COVID-19: A Magnifying Glass for the Mekong Region’s Data Ecosystem
By the time COVID-19 was announced as a Public Health Emergency at the end of January 2020, the need for more data to understand both the virus and its impacts was clear. Now, more than ever before, data are needed to support decision-making on developing effective solutions that meet the needs of all, including marginalised populations like migrant workers, Indigenous Peoples, ethnic and religious minorities, refugees, and women. Yet, long-standing gaps in the global data environment, highlighted by the current crisis, remain a hindrance to effective, evidence-based decision-making.
In the Mekong region, the data ecosystem is composed of government institutions such as National Statistics Offices (NSOs) and ministerial level departments as the primary and authoritative data producers and stewards, as well as not-for-profits, academia and the private sector who contribute formally and informally. Meanwhile, the nascent open data community contributes to citizen verification and oversight. Data constituents range widely, from journalists to academics, civil servants to students, business analysts and NGOs. From collection to analysis, their interactions with data, processes, technology and each other along each step of the way impacts decision-making. Using the pandemic as a magnifying glass, we take a look at how the data ecosystem impacts the region’s most vulnerable populations during the current pandemic.
An insufficient baseline
An assessment of the ability of the Southeast Asian countries, including the Lower Mekong Region, to track progress using the 2030 Agenda for Sustainable Development’s comprehensive framework for monitoring global development noted that there was insufficient baseline data in 2016. In the intervening half decade, this statistical gap has barely been bridged. When it became clear, early on in the pandemic, that COVID-19 would also impact the Lower Mekong Countries, this lack of data continued to be an issue. This was especially obvious when the Lower Mekong Countries reported surprisingly low numbers of infections, which critics questioned on the basis of data: limited testing in the region meant that there was no clear sense of actual impact. Furthermore, vulnerable populations are conspicuously absent from the data value chains, and this begins from design decisions that exclude them. Minority populations are neither consulted nor presented in data collected. Singapore’s initial success and subsequent downfall in effectively combating the virus is a high-profile example of the real-world impact of non-representative data.
While the best time to have collected baseline data is yesterday, data ecosystems would ideally be redesigned for data to benefit all by enabling inclusive representation of all vulnerable populations, including Indigenous Peoples and ethnic minorities. A significant hurdle that remains is the need to build trust between data producers, users and consumers. However, in the absence of capacity, political will, and time to redesign a responsive data ecosystem, independent contributions to these baselines are important, as discussed below.
Data quality, timeliness and interoperability
An unprecedented quantity of information about the pandemic has been spread incredibly quickly and widely, leading experts to term it an “infodemic”. As the pandemic wears on, the importance of data quality is coming to the fore as data analysis can only be as accurate as the data being used. While some of this information can clearly be classified misinformation or disinformation, others are not so clearly unreliable. For instance, even scientific models from leading institutions are limited by the data on which they rely, meaning that there is no clear sense of the actual impact of the virus in certain contexts. Low data quality circulating in an ecosystem that already lacks sufficient data has a significant impact on the Mekong Region due to poor governance and infrastructure, addressed below. Data are not disseminated in a timely manner, and there are clear gaps in the data, as well as inadequate disaggregation. NSOs whose statistical capacities have dropped in the last years use collection methods that are inherently biased as they may not include all vulnerable populations and are not disaggregated enough to distinguish true population demographics. Data are often made available in non-machine readable formats, increasing the risk of error in analysis and decreasing usability. This negatively impacts the ability of governments, organisations, and other data users to make decisions that accurately reflect real needs.
Improving data quality requires capacity development that increases engagement at all levels, as well as expanding the data ecosystem to allow for authoritative use of data derived from non-traditional sources. It is a challenging task and may require validating national statistical data with non-traditional approaches, such as semi-supervised machine learning or other crowdsourced approaches. The role of citizen science is being highlighted in the context of this lack of COVID-19 baseline data, although caution should be taken to ensure that this is being done responsibly. Open civic tech organisations, such as The Ananda in Myanmar and Sinar Project in Malaysia, can support responsibly governed data collaborations intent on filling in data holes, exemplified by Taiwan’s coronavirus response.
Data governance and infrastructure
Data governance frameworks vary widely. The Mekong region has little policy or legislation that deals with data governance. Thailand, the most advanced in the region in this respect, has promulgated a Personal Data Protection Act, but its enforcement may be delayed, due to the coronavirus. In the context of developing e-government, Vietnam has proposed 24 programmes in an effort to restructure government IT infrastructure; several of these programmes relate to developing data governance standards. However, these proposals remain in draft and under review. The other three Mekong countries do not yet have relevant legislation. Digital contact tracing solutions being rolled out in Thailand and Vietnam into this void means that there are many unknowns about how this data will be treated.
While the obvious solution is to develop policy and legislative data governance structures, more is needed. Data disregards national boundaries, and its use transcends borders. Without a full understanding of differences in the treatment of data across countries, the effectiveness of a particular course of action could be seriously impacted. As such, the need for consensus on collective data rights that also take into account structures for emergencies like COVID-19 is high. Yet, leaving data governance solely in the hands of those in power may result in a replication of more of the same, especially in regions like the Mekong where the incentive to protect the civic rights of vulnerable communities is low. Thus, community-led and rights-based approaches, such as Indigenous Data Sovereignty, may be a useful addition to the regional data ecosystem. Collectivising data governance and infrastructure, decentralised from national systems, can provide rights over collection, ownership, storage, sharing, and use that prioritize the needs of disenfranchised communities while informing contextually appropriate responses to COVID-19.
Where to look next
The seriousness of the pandemic has warranted what might otherwise be considered a severe infringement of rights in normal times. Very sensitive personal telecommunications data has been used to track the movement of individual people — something that in many places would seem like an abhorrent invasion of personal rights and would normally not be shared. However, this is currently being done in order to stop the spread of the virus, sometimes effectively, but not without questions as to how it will be treated with regard to privacy, security and other concerns. The reality is that this pandemic has shone a light squarely on ethical concerns about data governance, even in environments with robust protections.
For the Lower Mekong Countries, the pandemic has also shown that there is still a long way to go before the regional data ecosystem will truly support the needs of its constituents, let alone what might be helpful globally. One way to bridge this gap is to continue working to make data responsibly available in accordance with best practices. The FAIR principles are one such standard, while the CARE principles take into account unique requirements for the responsible management of Indigenous Peoples’ data. Applied together, these principles should be considered a minimum required standard for making data openly accessible. Efforts are being made to publicly release more data by the governments of Vietnam and Thailand, although data is not yet made open by default across the region. And while the ability for Indigenous People and ethnic minorities to exercise their data rights remains hindered by national governments, data producers and users can still adhere to available best practices that respect the rights of Indigenous Peoples and minorities.
Regardless of any of these data-focused solutions, the development impacts of the virus must be addressed, including that humanitarian aid for the most marginalised populations cannot be disrupted. The threat of a hunger pandemic resulting from the COVID-19 pandemic is a very real possibility, with marginalised populations in the Mekong Region already feeling it. Focusing narrowly on using new technologies to collect more data for COVID-19, but in poorly regulated and low trust environments, may not serve all populations. Without a long-term view and careful use of tech tools, the most vulnerable will only be left to trail further behind.
— — —
Contributions and edits by Pyrou Chung, Director of the Open Development Initiative. To learn more about what the Open Development Initiative team has been up to, visit www.opendevelopmentmekong.net.
