Pediatric Debate

International Journal of Obesity (2011) 35, 1270–1276; doi:10.1038/ijo.2011.163; published online 9 August 2011

Can we modulate physical activity in children?

T J Wilkin1

1Department Endocrinology and Metabolism, Peninsula Medical School, Plymouth, UK

Correspondence: Dr TJ Wilkin, Department of Endocrinology and Metabolism, Peninsula Medical School, University Medicine, Level 7, Derriford Hospital, Plymouth PL6 8DH, UK. E-mail:

Received 31 January 2011; Revised 24 June 2011; Accepted 2 July 2011; Published online 9 August 2011.



Intuition tells us that physical activity is central to weight reduction in obese children. Evidence, on the other hand, suggests that increases in physical activity are difficult to achieve in the short term, and may not be possible in the long term. One explanation could be an ‘activitystat’, a feedback loop in the child's brain that controls physical activity according to a set point. This brief article, which argues that it may not be possible to modulate the activity of children, reviews the principles of feedback control as they apply to physical activity, discusses evidence for its central control, and demonstrates how a physical activity control loop might operate to defend the set point. Studies restricted to objective measurement suggest that the physical activity of children varies in a systematic, rather than random manner. It varies little from environment to environment, from year to year or from place to place. Where children undertake more activity at one time of day, they appear to compensate at another. Systematic variation of this kind implies control, and the control of physical activity appears to lie with the child, not with his environment. Perturbation (temporary change in response to disturbance) during short-term physical activity interventions may be mistaken for modulation (permanent change in set point), a fundamentally different response. Perturbation lasts no longer than the disturbance that causes it, and there is little evidence that interventions raise activity long term, if at all.


physical activity; activitystat; central control; modulation



It is universally acknowledged that physical activity is good for health.1 Whether measured as cardio-respiratory health (VO2 max) or metabolic health (insulin sensitivity), the indices are regularly higher in more active children.2, 3 It is important to stress this fact, because the debate around physical activity often confuses fitness with fatness. Although physical activity is good for the health of children, there is no evidence that physical activity reduces their fatness.4

There is another common misconception. Associations are just associations, nothing more. They say nothing about causality so that, whereas the more active child is undoubtedly fitter, there is no evidence from the association alone that the less-fit children can be made more active.

Success with physical activity interventions is the Holy Grail of exercise science, and the critical interpretation of data gathered in the quest is important. This article does not contest the link between physical activity and fitness, quite the reverse, but argues that studies which claim that intervention can modulate the physical activity of children merit closer inspection.


Homoeostasis, modulation and perturbation

The negative-feedback (homoeostatic) control loop is ubiquitous in nature and in man's attempts to emulate it. Its design aims to ensure that output matches the set point under all conditions, which it achieves with the help of three components—a dial, a comparator and a generator (Figure 1).5 The dial defines the set point, the comparator compares output with the set point and the generator responds to the difference between the two. The dial, thermostat and boiler of a domestic heating system correspond to the three components in a more familiar context.

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Generic negative-feedback control system. The activitystat hypothesis proposes that physical activity (output) is set in the hypothalamus and controlled by a neurohumoral loop whose generator (the locomotor system) is constrained by a control signal. The control signal is a function of the difference between cumulative activity and the set point, and assures compensation accordingly.

Full figure and legend (61K)

The output of a control system may change for one of the three reasons—change in set point (modulation), functional failure of the comparator or the generator (dysfunction), or environmental disturbance (perturbation). The systems engineer is familiar with the notion of homoeostatic perturbation, where some environmental disturbance (opening the windows wide on a winter day, for example) is sufficient to alter output without a change in the set point. But perturbation, by its nature, lasts only for as long as the disturbance which caused it, and output will return towards the set point once the disturbance responsible is removed (the windows are closed again). Furthermore, the extent to which a disturbance perturbs output depends on the efficiency of the loop (known as its gain). A loop of low gain will be more susceptible to disturbance (more easily perturbed, less able to defend its set point), where one of high gain would be correspondingly resistant (that is, difficult to perturb).

Modulation is different and the distinction is critical. Modulation implies a change in setting to which the healthy control system faithfully responds by adjusting its output accordingly (control loops are sometimes referred to as servo-systems, after the Latin servo–I serve). Modulation quite properly alters the output of the loop, does not imply dysfunction and would not normally be caused by a disturbance.

Programmes designed to raise the physical activity of children are disturbances, and the key consideration is whether they can be expected to modulate the child's activity long term, or whether they merely perturb it for as long as the programme is applied. Assuming for one moment that the physical activity of children is subject to feedback control of the kind outlined above, a physical activity intervention is unlikely to alter the activity of the child beyond the period for which it is in place. Unless modulated by a change in the set point, control loops oppose change. Indeed, the purpose of a control loop is to defend its set point, making experimental intervention difficult.

In Darwinian terms, the capacity of a control loop to defend its set point will have evolved in response to survival pressure. Accordingly, where mankind evolved in the face of privation, energy expenditure is likely to have emerged tightly controlled, and defence of the physical activity set point against random activity correspondingly robust. Weight gain, on the other hand, has not been a survival pressure until recently, so that the reverse—the ability to defend physical activity against random weight increase—may arguably be weak. The impact of body mass on physical activity will be examined later.


Environmental or biological?

Ask Joe Public what controls the physical activity of children, and you will likely get the ‘environmental’ response—local park amenities, school playing fields, school time allocated to physical education, sports clubs, parental encouragement, etc. The environmental response is intuitive and, to many, obvious. However, ask a biologist—and particularly, an evolutionary biologist—the same question and the answer will almost certainly be different. The ‘biological’ response is liable to view physical activity—the only readily modifiable component of energy expenditure—as crucial to energy balance and therefore almost certainly, subject to biological control.

The ‘appestat’—a centre in the hypothalamus that regulates energy intake—is a widely accepted construct for which there is compelling evidence.6 Thus, brain damage,7 leptin deficiency8 and other genetic conditions such as Prader–Willi syndrome9 are all well-recognised disorders affecting appetite control, which result in obesity. Rowland10 first used the term ‘activitystat’ to conceptualise a corresponding centre for the control of physical activity, and we and others have embraced the concept.11, 12, 13

For reasons outlined above, there is little reason for systems which limit body mass to have evolved, but persuasive reasons for those which limit physical activity to have done so. It is the purpose of this paper to present evidence from the literature that the physical activity of children is under central control, and therefore, difficult to modulate through intervention.


Order and randomness

Control implies order and command over randomness, that is, systematic variation. We therefore sought evidence in the first instance for systematic rather than random variation in physical activity among children, according to environment, time and population.


We compared the physical activity of 5-year-old EarlyBird children recorded by accelerometer over five school days and over two weekend days (Figure 2). The line of best fit coincides with the line of identity. Moreover, the correlation of r~0.5 between weekday and weekend day activity suggests that not only was physical activity in different environments the same, but also its variance among individual children systematic. Figure 2, however, relates to 5-year-olds, and not every study has reported the same in older children. Rowlands et al.14 used accelerometers to examine the activity patterns of older children (9–12 years) and recorded more activity on weekday than weekend days. The girls nevertheless remained systematically less active than the boys. Very recently, McManus et al.15 again reported more weekday than weekend day activity in Hong Chinese children. The analysis, however, was concerned with patterns rather than volume of activity, and recorded only school-time activity on weekdays. If compensation occurs in response to regulation of physical activity as we propose, restriction of the measurement to part of the day only will wrongly estimate the daily total. Notwithstanding, the report shows the same low coefficient of variation that characterises control. Steele et al.16 on the other hand, reported no differences from weekday to weekend in 9–10 year-olds, whereas Aznar et al.17 found differences only for MVPA.

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Relationship between the mean daily physical activity of 5-year-old EarlyBird children during the school week and the home weekend.

Full figure and legend (92K)


Correlations in physical activity, both year-on-year and relative to baseline, were retained throughout childhood (Table 1), the latter inevitably weakening over time. Systematic retention of activity patterns, despite the changes in school, home environment and biological development that characterise childhood, are consistent with central control. Activity among the girls fell rapidly over time, but the variance remained unchanged, which may be consistent more with biological modulation of the set point during female puberty, than with randomly changing environmental disturbance. Others have made similar observations using pedometers. In a study involving more than 1000 6–12 year-olds from around the world, Wickel et al.18 reported low within-child day-to-day variability.


The total weekly activity of the EarlyBird children at 6 years was compared with that of a cohort of similar age living in Glasgow, using the same type of accelerometer (Table 2). The activity recorded by the girls was systematically lower than that of the boys in both locations, but the gender-specific means were much the same and the variances statistically no different. Glasgow lies some 800km due north of Plymouth, has a different light/dark cycle, different school system and different social culture. The similarities in physical activity nevertheless suggest that not only is there systematic variation within childhood populations, but consistency between populations as well.


Compensatory responses

The activitystat principle assumes that activity is constantly monitored by neurohumoral circuits and adjusted over time to meet the set point. Accordingly, it would predict that activity imposed during one part of the day would be met by compensation at another. To detect such compensation, it would be necessary to compare two separate, but adjacent blocks of time, to confine the intervention to one of them and to establish whether compensation occurred in the other. We used this design to explore the impact of very different in-school physical activity programmes in three UK primary schools, comparing the in-school with the out-of-school activity in each case. The key issues were to ensure that the differences in physical activity in-school were sufficient between the schools to be clearly recordable, and that the differences were maintained throughout.

To this end, we used accelerometers to measure the physical activity of 230 pre-pubertal children who were attending three schools (S1, S2 and S3—referred to from here on as the ‘three schools study’) that allocated (imposed), respectively, 9.2, 2.2 and 1.6h weekly to physical education. The range was sufficient to record 64% more in-school physical activity in S1 compared with S2 or S3 (P<0.001). The outcome over the week was revealing. Where the children at S1 recorded substantially more physical activity during school hours, they compensated—with striking precision—out of school.19 Thus, when the in-school and out-of-school activities at each school were totalled to provide mean total daily activity, there were no differences between the three. The daily means were the same, their ranges (variances) were the same, but the opportunity provided in-school very different. There was a more than five-fold difference in physical education time between the schools and a four-fold (interquartile) range of activity among the children, but less than 1% of the variance in activity of the children could be accounted for by the school attended. We concluded that the range in physical activity universally observed in children is attributable, not to opportunity (the environmental argument), but to a range of hypothalamic set points (the biological argument).

Reproducibility is central to the scientific argument. Accordingly, we repeated the study on three subsequent occasions to test whether it was reproducible from school term to school term, and to examine the influence of seasonal confounders.20 The patterns remained unchanged, even after adjusting for daily rainfall and for daylight hours (Figure 3). The girls remained systematically less active than the boys, and differences in school-time activity were always finely compensated by corresponding differences out of school. These data again suggest that the physical activity of children varies systematically, rather than randomly, and is therefore likely to be under control.

Figure 3.
Figure 3 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Habitual physical activity of children from three schools offering 9.2h physical education time weekly (S1), 2.2h (S2) and 1.6h (S3). Both total weekly activity (TPA, upper panel) and its moderate-and-vigorous component (MVPA, lower panel panels) are shown. The data represent the age- and gender-adjusted means of all four school terms and their 95% CI. None of the individual terms differed from another. For full data, see Fremeaux et al.20

Full figure and legend (25K)


Intervention studies—animals

Wheel-running is a well-established means of investigating the habitual physical activity of rodents, rendered more sophisticated by the introduction of digital measurement, the motorised wheel to impose activity, and the metabolic cage to measure energy balance directly.21 If the close similarities in physical activity across human populations observed in our own studies suggest a genetically determined set point, the heritability of wheel-running behaviour among laboratory rodents imply it more strongly still.22, 23 Thus, the progeny of lines selectively bred for different wheel-running behaviour faithfully follow the bloodline, whatever the opportunity offered. Studies to establish the neurohumoral circuits involved have already begun to reveal key pathways.24, 25, 26


Intervention studies—humans

The intuitive, but unfounded, belief that physical activity can be used to reduce the weight of obese children implies that intervention programmes can successfully raise the level of activity in the first place. There are studies that appear to do so, but it is crucial to distinguish what may be short-term perturbation from long-term modulation.

A recent meta-analysis involving 3000 children (submitted for publication), which incorporated only studies which used objective measures of physical activity (not self-report) throughout waking hours, made several important observations. First, and most importantly, there was no overall effect on total activity. Second, the biggest effect sizes emerged from the shortest studies. Indeed, there was a clear inverse correlation between duration and outcome (Figure 4). Finally, and crucially, three of the longer studies recorded interim outcomes that were greater than the ultimate outcome. In other words, the response peaked and subsequently fell. Thus, ‘before and after’ measures alone might give the impression of sustained response, whereas the trend rises only to fall again. Two points are not sufficient to deduce trend and can be misleading if relied upon to do so. The data seems more consistent with perturbation than with modulation. Once the load of the intervention (disturbance) lessens (perhaps enthusiasm for it wanes), the level of activity appears to revert towards the set point.

Figure 4.
Figure 4 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Effect sizes of the 12 published physical activity intervention studies in children using objective measurement (accelerometers). The size of circle reflects the numbers in the study and the arrows trace the trends in two individual studies where interim measures were available.

Full figure and legend (62K)



We have to reconcile the compensatory behaviour observed among children in our three schools study with the sometimes substantial responses to intervention in other studies. How to explain effect sizes of up to 1.2s.d. following intervention, while arguing that children respond to a set point? There could be three possible explanations. First, the ‘intervention’ in the three schools study, although clearly recordable at each repeat measure, may not have been sufficient to perturb the loop, permitting precise compensation. Although the time scheduled for physical education in S1 was 9.2h, it still constituted only a small part (~10%) of the waking week. The studies achieving sizeable intervention effects, on the other hand, may have introduced more substantial disturbances that were capable of perturbing control. Second, the observations made in the three schools study were not ‘before and after’—as is the case for intervention studies—but ‘during’. It is possible that the ability to compensate occurs progressively as an adaptation, which could explain the inverse relationship in the intervention studies between effect size and duration. Accordingly, the children in the three schools study might represent children in the later, rather than the earlier, stages of adaptation. Finally, as suggested above, the failure of intervention studies to achieve (or to sustain) an effect of longer term might simply reflect a failure to maintain the intervention. It is important, nevertheless, not to rely on this last explanation uncritically, as many authors have. The three schools study suggests that compensation applies even where the disturbance is clearly maintained.



Although it is not always appropriately quoted, the literature suggests that it is difficult, even with resources that would not normally be available, to increase the physical activity of children. Even where this appears to have been accomplished, increases tend to be confined to the short term, and there is a telling inverse relationship between effect size and duration of the intervention. One explanation for these observations—and an important possibility—is that of perturbation. All control loops can be perturbed given sufficient disturbance, but perturbation is temporary and differs fundamentally from modulation.

Given the evolutionary imperative of limiting weight loss in the face of restricted nutrition, it might be predicted that tight control of physical activity (strong defence of the set-point) would be selected for in avoidance of unnecessary energy expenditure. The reverse, however, may not be true. Increased physical activity has not evolved as a response to weight gain in the same way as a reduction has to falling weight.27, 28 Indeed, the heavier child is less active, an association that has sometimes been uncritically interpreted to imply that children need only be made more active in order to lose weight.29 Analysis of longitudinal data over multiple time points—the only means of examining trends and their causal interactions—tells a different story. From such analysis, it appears that high body mass index is followed by less activity, but that less activity is not followed by higher body mass index.30 The implications are important. Raising the physical activity of children, even if it could be achieved, seems unlikely to impact on their weight, whereas reducing their weight through calorie restriction may permit more physical activity.

Energy expenditure is correlated with physical activity so that, although their resting energy expenditure may be little different, high activity children expend more energy. Furthermore, heavier people, who undertake less activity, nevertheless expend more energy to move around because of their weight.31 It is pertinent therefore to question whether physical activity is the feedback signal that limits further activity, or the energy spent. There is currently no answer, though the question may be important in the context of aerobic and anaerobic exercise.

The intervention now waiting to be trialled in children is not more activity while monitoring body mass index, which has been shown not to work, but less body mass while monitoring activity. The corollary, at least, has been tried experimentally in adults, where overfeeding is indeed followed by a reduction in physical activity.32 The activitystat hypothesis proposes that physical activity is readily perturbed by weight excess, because the circuitry that defends the set point for physical activity responds only weakly to the load imposed by weight gain. Hypotheses need mechanisms, and one explanation for the fall in physical activity consequent upon weight gain may be a loss of mitochondrial efficiency. Mitochondria are cellular organelles responsible for energy transfer and are particularly profuse in muscle. Overweight is universally associated with a loss of insulin sensitivity in muscle, which reduces the efficiency of mitochondria,33 leading to early fatigue.34 Of course, the string in this argument remains speculative until the trial has been done, but primary weight reduction could improve metabolic health in two ways—directly by improving insulin sensitivity, and permissively by rendering the child more tolerant to physical exercise.



It is important to measure physical activity objectively and to incorporate whole periods (that is, whole days or preferably whole weeks) in the analysis, because in-school measures alone do not register compensatory behaviour out of school and risk misinterpretation as a result. There is no evidence that we can modulate the physical activity of children, although it can clearly be perturbed. There is a danger that the success of some short term studies in raising physical activity is being misinterpreted as modulation when it is really perturbation which will last only for as long as the environmental disturbance that caused it. The direction of causality between physical activity and body mass is fundamentally important if, as our data and those of others suggest, weight gain disturbs activity. Weight reduction may be expected to permit more physical activity, not as a result of modulation, but by attenuating a disturbance that was causing perturbation.


Conflict of interest

The author declares no conflict of interest.



  1. Warburton DE, Nicol CW, Bredin SS. Health benefits of physical activity: the evidence. Can Med Ass J 2006; 174: 801–809.
  2. Armstrong N, Welsman J. Physiology of the child athlete. Lancet 2005; 366(Suppl 1): S44–S45.
  3. Metcalf BS, Voss LD, Hosking J, Jeffery AN, Wilkin TJ. Physical activity at the government-recommended level and obesity-related health outcomes: a longitudinal study (EarlyBird 37). Arch Dis Child 2008; 93: 772–777.
  4. Harris KC, Kuramoto LK, Schulzer M, Retallack JE. Effect of school-based physical activity interventions on body mass index in children: a metaanalysis. Can Med Ass J 2009; 180: 719–726.
  5. Wilkin TJ. Endocrine feedback control in health and disease. In: Bittar EE and Bittar N (eds). The Principles of Medical Biology. JAI Press Inc: Greenwich, CT, 1998, pp 1–28.
  6. Joliffe N. The appestat; the appetite-regulating mechanism. Merck Rep 1952; 61: 3–7.
  7. Uher R, Treasure J. Brain lesions and eating disorders. J Neurol Neurosurg Psych 2005; 76: 852–857.
  8. Farooqi IS, O'Rahilly S. Leptin: a pivotal regulator of human energy homeostasis. Am J Clin Nutr 2009; 89: 980S–984S.
  9. McAllister CJ, Whittington JE, Holland AJ. Development of the eating behaviour in Prader-Willi Syndrome: advances in our understanding. Int J Obes (Lond) 2010; 35 (2): 188–197.
  10. Rowland TW. The biological basis of physical activity. Med Sci Sports Exerc 1998; 30: 392–399.
  11. Wilkin TJ, Mallam KM, Metcalf BS, Jeffery AN, Voss LD. Variation in physical activity lies with the child, not his environment: evidence for an ‘activitystat’ in young children (EarlyBird 16). Int J Obesity (Lond) 2006; 30: 1050–1055.
  12. Eisenmann JC, Wickel EE. The Biological basis of physical activity in children: revisited. Pediatr Exer Sci 2009; 21: 257–272.
  13. Garland Jr T, Schutz H, Chappell MA, Keeney BK, Meek TH, Copes et al. The biological control of voluntary exercise, spontaneous physical activity and daily energy expenditure in relation to obesity: human and rodent perspectives. J Exp Biol 2011; 214 (Pt 2): 206–229.
  14. Rowlands AV, Pilgrim EL, Eston RG. Patterns of habitual activity across weekdays and weekend days in 9-11-year-old children. Prev Med 2008; 46: 317–324.
  15. McManus AM, W Chu EYW, Yu CCW, Hu Y. How children move: activity pattern Characteristics in lean and obese Chinese children. J Obesity 2011; 2011: 679328 online. doi: 10.1155/2011/679328.
  16. Steele RM, van Sluijs EM, Sharp SJ, Landsbaugh JR, Ekelund U, Griffin SJ. An investigation of patterns of children's sedentary and vigorous physical activity throughout the week. Int J Behav Nutr Phys Act 2010; 7: 88.
  17. Aznar S, Naylor PJ, Silva P, Pérez M, Angulo T, Laguna M et al. Patterns of physical activity in Spanish children: a descriptive pilot study. Child Care Health Dev 2011; 37: 322–328.
  18. Wickel EE, Eisenmann JC, Pangrazi RP, Graser SV, Raustorp A, Tomson et al. Do children take the same number of steps every day? Am J Hum Biol 2007; 19: 537–543.
  19. Mallam KM, Metcalf BS, Kirkby J, Voss LD, Wilkin TJ. Contribution of timetabled physical education to total physical activity in primary school children: cross sectional study. Brit Med J 2003; 327: 592–593.
  20. Fremeaux AE, Mallam KM, Metcalf BS, Hosking J, Voss LD, Wilkin TJ. Physical education time in school and its influence on total physical activity: repeated measures study in 8-10 year-old children (EarlyBird 46). Int J Obesity (London); e-pub ahead of print 15 March 2011, doi:10.1038/ijo.2011.52.
  21. Tou JCL, Wade CP. Determinants affecting physical activity levels in animal models. Exp Biol Med 2002; 227: 587–600.
  22. Lightfoot JT, Turner MJ, Daves M, Vordermark A, Kleeberger SR. Genetic influence on daily wheel running activity level. Physiol Genomics 2004; 19: 270–276.
  23. Leamy LJ, Pomp D, Lightfoot JT. An epistatic genetic basis for physical activity traits in mice. J Heredity 2008; 99: 639–646.
  24. Mathes WF, Nehrenberg DL, Gordon R, Hua K, Garland Jr T, Pomp D. Dopaminergic dysregulation in mice selectively bred for excessive exercise or obesity. Behav Brain Res 2010; 210: 155–163.
  25. Knab AM, Lightfoot JT. Does the difference between physically active and couch potato lie in the dopamine system? Int J Biol Sci 2010; 6: 133–150.
  26. Kalra SP, Kalra PS. Neuroendocrine control of energy homeostasis: update on new insights. Prog Brain Res 2010; 181: 17–33.
  27. Redman LM, Heilbronn LK, Martin CK, de Jonge L, Williamson DA, Delany DP et al. Metabolic and behavioral compensations in response to caloric restriction: implications for the maintenance of weight loss. PLoS One 2009; 4: e4377.
  28. Martin CK, Heilbronn LK, de Jonge L, DeLanyJP, Volaufova J, Anton SD et al. Effect of calorie restriction on resting metabolic rate and spontaneous physical activity. Obesity 2007; 15: 2964–2973.
  29. Ness AR, Leary SD, Mattocks C, Blair SN, Reilly JJ, Wells J et al. Objectively measured physical activity and fat mass in a large cohort of children. PLoS Med 2007; 4: e97.
  30. Metcalf BS, Hosking J, Jeffery AN, Voss LD, Henley W, Wilkin TJ. Fatness leads to inactivity, but inactivity does not lead to fatness: a longitudinal study in children (EarlyBird 45) Arch Dis Child; e-pub ahead of print 23 June 2010, doi:10.1136/adc.2009.175927.
  31. Pacy PJ, Webster J, Garrow JS. Exercise and obesity. Sports Med 1986; 3: 89–113.
  32. Levine JA, McCrady SK, Lanningham-Foster LM, Kane PH, Foster RC, Manohar CU. The role of free-living daily walking in human weight gain and obesity. Diabetes 2008; 57: 548–554.
  33. Szendroedi J, Roden M. Mitochondrial fitness and insulin sensitivity in humans. Diabetologia 2008; 51: 2155–2167.
  34. Keyser RE. Peripheral fatigue: high-energy phosphates and hydrogen ions. Phys Med Rehab 2010; 2: 347–358.


I am grateful to the EarlyBird team who have generated compelling and sometimes challenging observations over the years, and particularly to Brad Metcalf, MSc, who has spent long hours debating the physical activity data and their interpretation. I am similarly indebted to the people who generously fund the EarlyBird Study—currently the Bright Futures Trust, Nestec, Peninsula Medical School, Kirby Laing Foundation, League of Friends and the Earlybird Diabetes Trust.

Extra navigation