key: cord-0749431-bzurjp86
authors: Eyu, Patricia; Elyanu, Peter; Ario, Alex R.; Ntono, Vivian; Birungi, Doreen; Rukundo, Gerald; Nanziri, Carol; Wadunde, Ignatius; Migisha, Richard; Katana, Elizabeth; Oumo, Peter; Morukileng, Job; Harris, Julie R.
title: Investigation of Possible Preventable Causes of COVID-19 Deaths in the Kampala Metropolitan Area, Uganda, 2020-2021
date: 2022-05-17
journal: Int J Infect Dis
DOI: 10.1016/j.ijid.2022.05.033
sha: a6b05791d1c29c41ab1fc2f02e4598c32251ed22
doc_id: 749431
cord_uid: bzurjp86

Background Identifying preventable causes of COVID-19 deaths is key to reducing mortality. We investigated possible preventable causes of COVID-19 deaths over a 6-month period in Uganda. Methods A case-patient was a person testing RT-PCR-positive for SARS-CoV-2 who died in Kampala Metropolitan Area hospitals during August 2020-February 2021. We reviewed records and interviewed health workers and case-patient caretakers. Results We investigated 126 (65%) of 195 reported COVID-19 deaths during the investigation period; 89 (71%) were male and median age was 61 years. Ninety-eight (78%) had underlying medical conditions. Most (118; 94%) had advanced disease at admission to the hospital where they died. Forty-four (35%) did not receive a COVID-19 test at their first presentation to a health facility despite having consistent symptoms. Ninety-five (75%) needed intensive care unit (ICU) admission, of whom 45 (47%) received it; 74 (59%) needed mechanical ventilation, of whom 47 (64%) received it. Conclusion and recommendations Among hospitalized COVID-19 patients who died in this investigation, early opportunities for diagnosis were frequently missed, and there was inadequate ICU capacity. Emphasis is needed on COVID-19 as a differential diagnosis, early testing, and care-seeking at specialized facilities before illness reaches a critical stage. Increased capacity for intensive care is needed.

The coronavirus disease 2019 pandemic began in Wuhan City, China in December 2019 (WHO, 2020a) and quickly spread globally (WHO, 2020c) . In Africa, the first case was registered on 14 February 2020 in Egypt, and as of 20 April 2021, the continent had registered 3.2 million cases and more than 100,000 deaths (WHO, 2020c) . Uganda registered its first COVID-19 case on 21 March 2020. Although small numbers of cases occurred during March-August 2020, the first major wave of COVID-19 in Uganda occurred during August-December 2020, resulting in approximately 32,000 confirmed cases and 200 deaths. This was followed by a larger second wave during April-July 2021, comprising approximately 53,000 recorded cases and 2,400 registered deaths (GoU, 2020).

As the pandemic progressed in Uganda, public health efforts shifted away from detecting every case and towards reducing COVID-19-related morbidity and mortality. In Uganda, district surveillance focal persons (DSFPs) were initially responsible for case investigations, including investigations of COVID-19 deaths. However, during the first wave, the caseload rapidly exceeded the capacity of DSFPs for investigation of individual events. Anecdotal reports from the public suggested that many patients who died with COVID-19 infection in hospitals were not tested until they died. This indicated that there were challenges with accessing care early, with conducting early testing, or both, suggesting opportunities for interventions to reduce COVID-19 deaths.

A comprehensive understanding of specific modifiable factors associated with deaths is essential to reduce case mortality. We investigated the characteristics and possible preventable causes of COVID-19 deaths in Kampala Metropolitan Area during the first wave to recommend interventions to reduce COVID-19 deaths in the country.

This was a descriptive cross-sectional study conducted during February 21-March 20, 2021 in three of the four districts of the Kampala Metropolitan Area (KMA) (Kampala, Mukono, and Wakiso). The three districts where the study was conducted represented 30% of the COVID-19 deaths in Uganda that had been reported at the time. The KMA is currently the second-fastestgrowing urban area in Eastern Africa. It is an industrial, commercial, and educational center and is vital to Uganda's economic growth (Kasimbazi, 2018) .

At the time of the investigation, all persons confirmed to have SARS-CoV-2 infection who had underlying medical conditions putting them at increased risk of severe disease were encouraged to be hospitalized pre-emptively, even if they did not have symptoms (Margini et al., 2020a) . The purpose of this was to ensure they could access medical care should their condition deteriorate.

At the time of the study, reverse-transcription polymerase chain reaction (RT-PCR) was the only test being done to detect SARS-CoV-2 infections in Uganda. A case-patient was defined as a person with a PCR-positive nasopharyngeal or oropharyngeal swab for SARS-CoV-2 who died in a hospital in either Kampala, Wakiso or Mukono districts during 26 August 2020-7 February 2021.

At the time of investigation, KMA only had 3 designated public COVID-19 treatment units, which were in government referral hospitals and treated patients from throughout Uganda.

However, other private hospitals within KMA were also managing COVID-19 patients. Patients from all 33 hospitals in Kampala, Wakiso, and Mukono Districts that were treating any COVID-19 patients at the time, including government hospitals (n=6), private for-profit (PFP) hospitals (n=19), and private not-for-profit (PNFP) hospitals (n=8), were included in this study.

We designed and pretested an investigation tool in Open Data Kit (ODK) which we used to enter information about case-patients. This tool was used to capture data for confirmed COVID-19 deaths at all the 33 hospitals in the three districts from 26 August 2020 to 7 February 2021. We only captured data for case-patients who had sufficient data in their medical records and from interviews to provide a broad picture of their clinical course. A team of investigators from the Uganda Ministry of Health and the Uganda National Institute of Public Health reviewed mortuary records from the 33 hospitals to identify case-patients. We crosschecked the names of the case-patients in the national online laboratory results system to confirm the positive SARS-CoV-2 test result. For case-patients for whom we confirmed a positive result, we retrieved clinical files in the hospitals in which they died and abstracted clinical data and phone numbers of the next-of-kin. We made phone calls to the next-of-kin to make an appointment for a physical meeting. During the physical meeting, we interviewed the next-of-kin to obtain information on case-patient demographics and reconstruct a timeline of events from illness onset to death. We also reviewed the Medical Certificate of Cause of Death form to obtain information on the cause of death. Patients were eligible for inclusion if 'COVID-19' was listed as the primary or secondary cause of death on Part 1 of the form, or if they died while being treated for COVID-19 symptoms and had a positive RT-PCR for SARS-CoV-2 after their death. We reviewed village health team (community health worker) registers, if available for the villages of the case-patients to verify the data already collected. We also conducted interviews with the clinicians to understand their opinions on whether or not death of the case-patient could have been prevented if something had been done differently; if affirmative,

we asked about what could have been done differently. A preventable death was defined for the clinicians as a patient death that could have been avoided with more timely and/or effective health care interventions available in Uganda at the time.

Data from ODK were downloaded into a MS Excel format. Data were analysed in Microsoft Excel. Both demographic and clinical characteristics were descriptively summarised. Age was presented as median and interquartile range (IQR), and the remaining categorical variables were expressed as frequencies and percentages.

Illness severity at admission to the hospital in which the case-patient died was classified according to medical record data, and using definitions from the U.S. National Institutes of Health (National Institutes of Health, 2021). In brief, asymptomatic infection included individuals who tested positive for SARS-CoV-2 but had no symptoms consistent with COVID-19. Mild illness was defined as any of the signs and symptoms of COVID-19 (e.g., fever [≥39.4°C], cough, sore throat, malaise, headache, muscle pain, nausea, vomiting, diarrhoea, loss of taste and smell) but no shortness of breath, dyspnea, or abnormal chest imaging. Moderate illness included evidence of lower respiratory disease (patient with pneumonia without features or signs of severe pneumonia) during clinical assessment or imaging and oxygen saturation (SpO2) ≥94% on room air. Severe illness was SpO2 <94% on room air at sea level, a ratio of arterial partial pressure of oxygen to fraction of inspired oxygen (PaO2/FiO2) <300 mm Hg, respiratory frequency >30 breaths/min, or lung infiltrates in >50% of the lung. Critical illness included respiratory failure, septic shock, and/or multiple organ dysfunction.

The Kampala Metropolitan Area had 195 (67%) of the 290 reported COVID-19 deaths in the country from 26 August 2020-7 February 2021. Among these, 126 (65%) had sufficient clinical data available to be eligible for inclusion. Of the 126 case-patients, 89 (71%) were males. Their median age was 61 years (IQR: 50-71); 78% were ≥50 years of age; 88% were Ugandan citizens and 7% Indian citizens.

Of the 126 deaths investigated, 63% were among case-patients at three hospitals with the highest patient loads. Ninety-eight (78%) case-patients had underlying medical conditions; hypertension (54%) and diabetes (45%) were the most common. Most (90; 71%) case-patients had cough. Ninety-eight (78%) case-patients sought care at a formal health facility at onset of symptoms. Nearly all had severe (51%) or critical (43%) disease on admission to the hospital in which they died (Table 1) . One case-patient with diabetes was hospitalized after testing positive for SARS-CoV-2, despite an absence of symptoms at the time of his admission.

However, he subsequently developed severe disease and died.

Eighty-two (65%) case-patients received a COVID-19 test the first time they interfaced with a health facility for their illness; of these, 91% tested positive. Among the 44 (35%) case-patients who did not receive a test at their first interface with a health facility, failure to suspect COVID-19 by the healthcare worker was the most common reason for not receiving testing (20; 45%), followed by a lack of a testing capacity at the facility (18; 41%) ( Table 1) . Among the 20 casepatients in whom COVID-19 was not suspected, 18 (90%) had symptoms that could have indicated COVID-19 infection (cough, difficulty breathing, chest pain, headache, runny nose, fever, and/or general weakness).

SARS-CoV-2 test results were available for 81% of case-patients on the same day (19%) or one day after (62%) their sample was collected. Thirty (24%) case-patients died within 3 days of first seeking care (Table 1) .

Ninety-five (75%) case-patients needed intensive care unit (ICU) admission (defined as those who were oxygen-desaturated and having severe respiratory distress indicating a need for mechanical ventilation), of whom only 46 (48%) were able to access it. Ninety-nine (79%) case-patients needed high dependency care unit (HDU) (defined as those who were oxygendesaturated but could use normal oxygen masks or continuous positive airway pressure

[CPAP]) admission, and of these, 89 (90%) received it when it was needed. (Table 2 ). Only 46 (61%) of case-patients who needed mechanical ventilation were able to obtain it. Among those who did not receive it, 20 (27%) did not get it due to lack of ventilators or ICU space at their facility (Table 2) .

Among 62 (49%) case-patients with a Medical Certificate Cause of Death form, 50% had COVID-19 listed as primary (26%) or secondary (26%) cause of death (Table 3) . Among 71 patients for whom clinicians felt that death could have been prevented, 36 (51%) of the deaths were considered to have been preventable if ICU space had been available; 23 (33%) were thought to have been preventable if the patient had sought care earlier (33%) ( Table 3 ).

In our evaluation of COVID-19 hospital deaths in the Kampala Metropolitan Area during August 2020-February 2021, we found several modifiable factors present among COVID-19 deaths.

These included missed opportunities for diagnosis and insufficient capacity for advanced patient care. Identifying such factors is critical to designing targeted interventions to reduce COVID-19 deaths. Our data highlight the need for increasing the advanced patient care capacity in the country and enhancing public awareness on the importance of prompt presentation for care if they become seriously ill with symptoms of COVID-19.

In this investigation, receiving hospital care late in the course of illness was frequent among fatal cases. Indeed, more than 80% of patients in this investigation had severe or critical illness by the time they were hospitalised, and one-quarter died within 3 days of seeking care. It is not clear if this was related to late care-seeking by individuals, or if care was sought earlier and patients were turned away. However, ensuring early care-seeking is a well-described challenge in African settings, where patients frequently employ home remedies or traditional healers before seeking more advanced care, even when illness is serious (Wasswa, 2021 , Mpimbaza et al., 2019 , Kasujja et al., 2021 . Beyond this, the cost of hospital care presents a challenge for many people living in low-resource settings, which can discourage early presentation (Peters et al., 2008) . Widespread publicization of the high costs of hospitalization dissuaded at least some patients from seeking hospital care in Uganda (Wasswa, 2021 , Athumani, 2021 , Agiresaasi, Muhumuza, 2021 until late in their course of illness.

Early detection of COVID-19 and effective, timely treatment are important to enable supportive care that can promote patient survival (Skrip et al., 2020 , Hindu, 2021 , Cobre et al., 2020 , WHO, 2021 . The emergency use authorization of the antiviral drugs Paxlovid and molnupiravir in late 2021 to treat COVID-19 has increased the importance of early presentation -both drugs need to be taken within 5 days of symptom onset -and future treatments are likely to make early detection even more relevant (Sara, 2021 , FDA, 2021b , FDA, 2021a . However, many patients in our investigation were not tested at their first presentation for COVID-19. While appropriate supportive treatment for patients in respiratory distress does not necessarily depend on a COVID-19 diagnosis, early during the pandemic in Uganda, lifesaving supportive equipment such as ventilators were largely shifted to COVID-19 treatment units (MoH, 2022 , Margini et al., 2020b , for which admission was restricted to persons who had a positive COVID-19 test (Margini et al., 2020b) . Compounding this issue, both having sufficient test kits and distributing them to facilities presented challenges in Uganda (Padula, 2020 , IOM, 2020 , Wetaya, 2020 as well as in other African countries (Kobia and Gitaka, 2020, Voice of America, 2020) . Among patients in our investigation who were not tested at their first presentation, 41% were not tested due to a lack of testing capacity at the facilities where they presented.

However, the primary reason for not testing patients at their first presentation to a health facility in our investigation was that COVID-19 was not suspected as a differential diagnosis, despite nearly all the patients having symptoms consistent with COVID-19 disease. Equipping lower-level health facilities with rapid diagnostic test kits and training clinicians to have a higher index of suspicion for COVID-19 could have facilitated earlier diagnosis and helped to ensure patients who needed care were able to receive it.

Absence of or limited ICU space was faced by approximately half of the case-patients in this investigation. This is a common problem across low-income countries in Africa (Siaw-Frimpong et al., 2021 , Barasa et al., 2020 : estimates from a study on critical care capacity in 54 African countries in 2020 found an average of 3.1 ICU beds per 100,000 people, fewer than half the beds per 100,000 that exist in upper-middle-income countries (Craig et al., 2020) . An assessment of the ICU bed capacity in Uganda just before the pandemic indicated that 12 out of 14 ICUs reviewed in public facilities were functional, but among these 12 ICUs, only 55 beds were available, for a ratio of 1.3 ICU beds per million population (Atumanya et al., 2019) . In contrast, the World Health Organization recommends that countries have a ratio of 1 to 3 ICU beds per 10,000 persons (Palamim and Marson, 2020) . While the presence of surge capacity is critical to an effective COVID-19 response (Aziz et al., 2020 , Blumenberg and Hendrickson, 2020 , Health., 2020 , it is challenging to consider surge capacity when basic intensive care needs remain unmet. Although efforts on the part of the Uganda Ministry of Health expanded the ICU bed capacity to 97 available beds in the country as of September 2021 (PE, personal communication), this still only yields a ratio of approximately 2.2 beds per million population.

One-third of patients who needed mechanical ventilation were not able to access it, primarily due to lack of ventilators or ICU space. This, too, is a common problem in sub-Saharan Africa (Ruth. and Marks, 2020) . A 2020 report showed that 41 African countries had fewer than 2,000 functional ventilators in total (WHO, 2020b). The ventilator-to-population ratio in Africa was lowest in East Africa, at only 0.23 ventilators per 100,000 people (Craig et al., 2020) . Even when ventilators are available, oxygen demand can present challenges. A media report in June 2021 showed that the demand for oxygen by COVID-19 patients had outpaced the manufacturing capacity in Uganda (Olukya, 2021) . While the Uganda Ministry of Health acquired 109 critical care ventilators in 2021 (Dr. Mary, 2021) , and an increase in medical oxygen production and donations of oxygen gas cylinders to support the COVID-19 response (ReliefWeb., 2021 , UNICEF., 2021 , there is still a shortage of trained staff to operate them. In 2019, there were 171 nurses working in 12 ICUs with only 13 that had been formally trained to work as critical care nurses (Atumanya et al., 2019) . Recent advertising to hire intensive care specialists (Dr. Mary, 2021) might help address some of these shortages.

Our investigation faced a number of limitations. First, we did not have a control group with whom to compare the deaths presented in this paper. Thus, we cannot identify some exposures, such as late presentation, as risk factors for death. Second, we did not capture the patients' condition at first presentation to a health facility nor did we capture if the first facility was the one in which they died, and thus cannot evaluate their care-seeking behaviors, if any, early in the course of their illness. Third, this investigation was retrospective and subject to both recall bias and/or perhaps social desirability bias (interest in saying the 'right thing') during the interviews. For example, very few patients reported going to a traditional healer, yet this is a common practice in many parts of Uganda. We tried to mitigate this issue by using a variety of data sources to triangulate information, but it is possible that there are some inaccuracies. Fourth, we conducted this investigation in only three districts of the country, all in the capital region, and thus the data are not representative of the entire country. However, it is worth noting that the challenges identified would only be expected to be greater in other areas of the country, where access to care is less and transport and diagnostics are more difficult to obtain. Thus, it is likely that the recommendations that apply to the Kampala Metropolitan Area may apply more strongly to other areas of the country. Fifth, not all patients had sufficient data to build a complete picture of their clinical course and death, which might result in bias if these patients were qualitatively different from those with data. Sixth, we asked clinicians' opinions on whether or not deaths were preventable; these were anecdotal and not based on any validated metrics. Finally, we were unable to capture community deaths, which are undoubtedly different from facility deaths. Thus, our data apply only to patients who were able to reach the hospital in this area.

In conclusion, we identified late receipt of specialized care, missed opportunities for diagnosis, lack of ICU space, and lack of mechanical ventilation as possible contributing factors to COVID-19 patient deaths in Kampala Metropolitan Area during the first wave of the pandemic. Our data suggest that there might be value in making the public aware of the importance of seeking care before patients reach a critical stage of COVID-19 disease, to both improve outcomes and reduce the impact of critical cases on the limited ICU resources. Clinicians across all health facility levels could be re-sensitized to consider COVID-19 as a differential diagnosis and test early when patients present with consistent symptoms. In addition, plans for surge capacity in ICUs and with hospital staff could facilitate improved outcomes during future waves of infection. Signs and symptoms at first presentation to health facility

COVID Patients Saddled With Steep Bills at Private Hospitals [Online

Court Orders Ugandan Government to Regulate COVID-19 Treatment Costs

Assessment of the current capacity of intensive care units in Uganda; A descriptive study

Managing ICU surge during the COVID-19 crisis: rapid guidelines

Assessing the hospital surge capacity of the Kenyan health system in the face of the COVID-19 pandemic

Risk factors associated with delay in diagnosis and mortality in patients with COVID-19 in the city of Rio de Janeiro

Estimating critical care capacity needs and gaps in Africa during the COVID-19 pandemic. medRxiv

ICUs in Uganda: An Interview with Dr. Mary Nabukenya

Update: FDA Authorizes Additional Oral Antiviral for Treatment of COVID-19 in Certain Adults

Update: FDA Authorizes First Oral Antiviral for Treatment of COVID-19 [Online]. US Food and Drug Administration

Ministry of Health, Uganda: Government of Uganda

COVID deaths due to late detection

Urban Expansion in the Greater Kampala Metropolitan Area

Understanding the diagnostic delays and pathways for diabetes in eastern Uganda: A qualitative study

COVID-19: Are Africa's diagnostic challenges blunting response effectiveness? 3, 4

Uganda"s Emergency Response to the COVID-19 Pandemic: A Case Study

Case Study: The Initial COVID-19 Response in Uganda

Uganda's Emergency Response to the COVID-19 Pandemic: A Case Study

Pillar: Case Management

Understanding the context of delays in seeking appropriate care for children with symptoms of severe malaria in Uganda

In Uganda, disputes over bills mark chaotic COVID-19 care

NATIONAL INSTITUTES OF HEALTH, N. 2021. Clinical Spectrum of SARS-CoV-2 Infection. COVID-19 Treatment Guidelines. United States: National Institutes of Health

Uganda hospitals run short of oxygen amid spike in COVID-19 cases

Why Only Test Symptomatic Patients? Consider Random Screening for COVID-19. Appl Health Econ Health Policy

COVID-19 -The Availability of ICU Beds in Brazil during the Onset of Pandemic

Poverty and Access to Health Care in Developing Countries

UK officially hands over life-saving oxygen cylinders to support Uganda's COVID-19 response

10 African Countries Have No Ventilators. That"s Only Part of the Problem

What doctors wish patients knew about COVID-19 oral antivirals

Capacity of intensive care units in Ghana

Clinical management and mortality among COVID-19 cases in sub-Saharan Africa: A retrospective study from Burkina Faso and simulated case analysis

First UNICEF Oxygen Plant-in-a-Box heads to Uganda to help with COVID-19 response

African Nations Try to Overcome Shortage of COVID-19 Test Kits

Covid-19: Uganda"s low inpatient numbers mask high community infection as desperate patients turn to herbs

Ugandan scientists develop quick, low-cost COVID-19 test kits

Archived: WHO Timeline -COVID-19

COVID-19 pandemic expands reach in Africa

Online: World Health Organisation

WHO. 2021. Promoting Cancer Early Diagnosis

We acknowledge Uganda Ministry of Health and Baylor College of Medicine Children's Foundation Uganda for the technical and logistical support.

The authors declare that they donot have any conflicts of interest.

The Uganda Ministry of Health (MOH) gave the directive to conduct epidemiological investigations and response to COVID-19 disease in the country. We sought permission from the local council authorities of residences of the case-patients to undertake the investigation.We obtained verbal consent from the next-of-kin before conducting the interviews. The questionnaires were kept under lock and key to prevent disclosure of personal information of the respondents to individuals who were not part of the investigation. This activity was reviewed by CDC and was conducted in a manner consistent with applicable federal law and CDC policy. § §See e.g., 45 C.F.R. part 46, 21 C.F.R. part 56; 42 U.S.C. §241(d); 5 U.S.C. §552a; 44 U.S.C. §3501 et seq.

We did not have funding to conduct this study.

Don't know 2 2

Received mechanical ventilation 46 61Not received due to lack of ventilators or ICU 20 26Not received due to lack of oxygen 2 3Not received because no space in ICU 5 6Not received because patient rejected ventilation 2 3

Cardio-pulmonary resuscitation 41 65Oxygen therapy 22 35HDU: High Dependency Unit, ICU: Intensive Care Unit If there was earlier diagnosis 3 4If comorbidity was properly managed 3 4If ventilation was not declined 2 3If laboratory results were received early 2 3If patient adhered to medical advice 1 1If the referral process was faster 1 1 ICU: Intensive Care Unit, HDU: High Dependency Unit