COVID-19 vaccination effectiveness in the population of Friuli Venezia Giulia, North-East Italy. Control of bias associated with divergent compliance to policies in a test-negative case-control study.
Alfa variant
COVID-19
Delta variant
Mediation analysis
Omicron variant
Selection Bias
Vaccine effectiveness
Journal
BMC public health
ISSN: 1471-2458
Titre abrégé: BMC Public Health
Pays: England
ID NLM: 100968562
Informations de publication
Date de publication:
11 Dec 2023
11 Dec 2023
Historique:
received:
24
08
2023
accepted:
16
11
2023
medline:
12
12
2023
pubmed:
12
12
2023
entrez:
11
12
2023
Statut:
epublish
Résumé
Vaccine effectiveness (VE) studies consolidate knowledge of real-world effectiveness in different contexts. However, methodological issues may undermine their conclusions: to assess the VE against COVID-19 within the Italian population, a specific threat to validity is related to the consequences of divergent compliance to the Green Pass policy. To address this challenge we conducted a test negative case-control (TNCC) study and multiple sensitivity analysis among residents aged ≥ 12 in Friuli Venezia Giulia Region (FVG), North-east Italy, from February 1, 2021 to March 31, 2022. Information regarding 211,437 cases of COVID-19 infection and 845,748 matched controls was obtained from the regional computerized health database. The investigation considered: COVID-19 infection, hospitalization, and death. Multiple conditional logistic regressions adjusted for covariates were performed and VE was estimated as (1-OR COVID-19 Full-cycle VE against infection decreased from 96% (95% CI: 96, 97) in the Alpha period to 43% (95% CI: 42, 45) in the Omicron period. Booster dose raised the protection in Omicron period to 67% (95% CI: 66, 67). Against the evasive Omicron variant, the protection of the booster dose was 87% (95% CI: 83, 90) for hospitalization and 90% (95% CI: 82, 95) for death. The number of swabs performed was included as a covariate in the adjustments, and the mediation analysis confirmed that it was a strong mediator between vaccination and COVID-19-related outcomes. The study suggests that, under similar TNCC settings, mediation analysis and adjustment for number of diagnostic tests should be included, as an effective approach to the challenge of differential testing behavior that may determine substantial selection bias. This correction allowed us to align with results from other studies that show how full-cycle VE against infection was initially high but decreased over time by variant circulation, counterbalanced by booster dose that raised protection across variants and outcome severity.
Sections du résumé
BACKGROUND
BACKGROUND
Vaccine effectiveness (VE) studies consolidate knowledge of real-world effectiveness in different contexts. However, methodological issues may undermine their conclusions: to assess the VE against COVID-19 within the Italian population, a specific threat to validity is related to the consequences of divergent compliance to the Green Pass policy.
METHODS
METHODS
To address this challenge we conducted a test negative case-control (TNCC) study and multiple sensitivity analysis among residents aged ≥ 12 in Friuli Venezia Giulia Region (FVG), North-east Italy, from February 1, 2021 to March 31, 2022. Information regarding 211,437 cases of COVID-19 infection and 845,748 matched controls was obtained from the regional computerized health database. The investigation considered: COVID-19 infection, hospitalization, and death. Multiple conditional logistic regressions adjusted for covariates were performed and VE was estimated as (1-OR COVID-19
RESULTS
RESULTS
Full-cycle VE against infection decreased from 96% (95% CI: 96, 97) in the Alpha period to 43% (95% CI: 42, 45) in the Omicron period. Booster dose raised the protection in Omicron period to 67% (95% CI: 66, 67). Against the evasive Omicron variant, the protection of the booster dose was 87% (95% CI: 83, 90) for hospitalization and 90% (95% CI: 82, 95) for death. The number of swabs performed was included as a covariate in the adjustments, and the mediation analysis confirmed that it was a strong mediator between vaccination and COVID-19-related outcomes.
CONCLUSIONS
CONCLUSIONS
The study suggests that, under similar TNCC settings, mediation analysis and adjustment for number of diagnostic tests should be included, as an effective approach to the challenge of differential testing behavior that may determine substantial selection bias. This correction allowed us to align with results from other studies that show how full-cycle VE against infection was initially high but decreased over time by variant circulation, counterbalanced by booster dose that raised protection across variants and outcome severity.
Identifiants
pubmed: 38082276
doi: 10.1186/s12889-023-17244-9
pii: 10.1186/s12889-023-17244-9
pmc: PMC10714502
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
2476Informations de copyright
© 2023. The Author(s).
Références
Lancet. 2021 Jun 26;397(10293):2461-2462
pubmed: 34139198
Am J Epidemiol. 2016 Sep 1;184(5):345-53
pubmed: 27587721
Vaccine. 2021 Jul 5;39(30):4013-4024
pubmed: 34119350
Euro Surveill. 2022 Apr;27(16):
pubmed: 35451363
Vaccines (Basel). 2022 Apr 22;10(5):
pubmed: 35632418
Eur Rev Med Pharmacol Sci. 2022 Jan;26(2):643-652
pubmed: 35113440
Vaccines (Basel). 2021 Jun 10;9(6):
pubmed: 34200538
BMJ. 2022 Feb 10;376:e069052
pubmed: 35144968
Vaccines (Basel). 2022 Mar 18;10(3):
pubmed: 35335101
Euro Surveill. 2022 Jan;27(1):
pubmed: 34991780
BMJ. 2016 Feb 25;352:i969
pubmed: 26916049
Emerg Microbes Infect. 2022 Dec;11(1):2383-2392
pubmed: 36069511
Nat Med. 2022 May;28(5):1063-1071
pubmed: 35189624
Am J Epidemiol. 2016 May 1;183(9):869-70
pubmed: 27076669
BMJ Open. 2017 Dec 26;7(12):e019503
pubmed: 29282274
Euro Surveill. 2020 Dec;25(49):
pubmed: 33303064
Int J Epidemiol. 2016 Dec 1;45(6):2060-2074
pubmed: 27453361
Nat Commun. 2020 Nov 12;11(1):5749
pubmed: 33184277
BMC Med Res Methodol. 2021 Oct 25;21(1):226
pubmed: 34689754