The Case for Technobiology: A Complement to Biotechnology

2.1 Integrated

Unfortunately, in spite of the provisions included in Public Law 111–148, the major healthcare reform bill signed by President Obama on March 23, 2010, the US healthcare remains in disarray — Essentially, a non-system. The US healthcare is a good example of a people-intensive, expensive and, yet, among the least effective system for a developed country; a minority of the population receives excellent care, while an equal minority receives inadequate care^[4]. This situation is not due to a lack of well-trained health professionals or to a lack of innovative technologies; it is due to the fact that it is based on a fragmented group of mostly small, independent providers driven by cost-obsessed insurance companies^[5]. As a consequence, an integrated healthcare system must be developed and implemented, one requiring the participation and support of a large number of stakeholders (i.e., consumers, doctors, hospitals, insurance companies, etc.). For example, patients must take increased responsibility for their own healthcare in terms of access to and use of validated information. In sum, U.S. healthcare must become a co-produced, technobiology-oriented system.

Moreover, by employing the integrative approach of systems engineering, the US healthcare imbalance can be better equilibrated, resulting in a more efficient and effective system. In this regard, payment for care should be more weighted to output or value provided than to input or activities undertaken. Additionally, such production techniques as cross-training and multi-use of facilities can enhance both efficiency and effectiveness. All of these issues must be taken into consideration as President Obama and Congress seek to improve US healthcare; public plans (e.g., Medicare and Medicaid) must be integrated with a plethora of private plans (that are, unfortunately, neither affordable nor portable) in order to achieve healthcare coverage for all US residents.

As mentioned earlier, a service system like healthcare is actually an integration or combination of three essential components — People, processes and products. In particular, people can be grouped into those demanding services (i.e., consumers, users, patients, buyers, organizations, etc.) and those supplying the services (i.e., suppliers, providers, clinicians, servers, sellers, organizations, etc.); although sometimes ad hoc, processes are primarily procedural (i.e., standardized, evolving, decision-focused, network-oriented, etc.) and/or algorithmic (i.e., data mining, decision modeling, systems engineering, etc.) in structure; and products can be physical (i.e., facilities, sensors, information technologies, etc.) or virtual (i.e., e-commerce, simulations, e-collaboration, etc.) in form.

Given the co-producing nature of services, it is obvious that people constitute the most critical component or element of a service system. However, because people are so unpredictable in their values, behaviors, attitudes, expectations, and knowledge, they invariably raise the complexity of a service system. Moreover, the multi-stakeholder - and related multi-objective - nature of such systems serve to only intensify the complexity level and may render the system to be undefinable, if not unmanageable. Human performance, social networks and interpersonal interactions combine to further aggravate the situation. Other barriers to co-production are considered by Parasher et al.^[6]. Processes that underpin system integration include standards, procedures, protocols, and algorithms. By combining or integrating service processes, one could, for example, establish a “one-stop shopping” approach, a highly desirable situation for the consumer or customer. However, integration of financial services has resulted in giant banks (e.g., Citigroup); integration of home building goods and services has resulted in super box stores (e.g., Home Depot); and integration of software services has resulted in complex software packages (e.g., Microsoft Office). Integration also enhances system efficiency, if not its effectiveness. For example, the radio frequency location and identification (RFLID) tag — or computer chip with a transmitter — serves to integrate the supply chain. Unfortunately, as supply chains become more electronically integrated and interdependent, cyber security becomes more problematic.

As indicated in Table 3, service system integration can occur over the physical, temporal, organizational and functional dimensions. Physical integration can be defined by the degree of systems co-location in the natural (i.e., closed, open, hybrid), constructed (i.e., goods, structure, systems) or virtual (i.e., service, simulated, e-commerce) environment. An urban center’s infrastructures (e.g., emergency services, health services, financial services, etc.) are examples of a constructed environment. Over time and with advances in information technology and the necessity for improved efficiency and effectiveness, these infrastructures have become increasingly automated and interlinked or interdependent. In fact, because the information technology revolution has changed the way business is transacted, government is operated, and national defense is conducted, the US President^[7] singled it out as the most critical infrastructure to protect following the disaster of 9/11. Thus, while the US is considered a superpower because of its military strength and economic prowess, non-traditional attacks on its interdependent and cyber-underpinned infrastructures could significantly harm both the nation’s military power and economy, including that of the healthcare sector. Clearly, infrastructures, especially the information infrastructure, are among the nation’s weakest links; they are vulnerable to willful acts of sabotage, if not invasions of privacy. Moreover, this interdependency contributed significantly to the 2008 economic disaster or recession. Recent efforts at imbuing infrastructures with “intelligence” make it increasingly feasible to address the safety and security concerns, allowing for the continuous monitoring and real-time control of critical infrastructures.

Temporal integration can be defined by the degree of systems co-timing from a strategic (i.e., analytical, procedural, political), tactical (i.e., simulation, distribution, allocation), and operational (i.e., cognition, visualization, expectation) perspective. Expectation, for example, is a critical temporal issue in the delivery of services. More specifically, since services are to a large extent subject to customer satisfaction and since, as Tien and Cahn^[8] postulated and validated, “satisfaction is a function of expectation,” service performance or satisfaction can be enhanced through the effective “management” of expectation. When applied to healthcare, however, it may be difficult, if not impossible, to manage a patient’s expectation under certain emergency situations.

Organizational integration can be defined by the degree of systems co-management of resources (i.e., people, processes, products), economics (i.e., supply, demand, revenue), and management (i.e., centralized, decentralized, distributed). In regard to management integration, Tien et al.^[9] provide a consistent approach to considering the management of both goods and services — By first defining a value chain and then showing how it can be partitioned into supply and demand chains, which, in turn, can be appropriately managed. Of course, the key purpose for the management of supply and demand chains is to smooth-out the peaks and valleys commonly seen in many supply and demand patterns, respectively. Moreover, real-time mass customization occurs when both supply and demand chains are simultaneously managed. The shift in focus from mass production to mass customization (whereby a service is produced and delivered in response to a customer’s stated or imputed needs) is intended to provide superior value to customers by meeting their unique needs. It is in this area of customization — where customer involvement is not only at the goods design stage but also at the manufacturing or co-production stage — that services and manufacturing are merging in concept, resulting in a “servgood” (which is further defined in Section 3).

Functional integration can be defined by the degree of systems co-functioning in regard to input (i.e., location, allocation, re-allocation), process (i.e., informatics, feedback, control), and output (i.e., efficiency, effectiveness). From an output perspective, for example, it is obvious that a system should be about integrating and enhancing efficiency and effectiveness, the twin pillars of productivity. However, it should be noted that efficiency is mostly related to the manner in which the supply chain has been designed for optimal operation, while effectiveness is mostly related to the manner in which the demand chain has been responsive to consumer needs.

Again, healthcare — as a service system — must be integrated in regard to people, processes and products, as well as over the physical, temporal, organizational and functional dimensions. It is obvious that designing an efficient and effective healthcare system is not easily accomplished; socialistic systems like Sweden’s cost too much, while capitalistic systems like the US’s are both high cost and not fair. New design approaches are required. The information technology revolution has permitted the analysis part of system design to be largely undertaken by computers; it allows for a simulated and collaborative redesign process to occur — until a satisfactory design is achieved which meets specified performance (e.g., morbidity, mortality, cost, etc.) criteria. The resultant and integrated healthcare system must be a comprehensive, interoperable system of systems. Perhaps the best example of an integrated healthcare system is that proposed by Goldschmidt-Clermont et al.^[10]; based on the autonomic computing initiative, they propose an autonomic healthcare system that combines the existing hospital information technology with operational processes to bring down the barriers among different specialties and to improve the quality of care being provided. As a final point, it should be noted that the human body is itself an amazingly integrated system of systems, one that should not, if possible, be perturbed by surgeries and other intrusive treatments; perhaps the clinical specialist that should command the highest salary is one who focuses on the body as a system of systems.

2.2 Adaptive

Because a service system is, by definition, a co-producing system, it must be adaptive. Adaptation is a uniquely human characteristic, based on the technobiology approach of decision making. (Indeed, designing a healthcare system is about making decisions or choices about the system’s characteristics or attributes.) Tien^[11] underscores the difference between data and information, especially from a decision making perspective. Data represent basic transactions captured during operations, while information represents processed data (e.g., derivations, groupings, patterns, etc.). Clearly, except for simple operational decisions, decision making at the tactical or higher levels requires, at a minimum, appropriate information or processed data; moreover, electronic data technology has transformed large-scale information systems from being the “glue” that holds the various units of an organization together to being the strategic asset that provides the organization with its competitive advantage.

The fact remains that data — both quantitative and qualitative — need to be effectively and efficiently fused and analyzed in order to yield appropriate information for informed or intelligent decision making in regard to the design, production and delivery of goods and services, including healthcare. At present, decision modeling methods suffer from two shortcomings. First, most of the available — especially optimization — methods are only applicable in a steady state environment, whereas in the real-world, all systems are in transition. (Note that steady-state, like average, is an analytical concept that allows for a tractable, if not manageable, analysis.) Second, most of the available methods are unable to cope with changing circumstances; instead, we need methods that are adaptive so that decisions can be made in real-time, as is required in most healthcare situations. Thus, non-steady-state, adaptive decision methods are required. More importantly, real-time decision modeling is not just about speeding up the models and solution algorithms; it, like real-time data fusion and analysis, also requires additional research and development.

Again, the real-time nature of co-producing services — especially human-centered services that are computationally-intensive and intelligence-oriented — requires a real-time, systems engineering approach. Ethnography, a branch of anthropology that can help to identify a consumer’s unmet needs, is being used to spot breakthrough product and service innovations. Another critical aspect of systems engineering is system performance; it provides an essential framework for assessing the decisions made — in terms of such issues as satisfaction, convenience, privacy, security, equity, quality, productivity, safety and reliability. Similarly, undertaking systems engineering within a real-time environment will require additional thought and research.

Human interface is another essential element of an adaptive service system; it is actually a critical tool in systems engineering. Such interface could include the interactions between and among humans and software agents, machines, sub-systems, and systems of systems. Human factors constitute a discipline that deals with many of these interactions. However, another critical interface concerns how humans interact with data and information. In developing appropriate human-information interfaces, one must pay careful attention to a number of factors. First, human-information interfaces are actually a part of any decision support model; they structure the manner in which the model output or information is provided to the decision maker. Cognition represents the point of interface between the human and the information presented. The presentation must enhance the cognitive process of mental visualization, capable of creating images from complex multidimensional data, including structured and unstructured text documents, measurements, images and video. Second, constructing and communicating a mental image common to a team of, say, clinicians and nurses could facilitate collaboration and could lead to more effective decision making at all levels, from operational to tactical to strategic. Nevertheless, cognitive facilitation is especially necessary in operational settings which are under high stress. Third, cognitive modeling and decision making must combine machine learning technology with a priori knowledge in a probabilistic data mining framework to develop models of, say, a nurse’s tasks, goals, and objectives. These user-behavior models must be designed to adapt to the individual decision maker so as to promote better understanding of the needs and actions of the individual, including adversarial behaviors and intents.

More importantly and as detailed in Table 4, service system adaptation can occur over the monitoring, feedback, cybernetic and learning dimensions. Monitoring adaptation can be defined by the degree of sensed actions in regard to data collection (i.e., sensors, agents, swarms), data analysis (i.e., structuring, processing, mining), and information abstraction (i.e., derivations, groupings, patterns). Data are acquired by sensors, which could be in the form of humans, robotic networks, aerial images, radio frequency signals, and other measures and signatures. In regard to patients, for example, sensors which monitor their vital signs are essential, as are verbal inputs from the patients themselves. More recently, data warehouses residing in the cloud are proliferating and Big Data mining techniques are gaining in popularity. However, no matter how large a data warehouse and how sophisticated a data mining technique, problems can occur if the provided data do not possess the desired attributes of measurability, availability, consistency, validity, reliability, stability, accuracy, independence, robustness, completeness and timeliness.

Feedback adaptation can be defined by the degree of expected actions based on standardized (i.e., pre-structured, pre-planned), procedural (i.e., policies, standard operating procedures), and algorithmic (i.e., optimized, Bayesian) approaches. In general, models underpin these approaches. As an example, Kaplan et al.^[12] have developed a set of complex models to demonstrate that the best prevention approach to a smallpox attack would be to undertake immediate and widespread vaccination. Unfortunately, models, including simulations, dealing with multiple systems are still relatively immature and must be the focus of additional research and development. Such system of systems models are quite complex — especially in regard to a possible misalignment of the underlying system objectives — and will require a multidisciplinary approach.

Cybernetic adaptation can be defined by the degree of reactive actions that could be deterministic (i.e., known states, deterministic actions), dynamic (i.e., known state distributions, dynamic actions), or adaptive (i.e., unknown states, adaptive actions). Cybernetics is derived from the Greek word “kybernetics”, which refers to a steersman or governor. Within a system, cybernetics is about feedback (through evaluation of performance relative to stated objectives) and control (through communication, self-regulation, adaptation, optimization, and/or management); thus, cybernetic adaptation refers to actions that are undertaken based on an assessment of the feedback signals and then taking corrective steps to control the system so as to achieve the desired system objectives. As an example, autopilots — which are programmed to deal with deterministic and dynamic situations — can, for the most part, take off, fly and land a plane; yet, usually two human pilots are also on the plane, just in case an unknown state occurs and the adaptive judgment of a human pilot is required. Clearly, a trained human — like a clinician or surgeon — is still the most adaptive controller, although machines are becoming more ‘intelligent’ through adaptive learning algorithms.

System control is perhaps the most critical challenge facing designers of multiple, interrelated systems or system of systems (SoS). Due to the difficulty, if not impossibility, of developing a comprehensive SoS model, either analytically or through simulation, SoS control remains an open problem and is, of course, uniquely challenging for each application domain, especially in regard to healthcare. Moreover, real-time control — which is required in almost all application domains — of interdependent systems poses an especially difficult problem. The cooperative control of a SoS assumes that it can be characterized by a set of interconnected systems or agents with a common goal, which unfortunately is not the case with most healthcare systems. Classical techniques of control, optimization and estimation could be used to create parallel architectures for, as an example, coordinating numerous sensors. However, many issues dealing with real-time cooperative control have not been addressed, even in non-SoS structures. For example, one issue concerns the control of a SoS in the presence of communication delays among the SoS sub-systems.

Finally, learning adaptation can be defined by the degree of unstructured actions based on cognition (i.e., recognition-based, behavioral), evidence (i.e., information-based, genetic), and improvisation (i.e., experience-based, evolutionary). Learning adaptation is mostly about real-time decision making at the operational level. In such a situation and as indicated earlier, it is not just about speeding up steady-state models and their solution algorithms; in fact, steady-state models become mostly irrelevant in real-time environments. Instead, it concerns reasoning under both uncertainty and severe time constraints. The rise of cloud computing — whereby the vast array of servers (including laptops and smartphones) can be connected to data and algorithms almost anytime and anywhere — is becoming an essential tool to computational decision making, including improvisation.

At all levels of healthcare decision making, there are a spectrum of possible methods that can be utilized, ranging from adaptive — instead of randomized — medical trials, to autonomous control, to virtual-touch tools, to genetic algorithms, to improvisation, all able to cope with imprecision, uncertainties and partial truth. As an example, shared or informed decision making is becoming more popular and, as a result, patients are electing to have fewer surgeries for clogged arteries (when they are informed that, except for reducing chest pains, drugs are just as effective as angioplasty surgeries — with balloons and stents — in preventing heart attacks and death), prostate cancer (when they are informed that 97 percent of men with prostate cancer die of some other cause), and herniated discs (when they are informed that the outcomes are the same whether they have surgery or not). Moreover, the methods can be used to process information, take into account changing conditions, and learn from the environment; thus, they are adaptive and, to a large extent, responsive to a data stream of real-time input. In a fully integrated and adaptive SoS, each system must be able to communicate and interact with the entire SoS, without experiencing compatibility issues.

2.3 Evidence-Driven

The venue for evidence is increasingly connected with the emergence of e(lectronic)-services; they include, as examples, medical records, financial services, banking, airline reservation systems, and consumer goods marketing. As discussed by Tien and Berg^[13], e-service enterprises interact or “co-produce” with their customers in a digital (including e-mail and Internet) medium, as compared to the physical environment in which traditional or bricks-and-mortar service enterprises interact with their customers. Similarly, in contrast to traditional services which include low-wage “hamburger flippers”, e-services typically employ high-wage earners and services that are more demanding in their requirements for self-service, transaction speed, and computation. In regard to data input that could be processed to produce information that, in turn, could be used to help make informed service decisions, it should be noted that both sets of services rely on multiple data sources; however, traditional services typically require homogeneous (mostly quantitative) data input, while e-services increasingly require non-homogeneous (i.e., both quantitative and qualitative) data input. Paradoxically, the traditional service enterprises have been driven by data, although data availability and accuracy have been limited (especially before the pervasive use of the Universal Product Code — UPC — and the more recent deployment of radio frequency location and identification — RFLID — tags).

Consequently, while traditional services are based on economies of scale and a standardized approach, e-services emphasize economies of expertise or knowledge and an adaptive approach. Another critical distinction between traditional and e-services is that, although all services require decisions to be made, traditional services are typically based on predetermined decision rules, while e-services require real-time, adaptive decision making; that is why Tien^[11] has advanced a decision informatics paradigm, one that relies on both information and decision technologies, in real-time. High-speed Internet access, cloud computing, wireless networks, electronic sensors and ever-smarter software are the tools for building a global services economy. Thus, in e-commerce, a sophisticated and integrated service system combines product (i.e., good and/or service) selection, order taking, payment processing, order fulfillment and delivery scheduling into a seamless system, all provided by distinct service providers; in this regard, it can be considered to be a system of different systems.

Although the comparison between services and physical goods highlights some obvious methodological differences, it is interesting to note that the physical goods depreciate with use and time, while the virtual service assets are generally reusable and may in fact increase in value with repeated use and over time. The latter assets are predominantly processes and associated human resources that build on the skill and knowledge base accumulated by repeated interactions with the service receiver, who is involved in the co-production of the service. Thus, for example, a surgeon should get better over time, especially if the surgical procedure is repeated. Indeed, clinical productivity increases for an average physician, from the dawn of a career to almost the end of a career, with a slight slowing down towards the end. Likewise, while most US physicians practice at a financial loss over the first few years, they progressively improve their financial standing over the course of their career.

Another critical influence on services is the computationally-driven move towards mass customization. “Customization” implies meeting the needs of a customer market that is partitioned into an appropriate number of segments, each with similar needs (e.g., Amazon.com targets their marketing of a new book to an entire market segment if several members of the segment act to acquire the book). “Mass customization” implies meeting the needs of a segmented customer market, with each segment being a single individual (e.g., a tailor who laser scans an individual’s upper torso and then delivers a uniquely fitted jacket). And “real-time mass customization” implies meeting the needs of an individualized customer market on a real-time basis (e.g., a tailor who laser scans an individual’s upper torso and then delivers a uniquely fitted jacket within a reasonable period, while the individual is waiting). To a large extent, one could argue that healthcare is, almost by definition, a real-time mass customized service; however, at the present time, healthcare is not computationally-driven and therefore not cost-effective. Nevertheless, it is interesting to note that in regard to customization and in relation to the late 1700s, the U.S. is in some respects going “back-to-the-future”; thus, advanced technologies are not only empowering the individual but are also allowing for individualized or customized goods and services. Moreover, when mass customization occurs, it is difficult to say whether a service or a good is being delivered; that is, an expert-guided robotic surgery — like a uniquely fitted jacket — can be considered to be a co-produced service and good or, as defined in Section 3, “servgood”.

Clearly, healthcare needs to transition from being a traditional (although high-wage) service to an electronic-based service industry, one relying on digital media for such activities as real-time access to patient records and consultation at-a-distance. Additionally, healthcare must adopt some of the methods that have enabled manufacturing to be efficient (e.g., reduced cycle time, improved quality, etc.), while focusing on service effectiveness (e.g., maintaining a high standard of co-production, meeting patient expectation, etc.). Most importantly, healthcare must be adaptive and must customize their treatments to the needs of their patients, ranging from evidence-based protocols to personalized medications.

3.1 Information

Clearly, with all of the data being generated by smart sensors or connected servgoods, there are yottabytes (1000**8 bits) of data available, even just in the medical or healthcare domain. Such Big Data — A term applied to data sets whose size is beyond the ability of available tools to undertake their acquisition, access, analytics and/or application in a reasonable amount of time — must be transformed into useful information. Whereas Tien^[11] forewarned about the data rich, information poor (DRIP) problems that have been pervasive since the advent of large-scale data collections or warehouses, the DRIP conundrum has been somewhat mitigated by the Big Data approach which has unleashed information in a manner that can support informed — yet, not necessarily defensible or valid — decisions or choices. Big Data is poised to add greater value to businesses (which can fathom their own transactional data to detect patterns suggesting the effectiveness of their pricing, marketing and supply chain strategies); to understand planet earth (which is being extensively monitored on the ground, in the air and in the oceans); to solve science and engineering problems (which have become more data- or empirically-driven); to enhance the World Wide Web (which is amassing terabytes of textual, audio and visual material that is becoming widely available through search engines like Google, Yahoo and Bing); to aid national security agencies (which are collecting and mining satellite and thermal imagery, audio intercepts and other readily available data and information); and, more specific to our current focus, to support modern medicine (which is collecting and mining large amounts of image scans and genetic markers). In a recent special issue of the Spectrum, IEEE^[17] employs Big Data to hack the human operating system code (similar to what medicine is attempting to do in understanding the human body); a series of articles reports on first reading the code, then analyzing the code, and finally attempting to change the code through microbots, cyborgs, Watsons and brain hacking.

While Big Data allows for a) better integration or fusion and subsequent analysis of quantitative and qualitative data, b) better observation of rare but great impact events or “black swans”^[18], c) greater system and system-of-systems efficiency and effectiveness; d) better evidence-based — “data rich, information unleashed” (DRIU) — decisions that can overcome the prejudices of the unconscious mind, and e) more messy findings but yet good enough to support informed decisions, there are still some critical concerns^[19]. Table 6 provides a summary of the concerns, in terms of the four data processing components: acquisition (i.e., Big Data does not imply big/complete understanding of underlying problem; Big Data quantity does not imply Big Data quality; and Big Data sample does not imply a representative or even a complete sample); access (i.e., Big Data’s on-demand accessibility may create privacy or Big Brother concerns; Big Data’s real-time abilities may obscure past and future concerns; and Big Data’s cyber-security concerns may overlook personal-security concerns); analytics (i.e., Big Data’s inherent messiness may obscure underlying relationships; Big Data’s correlational findings may result in unintended causal consequences; and Big Data’s unleashing of information may obscure underlying truths); and application (i.e., Big Data’s feasible explanations may obscure more probable explanations; Big Data’s evidence-driven findings may obscure underlying factual knowledge; and Big Data’s subjective, consumer-centric findings may obscure simpler objective findings).

Other concerns include surveillance by autocratic governments and processing data in an increasingly unfocused, unproductive and generally “shallow manner”^[20]. Moreover, even Google’s vaunted flu prediction algorithm (which in 2009 was able to predict and locate the H1N1 flu spread on a near real-time basis) failed in 2012; it was predicting more than double the proportion of doctor visits for influenza-like illnesses than the Centers for Disease Control and Prevention (which bases its estimates on a survey of clinics located throughout the US) — underscoring the fact that correlation does not necessarily imply causation. Of course, the potential Big Data concerns or problems can be mitigated with thoughtful and effective approaches and practices; for example, legislation could be promulgated and passed to forbid the invasion of privacy and to dispense severe sanctions against those who break the law or who knowingly publish false findings. Alternatively, a watch dog organization can be created to discover such findings, much like the recently established METRICS (Meta-Research Innovation Centre at Stanford) which mission is “identifying and minimizing persistent threats to medical research quality”.

3.2 Instrumentation

More recently, much of the focus on instrumentation in the healthcare area is on wearables, including smart phones and devices, self-powered sensors in clothing, and even tattoos printed on the body, all to monitor vital signs — including blood pressure, glucose, body temperature, respiratory rate, airborne pollutants, and ambient temperature. (Parenthetically, it should be noted that some of wearables may cause interference; for example, some tattoos can block light from a smart watch sensor.) Interestingly, the psychology behind wearables is that the obtained data would help the wearers to achieve, if not exceed, some daily goal of fitness. However, according to market researchers, while at least 1 in 10 Americans own a tracking device like Fitbit and Nike Fuelband, more than a third of those who possess such a gadget abandon them within a few months. The reasons for abandonment include loss of interest, discomfort, and meaningless output (e.g., after a few days, the wearer has a good feel of when the daily goal is met and/or exceeded).

More sophisticated sensors are forthcoming. Retinal imaging and the ability to view the underlying blood vessels can reveal the presence of several disorders, including heart disease, strokes, diabetes, multiple sclerosis, and Alzheimer’s. In the near future, retina selfies may help physicians detect such disorders in a timely manner so that appropriate preventative treatments can be successfully prescribed.

Experimental psychologists are now studying ways of motivating individuals beyond the feedback of data from wearables; instead, they must be motivated on a subconscious level to take charge of their health. Actually, the popularity of wearable trackers in the recent past has been partly due to outside incentives, including insurers providing discounts on premiums if vital signs improve, gyms installing televisions in front of treadmills, instructor-facilitated exercises (e.g., Pilates), game-directed exercises (e.g., Zombies, Run!), and social media-reinforced goal setting. In sum, successful wearables should support people making commitments to themselves and others.

3.3 Insertion

More and more of the sensors, medicines and replacement parts that are inserted into the body are being made not only with new and compatible materials but also employing a technology-driven, layer-by-layer assembly of thin films, sometimes referred to as 3-D Printing. In this manner, one can have nanometer control of each film’s material and thickness for a range of applications, including membranes, catalysis, biomedicine, and optics. The assembly technologies — including dipping, de-wetting, centrifugation, spinning, spraying, atomization, electrodeposition, electro-coupling, microfluidics, and filtration — can also have an impact on the performance and physicochemical properties of the nanofilms. Scaling up layer-by-layer assembly or printing is crucial for carrying out real-world applications. Also, the assembly technologies must be compared and assessed to gauge their ability to coat particulate substrates.