggregation and analysis are used to detect communicable diseases, and to support follow-up on damage, loss, mental health diagnoses, and chronic conditions. [15][16][17]35,36 In addition to the resilience provided by cloud-based electronic health records, they aid the monitoring and evaluation of health status of registered refugees over the longer term, potentially improving continuity of care, especially in protracted crises (See Table 2; Refugee Assistance Information System). [37][38][39] Technologies intended to overcome language and cultural barriers include websites and spoken language translators to increase communication and familiarize healthcare providers with cultural beliefs of refugee populations (See Table 2; Ethnomed). 40,41 Affordable training and education applications can cover all stages of humanitarian programming. The International Federation of the Red Cross, for example, trains its volunteers via an online platform that provides modules and a space for online dialogue and peer review. 5

Technology Used by Populations Affected by Humanitarian Crises
Although technological initiatives used by displaced populations do not have direct health objectives, they can have indirect effects on health and its physical, social, and economic determinants (See Figure 2). Table 3 summarizes these technologies, the context in which they have been used, their purpose, impact, and the determinants of health that they address.
Education, social support networks, and healthy child development Low-cost computers and digital classrooms have increased access to education (including health education) for affected populations in low resource settings. Information portals with preloaded content enhance access to information that addresses human rights, protection from violence and abuse, and computing skills (See Table 3; ideas box). 18 By learning collectively through creative media expression and by supporting those who are illiterate, technologies such as digital storytelling provide children with a sense of social inclusion and community interaction to promote mental health, well-being, and healthy child development (See Table 3; Voices Beyond Walls). 42 Feedback from users of such programs suggests the use of simpler activities prior to complex forms of expression. This might mean roleplay, dance/movement, and story boarding around topics not tied to struggles of displacement. 42 Use of SMS during natural disasters can facilitate social support networks. To save lives and property, one initiative received inquiries through SMS and created interactive live radio and television programs for risk coordination and communication (See Table 3; TERA). 5 Improving income, social status, access to health services, and the physical environment In emergencies, electronic voucher programs have played a pivotal role and gained recognition as a dignified form of humanitarian assistance that gives affected populations the liberty to use cash for what they deem important (i.e. food, shelter, health, pharmaceuticals, water, sanitation), thereby improving income, social status, and access to health services (See Table 3; WFP's e-voucher program). 43,44,50   43,45 Social media platforms such as Facebook and Twitter pages for refugee camps allow information sharing among camp residents (See Table 3; Zaatari camp Facebook page). 46 These platforms have been used for information dissemination about services provided in camps (including health), scheduling of aid distribution, advertising available job opportunities, and for weather forecast warnings. 46 Personal health practices and coping skills Thus far, the technologies mentioned have been integrated in marginalized populations with outcomes indirectly impacting health. Whereas personal fabrication technologies have the potential to empower those who have been injured in conflicts, perhaps to 3D print their own prosthetic body parts (See Table 3; Refugee Open Ware). 47

Discussion
Although many digital technologies do not have direct health objectives, they may have indirect effects on health and its physical, social, and economic determinants. For technologies used by humanitarian actors, our framework categorizes initiatives by the stage in the humanitarian program cycle. We recognize that this limited view of humanitarian action may not apply in protracted emergencies where the lines between preparedness, response, and recovery are blurred. Phases overlap, sometimes for decades, such as is the situation in many long-term crises.
Our results highlight the fact that digital technologies facilitate communication, coordination, and the collection and analysis of large amounts of data, enabling timely responses in humanitarian contexts, particularly sudden-onset disasters. The impacts of these technologies most often include their ability to: assess risk and identify the number and location of people with specific health needs, including health services and assistance; to maintain health records; and to enable follow-up, continuity, and improve quality of care. They have also provided a platform for the remote training of healthcare professionals and humanitarians, particularly in hard to reach locations, plus facilitated the coordination and logistics of humanitarian assistance with implications for the determinants of health.
The rise of internet activism around issues of social justice, human rights, education, health equity, and sustainable environments are likely to continue to raise awareness, and pave the way for social change and for health promotion for affected populations.
Our attempt to examine the impact of these technologies on the health and well-being of crisis-affected populations has, however, been hampered by the lack of impact assessment in the design and use of these technologies. Here we identify several gaps and concerns with current digital humanitarian practices:

Lack of evaluation
While surveillance systems have been routinely evaluated, including their impact on health outcomes, very few evaluations exist of other uses of digital technologies in humanitarian contexts. Although many technologies promise social change, evaluation models have not been integrated in humanitarian intervention technologies and remain underdeveloped. 46 Most other technologies discuss purpose and lessons learned with no other evaluative framework. 48 Specified criteria for assessment were often described, but for many technologies, we found neither methodologies nor evaluation reports. 18,[49][50][51]63 Where methods were described, deficiencies included shortcomings in the reporting of denominators and response rates. 15,63 The elusive nature of displaced populations as well as uncertainty and unpredictability of events may be underlying reasons discouraging investments in evaluation. 5,43,46 Rigorous monitoring and evaluation should, however, be conducted to inform funding priorities, and enhance learning for the design of future digital initiatives, as well as to provide an understanding of pathways by which these initiatives can improve health. Process, impact, and outcome indicators would need to be integrated at all levels of design.
Our review of web-based technologies for health promotion suggests that given the complex settings in which health technologies are deployed and disseminated, evaluations of health outcomes and impacts may be very difficult. 52 Humanitarian health innovations, however, might adopt action-oriented research frameworks that integrate evaluation of the design, development, and deployment processes. Health promotion research models such as the Spiral Technology Action Research and the e-Technology Participatory Action Research, emphasize the importance of engaging and working with user communities to identify and clarify the goals of the technology and of reporting back to inform its design iteratively. 53,54 A paternalistic approach Most technologies cited in this review are tailored to meet the needs and functions of humanitarian organizations; to be top-down paternalistic innovations. Within this approach, there is an inherent assumption that humanitarian actors collect data to understand populations' needs, without explicit involvement of the target population. 55,56 Big datasets, however, do not necessarily lead to improved response, particularly if data management and analysis capacities are insufficient. 56 Even when technologies are intended to be used by affected populations (See Table 3), involvement is often superficial. It usually takes place in the data collection phase, not in data analysis and interpretation. We only found one initiative 47 that included the affected population in the design process. Data were rarely made available as a resource for the affected communities themselves. Even with more inclusive design processes, a balance must be struck between designers creating products in a paternalistic manner and the use of potentially invasive participatory design methods which may overburden communities in the process. 57 Technology designers need to gain a deeper understanding of the population's context, health beliefs, and use of technology. Over-reliance on technology raises resilience challenges. Will health be detrimentally affected instead of improved? 55 Despite pressure to introduce and experiment with new technologies, these should be balanced with alternatives for action.

Issues of privacy and equity
As technology continues to facilitate the work of organizations that respond to humanitarian crises, issues of privacy and equity need attention. People caught in humanitarian crises fear having their personal information leaked or their location identified. They raise questions about privacy and fears of being tracked. 43 Data security and privacy should engender trust of the affected communities, to protect the vulnerable and ensure their rights to assistance. 56 The political blockade of information, such as internet censorship and surveillance, as well as intermittent or nonexistent access to internet and mobile data connections can hinder technologies and information dissemination. 46 Unequal access to technology therefore remains a key challenge. Socioeconomic, age, and gender gaps in technology use and access are well known. 5 They further magnify disparities in access to information relevant to health. 5 Similarly, inequities at national levels, where local organizations and governments in low-income countries face challenges in accessing financial, human, and technological resources, reinforce unequal power relations. 5 Crowdsourcing might redress this imbalance, by diffusing knowledge ownership beyond humanitarian actors. Where government disaster responses have led to the rise of new decentralized crisis mapping platforms, this may be the foundation for greater community participation in humanitarian action. 56,58 Recommendations and conclusions Rigorous evaluation of digital technologies in humanitarian crises must be integrated into design and deployment, to strengthen the evidence base for decision-making, including their impacts on health. Increased awareness of the criticisms and potential risks of digital technology use in humanitarian contexts is needed to ensure that technology is used responsibly.
Future work must strive to create a space for dialogue between technology designers and the populations affected by humanitarian crises to achieve increasing relevance and sustainability of innovations. 62 Care should be taken to ensure that participation is not just a token. Indeed, the culture of technology use by affected populations is on the rise. In many cases, the affected populations are the innovators. They should be supported as they adapt and appropriate digital technologies to their own needs. material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons. org/licenses/by/3.0/