Drivers of the Second Wave and clinical characteristics of COVID-19 cases in Uganda: A Retrospective Study of Conrmed SARS-CoV-2 cases, March-June, 2021

The COVID-19 continued to pose several public health, social, economic challenges and the drivers for the occurrence of different COVID-19 waves remains undocumented in Uganda. We conducted a cross-sectional population-based survey among recovered COVID-19 cases to establish the drivers of SAR-CoV-2 infections. We performed a retrospective study and interviewed 1120 recovered COVID-19 cases from 10 selected districts in Uganda. We further conducted 38 Key Informant Interviews of members of the COVID-19 District Taskforce and 19 in-depth interviews among COVID-19 survivors from March to June, 2021.


Characteristics of COVID 19 cases
a. Social demographics: Of the COVID-19 cases interviewed, more than half 51.5% (577/1120) were females. We found increased numbers of cases across all age groups and more occurrences among the young and middle age groups (30-39 years) at 26.8% (300/1120). Overall, we found increased cases up to 62% among age group 39 years and below ( gure 2). When we adjusted for age, the majority of the cases were among 40 years and above.
c. Admission Status Only 9.1%, of the COVID-19 positive cases were admitted to health facilities (Table 1). According to age group, most cases 31.3% (130) with underlying conditions were aged 40 years and above. However, an increased number of young people (13-39 years) cases ranging from 13% (12) to 21% (63) reported having underlying conditions ( Table 2). Among those cases aged 40 years and above, 31.3% (130) had underlying conditions and many of them were admitted either required oxygen, ventilation or admission to ICU as summarized in Table 3. The most commonly encountered underlying conditions were high blood pressure, diabetes and asthma. High blood pressure and diabetes were signi cantly associated (P<0.001) with low levels of survival.  (Table 6). vaccinated with at least one dose of the vaccine were 6 times more likely to survive compared to those not vaccinated as per adjusted prevalence ratio 6.1 (3.24-11.57) ( Table 4). A small proportion (17.8%, 199/1120) were asymptomatic (Table 1). At bivariate analysis, results showed that; not seeking care (CPR 1.99, P-value 0.003), not being admitted (CPR 2.15, P 0.013) and other household members not having symptoms (CPR 1.52, P 0.001) were positively associated with being asymptomatic among the COVID-19 cases (Table 5). While a household size of greater than 9 members (CPR 0.63, P 0.025), and having contact with others two weeks before testing (CPR 0.38, P 0.000) were likely to be symptomatic among the COVID-19 cases. The details of the bivariate analysis (  (Table 1). The contacts with COVID-19 like symptoms persons were mainly with family members and friends (54.9%), co-workers (25.2%), and classmates for students (9.0%) as per stamen quotes (table 8). Burial places were also reported to have contributed to further spread as bereaved members of the families and communities insisted on cultural practices of burials that increased congestion without adhering to recommended SOPs (Table 8).
c. Myths, misconceptions, and misinformation: Respondents of the KIIs reported that some members of the communities studied believed there was no COVID-19 and therefore ignored the instructions for observation of SOPs while others depended on the fake social media news to inform their response to instituted SOPs (Table 8).

d. Politics:
Respondents reported that the political season that started in August, 2020 till February, 2021 as the country was undergoing the rst wave of COVID-19 could have facilitated wide community transmissions triggering the second wave of COVID 19 (Table 8).
e. Schools: Respondents mentioned that the number of COVID-19 cases among the on-going school pupils and students had increased and were being under reported. Upon closure of the schools as part of the lockdown instituted early in June 2021, some pupils and students from boarding schools who were asymptomatic for COVID-19 returned home and unknowingly infected members of their families (Table  8).
f. Weak Health Systems: Respondents reported that inadequate resources in most health facilities across the country led to most of the con rmed COVID-19 cases being sent home for HBC management. This led to having so many positive COVID-19 cases in the communities that might have led to rapid spread of infections in the community highlighted by high positivity rates recorded in June, 2021 (table 8). Furthermore, respondents castigated that lack of resources for nationwide sensitization, asymptomatic infected health workers returning to their families, congestion in lower health facilities, limited testing centers, congestion at various trade points exchanging money and goods were among the drivers of COVID-19 in the second wave ( Table 8).

g. Stigma to COVID-19
Respondents reported several non-disclosures of the COVID-19 status especially among asymptomatic cases to avoid being discriminated and harassed in the community leading to indirect exposure of the virus to other unsuspecting members of the community (Table 8).

Discussion
In this study, we assessed the drivers of the second wave of SARS-CoV-2 infections between March and June 2021 from 10 districts in Uganda. In the second wave of COVID-19, we found a slightly higher proportion of female cases compared to males. Our results represent a shift from the rst wave where males were mostly affected (12,13) as has been reported elsewhere (14,15). In our study, the change in gender infection status with more females being infected and together with their social roles in families and communities facilitates close interactions with households and communities with more likelihood of increasing transmissions. We further noted increased cases among all age groups with more cases recorded in the young people aged 19 to 39 years that constituted the highest percentage (62%) of infections in the second wave. Again, our results re ect a change in the risk groups in the second wave where young people including the school going age were infected and probably escalated the spread infections in their communities. Previously in several studies, the virus was more reported in adults aged 40 years and above including disease severity presentation (16). In our current study, we found that the virus was affecting all age groups especially the young ones. We also report mortalities ranging from 1-3.7% among the infected young ones aged 13 to 39 years which was not the case in the rst wave. We strikingly noted high cases of underlying conditions (high blood pressure and diabetes) among the young COVID-19 positive cases aged 20-39 years. This observation is surprising and may explain the increased numbers of severe cases and hospitalizations observed and reported in the second wave. Whereas it has been severally reported that COVID-19 remains limited in young ones in terms of numbers, disease presentation and clinical outcomes, our study suggests otherwise. We think, there has been limited attention and focus to this age group as most cases would probably remain asymptomatic and rarely tested. During KIIs, it was reported that the COVID-19 positivity rate was high up to 70% among students returning from boarding schools upon closure of schools in the second lockdown in June, 2021. Hence, our results call for a shift in outbreak response strategies to address the current disparities and prioritize women and young generations for interventions like vaccinations and speci c awareness messages targeting this category to prevent further widespread of infections.
We found that the majority of the cases reported having several and varying symptoms during the course of the disease where most of them reported cough, headache, runny nose, fever, and general body weakness as previously reported (17,18). We further observed poor healthcare seeking behaviours among the COVID- 19  Our study further noted that a signi cant number of respondents pointed to health workers as a possible driver of the second wave because they lacked adequate Personal Protective Equipment which compromised their work and exposed them to the risk of being infected. The exposed health workers before testing positive continued to interact with other patients, members of their families and communities, an exposure factor for virus transmission. One more critical area of concern identi ed during our study, was social gatherings that continued to take place unbated despite of government directives on social gatherings like burials, weddings, churches, bars and restaurants, salons, markets, public transport and schools. SOPS like wearing of facemasks, social distancing of at least 2 meters, minimum numbers recommended of some social functions and hand washing with soap/sanitizers were not being observed, ignored or even completely forgotten. Respondents of KIIs and in-depth interviews castigated that the non-adherence to SOPS for social gatherings accelerated the number of cases in most communities observed in the second wave. Even the schools that were opened in the staggered manner with prior preparations and clear instructions to curtail transmissions within the schools, became a seed bed for COVID-19 transmissions. The schools aunted instructions and some concealed information of COVID-19 cases for fear of being closed. By the time the schools were closed again in June, 2021, the cases both identi ed and unidenti ed were very high and further contributed to community transmissions upon returning home. As much as our school situation and operational settings may be different with so many boarding schools compared to other regions of the world, schools (students and teachers) had been reported as one of the super spreaders of SARS-CoV-2 (21,22). The social gatherings were further fueled by stigma, social media misinformation and falsi cations that circulated widely about COVID-19 that affected many of the instituted prevention measures as also reported elsewhere (23

Study Location
The study was conducted in 10 selected districts in Uganda. Uganda is a land locked country that lies between 1 0 29' South and 4 0 12' North latitude, 29 0 34' East and 35 0 0' East longitude (27).
Uganda has a population of 41.6 million people based on the Uganda National Household Survey (UNHS) conducted in 2019/20 by the Uganda Bureau of Statistics (UBOS). More than half (54%) of the population is below 18 years of age. Uganda, just like other Sub-Saharan African countries, has a weak health care system characterized by low clinician to patient ratio, limited laboratory capacity, poor administration, and limited resources (28, 29).

Study Setting
In this study, we selected 10 districts ( gure 1) representing the main geographic regions that had the highest number of COVID-19 cases as reported by the MOH (5). The selected districts were the border districts (Busia and Tororo) with Points of Entry (PoE); major road highways for transit of cargo across districts (Mbale, Gulu, Luwero, Soroti, and Moroto districts); and highly populated regional city districts (Wakiso, Gulu, Mbarara and Kampala) (30, 31).

Study population
The study population included patients or care givers (especially for children below 18 years) of people who had suffered and recovered from COVID-19, either after HBC or discharge from health facilities. The retrospective cross-sectional study was done as part of the outbreak investigation from March to June 2021 for PCR/RDT con rmed cases.
Sample size and sampling procedure Sampling: We selected 10 districts based on their high population densities, high incidences of COVID-19 cases from March 2021 to June 2021 exceeding 300 cumulative cases in the study period, and having PoEs within the districts. We obtained data on COVID-19 positive RDT/PCR results from the MOH COVID-19 lab investigation forms available at respective health facilities in the study districts (sample form appendix 1). The information extracted was then used to systematically sample (Table 7) and locate the recovered COVID-19 cases who were interviewed in the community.  The eld team further accessed laboratory investigation forms of the COVID-19 PCR and RDT positive cases from the laboratories of the selected health facilities in each of the selected districts to extract data on variables such as; social demographics and clinical symptoms. The collected information was then used to locate the recovered COVID-19 cases in respective communities guided by the Community health workers. The selected cases were called via telephone to arrange appointments before the visits. On the day of the visit, the eld team consented the participant and then collected data from each participant using a community COVID-19 case questionnaire that was adopted from the MOH standard tool which assessed the social demographics and clinical characteristics of the COVID-19 positive cases. All the data collected on tablets was uploaded daily onto a mWater portal server secured with passcodes that was only accessed by the principal investigators.

Analysis
The collected data was exported and cleaned using MS Excel 2016 (Microsoft Corporation, Redmond, WA) and analyzed using STATA 15.0 statistical software (StataCorp, Texas USA). Descriptive analyses were performed for demographic characteristics, and clinical characteristics of the COVID-19 cases were presented as frequencies, proportions and means where appropriate. Being either symptomatic (coded 0) or asymptomatic/not symptomatic (coded 1) was considered as the outcome variable. To assess the association between the outcome variable and the explanatory variables, we considered a generalized linear model of the Poisson family with a logarithm as the conical link function with a robust error variance. This resulted into Crude Prevalence Ratios (CPR) at 95% con dence intervals. Furthermore, variables with a threshold P-value less than 0.05 (P-value<0.05) at bivariate analyses were subjected to the multivariable regression analyses to adjust for confounding, thus establishing Adjusted Prevalence Ratios (APR). At multivariable analysis, only variables with a P-value less than 0.05 were considered signi cant. Both the CPR and APR have been reported.
i. Qualitative data Data collection: An in-depth and key informant guide (appendix 3) was used to conduct interviews with members of the communities in the selected districts who had contracted COVID-19 and the DTF members respectively. The main theme explored was drivers of the COVID-19 transmissions and spread during the second wave. Consent was obtained verbally before the interview of the respondents who were purposively selected for KIIs and in-depth interviews. From each district, 4 respondents (2 male and 2 female) who had contracted COVID-19 were interviewed during in-depth interviews. Both the KIIs and Indepth interviews were recorded using smartphones and tablets and the audios transcribed verbatim into Microsoft Word that were only accessed by the study team. The transcripts were proofread while listening to the original audios by the two-independent researchers for accuracy and consistence.
Analysis: Qualitative data was analyzed using manual thematic analysis, diverging, converging and emerging themes with representative quotes that were obtained during the analysis. The outputs of these ndings are presented in increasing steadily. Permission was also obtained from districts' leadership at sub national level who provided letters of administrative clearances. All study participants consented to participate in the study and the data obtained was secured and kept under lock and key. The eld team individually signed the con dentiality agreement before commencement of the study.

Consent for publication
Not applicable

Availability of data and materials
All the project materials and data about this project are available. These can be accessed by contacting the rst author (Abel Wilson Walekhwa) on wabelwilson@gmail.com.

Competing Interests
All authors declare no nancial and non-nancial competing. All authors con rm that we have had full access to all the data in the study and accept responsibility to submit for publication.

Funding
The funding for this outbreak Investigation was obtained from statehouse, Uganda under the Presidential Scienti c Initiative on Epidemics (Epidemics Unit). The Epidemics Unit is mandated to collect, analyze all epidemiological data to inform national policies in the control and management.

Acknowledgement
We would like to thank all the Respondents for providing the necessary information that made this study a success. Further appreciation goes to the district leadership who actively participated in this study. We   "…you see when home-based care was introduced, the situation got out of hand because after testing positive for COVID-19, people went back to their workplaces such as the salon or shop even though they had mild symptoms like; flu, fever or cough. So, I think that is the biggest driver…." Assistant District Health Officer, District A.
"I was going to be put on oxygen but it was a lot of money, they were asking me 1.5 million per day yet we didn't know how long I was going to be there, whether a week. And so, the health worker said that, "my advice to you is if you can get some money to buy this medicine, you treat yourself from home." So, I went back home. I am here with my wife who also contracted COVID-19 and she got treatment and recovered." IDI Male, District W.
"Since I wasn't that severely sick, I decided to stay home under home based care." IDI 1, District A.
"…the good thing I was asymptomatic so I isolated from home." IDI 2, District A.
Social Gatherings "…We have recorded quite a number of COVID-19 cases from our markets …I'm talking about vendors, not even the customers but the vendors…" DHO, District W.
"You know when the president allowed food markets to operate but was not strict on this issue of attendance in the market. The vendors still moved to their homes and interacted with people in the market then they go and stay with their families…." LC5, District S. "…And then also bars, you find that most of them are still stealthily operational and those are areas that increase spread faster." LC5, District S. "...Last time they said that bars are closed but they are very open, a drunkard can't put on a mask. So, all these things lead to an increase in cases. One person can infect 100 people when they are together...." IDI Female, District M. "...They tell them …not to go for burials, and they don't listen because they want to go and bid farewell to their community member and you just wonder because the person has died of COVID-19. And we know that at the burial of a COVID victim, the chances of having other infected persons are high hence spreading infection in the whole community. There was a burial around here and people slept over saying that it's impossible not to do it. Even if you advise them to let a few stay, they don't listen and you wonder if they are all going to be tested or not and you know at least 10 are infected. Because they cuddled the widow, welcomed her with hugs, and they sat in house...." VHT, District A.

Myths, Misconceptions
and Misinformation "...the community in the district still say there is no COVID…and they do not put on masks…." IDI Male, District B.
"...Another driver is that information and technology that has given freedom to people to publish anything on COVID-19 yet social media tends to be highly consumed by the community…." District Surveillance Focal Person, District T.

Politics
"...you see, the campaigns were the key drivers of the second wave…." District Medical Officer, District P.

Schools
"… The number of cases that came from school are increasing because in terms of positivity the rate was at around 70% in that out of every ten individuals we were testing from the schools, seven were positive." DHO, District W.
Weak Health Systems "...the factor of inadequate resources to confine positive cases became key in spreading the infection...." Resident District Commissioner, District H. "…inadequate resources for sensitization because the rural populace took it as a disease for the urban. "That's your disease." And indeed, if you go to the rural areas, there's totally no SOPs observed. So inadequate sensitization in that regard." District Health Officer, District S.
"…we have seen health workers themselves getting infected. Aaaah. maybe they are not protected, health workers some of them don't have PPEs, they don't have what to use, they don't have gloves, uhmm, and yet they really see patients. For that reason, we have seen health workers who have tested positive. Probably they are also the agents of spreading the disease." District Laboratory Focal Person, District A. "...On the medical perspective, it is lack of machines to use, the PPEs and the rest. That has been one of the factors. You find us working 3 to 4 days but without a mask.…" Laboratory Focal Person, District M.
"…….just before we got the kits yesterday, people we saw had all the signs of COVID, they wanted to test but they could not test, so aanha limited availability of testing points could also have been a driver because some people have signs but for as long as they have not tested positive they will not isolate. They put others at risk and yet they know their status…." Assistant District Health Officer, District A.
"…Then also in our health centers there is a lot of congestion. These are areas that increase spread faster. Most health centers IIs and IIIs offer free health services so whoever believes has a challenge goes there…they tend to handle other cases such as malaria, they don't handle COVID. Regardless a patient can spend the whole day with people at the facility hence spreading the disease…." LC5, District S. Money (Exchange of goods) "...the main cause might be money since exchange of money from one person's hand to another happens when we are buying and/or selling stuff leading to infection. Also, there are some people, including us who keep their masks in the same bag where money is kept, money one has just received from somebody they do not know…." IDI Male, District W.

Stigma
"…the stigma is a big driver among those who know their status and they don't want others to know. So, in that state of hiding, they infect many...." Assistant District Health Officer, District A.
"...when one gets it, he just hides off, so you will find that the whole family is being… infected, that is what is causing the problem in the community...." IDI Male, District B. Figure 1 Location of study districts in Uganda.

Figure 2
Adjusted age distribution of study participants Symptoms experienced during illness

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.