Acute sickness
An illness or injury which caused the informant to cut down on any of the things he or she usually does about the house, at work or school or in his or her free time (in the two weeks prior to the interview).
Age standardisation
When proportions are compared across different sub-groups in respect of a variable on which age has an important influence, any differences in age distributions between sub-groups are likely to affect the observed differences in the proportions of interest. The objective of the direct age- standardisation procedure used in this report was to enable proportions to be presented across sub-groups after adjusting for the effect of age. Direct standardisation estimates the values of the proportions of interest where the compared sub-groups have been adjusted to the same age distribution. However, it should be stressed that age-standardised proportions provide only a summary reflecting the average relationship between the variables across all age bands and that age-standardisation adjusts only for age and not for other factors that may affect the variable of interest.
Age-standardisation was carried out (separately for men and women) by ten-year age groups. The age distribution of sub-groups was adjusted to the overall (or average) weighted distribution by age.
The age-standardised proportion p' was calculated as follows, where pi is the age specific proportion in age group i and Ni is the standard population size in age group i:
Therefore p' can be viewed as a weighted mean of pi using the weights Ni. The age groups were: 16-24, 25-34, 35-44, 45-54, 55-64.
Alcohol consumption
See Volume I, Chapter 5.
Analysis of variance
One-way analysis of variance is a statistical technique for testing whether the means of sub-groups of a population differ significantly from one another.
Angina
See Volume I, Chapter 10.
Anthropometric measurements
For body mass index (BMI) and waist-hip ratio definitions see Volume I, Chapter 7.
Blood pressure
See Volume I, Chapter 6.
Blood analytes
For total cholesterol, HDL-cholesterol, fibrinogen, ferritin, haemoglobin, vitamins A, C and E and carotenoids, see Volume I, Chapter 9. For cotinine, see Volume I, Chapter 4. For gamma gt, see Volume I, Chapter 5.
Body Mass Index
See Volume I, Chapter 7.
Breathlessness
See Volume I, Chapter 8.
Cardiovascular disease and related conditions
See Volume I, Chapter 10.
Cardiovascular disorder
Those classified as having any CVD disorder are those who said they had ever been diagnosed by a doctor as having had angina, heart attack, stroke, heart murmur, abnormal heart rhythm, other heart trouble (hypertension or diabetes). Women who have hypertension, heart murmur or diabetes only when pregnant are not classified as having any CVD disorder.
Cholesterol
See Volume I, Chapter 9.
Cigarette smoking
See Volume I, Chapter 4.
Claudication
See Volume 1, Chapter 10.
Cotinine
See Volume I, Chapter 4.
Ferritin
See Volume I, Chapter 9.
Fibrinogen
See Volume I, Chapter 9.
Gamma gt
See Volume I, Chapter 5.
Gastroenteritis
See Volume I, Chapter 11.
Geometric mean
The geometric mean is a measure of central tendency. It is sometimes preferable to the arithmetic mean, since it is less influenced by very large outliers in the distribution. The geometric mean of a continuous variable is calculated as:
GHQ12
See Volume I, Chapter 12.
Haemoglobin
See Volume I, Chapter 9.
High blood pressure
Informants were classified in one of four groups based on their Systolic (SBP) and Diastolic Blood Pressure (DBP) readings and current use of anti-hypertensive medication. For a definition of high blood pressure (informants who were normotensive treated, hypertensive treated, or hypertensive untreated) see Volume I, Chapter 6.
Ischaemic heart disease
Ischaemic heart disease includes those who reported previous diagnosis of heart attack or angina.
Linear regression
Linear regression was used to investigate the linear association of two or more factors (‘independent’ or ‘predictor’ variables) with a continuous variable (‘dependent’ or ‘outcome’ variable), such as blood pressure. The independent variables can be continuous or categorical (grouped) variables. The parameter estimates for a particular variable from a linear regression model give an estimate of the association of that variable with the outcome variable, adjusted for all other variables in the model. For example, linear regression was used to assess the association of social class with systolic blood pressure, after adjusting for age (see Table 6.16).
For a continuous independent variable, the regression coefficient is the change that is predicated in the mean of the outcome variable for a one unit change in the independent variable, adjusted for all other variables in the model.
Parameter estimates for categorical independent variables have been presented in two ways. The standard method defined one category of a categorical independent variable as a baseline or reference category and compared all other categories to this reference category. Therefore there is no parameter estimate for the reference category and estimates for all other categories give the predicted mean difference in the outcome variable between each category and the reference category, adjusted for all other variables in the model. In terms of the above example, the parameter estimate for Social Class IIINM would give the difference in mean systolic blood pressure between respondents in Social Class IIINM and the reference category (usually Social Class I), after adjusting for age. An alternative method of presenting results for categorical independent variables was used for region where there was no obvious reference category. In this method an estimate for a given category of a categorical independent variable gives the deviation in the mean of the independent variable for that category compared to the overall mean.
The statistical significance of independent variables in models was assessed by the F-ratio and its associated p value. 95% confidence intervals were also calculated for parameter estimates. These can be interpreted as meaning there is a 95% chance that the given interval for the sample will contain the true population parameter of interest. In linear regression a 95% confidence interval which does not include zero indicates that the given parameter estimate is statistically significant.
Reference: Weisberg, S. Applied linear regression. John Wiley & Sons, New York, 1985.
Logistic regression
Logistic regression was used to investigate the association of two or more independent or predictor variables with a two-category (binary) outcome variable. The independent variables can be continuous or categorical (grouped) variables. The parameter estimates from a logistic regression model for each independent variable give an estimate of the association of that variable with the outcome variable, adjusted for all other independent variables in the model. For example, logistic regression was used to assess the association of social class with CVD condition, after adjusting for age.
Logistic regression models the log ‘odds’ of a binary outcome variable. The ‘odds’ of an outcome is calculated as the probability of its occurring divided by the probability of its not occurring. The parameter estimates obtained from a logistic regression model have been presented as odds ratios for ease of interpretation.
For continuous independent variables, the parameter estimate gives the change in the odds of the outcome occurring for a one unit change in the value of the predictor variable.
Parameter estimates for categorical independent variables have been presented in two ways. The standard method defines one category of the categorical variable as a baseline or reference category and compares all other categories to this reference category. Therefore there is no parameter estimate for the reference category and odds ratios for all other categories are the ratio of the odds of the outcome occurring between each category and the reference category, adjusted for all other variables in the model. In terms of the above example, the odds ratio estimate for Social Class IIINM can be interpreted as the likelihood of having a CVD condition for respondents in Social Class IIINM compared to those in the reference category (usually Social Class I), after adjusting for age. An alternative method of presenting results for categorical independent variables was used for region where there was no obvious reference category. In this method the odds ratios for a given category of a categorical independent variable gives the change in the odds of the outcome occurring compared to the overall odds.
The statistical significance of independent variables in models was assessed by the likelihood ratio test and its associated p value. 95% confidence intervals were also calculated for the odds ratios. These can be interpreted as meaning that there is a 95% chance that the given interval for the sample will contain the true population parameter of interest. In logistic regression, a 95% confidence interval which does not include one indicates the given parameter estimate is statistically significant.
References: Norusis MJ. SPSS for Windows: Advanced statistics release 6.0. SPSS Inc, Chicago, 1993. Hosmer DW Jr. and Lemeshow. Applied logistic regression. John Wiley & Sons, New York, 1989.
Lung function
For definitions of the three measures of respiratory function (FEV1, FVC and PEF), see Volume I, Chapter 8.
Mean
Unless otherwise stated, the ‘means’ presented in the report are arithmetic means: the sum of the values for all cases divided by the number of cases. See also ‘geometric mean’.
Median
The central value of an ordered set of observations which divides the set into two equal parts such that half the cases have values below the median and half the cases have values above the median.
MRC Respiratory
Questionnaire
See Volume I, Chapter 8 for definitions relating to the measures obtained from the MRC Respiratory Questionnaire (breathlessness, phlegm and wheeze).
Multiple linear
regression
See linear regression.
Percentile
The value below which a specified percentage of values in an ordered set of observations fall. For example, the 20th percentile is the value below which 20 percent of the cases lie, while 80 percent lie above. The 50th percentile is the median.
Physical activity
For physical activity classification and summary measures (maximum intensity level and frequency-intensity activity level) see Volume I, Chapter 2.
p value
A p value is the probability that a difference or association as or more extreme than that observed would arise if there is no actual difference or association in the population. A p value of less than 5% is conventionally taken to indicate a statistically significant result (p<0.05). It should be noted that the p value is dependent on the sample size, so that with large samples differences or associations which are very small may still be statistically significant. Results should therefore be assessed on the magnitude of the differences or associations as well as on the p value itself. The p values given in this report are based on the assumption of a simple random sample and do not take into account the complex sampling design of the survey.
Quintile
Quintiles are percentiles which divide an ordered set of observations into fifths, i.e., the 20th, 40th, 60th and 80th percentiles.
Region
Regional analyses are based on the 15 Health Boards grouped in the following seven regions: Highland & Islands; Grampian & Tayside; Lothian & Fife; Borders, Dumfries & Galloway; Greater Glasgow; Lanarkshire, Ayrshire & Arran; and Forth Valley, Argyll & Clyde.
Rose angina
See Volume I, Chapter 10.
Social class
Social class of chief income earner
Social class was assigned on the basis of the occupation of the ‘chief income earner’ within the informant’s household. ‘Chief income earner’ was defined as the person within the household with the largest income, whether from employment, pensions, state benefits or any other source. Social class was based on the Registrar General's Standard Occupational Classification (Vol 3. OPCS, London: HMSO 1991). Occupations are assigned to six social class categories:
Social Class
Occupations
I
Professional occupations
II
Managerial and technical occupations
III
Skilled occupations
(IIINM) non-manual
(IIIM) manual
IV
Partly skilled occupations
V
Unskilled occupations
In some analyses, Social Classes I and II and Social Classes IV and V have been combined. In others, I, II and IIINM have been combined under the heading of ‘non-manual’, while IIIM, IV and V have been combined under the heading of ‘manual’
Where the chief income earner was not the informant, the social class of the chief income earner was derived from information obtained from the informant about the chief income earner’s occupation. Chief income earners who were in the armed forces, whose occupation was not adequately described or who were full-time students were not allocated a social class and are not shown separately in the tables. They are, however, included in the total column.
Own social class
In some analyses, social class is based on the occupation of the informant rather than on the occupation of the chief income earner within the informant's household.
Waist-hip ratio
See Volume I, Chapter 7.