Intended for healthcare professionals


Effect of socioeconomic group on incidence of, management of, and survival after myocardial infarction and coronary death: analysis of community coronary event register

BMJ 1997; 314 doi: (Published 22 February 1997) Cite this as: BMJ 1997;314:541
  1. Caroline Morrison, consultant in public health medicinea,
  2. Mark Woodward, senior lecturer in statistical epidemiologyb,
  3. Wilma Leslie, senior research nursea,
  4. Hugh Tunstall-Pedoe, professorc
  1. a MONICA Project, Royal Infirmary, Glasgow G31 2ER
  2. b Department of Applied Statistics, PO Box 240, University of Reading, Reading RG6 6FN
  3. c Cardiovascular Epidemiology Unit, Ninewells Hospital and Medical School, Dundee DD1 9SY
  1. Correspondence to: Dr Morrison
  • Accepted 6 December 1996


Objective: To investigate the effect of socioeconomic group (with reference to age and sex) on the rate of, course of, and survival after coronary events.

Design: Community coronary event register from 1985 to 1991.

Setting: City of Glasgow north of the River Clyde, population 196 000.

Subjects: 3991 men and 1551 women aged 25-64 years on the Glasgow MONICA coronary event register with definite or fatal possible or unclassifiable events according to the criteria of the World Health Organisation's MONICA project (monitoring trends and determinants in cardiovascular disease).

Main outcome measures: Rate of coronary events; proportion of subjects reaching hospital alive; case fatality in admitted patients and in community overall.

Results: Event rates increased with age for both sexes and were greater in men than women at all ages. The rate increased 1.7-fold in men and 2.4-fold in women from the least (Q1) to the most (Q4) deprived socioeconomic quarter. The socioeconomic gradient decreased with age and was steeper for women than men. The proportion treated in hospital (66%) decreased with age, was greater in women than men, and decreased in both sexes with increasing deprivation (age standardised odds ratio 0.82 for Q4 v Q1) Case fatality in hospital (20%) increased with age, was greater for women than men when age was standardised, and showed no strong socioeconomic pattern. Overall case fatality in the community (50%) increased with age, was similar between the sexes, and increased from Q1 to Q4 (age standardised odds ratio 1.12 in men, 1.18 in women).

Conclusions: Socioeconomic group affects not only death rates from myocardial infarction but also event rates and chance of admission. This should be taken into account when different groups of patients are compared. Because social deprivation is associated with so many more deaths outside hospital, primary and secondary prevention are more likely than acute hospital care to reduce the socioeconomic variation in mortality.

Key messages

  • Socioeconomic variation in rates of coronary events was greater for women than men

  • The largest social class gradient was in the proportion of deaths occurring outside hospital

  • Overall, 68% of all people who died of coronary events did so before admission

  • Acute hospital care applied to only 66% of all cases and therefore could affect only 32% of all deaths

  • Reduction in socioeconomic variation in mortality from coronary disease is best addressed by reducing the variation of event rates-that is, by primary and secondary prevention

  • Allocation of resources for reduction of coronary mortality should take account of social class differences and the relative potential effect of hospital care and primary and secondary prevention


Age, sex, and socioeconomic group all affect a population's mortality from coronary heart disease.1 2 3 Population registers have been used to document the effect of age and sex on event rates and case fatality in a community.4 5 6 Studies limited to patients admitted to coronary care units or hospital wards 7 8 9 10 or to those in drug trials related to myocardial infarction11 12 have reported the relation between case fatality in hospital, age, and sex. Social class may be related to event rates,13 14 15 16 17 delay in seeking medical care during myocardial infarction,18 19 and case fatality in hospital.20 We investigated the relation of socioeconomic group to four things-event rates, the proportion of people reaching hospital alive, mortality in hospital, and overall case fatality-in men and women who had fatal and non-fatal consecutive events over a period of seven years in a defined geographical population. Wilhelmsen and Rosengren recently showed the surprising lack of data on these things.21

Subjects and method

The Glasgow MONICA coronary event register comprises all cases of myocardial infarction and coronary death occurring in men and women aged 25-64 years resident in north Glasgow. The methods of ascertainment and investigation have been described previously.5 Ascertainment derives from hospital discharge data, the registrar general, and other, minor, sources. We validated the events ascertained using the criteria of the World Health Organisation's MONICA (monitoring trends and determinants in cardiovascular disease) project,6 and these depend on symptoms, electrocardiographic evidence, cardiac enzyme concentrations, and necropsy reports. Diagnostic information was gathered from many sources, including hospital case notes, general practitioners, and police reports to the procurators fiscal (the medicolegal authority in Scotland). Survival was determined at 28 days from the onset of the event. For this analysis we used all events classified as definite and all fatal events classified as possible and unclassifiable.6 This definition excludes non-fatal admissions without categorical evidence of infarction.

Socioeconomic group was derived for the postcode of residence at the time of the illness using the Carstairs and Morris deprivation score2 as updated by McLoone using 1991 census data.22 In this study we ranked postcodes in the order of the score. The population aged 25-64 for each postcode sector was obtained from the small area statistics from the 1991 census, and the total study population was divided into quarters (socioeconomic quarters) according to the updated Carstairs and Morris score rankings (Q1 is the least deprived quarter and Q4 the most deprived). We used the populations of the Carstairs and Morris quarters at the 1991 census as the denominators for calculation of age standardised rates. The specific populations in each socioeconomic quarter for each year were not available. The denominator for case fatality in hospital was all those who reached hospital alive. It did not include those transported to hospital but found to be dead on arrival.

Age standardised event rates were calculated as weighted sums of Poisson variables using the world standard population as the reference population.23 Age adjusted relative risks by socioeconomic quarter were calculated from Poisson regression models. Age standardised percentages of people with events who reached hospital alive, of those who died after reaching hospital alive, and of those dying overall were calculated as weighted means of binomial variables using the Glasgow male population of all events in this study as the reference population. Age adjusted odds ratios by socioeconomic quarter were calculated from logistic regression models. Age standardisation and age adjustment of relative risks were carried out using 10 year age bands. Age adjustment of odds ratios used one year age bands.


Population distribution

The estimated study population for 1991 was 96 063 men and 99 768 women. The total population in the age group 25-64 years declined each year from 1985 until a small upturn in 1991.

The socioeconomic distribution of the population of north Glasgow is skewed towards greater deprivation compared with Scotland overall. In north Glasgow 68% of the population live in postcode sectors with scores below a cut off point that defines 18% of the Scottish population when applied nationally. The populations of each deprivation quarter of the population showed different age structures both between the sexes and between quarters for each sex (fig 1), although there was no systematic pattern from Q1 to Q4.

Fig 1
Fig 1

Distribution of Glasgow MONICA population by age, sex, and socioeconomic quarter (Q1 is least deprived)

Event rate

Figure 2 shows the aggregate numbers of events at various points in the pathway through myocardial infarction and coronary death in the Glasgow MONICA register from 1985-91. Deaths outside hospital accounted for 68% of all deaths. The community case fatality at 28 days (all deaths, both in and out of hospital, related to all events) was 50%.

Fig 2
Fig 2

Pathway through myocardial infarction and coronary death in Glasgow MONICA project, 1985-91

More men than women had events (3991 v 1551). The age distribution of events for men and women is shown in table 1. Event rates were much greater for men than women for each 10 year age band. Both sexes showed an increase with age. The distribution of the events by deprivation quarter is also shown in table 1. For both men and women the risk of myocardial infarction increased with increasing deprivation. The socioeconomic gradient was steeper in women (P=0.01). The effect of deprivation was highly consistent across all age groups in both sexes. The socioeconomic gradient decreased with age for men and women, although men in the youngest age group and women in the two youngest age groups were too few to provide meaningful comparisons.

Table 1

Numbers of events (MONICA definition) by age, sex, and socioeconomic quarter. Values in parentheses are age specific relative risks unless stated otherwise

View this table:

Arrival at hospital

A total of 3627 (65%) subjects arrived at hospital alive. Table 2 shows the variation by age and sex. For both men and women there was no obvious pattern with age except that the oldest group of men had less chance than the younger men of being admitted. Only older women were more likely to reach hospital alive than men. The variation by socioeconomic quarter showed a decreasing proportion reaching hospital alive (Q1 to Q4) for men and women, with no obvious difference in the socioeconomic gradient between the sexes.

Table 2

Numbers (percentages) of people with coronary events who were alive on arrival at hospital by age, sex, and socioeconomic quarter

View this table:

Fatality in admitted patient

In all, 854 of the 3627 (22%) of patients who reached hospital alive died within 28 days. The case fatality in hospital of all those reaching hospital alive-including those who died in an accident and emergency department-was 21% (769/3627); it was 20% (708/3566) of all those reaching a hospital ward alive (fig 2). Two per cent (83/3566) of those admitted to hospital were discharged alive but they died within 28 days from the onset of the illness; 2%(70/3566) were alive at 28 days, although they were still in hospital.

Table 3 shows the case fatality in hospital for each age group and for both sexes. It increased with age and more steeply in women than men. Age standardised case fatality rose consistently with increasing deprivation for men. In women the pattern was not consistent but a contrast was seen between the most deprived and least deprived quarter.

Table 3

Numbers (percentages) of patients dying of coronary event within 28 days among those reaching hospital alive by age, sex, and socioeconomic quarter

View this table:

Community case fatality (in and out of hospital

The case fatality in the community overall was 50% (2769/5542) We previously showed the remarkable similarity in case fatality in men and women at 28 day even when age was standardised5; table 4 shows other variations. Case fatality increased with age in both men and women. Age standardised case fatality rose consistently with increasing deprivation in men; the pattern was not consistent in women but showed an increase between the least deprived and most deprived quarter.

Table 4

Numbers (percentages) of patients dying of coronary event within 28 days among all those who had events by age, sex and socioeconomic quarter

View this table:


The population of north Glasgow is skewed to the lower end of the socioeconomic scale, but we still found socioeconomic gradients in myocardial infarction among people who overall are least advantaged. This socioeconomic variation is seen in event rate and most obviously in the proportion of people reaching hospital alive. No variation is seen in case fatality in hospital. These findings are reflected in the socioeconomic variation in the proportion of people dying outside hospital and the case fatality in the community overall. Socioeconomic disadvantage thus increases the chance of a person having a myocardial infarction, decreases the chance of reaching hospital alive, and increases the chance of dying during the attack. This gradient is found in both women and men.

Event rate

We found a large socioeconomic variation in the event rate in men and women. This agrees with studies using various markers of socioeconomic group which show that deprivation, a shorthand for poorer education, poorer housing, increasing poverty, and lack of employment, relates to increasing event rates.14 15 16 Furthermore, socioeconomic variations in the extent of coronary risk factors in a population are associated with socioeconomic variation in disease prevalence.24 This supports the idea of reducing the unequal distribution of risk as a prime objective for local and national health promotion activities.

Arrival at hospital

The socioeconomic differences we and others have found may reflect poorer awareness of the importance of symptoms such as crushing chest pain or arm pain among those who are less well educated.18 19 Alternatively, the proportion of sudden deaths may rise by association with cigarette smoking.25

Seventy one per cent of all deaths in men but only 63% of deaths in women occurred outside hospital, reflecting the greater proportion of women arriving alive at hospital. A similar socioeconomic gradient was seen in men and women. We have already shown that women are more likely to reach hospital after a call to the family doctor, whereas men are more likely to go directly.5 Perhaps referral patterns also vary subtly according to patients' backgrounds.

Only two thirds of patients were admitted to hospital. The remaining third died outside hospital, limiting the impact of treatments in hospital.

Death rate in admitted patient

We found no social class variation in case fatality in hospital. This is consistent with the Scottish Office's data on mortality in hospital 30 days after a myocardial infarction, which show little variation when standardised for social class.26 A greater case fatality in hospital has been reported among people who are less well educated20 and among African-American women (with higher unemployment rates) compared with white women,27 and this was only partly explained by differences in case mix. Our findings may reflect a lack of bias in the care given to different socioeconomic groups. The case mix of hospital admissions may vary by socioeconomic quarter.20 Severely ill people from deprived background might be more likely to die outside hospital because of delays in admission. If care shows a socioeconomic bias against deprived people this might cancel out a better case mix.

Cardiologists and general physicians differ in their management of patients during myocardial infarction28 and unstable angina.29 Do they differ in how they care for patients with myocardial infarction of different socioeconomic groups? Socioeconomic bias applies to angiography for men in Glasgow30 and elsewhere31 but not to invitation to rehabilitation after myocardial infarction.32 Furthermore, men and women are treated similarly once they are receiving care for myocardial infarction.5

Community case fatality

The increase in overall case fatality with deprivation may have several explanations. Non-fatal events in deprived people may be less frequently recognised-that is, the denominator for case fatality in the more deprived groups may be comparatively more incomplete. Death, however, is enumerated similarly for all social groups. Any undercount in the denominator would imply an underestimate of the socioeconomic gradient.

Several factors could explain the variation in the numerator between socioeconomic groups. Firstly, the number of concomitant illnesses, particularly respiratory disease, increases with increasing deprivation.33

Secondly, coronary disease itself may be different in different socioeconomic groups, perhaps manifesting more often among deprived people as sudden death through the mediation of factors such as higher rates of cigarette smoking.25

Thirdly, there may be a social class gradient in the ability to heal or to ward off insults to various organs. In Glasgow recovery from various surgical procedures for cancer is worse among people who are deprived after adjustment for stage of disease and treatment.34

Finally, the potential for resuscitation, available for 38% of people who died outside hospital in north Glasgow, was fulfilled in less than a third of that number.35 This proportion is much greater elsewhere.34 The advent of Heartstart Scotland-the equipping of all emergency ambulances with semiautomatic defibrillators-has had little effect on successful resuscitation outside hospital in north Glasgow35 in contrast to other places.36 Exploration of the possible reasons for these differences is required to maximise the benefit of available services such as Heartstart Scotland.

The quality of Glasgow MONICA registration data in terms of completeness, accuracy, and consistency over time has been documented.5 6 Carstairs and Morris deprivation scores are better than occupational classification in discriminating between deprived socioeconomic groups as many of those on the register are unemployed.5 Postcode of residence, but not occupation, is routinely coded on both death certificates and hospital discharge data. The Carstairs and Morris score has been criticised for its complexity and for not relating to people but to geographical areas. However, it remains useful in describing variation over a wide range of morbidity, mortality, and other health related population measurements.2 38 39 Calculation of rates for deprivation quarters of the population using the 1991 census population as denominator for all of the data from 1985-91 might be criticised. However, there are no intercensal estimates made at that population level. As our data refer to 1985-91 we used the closest census population, that of 1991.

Implications and conclusion

Tackling inequalities in health has only recently received government emphasis.40 Reductions in mortality from coronary heart disease principally reflect reduction in risk factors in a population.41 Socioeconomic variation in health and disease has been recognised for hundreds of years.42 Action to reduce the variation has many years of inaction at all levels to redress.

Disease registers are valuable datasets for exploring variation in diseases in a population. Examination of subgroups, such as patients admitted to a coronary care unit or a trial, will always give a biased picture when 34% of people never reach hospital alive and two thirds of deaths occur before hospital admission. Those allocating scarce healthcare resources should therefore consider socioeconomic variation not only in community death rates but also at other points in the disease process such as the chance of dying outside hospital and therefore of reaching hospital care alive.

We have shown that the greatest socioeconomic variation in death is during the prehospital phase of myocardial infarction and coronary death. Thus treatments applied equitably in hospitals across socioeconomic groups during acute myocardial infarction will have little impact on socioeconomic variation in death rates. If the deaths outside hospital were not inevitable and patients reached hospital and received hospital treatment then acute care could have an impact on the socioeconomic variation in coronary mortality.

Further investigation is required to understand the reasons for the differences in prehospital mortality in terms of different patterns of accessing care, differences between socioeconomic groups in treatment before the attack, and socioeconomic gradients in disease severity before strategies can be devised to address this aspect of socioeconomic variation in myocardial infarction and coronary death.


We thank K Barrett, C Brown, C Bauwens, H Bilkhu, C Bowman, B Fitzpatrick, J Graham, M Hastings, M Irving, E Kesson, W Leslie, M-K McCluskey, W Millar, M Mitchell, J Palmer, M Robb, M Sharkey, M Shewry, M Thornton, W Tunstall-Pedoe, A Urie, and G Watt for their contribution to establishing the register, compiling the manual of operations, and collecting, coding, managing, and checking and verifying the data. We would be unable to maintain the register without the unfailing goodwill of the general practitioners of north Glasgow. We are also grateful for the support of ISD, the Information and Statistics Division of the Common Services Agency; hospital records officers and their staff in Glasgow Royal Infirmary and University NHS Trust, West Glasgow Hospitals and University NHS Trust, and Stobhill Hospital Trust; and other records departments throughout the United Kingdom. We thank the staff of the deaths unit of the Office of the Procurator Fiscal, Glasgow, the staff of offices of other procurators fiscal throughout Scotland, and staff of coroner's offices in England for their willing cooperation. Although we are unable to acknowledge them individually, many other people and agencies have generously supported the work of the Glasgow MONICA project register. The views expressed in this paper are ours alone and do not necessarily reflect those of the funding body or of those acknowledged above as previous or current members of staff.


  • Funding The Scottish MONICA project was funded by grants from the Chief Scientist Office of the Scottish Office Home and Health Department.

  • Conflict of interest None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.