Abstract
Objectives: To report current levels of obesity and associated cardiac risk using routinely collected primary care computer data. Methods: 67 practices took part in an educational intervention to improve computer data quality and care in cardiovascular disease. Data were extracted from 435 102 general practice computer records. 64.3% (229 108/362 861) of people age 15 y and older had a body mass index (BMI) recording or a valid height and weight record that enabled BMI to be derived. Data about cardiovascular disease and risk factors were also extracted. The prevalence of disease and the control of risk factors in the overweight and obese population were compared with those of normal body weight. Results: 56.8% of men and 69.3% of women aged over 15 y had a BMI record. 22% of men and 32.3% of women aged 15 to 24 y were overweight or obese; rising each decade to a peak of 65.6% of men and 57.5% of women aged 55 to 64 y. Thereafter, the proportion who were overweight or obese declined. The prevalence of ischaemic heart disease, diabetes mellitus and hypertension rose with increasing levels of obesity; their prevalence in those who are moderately obese was between two and three times that of the general population. Systolic and diastolic blood pressure, blood glucose even in non-diabetics, cholesterol and triglycerides were all elevated in the overweight and obese population.
Conclusion: Based on the recorded data over half of men and nearly half of women are overweight or obese. They have increased cardiovascular risk, which is not adequately controlled by current practice.
Introduction
Obesity in England is thought to be expanding at a faster rate than in other parts of Europe, and was estimated in 1998 to have cost the health service at least £½ billion per year in direct costs, about 1.5% of total NHS expenditure; with a further £2 billion in indirect costs Citation[1]. Obesity is defined as a condition where excessive adipose tissue is present to the extent that health may be impaired Citation[2]. Although it has been argued that waist circumference may better reflect adipose tissue deposits Citation[3], body mass index (BMI, weight in kilograms divided by height in metres squared) is the most used, and useful, to define obesity Citation[4]. Overweight or “pre-obesity” is defined as a BMI of 25 to 29.9 kg/m2 and obesity as 30 kg/m2 or over Citation[2]. Obesity is subdivided into “moderate” or class 1, where BMI is 30 to 34.9 kg/m2, “severe” or class 2, where BMI is 35 to 39.9 kg/m2, and “morbid”, where BMI is 40 kg/m2 or more Citation[2], Citation[5]. Obesity is associated with non-insulin-dependent diabetes, cardiovascular disease and hypertension; it is also associated with other non-cardiovascular diseases, though these were not examined in this study Citation[6]. BMI peaks in the sixth or seventh decade, and it has been speculated that obese people with co-morbidities or complications do not reach very old age Citation[7].
The Primary Care Data Quality Programme (PCDQ) is an educational intervention which improves general practice data quality through a series of targeted disease-focused programmes. The largest of these is its cardiovascular programme Citation[8]. One question that often arises is whether an association between obesity and cardiovascular risk is apparent in routinely collected general practice data; what proportion of those with cardiovascular disease are obese; and how well are their risk factors managed. We therefore re-analysed the data from pilot practices visited between November 2003 and February 2004, to provide baseline data against which practices could compare their own data.
Methods
This study was carried out in 67 general practices, with a total registered population of 435 102 (mean practice size 6494, range 1890–14 701). They were all participants in the PCDQ programme. The PCDQ programme works with individual practices and localities, extracting anonymized computer data from participating practices. English general practice is highly computerized, so data are readily available. The data are aggregated, analysed and fed back. Practitioners draw their own conclusions about where data quality might be improved, and where there is scope to improve data quality and the quality of care. One of the principal learning opportunities created is the Data Quality Workshop, which ideally runs as a half-day session. After an initial plenary, practitioners work in small groups to decide what action might be taken as a result of the data presented.
The practices were drawn from existing PCDQ programmes in three English regions, with the intention of forming an overview of the national prevalence. Thirty-two practices (43% of the sample) were in London (average list size of 5830), 21 (32% of sample) were located in the south east (average list 6641) and 14 (25% of the sample) from the north of England (average list 7792). The male-to-female ratio was 49.5:50.5, a slightly higher proportion of men compared to the national average of 48.7:51.3 Citation[9]. There were also more people in the first half of their working life than the national average: 1.17 times the number of 35- to 45-y-old males, and 1.10 times more females. There is a corresponding reduction in the number of children aged 0–15 y. A population pyramid shows how the study population compared with the English population ().
The data were collected between November 2003 and March 2004. We extracted the latest data recorded prior to the collection date. We did this with the written consent of the practices. The practices were participating in the PCDQ programme Citation[8], and had volunteered to test an enhanced version of the programme. Generally, localities took part in the programme because they wished to improve the cholesterol management as well as the general cardiovascular health of their population with coronary heart disease Citation[8]. These practices did not have a particular interest in obesity, and joined the programme on a locality basis. Enhancements were being made because the existing programme did not cover all the cardiovascular indicators in the new general practice contract Citation[10], and participants were keen to receive more feedback about their rate of progress towards its financially incentivized targets. During this period, a data collector visited each practice and collected information on demography, BMI and height and weight recording, cardiovascular diagnosis, and risk factors from their computer systems using the health query language MIQUEST (Morbidity Information and Export Syntax, Box 1) Citation[11]. The use of the MIQUEST query tool allowed each practice's data to be extracted in a consistent fashion. We collected clinical data which had been recorded as structured or coded data within the practice computer system. We did not collect narrative text. The data were aggregated in a customized Microsoft Access database to produce a single file for analysis purposes.
Box 1. Dataset.
Only anonymized data were collected from the practices using MIQUEST. The data extracted consisted of the date that the data were recorded and the code used, usually a read code, e.g. “12/2/2004, 1371” means that the “never smoked tobacco” code was recorded for that patient on 12 February 2004. Where there was an associated numeric date, code and numeral were extracted, e.g. “12/2/2004, 44p6, 3.1” means that 3.1 mmol/l LDL cholesterol was recorded on 12 February 2004.
The data had to be extensively cleaned and processed Citation[12]. Data reliability was tested by careful examination of the population denominator, the diagnostic and other clinical data due to be used in the study Citation[13], Citation[14]. We adopted a pragmatic approach to this study and included practices with a recorded prevalence of coronary heart disease of over 2%, who also had at least 50% of patients with a cholesterol measure of whom at least 50% were prescribed a statin. We knew from experience gained in a 2.4 million-person study that a combination of these three indicators suggested that they were engaged in diagnostic coding, electronic recording of pathology test results and computerized prescribing Citation[15].
We calculated the prevalence of overweight and obese people based on their BMI or, where this was absent, from their height and weight. We standardized the prevalence using the European standard population, because this has now become the standard comparator used within the UK health service Citation[16]. We also looked at the prevalence of cardiovascular co-morbidities and risk factors by gender and age. We analysed in 10-y bands from age 15 y upwards, grouping together the population over 85 y.
We then formulated audit criteria (Box 2) based upon best practice guidance Citation[17], Citation[18], in conjunction with representatives of the practices involved. Their purpose was to see what proportion of the population we could calculate BMI for and look at their level of cardiovascular disease and risk. Whether or not there was a causal link, practitioners wanted to know how much more likely overweight and obese people were to have cardiovascular disease and risk factors. It was agreed that control of blood pressure and lipid management in the overweight and obese would be compared using the standards set out in the UK National Service Framework for Coronary Heart Disease Citation[19]. The proportion of patients “above target” was to be identified, irrespective of being on treatment or not. We also extracted blood glucose information; data in practice computer systems are not flagged to indicate whether they were random or fasting samples.
Box 2. Audit criteria.
Data were analysed in SPSS version 12. An independent samples t-test was used to test whether the difference in the means between normal and overweight and obese for continuous variables was statistically significant and the Pearson χ2 test to compare the distribution of categorical variables. Mean and standard deviation (SD) are used to describe normally distributed variables, and the standard error of the mean (SEM) and confidence intervals (CI) to describe the precision of the population parameters.
Results
BMI was extensively recorded in the adult computerized medical record. Nearly three-quarters of women between 25 and 85 y had a recorded BMI. There were lower levels of data recording for men (56.5% of over 15 y, compared with 69% of women). Fifty-five per cent of 25- to 35-y-old men had a BMI recorded and the level of recording rose with increasing age, reaching just under three-quarters by age 85. Relatively few BMIs had to be derived from height and weight other than in the 15- to 25-year-old age group.
Nearly half (47.7%) of the standardized population was overweight or obese; the proportion was 55% based on the raw data. Thirty-nine per cent of men were overweight compared with only 29% of women. However, 17.2% of women were obese compared with only 15.5% of men. Men and women had comparable proportions who were moderately obese (BMI 30 to 34.9 kg/m2) and there were nearly twice as many severely (BMI 35 to 39.9 kg/m2) and morbidly obese (BMI ≥ 40 kg/m2) women as men. In men, the proportions who were overweight or obese peaked in one of two decades: either between 45 and 54 y, or 55 and 64 y. In females, the comparable peaks came 10 y later: either between 55 and 64 y, or 65 and 74 y ().
Table I. Percentage of people with a BMI recording who were overweight and obese. Raw data and data standardized to the European population.
The mean BMI for males over 15 was 26.0 (SD 4.63, SEM 0.146) kg/m2 and for women was 25.7 (SD 5.51, SEM 0.154) kg/m2. The peak in the mean BMI mirrors the peak proportions of overweight and obese patients. The mean BMI for men aged 15 to 24 was 21.7 kg/m2 rising to 27.3 kg/m2 between 55 and 64 y of age, then falling off again. For females, the corresponding figures were 23 kg/m2, rising to 27.12 kg/m2 one decade later (age 65 to 74).
Ischaemic heart disease (IHD), diabetes mellitus and hypertension were all more prevalent with increasing obesity (). These differences were all highly significant (t-test, p<0.001). However, the prevalence of IHD fell in men with morbid obesity, and in females with severe and morbid obesity. IHD and diabetes were more prevalent in overweight and obese men, but hypertension was more prevalent in females. The mean BMI for women with diabetes is 29.9 (SEM 0.086, SD 6.47) kg/m2 and for men 28.7 (SEM 0.063, SD 5.19) kg/m2; there was almost no gender difference in BMI in IHD and hypertension.
Table II. Change in the prevalence of ischaemic heart disease, diabetes and hypertension with increasing obesity.
The proportion of people with one co-morbidity increased with age and level of obesity and the same pattern was repeated with two, three and four co-morbidities (). (Diagnosis of stroke or transient ischaemic attacks and peripheral vascular disease were included with IHD, diabetes and hypertension as possible co-morbidities.) In most of the sections of there is a fall in the proportion of patients with morbidities in the over-85s with severe or morbid obesity.
Table III. Percentage of each age band and level of obesity for people with one to four co-morbidities.
The completeness of cardiovascular risk factor data was increased in those with a BMI record and higher still in people who were overweight or obese. Data were more complete in females than in males and with increasing age. The quality of risk factors recording was higher for overweight and obese people compared to data quality levels in the rest of the population. Data recording for smoking and blood pressure was better for both men and women between 25 and 85 y in overweight and obese people. Cholesterol and triglyceride were also better recorded in the obese and overweight; the data quality levels peaked in the 65 to 75 age group for both men and women.
Systolic and diastolic blood pressure, glucose (in non-diabetics), cholesterol, and triglycerides were generally higher in those who were overweight and obese. All these variables were higher in men than women and generally increased with age, with the difference between the non-overweight and obese populations being significant (t-test, p<0.001.) The exceptions were diastolic blood pressure, which fell in people over age 65 y and cholesterol in obese men, which was not statistically significantly different from the non-obese (see ).
Table IV. Mean values of risk factors.
A higher proportion of overweight and obese patients was recorded as smokers than those with a normal body weight. However, this may be accounted for because the data for overweight and obese patients was much more complete. The peak decade for male smoking was 34 to 45, and for females 25 to 34 ().
Figure 2. Proportion of obese people who smoked compared with overweight people and the general population.
![Figure 2. Proportion of obese people who smoked compared with overweight people and the general population.](/cms/asset/ceb0570f-6095-4df8-b9f1-a66b63c5e941/igen_a_175685_f0002_b.gif)
The proportion of overweight and obese people above targets for lipids and blood pressure, set out for cardiovascular health Citation[10], Citation[19], were generally two to three times greater than in those of normal BMI. The differences in blood pressure were highly significant (χ2 test, p<0.001), the changes in cholesterol and LDL cholesterol much less so, particularly in males ().
Table V. The proportion of overweight and obese people above national targets for lipid and blood pressure management.
Discussion
Principal findings
Where data are recorded, over half of men and nearly half of women were overweight or obese. A greater proportion of men were overweight, whereas more women were obese. Obesity peaked between 45 and 55 in men, and 10 y later in women, after which it began to fall. Cardiovascular morbidity, diabetes, hypertension and cardiovascular risk factors were all more common with increasing obesity. Our data suggest that smoking was not increased in overweight or obese people; though computer data were more complete in these groups.
Implications for practices
The data demonstrate that, despite current levels of treatment, cardiovascular risk was increased and risk factor control worse in these groups of people. General practitioners and primary healthcare teams need to treat their risk factors more aggressively.
Limitations of the study
The population did not precisely match the national population. Practices generally joined on a locality basis, following the advice of their local primary care organization. Therefore, whilst there was good representation of different sizes of practices, this study population is weighted towards the more affluent south of England.
The data collected were not exhaustive. We did not search for obesity diagnosis codes, or drugs for obesity, we simply extracted or derived BMI. Heights in general practice may not be completely accurate; often they are reported verbally by patients. Elderly people lose vertical height.
Only 63.1% of the total population had a BMI record. Overall the overweight and obese had higher levels of data recording. This better data quality may be accounted for by their increased rates of attendance Citation[20]; though it is impossible to do more than speculate about missing data.
No account was taken of people taking medication to control their blood pressure or cholesterol. It is possible that, if these patients were analysed separately, the rise in blood pressure and lipids may have been even greater in the overweight and obese groups. A survey of this type does not demonstrate a causal link between obesity and cardiovascular risk.
Comparison with the literature
Primary care professionals face the challenge of learning how to manage cardiovascular risk in overweight and obese individuals as well as they do the non-obese. This is recognized to be a problem across Europe Citation[21]. Exercise campaigns have been found to increase physical activity, but only have a marginal effect on body weight Citation[22]. Dietary advice from GPs has been found to be more effective than that from dieticians, but the effect is small Citation[23]. GPs may be unsure how to give nutritional advice most effectively Citation[24]. A trial is currently being conducted to see if Internet support might improve obesity management Citation[25].
These data suggest that the epidemic in obesity predicted by the National Audit Office for the UK appears not to have occurred Citation[1]. The data are consistent with those reported in other studies: the male/female ratio and age-related peaks are similar Citation[5], Citation[7], Citation[26]. The associations with diabetes, heart disease and hypertension Citation[27] have also been reported Citation[6], Citation[28]; these data illustrate that these problems remained unaddressed. Obesity has been identified as an important predictor of prevalence of diabetes Citation[29] and the extent of glycaemic control Citation[30].
Call for further research
Further research is needed to understand whether the decline in obesity with advancing age, and the fall in the prevalence of IHD in those with morbid obesity, is due to increased mortality of overweight and obese people with risk factors, i.e. “survival bias”. Practitioners need to know whether tackling the obesity or the cardiovascular risk (or both) is the most appropriate strategy. We also need to understand why obese patients are not having their modifiable risk factors managed as well as the non-obese. Is this because of a lack or resources, knowledge, or some other cause?
Conclusion
Obesity is associated with increased cardiovascular morbidity and risk. Despite this being known, modifiable cardiovascular risk factors remain suboptimally controlled in this group of patients compared with the general population. Primary care needs effective strategies to address this unmet need.
The participating practices are acknowledged for their participation in the PCDQ programme. The Primary Care Data Quality (PCDQ) coronary heart disease programme is funded by an unconditional educational grant from MSD.
References
- National Audit Office. National Audit Office. Tackling obesity in England. The Stationery Office, London 2001
- World Health Organisation (WHO). Obesity: preventing and managing the global epidemic. WHO, Geneva 1998
- Lean M, Han TS, Seidell JC. Impairment of health and quality of life in people with large waist circumference. Lancet 1998; 351: 853–6
- Bray GA. Don't throw the baby out with the bath water. Am J Clin Nutr 2004; 79: 347–9
- Department of Health. Health survey for England – cardiovascular disease. London: Stationary Office; 1999. Available at URL:, , http://www.archive.official-documents.co.uk/document/doh/survey98/hse98.htm.
- Bray GA. Medical consequences of obesity. J Clin Endocrinol Metab 2004; 89: 2583–9
- Inelmaen EM, Sergi G, Coln A, Miotto F, Peruzza S, Enzi G. Can obesity be a risk factor in elderly people?. Obes Rev 2003; 4: 147–55
- de Lusignan S, Hague N, Brown A, Majeed A. An educational intervention to improve data recording in the management of ischaemic heart disease in primary care. J Public Health (Oxf) 2004; 26: 34–7
- National Statistics. Census 2001: England population. Available at URL:, , http://www.statistics.gov.uk/census2001/pyramids/pages/64.asp.
- NHS Confederation. The New General Medical Services Contract. Annex A: quality indictors – summary of points. Available at URL:, , http://www.nhsconfed.org/docs/annex_a_quality_indicators.doc.
- St. George's Hospital Medical School. Primary Care Data Quality Project. What is MIQUEST. Available at URL:, , http://www.pcdq.org/miquest.htm.
- van Vlymen J, de Lusignan S, Hague N, Chan T, Dzregah B. Ensuring the quality of aggregated general practice data: lessons from the Primary Care Data Quality Programme (PCDQ). Stud Health Technol Inform 2005; 116: 1010–5
- Thiru K, de Lusignan S, Sullivan F, Brew S, Cooper A. Three steps to data quality. Inform Prim Care 2003; 11: 95–102
- Thiru K, Hassey A, Sullivan F. Systematic review of scope and quality of electronic patient record data in primary care. Br Med J 2003; 326: 1070
- de Lusignan S, Dzregah B, Hague N, Chan T. Cholesterol management in patients with ischaemic heart disease: an audit-based appraisal of progress towards clinical targets in primary care. Br J Cardiol 2003; 10: 223–8
- Commission for Health Improvement. NHS performance indicators direct standardisation methodology. Available at URL:, , www.chi.nhs.uk/Ratings/ Downloads/direct_standardn_meth.pdf.
- SIGN (Scottish Inter-collegiate Guidelines Network). Publication number 8. Obesity in Scotland: integrating prevention with weight management. Available at URL:, , http://www.sign.ac.uk/pdf/sign8.pdf.
- PRODIGY (Prescribing Rationally in General Practice). PRODIGY guidance – obesity. Available at URL:, , http://www.prodigy.nhs.uk/guidance.asp?gt = Obesity.
- Department of Health. National service framework for coronary heart disease. (Ref: 166021P30K). London: Department of Health; 2000. Available at URL:, , http://www.dh.gov.uk/policyandguidance/healthandsocialcaretopics/coronaryheartdisease/fs/en.
- Frost GS, Lyons GF., Counterweight Project Team. Obesity impacts on general practice appointments. Obes Res 2005; 13: 1442–9
- Macdonald R. European GPs don't know how to deal with obesity epidemic. Br Med J 2005; 330: 1348
- van Sluijs EM, van Poppel MN, Twisk JW, Chin A, Paw MJ, Calfas KJ, van Mechelen W. Effect of a tailored physical activity intervention delivered in general practice settings: results of a randomized controlled trial. Am J Public Health 2005; 95: 1825–31
- Olsen J, Willaing I, Ladelund S, Jorgensen T, Gundgaard J, Sorensen J. Cost-effectiveness of nutritional counseling for obese patients and patients at risk of ischemic heart disease. Int J Technol Assess Health Care 2005; 21: 194–202
- Hankey CR, Eley S, Leslie WS, Hunter CM, Lean ME. Eating habits, beliefs, attitudes and knowledge among health professionals regarding the links between obesity, nutrition and health. Public Health Nutr 2004; 7: 337–43
- Kirk SF, Harvey EL, McConnon A, Pollard JE, Greenwood DC, Thomas JD, et al. A randomised trial of an Internet weight control resource: the UK Weight Control Trial [ISRCTN58621669]. BMC Health Serv Res 2003; 3: 19
- Elia M. Obesity in the elderly. Obes Res 2001; 9 Suppl 4: 244S–8S
- Carretero OA, Oparil S. Essential hypertension part I: definition and etiology. Circulation 2000; 101: 329–35
- Kopelman PG. Obesity as a medical problem. Nature 2000; 404: 635–43
- Goyder E, Hammersley V. Explaining variations in reported diabetes prevalence in general practice: how much variation is explained by differences between populations?. Br J Gen Pract 2003; 53: 642–4
- Ciardullo AV, Daghio MM, Brunetti M, Bevini M, Daya G, Feltri G, et al. Characteristics of an Italian diabetic population and long-term changes in glycaemic control at varying body weight categories. Soz Praventivmed 2003; 48: 186–90