Brazilian municipal health policies during the COVID-19 pandemic
Data files
Dec 13, 2024 version files 2.65 MB
-
bank_measues_complete_2.Rda
2.40 MB
-
codebook_complete_2.csv
146.41 KB
-
participation.csv
1.46 KB
-
README.md
10.49 KB
-
README.txt
10.65 KB
-
table_contention.csv
85.57 KB
Abstract
Brazil was one of the countries most impacted by the COVID-19 pandemic in Latin America and the world considering the number of cases, deaths, and the duration of lockdowns. Between 2020 and 2022, both pharmacological and non-pharmacological interventions (NPIs) were adopted at the municipal level, with 5,568 municipalities and the Federal District taking health-related actions. We present a new dataset revealing the complexity of this situation by reporting data based on thirty-seven surveys taken by mayors between 23 March 2021 and 24 March 2022.
The number of participating municipalities in each survey varied over time. The database indicates in which rounds each municipality participated. The minimum number of participating municipalities was 1,328 (23.8%), while the maximum reached 3,591 (64.49%), showing significant variation. The median was 2,461 (44.19%), and the mean of 2,482 (44.57%) suggests that, in general, municipal participation was close to the median, suggesting the data are representative. Finally, the first quartile was 2,063, and the third quartile was 2,831. The table titled “participation” presents the participation percentages for each of the rounds.
This dataset deals with the need to monitor and share information about fragmented policies designed to tackle health crises like the COVID-19 pandemic. Quantifying these initiatives and how they varied across municipalities can help us to understand the effectiveness of interventions in reducing virus transmission. We offer information over time on a series of measures to encourage social distancing, implement the vaccination programme, provide infrastructure to treat infected people, and facilitate how local governments would eventually ease these measures. This information can contribute to the institutional learning of health systems worldwide, assisting in future situations where a highly contagious virus challenges society.
README: Brazilian municipal health policies during the COVID-19 pandemic
https://doi.org/10.5061/dryad.v6wwpzh5h
Description of the data and file structure
This dataset gathers information on the processes and activities of the pandemic response in Brazil, as well as the epidemiological outcomes of the COVID-19 pandemic in Brazilian municipalities.
Four documents are available: the database (bank_measures_complete_2.Rda*), the codebook (codebook_complete_2.csv), the table with the municipal participation rate (participation.csv), and a table detailing which questions are included in each round of the surveys (table_contention.csv*).
Further details about each document are provided below.
bank_measues_complete_2.Rda
This is the database containing all the questions and all the rounds conducted with the municipalities. The document codebook_complete.csv provides the codebook for all the variables present in the referred database.
codebook_complete_2.csv
It provides a detailed description of the information and the explanations of the respective codes present in the bank_measures_complete.Rda database.
Detailed information in the file codebook_complete.csv:
- group: The "group" variable gathers a set of questions related to the same analytical dimension.
- letter: For each "group", that is, for each dimension that groups a series of questions, a letter from the alphabet is assigned as a code.
- code: This variable indicates the number that the corresponding information represents in the bank_measures_complete.Rda database.
- questions: It presents the description of the question asked, which is recorded in the bank_measures_complete.Rda database.
- variables: This variable indicates the code of the corresponding question in the bank_measures_complete.Rda database.
- values: It provides the explanations for each coding associated with the values assigned to a given response.
participation.csv
This table presents the percentages of municipalities that participated in each round of interviews, considering the total number of municipalities in Brazil.
The minimum number of participating municipalities was 1,328 (23.8%), while the maximum reached 3,591 (64.49%), indicating significant variation in participation. The median was 2,461 (44.19%), and the mean, at 2,482 (44.57%), suggests that, overall, the participation of municipalities was close to the median, reinforcing the representativeness of the data. Finally, the first quartile was 2,063, and the third quartile was 2,831, highlighting the asymmetric distribution of participation among municipalities.
Detailed information in the file participation.csv:
- Interview sequence: In this variable, the order of interviews is presented in ascending order, ranging from a scale of 1 to 37.
- Interview week: The date on which the interviews were conducted.
- Responding municipalities: It presents the number of municipalities that participated in the mentioned round of interviews.
- Total municipalities: Total number of municipalities in Brazil.
- % of respondents: The percentage of municipalities that participated in each round of interviews relative to the total number of municipalities.
table_contention.csv
The document provides the total number of occurrences of each question throughout the 37 rounds and lists the rounds in which this question appeared. Over time, different questions were asked to the mayors that reflect different stages of the pandemic, and the table in question shows the frequency of each one, as well as indicating in which specific rounds they were conducted
Detailed information in the file table_contention.csv:
- group: The "group" variable gathers a set of questions related to the same analytical dimension.
- letter: For each "group", that is, for each dimension that groups a series of questions, a letter from the alphabet is assigned as a code.
- code: This variable indicates the number that the corresponding information represents in the bank_measures_complete.Rda database.
- questions: It presents the description of the question asked, which is recorded in the bank_measures_complete.Rda database.
- variables: This variable indicates the code of the corresponding question in the bank_measures_complete.Rda database.
- how_many_times_it_appears: This variable presents the number of times a specific question was asked across the 37 rounds of interviews. Thus, a question could have been asked at least once or up to 37 times, corresponding to the total number of rounds conducted.
- 23-25/03/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 29-31/03/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 05-08/04/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 12-15/04/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 19-22/04/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 26-30/04/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 03-06/05/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 10-13/05/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 17-20/05/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 24-28/05/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 31/05/2021-02/06/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 02-07/06/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 14-17/06/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 21-24/06/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 28/06/2021 - 02/07/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 05-08/07/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 12-15/07/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 22-25/07/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 26-29/07/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 04-05/08/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 09-12/08/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 16-19/08/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 23-26/08/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 30/08/2021 - 02/09/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 13-16/09/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 20-23/09/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 27-30/09/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 04-07/10/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 18-22/10/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 25-28/10/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 16-19/11/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 06-09/12/2021: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 10-13/01/2022: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 31/01/2022 - 03/02/2022: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 14-17/02/2022: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 07-10/03/2022: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
- 21-24/03/2022: "Yes" indicates that the question is present in this specific round, while "No" means that the question was not asked in this round.
Methods
Information on local NPI policies related to COVID-19 was collected through a telephone survey conducted directly with mayors, who had the option of receiving a password-protected link to respond to the online questionnaire later or to update previous responses. We focused on information concerning three essential dimensions related to the pandemic response: the monitoring of restrictive measures, infrastructure to treat infected people, and the implementation of the vaccination programme. We have included the week that respondents received the questionnaire, the initial date the questionnaire was presented to respondents, and the final date of questionnaire submission.
We collaborated with the Brazilian Confederation of Municipalities (CNM) to collect these data. The cooperation was formalised in a meeting with the CNM on 9 April 2020, and a written agreement was signed by the first and last authors of this article. The authors were given permission to describe, publish, and analyse the dataset. Prior to this current dataset with information from 2021 and 2022, the first and last authors of this dataset had already shared an initial dataset with lockdown measures in Brazil that refer to a survey conducted on October 19 2020, https://doi.org/10.5061/dryad.vdncjsxs2. Similar to our previous dataset that refers to a single survey in the initial days of COVID-19 pandemic, the data we now share on 37 surveys, are freely available to the public and to other academics for analysis.
As Brazil’s largest association of municipalities, the CNM has the email and phone numbers of all elected mayors in the country; the wide reach of the association makes it an ideal partner for large-scale data collection. The partnership was established to study the impact of decentralised measures in Brazil and the effects of decentralisation on the spread of infectious diseases. After the establishment of the cooperation agreement, the CNM added other questions to the questionnaire that were of interest to its municipal monitoring, such as questions related to the possible impact of the pandemic on municipal budgets.
Thirty-seven rounds of questionnaires were conducted, totalling 239 questions. Our database has 15 columns related to municipal identification, 4 on waiting for a bed, 6 on stock and yield of vaccines, 5 on intubation sets, 6 on restrictive measures, 3 on oxygen stocks in hospitals, COVID-19 centres, and other facilities, 4 on the stock of vaccines, 1 on financial resources, 3 on the situation in the UPAs, 3 on the social consequences of the pandemic, 2 on the care of people with health consequences resulting from COVID-19 infection, and 198 on vaccination.
Usage notes
As not all municipal authorities answered all the questions, we suggest that users of this dataset consider additional sources of information to document the implementation of missing policies, preferably using official sources of information such as local decrees. However, as decrees are not always available online, secondary sources, such as media reports, may need to be consulted.
We invite researchers to use these data to deepen their understanding of the pandemic and support health policymakers’ efforts in other health emergencies. Additionally, by combining our database with other government sources, such as those from the Superior Electoral Court, we offer tools to investigate effects of politics on public health policies during the pandemic and thus generate institutional learning for health systems, empirically demonstrating how political decisions influence public health policies. For example, we can analyse how the alignment of a particular mayor with the former president, elected in 2018, their party affiliation, and other such factors affected political decisions relating to pandemic response measures.