Wellness Initiatives: Benefits and Limitations
Bibliographic record
Abstract
In the last decade, the public has become increasingly aware and concerned regarding their susceptibility to common conditions such as obesity and diabetes. Groups of scientists have founded initiatives that harness this desire to stay disease free and have developed public health projects using a large population base. This novel approach relies on a wide volunteer group of individuals who allow themselves to be monitored and studied over a long time period (ranging from a few months to many years). Data based on gene sequencing, sleep cycles, diets, activities, and blood sample analyses are recorded on a regular basis. By following each volunteer participant, whether they progress to disease or remain healthy, scientists hope to find lifestyle characteristics and corresponding microbiomes or genes that are key to living well. Projects such as the 100K Wellness Project, Lake Nona Life Project, and Google's Baseline Study have well-established participant groups and aim to use this information to benefit the greater public and/or to make a profit. Because individuals vary greatly, both in biological composition and lifestyle, data collected from a ranging population may give scientists a more comprehensive data source. These data, in turn, are used to construct a personalized lifestyle plan for customers of wellness companies such as Arivale or similar company models. A customized consultant is assigned to the consumer and will potentially give advice based on monitored parameters. Technology has built another level of data monitoring; as part of the Google Baseline Study, GoogleX has created contact lenses that constantly monitor glucose concentrations and smart trackers that can alert users of imminent heart attacks. Similar technologies such as Samsung's Simband constantly record and transmit information to a larger database, all as part of mass health studies. This recording and transmission of information, however, makes some users uncomfortable about the prospect of data leakage or other loss of privacy that may harm them. Here, we explore this rapidly developing sector where technology and mass population testing crosses. We asked 4 experts to give their input on the development potential and perceived drawbacks of this industry. Somebody said that the best definition of health is when you do not think about your health. Do you believe that scrutinizing your health will generate anxiety or relaxation? Nigel Paneth: The answer to this question is to a large extent dependent on the individual. Some people, particularly when young, are relatively indifferent to factors that threaten health but do not produce symptoms. At the other extreme, some people are excessively focused on their health risks, which are often subject to exaggeration by the media covering health issues. Most people who prefer not to dwell on their health risks are unlikely to be interested in monitoring or otherwise scrutinizing their own health. The remainder—likely a small minority of the population, and concentrated in the educated classes—may find such monitoring interesting and potentially useful. The first condition mentioned in the preface to this Q&A article discussing public awareness and concern is obesity and the second is diabetes. Is there anyone who needs special monitoring to determine that they are obese? And diabetes can be diagnosed with a simple blood sugar measurement. How would such conditions benefit from complex biochemical monitoring? Henrik Vogt: When I was younger, I experienced health anxiety, and at that point such scrutiny would have generated worry, lost health, and lost youth as different findings had to be ruled out as dangerous. Now, I believe I would be less affected. As a doctor, I have become more relaxed towards most bodily symptoms and findings, trust myself as a generally self-maintaining entity, and live with the risk that comes with being alive. I definitely do not think I would be more relaxed, because I am relaxed in the first place (and should be as a healthy 39-year-old). The technology would, however, generate in me an awareness of risk, and if health equals the absence of thinking about health and risk, it would, by definition, be impossible to feel completely well anymore. Jerry Yeo: In general, I would say scrutinizing health tends to generate some level of anxiety in most people. For example, if I subject myself to blood screening tests, I will naturally be wondering which of those tests will return as abnormal and what the consequences will be. Although I am relatively healthy, if my blood screening test shows an abnormal value, I will naturally seek to do more follow-up testing to make sure that I don't have an undiagnosed disease. Stephen Master: I think that the answer to this question will depend entirely on the individual and their life experiences. Of course, as was pointed out during the early years of genetic susceptibility testing, everyone wants to receive a reassuring answer. However, individuals vary widely in how well they can evaluate equivocal (or possibly abnormal) results. Even some medical trainees may exhibit hypochondriasis until they have gained a sense of clinical perspective (the so-called “second-year medical student” syndrome), and a little knowledge can be dangerous. For others, however, a focus on finding better ways to monitor health may be just as normal as trying to measure their financial or career success. In your opinion, will patients actually follow instructions from coaches and make lifestyle changes? Nigel Paneth: Practitioners know that it is very hard to convince people to undertake lifestyle changes. As a cardiologist friend of mine once said, “The most I can achieve with my patients is to get them to quit smoking and take their medicines.” On the other hand, there is evidence that focused and detailed advice, such as from dieticians, can be adopted, but this is in relation to clear health risks such as obesity and diabetes. If established clinical disease engenders lifestyle changes only with difficulty, then how likely is it that asymptomatic biochemical abnormalities will? The few people who are willing to undertake intensive monitoring are probably those most likely to be coached into lifestyle changes. But for the majority of the population, this is unlikely to be an effective strategy. Henrik Vogt: Some no doubt will, but neither previous evidence on general health checks and lifestyle counselling, nor new evidence on the impact on communicating genetic risks on risk-reducing behavior, suggest that the general population will change behavior substantially. This makes the promise that great health benefits and cost-reductions are just around the corner quite dubious. Advocates of this “surveillance medicine” suggest that even more monitoring and constant feedback will change this picture, but a recent randomized controlled trial performed by researchers at Scripps suggested no differences in healthcare utilization or costs from an intervention employing different monitoring devices in people with common, chronic disease. However, advocates of so-called P4 medicine (predictive, preventive, personalized, and participatory) seem to continue with their bold promises as if this evidence is not there, suggesting that they are partly based on faith. Jerry Yeo: I believe that, in general, patients who have paid for an assessment of their health are more likely to listen to their coaches and to alter their lifestyles towards better health. However, there may be some variability in responses that is dependent on the level of difficulties in lifestyle alterations to achieve the intended outcomes. More motivated individuals are more likely to follow through while less motivated ones may find the prescribed changes too difficult to achieve. Thus it is a multifactorial consideration, including the patient's perceived risk of not following the recommendations based on the screening results. Stephen Master: Making lifestyle changes can be hard, and even direct advice from a physician is no guarantee that individuals will follow basic, well-established diet and exercise guidelines (I do not exempt myself from this difficulty). On the other hand, there are clearly patients who have had, for example, clinically significant cardiac events and have made radical behavioral changes as a result. Since detailed monitoring leads to personalized rather than general advice, I think that it is worthwhile to try and identify those patients who may benefit from the motivation that is provided by wellness data. Is there the potential that findings from such programs could be harmful to the individuals and in what ways? Nigel Paneth: A subset of patients willing to be monitored will have their health concerns exacerbated and may make untoward demands on the healthcare system. There is also the very real risk of excessive intervention, either for diagnosis or therapy, with the possibility of side effects. Every practitioner has seen examples of this, which multiply in proportion to the extent of testing of asymptomatic patients. To quote another internist friend of mine, “no one is really healthy—they just haven't been tested enough.” Henrik Vogt: Yes. Firstly, there are false-positive test results indicating that something is wrong when it in fact it is not. Secondly, there is overdiagnosis, meaning findings that represent a real abnormality, but that will actually never pose an actual health problem. Such findings can cause unnecessary worry and initiate a spiral of further testing and treatments with potential side effects and costs. It is hard to know who is overdiagnosed at the point of diagnosis, meaning who with an abnormality will actually suffer and who will not. Theoretically, these people can only lose from diagnosis and treatment. While concerns about anxiety as a result of predictive testing may not be sufficiently supported in isolated case control trials, screening may still instill a more general awareness of risk or suboptimal health in the population that is supposed to be alleviated by this testing. Jerry Yeo: Practitioners of laboratory medicine are well versed with the concept that the more you test (as in health assessment screening) a healthy population, the higher the risk of getting false-positive results. False-positive results can provoke anxiety and depression, leading individuals to seek medical help, which undoubtedly results in more diagnostic tests that are costly and may be unnecessary. Others may choose to believe the screening results and immediately seek treatment to ameliorate a perceived pathological condition. Additional testing and treatments may have associated risks and side effects that could lead to physical and psychological harm to the individual. Stephen Master: For low-prevalence disease, a test needs extraordinary performance to provide any predictive value. Given the reported deficiencies within our own medical community when it comes to understanding positive predictive value, I am concerned that including more rare diseases within a “wellness profile” would be actively misleading. While it is possible that this problem could be reduced through an intensive focus on reporting tools that help patients understand such data, I think that the near-term risks of misinterpretation need to be carefully considered. Should participants worry about breach of privacy from constant monitoring through smartphones? Nigel Paneth: In today's world, one must be worried about breaches of privacy every time one makes a connection with any electronic device. Henrik Vogt: Yes, information is volatile and there is no way one can be completely sure that it does not end up in the wrong hands, especially in countries with poor respect for individual freedom and human rights. To speculate about the future: Our society is concerned about 2 major fears or risks, terror attacks and the failure of our own bodies. A confluence of these 2 fears and the technologies we have developed for prediction and control of them may lead to the ultimate societal surveillance. In the future, what may be hacked from our smart-phones will not be a few data, but a computational avatar based on big data about most aspects of your being (“avatar” is actually a metaphor used by the European Digital Patient project to describe a future virtual mirror image of all of us). A state or company that wishes to monitor people, for example with reference to risk of terror, could access such an avatar containing our physiology, sociometrics, psychometrics, and environmental exposures. Jerry Yeo: Absolutely! Anyone who uses smartphones or wireless devices should be concerned about the risk of being hacked. Constant monitoring of health data via the Internet or wireless devices can be intercepted by unscrupulous parties, who can then use the information for various nefarious purposes, ranging from blackmail/ransom to theft of personal identity. Determined hackers will attempt to steal these data (especially from celebrities or government officials) and either leak embarrassing information or use them inappropriately for commercial purposes. Stephen Master: There are substantial privacy implications in at least 2 areas. First, the inherent volume of data may reveal fundamental aspects of the individual's health that would not be gleaned from a simple sample at one time point. Second, location, activity, and health data may reveal non–health-related aspects of a person's life (for example, when they typically leave their house to go running). Societal views on the level of privacy risk may many of use the fact that we know that Google our to provide the privacy risks inherent in any electronic monitoring should be clearly out by and there be of lifestyles that are you Nigel Paneth: of the of the monitoring is the time period that is health risks are in very early In for most chronic diseases of in the world, are any clinical The that we can these years they is We are more likely to be to the of clinical disease, as when we or These conditions are often many years the of clinical disease. How in are we likely to lifestyle risk factors of of which we are by using intensive monitoring? We will not early that we know for are major of disease life and the and These factors may not be by participants in the or may not be The to imminent heart attacks not symptoms that would the to a be but we would need to know the screening of such predictive and whether such Henrik Vogt: could no doubt measure substantial aspects of the which by definition every an individual is to from to including and psychological However, the that one can measure every of human which is and by an what in as of It is to go into an about it is impossible to measure However, it should really be the other way who that one may measure every of human are a very bold which also human to something and have a of to Jerry Yeo: I am not a but I would state that, in general, lifestyle such as and and a healthy life of and and can be but in a more The question is how are these data of lifestyle the data are dependent on the construct used by the For example, a construct used in our may not have the meaning in another and it a very different However, there is no doubt that these environmental factors have large effects in to various factors in the health of an individual. Stephen Master: As mentioned the effects of a patient's medical may be difficult to using In a substantial of what is and is not on the that are at any At there is no that I am aware of to blood However, this is something that could be in aspects of an individual's health, such as and behavioral are more difficult to in a “wellness However, to the extent that the aspects of human behavior it may be possible to personal that have an on health. Should public health or this of If would the have an health to those who Nigel Paneth: It should only be by if it is to be First, whether it is then of If it should be by If not it should not be Henrik Vogt: If it and was to be and for it should be It would be that some individuals could better health than from I would worry that a medicine that could provide the with substantial over the poor in of and of life would an that the Jerry Yeo: If there is evidence medicine to better health then it is a that should be by a of public health for those who it and for those who are it will cause a health where the less population sector will have higher risk of health a on public health that will be by society as a Stephen Master: If there are data to it should be is less than treatment. In the case of (and even large there to be an this fact and healthy the of should lead them to this technology if it is to population health. Is there evidence that such initiatives produce more than Nigel Paneth: There are no programs of intensive such as are in the preface to this Q&A which we can determine benefit or it does not seem that programs such as the 100K Wellness and have been up as to produce data. to be If to should the participants be in of the that they will be for and health These are people who will have better health even the I worry that will be made based on the better health of the monitored that are in fact not to the monitoring Henrik Vogt: On the is is the most of human life in As for example, by the by the screening and medicine are at a point where it is whether benefits really the and costs. The promise of P4 medicine is that more medical testing, data, and intervention will produce less and harm in well people. It will take years to if and how that is very little has been from the Wellness Project, it what that is the Jerry Yeo: I think there a of evidence that mass screening of a healthy population leads to more than In there is evidence of the (as to However, with in and the to and big data, it is possible that in the future complex data many factors could be used to generate predictive for individuals that will lead to better health while harmful effects. Stephen Master: not aware of data that this However, the potential for new of data for example, on changes rather than on the to a population and their effects on human health, it is to that to determine the effects of these initiatives would be for Should data from these initiatives be for public Nigel Paneth: The of a to the up in no of the intervention, absence of data on time of me about the of data for any Henrik Vogt: It should be up to the individual However, I find it that advocates of P4 medicine find it to that patients must understand that it is their societal to make their data to for example of future that will be at a loss if they do not. This more it should also be in of the privacy concerns and the fact that the of healthcare is not an but a Jerry Yeo: Yes, public access will further and of knowledge that can benefit society as a However, data in the public could be used by or unscrupulous individuals to make to to the general who may not be to understand the and of the data. with the of on the Internet and the via the public could be by being used to the public into Stephen Master: Yes, with Although the of the data would be by including individual is detailed medical data, it increasingly whether it is still possible to a data not as concerned about being from the many and we on to out from them. more concerned about data with respect to Some may be but is still likely to If not an would you to such programs and could you on your Nigel Paneth: I would not to such a for free or for until it had been to be effective at disease. We are a long way from that not clearly towards a of the of this of Henrik Vogt: in I would to my and on other the But of course, it would be interesting to my own physiology, and to more about how I from a I think many people monitoring their are motivated more by a sense of and than the prospect of better health. The has even the to as a can be from scrutinizing own computational can be seen as a of medical there is a and Jerry Yeo: I would to such a if evidence to the benefits (and of the my health If the screening of risk to my health that could be by changes in lifestyle or being on a I would definitely to In if the screening a genetic condition that I will with no I would such all it does is to cause great psychological any possibility of on to any screening I would to know the of assessment as well as the of the testing Stephen Master: long as there privacy and I was to access the data, To this in I would also volunteer to get my my or any of other biological tests I that this may not change a in my or life on any of this information a healthy of clinical and could be actively to my health. At the end of the I am just
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.104 | 0.223 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.003 | 0.007 |
| Science and technology studies | 0.006 | 0.004 |
| Scholarly communication | 0.011 | 0.014 |
| Open science | 0.005 | 0.010 |
| Research integrity | 0.003 | 0.005 |
| Insufficient payload (model declined to judge) | 0.034 | 0.004 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".