key: cord-1012493-mxzdbvea authors: Rahaman Khan, H.; Howlader, T.; Islam, M. M. title: Battling the COVID-19 Pandemic: Is Bangladesh Prepared? date: 2020-05-05 journal: nan DOI: 10.1101/2020.04.29.20084236 sha: 6bec9ae984b9c9c04710311e96c6324559d37bea doc_id: 1012493 cord_uid: mxzdbvea Following detection of the first few COVID-19 cases in early March, Bangladesh has stepped up its efforts to strengthen capacity of the healthcare system to avert a crisis in the event of a surge in the number of cases. This paper sheds light on the preparedness of the healthcare system by examining the spatial distribution of isolation beds across districts and divisions, forecasting the number of ICU units that may be required in the short term and analyzing the availability of frontline healthcare workers to combat the pandemic. As of April 18, COVID-19 cases have been found in 53 of the 64 districts in Bangladesh with Dhaka District being the epicenter. Seventy-five percent of the cases have been identified in 5 neighboring districts, namely, Dhaka, Narayanganj, Gazipur, Narsingdi and Kishoreganj, which appear to form a spatial cluster. However, if one takes into account the population at risk, the prevalence appears to be highest in Narayanganj, followed by Dhaka, Gazipur, Narsingdi and Munshiganj. These regions may therefore be flagged as the COVID-19 hotspots in Bangladesh. Among the eight divisions, prevalence is highest in Dhaka Division followed by Mymensingh and Barishal. The number of cases per million exceeds the number of available isolation beds per million in the major hotspots indicating that there is a risk of the healthcare system becoming overwhelmed should the number of cases rise. This is especially true for Dhaka Division, where the ratio of COVID-19 patients to doctors appears to be alarmingly high. Mymensingh Division also has a disproportionately small number of doctors relative to the number of COVID-19 patients. Using second order polynomial regression, the analysis predicts that even if all ICU beds are allocated to COVID-19 patients, Bangladesh may run out of ICU beds soon after May 4, 2020. We conclude that in spite of a significant increase in hospital capacity during 2005-15 and a 57% rise in the number of doctors during the same period, the healthcare system in Bangladesh and Dhaka Division in particular, may not be fully prepared to handle the COVID-19 crisis. Thus, further steps need to be taken to flatten the curve and improve healthcare capacity. have already spread to 53 of the 64 districts; 6059 isolation beds are ready; 595 doctors, 546 nurses, 130 medical technologists and 350 other healthcare staff are allocated to treat COVID-19 patients (IEDCR, 2020) . The burning question is, 'Are these preparations adequate to fight the pandemic in Bangladesh?' On March 7, when more that 100,000 people of 100 countries were affected by COVID-19, the World Health Organization (WHO) declared COVID-19 as a global pandemic (Shaw et al., 2020) . Since then, majority of countries have imposed travel bans and lockdown of cities to stop the spread of COVID-19 and healthcare systems have taken unprecedented measures to expand their capacities to handle the surge. For instance, on 3rd February, China started sanitizing its city streets, public places, parks etc and later they introduced QR code for all residence to separate infected and non-infected people. Immediately after the outbreak, Wuhan city of China built two hospitals (30 ICU units and 1000 patients capacity) and several mobile cabins within a short period of time to meet excessive demand of hospital beds (Gao & Yu, 2020) (Christopher et al., n.d.) . Meanwhile, the Ministry of Foreign Affairs, South Korea introduced government-ment implemented distribution of protective equipment through national security card which prevented panic buying of people [Wang et al. (2020) , Shaw et al. (2020) , Wang et al. (2020) ]. Taiwan implemented its previous SARS epidemic experience to tackle the challenge by implementing modified Traffic Control Bundling (TCB) tool which provided small infection rate among health professionals during SARS outbreak in 2003 (Schwartz et al., 2020) . After mainland China, Europe has become the active center of COVID-19 -according to WHO's declaration as of 13 March 2020 whereas the first case was reported in France on 24 January 2020 (Spiteri et al., 2020) . About 47 laboratories in 31 of total 44 European countries had diagnostic capacity of COVID-19 by 31 January 2020 with a minimum facility of 8275 tests per week [Spiteri et al. (2020) , Reusken et al. (2020) ]. Germany made its first COVID-19 crisis response by closing border with neighboring countries and prohibiting the export of medical logistics including mask and other protective equipment (European Policy Institutes Network, 2020) . The country has been expanding its intensive care facilities since January 2020, from 28000 intensive care beds to 40000 intensive care beds by far (Bennhold, 2020) . On the other hand, Italy has suffered due to lack of effective measure and immediate response to the emergency, supply of protective equipment and inadequate medical stuffs. On March 11, one fifth of its ICU beds (1028 of 5200) were filled with COVID-19 patients (Remuzzi & Remuzzi, 2020) . The country has total 5256 American Health Association (AHA) registered hospitals running with 534964 stuffs where 2704 of these hospitals provided ICU services through 96596 ICU beds. The Society of Critical Care Medicine advocates hospitals to implement a tiered staffing strategy during COVID-19 pandemic (Halpern et al., 2020) . In recent days COVID-19 infection has been increasing in South Asian countries. Despite having limited resources, Bhutan has firm control so far by early-sealing of its borders and monitoring the status regularly (Diplomat Risk Intelligence, 2020). For Bangladesh, COVID-19 is a humanitarian crisis with a public health dimension (World Economic Forum, 2020) . While there are important lessons to be learnt from China, Italy, UK, USA, Spain, France and other developed countries, it may not be possible to adopt many of their policies in Bangladesh due to scarcity of resources. Developed countries have been able to invest heavily in their healthcare systems and this has enabled them to respond effectively to the COVID-19 pandemic. China, for example, has a health 4 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 5, 2020. . expenditure per capita that is 10 times that of Bangladesh (World Economic Forum, 2020) . With limited resources, expanding healthcare capacity remains a challenge for Bangladesh. In 1980 there were only 28 ICU beds in Dhaka city. Since then the number of ICU beds has gradually increased (Nafseen, 2018) . There are about one hundred hospitals with ICU facilities in Bangladesh and 80% of them are located in Dhaka ("Message from president. Criticon Bangladesh 2018", 2018). Hospitals in Bangladesh currently have a total of 1,169 ICU beds. Out of these, 432 are in government hospitals and only 110 are outside the capital Dhaka, and 737 are in the private hospitals (Khan & Hossain, 2020). These numbers are reasonably low if considered against a population of 170 milion ("The Daily DhakaTribune", March 21, 2020). According to (Kennedy & Pronovost, 2006) , the total number of ICU beds in a hospital should be between 5% and 12% depending on the care given by the hospital. In 2017-18, the total number of beds in hospitals were 1,27,360. Out of these, 48,934 were in government hospitals and 78426 were in private hospitals [ (Nafseen, 2018) . According to ("The Daily DhakaTribune", March 21, 2020), Bangladesh currently has a total 141,903 hospital beds or 0.84 beds per 1000 people. Whether these resources are sufficient to tackle the COVID-19 pandemic requires a more indepth analysis. This paper tries to answer this question by examining the capacity of the healthcare system in Bangladesh in relation to the COVID-19 pandemic. The data are collected from two difference sources in Bangladesh. Information on number CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 5, 2020. . . Moreover, all of the maps presented in this study are produced by "mapReasy", an R package for producing administrative maps Islam et al. (2017) . We produced a number of spatial graphs with the above R package. We also predicted the number of COVID-19 infected people for a short period of time using second order polynomial regression. about 45% of all cases in the country up till April 18. The contagion has spread to all districts that share a border with the epicenter. However among these, Dhaka's neighboring district to the east, Narayanganj, appears to be the hardest hit registering 309 cases. Three neighboring north eastern districts, namely, Gazipur, Narsingdi and Kishoreganj, have also seen alarming numbers of cases. These five regions seem to form a spatial clus-6 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 5, 2020. . ter that accounts for 78% of all cases in Bangladesh. Other emerging hotspots include the port city of Chattogram, which is a major business center and Munshiganj district, which shares borders with Dhaka and Narayanganj. It is important to note, however, that the above results do not take into account the population size of each district, and therefore may not reflect the true scenario regarding the levels of infection in the populations. Figure 1 (right panel) shows the spatial distribution of laboratory confirmed cases of COVID-19 per million across districts. Narayanganj now replaces Dhaka as the number one hotspot with 88-90 cases per million while Dhaka is second with 64-66 cases per million. The third most affected area is Gazipur with 37-39 cases per million followed by Narsingdi which has 34-36 cases per million. In contrast to the previous graph, Figure 1 (right panel) indicates that Munshiganj has a higher number of cases per million than Kishoreganj. Furthermore, Chattogram is now no longer among the emerging hotspots having only 4-6 cases per million. Bangladesh is divided into 8 administrative divisions. Figure 2 (left panel) shows the spatial distribution of laboratory confirmed cases of COVID-19 according to division. Dhaka Division is the epicenter with a total of 1606 cases up till April 18. The next highest number of cases have been reported in Chattogram Division, which is located to the 7 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 5, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 5, 2020. . https://doi.org/10. 1101/2020.04.29.20084236 doi: medRxiv preprint the number of available isolation beds are highest in Sherpur, Sylhet and Kishoreganj. Among the southern districts, Barishal, Chattogram and Rangamati possess the highest number of isolation beds, i.e. between 201-300, followed by Bhola, which has 176-200 isolation beds. The distribution of isolation beds suggests that some of the districts with high prevalence of confirmed cases could be at a risk of experiencing shortages should the number of cases rise. For instance, Naranganj district, which is the second major hotspot after Dhaka city with 309 cases has between 126-150 isolation beds. Other COVID-19 hotspots, namely, Gazipur and Narsingdi, which have already registered 161 and 93 cases, respectively, have less than 50 isolation beds. Emerging hotspots such as Munshiganj, Madaripur and Gopalganj, which have between 20-39 confirmed cases also have less than 50 isolation beds. In contrast, health system facilities in districts such as Bhola, Cox's Bazar, Naogaon, Joypurhat and Bogura are less likely to be overwhelmed as the number of cases (< 3) is far less than the available number of isolation beds (between 151 − 200). However, this is true only as long as the number of cases remains fairly low. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 5, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 5, 2020. . Figure 6 shows that 11 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 5, 2020 . . https://doi.org/10.1101 One alarming finding is the disproportionately small number of doctors in relation to the number of cases in Dhaka Division resulting in a very high ratio. The implication is a high risk of the healthcare system being overwhelmed due to acute shortage of doctors, who form the frontline defence in the fight against the pandemic. It appears that Mymensingh Division may also be in a vulnerable position due to the number of doctors being nearly a quarter of the number of COVID-19 patients. Chattogram Division, which has the second highest number of cases after Dhaka, appears to have a ratio just less then 1 but the ratio could easily increase if the number of cases rise and more doctors cannot be provided to treat these cases. A similar situation exists in Barishal Division as well. 12 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 5, 2020. . On the otherhand, Khulna, Rajshahi and Sylhet have much smaller ratios and therefore appear to be in a better position to handle the crisis. We analyzed the COVID-19 data of Bangladesh for predicting the number of infected cases by following the second order polynomial regression as suggested by (Khan & Hossain, 2020), (Khan, 2020). Polynomial regression has been used in forcasting COVID-19 diseases along with for fitting trends by many researches [Pandeya et al. (2020) , Johannes (2008), Howard (1943) ]. During an epidemic, projections on the number of patients who may require intensive care are very useful to hospital administrators for planning. We therefore use the data till April 18 to obtain projected number of ICU patients in 13 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 5, 2020. . https://doi.org/10. 1101 Bangladesh. According to (Khan & Hossain, 2020) , this projection is suitable for a short term period and hence projections were made for the period April 19 to May 18, 2020 based on the data reported from April 1 to April 18. Table 1 shows the projected numbers of infected people, hospital admissions, and ICU patients along with their associated 95% confidence intervals. The number of hospital admissions are estimated based on the findings in [Phua et al. (2020) ] that 20% of infected patients require hospital admissions. Indirectly, this gives us estimates for the required number of isolation beds assuming an admitted person requires a single bed. Furthermore, according to (Phua et al., 2020) , 12% of all reported cases need ICU admissions. This result was used to obtain the projected numbers of ICU patients reported in Table 1 These findings sound the alarm that the healthcare system may not have the capacity to handle critical Covid-19 patients due to insufficient number of ICU beds. Following detection of the first few COVID-19 cases in early March, Bangladesh has stepped up its efforts to strengthen the capacity of the healthcare system to avert a crisis in the event of a surge in the number of cases. This analysis sheds light on the preparedness of the healthcare system in terms of the spatial distribution of number of isolation beds, availability of ICU beds and the availability of frontline healthcare workers to combat the pandemic. COVID-19 cases have been found in 53 of the 64 districts in Bangladesh as of April 18. Seventy-five percent of the cases have been identified in 5 neighboring districts that appear to form a spatial cluster of COVID-19 cases. These regions include 14 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 5, 2020. . Dhaka, Narayanganj, Gazipur, Narsingdi and Kishoreganj with Dhaka being the epicenter. However if one takes into account the population at risk, the prevalence appears to be highest in Narayanganj, followed by Dhaka, Gazipur, Narsingdi and Munshiganj. These regions may therefore be flagged as the COVID-19 hotspots in Bangladesh. Among the eight divisions, prevalence is highest in Dhaka Division followed by Mymensingh and Barishal. Yet Dhaka Division has the lowest number of isolation beds per million of its population. A similar situation is seen in Mymensingh Division which also has high prevalence of the disease but a comparatively small number of isolation beds per million. On a finer resolution, district-wise comparisons reveal that the epicenter, Dhaka District, with 64-66 cases per million has one of the lowest, i.e. between 11-20, isolation beds per million. Narajanganj, which is the major hotspot with 88-90 cases per million has only 31-40 isolation beds per million. The third hotspot is Gazipur with 37-39 cases per million and less than 10 isolation beds per million, which is the lowest among all the districts. Narsingdi, which is the fourth largest hotspot also has a small number of isolation beds relative to the number of cases per million of its population. These figures indicate that there is an elevated risk of the healthcare system becoming overwhelmed in the major hotspots in Bangladesh. On the positive side, the hilly district Rangamati has the highest number of isolation beds and lowest number of cases per million in Bangladesh. With regard to availability of healthcare resource persons, the analysis finds that doctors form the largest proportion of healthcare staff in Bangladesh in five of the eight divisions. The proportions of doctors and nurses are nearly equal in Khulna and Sylhet Divisions. Rajshahi Division appears to be an outlier with the proportion of doctors being less than both the proportion of nurses and the proportion of other healthcare staff. An alarming finding is the very high ratio of COVID-19 patients to doctors in Dhaka Division. Mymensingh Division also has a disproportionately small number of doctors available to treat COVID-19 patients. The analysis has revealed that the numbers of isolation beds and doctors available to treat COVID-19 patients are worryingly low in the major hotspots in Bangladesh. This finding has important implications for policy. In a densely populated country like Bangladesh where majority of the people are poor, it is difficult to effectively enforce social distancing 15 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 5, 2020. . https://doi.org/10. 1101 and other preventive measures. Thus, there is a fear that the number of cases could flare up at any time during the course of the pandemic. In these circumstances, the healthcare system must be adequately prepared to face the looming crisis by creating excess capacity and mobilizing resources to the most affected areas. Based on the analysis, we recommend that the number of isolation beds in Dhaka District (which includes Dhaka city) as well as Gazipur District be tripled so as to be on par with the number of cases. In Narayanganj District the number of isolation beds should at least be doubled. Projections indicate that there is a risk of a deficit in the number of ICU beds for critical patients. Thus, increasing the number of ICU units should also be a priority. Strategies for increasing capacity could include freeing up more beds for COVID-19 patients in hospitals, increasing the number of COVID-19 designated healthcare facilities by roping in private hospitals and clinics, and repurposing government buildings and setting up camps as an emergency measure should the outbreak go out of hand. With the number of doctors available to treat COVID-19 patients being dangerously low in Dhaka Division, we recommend increasing the number of doctors by several folds. This could be done by recruiting new doctors in hospitals including final year medical students and calling retired doctors back to work. There have been reports that some doctors avoid work due to fear of contracting the virus due to lack of PPE. As a result, both COVID-19 patients as well as patients with other medical conditions may not receive adequate medical care. Lack of PPE may therefore be identified as one of the leading causes of shortages in medical personnel. We recommend that in addition to offering attractive incentives to frontline healthcare workers, the government must by all means ensure that medical personnel on duty receive PPE and training on how to use them. This is one of several measures that need to be taken to ensure the safety of healthcare workers and their families so as to maximize their participation in the fight against the pandemic. COVID-19 preparedness in Bangladesh must include an effective and feasible plan of action in which guidance on resource management and communication and coordination between national health stakeholders are of utmost importance. If preparations are taken in a pragmatic way, Bangladesh will win the COVID-19 war with less fatalities since it has a strong network of community health workers, a history of success in public-private partnerships during emergencies like floods and cyclones, and people with incredible levels of resilience. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 5, 2020. . Population projection of Bangladesh: dynamics and trends 2011-2061. Bangladesh Bureau of Statistics (BBS), Statistics and Informatics Division A German Exception? Why the Countrys Coronavirus Death Rate European Policy Institutes Network Public governance mechanism in the prevention and control of the COVID-19: information, decision-making and execution US ICU resource availability for COVID-19 Fitting polynomial trends to seasonal data by the method of least squares Institute of Epidemiology Disease Control And Research Bangladesh Coronavirus Update mapReasy: Producing Administrative Boundary Map with Additional Features Embedded Bangladesh Society of Critical Care Medicine Critical Care Medicine: Bangladesh Perspective Seir and regression model based COVID-19 outbreak predictions in india Intensive care management of coronavirus disease 2019 (COVID-19): challenges and recommendations COVID-19 and Italy: what next? The Lancet Laboratory readiness and response for novel coronavirus (2019-nCoV) in expert laboratories in 30 EU/EEA countries Protecting health care workers during the COVID-19 coronavirus outbreak-Lessons from Taiwan's SARS response Governance, technology and citizen behavior in pandemic: Lessons from COVID-19 in East Asia others (2020). First cases of coronavirus disease 2019 (COVID-19) in the WHO European Region Response to COVID-19 in Taiwan: big data analytics, new technology, and proactive testing The World Bank data We declare that we have no competing interests. There is no funding for this study. The working data set used for this study has been submitted to the journal as additional supporting file.