We could learn much more from 1918 pandemic – the (mis)fortune of research relying on original death certificates

Abstract

The analysis of historical death certificates has enormous potential for understanding how the health of populations was shaped by diseases and epidemics and by the implementation of specific interventions. In Brazil, the systematic archiving of mortality records was initiated only in 1944 – hence the analysis of death registers prior to this time requires searching for these documents in public archives, notaries, parishes and especially ancient cemeteries, which are often the only remaining source of information about these deaths. This article describes an effort to locate original death certificates in Brazil and document their organization, accessibility and preservation. To this end, we conducted an exploratory study in 19 of the 27 Brazilian states, focusing on the period surrounding the 1918 influenza pandemic (1913 to 1921). We included 55 cemeteries, 22 civil archives, and 1 military archive. Apart from few exceptions, the results show the absence of a curatorial policy for the organization, access or even physical preservation of this material, frequently leading to unavailability, deterioration and ultimately its complete loss. This study indicates the need to promote the preservation of a historical heritage that is key to understanding historical epidemiological patterns and human responses to global health threats.

Introduction

The use of death records to study population mortality patterns can be traced back to the work of John Graunt, known as the father of demography. Graunt analyzed 50 years of vital statistics data from London extracted from the Bills of Mortality, and in 1662 published his findings in a book entitled “Natural and Political Observations Made Upon the Bills of Mortality”, with profound impact in the field of demography [1] [2]. Although holding great potential for the understanding of disease patterns and the ways whereby specific interventions have affected disease burden and transmission, the study of historical vital statistics data has been, however, largely neglected within the public health sciences. Largely to blame is the difficulty to locate, access and preserve what are most often extremely fragile paper collections, kept in less than ideal conditions. This is particularly true in developing nations, where existing historical data is seldom preserved or recording keeping is relatively recent [3] [4].

In Brazil, the first compilation of mortality records took place in 1944, and was aimed at the publication of annual cause-specific mortality statistics for all Brazilian state capitals in the period from 1929 to 1939 [5]. Although the Brazilian civil death registry was established in 1888 [6], it was only in 1939 that the government engaged in the administration of public records (Decree 4851 of 1939). Mortality data from 1979 onward is now available in electronic format through DATASUS, the National Health System database of the Brazilian Ministry of Health, and indexed following the International Code of Diseases [7]. The fact that these publicly available datasets can now be accessed online for consultation is the result of an exemplary and pioneering initiative in Latin America that has allowed local and foreign researchers to conduct quantitative studies with potential public health policy implications.

The analysis of death records prior to 1979, however, still depends on the active search for, and access to, primary data sources. The availability of mortality records prior to 1939 is even more precarious. Death certificates from this period are usually located in public archives, government offices, parishes or even ancient cemeteries. Unfortunately, as we will characterize in this study, the search for such documentation is often hindered by administrative, institutional and managerial obstacles, frequently preventing the completion of investigations that rely on these data.

Here we explore and document the availability and preservation of historical death certificates from the beginning of the 20^th century, in the period coinciding with the 1918–1920 influenza pandemic, the so-called ‘Spanish flu’ (responsible for mortality figures higher than those experienced in battle during the World War I). Although the mortality impact of this pandemic is well documented in the ‘Old World’ [8] [9] [10] [11], few studies have addressed it in tropical or subtropical regions. One exception was an investigation conducted in a subtropical Brazilian city (Florianopolis, Santa Catarina) [12] [13]. The analysis of death certificates available from the municipal archive of this city made it possible to reveal novel patterns of the 1918 pandemic at this latitude, as well as the discovery of forgotten aspects of local and regional history. The experience and findings obtained with the latter study motivated its expansion throughout Brazil, to compare the epidemiological features of the pandemic across different regions of the country. This article describes the efforts to collect historical death certificates in several Brazilian cemeteries and archives, as a means of documenting the availability, organization, and preservation of these documents, as well as to raise awareness on the lack of national policies for the preservation of these valuable data sources. Possible solutions and suggestions for improving this situation are discussed.

Material and methods

To determine the availability of vital statistics data from 1917 to 1921, we initially conducted an exploratory survey from March through May 2011 with intentional sampling in the capitals of 19 Brazilian states located in the North, Northeast, Midwest, and Southeast regions of Brazil (Acre, Alagoas, Amapá, Amazonas, Bahia, Espírito Santo, Maranhão, Mato Grosso, Mato Grosso do Sul, Minas Gerais, Para, Paraíba, Pernambuco, Piauí, Rio Grande do Norte, Rondônia, Roraima, Sergipe and Tocantins). Our aim was to obtain information on the existence of (and accessibility to) death records in 22 public archives and 55 cemeteries, which (given their antiquity) could possess death records for this period. All institutions were initially contacted by telephone during office hours. Repeated attempts (up to 15) were made at various office hours in different days. When no contact by telephone was possible, we tried to reach the institutions by email.

One of us (FN) personally traveled to two cities (see results), where confirmation of the existence of the death records for the period studied had been obtained. Another one of us (WJA) traveled to Rio de Janeiro (the capital of Brazil in the beginning of the 20^th century) to consult the Archives of the Brazilian Navy for records of the 1918 casualties among the Brazilian troops that participated in WWI.

We tabulated and quantified the success rate of locating and accessing the mortality records in those localities at each stage of the research process.

Results

Several difficulties were encountered to obtain contact information for the 55 cemeteries investigated. Contact information available in directories, lists and websites was often out of date, wrong or incomplete.

The outcomes of the contact attempts made with each of the institutions are summarized in Table 1. Success rate, as measured by the proportion of institutions with death certificates available for consultation, was only 11%. The rate of success varied substantially across institution types, ranging from 100% for charitable religious institutions to 2% for cemeteries. A detailed account of the process to access this material follows in the remainder of this section.

Table 1.

Success at locating historical death certificates for the period from 1917 to 1922 in several Brazilian depositary institutions

(Among those for which contact was successful)

Type of institution	Investigated	Contact successful	Staff did not know about the death certificates	Institution did not have death certificates	Certificates from 1917 to 1922 not found	Certificates available for consultation	Success rate (% institutions with files available for consultation)
Cemetery	56	33	6	9	22	1	2%
Public archive	30	25	6	13	1	7	23%
Religious charity	2	2				2	100%
Library	2	1	1	1			0%
Navy	1	1				1	100%
TOTAL	90	61	13	23	23	10	11%
Percentage		68%	21%	38%	38%	16%

Our systematic attempt to make telephone contact with the administrators of the 55 cemeteries was successful in 58% of the cases (33 cemeteries). For the remainder cemeteries for which telephone contact was not possible, only 6 (21%) replied to our email contact.

Of the 90 employees contacted in the cemeteries, 32 (35%) were able to provide some information about the mortality records from the period investigated. Two employees informed that the record books were in advanced stage of deterioration, and one employee mentioned that the records would not be available for consultation as they had been sent to restoration. Eight employees provided the contact of a supervisor and four granted us permission to consult the files containing the death certificates. One employee informed us that the vital statistics data in their possession was in an advanced stage of deterioration and with missing years (they could not locate the records for the years investigated). Another supervisor informed us that the record books had been sent to a restoration company, but no information was available on the years covered by those records. Two staff members instructed us to provide a formal request to the cemetery director to authorize consultation of the data sources – which we did, but without a positive response.

Information about the mortality records was obtained from 25 public archives, of which: 12 informed that they did not have death records in their collections but indicated other possible locations in the region that could have them (e.g., cemeteries, buttresses, city hall, public archives), 11 had burials starting later than our period of interest, and 4 had no death records in their collections (only documents such as old newspapers and government documents). Additionally, three clerks agreed to check whether the collections included the period of interest, one clerk reported that death records remained with the funeral company responsible for the burials and one clerk reported that the records remained in the church where the deathwatch would have taken place.

Overall, certificates for the period of interest were available in only four state capitals: Belém, Salvador, Belo Horizonte and Vitória. In these capitals, burial books and death certificates from 1913 to 1921 were deposited in public archives, old cemeteries, churches and in some cases in notary offices.

Our only consultation with the Navy (Rio de Janeiro) was successful, with some documents scanned and sent to us. We were informed that the Archives of the Brazilian Navy were available for consultation, and assistance would be provided upon arrival at their facilities.

Consultation of documents in situ

The historic certificates were perfectly organized and preserved by the Municipal Archive of the city of Belo Horizonte (APCBH). In this institution, the employees offered full support for the consultation of the mortality records. In contrast, our experience in Belém do Pará is illustrative of the range of problems encountered even in the rare event of successfully locating these historical documents. We were informed by the Santa Isabel cemetery staff that their entire old registry had been lost, and only records for the period 1963–1990 were preserved. These records were also not organized, complicating further their inspection. Moreover, two employees of the cemetery staff also indicated that the records were kept currently were in a dangerous area of the city, in a facility that had no security, which would complicate our visit to access the records. The records corresponding to deaths from another region of the city did exist and were kept at the 1st Notary’s Office. Unfortunately, when we visited this office, we were denied access to the documents, even after including in our request a signed support letter from a research collaborator from Evandro Chagas Institute (an internationally renowned scientific institution in Brazil in charge of studying infectious diseases, with its headquarters nearby). The trip to Belém was not completely disappointing as we also visited the Metropolitan Curia (Archdiocese of Belém), which informed us that they had the mortality records of the period of interest, and would grant us access to conduct our research (although we could not access them because the building where the certificates were deposited was under renovations at the time of the trip).

The trip to the Archives of the Brazilian Navy in Rio de Janeiro was very positive. Documents were preserved in perfect conditions, the staff was very helpful and access to all documents needed was granted without problems.

Discussion

Vital statistics records are extremely valuable data sources to help reconstruct the history of communities, cities, states, and nations. They are often the only testimony to the existence of individuals, with information that allows us to infer key aspects of social, cultural and demographic processes and past epidemiological patterns [14] [3] [4]. Hospital archives, for instance, have been deemed as a key source of knowledge about “family constitution, migratory movements, feeding practices, conflicts between the various arts of healing, among other topics”, and even as legal documents [15].

Similarly, the analysis of historical mortality time series is critical for medical and epidemiological research, as it enables understanding how population trajectories have been molded by diseases, epidemics, catastrophes, conflicts as well as by human interventions (such as the introduction of vaccines). Through examination of death certificates it is possible to infer, for example, how infectious diseases affected infant mortality rates in the past, how epidemics and sores have spread from one place to another, and how vulnerability to certain diseases has evolved over the years [14] [3] [4].

Although death certificates are of great historical importance, apart from notable exceptions, we witnessed severe shortcomings in curatorial and preservation efforts for the period around the 1920s in Brazil. Locating these data sources is difficult as they are dispersed, with no apparent pattern, among various institutions. In most cases, it was not even possible to identify the whereabouts of such records. We are not aware of the existence of a national survey or catalog of old record books, nor is their information – to the best of our knowledge - transcribed or copied for preservation. This situation casts serious concerns on the ongoing irreversible loss of the valuable information available in these documents.

In the rare cases where data sources were located, their state of conservation and accessibility varied greatly depending on the city studied, the year, and the institution responsible for its preservation. In the most positive situations, access to the records followed a well-established protocol, which included the exhibition of the catalog by the staff member in charge of their curation, guidance on the appropriate handling of the documents, permission to carry out the digitization of information through photographs, and the request of the appropriate credits to the institution in resulting publications. The positive experiences in the Municipal files of Florianopolis or the Archives of the Brazilian Navy in Rio de Janeiro are very illustrative of the latter description. As a result, in both cases research findings of international interest were revealed [12] [16]. On the other side of the spectrum, several attempts at locating the historical data have either revealed the abandonment and clear deterioration of these documents, or the creation of bureaucratic obstacles for consultation (as in the case of the notary’s office mentioned above, thus evidencing the treatment of public domain material [17] as private property).

Old death records have a value that goes beyond their historical interest; they allow, for example, quantitative impact analyses that may shed new light on the mechanisms underlying the evolution and transmission of diseases that still afflict our populations today. Problems of similar nature were also detected with data from hospital certificates [15]. For such documents to be investigated, however, systematization of access to such information is necessary.

Brazilian legislation recognizes the importance of this patrimony, and its responsibility in ensuring its preservation: “It is the duty of the Public Power to document management and special protection of archival documents as an instrument to support administration, culture, scientific development and [17] is “the body linked to the National Archive, which will define the national policy of archives.” Here we take advantage of the occasion to suggest to the National Council on Archives (CONARQ) the creation of a national catalog to prevent the rapid and irreversible deterioration of these unique sources of information and implement further measures that go beyond promoting the preservation of the original documents. We suggest implementing a project for the comprehensive digitization of such historic material to ensure the safeguarding and accessibility of this information. Given the magnitude of the task that would entail locating and digitizing the certificates, it will probably be necessary for various social actors to be invited to contribute - not only staff members from cemeteries, churches, hospitals and public archives, but also researchers and historians. We again can cite the example of our previous study in Florianópolis [12], in which the digitalized version of all records consulted (achieved with the aid of an commercial digital camera) was delivered to the municipal archive, and subsequently incorporated as a new data source at zero cost for this public institution.

Preserving, organizing and creating the means to ensure access to these documents can serve a variety of functions, not only in the fields of epidemiology and medicine, but also as a valuable source for understanding our historical, social, and demographic processes. In many cases, these records also represent our only source of information from past generations and the testimony of the existence of individuals who are no longer among us [4].