Thursday November 3 2011
Diet is just one factor that affects rates of death in the UK
“Thousands of Welsh, Scottish and Irish lives could be saved” by following the average English diet, reported the Daily Mail. The newspaper also said that “as many as 80% of preventable deaths from the biggest killer diseases would be eliminated if the rest of the UK followed England’s nutritional habits.”
Scotland, Wales and Northern Ireland are known to have higher rates of death from cardiovascular disease and cancer than England. This news story is based on a study that estimated how many excess deaths from diet-related illnesses could be prevented by changing eating habits to those of England. Using survey data, researchers estimated the average quantity of 10 different types of foods that people in each country ate. They also collected data on the number of deaths due to cardiovascular disease and certain cancers in England, Wales, Scotland and Northern Ireland. Entering this data into a statistical model, they estimated how many of these deaths could have been prevented or delayed if all the countries ate the average English diet.
This study estimated the potential impact of changing dietary habits. It supports the general health message that a balanced diet, high in fruit and vegetables, can reduce overweight, obesity and associated chronic health problems. The model is based on estimates, so the study’s findings about the impact of making these alterations should be interpreted cautiously. Also, as the researchers highlight, diet does not provide the whole answer and much of the difference in death rates between the countries is likely to be due to other non-dietary risk factors. Factors such as smoking, alcohol consumption and exercise may account for some of the differences in mortality among the countries.
Where did the story come from?
The study was carried out by researchers from Oxford University and the John Radcliffe Hospital, and was funded by the British Heart Foundation. It was published in the peer-reviewed journal BMJ Open.
The media’s reporting of this research was generally accurate, although few stories pointed out that the research was based on modelling techniques. The Daily Mirror’s headline – “English diet saves thousands of lives – in theory” - does indicate the theoretical nature of the results.
What kind of research was this?
This research used data from previous cohort and case-control studies to generate a model that would estimate the impact on mortality rates of altering the average diet in Scotland, Northern Ireland and Wales. The researchers say that there is geographical variation across the UK in the number of cases and deaths due to coronary heart disease, strokes and certain cancers. Deaths due to these causes are often referred to as avoidable mortality, as they can be prevented through timely and appropriate changes to lifestyle habits or through medical intervention.
Models can be a useful way of combining data to estimate the impact that modifying risk factors may have on a given outcome. However, they can only provide a simplified picture of the development of illness, and cannot account for all relevant variables. As such, their results should be interpreted cautiously.
What did the research involve?
Researchers used previous cohort and cross-sectional studies to generate a model which represented the impact of diet on the relative risk of developing coronary heart disease (CHD), stroke and certain cancers. They aimed to use this model to determine whether differences in eating habits accounted for the variation in mortality due to these causes across the UK. The researchers refer to this variation as a “mortality gap”.
The model, called DIETRON, quantifies the change in death due to CHD, stroke and 10 diet-related cancers at a population level. The cancers included in the model were mouth, oesophagus (throat), stomach, lung, colon (large bowel), gallbladder, pancreas, breast, endometrial (lining of the uterus) and kidney cancers.
The researchers used data from the annual Family Food Survey to define the average energy and nutritional intake in England, Wales, Scotland and Northern Ireland. They included 10 variables from this survey in their model:
- total energy intake (kcal per day)
- fruit (grams per week)
- vegetables, excluding potatoes (grams per week)
- salt (grams per day)
- total fat (% of total energy)
- saturated fat (% of total energy)
- polyunsaturated fat (% of total energy)
- monounsaturated fat (% of total energy)
- dietary cholesterol (% of total energy)
- non-starch fibre (grams per day)
Data from national registries were used to determine age- and gender-adjusted mortality due to CHD, stroke and diet-related cancer. The researchers defined the mortality gap between England, Wales, Scotland and Northern Ireland as the difference in the observed number of deaths in these countries and the expected number of deaths if each had mortality rates equal to England’s (i.e. the excess deaths in these countries).
Entering the dietary and mortality data into the model provided estimates of the annual number of deaths that could be delayed or averted in Wales, Scotland and Northern Ireland if those countries followed a similar diet as the English. The researchers divided this number by the previously defined mortality gap to calculate what percentage of the gap could be “closed” by aligning dietary habits with those seen in England.
There was uncertainty surrounding the data used to build the model. To account for this, the researchers ran an analysis that included both the relative risks and associated 95% confidence intervals from observational studies, and estimated the uncertainty surrounding the model’s results.
What were the basic results?
When analysing differences in average national diet, the researchers found that:
- Scotland and Northern Ireland had, on average, a poorer diet than England, as indicated by higher saturated fat and salt consumption, and lower fruit and vegetable consumption.
- Wales had fewer differences with the English diet than Scotland or Northern Ireland, and the average Welsh diet contained more vegetables, saturated fat and salt than the English diet.
When analysing mortality data from the four countries, the researchers found that the mortality gaps (excess deaths) for CHD, stroke and diet-related cancers between 2007 and 2009 were:
- 15,719 deaths for Scotland
- 2,329 deaths for Northern Ireland
- 3,723 deaths for Wales
The model estimated that changing to an English diet would result in:
- a 40% reduction in the mortality gap for Scotland (95% confidence interval [CI] 33% to 51%), equivalent to avoiding 6,353 deaths from CHD, stroke and diet-related cancers
- an 81% reduction in the mortality gap for Northern Ireland (95% CI 67% to 99%), equivalent to avoiding 1,890 deaths from CHD, stroke and diet-related cancers
- an 81% reduction in the mortality gap for Wales (95% CI 62% to 108%), equivalent to preventing or delaying 3,005 deaths
For deaths from CHD alone, the model estimated that changing to an English diet would lead to:
- a 58% reduction in the mortality gap for Scotland (95% CI 47% to 72%)
- an 88% reduction in the mortality gap for Northern Ireland (95% CI 70% to 111%)
- an 88% reduction in the mortality gap for Wales (95% CI 69% to 118%)
The dietary factors found to be most associated with the mortality gap in Scotland, Northern Ireland and Wales were total energy intake and the amount of fruits and vegetables eaten.
How did the researchers interpret the results?
The researchers concluded that diet substantially affects national differences in deaths due to CHD, strokes and diet-related cancers. They say that “improving the average diet in Wales, Scotland and Northern Ireland could have a substantial impact” on reducing the differences in the number of deaths from chronic diseases between these countries and England, especially deaths due to coronary heart disease.
This study suggests that improving the average diet in Wales, Scotland and Northern Ireland could reduce variations in disease mortality rates across the UK. This modelling study was based on observational data, so the results should be interpreted cautiously and considered as theoretical estimates only.
The study had several limitations, stemming from weaknesses in modelling as well as limitations to the underlying observational research:
- Models rely on theoretical scenarios, and can only estimate how diseases occur and progress in the real world. Multiple factors contribute to the development of the illnesses examined here, and diet is only one of them. Smoking, drinking alcohol, exercise habits and genetics are all risk factors for cardiovascular disease, strokes and certain cancers. The researchers point out that much of the difference between Scotland and England in terms of deaths due to these illnesses is explained by risk factors other than diet.
- Only 10 risk factors, all dietary, and 10 causes of death were included in the model. This simplifies the complex relationship between diet and mortality.
- The uncertainty analysis attempted to account for uncertainty in the relative risks used to construct the model, but did not account for the uncertainty surrounding dietary estimates derived from the Family Food Survey. The researchers say that this may have led them to underestimate the uncertainty surrounding the model’s estimates.
- The data entered into the model were based on previous observational studies. The researchers say that it is not possible to completely account for factors in these original studies that may have influenced the model’s estimates (confounding factors).
Despite the limitations of building models based on observational studies, the results indicate that small, achievable changes to the average diet in the UK could reduce the number of deaths due to CHD, strokes and certain cancers. This could close the gap in mortality rates between England and the rest of the UK. The researchers say that the biggest impact would come from reducing the amount of saturated fat consumed, and increasing the amount of fruits and vegetables eaten. They also suggest that financial incentives could be offered for changing dietary habits.
Risk factors such as smoking, alcohol and exercise may account for some of the differences in mortality seen among the countries.