Rose revisited: a “middle road” prevention strategy to reduce noncommunicable chronic disease risk

886 Bulletin of the World Health Organization | November 2007, 85 (11) Rose revisited: a “middle road” prevention strategy to reduce noncommunicable chronic disease risk Wendy J Brown,a Richard Hockey b & Annette Dobson b Perspectives a School of Human Movement Studies, University of Queensland, St Lucia, QLD 4072 Australia. Correspondence to Wendy J Brown (e-mail: wbrown@hms.uq.edu.au). b School of Population Health, University of Queensland, Australia. doi: 10.2471/BLT.07.041566 (Submitted: 19 February 2007 – Accepted: 2 May 2007) In light of worldwide concern about the obesity crisis and prevention of noncommunicable chronic disease, it is timely to revisit the principles advo- cated by Geoffrey Rose.1 The essential tenet of his work is that while strategies that focus on high-risk individuals (for example, weight-loss clinics for obese people) may help these people reduce their risk of chronic disease, the impact on the total burden of disease at the population level may be disappoint- ing. This is because numerous cases of risk-factor-related health problems may arise among the many people who are in the middle of the risk distribution. In contrast, by lowering the risk across the whole population, the numbers of attributable cases of disease are signifi- cantly reduced.1 Although this principle is well documented for conditions like hyper- tension, which have a relatively direct or linear relationship with risk factors such as body mass index (BMI), it is unclear for conditions like diabetes, where incidence rises sharply among people who are in the overweight and obese categories of BMI. To explore this issue, we used data from eight years’ follow-up of middle-aged women in the Australian Longitudinal Study on Women’s Health2 to estimate the reductions in incidence of hypertension and diabetes that would result if the BMI distribution were shifted to the left in various ways. The Australian Longitudinal Study of Women’s Health Participants were randomly selected from the national Medicare health insur- ance database (which includes all perma- nent residents of Australia regardless of age, including immigrants and refugees) with intentional over-representation of women living in rural and remote areas. Further details of the recruitment methods and response rates have been described elsewhere.2 The study collects self-reported data using mailed surveys at 2- to 3-year intervals from about 40 000 women living in all Australian states and territories. The surveys include questions about: health conditions, symptoms and diagnoses; use of health services; health-related quality of life, in- cluding measures of physical and mental health; social circumstances, including work and time use; demographic fac- tors; and health behaviours. Informed consent was obtained from all partici- pants in 1996, with ethical clearance by the University of Newcastle, Australia. This paper includes data from 13 716 women in who were aged 45–50 at the time of the first survey in 1996. The women were asked to report their height and weight at each survey. BMI was calculated as reported weight (kilograms) divided by the square of reported height (metres).2 At each sur- vey women were asked if they had been told by a doctor that they had any of a list of conditions, including hyperten- sion and diabetes. At survey 1 (1996) they were asked if they had ever had a diagnosis of hypertension or diabetes. At surveys 2 (1998), 3 (2001) and 4 (2004) they were asked whether they had been diagnosed with each condition in the time period that had elapsed since the previous survey. Modelling different prevention approaches BMI at survey one is presented as a simple frequency distribution in Fig. 1. After excluding data from women who reported having hypertension (n = 2859) or diabetes (n = 395) at survey 1, incidences of hypertension and diabetes (1996–2004) were calculated and superimposed on the 1996 BMI distribution (Fig. 1). Mean BMI was 25.8 (standard deviation, 5.13) kg.m–2. Hypothetical reductions in the in- cidence of each condition were then modelled for: (1) a 1-unit reduction in BMI in the whole population (the whole-population strategy); (2) a 3-unit reduction in BMI for women in the top 20 percent of the BMI distribution (BMI > 29; the high-risk strategy); and (3) a 2-unit reduction in BMI for women in the top 50% of the BMI dis- tribution (BMI  24; the “middle road” strategy). These analyses were performed using SAS, version 9.1.2. Whole-population or high-risk approach? Our calculations showed that if the entire BMI distribution was shifted to the left by 1 unit (the whole-population strategy), then incidence of hyperten- sion and diabetes would be reduced by 10.3% and 13.4% respectively. In con- trast, if BMI were shifted to the left by 3 units in the top 20% of the distribution (BMI > 29) (the high-risk strategy), the corresponding reductions in incidence would be 7.3% for hypertension and 16.8% for diabetes. It is clear that for hypertension, the whole-population strategy of achieving a small change in everyone would result in a greater re- duction in cases (10.3%) than would the high-risk strategy of larger changes among those in the top 20% of the BMI distribution (7.3%). However, for diabe- tes, where the risk increases much more at the high end of the BMI distribution, a greater change among those most at risk would result in a greater popula- tion risk reduction (16.8% compared with 13.4%). 887Bulletin of the World Health Organization | November 2007, 85 (11) Perspectives Rose revisited: a “middle road” prevention strategyWendy Brown et al. References 1. Rose G. The strategy of preventive medicine. Oxford: Oxford University Press; 1992. 2. Lee C, Dobson AJ, Brown WJ, Bryson L, Byles J, Warner-Smith P, et al. Cohort profile: the Australian longitudinal study on women’s health. Internat J Epidemiol 2005;34:987-99. 3. Hill JO, Wyatt HR, Reed GW, Peters JC. Obesity and the environment: where do we go from here? Science 2003;299:853-855 4. Hill JO. Preventing excessive weight gain. Obesity Res 2005;13:1302. Fig. 1. Distribution of BMI in 13 716 Australian women aged 45–50 years in 1996 3000 2500 1000 500 0 Fr eq ue nc y 80 0 Ra te p er 1 00 0 pe rs on -y ea rs Incidence of hypertension between 1996 and 2004 1500 2000 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 BMI 10 20 30 40 50 60 70Incidence of diabetes between 1996 and 2004 BMI Projected BMI distribution under the �middle-road� strategy targets The “middle road” approach However, when we modelled a shift in the BMI distribution of 2 units in the top 50% of the BMI distribution (BMI  24) – the middle road strategy – there was a reduction in incidence of 12.3% for hypertension and 23% for diabetes. Hence for both hypertension and diabe- tes, the middle road strategy would result in the greatest reduction in both these conditions in this population of middle- aged women. The white bars in the Fig. 1 indicate the new population distribution of BMI, if this could be achieved. Like many others, we were under the impression that Rose’s hypotheses related only to the high-risk and popula- tion approaches. On revisiting his text, we found Rose had noted that while the range of variation in many health risks is quite stable (that is, when the population average shifts, the dispersion remains fairly constant), this was less true for body weight. He pointed out that when the population average weight rises, there is a disproportionate increase in “exceptional” obesity. He suggested therefore that weight reduction in the entire upper half of the distribution, without any slimming among those on the “thin side of average”, would be particularly welcome, because of the “J-shaped” relationship between body weight and total mortality.1 Lessons learned Rose’s advice on the benefits of weight reduction for the upper half of the BMI distribution is borne out by our prospec- tive data. Although our estimates are based on self-reported data and require replication with more objective mea- sures, they confirm that a middle road strategy may be optimal for combating the growing rates of chronic illness as- sociated with increasing BMI. Using our data, we estimate that if all women with BMI  24 reduced BMI by 2 points (approximately 5.5 kg), the incidence of hypertension would be reduced by 12% and the incidence of diabetes would be reduced by 23%. Assuming that each kilogram of stored fat is equivalent to 7000 kcals (29 400 kJ), and that this energy is converted with an efficiency of 50%, then closing the “energy gap” by 200 kcals per day through an additional 20 minutes or 2000 steps of brisk walking and re- ducing energy intake by 100 kcal (the equivalent of one chocolate biscuit) ev- ery day for one year would substantially contribute to achieving this goal.3,4 As suggested in the WHO Global Strategy on Diet, Physical Activity and Health (DPAS) (resolution WHA57.17, May 2004), these data suggest that small changes in two of the main risk factors could result in significant reduc- tions in risk of chronic noncommuni- cable diseases across populations. The level of change suggested here is likely to require considerable social, cultural and environmental support to encour- age more active living among today’s time-pressured women. ■ Competing interests: None declared. BMI, body mass index.