# 2.1 Measures of mortality and morbidity

## 2.1.1 Mortality statistics

Mortality data from death certificates and from census and population registers are routinely collected; from these the death rate in a population can be calculated. To calculate a death rate the number of deaths recorded is divided by the number of people in the population, and then multiplied by 100, 1,000 or another convenient figure.

The *crude death rate* shows the number of deaths in the total population and, for the sake of manageability, is usually calculated per 1,000. It is calculated as follows:

Crude death rates do not show the burden of deaths in particular groups in the population. For example, one might assume that a town such as Bournemouth is an unhealthy place because it has a high crude death rate, but on closer examination this is found to be due to the fact that it is a popular place to retire to and so has a high proportion of older people. To counter this problem, *age-specific* rates can be calculated as follows:

As well as straightforward age-specific rates, certain special age rates can be calculated which are of particular importance in public health. The infant mortality rate (IMR) is used as an indicator of the overall health of a nation or community because this rate correlates well with young adult mortality, but is more sensitive to socio-economic and environmental improvements as well as to improvements in healthcare. The IMR is calculated as follows:

Other special rates of deaths in infants under one year of age include the annual stillbirth rate, late foetal deaths after 24 weeks of gestation and the annual perinatal mortality rate (stillbirths plus deaths in the first week of life), and are calculated in the same way as infant mortality. Figure 9 shows these rates for the UK in 2003.

*Source: Office for National Statistics; General Register Office for Scotland; Northern Ireland Statistics and Research Agency*The death rates discussed above are based on all causes of death. The cause-specific death rate is used to calculate how many deaths occurred from specific diseases such as cancer or heart disease. This would be calculated as follows:

The calculations made so far can provide the overall crude death rate for a population, the death rate for different age groups and deaths from different causes, but do not allow for a comparison to be made between one part of the country and another. For instance, as we noted above, south coast resorts such as Bournemouth have a preponderance of older people and consequently a high death rate, whereas a population with a high proportion of young people, as in a new town, is likely to have a low death rate. A direct comparison of crude mortality rates for the two localities would obviously produce a distorted picture. So the death rate for a specific condition in a particular area may be higher than the national average simply because the area contains relatively more residents in a susceptible age group than the national population. *Standardised Mortality Ratios* (SMR) can compare mortality rates between different geographical areas, taking age differences into account. (How an SMR is calculated in given in Box 1.) So despite the very different age structures of the populations involved, regional comparisons can be made, as Figure 2 demonstrates for cities in Wales.

### Box 1 Calculating a Standardised Mortality Ratio

Death rates for age groups (or other groups) in the standard [national] population are multiplied by the population of the same groups in the study population. This produces an ‘expected’ number of deaths representing what the number of deaths in the study population would have been, if that population had the same death rates as the standard population. The observed (or actual) number of deaths in the study population is then divided by the total expected number and multiplied by 100. This produces an SMR. The standard population always has an SMR of 100, with which the SMR of the study population can be compared. The SMR figure is actually a percentage. This means that if the study population’s SMR is 130, its death rate is 30% higher than that of the standard population. If the study population’s SMR is 86, then its death rate is 14% lower than that of the standard population.

A somewhat tongue-in-cheek study reported in the *British Medical Journal* (Crayford et al., 1997) used SMRs to study the death rates of characters in soap operas on British television. They found that being a character in *EastEnders *was the most dangerous job in Britain, as Table 1 shows.

### Table 1 Standardised Mortality Ratios for various high-risk groups in comparison with the general population

SMR | |
---|---|

EastEnders characters | 771 |

Formula One drivers | 581 |

Coronation Street characters | 353 |

Oil rig divers | 235 |

Bomb disposal experts | 196 |

Steeplejacks | 148 |

General population | 100 |

### Activity 1 SMRs in Camden, London

You have been introduced to the ways in which epidemiology uses mortality data to compare death rates between different groups in the population by age, sex, class and geographical area. This is done by calculating the Standardised Mortality Ratio (SMR). The SMR is a much-used statistic in public health. You will now carry out a calculation of an SMR.

Look at Table 2 below and read through again the explanation of how to calculate an SMR given in Box 1 above. Then, from the data presented in Figure 4, calculate the SMR for male mortality from ischaemic heart disease (IHD) in Camden and compare it with that for males aged 35-74 in Great Britain as a whole

The SMR can be calculated either for all-cause mortality or for specific diseases. It can be calculated for all ages, or for any range of ages (e.g. 35–74 years), and over any specified period of time. Usually it is calculated for males and females separately, since men and women often have different mortality experience.

Whatever the precise details, the SMR is always calculated in the same general way. Age-specific death rates for the reference population (generally the country as a whole) are applied to the area under study (as in Table 2, men in Camden) to obtain the expected number of deaths in each age group (i.e. the number of deaths expected if national death rates were seen in Camden). The total number of expected deaths in all age groups is then compared with the number of deaths actually observed in Camden. An SMR of 100 suggests that the overall mortality rate for the study area (Camden) matches that of the country as a whole; a figure of more than 100 suggests that the local area has a mortality rate above the national average, and so on.

#### Table 2 A Comparison of male mortality from ischaemic heart disease (IHD) in Camden with that in Great Britain as a whole (for ages 35–74 years)

Age group | National rate (IHD) (GB men) | Camden population (men) | Camden expected No. deaths | Camden observed No. IHD deaths (men) | |||

35–44 yrs | 2.6/1000 | x | 9,410 | = | 24.4 | 20 | |

45–54 yrs | 12.7/1000 | x | 8,700 | = | 110.4 | 109 | |

55–64 yrs | 35.9/1000 | x | 8,548 | = | 306.7 | 306 | |

65–74 yrs | 79.4/1000 | x | 6,892 | = | 547.4 | 456 | |

Total expected | 988.9 | Total observed | 891 |

#### Comment

If the SMR is the observed deaths divided by the expected deaths multiplied by 100 then the SMR for male mortality from ischaemic heart disease in Camden for men aged between 35 and 74 is: 891 divided by 989 x 100 = 90. So the IHD mortality in Camden for men aged between 35–74 appears to be below the national average with an SMR of 90.

The SMR can be calculated either for all-cause mortality or for specific diseases. It can be calculated for all ages, or for any range of ages (e.g. 35–74 years), and over any specified period of time. Usually it is calculated for males and females separately, since men and women often have different mortality experience.

Whatever the precise details, the SMR is always calculated in the same general way. Age-specific death rates for the reference population (generally the country as a whole) are applied to the area under study (as in Table 2, men in Camden) to obtain the expected number of deaths in each age group (i.e. the number of deaths expected if national death rates were seen in Camden). The total number of expected deaths in all age groups is then compared with the number of deaths actually observed in Camden. An SMR of 100 suggests that the overall mortality rate for the study area (Camden) matches that of the country as a whole; a figure of more than 100 suggests that the local area has a mortality rate above the national average, and so on.

So far we have been focusing on mortality data, but how is the distribution of morbidity measured?