Table 1 Prior research on the topic

1. Bystrova (1965);

2. Bednyi et al. (1980,1981);

3. Ovcharov, Bystrova (1982) *studies cited by Meslé et al. (1996) [31]

1. Cities of Tula, Novomoskovsk, Tambov, Michurinsk in the early 1960s

2. “Different regions of Russia” (not specified), 1979.

3. Belarus and Turkmenistan, 1981–1982.

To assess the accuracy of cause-of-death diagnostics and coding in the USSR.

- Re-inspection of medical death certificates.

- Number of deaths in the sample, age and sex distributions not described.

- Against a widely held view, no evidence was found of substantial overestimation of cardiovascular mortality in the USSR and Russia.

- High percentages of false diagnoses for certain circulatory causes with compensation within the same ICD chapter were found. The total for the whole class was trustworthy.

- The lowest percentage of misclassified cases was found for neoplasms and external causes; the highest was found for digestive and respiratory diseases.

Shkolnikov, Chervyakov et al. (2000) [34]

Cities of Udmurt Republic, 1998-1999

To inspect the quality of cause-of-death coding under the conditions of the “mortality crisis” of the 1990s.

- Analysis of 1,023 medical death certificates of males aged 20–55.

- Re-coding of 782 deaths by an independent and experienced expert.

- For 98 % of deaths the ICD chapter of the initial underlying cause coincided with the cause specified by the expert.

- The level of concordance decreased at the lower level of cause-of-death aggregation.

- The most significant misclassification was found within the chapter of circulatory diseases and within the chapter of external causes of death.

Vaysman (2013) [40]

Tula obl., 2008, 2012

To analyze the accuracy of the reporting of causes of death in the Tula region.

- Re-coding of a sample of 148 medical death certificates.

- Share of deaths was attributed to circulatory diseases decreased from 66.2 % to 56.8 % after re-coding.

- Share of deaths from endocrine disorders, neoplasms, digestive system disorders increased.

1. Meslé et al. (1996) [31]

2. Shkolnikov et al. (1996) [32, 33]

1. Russia as a whole, 1965–1994

2. Russia as a whole, 1970-1993

To analyze the different aspects of mortality for a long-term period, and to explore the components of the health crisis in Russia (USSR).

- The method of reconstruction [45, 46, 58].

- Visual inspection of cause- and age-specific mortality trends.

- Overestimation of the cardiovascular deaths total was not confirmed.

- CVD mortality may even be underestimated among those aged 80+ due to excess use of “senility.”

- Some misclassification found for several causes within the class of circulatory diseases.

Shkolnikov et al. (2002) [35]

Izhevsk, 1998-1999

To determine whether deaths from alcohol poisoning were being misclassified as cardiovascular deaths.

- 309 deaths of males aged 20–55 with necropsy records were checked for blood alcohol concentration (BAC).

- No evidence that alcohol poisonings were misclassified as cardiovascular deaths was found.

- Among 10 deaths (3.2 % of the sample) with potentially lethal BAC (>4 g/L) seven were coded as accidental poisonings, two as suicides, one as a traffic accident.

- No deaths with BAC > 4 g/L were coded as cardiovascular diseases.

Zaridze et al. (2009) [37]

Barnaul, 1990-2004

To determine the role of alcohol in unusual fluctuations in mortality in Russia.

- Inspection of cause-specific mortality trends in Russia in 1991–2006.

- BAC values and official diagnoses for 24,836 forensic autopsies in 1990–2004 in the city of Barnaul.

- A post-mortem potentially lethal BAC (>4 g/L) found for 14 % of deaths of males aged 35–69 officially recorded as deaths from cardiovascular causes.

- During the period of large mortality fluctuations in Russia in the 1990s, mortality from myocardial infarction (MI) did not follow the fluctuating trends of other cardiovascular causes. A low percentage of the deaths with lethal BAC levels were found among MI deaths.

- Abrupt changes in the death rates from many circulatory causes may be caused by the misclassification of alcohol poisonings.

Leon et al. (2010) [36]

Izhevsk, 2003-2005

To find a link between alcohol and mortality among men of working age.

- Inspection of cause-specific mortality trends in Russia in 1980–2007.

- Case–control study (1750 cases – deaths of males aged 25–54; 1750 controls – live men) with interviews.

- Deaths subjected to forensic autopsy checked for BAC.

- Criticism of the results of Barnaul study (Zaridze et al., 2009): 1) mortality from cerebrovascular diseases fluctuated significantly in the 1980s and the 1990s, in tandem with mortality from alcohol causes; yet the percentages of deaths with BAC >4 g/L were very low for this group of causes in both Barnaul and Izhevsk. These findings contradict the evidence from the Barnaul study; 2) during the years immediately following the transition to the RC-1999 there was an artificial conflation of mortality from ischemic and non-ischemic heart diseases in Barnaul. Thus, the mortality trends in Barnaul were not nationally representative.

- In Izhevsk, only 5 % of deaths certified as deaths from circulatory diseases showed BAC higher than 4 g/L at post-mortem autopsy.

Sidorenkov et al. (2011) [24]

Arkhangelsk, 2008-2009

To determine whether deaths from alcohol poisoning were misclassified as cardiovascular deaths.

- All deaths at ages 30–70 from cardiovascular diseases subjected to forensic autopsy checked for BAC.

- No evidence of alcohol poisonings being misclassified as cardiovascular deaths was found.

- Only six cases of deaths with BAC higher than 4 g/L were certified as cardiovascular deaths.

Gavrilova et al. (2008) [27]

Russia as a whole, 1991–2005; Kirov and Smolensk regions, the city of Moscow, 2003

To investigate which causes of death are hidden under ill-defined conditions.

- Descriptive statistical analysis of mortality from ill-defined conditions in Russia.

- Analysis of various types of information from medical death certificates.

- The instructions on death certification in Russia encourage medical practitioners to use unspecified diagnosis in medical death certificates.

- A significant fraction of deaths from ill-defined conditions at working age are deaths from external causes (including violence) hidden under other diagnoses.

Lopakov (2011) [39]

Kaluga region, 2002

To find mistakes in medical death certificates.

- Analysis of 419 medical death certificates.

- Mistakes in medical death certificates were found.

- The shares of “other” and “unspecified” diagnoses were too high.

Roschin et al. (2013) [41]

- Three hospitals in the Moscow region, 2002

- Tver, Tula, and Kaluga regions, 2002

To assess the accuracy of reporting diabetes in medical death certificates.

- Comparison of medical death certificates and medical records for individuals who died in hospitals.

- Analysis of the information reported in medical death certificates

- While 25 % of the deceased in hospitals had diabetes, this diagnosis was not specified in death certificates (other than in a few cases in which diabetes was selected as an underlying cause of death). The frequency of the reporting of diabetes in medical death certificates did not correspond to its real prevalence in the population.

Pridemore (2003) [42]

Russia as a whole 1987–1998; 78 regions of Russia 1994-1998

To evaluate the homicide reporting in Russia.

- Comparison of two sources of homicide estimates in Russia: data from the vital statistics and data from the Ministry of the Interior.

- Disparities across regions in the reporting of homicides were found in the mortality and crime data.

- In the majority of regions (66 of 78) the number of homicides in the vital statistics was higher than the number in the crime statistics, though the magnitude of the difference varied.

- Opposite ratios were observed in 12 regions.

Nemtsov (2003) [43]

77 regions of Russia, 1990-2001

To estimate alcohol-related mortality in the Russian regions.

- Analyzing the mortality from alcohol-related causes in regions.

- Comparison of mortality levels from alcohol poisonings and alcohol psychoses across the regions.

- Among all alcohol-related causes of death acute alcohol poisonings had the highest variability across regions.

- The mortality levels and the dynamics of the alcohol poisonings and alcohol psychoses did not correspond to each other in many regions; this finding contradicts the current understanding of the link between these causes, and can thus be regarded as a statistical artifact.