In 1971, after more than thirty years in academic medicine--including twenty years as chair of the Department of Surgery at SUNY Downstate--Dennis felt ready for a change. When the National Heart and Lung Institute (NHLI) at the National Institutes of Health asked him to take over the direction of its Division of Technological Applications (DTA), which was responsible for coordinating and overseeing the development of total and partial artificial hearts and other cardiac assistance devices, he accepted. He believed the job would provide an opportunity to apply his expertise in cardiac surgery, blood flow dynamics, and cardiac assistance devices, coupled with his extensive administrative experience, in a new environment.

The Division of Technological Applications was an ambitious biomedical research program, conceived in the era of technological optimism generated by America's successful space program. In the early 1960s researchers such as Michael DeBakey, Adrian Kantrowitz, Willem Kolff, and others were making progress developing cardiac assist devices. To many, it seemed increasingly possible that, with sufficient expertise and funding, commercial models might be achieved within a decade. Like space flight, development of a mechanical heart seemed to be essentially an engineering problem. If the resources of private industry could be marshaled under a dedicated federal program, the major difficulties--reliable pumps, implantable power sources, and, for all components, materials compatible with blood and other tissues--could be resolved. And the need for such devices was great; despite decades of steady progress in medicine (including cardiovascular surgery), heart disease still caused hundreds of thousands of deaths each year. In 1963, the Advisory Council of the National Heart Institute (later the NHLI), at DeBakey's suggestion, recommended that NIH prioritize the development of an artificial heart. With initial Congressional funding in 1964, the Artificial Heart Program was established, and six companies were contracted to assess the feasibility of an implantable artificial heart from an engineering standpoint. The summary report of these studies, released in 1966, recommended a systems-oriented effort to develop "a family of temporary emergency devices, long-term partial heart replacement devices, and total artificial hearts."

Accordingly, the NIH master plan for the artificial heart called for competing private firms to design, test, and develop component parts and materials for the devices, along with specifications for hardware, installation facilities, testing, and training facilities, by the end of 1967. During the next phase, to be completed by the fall of 1969, the winner of the initial competition would carry out the actual design and development of a number of artificial heart prototypes. In the final phase, the specifications would be completed for mass-production, installation, and maintenance of artificial hearts--by 1970.

This timetable soon proved to be wildly optimistic. The program foundered during its first few years for many reasons. NIH director James Shannon, though publically supportive of the initiative, was skeptical of the narrow focus on the total artificial heart, and modified the program so that some of its funding could be used for other cardiovascular research, leaving less for engineering work. The two primary technical challenges--finding biocompatible materials for blood pumps and energy systems suitable for implantation--were formidable. Because a nuclear power supply, using plutonium 238, seemed to be the only feasible option for the latter, the Atomic Energy Commission became involved in the power system work. Meanwhile, the first heart transplants were performed in 1967-68, suggesting an alternative to artificial hearts, and the first implantation of an artificial heart prototype by Denton Cooley in April 1969 (regarded by some surgeons as premature and ill-considered) raised concerns about the appropriateness of the program's goals. A 1969 advisory panel recommended broadening the program to include left ventricular assist devices and intraaortic balloon pumps. The artificial heart program became the Medical Devices Applications Branch of the National Heart and Lung Institute, and in 1972, the DTA.

When Dennis arrived in March 1972, the work on artificial hearts and other circulatory support devices was still going slowly; there had been some progress during the first five years of development work, but a reliable, long-term heart substitute hadn't yet been realized. Congressional enthusiasm for the program was waning, as were the generous federal budgets of the 1960s. Dennis was asked to evaluate the program's operation, identify sources of failure, and make the necessary changes. The task did indeed call on all his expertise and experience.

Dennis found many obstacles to full effectiveness in the program. The most significant was its basic systems engineering structure. As he noted in his final evaluation, the systems engineering approach worked very well in projects in which the basic scientific and technical knowledge were already established. The NIH quest for a usable artificial heart, in contrast, was begun with seriously inadequate knowledge of physiological and hematological problems, among other biological issues, and of synthetic polymer science. And often, the available basic knowledge was not used, because the medical and engineering researchers didn't work together. Thus, for example, implantable devices were being developed with little consideration of basic factors such as the device's size, shape, and weight, or the need to dissipate heat generated by its operation. Likewise, the competitive contracts arrangement discouraged the open sharing of information between the groups working on various aspects of device development, so that efforts were duplicated, and unproductive work was continued rather than changing direction in response to new data. The problem of knowledge was exacerbated, Dennis added, by the program's dysfunctional management structure: the staff of 6 to 8 physicians and engineers was far too small to adequately plan, organize, and manage the program, monitor more than 100 ongoing contracts scattered across the country, and issue requests and review proposals for new contracts, among other duties. Collectively, the staff often lacked the expertise needed to make good decisions, and even when they did have such expertise, it was not always applied. Thus, for example, expensive contracts were awarded and renewed with only token reviews by NHLI staff, while less costly contracts for exploring essential physiological problems were terminated. To make progress, Dennis said, the program would need contract monitors with expertise in cardiovascular physiology and surgery, hematology, pathology, radiological physiology, mechanical, electrical, nuclear, and hydraulic engineering, biochemistry, and polymer chemistry. These experts should be not just administrators, but investigators involved in the work in progress. (Dennis himself had hoped to do some hands-on research in addition to his other duties, but found no lab space available at NIH.) Unfortunately, Dennis noted, current federal government hiring and salary restrictions, as well as travel restrictions, blocked his ability to make such improvements. Finally, the program would have to acknowledge that much basic science work needed to be done to inform the engineering work, and that the goal-oriented contractual approach was "an extravagant and unproductive way" to fund and manage such investigations.

For Dennis, his time at NIH was intensely frustrating. He succeeded in identifying many problems with the DTA program, weeding out non-productive contracts, replacing some of the DTA staff, and formulating a proposal for reorganization. But in August 1973, the DTA was disbanded due to budget cuts, and Dennis was given the title of Special Assistant for Technology in the office of the NHLI director. He continued to coordinate technological development, but was "in essence kept busy with a panorama of activities for which he [was] not particularly well-grounded while his areas of expertise and the basis for [his] recruitment [were] ignored." He resigned from his NIH post in December 1974, and returned to academic medicine, this time at SUNY Stony Brook, where he hoped to get back to clinical work and teaching, and "get down to brass tacks again in the laboratory."