LEADS for Health Security and Resilience Consortium LEADS for Health Security and Resilience Consortium
Frequently Asked Questions (FAQs)
About the Project
L4H or the LEADS (Leading Evidence Based Actions through Data Science) for Health Security and Resilience Consortium is a collective of professionals from varying disciplines related to data sciences and public health, that was formed to assist in the response against public health threats, particularly in the context of COVID-19. The consortium utilizes the expertise and experience of its members to develop tools that could help decision-makers and local governments arrive at informed plans and policies for their disease mitigation and control measures.
The L4H consortium currently provides two dashboards that show separate sets of relevant information: 1) the Time Varying R Dashboard, which contains updated monitoring graphs of the effective reproductive number, case fatality rates, and doubling rates, and 2) the CO-INFORM Risk Dashboard.
Anyone can use the L4H tools. Although originally designed for decision-making at the level of individual Local Government Units, these tools are open for anyone who may find them relevant and useful for health resilience and security.
Responding against COVID-19 involves a great deal of data management and decisions relating to how data is interpreted. L4H tools provide the convenience of putting in map or graph form the various information that would assist in forming policy recommendations and decisions that would curb the spread of disease. It hopes to present in simple terms the means by which we measure, for example, 1) the level of transmissiblity of COVID in a given area at a given time, and 2) the level of risk that a certain locality possesses in the context of disease spread.
Amidst the barrage of graphs and models being produced by different groups, what served as the primary impetus for L4H to convene was the presence of some other models that while well-meaning, seemed to somewhat overlook particular aspects of public health, epidemiology, and data analysis in their presentation. The strength of the collaboration of L4H is that there is input from various experts in these intersecting but discrete fields who can countercheck each other's efforts and collegially come up with a sound and comprehensive data output.
Data sources for the models consist of public databases of Johns Hopkins University and the Philippine Department of Health (DOH)
The L4H is very open to collaborations with Local Government Units to assemble a package of relevant and useful datasets that can inform policy decisions regarding COVID and health security and resilience. Please feel free to contact us at the details below.
Time Varying R Dashboard
The Time Varying R Dashboard shows a series of graphs documenting the trends of COVID in the Philippines. In particular, it documents the trend of effective reproductive number (Rt), the delay-adjusted Case Fatality Rate (CFR), Crude Recovery Rate, Confirmed Cases, Reported Deaths, Reported Recoveries and their respective doubling times, and comparative case counts.
The basic reproduction number (R0) is used to measure the intrinsic transmission potential of a disease when first introduced in a full susceptible population. It is the average number of secondary infections produced by a TYPICAL case of an infection in a population where EVERYONE IS SUSCEPTIBLE. R0 is affected by the following factors: (1) rate of contacts in the host population; (2) probability of infection being transmitted during contact; and, (3) duration of infectiousness.
For example, if the R0 for COVID-19 is 2.5 when first introduced in the Philippines, then we would expect that each new case (particularly the index case or patient zero) to produce 2.5 new secondary cases assuming all of his/her contacts are fully susceptible to the disease. Note that moving forward, computation of R0 excludes new cases produced by the secondary cases (i.e., we are only looking at the first generation of offsprings).
The effective reproduction number (Rt) is the instantaneous transmission potential of a disease at current time t. It is the average number of secondary infections produced by a TYPICAL case of an infection during the prevailing conditions in a population. In other words, Rt is the average number of secondary cases per infectious case in a population made up of both susceptible and non-susceptible hosts, where non-susceptibility can be due to immunity, changes in behaviour leading to reduced probability of getting infected such as reduced contact rates or observation of proper hygiene. Rt can be crudely estimated by the product of the basic reproductive number (R0) and the fraction of the host population that is susceptible (x), i.e. Rt = R0x.
This metric is the one which can be monitored in real-time to evaluate the current transmission potential of a disease to evaluate strategies and programmes to reduce, or even eliminate, the disease incidence including case isolation or increased case detection capacity and treatment. In other words, this is the metric which we want to reduce to less than 1 to be able to say that we have successfully controlled the disease, or even eliminate the disease from a population. Note, however, that estimation of this metric is very sensitive to reporting fidelity, lags, and changes. For examples, please refer to our Rt dashboard.
While the R0 is a summary statistic that represents the average number of individuals that one infected person can transmit the disease to (ie, an R0=2 means that an infected person will, on average, transmit the disease to two other individuals in that population), the Rt, which is being used in the Time Varying R Dashboard, additionally takes into account the elements of time and interventions being implemented to control the disease. This allows for monitoring of ongoing dynamics of transmission and infection at any given point in time. The Rt reflects real-time transmissibility of the disease and varies depending on the effectivity of control measures, where an Rt > 1 would indicate that the epidemic is still causing transmission and therefore is not under control, an Rt = 1 means that at a given point an infected person is only transmitting to one secondary case, while an Rt < 1 would show decelerating transmission activity at a certain time.
The basic reproductive number R0 is a fundamental measure used in epidemiology to quanitfy the potential for transmission of a disease in a population. However, R0 cannot reflect the time-varying nature of an epidemic. Time varying R or Rt, on the other hand, can track in real-time the evolution of transmission dynamics
The target Rt that we would like to see is an Rt < 1, which would suggest that the community has started to effectively control the spread of COVID-19.
The Time-Varying R Dashboard is updated daily to reflect real-time trends.
There is general agreement that a "flattening of the curve" of the pandemic is a reasonable measure of the control over the widespread transmission of disease. What this entails is a suppression of the tendency of a graph to proceed in exponential increase of number of cases.
While in general terms this would mean a stabilization or even decrease in the daily number of cases reported, monitoring raw numbers of cases may be tedious and fails to account for trends in data. A good indicator would be one that can demonstrate not only disease trajectory, but also the risk of transmission in certain communities. It is also important to note that quality of data is crucial, and that availability of information on the "denominator" or the actual number of cases being detected (as influenced by testing measures, detection, and contact tracing) will dictate the accuracy and usefulness of whichever indicator will be utilized.
Herd immunity is what happens when a threshold number or proportion of people in a population develop immunity against a disease, thereby producing a level of protection for the rest of the population. The presence of herd immunity in a group of people (brought about by either vaccination or built immunity due to infection) would result in a slowing down of the disease transmission yielding a flattening of the curve, however a very large proportion of people would have to be either immunized or infected in order for this effect to be achieved.
Increasing COVID testing and increasing yield of test results would produce more accurate data. If there are not enough tests being done and read, there is a risk of misrepresenting the actual number of cases in a population, especially considering the presence of asymptomatic individuals. The proportion of deaths and disease severity may also be skewed by inadequate testing.
CO-INFORM Risk Dashboard
The CO-INFORM Risk Dashboard provides a visualization of a computed level of risk of a certain area (up to the Provincial/City Level) to the effects of COVID. The CO-INFORM Risk Heat Maps show how susceptible localities may be to the disease, which can be a basis for drafting plans to reduce the risk and mitigate future impacts. As a tool for communication, these maps may be used to convey to local stakeholders the areas that would need proactive intervention and multi-sectoral collaboration. As a model, the CO-INFORM Risk Index utilizes multiple indicators from varying dimensions to comprehensively evaluate the level of risk across the country.
Broadly, the CO-INFORM index is composed of the dimensions of Hazard & Exposure, Vulnerability, and Resilience. Each of these dimensions, in turn, are measured by indicators such as the number of infected individuals, level of sanitation in the community, non-communicable disease burden, proportion of vulnerable populations, poverty incidence, health system capacity, etc. These multi-dimesional indicators are meant to holistically and adequately ascertain how a population within a certain jurisdiction will fare given the risk of COVID in that area. The illustration of this level of risk is envisioned to provide a sense of how local governments can adapt and proactively manage their risk factors. By identifying the specific parameters that siginificantly alter their risk index, data-driven directives and policies can be pursued.
The data sources for the indicators are from publicly accessed databases and reports, as well as DOH data.
The CO-INFORM model and its results are largely reliant on available data. If there are discrepancies with more concurrent data present in a certain locality, this is primarily due to lack of access to said data. If there are localities that can provide their own updated information on the parameters required, these limitations in datasets can be managed, and the model may be able to reflect more localized and accurate risk indices.