Danish Centenarians after 1800

by Axel Skytthe & Bernard Jeune

Many tales have been told about very old people, and the existence of centenarians has rarely been questioned. The number of reported centenarians through time, however, has been rather low until the middle of this century, when the number of centenarians increased dramatically (Kannisto 1994, Vaupel and Jeune, this monograph). Even though the number of reported centenarians in the past has been low, the reported ages at death of the oldest centenarians have been extremely high, e.g. the famous Danish-Norwegian sailor, Christian Drakenberg, who died in 1772 at the age of 146 (Bowerman 1939). The high ages of these very old centenarians have however been questioned by Thoms and others (Thoms 1873, Bowerman 1939, Ørberg 1972).

        The highest probable age at the present level of mortality and the current size of the world's population is about 120 years for a woman (Kannisto and Thatcher 1993). This corresponds to the age of the oldest authenticated person in the world, Madame Jeanne Calment, who celebrated her 120th birthday in February 1995 (Allard et al. 1994). But if the mortality level is higher, and the population is smaller, as in older days, then the question arises: When did the first centenarian emerge? In this chapter we try to answer this question with regard to the emergence of the first Danish centenarian.

Data

Statistics on demographic data in Denmark are based on clerical registration in parish registers.

        The first parish register appears at the beginning of the 17th century. In 1646 the King issued a decree by which the vicar in every parish was ordered to keep records of the date and the number of births, marriages and deaths in his parish. There were, however, no instructions on how the records were to be kept, and the quality of the records for the next 175 years varies. In 1812 a new decree stated that the registers should be kept in special pre-printed books with 7 different forms, including one for dead males and one for dead females. Since that year this has with a few minor changes been the standard for keeping records on births, marriages and deaths, although it took some years to introduce the forms all over the country (Ørberg 1982).

        The collection of demographic data goes back to the 17th century. The collection was, however, sporadic and only carried out in some parts of Denmark. From 1735 reports on births and deaths from the parishes each year were sent to the central administration ("Commerce-kollegiet"). Until 1774 only the total number of deaths were reported, not in age categories. From 1775 to 1834 the number of deaths was given in 10 year age groups with "100 years and above" as the highest category, but the quality of the reported data was not very good. The instructions on how to count the deaths were not clear: Did 'the year' mean the calendar year or the ecclesiastical year. Furthermore, where should a person dying at the age of 80 be included - the category headings were "From 70 to 80 years" and "From 80 to 90 years".

        Graduately the accuracy was improved by more detailed instructions on how to count the deaths. In 1835 the deaths were given in 5 year age groups, and from 1870 deaths were reported by single year of age with the exception that deaths at an age of 100 years or above still were reported as a total.

        Since 1943 all deaths in Denmark are registered at DIKE (Danish Institute of Clinical Epidemiology) in a computerized form, and deaths can therefore be given by exact age after this year.

        The first census in Denmark to report centenarians was in 1801, when 5 female centenarians were found. Censuses on a regular basis were held from 1834 with the next in 1840, and from that year a census was held every 5 or 10 years.

        Before 1870 the census lists were completed by local civil servants in Copenhagen and other towns and by the vicar in the country. Only the head of a family had to go to the vicar and inform him about his family. From 1870 to 1930 each household or family in Copenhagen and towns had to complete a list of all persons in that family or household, while in the country the local parish council was in charge of completing the lists. From 1935 to 1970 the census lists were checked by the municipal public registers, which held information about every person living in a district (Det statistiske Departement 1966). In 1968 the CPR (Central Person Registry) person number system was introduced in Denmark, and the last paper-based census was held in 1970 (Danmarks Statistik 1994).

        The collection and treatment of statistical and demographic data were in the hands of very few persons in the central administration until the middle of the 19th century. A separate office ("Tabelkontoret") was established in 1797, but it was closed in 1819 because of difficulties in making the required tables, internal arguments among staff members and the Danish national bankruptcy in 1814.

        From 1834 to 1849 the collection and treatment of statistical data were administered by a commission with members from various offices of the central administration. The staff involved was rather small: In 1848 the staff included the 5 members of the commission and 6 scribes. The commission had to deal with all statistical requests, and demographic studies were only a part of their work. During the 1840's criticism from scientific statisticians increased, and this lead to the establishment of an independent office for treatment of statistical data in 1849 (Holck 1901).

        The data used in this chapter on number of deaths and population were taken from published tables by Johansen (1975) (for the period 1775 - 1800), and from different published tables by the Statistical Office (Tabelkommisionen, Statistisk Bureau and Danmarks Statistik). Census counts were also taken from published tables by the Statistical Office. A complete list of the publications used can be obtained from the authors.

        The estimates of population of centenarians on January 1st are taken from a database at Odense University on Danish mortality 1870-1992 (Skytthe et al. 1994). In this database populations over 80 are calculated by the extinct cohort method devised by Vincent (1951). From 1943 the number of deaths above 100 are given in exact ages, but before that year the age at death above 100 is estimated using the mortality rates of Sweden for the corresponding period.

Trends in the number of centenarians in Denmark 1775 - 1993

        Table 1 shows the number of reported deaths of centenarians in Denmark since 1775 in groups of 10 years. In this 200-year period the Danish population increased from approximately 1 million to 5 million (Danmarks Statistik 1994). The numbers before 1801 are, however, minimum numbers, because data are missing for some parts of Denmark for certain years. It may therefore be assumed that the number of reported deaths of centenarians was over 100 per decade or in average over 10 per year in the period 1775-1814.

        It is evident that the number of reported deaths of centenarians declines dramatically in first half of the 1800s from over 100 per decade to a minimum of 20-30 deaths per decade in the period 1854-94. This decline is especially evident for males. There are more reported female than male centenarians in all decades with a variation of the sex ratio between 1.4 and 5.6. The sex ratio is especially high in the period with the minimum of reported deaths among centenarians.

        This decline in the first half of the 1800s is also evident when the number of reported deaths of centenarians is calculated per million population - from over 10 to under 2 per million (see Figure 1). Figure 1 also shows population data on centenarians since 1870 when we had the possibility of estimating the number of living centenarians per year based on Vincent's extinct cohort method (Vincent 1951). It shows that the number of estimated living centenarians did not exceed the number of reported deaths of centenarians until 1940.

        The level of centenarians seems almost constant in the period from 1860 to 1930. From around 1965 the number of centenarians increases tremendously to a number of 73 per million in 1993. In this last period the estimated number of living centenarians increases faster than the number of deaths of centenarians, ending with a twofold difference. This rapid development reflects the decrease of the mortality rate of centenarians shown by Kannisto (1994).

        Figure 2 shows three different ways of illustrating the number of centenarians in the period 1800 to 1960. The filled squares represent census reporting of the number of centenarians per million. The number of centenarians reported at censuses shows a dramatic fall in a very narrow period from 1845 to 1860 and continued decline to 1900.

        It is hard to believe that this decline is due to deterioration of living conditions; industrialisation in Denmark only began slowly in the latter half of the 1800s. A plausible explanation is a more careful validation of the reported ages at census after the establishment of Statistisk Bureau in 1849. Understanding the importance of correct data and awareness may also have contributed to a more careful registration of age, when the census lists were filled, and also when a death was reported to the local vicar.

        The third curve, represented by stars, shows estimates of the centenarian population on January 1 based on the extinct cohort method. The difference between the estimated population and the census count could in part be due to the fact that the mortality rates used in calculation of deaths when these were not reported were taken from Swedish data. But it could also be a consequence of over-reporting of centenarians in the censuses due to age-exaggeration.

        Though we only have estimates of the population of centenarians on 1 January, the following 2 figures are based on these, because this is the best available measure of the prevalence of centenarians at a given time.

        On the basis of the estimated number of centenarians from different periods three exponential regression lines were fitted (Figure 3). Assuming an increase in the population size which in itself could explain an increase in centenarians, each regression line has been extrapolated backward in time to where it crosses 1 centenarian. This would give an indication of the year when the first centenarian emerged in Denmark if the conditions had been the same earlier as in the period the regression is based on.

        If the fit is based on the period from 1925-59, the first centenarian would have emerged after 1900 (i.e. 1905). Accepting the population estimates as being correct and assuming that the increase from 1870 to 1924, which as shown earlier is very slow, corresponds with the "natural" increase in the last century, then extrapolation back in time would indicate the "true" year of the first centenarian. This year turns out to be year 1795, in good correspondence with the deliberately provocative allegation - no centenarians before year 1800, at least in Denmark - which was proposed before this analysis was done (Jeune 1994).

        Assuming that the extrapolated curve in Figure 4 represents the true trend in centenarians, most of the reported deaths as centenarians in the hundred year period 1760-1860 must have been assigned too high an age at death.

The Danish Centenarian Registry

We have decided to establish a database of Danish centenarians with a verified age in order to determine the exact proliferation of centenarians in modern time and to determine when the first 100-year-old individual, the first 101-year-old, the first 102-year-old, etc. emerged in Denmark. At this moment we only know that the first verified 110-year-old woman emerged in Denmark on 26 November 1994.

        The identification of centenarians is central in establishing the database. As a starting point we use the number, given in the published statistics of deaths, with an age of 100 years or higher as a key to how many centenarians we expect to find each year. Each alleged centenarian has then to be identified. This is an easy task as far as the last 50 years are concerned. From the Central Person Registry it is possible to identify all living centenarians, and also those who have died since 1 April 1968. Further, the Death Certificate Registry at DIKE, which was started in 1943, includes all dead centenarians since 1943.

        But prior to that time it is more difficult. No easy way to identify centenarians before 1943 exists. From the beginning of the 19th century to 1868 and a single year 1887, the original reports of number of births, marriages and deaths which were sent from the parishes to the Statistical Bureau exist in the National Archives ("Rigsarkivet") in Copenhagen. By carefully going through these it is possible to locate the parishes that reported dead centenarians, and then check the death (or burial) registers from these parishes.

Table 2 : Sources for identification of centenarians

Sources for identification of centenarians

 From 1968  Central person register (CPR)

 From 1943

 Death certificate register (DIKE)
 Approx. 1800 - 1868 and 1887  Original reports from parishes to the Statistical Bureau
 From approx. 1650 - 1890  Indexes on deaths at some regional archives
 1890 - 1942  Difficult ! No yearly reports available, only census lists

        In each of the 4 regional archives, where the parish registers are archived, indexes on marriages and deaths exist for some of the parishes, covering the period from beginning of the parish register, about 1650-1700, to about 1890. By looking through these indexes it is possible to catch almost every alleged centenarian up to 1891. This has been done for approximately 3/4 of the parishes in the county of Funen, where indexes exist. Parish registers for the remaining 1/4 of parishes were searched, resulting in a total of 275 reported centenarians. Assuming that Funen corresponds to 1/8 of the Danish population at any time, this would indicate a total number of 2,200 alleged centenarians before 1890 in Denmark.

        For the years between 1891 and 1942 it is necessary to rely on the kind cooperation of other people. We have contacted genealogists, local historical archives in Denmark and others, who might know anything about very old people.

        Until now we have focused on the period 1848 - 68, mainly because of the existence of the original reports, but also because this is the period where the number of reported centenarians is about the same as indicated by the extrapolated curve.

        Approximately 5,000 alleged centenarians are expected to be included in the database. Of these 2,400 have been recorded as centenarians since 1968, and approximately 500 are from the period 1835 - 1942. Until now (Dec. 1994) we have identified approximately 350 alleged centenarians before 1943, so we still have some work to do.

Table 3 :Four levels of certainty of age

Verification of Age

Level

Criteria

D

 Death date and age (no verification)

C

 Birth registration

B

 Life story

A

 Family reconstructed

        A very important aspect of establishing the database is to verify the age of alleged centenarians. As an indicator of how well documented the age is, we use 4 levels of certainty of the age. The lowest level, D, is simply a reported age of 100 years or more. This is, of course, insufficient. As a minimum the birth (or baptism) must be documented by a registration in a parish register; this is level C.

        It is, however, still necessary to have more information in order to verify the age with certainty. You must be able to reconstruct at least part of the life history of the person with data from other sources such as appearance in census lists, confirmation, military service etc. A verification at this level is classified as level B.

        In order to be sure of avoiding any namesakes, it is in addition necessary to find all brothers and sisters because of a fairly common practice of naming a newborn baby after a predeceased brother or sister. Thoms gave several examples of this in his book (Thoms 1873), and Johansen (1975) has also observed this in his work on the Danish population in the 18th and 19th century.

        If we want to follow Thoms "species of evidence" we must therefore go to the highest level of verification, level A, which requires the total reconstruction of the family of the centenarian, i.e. date of birth and death of both parents, their wedding date, the name and date of birth and deaths of all brothers and sisters. All these data must be found and checked which is very time-consuming.

        The verification of age is a slow process compared to the identification. However, at the identification we attempt to verify the age to at least level C. Verification to level B is often possible because of the possibility of using census lists in order to find the place of birth.

        The intention is to verify all centenarians, who died in the period 1835 to 1943, to at least level B. If verification to this level is satisfactory, we will consider the person in question as a true centenarian, knowing the possibility that we might include some false centenarians due to namesaking. Centenarians, who died after 1943, will in general be considered as true centenarians, when they are verified to level C. We consider the information in the Death Register at DIKE as being valid because of the long-standing tradition for registration of demographic events in Denmark. However if we find any difference in the number of centenarians reported by the official statistics and the number reported from DIKE, every individual dying that year will be checked and verified to at least level B.

        Verifying the age of an alleged centenarian before 1840 is increasingly difficult, partly because of the quality of information in parish registers, partly because of an increasing probability that parish registers do not exist, when you go back to the eighteenth and seventeenth century, mainly lost due to fires at the vicar's residence. Finally it can be very difficult to locate the place of birth, when there is no supplementary information from, for instance, censuses.

        The top level verification is very expensive to do and will not be done on all alleged centenarians, but only on those at extreme ages (the 5 or 10 oldest centenarians each year as well as all above 105 in the whole period and centenarians above 103 before 1940).

        Finally an example of the verification process: In a little parish on Fyn, Rorslev near Middelfart, a man by the name Mads Pedersen Ribe died 24 April 1864 at the age of 109 years. Obviously the vicar also thought that this was an extreme age, because in the burial register he added "Born in Ribe 24 June 1755". Looking in the parish register for Ribe Cathedral we do find an entry saying " .. 24 June 1755 Mads, son of Peder Madsen, soldier". But that is all. There is simply no information on his mother.

        Furthermore - it is not possible to find Mads Pedersen Ribe in any of the census lists from the parish in the period 1834-60. This means that either he has arrived at the parish after 1860 - at the alleged age of 105 - or he has lived as a recluse without contact to others for many years. In the probate register it is stated "Pauper Mads Pedersen Ribe has died at the age of 109 - no possessions, no information about heirs". Thus the status of Mads Pedersen Ribe as a centenarian can only be verified to level C. This is not sufficient to consider Mads Pedersen Ribe a true centenarian.

Literature


Updated by V. Castanova,   March 2000