Stories about mechanization of industry and the effect of industrialization on workers are centuries old. Ever since the Luddites attacked mechanical looms during the Industrial Revolution, battles have been waged over the social effects of technology and the role it should play in the marketplace.
Last week, the online magazine Slate.com published a series on robotics and how they are infiltrating various professional sectors. Slate’s excellent technology writer, Farhad Manjoo, described how law, pharmacy and journalism have hosted efforts to increase efficiency, productivity and results by applying software-based solutions. Since the state of medical technology has exploded recently, medicine received its own article in the series.
The author astutely describes how software packages are helping pathologists read Pap tests, radiologists read mammograms, and surgeons perform certain types of surgery more safely with robotic help. By increasing the efficiency by which one physician operates, Manjoo argues, the total number of physicians needed to do the work will drop.
However, at least to date, technology has not reduced the number of human jobs in the medical system. Quite the opposite – demand for the services of those specialties that Manjoo discusses have flourished. When going through rotations in radiology as a medical student (it was one of the specialties I considered pursuing), a common aphorism was that demand for radiology services increases at a rate of around 5% per year. This round number closely matches those estimates produced in studies. (1,2)
As imaging and diagnostic technology gets better, and even as these improvements simultaneously increase the amount of work that one radiologist can do herself, use of these services correspondingly becomes the standard of care. A generation ago, a general surgeon who did not have a significant proportion of his appendectomy specimens turn out to be benign (i.e. ex-post-facto unnecessary surgery) wasn’t doing his job properly, since he’d also be missing out on a significant portion of diseased appendices. Nowadays, surgeons more often wait for radiologic confirmation of appendicitis before proceeding to surgery. So sure, computerized tomography has increased the number of scans that a radiologist can read, but it’s also cause a commensurate increase in the demand for that radiologist’s services.
In oncology, there are obvious improvements made by technological advancements. On the radiation side, linear accelerators have replaced the older Cobalt-60 machines. Improvements like CT/MRI imaging, multileaf collimators, and stereotactic modalities allow for more conformal treatments and improve the therapeutic ratio. But so far, the developments have expanded what physicians can do to treat cancers, at least balancing out the increased supply of service with increased demand for those services. Thanks to medical advances in all fields (technologic and not), people are now living longer than ever and more patients are being diagnosed with and treated for cancer. And the notion that even semi-autonomous robots will soon doing the work once performed by multiple oncologic surgeons is farfetched at best. For the foreseeable future, technological advances are needed simply to allow practitioners to keep up with burgeoning demand and to offer patients the most up-to-date and effective treatments.
A more realistic near-term scenario is one mentioned in Manjoo’s article – the takeover by technology of certain aspects of oncologists’ jobs, leaving certain other facets in the hands of people. The scope of practice could change, freeing up more time for other tasks. For example, robotic assistance for surgeries could dramatically reduce waiting times for patients between diagnosis and procedure. Radiation contouring and planning has already been improved greatly by improved software packages. Improvements in image recognition software could easily continue to the point where most, or even all, contouring and treatment planning is initially completed electronically. The radiation oncologist would still be responsible for reviewing plans and making adjustments based on the patient’s clinical parameters.
Treatment decisions are not made in a vacuum. Adjustments must always be made based on patients’ goals, values, and responses to treatment. Completely leaving these decisions to computers would certainly invite legal liability issues should anything go wrong. Certainly trained and certified physicians need to be driving and supervising the operations of the oncology centre, even if more of the nuts-and-bolts work is done by computers.
Added improvement in automated treatment planning would greatly increase productivity and help keep up with demand for oncology services. But there would still be a role for clinicians to see patients, discuss treatment options with them, answer questions, and adjust the software-derived plans as would often be needed. The amount of time dedicated to the different roles of an oncologist may change – less time devoted to planning and executing treatment, and more time discussing decisions with patients.
The day may come when technology becomes so refined that it replaces a significant number of physicians in these “technical” specialties. That day is not now, nor does it appear to be on the horizon.
1. Maitino AJ et al. Radiology. 2003 Apr;227(1):113-7.
2. Smith-Bindman R et al. Health Affairs. 2008. 27(6):1491-1502.