A not-so-recent paper estimated that for every £ spent on cardiovascular research, there was a continuing 9% savings in health costs 17 years later. If the time from the research bench to the bedside was reduced to 10 years, those continuing savings rose to 13%. Why does it take so long to translate medical science into medical care?
Medicine is notoriously slow to change, which is, in many ways, a good thing because it allows time for a concept to be tested and shown to have value. The usually 10-year process of developing a vaccine was short-circuited when COVID came to call, and Operation Warp Speed brought one to market in just a few months.
The translation of research to action has become a field of study, implementation science. For medicine, there are two distinct phases. The first, “translation to humans, is when basic science identifies a potential clinical good, be it a medication or surgical procedure. The second is “translation to patients,” as that new clinical good is tested and found to be of value or not. For context, Operation Warp Speed accelerated the second phase of the translation. There is another less spoken of phase, “translation to practice,” were guidelines and standards of care become woven into the fabric of day-to-day practice.
This not-so-current study sought to identify the length of those waiting periods based on a meta-analysis of empirical studies of lag times. They found a variety of measures that might be used, all yielding differing lag times. But they did make some general observations.
- The typical lag time from bench to bedside approach 17 years.
- Negative results took a year longer to be published than positive results – the result presumably of publication bias.
- Lag time varied by domain and even within a domain. In neonatology, using artificial surfactants to improve breathing took 13 years to become standard care, and parenteral nutrition to support an immature digestive system 21 years.
“Each activity involves a lag, either because the effort required for carrying out the task or as a result of non-value adding waits.”
Over my clinical career in surgery, I have had the opportunity of experiencing translation into practice several times. Each has assumed its own trajectory with and without those non-value-added waits. My experience may inform you as to what those waits might be.
Laparoscopic surgery: the first shift
The widespread adoption of laparoscopic surgery from gynecology to general surgery set the stage for minimally invasive surgical care and surgical robots. The first laparoscopic operation widely adopted by general surgeons was laparoscopic cholecystectomy, the removal of the gallbladder, a common procedure. But an entire surgical workforce required retraining to carry out this new technique. General surgeons in practice quickly developed short training programs with gynecologists already conversant in the methods. Hospitals developed mentoring and accreditation procedures; laparoscopic cholecystectomy became the standard of care in about two or three years.
The rise of minimally invasive vascular surgery
Endovascular surgery swapped long incisions and operating times for needle sticks, wires, balloons, and stents. Once again, an entire workforce needed to be converted, but unlike the overwhelming response to laparoscopic surgery, that conversion took most of a surgical generation.
The techniques involving wires and catheters “belonged” to the cardiologists busy applying them to coronary arteries. Radiologists also were quite conversant with those methods, using them in performing arteriography, the mapping of blood vessels. It was not a giant leap for those radiologists to realize that they might just as well be dilating an artery “while they were there.” Two financial factors partly fueled their interest and enthusiasm. Radiologists never admitted patients to the hospital; they provided images, care was left to other physicians – so caring for the patients whose arteries were fixed was not their problem, even though the fee for the fixing was theirs, a sweet deal. At the same time, non-invasive imaging with CT and MR scans was beginning to eat away at the arteriography business, so finding new uses for those skills was imperative.
Oddly enough, unlike gynecologists, radiologists did not feel these techniques should be shared with their surgical colleagues. Several years of turf wars followed as vascular surgeons found ways to train and accredit themselves. Cardiologists also reasoned that a coronary artery was just a more specific example of any artery, so their training was sufficient to expand their practices to peripheral arteries, the domain of the vascular surgeon. [1] It became a three-way turf war in many institutions, each specialty claiming its expertise. That added five years to the roll-out of endovascular care on a routine basis.
The early poor results of these techniques compared to open surgery slowed the adoption of endovascular techniques in vascular surgery. A generation trained in doing operations was not about to abandon those treatments for simpler ones with poorer results. Overall, between turf wars and improvement in patient selection and therapy, endovascular surgery required a generation of at least 15 years. There were two exceptions, the application of endovascular care to aortic aneurysms and surgery for strokes, carotid endarterectomy.
Endovascular aneurysm repair and Carotid stenting
Repair of an aortic aneurysm, a weakness in the wall of the aorta which, if ruptured, is essentially fatal, was a three-to-four-hour operation, an 10-12” incision, and a week in the hospital. The development of a stent graft that would provide the necessary repair in half the time, with two 2” incisions and an overnight stay in the hospital, transformed aneurysm surgery. What patient would not want the less invasive alternative? And what physician could afford only to be able to do the more invasive procedure when the doc next door could offer a less invasive option?
There was a rush for training; you had to be able to offer endovascular aneurysm care even if you were not interested in other endovascular care. In this instance, two hurdles delayed translation. First, a mentoring and accreditation program by the hospitals often devolved into turf wars where these procedures were collaborations with radiologists who knew about the wires and catheters and surgeons trained in using the stent grafts.
Second was a bottleneck in training on using the stent graft itself. These devices are complex in their deployment into the aorta, you need some hands-on experience, and the device manufacturers who had the most to lose from poorly implanted devices required their sign-off if you wanted to use their equipment. As the waiting line for those three-day introductory courses and observation of a live case grew, manufacturers prioritized their best customers to the front of the line. A community-based surgeon might well wait a year to get the initial training. As a trained surgical workforce expanded, the device manufacturers allowed partners to share with one another. Endovascular repair of aneurysms took a bit longer than laparoscopic cholecystectomy but not as long as endovascular repair of peripheral arteries.
And then, there was stenting of the carotid artery. Atherosclerotic material breaking free from your carotid artery can travel deep into your brain’s circulation, causing a stroke. The surgical treatment is to remove atherosclerotic material from the artery. Stenting involves placing a “chicken-wire” like structure over the material to keep it in place. Carotid stenting, closely following endovascular aneurysm repair, was anticipated to become the standard of care within about eight years. Thirty years later, it is used only about 25% of the time.
Among the reasons for its much slower adoption is that it does not seem to confer any advantage over traditional surgery in preventing a stroke. It has a slightly lower procedural mortality than endarterectomy, which is already hovering below 1%, but there is a higher risk of a peri-procedural stroke. The second reason is more technical; navigating a wire and catheter from the groin to the middle of the neck requires more experience and better patient selection. The actual process of implanting the stent requires more skill because of the risk of small bits of the artery breaking off and causing a stroke (athero-embolism). As a result, few physicians are willing to take the additional risk when there is a safer alternative. That this alternative may require an incision and overnight hospital stay is of secondary importance.
Implementing new treatments has several non-value-added waits. Some are turf battles; others are about training a sufficient workforce to provide the new care. Workforce development is often slowed by accreditation issues and the significant variance between the results in a carefully controlled Phase III study and the consequences when most practitioners have to climb a learning curve.
[1] Interestingly, while the cardiologists felt that if you could treat a coronary artery, you could treat any vessel, they firmly believed that treating coronary vessels was restricted to their specialty, and vascular surgeons need not apply.
Source: The answer is 17 years, what is the question: understanding time lags in translational research Journal Royal Society of Medicine DOI: 10.1258/jrsm.2011.110180
