Development of a tool to prioritize the monitoring of COVID-19 patients by public health
teams
Andres I. Vecino-Ortiz, MD MSc PhD*.
Department of International Health, Johns Hopkins
Bloomberg School of Public Health, Baltimore, MD USA.
Nicolás Guzman-Tordecilla ND MSc.
Institute of Public Health, Pontificia Universidad
Javeriana, Bogotá, Colombia.
Yenny Fernanda Guzmán Ruiz MD
. Department of Health Services. University of Washington.
Seattle, WA USA.
Rolando Enrique Peñaloza-Quintero, DDS PhD
. Institute of Public Health, Pontificia
Universidad Javeriana, Bogotá, Colombia.
Julián A. Fernández-Niño
Ministry of Health
and Social Protection
of Colombia and
Departmen t of Public Health, Univ ersidad del Norte. Bogota, Colombia .
Fernando Ruiz Gomez
. Ministry of Health and Social Protection of Colombia. Bogota,
Colombia.
Antonio J. Trujillo
. Department of International Health, Johns Hopkins Bloomberg School of
Public Health, Baltimore, MD USA.
*Corresponding author
Department of International Health.
Johns Hopkins Bloomberg School of Public Health
615 N. Wolf Street.
Suite E8620
Baltimore, MD. USA
avecino1@jhu.edu
. CC-BY-ND 4.0 International licenseIt 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 (which was not certified by peer review)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 April 11, 2021. ; https://doi.org/10.1101/2021.04.08.21254922doi: medRxiv preprint
NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.
Abstract
Background:
In the context of the COVID-19 pandemic, public health teams have struggled to
conduct monitoring for confirmed or suspicious COVID-19 patients. However, monitoring these
patients is critical to improving the chances of survival, and therefore, a prioritization strategy
for these patien t s is warranted. This study developed a monitoring algorit hm for COV ID -19
patients for the Colombian Ministry of Health and Social Protection (MOH).
Methods:
This work included 1) a literature review, 2) consultations with MOH and National
Institute of Health officials, and 3) data analysis of all positive COVID-19 cases and their
outcomes. We used clinical and socioeconomic variables to develop a set of risk categories to
identify sev ere case s of CO VID-19.
Results:
This tool provided four different risk categories for COVID-19 patients. As soon as the
time of diagnosis, this tool can identify 91% of all severe and fatal COVID-19 cases within the
first two risk categories.
Conclusion:
This tool is a low-cost strategy to prioritize patients at higher risk of experiencing
severe CO VI D-19 . This tool was develop ed so public heal th team s can focus thei r scarce
monitoring resources on individuals at higher mortality risk. This tool can be easily adapted to
the context of other lower and middle-income countries. Policymakers would benefit from this
low-cost strategy to reduce COVID-19 mortality, particularly during outbreaks.
Keywords:
Monitoring, risk, prioritization, COVID-19
. CC-BY-ND 4.0 International licenseIt 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 April 11, 2021. ; https://doi.org/10.1101/2021.04.08.21254922doi: medRxiv preprint
INTRODUCTION
The novel coronavirus SARS- Co V-2 (COVID-19 ) spre ad across the world and plac e d many
health systems under unprecedented strain. During the first months of the pandemic, a rush to
understand early treatment options for COVID-19 was made. Towards the midst of 2020, there
was a relatively good understanding of early strategies to reduce the disease's lethality to
complement and improve the eff icacy of more traditional public health interventions as social
distance, lockdowns, hand washing, and contact tracing
1,2
. However, overwhelmed health
systems struggled to provide appropriate monitoring to all identified cases, particularly during
outbreaks. Moreover, monitoring is challenging because it requires high upfront investments
in infrastructure, planning, trained personnel, and informat ion systems. Despite the high
positive exte rnaliti es, monitoring COVID - 19 confron ts organizational barrier s that slow down
its implementation.
The severity of COVID-19 infection is heterogeneous, and the fatality rate in the general
population is estimated around 1%
3
. Some individuals are asymptomatic or experience mild
flu -like s ymptoms, but oth ers suffer a severe illn ess that can lead to death. The literatu re
shows a higher probability of seve re disease and death s among older indi viduals with
comorbidities such as diabetes, cardiovascular diseases, obesity, and respiratory diseases
4
.
COVID-19 cases with these conditions benefit from early and const ant monitoring to seek
advanced heal th care as soon as possible . However, given the limited re sources available to
monitor individuals, priori tizing severe disea se risk categ o rie s is an urgent need. Suc h need i s
particularly true during out breaks when the pressure on health systems is higher. Especially
during high-demand times, it is necessary to ident ify early those with the highest probability of
complications to receive faster life-saving care.
In this study, we developed and tested a s trategy to prioritize monitoring systems for COVID-
19 patients by directing monitoring resources to those in need. The study describes the
development of this tool. This tool can be of use by other lower and middle-income countries
(LMIC). This tool will also benefit other countries by reducing negative externalities towards
other health care services.
. CC-BY-ND 4.0 International licenseIt 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 April 11, 2021. ; https://doi.org/10.1101/2021.04.08.21254922doi: medRxiv preprint
METHODS
In this work, we first developed a liter ature r eview to identify relevant variables for the
prioritization tool. Based on the predictors identified during the review, we conducted the
preliminary design of the prioritization tool in consultation with officials from the Colombian
Ministry of Health and Social Protection and the National Institute of Health of Colombia (MH-
NIH). Finally, we conduct ed the data analy si s to asse ss the validity of the prioritiza tion
algorithm.
Literature review
The literature review aimed to iden tify effect sizes for variables that predict a highe r likelihood
of severe or fatal COVID-19. We conducted a MEDLINE and Google Scholar review of papers in
English and Spanish, in both preprint and peer-reviewed journals, and published before
October 1, 2020. The following search terms were used in English:
Novel Coronavirus,
Epidemic, Pandemics, Sars-Cov-2, COVID-19, Coronavirus Infections, Complications, Mortality,
Intensive Care Unit, Disease Severity, Mortality rate, Fatality rate, Deaths, Severity, Obesity,
Body Mass Index, Total Body Weight, Smoking, Comor bidity, Cerebrovascular, M eta- Analysis .
In
Spanish, the terms were: Nuevo coronavirus, Epidemia, Pandemia,
Sars-Cov-2, COVID-19,
Infección por
coronav irus, Complica cio nes, Mortalida d, Unidad de Cuida do intensivo, UCI,
severida d de la enfer med ad, Tasa de m ortalid ad, Tasa de fatalidad, Muerte , Se veri dad,
Obesidad, Índice de masa corporal, Exceso de peso, Sobrepeso, Fumar, Comorbilidad,
Enfermeda d Cerebr ovascular, Met a- aná lisis.
Moreover,
we identified the papers cited in the reference list to widen the search.
The inclusion criteria for the selection of papers were:
1.
Paper s describing severe or fatal COVID -19 (defined as a patien t requiring
admission to an inten sive care unit, intu batio n, or dying from COVID-19 ).
2.
Papers including risk and predictor factors that have been measured
empirically or have been described as high-risk factor for severe or fatal COVID-19.
. CC-BY-ND 4.0 International licenseIt 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 April 11, 2021. ; https://doi.org/10.1101/2021.04.08.21254922doi: medRxiv preprint
The exclusion criteria for the selection of papers were:
1.
Papers were abstracts, conference proceedings, or book chapters
2.
Paper s that did not di rectly add re ss risk factor s f or sever e or fatal COV ID -19.
3.
We did not exclude papers based on geography, gender, or age.
Our initial search identified 8,319 papers. Excluding 1,081 duplicated manuscripts, our final
search identified 7,238 papers. Two reviewers screened titles and abstracts yielding 194
papers for full-text review, and selected 12 papers for inclusion. The search strategy can be
seen in figure 1. In these papers, we obtained a list of all potential variables with more
predictive value .
<INSERT FIGURE 1>
Data sources
We used officia l combined da ta for all 1) confirmed, 2) suspected, and 3) negativ e Coronav irus
cases in Colombia in the general population (excludes specific closed populations including
military, police per sonn el, and institut ion alized individuals, which rep rese nt a smal l percenta ge
of the population) through the Ministry of Health (MOH) and National Health Institute (NIH)
platforms (SEGCOVID19 and SIVIGILA, respectively), between March 6 and October 11, 2020.
This dataset ascertains data on all individuals' health, demographic, and socioeconomic
characteristics with a confirmed, possible or negative diagnosis of COVID-19
5
.
SEGCOVID19 is the main platform where all the data from the national contact tracing
program (called PRASS, the Spanish acronym for Sustainable Program for Tests, Tracing, and
Selective Iso lation) is collected and consolidated. This syste m collects information from both
notifications issued by the National Health Institute system (SIVIGILA) and cases and contacts
identified by insurers. The system is critical to trace transmission chains and identify suspected
contacts. SEGCOVID19 also collects information from all tests performed in the country and
gathers data from databases recording health care service provision, civil registration, and vital
statistics records. We used both records to ensure that we captured as many variables as
possible from every single case so that the prediction model would be more comprehensive.
. CC-BY-ND 4.0 International licenseIt 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 April 11, 2021. ; https://doi.org/10.1101/2021.04.08.21254922doi: medRxiv preprint
