Subscribe to DSC Newsletter

Thermometers and scales to measure weight appeared in retail outlets long ago. Blood pressure monitors perhaps came later. Pedometers and heart-rate monitors seem more recent - possibly closer to my time. I saw several devices while doing this blog intended to electronically record among other things hours of sleep; these are designed to be worn on the body all the time. A couple of weeks ago, I bought something to give the heart rate and blood oxygen saturation level. I consider it a real game-changer due to its ease of use. I just clip it to a finger. I can take readings a few times a day without attracting attention. So it is something "secret." I guess this is the real essence of the change. I control the device. The data belongs to me. I hide it. We live in a world where the average person will soon have access to large amounts of personal health data. This is not data from clinical trials or field studies involving many people. If a person diligently maintains records day after day, he or she would quickly amass more personal health data than any doctor. How can this be given that a doctor has many patients? My emphasis is on the word "personal." A doctor has all sorts of data but little about me specifically. His or her advice is premised on normalcy and uniformity - that I am rather like everybody else. This is how a doctor can give advice without knowing much or anything about my body "specifically." Prior to the development and widespread availability of personal health monitoring devices, a person had little choice but to accept a doctor's authority. In this blog, I don't intend to question medical professionals. But I want to discuss the emerging role of data derived from individual experiences.

Many people do share some physical and mental similarities. For instance, a significant number of people can get to places by walking, made possible of course by their common ability to walk. Buildings can be designed to satisfy their needs; however, those that are mobility impaired might experience difficulties if neglected from the process of budgeting and the allocation of resources. Mass-consumerism is premised on the notion that, despite their individual and possibly unique traits, there are somewhat homogenous segments of the market responsive to mass-products and mass-marketing. In a trip to the Philippines, I can still remember going to a tailor as a child and having clothes specially made to fit me. It seemed like such an interesting concept. Of course these days, most people buy things off the shelf. Businesses are optimized to serve the needs of mass populations. I believe that much the same can be said in relation to the medical industry. The data gathered from people seems to play a role in social engineering: the footprint of any individual must yield to the tracks left by the herd or the herded. Data contributes to the herding process: for instance, it helps to get people into and through the health care system. In particular, it gets the majority of people across by dealing with their most common ailments and health concerns. Data from individual experiences might receive less attention quite simply because it isn't feasible to give more - the same way it is costly and impractical for people to go to tailors to have clothes made. The emergence of personal health data creates potential conflict in relation to the existing power structure. To be an individual is to dispute the value of being part of the collective and those who profit from collective dynamics.

When we have systems designed to serve the majority, it can be annoying to belong to a marginalized or disenfranchised minority. However, in relation to health care, being disadvantaged statistically is particularly unsettling. Consider a scenario for instance where there is a distribution of residual genes from different types of humans in our current make-up. I know this must seem like a random interjection of sorts. I heard this argument on a Canadian radio program a few years ago. I'll share it with readers. Apparently, some people have more Neanderthal in them than others depending on their ethnic background. There is a genetic footprint from early humans to us today. The radio program mentioned other types of early humans that continue to persist in our genes.  The commentator suggested that the interaction between these different backgrounds can lead to competitive advantages and disadvantages depending on circumstances. I think a way of committing genetic "genocide" so to speak and ensuring purity is to impose a bias towards a normative. For instance, a health care system could be optimized to ensure that the healthiest "white people" live comfortably. I have the term in quotes since there is no single type. I am just making an argument. A "black person" might register as a sickly white person. I fall below the normative on several fronts. I'm short. I have flat feet. My eyesight is terrible. I snore. I learned to cope with my peculiarities. But compared to other people, I am not necessarily the finest human specimen. I wonder if a science can develop around my data as an individual. In environmental studies, there is the concept of a monoculture where homogeneity can sometimes exist in species. When a disease affects one person, it affects everybody due to their uniformity. Social engineering promoting the normative can compromise the viability of the species.

I mention in a number of blogs that I regularly maintain data pertaining to my personal health. I use a research prototype called Tendril, a program that is running this moment as I write. One of the joys of using this prototype is the relative absence of normatives. I guess such a statement is easy to misinterpret. Of course, I want high levels of oxygen in my blood. I want my weight to be as low as possible. I want my breathing to be clear and effortless. I would describe these as contextual references for the proper "direction" of data but not any specific "level." For example, I don't have a preset weight target. I am not trying to live up to an externally defined standard. I don't know what my ideal weight should be given my personal situation. Perhaps it takes a tremendous amount of discipline and faith to allow an algorithm to determine ideal metrics such as weight. I am committed to developing a science around myself. It is a science of me. But I am not actually making medical decisions. Over the quotidian routines of life, a person might make choices that some would say reflect personal tastes and preferences. I do certain types of exercises. I eat particular kinds of foods. I am fond of certain activities. Perhaps many people know what makes their own bodies happy. I sometimes don't. I use a computer program to build a database. I actually don't know the reasons behind the lists generated by the prototype. But I take the items appearing on these lists seriously. I am particularly observant of everyday items that are flagged as potentially hazardous. I'm not saying that the items are hazardous in relation to the normative majority. I don't have anything to say about the majority. The data is about me and my experiences.

Algorithmic Man - Unplugged

I think sometime in April of this year I wrote about how my life has become influenced by algorithms. I have become a product of mathematics. I present above some pictures of me over a 10-year period. The lower right image was taken just last month; it presents a version of me that interacts routinely with algorithms. There is a fair amount of technology supporting my day-to-day routines. I'm uncertain if the weight loss is apparent. I am about 30 pounds lighter compared to the preceding two snapshots. I don't credit my weight loss to algorithms mind you. I believe that my technology has helped me keep the weight off. Just this morning I ate a couple of chocolate bars. I had fried chicken this evening. I sit around behind a computer all day. So I need all the help I can get. I would say that my algorithmic contrivances have given me "true freedom." I manage the parameters of my computer programs. I let the software do the rest of the work. I spend time often wondering what the parameters should be. I think about what I would like to get out of myself psychologically, physically, and socially. "How would this person best fit in his current environment?" is the sort of question that I would ask myself. A person cannot physically evolve to adapt to the environment; but he or she can make adjustments to behaviour and decision-making.

I take a blood pressure reading every day usually late in the evening. So far, I have about 227 days of blood pressure data. I'll share some details on how the system operates to demonstrate my own management approach to personal health data. The worst reading occurred on June 24, 2014: 135 / 90. I think most people would consider this a high reading although not terribly so. The lowest reading occurred on May 13, 2014: 95 / 70. The distribution on the illustration below shows that my blood pressure fluctuates quite a bit from day to day. The numbers on the y-axis are not pressure readings but gradient values: 0 indicates the worst day; and the reading just about 100 is the best. Taken as a whole, there seems to be some improvement. In fact, the distribution appears to be lifting off the bottom the chart. For technical reasons that will be explained shortly, it is inappropriate for me to have as my ultimate objective "making the pressure as low as possible." Tendril is designed to let the pressure decline to an ideal point given the systemic impacts to other physiological metrics. I actually don't know what the "ideal blood pressure" for my body is, assuming it is static. My body will make that determination for me. I just have to be observant and sensitive to changes and outcomes. The science I build around my body is different than the science that would be applied to a large number of people. My placement on a statistical distribution is not in itself relevant. I don't have any external ideals. To the extent that it is possible to do so, given my level of personal autonomy, I try to influence the mathematics of my life.

My objective has never been to serve science per se but rather the needs of my body. So I don't deliberately experiment on myself to determine the validity of assertions the way medical researchers have sometimes taken chances or experimented on people. I make routine life choices. Should I have cheesecake or chocolate cake? Should I do sit-ups or crunches? Will I go to the library or to the grocery store? I take some food supplements although surprisingly few rather infrequently. I don't have a stance on supplements: e.g. "I make sure to take my supplements" or "I never take supplements." If something seems beneficial, I often take it again. If it seems hazardous and it cannot be reintroduced in non-hazardous terms, I generally avoid it. I would say that most of items I have purchased off-the-shelf have led to negative rather than positive outcomes. This is not to say that the items would be unbeneficial to others. But as I already pointed out - or it should be obvious by now - I keep exceptionally good records. I not only keep track of what I take, in some cases I actually note specific brands. I'm not afraid to say at some point, "It seems I wasted my money on this." However, money is never truly wasted. Determining that something is hazardous or unhelpful is valuable information.

I find myself wishing that I had come up with the technology maybe a decade earlier. Since I don't actually have significant knowledge of human physiology or drug treatments, I have a pure data science approach to personal health data. I can only learn from experience; by discovering things sometimes by accident; by making mistakes; and of course by guessing and deliberately taking risks. In some cases, a researcher might be able to analyze data without ever being close to its source. For instance, the data might be electronically conveyed from a remote location to a central research facility. I notice that at times, data scientists in particular seem quite concerned about the pace of disease; predicting the impacts of outbreaks; finding efficiencies in how people might be handling disease. The role of the data is different once disease becomes a lived experience. One starts to consider cures, remedial actions, and even lifestyle changes. The idea of publishing a paper or developing a drug treatment doesn't really enter the picture. So in a manner of speaking, the increasing availability of data to those most affected can help to restore the original purpose of medicine. It isn't to attract funding, bolster careers, or perpetuate industries. The original idea is to save lives. A life worth living is certainly a life worth saving.

Passion and Estrangement

I remember working as a summer student for the federal government. A rather frantic elderly gentleman came to my desk saying that he had to speak with somebody right away. I asked him how I could be of assistance. He said that he had a fantastic idea to remove contaminants from groundwater. He proposed freezing and excavating the water. In retrospect, now being aware of the high cost of excavation, the idea is problematic from a purely financial standpoint. Also, moving the problem doesn't actually resolve it; shipping contaminants is not without risk; plus there is the question of where to safely put it. But I thought to myself, I was in the presence of a highly motivated individual, practically evangelical in this convictions and beliefs. His ideas didn't exist only on paper. He lived his ideas; maybe he event dreamt about them. He was obsessed. I almost didn't have the heart to tell him that I was just a summer student. I didn't even know how to get to the bathroom let alone direct him to the most appropriate department. There is normally a separation that occurs when a person is detached from the underlying phenomena: it is a kind of data detachment or estrangement. One becomes focused on departmental objectives or what his or her profession might have to say. The terrain changes when the analyst is much closer to the phenomena - living within and enduring the data. I think we can expect more passion. Perhaps, there is a need for that fire to burn bright in people sometimes. So I felt myself privileged to meet somebody passionate about his ideas. People see different things in the numbers - see themselves playing different roles. It's like encountering a door to a different universe.

I mention the job for another reason. I was hired to set up a database to help the government with a program to clean up contaminated sites. I went through the process of survey development, distribution, database construction, and data collection. I trained a couple of people to handle the data-entry. Some might wonder what any of this might have to do with personal health. This is my main point, actually. I don't want it to seem like I'm preoccupied with medical professionals. Taken as a whole, these individuals are close to their markets and the people that they serve. But let's be open-minded and consider the clean-up of contaminated sites as a form of health care - because site contamination can lead to real-life health consequences. I was working on my degree in environmental studies. My personal focus was on public impacts. I seemed to me at the time that a $500 million program could have many positive "health" outcomes. I knew for instance that native communities sometimes came across issues relating to higher mercury levels in fish and unsafe drinking water supplies. One way to assess contaminated sites is in relation to their health consequences. I was naturally inclined towards this perspective. However, the department that I worked for was much more engineering-oriented. It is possible to allocate funds based primarily on containment and mitigation objectives - not necessarily taking into account local populations and health impacts. In other words, it is possible to examine a health care issue in purely engineering terms. As I have discovered more recently, health care can even be evaluated in completely budgetary and administrative terms.

If the data driving the allocation of funds were closer to those affected, the purpose of funding and its expected outcomes might be portrayed or interpreted differently. In my summer job, I was part of a big wheel mostly to help initiate the process of first-contact. My role was to compile a database of clean-up companies, their technologies, and capabilities. Those listed on the database represented the "target market" for the department. I don't want to over-generalize a complex situation; but I think that it is sometimes possible to do a lot, spend a lot, and in the end accomplish much less than expected. This can happen when there is an attempt to centralize problem-solving through the distribution of capital. The success of funding is inferred through the compliance of these disbursements to criteria. A program can be designed to deliver an institutional response. I would say this is true both from the standpoint of management and data collection. Bureaucratic agencies in particular - and I don't simply mean government - can carry out their affairs without ever considering the underlying impacts of their operations. They might not be judged by real or tangible success. I sometimes find myself among the first to criticize "neoliberal ideals" - such as those pertaining to performance measurements and competition. But I want to make it clear that my concern is more related to the methodology than intent. In fact, in a world of scarce resources, we should indeed be concerned about the effective use of capital. There are responsible and irresponsible ways of using capital.

I am therefore suggesting that effectiveness of health expenditure is achieved to some extent by reducing the distance between those portrayed in data and those collecting the data. Bringing personal health data closer to the average person is part of a much broader movement of liberalization decades in the making. I would argue that recent developments bringing data and people closer together represent aspects of our social evolution. Nobody is forced to remain compliant members of society. It makes a lot of sense to participate in the benefits. But those benefits are premised on institutional sensitivity to individual needs. As participation levels increase and there is a need to gain efficiencies, I believe that the risk of structural estrangement becomes more relevant; this can contribute to radicalization and loss of social commitment towards common goals and values. It might for instance become more difficult to fight domestic terrorism if people stop caring, or if they cannot meaningfully participate in data. However, even if one dismisses the risk of alienation, society must still confront the decline of public resources to support "diverse needs." As efficiencies increase, I suggest that the diversity of the needs served start to decline; and people become oppressed not so much by ideology but scarcity. The oppression starts by distancing the beneficiaries of society from the data intended to represent their interests.

Lens Leading to the Individual

In my personal health database, blood pressure is one of many concerns that I follow. This is not to say that my blood pressure has ever been much of a concern. I decided to be proactive in light of some family history. When data emerges from individual experiences, I get to decide which experiences to follow. Researchers might be interested in aspects of people related to specific areas of research and perhaps expertise. I don't diminish the importance of their concerns except that mine might differ. I sometimes wonder about the merits of certain areas of research; and I periodically question whether public funds are being used wisely. When the data is quite close to me, I am interested in all sorts of issues that never really get much public attention. Below, I provide a short listing of the main areas of improvement made possible through one's closer proximity to data.

1. Greater Expressiveness

I will share some of the more interesting health concerns that I follow on my research prototype: 1) enthusiasm and positive outlook; 2) inner peace; 3) ability to speak coherently; 4) sugar drain - e.g. felt shortly after eating a chocolate éclair; 5) something called "luck advantage"; 6) a condition that I describe as funny elbow - no need to elaborate; 7) head aches; and 8) sensitivity to humming noises or some might call it tinnitus. Old electronic equipment such as transformers can actually wake me up if left on. Or I might detect high-pitched sounds in particular locations. I understand that exposure to certain work environments and medications can lead to hearing abnormalities. I don't actually try to rationalize the causes. I currently follow about 100 different health contexts. I am really interested in my quality of breathing perhaps because some family members had issues. I incorporate their experiences into the construction of my health contexts, which in turn power Tendril's algorithms. My point really is that I have control over what gets included. It is a control that would not exist if I left it to chance in the hands of the broader research community. I have data, technology, and more choices.

2. Data Rainforest

When Sean Connery appeared in the movie Medicine Man, he was trying to protect a rare plant species that he thought could protect people from cancer. Of course, many natural habitats are being destroyed rapidly in order to accommodate competing demands such as land to raise cattle and for cities. I believe that when data is placed in the hands people - their own data in their own hands - some percentage will become creative in ways that cannot be anticipated. Fostering a movement to bring data closer to people is about tapping into power and imagination. The more inclusive we are as a society, the greater the potential benefits. But this isn't just a search for solutions. We create roads to help solutions rise up from the deeps of society. I remember Scrooge dealing with the ghosts of Christmas past, present, and future. I think that in many respects, there are always intercessors. Societies and organizations are given their fair share of creative people. But if left in a setting where they cannot contribute their talents, we sometimes miss opportunities that can never be recovered.  If deprived to tools and data, event their talents lose relevance.

3. Increased Responsiveness

Apart from controlling health contexts, I control what events from my day-to-day life get included as data. I don't regularly take any medication. I keep track of things like espresso, cookies, eggs, lotions, soaps, chocolate, chicken, and nuts. I note my exercises. Perhaps, more important than what I take and do are the things that I don't take and do. Because my life generates the data, I can be specific and spontaneous the way no outside researcher can be. The data can be whatever comes to mind or seems worth recording at the time. The ease at which I include all sorts of events is related to the technology that I use. However, the technology behaves the way it does because of my personal needs. My individual requirements have given shape to the database and its environment. The system might be completely different in the hands of a researcher motivated by departmental and professional priorities.

4. Capturing Systemic Benefits

For me, the term "systemic" is a specific algorithmic concept that can be incorporated into a computer program. As I mentioned earlier, I follow about 100 different health contexts. Sometimes, attempting to promote one health objective can detract from another. I can give an example just off the top of my head although it's not one that has affected me: a person might choose to lose a lot of weight in a short period of time without considering the immediate adverse impacts. Drastic weight loss can cause loss of energy and maybe even disorientation. On the other hand, he or she can focus on events that result in broad systemic improvement. I believe that access to systemic benefits increases when one is close to the data; this is due to the ability to rapidly detect and connect the impact of events to different health contexts. I call this the "Push Principle," and it is an important aspect of the prototype that I use. Indeed, when I decide to avoid something, it probably had a negative impact on me in a number of different ways. The things that I decide do often have many positive benefits. There is a kind of liberation that emerges from these dynamics - knowing that the technology is working on a systemic level - for it reduces the likelihood of unintentionally or accidentally causing harm.

5. Capturing Specific Benefits

I remember in introductory environmental toxicity being introduced to a measurement called the LD50 (the lethal dose that kills half the population in a sample). The rationale pertains to statistical distributions. Alcohol for instance is quite toxic - according my toxicity professor - in light of how the body responds to it. A certain segment of the population would experience adverse impacts right away even if just a small amount of alcohol is consumed. Some people are able to drink large amounts without feeling or getting ill. Then there is a spectrum of people between the extremes. Well, the same principle can be applied in relation to all sorts of substances and perhaps even life events. For instance, for some kids, certain interactions might not seem like bullying at all; but for others even the slightest conflict might qualify as bullying. In relation to health treatments, attempts are likewise made to arrive at generalizations. We get all sorts of debate for instance surrounding the question of whether or not coffee is good for "people." The idea of coffee being terrible for some people and great for others doesn't seem to fit the discourse. The individualization of health seems unprofessional perhaps because it is difficult for a profession to express an opinion in relation to particular individuals. Here then we encounter something of a structural limitation on the value of an aggregate perspective. Reducing the distance between people and their representation in data makes the data more relevant to the lives of individuals.

6. Departure from the Mass Consumerism

There is no shortage of talk on ways to make money from data. I believe that some ideas presume that similarities exist between large numbers of potential clients: i.e. there are untapped mass markets. In such a scenario, a major player in the data industry might attempt to take an entire market. Let's consider the possibility that clients might not be particularly similar. Further, to the extent that they are uniform, their common needs might already be served by existing products and services; this leaves only those opportunities related to areas of individualization. I believe that opportunities do most certainly exist, but development will emerge in response to the data. Normally a database gets constructed after a mass-market product is acquired: the data-gathering infrastructure along with the actual data collected must conform to the specifications and capabilities of the product. But I believe that in the future, some practitioners in the area of health might not be medical doctors at all but rather data specialists focused on the individual circumstances of clients. This is not to say that such specialists will give medical advice. They might help clients engage complex situations that must of course incorporate the advice of medical professionals.

7. Less Cost

I have faith in doctors in relation to certain types of problems. I believe however that lifelong health decisions generally fall outside their scope; and if these things were within their scope, there might be some reluctance among sponsors and insurers and pay for the costs. For example, should a person ride a bicycle? It is healthy to do, a doctor might say almost instinctively. That's not really true in a busy city. Riding a bicycle is quite dangerous if we consider pollution, cars, and poor road conditions. So will anybody pay a doctor to give advice about riding a bicycle? How about spending time on a computer; commuting to work; working as an accountant; wearing loose clothing for the winter? Long before situations in the lives of people become clinically relevant, there might be all sorts of important day-to-day life decisions. It is just extremely inconvenient and outside the scope of the medical profession to cover the quotidian aspects of life. I would say that at this time, nearly all of the data resources are mass public resources (such as the internet) and therefore somewhat ineffective to deal with highly individual problems. The move to bring data closer to the people sets the stage for more intervention opportunities.

8. Respect for the Distance

Around the world, there has been more collection of personal information by government agencies. Normally when people spend a lot of time watching reality television shows, we say that they might be wasting their time. We have been reducing the distance between governments and the data of its people; but the distance is increasing between people and their data. When 911 occurred, the lasting impact was on US intelligence. It was a successful attack on the support structures for metrics and detection. So what does this have to do with personal health data? Well, personal health data cannot fight terrorists or indeed many other things; but it like anything else can be configured to give governments an apparent mandate. That mandate starts to lose potency when the data becomes more closely connected to the individual; for there would be fewer aggregate details to extract. My rather abstract argument here is that when the distance between data and a regulator declines (the projection of proxy), the distance between the data and those being regulated increases (the articulation of phenomena), until the data itself become estranged and irrelevant. This is a principle relating to social disablement, which I believe is one of the most important issues pertaining to the use of massive data today. It is a complex ontological concern that I briefly touched in my introduction of the Universal Data Model (the "exclamation model") shown below. I will elaborate on this model in the future.

Liberty - Taking It Personally

In this blog, I juxtaposed between the alienated administrative use of data affecting public resources and the personal lived experiences of people. I said that health can be regarded purely from the perspective of departmental or professional needs, which can limit or place constraints on the meaning of the data and its relevance to people. I was once asked to come up with different ideas to quantify risk. There was some interest expressed not just in the cost of cleaning up contaminated sites but also taking into account softer issues such as liability - if this can be described as a soft issue - politics and other impacts. Some might be surprised to discover that the project was initiated and funded by the military. I think as a matter of survival, some agencies are at times forced to gather and consider diverse data in their assessment of risky situations. More often than not, the lives we lead tend to be estranged from the data regulating our day-to-day affairs. Others make decisions for us. They don't just perform a service. They make it possible for us to give up control.

Others determine our placement, the value of our involvement, really our very existence in the data that is gathered; for the data can exist without ever taking our needs into account. One of the reasons why it is so difficult to quantify different aspects of risk relates to the phenomenological nature of impacts and consequences. In this case when I mean phenomenological, I mean highly personal. This isn't simply a disposition that can be changed on a whim. Some of us literally have a low tolerance for toxicity while others can stand a great deal. Some people can fight through social barriers while others collapse under its weight. Phenomenology is not merely emotional but also mathematical - something that can be detected in the statistical distribution of internal responses to external stressors. Moreover, tolerance to external stress is transient and dynamic; it is dependent to some extent on the different support structures that exist in our lives at the time. The erosion of such supports can reduce one's tolerance.

I suggest that austerity and efficiency initiatives on a broader societal level can bring about a level of harm and neglect that hasn't been seen in centuries; and this is not a society that can exist in a stable form while being detached from the personal impacts and consequences to people. The administrative, alienated, and impersonal use of data can lead to adverse health consequences. If the idea is to cut back on the social support, it is necessary to return the power that people have over how they are manifested in their data. But apart from adapting to austerity, bringing data closer to the people is a way of restoring balance and preserving personal liberties. I believe this is the movement before us today and the challenge we are now facing.

Views: 330

Tags: administration, algorithms, alienation, applications, approaches, articulation, care, computer, experiences, government, More…health, management, military, models, optimization, participation, personal, phenomenology, planning, programs, projection, public, reification, services, strategies, studies, systems

Comment

You need to be a member of Data Science Central to add comments!

Join Data Science Central

© 2019   Data Science Central ®   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service