POLICY NOTE NO. 6 / 08 May 2020
Introduction
Our previous policy notes already highlighted some epidemiological metrics vital for monitoring the state of the country in its fight against COVID-19. These metrics provide context, give insight, and serve as a guide to help stakeholders take control of the disease and monitor our own progress. As we have repeatedly pointed out since Day 1, any analysis is only as good as the data that we have.
Relevant, and accurate data about COVID-19 and the resources the country has in the fight is important. The speed and timeliness of how such is released is equally critical. We do recognize the difficulties in the collection of detailed and timely records of COVID-19 data on a nationwide scale. However, no matter how difficult, these problems need to be addressed at the soonest possible time and should be of top priority. Here, we discuss prevailing data issues we have observed, and our recommendation for open data moving forward.
Data Sharing Practices
Collecting COVID-19 data has not been easy. Since the country finally accepted the presence of community transmission around the first week of March 2020, government’s data reporting protocols have changed far too often (Table 1). There was initially no standard time of the day for when DOH posted the official daily numbers. There was even a brief period when DOH stopped releasing updates, leaving people using the data in the dark.
Fortunately, the DOH has been open to feedback, and this has been reflected in the improvement in reporting. There is now a centralized resource for modelers to use as a common reference. There is also regularity and predictability on when reports are given, and this is commendable given that transparency and timeliness are key to managing any crisis situation, especially a pandemic.
However, there is still room for improving the quality of the data, and the process of collecting information for DOH’s data drop in the tracker. It is important to make sure that correct data is captured as swiftly as possible to minimize uploading of erroneous and anomalous, if not missing, data.
Data Accuracy and Integrity
The availability of accurate and relevant data is a basic requirement in managing any situation that requires urgent and targeted response. Almost three months since we had our first confirmed case in the person of a Chinese national on 30 January, we have yet to reconcile differences in numbers between DOH and LGU sources. For example, on 03 May 2020, DOH reported 7 deaths (28 recoveries) in Laguna, which was 22 deaths (65 recoveries) lesser than the provincial government’s official count.
Accuracy, however, goes beyond correctness in reporting aggregate numbers. Recent data drops by DOH revealed a number of alarming patient-level inconsistencies, if not gross errors. A quick comparison of the April 24 and April 25 data drops showed that forty-five (45) cases have changed sex from male to female or vice-versa; while 75 others had the data on age modified. This is on top of the 516 cases where the residence data was reclassified to another city, if not a completely imaginary city (i.e. a barangay or district) like what happened in the City of Manila (Table 2).
Related to the problem of accurate residential reporting is the handling of certain variables in the DOH data drop. For example, RegionRes is a variable for the region of residence and is coded in text such as “NCR”, “Region III: Central Luzon”, and so on. RegionPSGC is the region code based on the Philippine Standard Geographic Code [PSGC], compiled by the Philippine Statistics Authority, based on their mandate to prescribe “uniform standards and classification systems in the generation of government statistics to ensure harmonization and comparability of statistics in the country and at the international level” (PSA 2020). The DOH data drop introduced this variable last April 26, 2020 for regional, provincial, and city/municipal classification. Each region has a unique code in the PSGC; however, as seen in the table below, inconsistencies in the coding of the PSGC for cases within regions have been problematic. It is noted that the DOH Tracker uses the RegionRes variable for statistics, not the Region PSGC (Table 3).
There are other troubling anomalies in recent data drops of DOH. For example, 18 cases no longer have data on residence in the April 25 update. On the same date, the recovery dates of two cases were either missing or changed. One patient who reportedly died on April 24 is no longer dead the following day.
The DOH data drop is also inconsistent with its use of date formats, which makes it difficult for automated systems of extracting and updating data from case information. It has made the work of data analysis difficult because of these sudden changes (Table 4).
These lapses may seem small relative to the total size of data contained in the daily updates, but they have significant implications on the reliability of our scientific analyses on COVID-19. Patient case data is the keystone for effective and insightful metrics and analysis. The integrity of the data drops is particularly important given that no less than President Rodrigo Roa Duterte himself has said many times that the government’s decision on managing COVID-19 will be based on science. We fully support President Duterte on this call for science-based decisions, hence this statement.
Transparency and Accountability
We acknowledge the importance of data privacy as provided for in our existing laws such as the Data Privacy of 2012 (RA 10173) and the Mandatory Reporting of Notifiable Diseases and Health Events of Public Health Concern Act of 2018 (RA 11332), among others. However, there are important data that can already be anonymized and made available to serve public interest. For example, identifiers, such as employment information or specific addresses may be removed, but variables such as onset of symptoms, exposure history, co-morbidities, and whether they were medical front-liners or not are key inputs for modelers and statisticians to map the progress of our fight against COVID-19.
We are also aware that the DOH is already sharing government data with selected groups from the private sector. These organizations are bound by Non-Disclosure Agreements (NDA), as required by law. However, while it may be legally right, it does not serve public interest in this time of great need for accurate and timely information. For example, DOH restricts the analytics involved with patient statistics, even for some aggregates, which have implications on understanding IATF’s recommendations for placing some provinces under ECQ.
The COVID-19 pandemic requires a science-based approach, and science cannot exist in a vacuum. Any scientific output would benefit from cross-validation from peers, and if findings do not converge, we might be standing on shaky grounds. Such scientific rigor can only happen in an environment where data, especially government data, is made available to all relevant stakeholders. Entrusting government data to select private entities is inimical to public interest.
Call for Open Data and Scientific Cooperation
We understand that some data can only be shared internally (i.e. within the government) and are not fully open to the public. In this regard, we call on other agencies, to share relevant data that can help capable institutions make scientific assessments for discussions on the evolving crisis to come up with better peer-reviewed science. Regardless of technology, it is important that the reporting system be standardized and regularized, integrated into the existing data tracker as much as possible, and made open to the public.
We also call on private institutions to contribute to the COVID-19 related data already shared in the COVID-tracker Data Drop. We believe there are private corporations who possess data that can benefit all researchers cooperating in this fight.
Making all data sources open, while also being mindful of the same data privacy protocols that DOH is following, can further empower both official and unofficial stakeholders (i.e., commissioned and independent scientists and researchers, local governments officials, IATF/NTF decision-makers) in the battle ahead. This is important not only to inform our plans, but also to tell us how we are doing in the fight against COVID-19. This is particularly so in light of recent announcements by DOH and its private partners that we have already flattened the curve. Without access to full government data entrusted to select private sector groups, the task for an independent corroboration—the hallmark of any scientific undertaking—becomes impossible, to the detriment of public welfare and interest.
This call for open data is in line with the UNESCO call for open science and reinforced scientific cooperation. According to UNESCO, it is imperative now more than ever to strengthen/build international inter-continental and national scientific cooperation between scientists, decision/policy makers, private practitioners, industries and health professionals and civil society for a multi-dimensional approach to tackling the pandemic. This calls for open access to scientific knowledge and know-how, data sharing and evidence-based policy and decision-making.
Nowhere is the need for Open Data as clearly manifested than in the current COVID-19 crisis. In preparing for, responding to, and recovering from the impacts of health hazards or any natural hazard for that matter, data must be used to generate knowledge. If we keep our information in silos, our collective efforts and perspective of the situation narrows and so do our chances to maintain and preserve public health and security. Ultimately, because the battles ahead will no longer be just about health, this call for more open data sharing is a call to other sectors as well. We need to resolve our data issues posthaste to secure public trust in the plans, decisions, and pronouncements of the government and its private partners.
For questions or clarifications related to the technical or other aspects of this policy note, please send an email to upri.covid19＠up.edu.ph. Scientific reports related to this statement will be posted in the endcov.ph site.
Download a copy of the Policy Note No. 6 here.
