Background Excess all-cause mortality has been used in many countries as an estimate of mortality effects from COVID-19. What was the excess mortality in Israel in 2020 and when, where and for whom was this excess? Methods Mortality rates between March to November 2020 for various demographic groups, cities, month and week were compared with the average rate during 2017–2019 for the same groups or periods. Results Total mortality rates for March–November were significantly higher by 6% in 2020, than the average of 2017–2019, 14% higher among the Arab population and 5% among Jews and Others. Significantly higher monthly mortality rates were found in August, September and October by 11%, 13% and 19%, respectively, among Jews and Others, and by 19%, 64% and 40% in the Arab population. Excess mortality was significant only at older ages, 7% higher rates at ages 65–74 and 75–84 and 8% at ages 85 and above, and greater for males than females in all ages and population groups. Interestingly, mortality rates decreased significantly among the younger population aged under 25. The cities with most significant excess mortality were Ramla (25% higher), Bene Beraq (24%), Bat Yam (15%) and Jerusalem (8%). Conclusion Israel has seen significant excess mortality in August–October 2020, particularly in the Arab sector. The excess mortality in March–November was statistically significant only at older ages, over 65. It is very important to protect this susceptible population from exposure and prioritize them for inoculations. Lockdowns were successful in lowering the excess mortality. The excess mortality is similar to official data on COVID-19 deaths.

BackgroundRegional variations in mortality can be used to study and assess differences in disease prevalence and factors leading to disease and mortality from different causes. To enable this comparison, it is important to standardize the mortality data to adjust for the effects of regional population differences in age, nationality and country of origin.MethodsStandardized mortality ratios (SMR) were calculated for the districts and sub-districts in Israel, for total mortality by gender as well as for leading causes of death and selected specific causes. Correlations were assessed between these SMRs, regional disease risk factors and socio-economic characteristics. Implications for health policy were then examined.ResultsTotal mortality in the Northern District of Israel was not significantly different from the national average; but the Haifa, Tel Aviv, and Southern districts were significantly higher and the Jerusalem, Central, Judea and Samaria districts were lower.Cancer SMR was significantly lower in Jerusalem and not significantly higher in any region. Heart disease and diabetes SMRs were significantly higher in many sub-districts in the north of the country and lower in the south. SMRs for septicemia, influenza/pneumonia, and for cerebrovascular disease were higher in the south. Septicemia was also significantly higher in Tel Aviv and lower in the North, Haifa and Jerusalem districts. SMRs for accidents, particularly for motor vehicle accidents were significantly higher in the peripheral Zefat and Be’er Sheva sub-districts.ConclusionThe SMR, adjusted for age and ethnicity, is a good method for identifying districts that differ significantly from the national average. Some of the regional differences may be attributed to differences in the completion of death certificates. This needs to be addressed by efforts to improve reporting of causes of death, by educating physicians.The relatively low differences found after adjustment, show that factors associated with ethnicity may affect mortality more than regional factors. Recommendations include encouraging good eating habits, exercise, cancer screening, control of hypertension, reduction of smoking and improving road infrastructure and emergency care access in the periphery.Electronic supplementary materialThe online version of this article (doi:10.1186/s13584-017-0164-1) contains supplementary material, which is available to authorized users.