Medical Device Technology Forecast: Discussion
Although a large number of topics emerged from this survey, six major trend categories circumscribe all of the product-type examples. These same categories encompass all of the generic technologies except "infection control" (which elicited responses too heterogeneous for analysis), and "virtual reality" (which participants viewed as an educational tool, not a clinical one). These trend categories are:
- Computer-related technology
- Molecular medicine
- Home- and self-care
- Minimally invasive procedures
- Combination device/drug products
- Organ replacements and assists
The first two of these trend categories comprise developments grounded in scientific advances; the second two in growing delivery modalities; and the last two in specific product-types.
1. Computer-related Technology
Computer-related technologies cited in the survey (Table 1) include computer-aided diagnosis, intelligent devices, biosensors and robotics (which panelists associated with intelligent devices), and networks of devices. ("Telemedicine" and self-diagnosis technologies are discussed under "Home- and self-care" below.)
Specific product-types cited by participants as examples of these technologies (Table 2) are integrated patient medical information systems, patient smart-cards, clinical lab robotics, computer-aided clinical lab systems, biosensors, and robotic surgery.
Current computer systems began with the vacuum tube (1907) which enabled the first automatic electronic digital computer, the Eniac (1946). Similarly, the transistor (1948) enabled the integrated circuit microchip (1959), the microprocessor comprising an entire computer processor on one chip (1971), and the first miniaturized 'personal' computers by Altair (1975) and Apple (1977). The resulting applications in the clinical community were administrative patient data bases (1970s), computerized medical diagnosis programs (ca. 1970), and 'computerized' medical devices (1960s and 1970s). , , , 
All of the Table 1 technologies encompassed by this category were judged likely to experience significant development within the next five (and ten) years that would result in new products for clinical use.
On a 'systems' level, the survey participants projected very significant developments regarding integrated patient medical data bases (including patient 'smart cards'). The driving forces were described as cost-reduction pressures, business and financial planning, interest in clinical 'outcomes' data, and computer support for clinical decision making. These views are supported by other analyses in the literature. 
Participants, however, had divided expectations regarding the future of computers in clinical decisionmaking. Clinicians' projections were generally more conservative than engineers'. Similarly, the literature reflects some moderation evolving among researchers in this field, suggesting a support role for future computer models rather than the full-fledged automatic consultations envisioned by some early exponents. On the other hand, participants did generally anticipate an increasing trend toward reliance on automated analysis in the clinical laboratory. 
An escalating trend toward microprocessor-based intelligent devices was generally anticipated. Commonly cited examples were cardiac and drug-delivery implants, as well as 'smart' robotics used in minimally invasive surgery. The accelerating prospects for such medical devices are ultimately grounded in the doubling of microprocessor power approximately every 18 months in accordance with "Moore's Law". The chip technology described by "Moore's Law" will probably continue to grow in this exponential manner for at least another decade, after which limitations may arise due to quantum effects. The growth of intelligent devices is likely to mirror that growth in microprocessor technology.
Survey participants also anticipated miniaturized biochemical and optical biosensors in these intelligent devices, sometimes in integrated "sensor fusion" configurations. The literature, too, reflects a substantial research interest in bioanalytic, electrochemical, and optical sensors in intelligent biomedical applications. ("Robotics" are discussed under "Minimally Invasive Procedures" below.) 
Finally, some participants noted the potential for microprocessor-based intelligence to promote a trend toward customizing device performance to the needs of individuals and of specialized patient groups with common clinical characteristics.
2. Molecular Medicine
The technologies (Table 1) cited in the survey under this category are genetic diagnostics, genetic therapy, and tissue-engineered devices. Except for tissue-engineered devices (discussed chiefly under "Organ Substitutes" below), the only specific medical-device example of those technologies in Table 2 is biosensors.
Current developments in genetics were anticipated in a number of developments including Nobel prize-winning research on the role of chromosomes in heredity (1933), and on the structure of DNA (1953). Seminal landmarks include the development of amniocentesis to detect genetic disorders (1952), the demonstration that eukaryotic DNA fragments could be inserted into bacterial cells and reproduced (1973), the development of hybridoma technology to produce unlimited quantities of monoclonal antibodies (1975), the use of antisense DNA to modulate gene expression (1978), the initiation of the Human Genome Project (1986), the discovery of a gene associated with Duchenne muscular dystrophy (1986), the development of automated DNA sequencing methods combining fluorescence-based enzyme techniques with laser instrumentation and programmable pipetting robots (1990s), and the first published account of successful gene therapy for a cholesterol-related disorder (1994). , , , 
In scoring the technologies, participants expected significant developments leading to clinical applications in both genetic diagnosis and tissue-engineered devices over the next five to ten years. This expectation clearly derives from the ongoing Human Genome Project which has targetted the goal of sequencing the 3,000,000,000 monomers comprising the 80,000 genes in the human genome by 2005. Participants were much more guarded about the prospects for clinical delivery of genetic therapy even over a ten year period, however.
For genetic diagnosis, participants' interviews projected accelerating growth especially for single-gene disorders such as cystic fibrosis. The primary medical hardware cited was the DNA microchip sensor device. 
Genetic therapy elicited somewhat less optimism in interviews and group discussions. With over 4000 known human genetic diseases, participants did expect an intense interest in this field during the next decade, but most believed the likelihood of a major clinical impact during that period to be only moderate. Some participants noted the potential for in vivo delivery of tissue-engineered genetic therapies through implantion of sequestered tissue-engineered cells in encapsulated form. Investigation of such polymer-cell implants is discussed in the literature, as well. 
Cancer was cited by participants as a likely focus of both diagnostic and therapeutic genetic techniques, partly because of the large patient population. The literature, too, reflects research interest in genetic techniques such as tumor vaccines, "suicide" genes, and tumor suppressor genes to treat diseases including certain leukemias, brain tumors, carcinomas, melanomas, and retinoblastomas. 
3. Home- and Self-Care
Generic technology areas (from Table 1) included in this trend are home/self monitoring and diagnosis, home/self therapy, and telemedicine. Specific product examples (Table 2) encompassed by this category are home diagnostics and telemedicine for patients in the home.
This trend encompasses such milestones as the introduction of home dialysis (1964) and a broad variety of other devices. More fundamental, though, is the fact that nearly all U.S. health care was performed at home by nonprofessionals until around 1900. Before then most medical care was provided at home by relatives and neighbors; physicians played a small role in the care of the average patient. Indeed, until state licensing boards appeared in the 1880's anybody could call him or herself a 'doctor' in the U.S. It was also not until around 1900 that hospitals began to grow into their modern form, due to the emerging need for specialized facilities to house antiseptic surgical suites with anesthetic equipment, and the unique requirements of the new field of radiology. Today, home- and self-care are re-emerging in response to cost-containment pressures resulting both from the explosive growth of medical science since 1950, and from Medicare and Medicaid funding legislation in 1965. This return to decentralized care is being catalyzed by the emergence of the Internet as an unprecedented conduit of health information to patients, and by the diffusion of inexpensive computer technology as an aid to medical decisionmaking by individual consumers. , 
Participants' scores indicate their expectation for significant developments leading to new products in each of the technologies in this category in both five- and ten-year intervals.
During discussions, participants generally envisioned this trend as important but unlikely to produce significant technical advances. It was perceived to be driven by considerations of cost and, to a lesser extent, convenience. The types of home diagnostics commonly envisioned were tests involving urine and blood chemistry, as well as drug concentrations -- particularly for elderly patients. Improved monitoring of glucose levels for diabetics was frequently mentioned. The most common form of home therapy cited was drug administration using simplified delivery techniques. Some participants noted the prospect of using home-based intelligent devices to modulate therapies and to "coach" patients. Several participants noted the possible use of relatively simple forms of telemedicine for home care, especially within the confines of a local or regional medical system. Interestingly, participants anticipated greater significance for this "low-technology" telemedicine application than for some other other "high-end" versions, perhaps because of potential interstate jurisdictional difficulties during the time period addressed by this study.
4. Minimally Invasive Procedures
Technology groups (Table 1) related to this category include minimally invasive devices, medical imaging, microminiaturized devices, laser diagnosis and therapy, robotic surgical devices and non-implanted sensory aids. The specific examples (from Table 2) cited by participants were minimally invasive cardiovascular and neurosurgery, laser surgery, robotic surgery, nanotechnology, endoscopy, functional and multimodality imaging, MRI, PET, and image contrast agents.
The invention of the stethoscope (1816) began a landmark change from diagnostic reliance on surface observations and patient reports to collecting data on internal events using nontraumatic methods. It was followed by such devices as ophthalmoscopes (1850), clinical thermometers (ca. 1850), sphygmomanometers (1896), ECG devices (1901), and EEG instruments (1929). , , 
Noninvasive radiologic imaging began immediately after the discovery of x-rays (1895), and eventually included PET (1951), ultrasonography (1968), CT (1971), and MRI (early 1980s). , , 
The early examples of modern optical endoscopy were laryngoscopes (1857), scopes for the rectum and vagina (1860s), cystocopes for the urinary bladder (1877), and arthrocopes (1918). Development of fiber optic imaging bundles (1950s) made possible the first really flexible endoscopes and revolutionized endoscopy, including a host of emerging therapeutic techniques (1950s) which would eventually include laparoscopic appendectomies, herniotomies, cholecystectomies and hysterectomies. After the invention of the laser (1960) and its introduction as a revolutionary surgical tool (1962), those optic fiber developments transformed its use, as well. , , , 
Early developments in minimally invasive cardiac surgery included the cardiac catheter (1929), the intra-aortic balloon pump (1961), and balloon angioplasty (1968). Landmarks in laparoscopic surgery include the first laparoscopic appendectomy (1983) and the first laparoscopic cholecystectomy (1987). , , , , 
Table 1 reflects the survey participants' strong view that every technology in this category will experience significant new developments during the next five and ten year periods leading to new clinical products, with two exceptions. Substantial developments were anticipated for microminiaturized devices, but only on a ten-year time scale. For nonimplanted sensory devices, participants generally envisioned only a modest chance for major innovations throughout the next decade (except for the specific example of hearing aids).
In interviews and group discussions, survey participants expressed an expectation of continuing advancements in endoscopic procedures including fiber optic laser surgery and optical diagnosis, smart miniaturized robotic devices, and a range of miniaturized devices. Clinically, most participants expected an emphasis on minimally invasive cardiovascular surgery and minimally invasive neurosurgery.
While increasing miniaturization of components was broadly anticipated, nanotechnology was seen as a separate issue. Although some participants believed nanotechnology might eventually alter the clinical landscape profoundly, views were divided on the likelihood of signficant developments over the ten year period covered by this survey.
Participants also predicted continuing advances in noninvasive medical imaging, including a trend to image-guided procedures. The most pronounced expectations were for developments in functional and multimodality imaging.
Finally, some participants observed that longer term trends might ultimately lead to non-invasive technologies. Such technologies would seek to direct energy (not material devices) transdermally to internal body structures for therapeutic interventions. Existing techniques that may point toward such future developments include ultrasonic lithotripsy and gamma knife technology.
5. Combination Device/Drug Products
Table 1 includes a single technology area in this category -- device/drug/biological products. Two specific examples were cited by participants and included in Table 2. These are implanted drug delivery systems (whose primary function is drug delivery) and drug impregnated devices (in which drug delivery is an adjunct to the device function).
The history of this category includes a variety of product-types, dating at least from the perfection of the hypodermic needle (1855). There are many modern examples of implanted delivery systems, such as the insulin pump (1980). One fundamental driving force for delivery systems has been the growth of new pharmaceutical products, especially since the dramatic expansion of drug research after 1945. That research has led to the synthesis and testing of millions of compounds for pharmacological and antimicrobial properties. Indeed, today much of that development is performed in automated computer-controlled systems, leading to an even greater acceleration of the process. , 
Table 1 shows that participants anticipated a strong likelihood of developments in this category over both five- and ten-year periods leading to new clinical products.
In the interviews, several survey participants characterized this as an very important area having major importance to a large group of patients. Generally, participants expected three types of developments. First, they anticipated development of new products designed for implanted delivery of insulin and other drugs. They pointed toward new implanted pumps, possibly intelligent devices with improved biosensors to monitor concentrations in body fluids and make dynamic adjustments in delivery rates. They also suggested the likely development of new polymeric timed-release devices which could improve the delivery of long-acting pharmaceuticals at optimized locations and rates.
Second, participants projected new developments in drug-impregnated devices. Examples included new types of cardiac implants with antithrombogenic drugs, as well as orthopedic implants with bacteriostatic coatings.
Finally, survey participants expected new developments in drug delivery systems to simplify reliable use by unsophisiticated patients in home settings, including the growing elderly population. Examples included nasal and inhalation products.
6. Organ Substitutes and Assists
The Table 1 technologies subsumed under this category are artificial organs, tissue engineered organs, and electrical stimulation. Specific product examples (Table 2) include bone, heart valves, heart pumps, cartilage, pancreas, blood vessels, kidney, skin, liver, eye, and regenerated nerve cells. Also included in the examples were cardiac, neural and neuromuscular stimulation.
There were few significant replacement or assistive devices before 1950 other than wooden legs, corrective glasses, dental prostheses, and (relatively unsuccessful) attempts to stabilize bone fractures with metallic implants. The few exceptions included the Drinker respirator, or "iron lung" (1927), and the first artificial kidney (1944). Even the latter was developed into a practical device only later for chronic use in a hospital setting (1959) and, later still, for home use (1964). , , , 
Artificial replacement implants largely began with the same development that enabled an explosion of donor-organ transplants -- the cardiopulmonary bypass unit (1951). This device, together with heparin and hypothermic surgical techniques, opened the way for donor-organ transplants of the kidney (1954), liver (1963), and heart (1967), and eventually even the intestines & pancreas (1990s). , , , 
Important early replacement and assistive devices included mechanical heart valves (1952) synthetic arterial grafts (1957), implantable pacemakers (1959), cemented total artificial hips (1960), pulsatile ventricular assist devices (1963), xenograft bioprosthetic heart valves (1965), and artificial hearts (1969). The development of many of these replacement and assistive devices has been motivated mainly by the severe shortage of natural donor organs. It has been estimated that each year at least 2 million U.S. patients receive artificial body parts while only 20,000 donor organs are available for transplant. , , 
Table 1 indicates that survey participants expected electrical stimulation technologies to continue to yield new developments in cardiac, neural, and neuromuscular applications leading to new clinical products over the next five to ten years. Significant new developments were also deemed likely for artificial (i.e., hardware) and tissue-engineered organs but only over the longer ten-year period.
Many future developments will be driven by the continuing dearth of natural donor organs. For the 260,000 potential liver transplant candidates in the U.S. approximately 3,500 receive liver transplants annually, and 25,000 liver patients die each year. About 10,000 renal transplants occur each year in the U.S. for a dialysis population exceeding 200,000, while approximately 40,000 kidney patients die. Many of these patients are transplant candidates. 
Survey participants projected a variety of specific products in this category, although most of these products would replace only a few of the most critical functions of the target organ. Some products were envisioned as primarily hardware-based devices such as the electrical stimulation devices, miniaturized assistive heart pumps, and portable hemodialysis units. Others were primarily tissue-derived products including cartilage and blood vessels, as well as improved bioprosthetic heart valves.
Many of the predicted examples, however, were hybrid hardware-tissue products including implantable bone and pancreas, and, eventually, kidney and heart valve replacements. Participant predictions and the scientific literature both reflect research on these devices using implanted filters, polymer-tissue composites, polymer- or hydrogel encapsulated cells, and cell-seeded synthetic scaffolds or hollow fibers with and without additional enzymes or pharmacological agents. Several participants predicted that such hybrid products were likely to play a dominant role in this category over the next decade.
Finally, participants identified three applications which they deemed important, but not likely to generate developments leading to new clinical products in the next decade. They were electronic ocular prosthetics, artificial livers, and nerve regeneration products. The latter two areas were envisioned as hybrid hardware-tissue devices.
This category poses especially difficult challenges in predicting the timing of new developments. Experience suggests that the difficulties entailed in developing artificial organs are often underestimated. These difficulties include the effort and the technical advances needed to pass from a 'proof of principle' prototype for animal evaluation to fabrication of a clinically acceptable system for human use. Among the bottlenecks frequently cited are full-scale design, cell procurement, cell survival, and device storage. 
In aggregate, these projections support a vision of the next decade with four discernible characteristics. First, medical hardware seems certain to become smarter. Devices and systems are likely to reflect a more sophisticated capability for intelligent behavior, and more mature information data bases to guide product performance. Second, smarter and simpler products will facilitate a growing trend to decentralization of care. Technology will support the cost- and convenience-driven diffusion of health care from the clinic to the home. Third, product development will increasingly blur the boundaries between biological systems on the one hand, and physical and engineering designs on the other. Integrated and hybrid approaches will play an expanding role. Fourth, technological developments will help to catalyze a trend toward greater precision in clinical interventions, both spatially and temporally. Reductions in invasiveness will probably mirror advances in miniaturization and improvements in early diagnosis.