Temporal Trends in Mortality in Portugal, 1913–2022

Abstract

Introduction

Demographic and epidemiological transitions have transformed mortality patterns, with declines observed up to the present. Long-term time series studies on mortality evolution in Portugal are scarce. We aimed to estimate and describe annual crude and age-standardized mortality rates (ASMRs) in Portugal from 1913 to 2022 and to analyse temporal trends from 1940.

Methods

A time series study was conducted using official statistics on deaths and resident population, between 1913 and 2022. Annual crude mortality rates (CMRs), overall and stratified by sex and age group, were calculated for the entire period. ASMRs were estimated for 1920, 1930, and the period from 1940 onwards. Trends in CMR and ASMR, overall and by sex, from 1940 onwards were analysed using joinpoint regression.

Results

Between 1913 and 2022, the annual CMR decreased by 42.1%. From 1940 to 2022, CMR decreased an average of 0.4% annually (95% CI: −0.5% to −0.3%). We observed a decrease during most of the study period for older age groups (≥60) and a substantial reduction in mortality in the youngest group (0–9 years), particularly in the early years under analysis and until the 1990s. The period 1940–1982 showed a decreasing trend, with a growing trend from 1982 to 2022. Sex-stratified analysis revealed different trends for men and women. Between 1920 and 2022, the overall ASMR declined by 70.9%. From 1940 to 2022, the ASMR had an average annual decrease of 1.4% (95% CI: −1.6% to −1.3%), interrupted in the 1960s and 2017–2022. Distinct temporal trends by sex were also identified between 1940 and 2022.

Conclusion

We showed an overall decreasing trend in CMR and ASMR, although with different magnitudes and trajectories during the study period. To the best of our knowledge, this is the first study to describe long-term trends in both crude and age-adjusted mortality rates in Portugal from 1940 to 2022, contributing to the knowledge on mortality.

Introduction

Over the past two centuries, demographic patterns have changed profoundly worldwide. Political, economic, and social changes have led to improvements in nutrition, sanitation, and living standards, alongside advances in medicine and public health. Together, these changes have contributed to reductions in mortality and increases in life expectancy. In countries and regions that have undergone the so-called epidemiological transition, the leading causes of death shifted from infectious diseases to predominantly non-communicable diseases [1, 2]. More recently, declines in mortality from specific diseases, particularly cardiovascular diseases, have been observed. These trends are associated not only with the factors described above but also with reduced exposure to certain risk factors, such as tobacco and alcohol consumption, as well as with developments in health care [3]. Nevertheless, these improvements have not been uniform at global, regional, or national levels, with persistent economic and social inequalities continuing to play a major role in shaping mortality differentials [4, 5].

Long-term assessment of mortality based on time series is essential for public health planning, as it allows the identification of patterns and trends in mortality. Numerous studies highlight this importance by analysing age-adjusted mortality rates in extensive time series. In addition to summarizing overall mortality trends, such studies enable the identification of distinct periods with different rates of mortality decline, thereby supporting the investigation of potential causal relationships with specific health determinants and interventions [6, 7].

Several studies have analysed the evolution of mortality in the Portuguese population. Mortality patterns have changed notably throughout the 20th and 21st centuries, with a decline up to the present [8, 9]. Until the mid-20th century, this paradigm shift was primarily driven by a substantial decrease in infant mortality. In more recent decades, however, the reduction in mortality has been observed predominantly among the elderly [8]. As a consequence, life expectancy has increased markedly, leading to an increasing concentration of deaths at older ages. Despite this evidence, long-term time series studies on mortality in Portugal remain relatively scarce, underscoring the need for continued research in this area [8–10]. Thus, the aim of this study was to estimate and describe crude mortality rate (CMR) and age-standardized mortality rate (ASMR) in Portugal from 1913 to 2022, and to analyse temporal trends from 1940 onwards.

Materials and Methods

We conducted an ecological, descriptive, time series study using data on deaths and the resident population in Portugal. Official statistics for Portugal by age and sex between 1913 and 2022 were used.

Data Sources

Data on deaths were obtained from Statistics Portugal according to the following:

• 1913 to 1925: Estatística do Movimento Fisiológico da População de Portugal (Statistics from the physiological movements of the Portuguese population) [11];

• 1929 to 1967: Anuário Demográfico (Demographic yearbook) [12];

• 1941 to 1979: Estatísticas Demográficas (Demographic statistics) [13];

• 1980 to 2022: Estatísticas de Óbitos (Death statistics) [14].

Population data were also obtained from Statistics Portugal, according to data availability:

• 1920 to 1939: Estatísticas Históricas Portuguesas (Historical Portuguese Statistics), estimated resident population data [15];

• 1920 and 1930: Recenseamentos Gerais da População (Census data) [16];

• 1940 to 1979: Anuário Demográfico, Estatísticas Demográficas (Demographic yearbook, Demographic statistics) [12, 13];

• 1980 to 2022: Estimativas Anuais da População Residente (yearly estimates of population) [17].

Population data disaggregated by age group were only available for 1920, 1930, and from 1940 onwards.

Mortality Rate Estimation

CMRs were calculated annually for the total population and stratified by sex and age group (10-year intervals up to 80+). ASMRs were estimated for 1920, 1930, and annually from 1940 onwards, using the direct method and the 2013 European standard population (ESP 2013) as the reference [18]. Annual ASMR estimates were only feasible for the specified years, as age-specific population data were available only for 1920, 1930, and 1940–2022. To compute ASMR, age groups were defined based on classifications available in official death and resident population data. For the census years 1920, 1930, and 1940, data were categorized using the following age bands: 0–9, 10–14, 15–19, 20–24, 25–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–79, and 80+. For 1941–1979, data were grouped in 5-year age groups aligned with the structure of the ESP 2013, specifically 0–4, 5–9, 10–14, 15–19, 20–24, 25–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–79, 80–84, and 85+. For 1980–2022, data were classified in broader age categories: 0–4, 5–9, 10–19, 20–29, 30–39, 40–49, 50–59, 60–69, 70–79, and 80+. To ensure comparability across periods, ESP 2013 age groups were aggregated when necessary to match the broader age bands. In addition, the open-ended age groups at both extremes were harmonized across periods: 0–9 and 80+ for 1920–1940, 85+ for 1941–1979, and 80+ for 1980–2022. ASMRs were computed using the same direct standardization procedure for all years.

Trend Analysis

Trends in crude and ASMRs were assessed using joinpoint regression, employing the Joinpoint Regression Program v5.1.0.0 [19]. This method identifies the years (joinpoints) at which changes in a time trend occur, calculating the annual percentage changes (APCs) for each time segment, as well as the average annual percentage change (AAPC) for the entire period under analysis. Statistically significant positive APCs indicate increasing trends, while statistically significant negative APCs support decreasing trends. Non-significant results indicate no evidence of an increasing or decreasing trend. Joinpoint regression analyses were restricted to 1940 onwards due to limitations in data availability, as population data by age group were only available for 1920, 1930, and from 1940 onwards. Analyses were conducted overall and stratified by sex.

In the model selection process, a maximum of seven joinpoints was allowed for the 83-year period (1940–2022), yielding up to eight segments, approximately one per decade. Recent studies have shown that long time series (80–100 data points) can support models with 6–8 joinpoints when penalized selection criteria such as the weighted Bayesian information criterion (WBIC) are applied, as these criteria strongly penalize unnecessary change points and reduce the risk of overfitting [20]. In accordance with this evidence, WBIC was used for model selection, which incorporates weights to adjust the evaluation of model quality based on the characteristics of the data [20]. To assess robustness, a sensitivity analysis was performed by restricting the maximum number of joinpoints to 5. The level of significance was set at 0.05.

Reporting and Reproducibility

The STROBE checklist was used as a guide and adapted where necessary to reflect the study design (shown in online suppl. Table S1; for all online suppl. material, see https://doi.org/10.1159/000552045). All files and procedures required to reproduce the analyses (joinpoint input files and detailed steps used to compute CMR and ASMR) are deposited in an open repository (OSF; https://doi.org/10.17605/OSF.IO/MB75E).

Results

Temporal Trends in the CMRs

Between 1913 and 2022, CMR decreased by 42.1%, declining from 2,059.1 in 1913 to 1,190.2 per 100,000 population in 2022 (shown in Fig. 1; online suppl. Table S2). The CMR reduction was not uniform. In fact, from 1982 onwards, the CMR gradually increased, reversing the previous declining trajectory. During the overall period, there were years with peaks, particularly in 1918 (4,138.1 per 100,000), but also in 1941 (1,730.5 per 100,000), 1956 (1,228.8 per 100,000), and 1969 (1,162.3 per 100,000).

Fig. 1.

CMR, overall and by sex, in Portugal, 1913–2022.

In males, CMR decreased by 44.0%, from 2,211.1 in 1913 to 1,237.7 per 100,000 population in 2022, while in females decreased by 40.3%, from 1,921.9 to 1,146.9 per 100,000 population. Although males consistently had higher crude mortality rates, the tendencies and peaks were similar for both sexes (shown in Fig. 1; online suppl. Table S2).

Analysis per age group (shown in Fig. 2) indicates a decrease throughout most of the study period for older age groups (60 and older). It also shows a substantial reduction in mortality in the youngest group (0–9 years) until the 1990s and particularly in the early years under analysis. Reductions in the remaining groups are also observable mainly until the 1960s. A similar pattern was observed in both sexes; however, in males, the CMR tended to be higher in age groups older than 0–9 years. The 50–59 age group consistently registered a substantially higher CMR in males than in females throughout the entire period.

Fig. 2.

CMR by age group, overall and by sex, in Portugal, 1940–2022. For readability, age groups are presented separately for those below 60 years (left panels) and those aged 60 years and above (right panels). Results are shown overall (top), for males (middle), and for females (bottom).

From 1940 to 2022, there was an average annual decrease of 0.4% in CMR (95% confidence interval [95% CI] −0.5% to −0.3%) for both sexes. CMR joinpoint analysis detected six distinct temporal trends during the study period, as a result of five turning points. There was a marked decrease in CMR between 1940 and 1954 (APC −2.8%; 95% CI: −3.3% to −2.4%), followed by a period of stabilization between 1954 and 1974 (APC −0.2%; 95% CI: −0.4% to 0.5%). Subsequently, from 1974 to 1982, there was a decline in the rate (−1.8%; 95% CI: −4.7% to −0.8%), which was interrupted by a period of increase from 1982 to 1996 (APC, 0.8%; 95% CI: 0.4% to 3.7%). The last declining time period took place from 1996 to 2011 (APC −0.5%; 95% CI: −2.4% to −0.1) and was then followed, from 2011 to 2022, by a period of increase (APC, 1.8%; 95% CI: 1.2% to 2.8) (shown in Table 1; Fig. 3).

Table 1.

Joinpoint analysis of crude and ASMRs by sex, in Portugal, 1940–2022

​	Mortality rate (per 100,000)		Total study period		Trend 1			Trend 2			Trend 3
	1940	2022	AAPC	(95% CI)	years	APC	(95% CI)	years	APC	(95% CI)	years	APC	(95% CI)
CMR
Overall	1,560.3	1,190.2	−0.4*	(−0.5, −0.4)	1940–1954	−2.8*	(−3.3, −2.4)	1954–1974	−0.2	(−0.4, 0.5)	1974–1982	−1.8*	(−4.7, −0.8)
Male	1,641.5	1,237.7	−0.4*	(−0.5, −0.4)	1940–1954	−2.8*	(−3.3, −2.4)	1954–1976	−0.1	(−0.2, 0.3)	1976–1979	−4.6*	(−5.7, −1.5)
Female	1,485.1	1,146.9	−0.4*	(−0.5, −0.3)	1940–1953	−2.8*	(−3.7, −2.2)	1953–1983	−0.6*	(−0.9, −0.4)	1983–2014	0.2	(−0.1, 0.3)
ASMR
Overall	3,008.3	935.3	−1.5*	(−1.6, −1.3)	1940–1962	−1.8*	(−2.2, −1.4)	1962–1967	2.8	(−2.6, 6.9)	1967–2017	−2.0*	(−2.2, −1.9)
Male	3,363.5	1,137.1	−1.4*	(−1.5, −1.3)	1940–1962	−1.4*	(−1.8, −1.1)	1962–1966	4.5*	(0.7, 7.8)	1966–2022	−1.8*	(−1.9, −1.7)
Female	2,748.5	774.1	−1.6*	(−1.7, −1.5)	1940–1960	−2.1*	(−2.6, −1.7)	1960–1974	0.3	(−0.4, 2.2)	1974–1980	−4.7*	(−8.5, −2.4)

​	Mortality rate (per 100,000)		Total study period		Trend 4			Trend 5			Trend 6
	1940	2022	AAPC	(95% CI)	years	APC	(95% CI)	years	APC	(95% CI)	years	APC	(95% CI)
CMR
Overall	1,560.3	1,190.2	−0.4*	(−0.5, −0.4)	1982–1996	0.8*	(0.4, 3.3)	1996–2011	−0.5*	(−2.4, −0.1)	2011–2022	1.8*	(1.2, 2.8)
Male	1,641.5	1,237.7	−0.4*	(−0.5, −0.4)	1979–1997	0.6*	(0.3, 1.4)	1997–2011	−0.6*	(−2.1, −0.2)	2011–2022	1.6*	(1.1, 2.5)
Female	1,485.1	1,146.9	−0.4*	(−0.5, −0.3)	2014–2022	2.4*	(1.3, 5.1)	–	–	–	–	–	–
ASMR
Overall	3,008.3	935.3	−1.5*	(−1.6, −1.3)	2017–2022	1.1*	(−1.8, 7.7)	–	–	–	–	–	–
Male	3,363.5	1,137.1	−1.4*	(−1.5, −1.3)	–	–	–	–	–	–	–	–	–
Female	2,748.5	774.1	−1.6*	(−1.7, −1.5)	1980–2017	−1.9*	(−2.2, −1.8)	2017–2022	1.3	(−1.4, 7.8)	–	–	–

* indicates p < 0.05.

Fig. 3.

Trend of CMR, including identified joinpoints, overall and by sex, in Portugal, 1940–2022. Data markers represent observed rates, and lines represent modelled rates. * indicates p < 0.05. APC, annual percentage change.

In males, the CMR decreased by an average of 0.4% annually (95% CI: −0.5% to −0.4%) from 1940 to 2022. During the study period, the joinpoint analysis identified six time segments. The main difference was the third trend from 1976 to 1979, which was shorter and showed a sharper decline (APC −4.6%; 95% CI: −5.7% to −1.5) (shown in Table 1; Fig. 3).

In females, CMR exhibited an average annual decrease of 0.4% (95% CI −0.5% to −0.3%). The joinpoint analysis detected three joinpoints leading to four distinct temporal trends. Female CMR was more stable than male CMR, with a continuous decline from 1953 to 1983 (APC −0.6%; 95% CI: −0.9% to −0.4%), and stabilization from 1983 to 2014 (APC, 0.1%; 95% CI: −0.1% to 0.4%) (shown in Table 1; Fig. 3).

Temporal Trends in ASMRs

Between 1920 and 2022, overall ASMR declined by 70.9%, from 3,223.1 to 935.2 per 100,000 population (shown in Fig. 4; online suppl. Table S3). The trend was generally decreasing, interspersed with brief periods of temporary increases, mostly up until the 1970s, as observed in 1941 (3,251.2 per 100,000), 1949 (2,930.5 per 100,000), 1956 (2,794.5 per 100,000), and 1969 (2,491.7 per 100,000). After 1976, there was a continuous and steady decline in mortality rates, with specific increases in certain years, such as in 2020 (956.7 per 100,000).

Fig. 4.

ASMR, overall and by sex, in Portugal, 1920–2022.

In the same period, male ASMR declined by 67.9%, decreasing from 3,543.6 to 1,137.1 per 100,000 population, while the female ASMR dropped from 2,969.4 to 774.1 per 100,000 population (73.9%). The rate was always higher in males than in females. Overall patterns were similar in both sexes (Fig. 4; online suppl. Table S3).

From 1940 to 2022, the ASMR had an average annual decrease of 1.4% (95% CI: −1.6% to −1.3%). Joinpoint analyses revealed four time trends. The initial period, from 1940 to 1962, demonstrated a declining trend (APC −1.7%; 95% CI: −2.2% to −1.4%), followed by a subsequent period from 1962 to 1967 with an observed increase (APC, 2.7%; 95% CI: −0.3% to 6.9%). Later, between 1967 and 2017, there was a long period of declining ASMR (APC −2.0%; 95% CI: −2.2% to −1.8%) that culminated in a final period of increase from 2017 to 2022 (APC, 1.0%; 95% CI: −1.7% to 7.7%) (shown in Table 1; Fig. 5).

Fig. 5.

Trend of ASMR, including identified joinpoints, overall and by sex, in Portugal, 1940–2022. Data markers represent observed rates, and lines represent modelled rates. * indicates p < 0.05. APC, annual percentage change.

In males, there was an average annual decrease of 1.4% in ASMR during the study period (95% CI: −1.5% to −1.4%). Joinpoint analysis identified three distinct time periods, identical to the total ASMR, except for the final segment spanning from 1966 to 2022, which showed a decreasing trend (APC −1.8%; 95% CI: −1.9% to −1.7%) (shown in Table 1; Fig. 5). Regarding females, the ASMR presented an average annual decrease of 1.6% in the study period (95% CI: −1.7% to −1.5%). Five trends were identified by the joinpoint analysis. Following a temporal trend identical to that observed in males, there was a subsequent period from 1960 to 1974, characterized by a plateau (APC, 0.3%; 95% CI: −0.4% to 2.2%), and a period from 1974 to 1980, marked by a clear decline (APC −4.7%; 95% CI: −8.5% to −2.4%). After a decline similar to that observed in males between 1980 and 2017 (APC −1.9%; 95% CI: −2.2% to −1.8%), there was a distinct final period from 2017 to 2022 characterized by an increase (APC, 1.3%; 95% CI: −1.4% to 7.8%) (shown in Table 1; Fig. 5).

Sensitivity Analysis of Trends in CMR and ASMR

In the primary analysis (maximum of seven joinpoints), the WBIC-selected model did not reach the predefined upper limit and retained five joinpoints for the mortality trend. To assess the robustness of these findings, we re-estimated the model, restricting the maximum number of joinpoints to five. The model remained identical to the primary analysis, yielding the same joinpoint locations and unchanged APC/AAPC estimates (shown in online suppl. Table S4).

Discussion

This study aimed to describe mortality rates in Portugal and their evolution over time in the period from 1913 to 2022. The results of temporal trend analyses showed an overall decreasing trend in both CMR and ASMR, although with different magnitudes and trajectories.

The results revealed a substantial decrease in CMR over the analysed period. This decline was not continuous and was marked by intermittent spikes in mortality, such as the peak observed in 1918, possibly due to the Spanish influenza [21]. In fact, the temporal trend analysis identified two periods with different directions. The period between 1940 and 1982 showed a decreasing trend in CMR, particularly steep from 1940 to 1959 (APC: -2.8% per year [95% CI: -2.4 to -3.3]), possibly attributable to the pronounced reduction in infant mortality, which declined from over 150 deaths per 1,000 live births in 1910 to 19.8 deaths per 1,000 live births in 1982 [10, 22]. As observed in other European countries, infant mortality in Portugal declined sharply throughout the 20th century, due to improved healthcare, maternal and child health programmes, vaccination, and better social conditions [23]. This decline, combined with a substantial increase in life expectancy at birth – from 51.4 years in 1940 to 72.7 years in 1982, likely contributed to the overall reduction in CMR [24, 25]. The period from 1982 to 2022 was characterized by an increasing trend in CMR, despite the continued decline in infant mortality, which reached its lowest value in 2021 (2.4 deaths per 1,000 live births) [22]. This trend may be partly explained by the progressive ageing of the population, with the ageing index rising from 47 older adults per 100 young people in 1982 to 184.4 per 100 in 2022 [10, 26]. This growth was irregular and was interrupted by a declining trend between 1996 and 2011, possibly due to continuous declines in mortality even in older groups and particularly among the oldest (80 and older). This might also be influenced by demographic shifts during this period, including an increase in net migration and the transition of the natural population balance from positive to negative from 2009 onwards [27, 28]. Although life expectancy at birth continued to increase, mainly due to declining mortality from circulatory system diseases, which has pushed mortality to older ages, it appears insufficient to reduce the mortality rate [8, 9, 29]. In recent decades, Portugal, like the rest of the European Union, has experienced an ageing population. In 2023, the proportion of individuals over 65 years old in Portugal was 24.0%, the highest in the European Union [30]. Mortality rates associated with ageing have been increasing in both sexes, following trends observed in countries such as Germany, France, and Spain [31].

When stratifying by sex, we were able to identify periods with distinct trends for men and women. In particular, among men, between 1954 and 1976, we observed no changes in mortality rates, followed by a steep decline between 1976 and 1979 (APC −4.6; 95% CI: −5.7 to −1.5). This is probably related to the large-scale emigration which occurred in young men during the 1950s–1970s when male emigration consistently exceeded female emigration (e.g., 51,660 men vs. 27,857 women in 1973), followed by the return of populations from former colonies after the Colonial War [32, 33]. In 1969, the migration growth rate was −2.4% compared to the previous year, whereas in 1975 it increased to 3.8% [34]. Following this period, the male mortality trend is rather similar to the overall trend. On the other hand, women present a different evolution in mortality, highlighting the importance of a stratified analysis. Among women, we observed an overall downward trend until 1983, at a slower rate from 1953. This indicates that the overall female mortality trend has not been as affected in the period of intense male emigration. It is also interesting to note that among women the downward trend is interrupted from 1983, first by a stable period until 2014 and then by an increase until 2022. This may indicate that changes related to an ageing population started earlier among women, similarly to what occurred in European countries such as Germany. Although since 2001 the gains in life expectancy at birth have been greater for men, in 2023 women still lived on average about 5 years longer than men (Portugal: 78.4 for men and 83.7 for women; Germany: 77.9 for men and 82.7 for women) [30, 33]. Some authors suggest that this increase may be related to a reduction in uptake of riskier behaviours in males, such as smoking, and a reduction in mortality caused by cardiovascular diseases, following a similar trend observed in other countries, like the USA [7, 8]. Thus, in 2023, continuing a trend of convergence observed over recent decades, the proportion of deaths in both sexes was nearly equal (men: 50.1%; women: 49.9%) [35].

Despite the ageing of the Portuguese population and other demographic changes influencing CMR, the results observed in the ASMR suggest that decreases in mortality, after removing age effects, were only interrupted in the 1960s and after 2017. The first period of interruption (1962–1967) likely reflects the effects of emigration, particularly among men [15, 28, 32]. The second period of ASMR growth (2017–2022) includes increases in mortality caused directly and indirectly by the COVID-19 pandemic [36–38]. The impact of the pandemic is undeniable, with monthly excess mortality peaking at 60.5% in January 2021, and a recent study reported an all-cause excess of 21,423 deaths in Portugal between March 2020 and December 2021, 90% of which were directly attributable to COVID-19 [30, 39]. Despite this, the joinpoint analysis did not indicate a clear change in the period 2017 to 2022. This might be due to effects focused on a short period of time. It is also noteworthy that there was no increasing trend in the rate for men between 2017 and 2022. On the other hand, this period starts a few years before the pandemic. Given the impact of the pandemic and the lack of excess mortality data for all causes prior to 2020, it is difficult to ascertain whether this increasing trend was indeed starting before the pandemic. Assessment of data for 2023 and beyond is required to understand this trend. To the best of our knowledge, this study is the first to describe long-term trends in both crude and age-adjusted mortality rates in Portugal from 1940 to 2022, contributing to the existing knowledge on mortality. A sensitivity analysis, restricting the maximum number of joinpoints to five, produced identical joinpoint locations and APC/AAPC estimates, supporting the robustness of the findings.

Yet, there are limitations worth noting. Firstly, annual data before 1940 were lacking. Annual ASMR estimates were only possible for selected years because age-specific population data were available only for 1920, 1930, and from 1940 onwards, which limited an even longer time series analysis. Secondly, changes in data collecteion and methods of collection might introduce differences in existing data. As we are analysing official statistics of a well-established vital statistics system, this effect is likely negligible. Thirdly, joinpoint analysis might not have detected changes in trends of small magnitude or short periods of time. Finally, our analysis has not included causes of death, which could shed further light on some of the trends identified. Given the ageing of the population, examining specific causes – such as neoplasms, which increase with age – may help explain some patterns. Cause-specific analyses may also clarify indirect effects of the COVID-19 pandemic on mortality. Future analysis of death rates by cause or by place of residence, including municipal-level data, may provide further insights into mortality in Portugal.

This study provides a comprehensive overview of the long-term evolution of crude and age-standardized mortality in Portugal, highlighting distinct dynamics across time. The findings support the interpretation that the long-term decline in mortality reflects reductions in deaths at younger ages, combined with increases in life expectancy, which have resulted in a growing concentration of deaths at older ages for both sexes. Future studies are needed to investigate the underlying causes of these trends. By identifying specific patterns and shifts in mortality rates, this study contributes to a deeper understanding of long-term mortality dynamics in Portugal and may help generate new research hypotheses. Furthermore, this knowledge may also support the design and implementation of public health strategies aimed at addressing the challenges associated with an ageing population.

Statement of Ethics

This study was approved by the Comissão de Ética para a Saúde do INSA, with the internal reference number 169/2024.

Conflict of Interest Statement

The authors declare that there are no conflicts of interest. Carlos Matias Dias was one of the editorial board members of the journal at the time of the submission.

Funding Sources

This study was not supported by any sponsor or funder.

Author Contributions

Carlos Aniceto, Andreia Leite, and Carlos Matias Dias: conception of the idea and methodology; Carlos Aniceto: software, formal analysis and investigation, and writing – original draft preparation and editing. Andreia Leite and Carlos Matias Dias: manuscript review and editing. Madalena Rodrigues: resources; Carlos Matias Dias: supervision. All authors read and approved the final manuscript.

Funding Statement

This study was not supported by any sponsor or funder.

Data Availability Statement

All data used in this study come from official sources and are publicly available. To ensure reproducibility, a minimally reproducible analysis package has been deposited in an open repository (OSF/Zenodo), including (1) the input files used in joinpoint; (2) code or detailed procedures describing the calculation of crude mortality rates (CMRs), age-standardized mortality rates (ASMRs), and the standardization steps; (3) an additional file providing links to the mortality data and resident population data obtained from public sources. The repository is available at https://doi.org/10.17605/OSF.IO/MB75E.