[Previous] [Contents] [Next]

The Scottish Health Survey 1998: Volume 1: Chapter 1

1 INTRODUCTION

Andrew Shaw

1.1 THE 1998 SCOTTISH HEALTH SURVEY

The 1998 Scottish Health Survey is the second of a series of surveys designed to make a major contribution to monitoring health in Scotland. The first survey took place in 1995¹, and the next survey is expected to be in 2001. Both the 1995 and 1998 surveys were commissioned by The Scottish Office Department of Health, now the Scottish Executive Health Department. They have been conducted by the Joint Health Surveys Unit of the National Centre for Social Research (formerly SCPR) and the Department of Epidemiology and Public Health at University College London (UCL).

People living in Scotland experience relatively high rates of mortality from major diseases compared with the populations of England and many other western countries. Deaths and ill-health from coronary heart disease, cancer and stroke are higher than elsewhere. However, statistics on the numbers and causes of death do not provide explanations for these differences. Other statistics, such as hospital admissions, are derived from people's contacts with the National Health Service, so tend to be concerned with limited, albeit more serious, manifestations of ill-health. Before the Scottish Health Survey there was no comprehensive picture of the health of the whole population, its biological characteristics or health-related behaviour. Nor, therefore, was reliable information available on trends in health and related behaviour.

The Scottish Health Survey was designed to overcome this lack of knowledge. The specific aims of the survey are:

1 to estimate the prevalence of particular health conditions in Scotland;

2 to estimate the prevalence of certain risk factors associated with these health conditions and to document the pattern of related health behaviours;

3 to look at differences between regions and between subgroups of the population in the extent of their having these particular health conditions or risk factors, and to enable comparisons with other national statistics and with people in England;

4 to monitor trends in the population's health over time; and

5 to make a major contribution to monitoring progress towards the Scottish health and dietary targets.

The White Paper Towards a Healthier Scotland², published in 1999, sets out a co-ordinated three level approach to better health -involving improving life circumstances, changing lifestyles and addressing priority health topics - with the overall aim of tackling health inequalities.3

Smoking, poor diet, lack of physical activity and alcohol misuse are among the key lifestyle behaviours identified in the White Paper as contributing towards poor health and early death. The Health Survey asks informants detailed questions about each of these.

Children's health, dental health, coronary heart disease and accidents are among the topics identified in the White Paper as priorities for action. Again, the Health Survey asks a series of relevant questions.

The Scottish Health Survey is used to measure progress towards a number of the White Paper's targets. However, the data on prominent concerns of public health policy represent a small fraction of the portrait of health in Scotland provided by the Health Survey. The purpose of this report is to sketch out this portrait; even though we present several hundred pages of tables and commentary, more detail about each topic awaits further analyses of this publicly-available data set.

1.2 OVERVIEW OF THE SURVEY DESIGN

The Scottish Health Survey consists of a number of core questions and measurements (such as height and weight), plus modules of questions on selected topics which could change from one survey to the next. In this respect, it is modelled on the annual Health Survey for England, which started in 1991. In practice, there was limited scope to change topics in the 1998 Scottish survey, since updated information was required for nearly all topics covered in 1995. Thus there continued to be special emphasis on cardiovascular disease and its associated risk factors. General health, asthma, accidents and dental health were among the accompanying modules.

The main innovation in 1998 was the introduction to the sample of children aged 2-15. In addition, the upper age limit for adults was extended from 64 to 74. As in 1995, approval was obtained from the Research Ethics Committees for all Area Health Boards in Scotland.

1.2.1 Sample design

The survey was designed to provide a nationally representative sample of the population of Scotland aged between 2 and 74 living in private households. Residents living in institutions (who tend to be older than people in private households) were excluded from the survey.

The survey is based on 'a stratified, multi-stage random sample'. This design involved first ordering all postcode sectors in Scotland by region and the Carstairs index of deprivation and then systematically selecting 312 of them. Within these sampling points, 15,288 addresses were selected from the Postcode Address File (PAF). At each residential address up to three households were selected randomly by interviewers, though selection of more than one household was rare. Within each household, one person aged 16-74 and up to two children aged 2-15 were selected randomly to be included in the survey.

One aim of the Health Survey is to compare regions within Scotland. Hence, regional sample sizes have to be sufficient for reliable comparisons to be made. It was not feasible to enlarge the total sample to interview a sufficiently large sample in every Health Board area. Therefore, for the purposes of the survey, seven 'regions' were defined by aggregating (mainly) contiguous Health Boards:

Highland & Islands (Orkney, Shetland, Western Isles)
Grampian & Tayside
Lothian & Fife
Borders, Dumfries & Galloway
Greater Glasgow
Lanarkshire, Ayrshire & Arran
Forth Valley, Argyll & Clyde

In order to provide sufficient sample sizes within each region, the two least populated regions - Highland & Islands and Borders, Dumfries & Galloway - were slightly over-sampled, while the other five regions were sampled (roughly) in proportion to their population size.

1.2.2 Fieldwork design

Each sampled address was sent a letter explaining the survey in advance of the first interviewer visit. Interviewing was conducted using Computer Assisted Personal Interviewing (CAPI). Children aged 13-15 were interviewed in the presence of at least one parent or guardian. Parents answered on behalf of younger children, who were nevertheless required to be present. In addition, those aged 8 and over were asked to complete a short paper questionnaire on more sensitive topics. There were four such questionnaires, administered according to age group (8-12, 13-15, 16-17 and 18-74). Interviewers were also responsible for measuring informants' heights and weights. Finally, they sought permission for a follow-up visit by one of the nurses specially trained to work on the survey.

The nurse collected information about use of prescribed medicines, vitamin supplements and nicotine replacements, and recent experience of gastro-enteritis. She then measured blood pressure, lung function, and waist, hip, mid-upper arm circumference and demi-span, as appropriate depending on the informant's age. With the written agreement of adult informants or a parent/guardian of children aged 11-15, small blood samples were taken by the nurse. Clearance for taking blood samples from children was not obtained in time for the start of fieldwork, and so not requested for all child informants. In a sub-sample of areas within an hour of a participating laboratory, an extra tube of blood was taken from adults to measure levels of vitamins A, C, and E and carotenoids. A saliva sample was collected from informants aged 4 and over.

Informants were told that participation at each stage of the survey, and each measurement, was voluntary and that they could choose to opt-out at any point.

The survey was designed to be conducted throughout a twelve month period, in order to allow for seasonal variations. Fieldwork began in April 1998 and the final 41 interviews were conducted in May 1999. The vast majority of interviews (96.4%) had been undertaken by the end of March 1999.

1.2.3 Survey response

Interview data was secured from 9,047 persons aged 16-74 and 3,892 children aged 2-15. Of these informants, 7,455 adults and 3,211 children were visited successfully by a nurse. Within these visits, at least one usable blood sample was taken from 6,178 adults and 466 children aged 11-15.

The total number of eligible households (those with someone aged 2-74) selected for the survey is not known exactly because no information at all was obtained from a small proportion of households. Therefore, participation rates are calculated by assuming that this group includes the same proportion of eligible households as the rest of the sample. On this basis, at least one interview was conducted at 77% of the eligible households.

The proportions of eligible adults and children who participated in various components of the survey are shown below. (Note that not all informants were asked to provide every measurement and that not all measurements and samples resulted in analysable data.) The figures for children are estimated on the basis that the average number of eligible children in non-participating households was identical to that in households which did take part. Figures for measurements among adults are also estimated on the basis that the eligible proportion among non-participants was the same as that among those who did take part.

SUMMARY OF PARTICIPATION BY ELIGIBLE SAMPLES

1.2.4 Data

The data collected in the field was subjected to an intensive editing process to produce a fully 'clean' final dataset. This covered coding of verbatim responses, including 'other' responses to pre-coded questions, as well as a series of consistency and plausibility checks on data values.

Since addresses and individuals did not all have equal chances of selection, the data have to be weighted for analysis. In addition, different rates of response by groups within the population, and random variation, may impact upon the extent to which the achieved sample represents the population aged 2-74. Hence, the weights applied to the survey data are calculated to ensure that the age and sex profile of the sample is identical to the known profile of the whole population.

As the Scottish Health Survey is a cross-sectional survey of the population at a particular point in time it can examine associations between current health and personal characteristics or types of behaviour. It cannot however determine causality, because current health may reflect past rather than present behaviour and other conditions.

The 1998 Scottish Health Survey data will be deposited at The Data Archive at the University of Essex, from where the 1995 data may also be obtained.

1.3 THIS REPORT

The 1998 Scottish Health Survey report consists of two volumes. This volume reports the survey findings (Chapters 2 to 14). The objective of each chapter is to describe the available data, to present detailed results in a set of tables and to provide a succinct and accessible commentary on these results. A summary of the main findings is provided at the beginning of each chapter.

Data for men and women are presented separately. The same is usually true for boys and girls. Survey variables are tabulated by age groups and, usually, region and social class. Other common analyses include comparisons with results from the 1995 survey and with those from the Health Survey for England. The latter includes comparisons with England as a whole, and with Northern England, mostly using 1998 data (but sometimes earlier data where that is the latest available). These analyses take account of differences between surveys in the coverage of adults. Comparisons with 1995 are limited to those aged 16-64, while adults aged 75 and over have been excluded from the data for England which is presented here. Readers are asked to note, therefore, that any further comparisons made simply by examining together the findings in this report and those presented in the 1995 report and the annual reports on England may well be invalid.

The second volume is the Technical Report, which documents the study design and the content and conduct of interviews and measurements, including those requiring blood samples.

The conventions used in the tables and reporting are described at the end of this Introduction.

1.3.1 Volume I: Findings

Chapter 2 Cardiovascular disease prevalence and risk factors

This chapter reports the prevalence of self-reported cardiovascular disease (CVD) conditions for the Scottish adult population (aged 16-74), examines CVD differences between regions and social classes, trends over time and comparisons with England. Findings on four blood analytes (total cholesterol, HDL-cholesterol, fibrinogen and C-reactive protein) associated with risk for cardiovascular disease are presented. The relationship between CVD and its risk factors such as obesity, smoking and drinking is also explored.

Chapter 3 Respiratory Symptoms

This chapter deals with asthma and asthma-related symptoms in adults and children. It reports on different respiratory symptoms: their presence, severity and impact on everyday life; use of health services; and, for adults only, variations in the prevalence of respiratory symptoms by region, degree of urbanisation, social class and cigarette smoking. Comparisons are made with results from the 1995/1996 Health Survey for England. The last section in this chapter is concerned with levels of serum total immunoglobulin E (IgE) and house dust mite specific IgE (HDM IgE) in children.

Chapter 4 Lung function

Lung function measurements, obtained from informants aged 7-74, are examined in this chapter. Three measures were collected: FEV₁ (Forced Expiratory Volume in the first second), FVC (Forced Vital Capacity) and PEF (Peak Expiratory Flow). Estimates of FEV₁, FVC and PEF are presented by age and height groups. For adults only, FEV₁ and FVC are reported by region, social class, smoking status and respiratory symptoms. Relative lung function levels are presented; FEV₁ results are compared with those from the Health Survey for England 1996.

Chapter 5 Accidents

This chapter presents findings for both children and adults on the incidence and characteristics of non-fatal accidents about which advice was sought from a doctor, nurse or other health professional. It covers the circumstances under which accidents occur and the personal characteristics of those having accidents. Accident rates are compared by social class and region within Scotland, and between Scotland and England. Data for England is for 1995/1996 for adults and 1995 to 1997 for children.

Chapter 6 Adult physical activity

Figures for physical activity among adults are presented first within four broad types of activity and then as summary measures based on all kinds of activity. The four activity types are activity at home (housework, gardening, DIY), walks of 15 minutes or more, sports and exercise activities, and activity at work. Frequency, duration and intensity of activity are all reported. Patterns in physical activity are examined by social class and region; results are compared with those for England.

Chapter 7 Child physical activity

The 1998 Scottish Health Survey collected details about the physical activity of children aged between 2 and 15 years in four main categories: sport and exercise, active play, walking and housework/gardening. This chapter reports the rates, durations and patterns of activity _ both within these four categories and in aggregate. It examines variations according to social class, region and other factors and makes comparisons between children in Scotland and England.

Chapter 8 Smoking

This chapter examines current and past smoking behaviour among those aged 8-74. It reports on daily consumption and tar levels among adult smokers and smoking status among children. Levels of the blood analyte serum cotinine are used to validate self-reports of smoking behaviour. Social class and regional variations are explored, along with the exposure of non-smokers to tobacco smoke; comparisons are made with 1995 results (for adults) and with England (1998).

Chapter 9 Drinking

This chapter explores alcohol consumption among adults and children aged 8-74. Adults' self-reported levels of weekly alcohol consumption are calculated and compared with those recorded in 1995. Problem drinking and drunkenness are then assessed. For children, experiences of drinking alcohol and levels of consumption are reported. Results are compared with those for both adults and children from the 1998 Health Survey for England. There is accompanying analysis of the blood analyte gamma gt.

Chapter 10 Eating habits

Summary information on the eating habits of adults and children was collected for a wide range of food types. This chapter describes informants' patterns of consumption of seven major food types: foods containing sugar, spreading and cooking fats, dairy produce, meat and poultry, fish, foods containing starch and fibre, and fresh fruit and vegetables. Use of salt and dietary supplements is reported and variations in eating habits by social class and region are explored. Finally, trends since 1995 (among adults) are discussed and comparisons for 1998 made with England.

Chapter 11 General health

This is a wide ranging chapter. For adults and children, it reports levels of self-assessed general health, longstanding illness or disability and acute sickness. The General Health Questionnaire (GHQ12) measure of psycho-social health is available for adults and children aged 13-15. All informants were asked about their use of GP and hospital services, as well as about their dental health. Nurses collected data on prescription medicines and gastro-enteritis from adults and children, while women informants provided information on use of contraceptive pills and hormone replacement therapy. Informants aged 13 and over reported on bladder problems and blood lead levels were measured among children aged 11-15.

Chapter 12 Blood Pressure

The purpose of this chapter is to describe blood pressure levels for adults and children aged 5 and over and the prevalence of high blood pressure for the adult Scottish population in 1998. The extent and 'success' of treatment is reported. Variations in blood pressure between regions and social classes are examined for both adults and children. Trends in adult blood pressure since 1995 are considered and comparisons are made between Scotland and England.

Chapter 13 Anthropometry

This chapter presents the second round of population-based data on anthropometric measures and iron status for adults aged 16-64 as well as new baseline data for children (2-15) and older people (65-74). It examines age-sex distributions of height, weight, body mass index (BMI), waist-hip ratio (WHR), demi-span (an alternative to height as a measure of skeletal size for adults aged 65-74), mid-upper arm circumference (in children aged 2-15) and iron status, indicated by serum haemoglobin and ferritin levels. Then, regional and social class variations in two widely used measures of obesity - BMI and WHR - are investigated. Finally, trends over time and comparisons with England are reported.

Chapter 14 Vitamins and carotenoids

This chapter reports levels among adults of vitamins A, C and E, as well as of carotenoids including b-carotene. Variations by social class, season, smoking and drinking are explored.

1.3.2 Volume II: Technical Report and Documentation

Chapter 15 Survey methodology and response

This is a detailed account of the design and conduct of the survey, including sampling and response. The calculation of weights required for data analysis is explained; the accuracy of results is discussed and illustrated through the presentation of estimated complex sampling errors and confidence intervals.

Chapter 16 Blood analytes, quality control and quality assessment

This chapter provides a description of the biochemical blood assay methods used, along with internal and external quality control procedures and results.

Appendices

Appendix A contains copies of the main fieldwork documents including the interviewer and nurse schedules. Appendix B provides the protocols used for taking blood samples and for all measurements, including height, weight, waist and hip circumferences, blood pressure, and lung function. Appendix C is a map of the areas covered by Health Boards in Scotland and Appendix D is a Glossary.

1.3.3 Notes to tables and analysis

1 The following conventions have been used within tables:

2 The row or column percentages may not add to 100% because of rounding or, of course, when more than one answer could be given to a question.

3 'Missing values' occur for several reasons, including: refusal or inability to respond; refusal to co-operate with an entire section of the survey (such as the nurse visit or self-completion questionnaire); and questions which were not applicable to informants. Missing values have been omitted from tables and analyses. However, those with missing data for a classification variable, such as social class, are included in Total columns in tables (unless, of course, data is also missing for the other variable of interest).

4 The tables show both the weighted and unweighted bases. The weighted bases are shown in order to reflect the true relative sizes of the groups analysed. The unweighted bases are shown because the accuracy of the results depends on their size. Note that tables comparing adults in Scotland and England show weighted bases only for Scotland, since weights are not applied to adult informants in the Health Survey for England.

5 The text of this report aims to draw out the most salient findings on the topics covered in the survey. It is not designed to be a comprehensive account of the data collected or, even, of over 300 tables presented here.

6 Tables comparing Scotland with England use 1998 data for both countries except where earlier years are indicated for England: Chapters 3, 4, 5 and, for children, 10, use data variously from 1995 to 1997.

7 In general, differences commented upon in the text are statistically significant, at the 95% level or higher. Note, however, that the standard tests used to assess significance do not take account of survey design factors which tend to increase the standard errors of estimates. Estimates of design factors and true standard errors for key variables may be found in Chapter 15. Note, too, that the presence of and commentary upon a clear trend within the data, for example increasing prevalence by age group, does not imply that figures for each adjacent age group differ significantly. Also, noteworthy differences may have been confirmed by combining classification categories in ways not shown in tables.

8 The standard errors which are presented in Volume I have been calculated by formulae for simple random samples which utilise weighted bases. They should be interpreted cautiously, given that more precise estimates of standard errors would account for stratification, clustering and weighting of the sample. As mentioned, some such estimates may be found in Chapter 15.

9 Various analysis techniques, including age standardisation and logistic regression, have been used in this report. Descriptions of these and other technical or medical terms may be found in the Glossary in Volume II.

[Previous] [Contents] [Next]