Who died in Turkey from the Covid-19 pandemic?
An analysis of excess mortality by age groups in selected Turkish provinces
The management of the SARS-CoV-2 pandemic in Turkey regarding the data on deaths and infections has been far from transparent. During the Summer of 2020, the Turkish Health Ministry changed how it published the daily infection data to attract more tourists. Later, when it changed to the old format, it started publishing disaggregated data by age group, but this was also stopped in Autumn 2020. Moreover, the Turkish Statistical Institute (TÜIK) has indefinitely postponed the publication of demographic statistics that would have allowed the calculation of excess deaths during this period. Also, no academic institution in Turkey has carried out studies of this type, unlike in other countries, out of fear of prosecution. In the absence of this crucial data, individual efforts like the one by software engineer Güçlü Yaman (later supported by the Turkish Medical Association or TTB) have been the only way to grasp the actual toll of the pandemic in Turkey.
Similarly, from April 2020 to March 2022, at NÂR Research & Consulting we kept an account of the excess mortality by tracking the 30 provinces / municipalities / towns that, in normal circumstances, accounted for more than half of the deaths by all causes in Turkey and that, at some point of the pandemic, published regular data on deaths. These were cemetery websites, province Health Directorate deceased lists or simple Town hall pages where the Mayor would publish condolence statements to the families of the deceased within its municipal boundaries. The data quality is mixed, but at least it gave us a tool to analyze the situation.
This lack of studies is what pushed me to carry out a small analysis on how the Covid-19 pandemic was affecting different age groups. The result is that in all age groups over 50 years, there has been a notable increase in mortality (between 40 and 60%), while excess mortality is more limited or null in those of younger ages. The following chart shows the tremendous lethality of the November-December 2020 wave (in which deaths doubled in many age groups) and how it moderates after the start of vaccination. However, with the Delta and Omicron Covid-19 variants, the number of deaths rose again, especially in age groups over 70 years
Here is how I did my analysis: I downloaded the data of the following 11 provinces’ (or districts within the province) death/cemetery registries (Bursa, Denizli, Diyarbakir, Gaziantep, Kahramanmaras, Kayseri, Kocaeli, Konya, Malatya, Nazilli-Aydin, and Sanliurfa). Unfortunately, the data of the biggest cities (Istanbul, Ankara, Izmir) don’t include reference to age because their municipalities interrupted the publication of age data in their online death registries, supposedly under Government pressure.
The selection of the provinces also allows us to have a fairly good geographical distribution of the data, with data coming from the most populated and economically developed Marmara Region, towns and cities from the Aegean region and Central Anatolia, and also from the less economically developed Southeastern Anatolia provinces. Unfortunately, there is no data on deaths including the age of the deceased from the Black Sea region or Eastern Anatolia provinces.
I selected the data from the first half (days 1st to 15th) of the April and November months of 2020 and 2021 which broadly correspond to the way up to the 1st, 2nd and 3rd (Alpha) waves and the plateau of the 4th (Delta) long wave of Covid-19 infections that Turkey has experienced. I compared it with the data of 2019.
Having scrapped all the data from these 11 provinces, I grouped them by age groups ranging from 1–9 to 90+. Deaths reported in these lists at 0 refer to deaths at birth or during pregnancy. First, I grouped them at the provincial level and then all together.
The first wave in Turkey was mainly limited to Istanbul and the Marmara region, so the overall effect on deaths is not perceptible in Chart 1. However, it can be appreciated if we focus on the Marmara region’s provinces (Bursa and Kocaeli).
The second wave was genuinely lethal. The vaccine rollout had not yet started, and the Turkish Government chose to withhold the actual number of cases and fatalities, despite medical unions and doctors complaining about the hospital situation.
However, the third Covid-19 wave, even if it had more cases detected than the previous one, had less mortality because the older population was recently vaccinated in a vaccination campaign that started early compared to other European countries, but its pace was later reduced.
Turkey started vaccinating its at-risk population (older people, people with chronic diseases and Health workers) in January 2021. At the beginning of the vaccine rollout, only inactivated virus CoronaVac, developed by the Chinese Sinovac, was available, although later, during the Spring, the mRNA vaccine developed by Pfizer–BioNTech was also made available, and by the end of the year, was the most used. Unfortunately, the protective effect of the first two vaccine doses vanished, specially in the CoronaVac case, and not all the target population received a booster (see these charts), so during Autumn 2021, we had another increase in deaths, especially within the older population.
Overall, this is the excess mortality per age group I have detected in the four different periods studied:
50–59: +523 (+51%)
60–60: +1,190 (+65%)
70–79: +1,567 (+60%)
80–89: +1,344 (+47%)
>90: +293 (+29%)
For the group 0 to 29 years old, there is no significant excess mortality or a deficit. In the case of 30–39 and 40–49 years old, there is a significant increase in relative terms (+40%), but as the absolute numbers are still low (100), we need more data to reach any conclusion.
This analysis was carried out at the end of 2021. Later, in March 2022, I added the data of the first 15 days of February to compare the effects of the Omicron wave, which produced a record number of infections. Due to lack of time, I could not scrap the February data of previous years, so the comparison quality is lower than in the previous cases. However, it also points to a worrying increase in mortality, especially for the aged over 70.
And that is it! This could be a good starting point for those who wish to study the effects of the Covid-19 pandemic in different age groups. Any comment or criticism is welcome.
— — — — — — — — — — — — — — — — — — — — — — — — —
The source of the data is the following:
For Bursa, Denizli, Diyarbakir, Kahramanmaras, Kayseri, Kocaeli, Konya and Malatya, I used data from the provincial death registries published in the E-Devlet platform of the Turkish State. https://www.turkiye.gov.tr/arama?aranan=Vefat%20Bilgisi%20Sorgulama
For Gaziantep, I used the provincial-level data of the Municipality https://www.gaziantep.bel.tr/tr/defin-listesi
For Sanliurfa, I used the online book of condolence at the Municipality website https://www.sanliurfa.bel.tr/kategori/80/0/taziye-defteri?page=1
For Nazilli, a district within the Aydin province, I used the Municipality online death registry https://ebelediye.nazilli.bel.tr/webportal/index.php?wwsayfa=72
