Other ways of looking at multiple causes of death
The cause of death varies depending on the way that the causes are assessed. The underlying cause approach is not sufficient to capture the involvement of certain causes, including potentially preventable health conditions, in causing death (Bishop et al. 2023).
If we are interested, for example, in understanding sepsis as a cause of death, we know that it was one of the most common causes recorded on death certificates in 2022. However, sepsis was not identified as a leading underlying cause of death. A more complete picture of sepsis as a cause of death requires looking at more than just the underlying cause.
Other ways of looking at multiple causes highlights the role (underlying, direct or contributory) most played by a specific cause of death, and the extent to which people die with a condition, compared to from the condition.
Health conditions can play different roles in causing death
Some conditions reported as causing death are more likely to be the underlying, rather than a direct or contributory cause. Figure 8.1 shows the role played by each cause of death according to broad cause groups.
When cancer, external causes, or neurological conditions were involved in causing death, they were usually the underlying cause (in 70%, 57% and 48% of their involvement, respectively). External causes, such as surgeries, can be involved as direct causes where the underlying condition requires medical intervention, for example, a kidney transplant for chronic renal disease, or a hip replacement after a fall.
Some causes were mostly involved as direct causes: injuries (88%), ill-defined causes (75%), infectious diseases (59%), respiratory diseases (48%) and skin diseases (45%). Infectious, respiratory, and skin diseases can often arise as a complication of the underlying cause and often in a medical setting. Injuries are almost always direct causes as these reflect the consequences of external causes.
Endocrine, musculoskeletal, and mental and behavioural conditions are usually reported as contributory causes (in 61%, 71% and 72% of their involvement, respectively).
More detailed explorations could demonstrate which underlying causes are most associated with the direct causes, and linked data could identify other relevant factors such as the setting where infections are acquired. Further investigation can also highlight how particular causes contribute to specific underlying causes.
Figure 8.1: Cause of death, per cent involvement by cause type, cause groups and specific causes, 2022
The per cent involvement in death for each cause group and specific cause changes by the role it plays as different cause types. It shows that for example, cancer, external causes and neurological conditions, when they are involved in causing death are more often the underlying cause. Endocrine, musculoskeletal, and mental and behavioural conditions are usually contributory causes, while injuries, unspecified or undetermined causes, infectious diseases and respiratory diseases are usually direct causes of death.
Notes
- Deaths registered in 2022 are based on the preliminary version and are subject to further revision by the Australian Bureau of Statistics.
- The percentages across cause type may not sum to 100 due to rounding.
Source: AIHW National Mortality Database; Table S8.1.
Comparing multiple causes to underlying causes
Another way to look at multiple causes is to assess the extent to which the cause is not the underlying cause. We can do this for specific causes by comparing the death rate based on multiple causes with the death rate based on the underlying cause. This measure is referred to as the ‘standardised ratio of multiple to underlying’ (SRMU) (Désesquelles et al. 2010). The calculation and interpretation of the SRMU is shown in Table 8.1.
The SRMU helps to understand the extent to which people die with the condition compared to from the condition. When a cause is more often an underlying cause than a non-underlying cause (for example, the SRMU is 2.0), it shows that people are more likely to die from the condition than they are to die with the condition.
| Underlying cause rate (A) | Multiple cause rate (B) | SRMU (B/A) | Interpretation |
|---|---|---|---|
| 150 deaths per 100,000 | 150 deaths per 100,000 | 1.0 | Most likely to die from, than with, the cause (Cause is always the underlying cause) |
| 150 deaths per 100,000 | 225 deaths per 100,000 | 1.5 | More likely to die from, than with, the cause |
| 150 deaths per 100,000 | 300 deaths per 100,000 | 2.0 | Equally likely to die from or with the cause |
| 150 deaths per 100,000 | 780 deaths per 100,000 | 5.2 | More likely to die with, than from, the cause |
| 0 deaths per 100,000 | 150 deaths per 100,000 | . . | Most likely to die with, than from, the cause (Cause is never the underlying cause) |
.. not applicable
Figure 8.2 shows the SRMU for different causes of death.
Which conditions are people more likely to die from rather than with?
Most cancers (with the exception of prostate, non-melanoma skin and some blood cancers) have an SRMU close to 1, suggesting people are more likely to die from, than with, the cancer. Cancers themselves are often the cause that led directly to the death without other causes recorded on the death certificate.
People are more likely to die from than with suicide, accidental poisoning, falls and road traffic injuries (SRMUs between 1.0 and 1.2). Motor neuron disease and influenza also have low SRMU (1.1 and 1.4 respectively); that is, people are more likely to die from these conditions, than with, the condition.
The SRMU for COVID-19 was 1.3 suggesting that people mostly died from, than with COVID-19. Over time, as more data becomes available, we can assess changes in the SRMU to better understand the role of COVID-19 in causing death.
Which conditions do people die equally from and with?
Coronary heart disease and dementia (the 2 leading underlying causes of death in Australia) had an SRMU of 2; that is people are equally likely to die from or with these conditions. It also suggests that these causes are involved at twice the rate indicated by the underlying cause.
Which conditions are people more likely to die with rather than from?
Chronic kidney disease, diabetes and COPD are among the 20 leading underlying causes. The SRMU for these causes was 5.0, 4.5 and 2.5, respectively (Figure 8.2). This suggests that people are more likely to die with, than from, these causes. The SRMU shows for example, that in every 5 deaths per 100,000 from chronic kidney disease, 4 deaths per 100,000 had this condition as a direct or contributory cause.
Causes with very high SMRUs included hypertension, delirium, depressive disorders, drug use disorders, medical complications, anxiety disorders, tobacco use disorders and immobility/bed-bound (Figure 8.2, Table S8.2).
Causes having the highest SRMU often reflect risk factor-related conditions. Drug use disorders for example, are highly associated with deaths due to suicide, accidental poisoning and chronic liver disease (for example, Case & Deaton 2015). Assessment of the association between causes was out of scope for this report.
Figure 8.2: Standardised ratio of multiple to underlying cause (SRMU) for selected1 causes of death, 2022
Analysis of the Standardised ratio of multiple to underlying cause (SRMU) shows people are more likely to die from cancers, most external causes, motor neurone disease, influenza and COVID-19, for example, as the underlying cause. People are more likely to die with, for example, hypertension, depressive disorders, drug use disorders and tobacco use disorders as non-underlying causes. People died from and with coronary heart disease and dementia at approximately equal rates.
Notes
- Dying from the cause refers to having the condition as underlying cause. Dying with the cause refers to having the conditions as a direct or contributory cause.
- Causes were selected to demonstrate the variation in the extent of dying with compared to dying from a cause. The SRMU is available for all causes assessed in Table S8.2.
- Deaths registered in 2022 are based on the preliminary version and are subject to further revision by the Australian Bureau of Statistics.
- SIDS: Sudden infant death syndrome.
- Infant & congenital excludes SIDS.
Source: AIHW National Mortality Database; Table S8.2.
Bishop K, Moreno-Betancur M, Balogun S, Eynstone-Hinkins J, Moran L, Rao C, Banks E, Korda RJ, Gourley M and Joshy G (2023) ‘Quantifying cause-related mortality in Australia, incorporating multiple causes: observed patterns, trends and practical considerations’, International Journal of Epidemiology, 52(1): 284–294.
Case A and Deaton A (2015) ‘Rising morbidity and mortality in midlife among white non-Hispanic Americans in the 21st century’, Proceedings of the National Academy of Science USA, 112(49):15078–83, doi:10.1073/pnas.1518393112.
Désesquelles A, Salvatore MA, Frova L, Oace M, Pappagallo M, Meslé F and Egidi V (2010) ‘Revisiting the mortality of France and Italy with the multiple-cause-of-death approach’, Demographic Research, 23:771–805.