COVID-19 FORECASTS IN THE PHILIPPINES: Sub-National Models for NCR and other Selected Areas

| Posted by UP Media and Public Relations Office

Forecast Report No. 6
(May 8, 2020)

 COVID-19 FORECASTS IN THE PHILIPPINES: Sub-National Models for NCR and other Selected Areas

 

Guido David, Ph.D.
Professor, Institute of Mathematics
University of the Philippines
Fellow, OCTA Research (www.octaresearch.com)

Ranjit Singh Rye, MPA
Assistant Professor,
Department of Political Science
University of the Philippines and
Fellow, OCTA Research (www.octaresearch.com)

Ma Patricia Agbulos, MBM
Associate, OCTA Research (www.octaresearch.com)

 

OVERVIEW

Nearly two months into the Enhanced Community Quarantine of Luzon, the pandemic is still spreading but at a manageable rate. In this report, we look at the state of the pandemic in the country and the provinces. However, the mathematical model used, with all its limitations and underlying assumptions, is still highly dependent on complete and valid data. Caution must be made to avoid complacency based on the results of a mathematical model. Trends may change at any time.

Figure 1 shows the cumulative Covid-19 cases in the Philippines since March 8. The spread of a pandemic would normally resemble exponential growth in the early stages. While the ECQ has helped control the spread of Covid-19, the number of cases appears to be increasing at a steady rate. The figure suggests that the curve has not flattened for the entire country (the estimated value of the reproduction number R for the entire Philippines is 1.2). This is not what we want when controlling the pandemic.At first glance, the ECQ is not working as well as we would like it to. However, there are a number of points that will explain what is going on and give us more confidence in the effectiveness of the ECQ.

1. Recent increase in Covid-19 testing has resulted in additional new cases.

2. Localized analysis of the pandemic reveals that the ECQ, has in fact, been very effective in certain areas. Other areas have had recent surges, thus affecting the overall totals. Localized analysis is provided in subsequent sections of this report.

3. Based on Department of Health (DOH) data, more than 1,000 cases are still up for validation.

 

Figure 1. Aggregate number of cases in the entire Philippines.

 

On the other hand, there are some factors that may also cause the number of cases in the Philippines to increase.

1. There are asymptomatic and mild Covid-19 cases that have not yet been tested. This highlights the need for mass testing.

2. Delays in reporting of data on transmission and number of deaths to the Department of Health. From our experience, some Covid-19 cases are reported late. Our estimate is that the number of new cases in a particular week may still increase by an average of 15% once all the data has come in.

Figure 2 shows the distribution of Covid-19 cases in the Philippines by age and gender.The distribution characteristics are similar between male and female Covid-19 patients. More males are affected by Covid-19 from ages 25 and up.

 

Figure 2. Distribution of Covid-19 cases in the Philippines by age and gender.

 

EVALUATION OF PANDEMIC MANAGEMENT PER PROVINCE

We use the following measures or grades in evaluating the state of the pandemic in each province (or city).

A – no new cases for the past 7 days

B – the average number of new cases for the past days is less than 1 per million of population, and the value of R (epidemiological reproduction number) is consistently less than 1 during the past 7 days. Those with less than 5 new cases over the past 7 days will also fall under this grade.

C – the value of R based on the most recent 7-day average is less than 1, but the average number of new cases for the past days is greater than 1 per million of population. Those with 5 to 14 new cases over the past 7 days will also fall under this grade.

D – the value of R based on the most recent 7-day average is greater than 1

We may consider Grade A as a province or LGU where there is no pandemic and there are no new cases of Covid-19 reported. A Grade B province or LGU is in a low-level pandemic state with few new cases. A Grade C province or LGU is where there is a pandemic that has stabilized or the curve has flattened, but the number of new cases is still considerable. In a Grade D province or LGU, the pandemic is still escalating and the curve has not yet flattened. The one per million cases per day is a benchmark used by South Korea in evaluating the level of a pandemic.

Inclusions in each grade are subject to verification of data as provided by the Department of Health (DOH). Analysis for select regions are provided in subsequent sections. The provinces (and LGUs) according to grade are:

 

Grade A

Ilocos Sur                                La Union                                Isabela
Abra                                         Benguet                                  Ifugao
Nueva Ecija                            Tarlac                                     Zambales
Marinduque                           Palawan                                  Romblon
Catanduanes                          Capiz                                       Negros Occidental
Negros Oriental                     Leyte                                       Northern Samar
Bukidnon                                Camiguin                                Lanao del Norte
Davao de Oro                         Davao del Sur (except Davao City)
Davao Oriental                       North Cotabato                     Cotabato City
South Cotabato                      Sultan Kudarat                      Agusan del Norte
Lanao del Sur                         Maguindanao                         Sulu

All other provinces with no reported cases of Covid-19.

 

Grade B

Ilocos Norte                           Pangasinan                            Cagayan
Nueva Vizcaya                       Bulacan                                  Cavite
Quezon                                   Rizal                                        Occidental Mindoro
Oriental Mindoro                 Camarines Sur                       Aklan
Antique                                   Guimaras                                Bohol
Western Samar                     Misamis Occidental              Misamis Oriental
Davao del Norte                    Davao Occidental

 

Grade C

Bataan                                   Pampanga                              Laguna
Albay                                      Iloilo

 

Grade D

NCR                                         Batangas                                Cebu
Zamboanga del Sur              Davao City

Those provinces in grade A have had no new Covid-19 cases for the past 7 days. Those in Grade B had a low number of new cases. Those in Grade C are close to flattening, or in the case of Albay and Iloilo, resurging but with a manageable number of new Covid-19 cases. Those in Grade D have a reproduction number R> 1.

 

SELECT REGIONS

We examine in detail the state of the pandemic in the following provinces or cities: NCR, Albay, Iloilo, Cebu (province), Cebu City, Zamboanga del Sur and Davao City. Figure 3 shows the number of new cases (over the past 7 days, from April 30 to May 6) for the select regions. Note that due to delays in reporting, the number of new cases may still increase by an average of 15%. NCR has the most number of new cases, followed by Cebu City. Cebu province reported many new cases from Mandaue, Lapu-Lapu and a few other LGUs. Most of the new cases in Iloilo are from Iloilo City. Legazpi accounts for most new cases from Albay.

 

Figure 3. Number of new Covid-19 cases (from April 30 to May 6) for Albay, Cebu City, Cebu province (not including Cebu City), Davao City, Iloilo, NCR and Zamboanga del Sur.

 

Figure 4 shows the calculated 7-day average of the reproduction number R. The reproduction number indicates the rate of spread of the pandemic. If R< 1, the curve is flattening. The value of R is not correlated with the number of new cases (for the past week), rather on the rate of change of new cases. This explains why Zamboanga del Sur, Cebu (not including Cebu City) and Iloilo have large values of R. These three provinces have seen a significant increase in the number of new cases over the past week. For example, the number of new cases in Zamboanga del Sur represents 74% of the total number of Covid-19 cases in the province, while the number of new cases in Cebu province is 75% of all Covid-19 cases in the area. Care must be taken into consideration to control the spread of the pandemic. For Albay, Cebu City, Davao City and NCR, the value of the reproduction number R is greater than 1. The continued implementation of the ECQ should help push this value below 1. Cebu (the province and Cebu City) and Zamboanga del Sur will need a more disciplined implementation of the ECQ to prevent the spread of the pandemic.

 

Figure 4. Reproduction number R for key areas in the Philippines where there is a high rate of transmission of Covid-19. Zamboanga del Sur, Cebu province and Iloilo all had recent surges in Covid-19 cases.

 

National Capital Region (NCR)

For the National Capital Region(NCR), the number of new cases per LGU is shown in Figures 5a (for District 1 and 4) and 5b (for District 2 and 3). Quezon City has the most number of new cases (for the week from April 30 to May 6), followed by Paranaque, Manila, Makati, Pasay and Taguig.

 

Figure 5a. Number of new Covid-19 cases (past 7 days) for NCR Districts 1 and 4. Paranaque, Manila and Makati have the most number of new cases in these two districts.

 

Figure 5b. Number of new Covid-19 cases (past 7 days) for NCR Districts 2 and 3.Quezon City has far and away the most number of new cases in NCR.

 

The reproduction number R for each LGU is shown in Figures 6a (for District 1 and 4) and 6b (for District 2 and 3). Malabon, Navotas, Paranaque, Muntinlupa and Pasay all have R> 1. Due to late reporting of a fraction of new Covid-19 cases, it is possible that Makati, Taguig and Quezon City have R close to or greater than 1. The best performing LGUs in terms of controlling the pandemic are Marikina, Mandaluyong, San Juan, Valenzuela and Pasig. Moreover, there are currently more than l,000 unvalidated data in NCR, so it is not yet clear to which LGUs these cases belong to.

 

Figure 6a. Reproduction number R for NCR Districts 1 and 4. Las Pinas, Pateros and Manila have the lowest value of R in these two districts. Paranaque, Muntinlupa, Pasay have R> 1.

 

Figure 6b. Reproduction number R for NCR Districts 2 and 3. Marikina, Mandaluyong, San Juan, Valenzuela and Pasig have the lowest value of R in NCR. Malabon and Navotas both have R> 1.

 

CALABARZON

On the outskirts of NCR, Calabarzon has been the region hit hardest by the pandemic next to NCR in Luzon. Based on DOH data, 80% of all Covid-19 cases are found in Luzon. Figure 7 shows the number of new cases (from April 30 to May 6) in the provinces of Calabarzon. Laguna has the most number of new cases, averaging 7 new cases per day, while Quezon has the least, with about 1 new case per day.

 

Figure 7. Recent (past 7-days) Covid-19 cases for provinces in Calabarzon.

 

Figure 8. Reproduction Number R for Calabarzon. Batangas still has a value of R> 1. Laguna and Quezon both have R< 1. Rizal and Cavite have the lowest values of R.

 

Figure 8 shows the reproduction number R for the provinces in Calabarzon. Only Batangas has R> 1 indicating the curve has not yet flattened in the province. Rizal and Cavite both have a low value of R, i.e. the curve is flattening in these two provinces.

CENTRAL LUZON

The number of new cases over the past 7 days for the provinces in Central Luzon  (except Aurora) is shown in Figure 9. No new cases were reported in Nueva Ecija, Tarlac and Zambales. We will check if there are late reports of new cases in these provinces by next week.

Figure 9. Number of new cases (over the past 7 days) for the provinces in Central Luzon except Aurora. No new Covid-19 cases have been reported in Nueva Ecija, Tarlac and Zambales.

 

Figure 10. The reproduction number R (over the past 7 days) for the provinces in Central Luzon (except Aurora). With no new cases, R is very low for Nueva Ecija, Tarlac and Zambales. Pampanga and Bulacan both have R > 1, i.e. the curve has not yet flattened in these provinces.

 

FORECASTS

Three forecasts are illustrated in Figure 11 for NCR. The current number of total cases in NCR is around 6,600. The first forecast uses the current trends based on the recent 7-day average of the reproduction number R, projecting the numbers to June 15. This forecast is the default, marked ECQ and in red. The second forecast assumes a continued ECQ and extrapolates R(assumed to decrease by 1% per day) over the forecast period. This forecast is denoted IMPROVING and marked blue. The third forecast assumes a relaxing of the ECQ, with an increase in R. The value of R increases from the current state. This forecast is denoted NO ECQ and marked in green. Note that with a continued ECQ, the number of Covid-19 cases in NCR should be around 12,000 by June 15 with 800 deaths. Without any changes in the trend, the projection is 14,000 cases by June 15 with 1,000 deaths. A premature relaxation of the ECQ in the NCR may result in an escalation of 24,000 cases and 1,700 deaths  by June 15, 2020.

 

Figure 11. Forecasts for NCR until June 15 using current trends in the reproduction number R with no improvements (ECQ), continued ECQ but with a value of R that is decreasing by 1% per day (Improving), and some form of a modified ECQ. With an improving ECQ, the projection is 12,000 total Covid-19 cases by June 15 with 800 deaths in NCR. With a modified ECQ, the projection is 24,000 cases by June 15 with 1,700 deaths.

 

SUMMARY RECOMMENDATIONS

Four grades were provided, based on the number of new cases (over the past 7 days), and the measured reproduction number. Grade A includes those with no new cases over the past 7 days. The list includes provinces under GCQ, and some provinces under ECQ, such as Nueva Ecija, Tarlac and Zambales. Grade B includes those with a low number of new cases, i.e. less than 1 new case per day per million of population. The list also includes many provinces under GCQ, but also includes provinces under ECQ, such as Bulacan, Cavite, Quezon and Rizal. Grade C includes many provinces under ECQ, and some provinces under GCQ that have some resurgence in Covid-19 cases. These provinces and cities are close to flattening the curve, but may still need time to ensure this (due to delays in some cases being reported). The list includes Bataan, Pampanga, Laguna, Albay and Iloilo. Many LGUs in NCR may actually be considered Grade C. This includes Marikina, Mandaluyong, San Juan, Valenzuela and Pasig. Of course, this is still subject to late reports of Covid-19 cases which may increase the number of new cases in those LGUs. Finally, Grade D are cities and provinces where the calculated R is still greater than 1. Continued implementation of the ECQ is needed in order to manage the spread of the pandemic. The list includes the entire NCR, Davao City, Cebu City and the province of Cebu, Batangas, and Zamboanga del Sur.

We emphasize that the number of new cases and low transmission rate (low reproduction number R) is just one of the factors to be considered in deciding whether to continue or lift the quarantine in an area. This factor is highly dependent on accurate data, and any delays in transmitting reports will affect the grade of a province or city. Erring on the side of caution is always recommended. In addition to the low number of new cases and low transmission rates, it is advisable that the province or LGU must also meet the minimum testing requirements and health safety measures put in place.

Given these scenarios, we recommend the following:

1. That the national government (NG) should consider extending the enhanced community quarantine (ECQ) beyond May 15, 2020 for the following areas: National Capital Region, Batangas, Cebu, Zamboanga del Sur, and Davao City.The NG may also consider extending the ECQ for Albay, Bataan, Iloilo, Laguna and Pampanga, depending on their health provisions. As shown above, there had been gains due to the ECQ. The goal is to sustain these gains until such time that it has scaled up and rolled out its programs and its initiatives for mass testing, contact tracing and isolation of infective individuals. We wish to emphasize that we are faced with a choice not between the economy and public health (lives vs livelihood), but between a less or more costly disruption to the lives of ordinary Filipinos. Based on our data, if the ECQ is lifted prematurely, we will be faced with another wave or a surge in transmissions that is certain to squander our gains forcing us to make further costly interventions and increasing the total economic cost and the number of lives lost.

It is in this light, that we exhort the government to reconsider its plan of relaxing the ECQ for selected cities within Metro Manila. We believe that it is premature to consider this proposal given the paucity of reliable data to support it at this time. Moreover, the proposal will not just be technically and administratively challenging to implement in an integrated structure like Metro Manila but it will also further burden our LGUs that are now focused on scaling up their testing, tracing and isolation capabilities as well as their local health capacities and facilities. We believe that this proposal will create additional risk as the relaxation of the ECQ will encourage increased mobility, a driver of the virus, in some cities in Metro Manila. It is possible that this may lead to increasing transmissions of COVID-19 infections and could possibly lead to more deaths. Our data suggest that a premature relaxation of the ECQ in the NCR may result in an escalation of 24,000 cases and 1,700 deaths by June 15, 2020.

We caution the government against the premature relaxation of the Enhanced Community Quarantine without substantial data and without the minimum health safeguards in place in affected areas regardless of the historical number of cases.

2. We laud the national government and the private sector for expanding testing capability all around the country over the past few weeks to reaching the goal of 8,000 tests per day. We continue to exhort the expansion of our testing capability to enable mass testing. We reiterate the strategic importance of the continued expansion of our testing capability. This will enable a greater number of our labor force to return to work and hasten economic recovery. The estimate for the United States contained in a roadmap published by the Edmond J. Safra Center of Ethics at Harvard University is between 2 to 6 percent of the population. This estimate, however, is applicable if the economy is going to re-open fully. Given that restrictions will be in place, a capacity similar to that of South Korea, at 15,000-tests per day for Luzon, would be optimal given that South Korea and the island of Luzon both have a population around 50 million people. Testing will help us determine the actual number of Covid-19 cases and prevent infective individuals, particularly asymptomatic cases, from spreading the virus. This will help our policy makers put in place the appropriate rules for each region or area.

3. Contact tracing is the weakest link in our national response to the Covid-19 pandemic. We continue to reiterate the need for an efficient and effective contract tracing system to track those who could have been exposed to the infective individual. To this end, the national and local government should consider hiring and training more personnel to do contact tracing. Where applicable, digital contact tracing apps should be used. While technology like peer-to-peer warning system and other apps are available, there might be limited utility in certain areas where smartphones are not prevalent. More than locating and identifying suspected cases, proper attention must be given to ensure that the person is not put in harm’s way by maintaining privacy. Mass testing without effective contact tracing will only increase the value of R and without stemming possible new infections especially once the restrictions are loosened.

4. We continue to reiterate our recommendation on the necessity of widely available supportive isolation programs and facilities (especially in the LGUs outside of Metro Manila) sufficient to accommodate suspected and confirmed cases with mild or no symptoms especially in areas that will loosen restrictions on mobility. Provisions for food and other forms of social amelioration for patients during the quarantine period might be needed to ensure that they will observe restrictions in place.

5. The ECQ is working and has been critical in reducing transmission and deaths due to Covid-19 in the Philippines. We need to sustain the efficient and effective implementation of quarantines all over the country despite the fact that they are costly and unsustainable initiatives in the short run. The national government needs to shorten the length of time of our quarantines to reduce its negative economic, governance and social impact.  To do so will require government to augment quarantines with massive testing, effective tracing and supportive isolation programs, as is the best practice around the world. Moreover, the government needs to sustain unprecedented levels of collaboration and cooperation between civil society and the private sector.  To do so, the government should ensure that quarantines implemented all over the country are humane and ethical, protective of civil rights and due process, and provide mechanisms that respect health and data privacy.

6. We also recommend that greater effort be given to the expansion of the health system capacity especially those needed to combat Covid-19 like personal protective equipment, ventilators, and isolation rooms. Considering that people afflicted with other diseases would also need a share of the health resources, shortage of health facilities and supplies could result in deaths not because of Covid-19 but due to being crowded out from the Covid-19 strained health system. We need to scale up hiring of health human resources to deal with the medium to long term impact of the pandemic.

7. We continue to reiterate with urgency for the Department of Health (DOH) to hasten the setting up of Covid-19 laboratories and the accreditation of other laboratories across the country. To this end, we are suggesting that the DOH consider utilizing laboratories hosted in DOH-owned hospitals and in state universities and colleges with medical schools and vacant land. This network of Covid-19 laboratories will help enable the massive testing needed in the months to come.

8. The implementation of the quarantine should be made more effective in scaling up the government’s information drive to inform individuals and business of their responsibilities during the new normal especially the guidelines issued by the various agencies.

9. We laud the government for extending the distribution and expanding the coverage of the social amelioration program. The government, however, needs to improve the implementation of this program in order to urgently and expeditiously distribute these resources including relief goods with focus on efficiency, effectiveness, and reduction of pilferage and corruption.

While we have collectively achieved a lot together, we must not be complacent as we are not where we are supposed to be as far as dealing with this pandemic. We are still in the early part of managing this crisis. It is against this backdrop, that we encourage ever greater cooperation and collaboration between government, business, and civil society moving forward.


Forecast Report No. 6-b
(May 11, 2020)

COVID-19 FORECASTS IN THE PHILIPPINES: Addendum to Report No. 6

Read the Addendum in full here.

SHARE ON
TwitterFacebook