Studies on Morbidities and Mortalities from COVID-19: Novel Public Health Practice During Pandemic Periods

MATERIALS AND METHODS

The current study focused on the investigative practice and outbreak analysis of Severe Acute Respiratory Syndrome Coronavirus 2 by the Ministry of Health professional authorities as part of a series of evaluations for this pandemic in various regions of the world.

Study area: The COVID-19-impacted population data were obtained from the South Sudan database for the daily cases and deaths. The study covered a total number of days of 1107 from the start date 03 January, 2020 till 13 January, 2023 of the observation and monitoring period. This is equivalent to 3 years and 11 days or 36 months and 11 days of continuous daily record.

Public database source: The Human Coronavirus 2019 records were obtained and processed from Humanitarian Data Exchange⁹. The dedicated country's dataset was processed using the Excel functions (Microsoft Office Professional Plus 2010 Version 14.0.726.5000).

Analysis type and metrics: This work used the newly detailed daily cases and fatalities-in chronological sequence-as significant public health measurements¹⁰. The South Sudan dataset was extracted for the filtered Excel database¹¹. The survey would include preliminary 3-D visualization coupled with Pareto analysis, time-series plots employing trending charts and model-fit screening to assess the dynamicity of cumulative morbidities and mortalities, as well as the chronological link between both indices¹².

Software and application: The research was carried out on a commercially available, user-friendly and cost-effective software platform^10-12. Microsoft office excel-generated database was processed chronologically to yield a daily transformed census of cumulative cases and deaths to be further analyzed using Minitab version 17.1.0 and CurveExpert version 1.40 to produce Contour, 3-D scatter plot associated with Pareto charting and process-behavior charts by the first program and the rational best curve-fitting by the second one.

Statistical analysis: The raw data obtained, including morbidity and mortality populations and their counts, was organized by time, date and season. The proportion of each group was calculated and the total cumulative percent was computed. This would yield a Pareto analysis. Drawing of the trending charts was conducted according to the software manual using Laney modification to correct for data dispersion. Modeling was done using the best curve fitting approach with prioritization selection based on best regression, least error and better residuals, especially after logarithmic conversion to the base ten. One was added to the affected population census before transformation to correct for the presence of zeros without severely distorting the originally recorded report.

RESULTS

The outbreak showed the 3D pattern in Fig. 1. The appearance demonstrated higher density in the middle of the investigation period, with a relatively lower census of morbidities and mortalities at the beginning of the investigation period and expressing the lowest magnitude at the end. A detailed examination was performed using the Pareto principle of 80% major contribution could be visualized in Fig. 2 and 3. From these figures, simplified Table 1 that has been created to demonstrate a focus group that details the most important times of the epidemic for morbidity and mortality counts in terms of seasons, years and months.

Exploratory process-behavior charts were used to track and show the pattern of the disease progression (Fig. 4). Time-series of the census illustrated several successive waves of the outbreak that appeared as peaks and clusters made of groups from adjacent sporadic aberrant counts. The death chart is correlated with the reported cases graph except for the last wave in morbidity which was not accompanied by the corresponding peak in the mortality graph.

Fig. 1(a-b):

3D visualization of the daily reported cases and fatalities from COVID-19 in South Sudan using (a) Contour plot and (b) 3D scatter plot (generated using Minitab version 17.1.0)

Fig. 2(a-h):

Pareto diagrams showing the main contributing periods in cases and deaths of COVID-19 in terms of seasons, years and months, (a) Cases by seasons, (b) Deaths by seasons, (c) Cases by years, (d) Deaths by years, (e) Cases of years by seasons, (f) Deaths of years by seasons, (g) Cases of months by seasons and (h) Deaths of months by seasons (generated using Minitab version 17.1.0)

The kinetics of the cumulative daily cases and deaths-in the transformed form-showed a significant pattern of the Morgan-mercer-flodin (MMF) model after screening for the best non-linear regression of the best curve fitting. Also, the MMF model showed an acceptable fit when used to describe the relationship between the two parameters of the epidemic. Figure 5 illustrated this curve fitting behavior and Table 2 showed the parameters of MMF equation:

where, a, b, c and d are coefficient data, x is the time in days and x is the census of cases or deaths.

Fig. 3(a-d):

Overall Pareto plots showing main contributing months and months/years in COVID-19 metrics, (a) Cases by months, (b) Deaths by months, (c) Cases by month per year and (d) Deaths by month per year (generated using Minitab version 17.1.0)

Fig. 4(a-b):

Exploratory attribute trending chart showing the pattern and excursions in the daily census of the reported illness and fatalities in South Sudan, (a) Process behavior chart of new cases and (b) Process behavior chart of new deaths (generated using Minitab version 17.1.0)

DISCUSSION

The present study of South Sudan is characterized by a diminishing number of reported death numbers that were attributed to COVID-19. Like other countries, the widely disseminated viral outbreak was accompanied by very low fatality rates¹³. The average rate of mortality from the mean affected population by COVID-19 is very low (0.75%) and the upper bound value is 1.27%. The worst year in the epidemic was 2021 followed by 2020 and 2022 with the worst season in morbidity being the winter while the worst season for mortality was the summer. An exception was found in the summer of 2020 which witnessed the worst level of the seasonal outbreak.

Logarithmic transformation of the datasets was found to be useful in decreasing data spreading and hence, the noise and error factors. Thus, the error factor in the record would be reduced significantly¹⁴. This was found to be useful in the result interpretation and modeling. Time-series plots could reveal the start time of the outbreak and the presence of low, medium and high-magnitude waves of the mass illness which could be found reasonably correlated for both outbreak indices (i.e., cases and deaths).

Fig. 5(a-c):

Modeling of COVID-19 logarithmically transformed, (a) Cumulative cases (CC) and (b) Cumulative deaths (CD) and (c) Kinetics with elapsed time and against each other(CurveExpert version 1.40)

Best curve fitting showed that the MMF equation was the preferred fit for the transformed disease indices with high regression and low standard error. The MMF model description of the morbidity vs mortality relationship showed that at the initial stage, death is lagging the detected illness¹¹. The second stage demonstrates a fast rise in both parameters after inflection point^15,16. The last stage represents the flattening phase where deaths begin to cease to increase with new cases^17,18. A similar pattern could be observed when observing the dynamic nature of both morbidity and mortality rates with time as a balanced saturation level could be found after reaching the community capacity based on its pre-existing conditions and actions put in place.

The novel perspective for examination and analysis of the pandemics in the present study is helpful for monitoring and tracking the progression and pattern of the outbreak by the workers in the health governmental authorities. Expanding the perception and knowledge about epidemic containment and resource management for the employees in the ministries of health is inevitable through the support of these complementary methodologies. A future enhancement-through further investigation-would be required to overcome the challenge of the presence of autocorrelation during the modeling of the cumulative counts of the morbidities and mortalities over time.

CONCLUSION

The present work provided novel and useful epidemiological analysis and descriptive study that could be implemented by public health professionals and the Ministry of Health authorities to study, analyze, control and predict the behavior of the outbreaks. This study helps to understand and set measures to contain the microbial disease by tracking and predicting the behavior and the pattern of the outbreak spreading. Further examination would be needed to overcome the challenges that might be confronted due to issues raised due to the potential autocorrelation of the residuals in the proposed model.

SIGNIFICANCE STATEMENT

Despite the limitation that could be encountered by modeling the time-series data, the present investigation provided a useful perspective analysis for health professionals and public health authorities for the study of the pandemics and quantitative evaluation of the disease magnitude and measurable metrics for the health risk to the affected herd within the community.