Covid-19 vaccination priorities defined on machine learning

Couto, Renato Camargos; Pedrosa, Tania Moreira Grillo; Seara, Luciana Moreira; Couto, Carolina Seara; Couto, Vitor Seara; Giacomin, Karla; Abreu, Ana Claudia Couto de

doi:10.11606/s1518-8787.2022056004045

Acessibilidade / Reportar erro

Brasil

Español English

sumário « anterior atual seguinte »

Sumário

Original Article • Rev. Saúde Pública 56 • 2022 • https://doi.org/10.11606/s1518-8787.2022056004045 copy

Covid-19 vaccination priorities defined on machine learning

Authorship SCIMAGO INSTITUTIONS RANKINGS

ABSTRACT

OBJECTIVE

Defining priority vaccination groups is a critical factor to reduce mortality rates.

METHODS

We sought to identify priority population groups for covid-19 vaccination, based on in-hospital risk of death, by using Extreme Gradient Boosting Machine Learning (ML) algorithm. We performed a retrospective cohort study comprising 49,197 patients (18 years or older), with RT-PCR-confirmed for covid-19, who were hospitalized in any of the 336 Brazilian hospitals considered in this study, from March 19th, 2020, to March 22nd, 2021. Independent variables encompassed age, sex, and chronic health conditions grouped into 179 large categories. Primary outcome was hospital discharge or in-hospital death. Priority population groups for vaccination were formed based on the different levels of in-hospital risk of death due to covid-19, from the ML model developed by taking into consideration the independent variables. All analysis were carried out in Python programming language (version 3.7) and R programming language (version 4.05).

RESULTS

Patients’ mean age was of 60.5 ± 16.8 years (mean ± SD), mean in-hospital mortality rate was 17.9%, and the mean number of comorbidities per patient was 1.97 ± 1.85 (mean ± SD). The predictive model of in-hospital death presented area under the Receiver Operating Characteristic Curve (AUC - ROC) equal to 0.80. The investigated population was grouped into eleven (11) different risk categories, based on the variables chosen by the ML model developed in this study.

CONCLUSIONS

The use of ML for defining population priorities groups for vaccination, based on risk of in-hospital death, can be easily applied by health system managers

COVID-19 vaccines, supply & distribution; Immunization Programs; Health Priorities; Machine Learning

Chronic health conditions	Patients, n/N (%)	Death, n_d/N (%)	Incidence of death, n_d/n (%)^a
Age Group
≥ 90 years	1,685/49,197 (3.4)	856/49,197 (1.7)	856/1,685 (50.8)
≥ 80 < 90 years	5,583/49,197 (11.3)	2,243/49,197 (4.6)	2,243/5,583 (40.2)
≥ 70 < 80 years	8,490/49,197 (17.3)	2,370/9,197 (4.8)	2,370/8,490 (27.9)
≥ 60 < 70 years	10,355/49,197 (21.1)	1,848/49,197 (3.8)	1,848/10,355 (17.8)
≥ 50 < 60 years	9,355/49,197 (19.0)	897/49,197 (1.8)	897/9,355 (9.6)
≥ 40 < 50 years	7,550/49,197 (15.4)	388/49,197 (0.8)	388/7,550 (5.1)
≥ 30 < 40 years	4,792/49,197 (9.7)	179/49,197 (0.4)	179/4,792 (3.7)
≥ 18 < 30 years	1,387/49,197 (2.8)	42/49,197 (0.1)	42/1,387 (3.0)
Sex
Female	21,875/49,197 (55.5)	3,958/49,197 (8.0)	3,958/21,875 (18.1)
Male	27,322/49,197 (44.5)	4,865/49,197 (9.9)	4,865/27,322 (17.8)
Obesity	6,891/49,197 (14.0)	1,410/49,197 (2.9)	1,410/6,891 (20.5)
Chronic renal failure with dialysis	1,224/49,197 (2.5)	906/49,197 (1.8)	906/1,224 (74.0)
Chronic renal failure without dialysis	2,278/49,197 (4.6)	978/49,197 (2.0)	978/2,278 (42.9)
Myocardial and valvular heart diseases, and arrhythmias	3,319/49,197 (6.7)	1,319/49,197 (2.7)	1,319/3,319 (39.7)
Chronic arterial hypertension	23,881/49,197 (48.5)	5,592/49,197 (11.4)	5,592/23,881 (23.4)
Diabetes mellitus	12,549/49,197 (25.5)	3,106/49,197 (6.3)	3,106/12,549 (24.8)
Neoplasms	1,607/49,197 (3.3)	651/49,197 (1.3)	651/1,607 (40.5)
Chronic respiratory diseases	4,195/49,197 (8.5)	1,065/49,197 (2.2)	1,065/4,195 (25.4)
Cerebrovascular disease and its sequelae	848/49,197 (1.7)	363/49,197 (0.7)	363/848 (42.8)
Degenerative diseases of the central nervous system	980/49,197 (1.9)	475/49,197 (0.9)	475/980 (48.5)
Psychiatric disorders	1,161/49,197 (2.4)	189/49,197 (0.4)	189/1,161 (16.3)
Thyroid disease	3,736/49,197 (7.6)	853/49,197 (1.7)	853/3,736 (22.8)
Anemias	603/49,197 (1.2)	250/49,197 (0.5)	250/603 (41.5)
Transplant recipients and patients depending on respiratory support equipment	401/49,197 (0.8)	139/49,197 (0.3)	139/401 (34.7)
Hemorrhagic hematologic disease	387/49,197 (0.8)	166/49,197 (0.3)	166/387 (42.9)
Patients without any of the 15 chronic health conditions listed above	15,812/49,197 (31.2)	1,185/49,197 (2.4)	1,185/15,812 (7.5)

Group	Incidence of death (%)	Sex	Age group	Chronic health conditions	p	Risk-of-death priority level (Min-Max incidence of death, %)	Risk category-population feature
Group I						Priority 1
grp32	58.8	1	7	1			Male and female, > 90 years old, with comorbidities.
grp31	50.2	0	7	1	> 0.05	48.7–58.8	Male, 80 to 90 years old, with comorbidities.
grp28	48.7	1	6	1
Group II						Priority 2
grp27	38.5	0	6	1			Female, 80 to 90 years old, with comorbidities.
grp30	37.2	1	7	0	> 0.05	34.2–38.5	Male and female, > 90 years old, no comorbidities.
grp29	37.2	0	7	0			Male, 70 to 80 years old, with comorbidities.
grp24	34.2	1	5	1
Group III						Priority 3
grp26	28.9	1	6	0			Female, 70 to 80 years old, with comorbidities.
grp23	27.6	0	5	1	> 0.05	25.1–28.9	Female, 80 to 90 years old, no comorbidities.
grp25	26.8	0	6	0			Male, 70 to 90 years old, no comorbidities.
grp22	25.1	1	5	0
Group IV						Priority 4
grp20	22.6	1	4	1	< 0.001	22.6	Male, 60 to 70 years old, with comorbidities.
Group V
grp21	18.5	0	5	0	> 0.05	Priority 5	Female, 70 to 80 years old, no comorbidities.
grp19	18.4	0	4	1		18.4–18.5	Female, 60 to 70 years old, with comorbidities.
Group VI						Priority 6
grp16	14.1	1	3	1	< 0.001	14.1	Male, 50 to 60 years old, with comorbidities.
Group VII						Priority 7
grp18	13.1	1	4	0	< 0.001	13.1	Male, 60 to 70 years old, no comorbidities.
Group VIII
grp15	11.4	0	3	1	> 0.05	Priority 8	Female, 60 to 70 years old, no comorbidities.
grp17	9.9	0	4	0		9.85–11.5	Female, 50 to 60 years old, with comorbidities.
Group IX
grp12	8.3	1	2	1		Priority 9	Male, 50 to 60 years old, no comorbidities.
grp8	7.9	1	1	1		5.5–8.5	Male and female, 18 to 50 years old, with comorbidities.
grp11	7.8	0	2	1
grp7	7.4	0	1	1	> 0.05
grp14	6.4	1	3	0
grp3	5.7	0	0	1
grp4	5.5	1	0	1
Group X
grp13	3.6	0	3	0		Priority 10	Male and female, 40 to 50 years old, no comorbidities.
grp9	2.5	0	2	0	> 0.05	2.2–3.7	Female, 50 to 60 years, no comorbidities.
grp10	2.2	1	2	0
Group XI
grp6	1.4	1	1	0		Priority 11	Male and female, 18 to 40 years old, no comorbidities.
grp2	1.4	1	0	0	> 0.05	0.1–0.2
grp5	1.2	0	1	0
grp1	0.1	0	0	0

Faculdade de Saúde Pública da Universidade de São Paulo Avenida Dr. Arnaldo, 715, 01246-904 São Paulo SP Brazil, Tel./Fax: +55 11 3061-7985 - São Paulo - SP - Brazil
E-mail: revsp@usp.br

Acompanhe os números deste periódico no seu leitor de RSS

[1] Correspondence: Renato Camargos Couto Faculdade de Ciências Médicas de Minas Gerais Alameda Ezequiel Dias, 275 - Centro 30130-110 Belo Horizonte, MG, Brasil E-mail: renatoccouto@gmail.com