COVID-19 epidemic in Malaysia: Impact of lock-down on infection dynamics

Naomie Salim; Weng Howe Chan; Shuhaimi Mansor; Nor Erne Nazira Bazin; Safiya Amaran; Ahmad Athif Mohd Faudzi; Anazida Zainal; Sharin Hazlin Huspi; Eric Jiun Hooi Khoo; Shaekh Mohammad Shithil

Based on the results, we can see that different models have slightly different prediction on the estimated COVID-19 peak dates and a sizeable variation in terms of the maximum number of people infected in Malaysia, as shown in Table 6 . Method 1 based on curve-fitting with probability density function estimated that the peak will be reached on 19 th April 2020 with an estimation of 5,637 infected persons. Method 2 based on SIR estimated between 630,000 to 800,000 persons will be infected at the peak of outbreak in between 20 th to 31 st May 2020 if MCO is in place. Method 3 based on Systems Dynamics estimated that 22,421 persons will be infected at the peak date around 17 th May 2020. The Mean Absolute Percentage Error (MAPE) was calculated using Eq. 4 and the accuracy reported in the table was calculated using Eq. 5 as reported in Section 3.4 above, based on actual data from 18 th March 2020 to 6 th April 2020. The models show different modeling have different estimation and vary in terms of difference from actual data produced by Malaysia Ministry of Health. Models to project the progress of COVID-19 in Malaysia is only as good as the data they are based upon. First, the data needs to be cleaned up so that it reflects the actual numbers. However, epidemiologists have expressed concerns that several factors are affecting the quality of the data. For instance, the number of infected persons per date must be accurate. This data is arguably not accurate when there is a backlog of 8,398 pending COVID-19 test results on April 5, 2020 in Malaysia [7] . As of April 5, a total of 51,937 people was tested, out of which 3,662 tested positive and 39,877 tested negatives. Besides a total of 3,662 confirmed coronavirus cases in Malaysia, the country also reported 61 deaths. Data such as actual dates of when COVID-19 tests are run and dates of when these test results come back is crucial to know the number of people infected on a certain date. Apart from backlog cases that were infected and tested some time ago, it is also uncertain on whether the figures reported each day have excluded data from repeated tests of patients currently under treatment.