Failure to validate the Health Survey for England physical activity module in a cardiac population

Abstract

Objectives

The Health Survey for England physical activity module interview, although not validated, is used as a performance indicator to see if people are achieving current physical activity targets and is the primary source of information for physical activity related policy making. The aim of this study was to assess the validity of the Health Survey for England physical activity module as a measure of physical activity in older adults with coronary heart disease.

Methods

Seventy-two older adults who had experienced a cardiac event completed the Health Survey for England physical activity interview and wore an accelerometer for seven consecutive days. Physical activity classification levels were derived from accelerometry and from the Health Survey for England physical activity interview, together with the number of episodes in which participants were moderately active for 30 min or more.

Results

The Health Survey for England physical activity interview exhibited high sensitivity (1.0) and specificity (0.76) for people engaged in high levels of physical activity, moderate sensitivity (0.40) and specificity (0.56) for people engaged in medium levels of physical activity and low sensitivity (0.35) and high specificity (0.92) for people engaged in low levels of physical activity. Compared with the accelerometer, the survey misclassified 63% of participants.

Conclusions

The Health Survey for England physical activity interview misclassified true activity levels in older adults with heart disease by overestimating actual activity in the less active participants. Similar biases on self-report physical activity measures have been demonstrated in the general population, suggesting that the Health Survey for England physical activity module interview may be providing inaccurate information on national activity levels.

Introduction

The Health Survey for England (HSE) is an annual, cross-sectional population interview-based survey which provides information on various aspects of the nation's health and health indicators including physical activity [1]. The HSE is a major source of data for assessing the prevalence of physical activity in England as important health policy documents and strategic targets are based on estimates from recent surveys [2], [3]. The questions used in the HSE physical activity module are derived from those originally used in the Allied Dunbar National Fitness Survey (1992) [4]. Surprisingly, neither survey has been validated. Amongst other information the HSE reports the percentage of the adult population achieving the current national recommendations of a minimum of 30 min of moderate intensity activity on at least 5 days per week [1]. The latest figures available reported that 37% of men and 24% of women met this criterion, but that this declined with age with 8% of men aged >75 years and 3% of women aged >75 years meeting the target activity guidelines [5].

Accurate measurement of physical activity by self-report is difficult and limited as factors such as social desirability, age, complexity of questionnaire, seasonal variation and length of period surveyed can influence responses to questions [6]. Direct measurement of physical activity by accelerometry may assist in reducing the limitations of self-report questionnaires. Free-living physical activity can be assessed on a minute-by-minute basis using accelerometry. Accelerometry provides a portable, simple, affordable (for small scale studies), objective and socially acceptable method of measuring physical activity [7].

As part of a study to validate an activity questionnaire for people with cardiac illness we administered both the Health Survey for England physical activity interview and accelerometry to participants who had a previous history of an acute cardiac event. This provided data that afforded the first evaluation of the validity of the Health Survey for England physical activity module, albeit in an older population with a chronic illness. To date, no studies have investigated the validity of the HSE for assessing physical activity levels in an elderly and chronic disease population, such as coronary heart disease (CHD). The aim of this study was to assess the agreement between physical activity classifications of ‘high’, ‘medium’, or ‘low’ as measured on the HSE compared with physical activity classification derived from accelerometry data in older adults with CHD.