ORIGINAL ARTICLE
The Risks of Innovation in Health Care Dieter R. Enzmann, MDa Abstract Innovation in health care creates risks that are unevenly distributed. An evolutionary analogy using species to represent business models helps categorize innovation experiments and their risks. This classification reveals two qualitative categories: early and late diversification experiments. Early diversification has prolific innovations with high risk because they encounter a “decimation” stage, during which most experiments disappear. Participants face high risk. The few decimation survivors can be sustaining or disruptive according to Christensen’s criteria. Survivors enter late diversification, during which they again expand, but within a design range limited to variations of the previous surviving designs. Late diversifications carry lower risk. The exception is when disruptive survivors “diversify,” which amplifies their disruption. Health care and radiology will experience both early and late diversifications, often simultaneously. Although oversimplifying Christensen’s concepts, early diversifications are likely to deliver disruptive innovation, whereas late diversifications tend to produce sustaining innovations. Current health care consolidation is a manifestation of late diversification. Early diversifications will appear outside traditional care models and physical health care sites, as well as with new science such as molecular diagnostics. They warrant attention because decimation survivors will present both disruptive and sustaining opportunities to radiology. Radiology must participate in late diversification by incorporating sustaining innovations to its value chain. Given the likelihood of disruptive survivors, radiology should seriously consider disrupting itself rather than waiting for others to do so. Disruption entails significant modifications of its value chain, hence, its business model, for which lessons may become available from the pharmaceutical industry’s current simultaneous experience with early and late diversifications. Key Words: Health care, change, innovation, risk, experimentation, evolution J Am Coll Radiol 2015;12:342-348. 2015 Published by Elsevier Inc. on behalf of American College of Radiology
EXPERIMENTS AND RISKS In facing change, a previous recommendation was for radiology to experiment in the realm of the “adjacent possible” [1]. The easiest adjacent possible experiments are variations of current practices and business models. An evolution analogy reveals two different kinds of experiments, early and late diversifications, which have different risk profiles. It behooves radiology to understand, look for, and distinguish them, certainly within the imaging domain, but also inside and outside health care. We shall see in radiology that variants of current business models are late diversifications with low risk, whereas new business models are early diversifications with high risk. This article suggests considering higher risk experiments. The clarion call to organizations, professional groups, and individuals for “adaptation and innovation on a a
Department of Radiological Sciences, David Geffen School of Medicine at UCLA, Los Angeles, California. Corresponding author and reprints: Dieter R. Enzmann, MD, David Geffen School of Medicine at UCLA, Department of Radiological Sciences, 924 Westwood Blvd, Suite 805, Los Angeles, CA 90024; e-mail: [email protected].
massive scale” in response to health care changes entails risk [2]. From a system point of view, whether you call it “adaptation” or “innovation,” it is fundamentally experimentation [1]. The experiments are just that— experiments whether one perceives them as precipitated by an event, the Patient Protection and Accountable Care Act, or driven by scientific development, such as deoxyribonucleic acid (DNA) sequencing. Evolution continually experiments to create new species or subspecies, which incur risk when facing natural selection. That determines survival, replication, or extinction. Adaptation or innovation experiments in health care, including new organizational structures, new business models, new care models, new IT solutions, new technologies, and so on, incur risk when facing market, patient, and professional selection forces, which determine their fate. I tap into evolution using species and subspecies as an analogy to new and variant business models respectively [3]. All business models must generate cash, and they do so by using different components to construct a value chain. Different business models are defined by significant differences in value chain components. If radiologists ª 2015 Published by Elsevier Inc. on behalf of American College of Radiology
342
1546-1440/15/$36.00
n
http://dx.doi.org/10.1016/j.jacr.2014.09.026
work in capitated organizations as a cost center, in private practices as a profit center, or as salaried employees of universities, they are in different business models even though their clinical work is similar. Cash is generated differently in each. Our interest is in experiments that are capable of altering business models. Some innovations, such as tracking inventory by using radiofrequency identifiers and MR fingerprinting, just change business or clinical practices without altering the business model. Some innovations sound like new business models, but are not. As spectacular as CT and MR technologies were at their introduction, they did not disrupt radiology’s business models. They produced value chain variants, but cash was still generated in the same way, albeit in higher amounts. The variants were subspecies rather than new species, because the acquisition device component was only a variant; it still generated images, only differently. Sometimes experiments string together components into an entirely new value chain. In the past, a business model built around an Internet search engine did not exist. Many experiments were run; some failed (Netscape), and some survived (Google). Some survived by completely reassembling the advertising value chain, going from print area to clicks. These were new species, new business models. In medicine, a new business model example would be concierge private practice, in which real cash is generated by monthly retainers independent of service volume. Human experiments are attempts to change in response to or in anticipation of change. Experiments carry different risks depending on their intrinsic characteristics and on whether one runs or merely observes them. This article concentrates on the intrinsic nature and provenance of experiments. Paradoxically, experimentation generates not only risks but also options to counterbalance them [1,4]. Health care systems increasingly experiment to create options. Anticipating the need to manage population health risk, some experiment in care delivery using accountable care organizations (ACOs), while others experiment in consolidation to increase size and scale [5]. Health care entrepreneurs experiment with new companies evidenced by a recent spike in health care and biotechnology initial public offerings [6]. Radiology should ask, What is the risk profile of experiments surrounding us? Are we participants or observers? Evolution reveals two fundamental types of experiments, “early diversification” and “late diversification,” which differ intrinsically in character and in risk. The evolution analogy reveals important distinctions between them. Some detail is necessary because their temporal Journal of the American College of Radiology Enzmann n The Risks of Innovation in Health Care
sequence relative to selection forces is important in assessing risk (Fig. 1). Early and late refer to the relationship of the two types of experiments to each other, not to any time interval from what may be perceived as a triggering event. Assigning causality in complex systems is problematic, and any set of conditions can precede either or both diversifications [1].
MULTISCALE EVOLUTIONARY ANALOGY Using varying terminology, temporal sequences of design change in the evolution of complex systems have been identified [7,8]. The order of experiments matters. One sequence is based on speciation, nature’s way to innovate. It starts with early diversification, during which a remarkably large number of “experiments” generate a wide range of new species (Fig. 1). Another sequence occurs at the molecular level, with the creation of many new molecules, termed neutral diversity expansion [8]. Dramatic early design expansions inevitably face selection pressures, which unfortunately end on a down note, with most new designs disappearing. Species experience a “decimation” stage, during which an equally remarkable burst of extinction occurs, reflecting the high risk to participants (Fig. 1). At the molecular level, this selection is termed selective diversity contraction [8]. By appearing at multiple scales, this temporal sequence reveals its generality. Species designs surviving decimation form the “early standardization” stage and ultimately transition to late diversification, during which additional experimentation is limited to variations of surviving designs [7]. The design range is much narrower than early diversification. Late diversification, not being followed by decimation and having a low extinction rate, poses lower risk to participants and to observers, with one important exception noted below (Fig. 1). Transitions between diversifications and extinctions form steps in the punctuated equilibrium (PE) pattern of system change, meaning that they are unpredictable, more sudden than expected, and usually of indeterminate causality [1]. Many lineages in complex systems, be they animal species, cars, computers, TVs, medical services, and so on, trace out this sequence in a PE pattern [9]. LATE DIVERSIFICATIONS Let us work backward because late diversifications are easier to understand. These experiments carry lower risk to participants because the designs, rather than being untested new ones, are ones already tested by natural selection, ie, decimation. They have early diversification 343
Fig 1. This diagram shows the evolutionary sequence of experiments (new species) starting from early diversification through decimation to early standardization and finally to late diversification. Early diversification features prolific experiments, which decrease substantially during decimation to result in a few survivors that constitute early standardization. The color and number coding emphasize the large number of early diversification experiments, decreasing to a small number of survivors, which are both disruptive and sustaining types. Late diversification consists of surviving experiments expanding within a narrower design range (decimal numbers). The sequence highlights the relative changes in the number, type, and fate of experiments, which are analogous to experiments in health care. The icons are meant to reflect environmental factors related to, but not causal of, diversifications.
precedents and become variants, which tend to expand current business models. Radiology, for example, has created such low-risk diagnostic and treatment variants by incorporating CT and MR technologies into imaging subspecialties, by developing musculoskeletal ultrasound, and by performing image-guided tumor ablation. Although CT and MR technologies were low-risk, late diversification for radiology participants, they posed higher risks to observers, neurology, and general surgery. A striking example of late diversification is consolidation. Although it includes a mild form of extinction, the resulting business models are just larger variants of the consolidating entities. Consolidation occurs on many scales: a large health system merging with smaller hospitals, academic medical centers merging into large health systems, or smaller radiology groups merging with larger ones [5,10]. Consolidation, be it in airlines, hospitals, companies, or radiology groups (or societies), is classic late diversification because the underlying business models remain stable, the variation being primarily increased size and scope. When a small radiology group is absorbed by a larger group, but the cash-generating profit 344
center model remains dollars per examination, it may feel “disruptive.” It is not in Christensen’s sense, because the cash-generating business model is unchanged. If a small group were absorbed into a capitation, cost-center, salaried-radiologist model, it would be “disrupted” because cash would be generated in ways other than dollars-perexamination “piecework.” At the system level, this is still late diversification because the group has simply joined an existing radiology model variant, which once was disruptive. Observers in late diversification can face high risk. Radiology’s cottage industry is waning.
EARLY DIVERSIFICATIONS Early diversifications are more complicated and of higher risk for two reasons: they have a high mortality rate because of follow-on decimation and because they have a higher likelihood of generating disruptive new designs. The rapid appearance of multiple new designs in any health care segment heralds early diversifications. It is prudent to be alert to their development because they are harbingers of potentially uncomfortable change by germinating truly disruptive business models. Journal of the American College of Radiology Volume 12 n Number 4 n April 2015
There is a simplified relationship between these diversifications and Christensen’s innovation concepts. Early diversifications are more likely to produce “disruptive innovations,” whereas late diversifications tend to generate “sustaining innovations” [11,12]. Disruptive innovations are new, different species that disrupt rather than fit into existing business models, just as new species do not stay within their ancestor species. They add considerable risk to the environment. Sustaining innovations, by being incorporated into existing business models, add less risk [11,12]. Christensen’s disruption and sustaining terminology refers primarily to the innovation’s effect on business models, its value chain components, and how cash is generated, not on the technologies per se and not on disruption of individual careers. Another feature of this early-late diversification sequence merits attention. Survivors of early diversification are likely to be a mixture of “sustaining” and “disruptive” experiments (Fig. 1). Sustaining ones can be incorporated into existing business models and, hence, are of low risk. Disruptive early diversification survivors are higher risk because they cannot be incorporated. When disruptive survivors enter late diversification, their disruptive effects are amplified. It is often the “later mover” that is highly successful as a presumably wiser variant of the “first mover” in early diversification. [13] This is also known as the “Columbus Effect,” one worth knowing [14]. As noted earlier, mere observers can face higher risk. Identifying and monitoring the temporal relationships of experiments is central to assessing their risks.
DIVERSIFICATION SEQUENCE AND TIMING The well-publicized dot-com era illustrates the diversification sequence. An initial burst of a wild number of creative dot-com companies, most with untested business models, formed the early diversification. This “boom” was followed by decimation, the dot-com “bust.” Some survivors were disruptive—Google, Amazon, Expedia, Facebook, and others. Late diversification appeared when dotcom businesses again expanded as variations on successful surviving designs. Late diversification variants, such as Amazon Fresh and Google Shopping Express, are still appearing [15]. Late diversification of disruptive models can bring new entrants into health care. Google, in exploring its adjacent possible (grand-scale computing), is attracting vast, new resources into health care by assembling knowledge generally residing outside traditional medicine [16]. Late diversification of disruptive survivors intensifies their disruption (see corporate teleradiology). Journal of the American College of Radiology Enzmann n The Risks of Innovation in Health Care
Late diversifications can be quite successful and profitable. A German group simply copies successful American dot-com companies and expands them around the globe [17]. This is classic late diversification, during which companies essentially copy decimation survivors. Pharmaceutical companies (pharma) have done well by expanding “me too” drugs, drugs that have survived decimation. Late diversifications are constantly playing out in different arenas, ranging from travel to pharmaceuticals to health care. While they may be of lower risk to participants, they can be high risk to observers. Radiology faced this sequence on a smaller scale when it encountered early diversification in teleradiology. Decimation has waned, leaving survivors as a mix of disruptive and sustaining models [18,19]. Sustaining models, having been incorporated into radiology practices (night coverage), are already entering late diversification with mobile variants [20]. A disruptive, corporate venturebacked form of teleradiology (VRad, Radiology Partners, etc) has survived. Although it seems to be an adjacent possible, its provenance reveals it to have arisen outside of radiology, and it has a different value chain that allows hospitals to essentially outsource radiology. This model, as with other disruptors upon entering late diversification, as evinced by consolidation efforts, could increase its disruption. Variants are appearing, which bundle radiology with other subspecialties, expanding the scope of outsourcing and clearly exhibiting a different value chain and disruptive business model. Complaining about disruptive survivors and hoping they go away is unrealistic. A compelling industry-wide, simultaneous interplay between early and late diversifications is unfolding in pharma [21,22]. Early diversification has formed in biotechnology, driven by the fundamental shift to the molecular redefinition of diseases and derivative-targeted therapies. Tiring of shrinking profits from late diversification me-too drugs, pharma is availing itself of proliferating biotechnology companies to rejuvenate lagging innovation in their own research and development [23]. Forming various relationships with biotechnology companies allows pharma to parry some early diversification high risk. This early diversification, based on fundamental changes in science and technology, creates innovation opportunities in drugs, biologics, and cell therapy for pharma, which is simultaneously engaged in a classic late diversification activity, consolidation, buying innovative companies and “asset-swapping” to increase economies of scale in more focused, specialized disease areas [21-24]. To what extent early diversification biotechnology survivors will sustain or disrupt pharma is not yet clear. 345
Health care sectors experience early and late diversifications at different times, sometimes concurrently. Consolidation of health care systems is late diversification, whereas ACO experiments are early diversifications [5,25]. The recent “explosion” of ACOs (>600) indicates early diversification in these business and patient care models. The risk is high, as a large number will fail during decimation [25]. The PE change pattern means timing of failure is problematic, but health care systems and radiology groups would be wise to track both sustaining and disruptive ACO survivors [26].
SPOTTING EARLY DIVERSIFICATION Where do other early diversifications lurk? Christensen suggests that new experiments will proliferate as increasingly sophisticated care migrates from expensive to less expensive venues; from hospitals to clinics, to offices, to homes, and eventually to mobile sites, spurring decentralization and deprofessionalization of medical care [27]. Technology, medical skills, and medical knowledge will move centrifugally from long-lived standard models. Radiology has experienced this with handheld ultrasound. Early diversification will become more apparent as new patient care designs appear outside traditional venues [28,29]. Retail medical clinics in drug stores, “hospitals at home,” and “home palliative” care are early examples. Retail medicine is disruptive. Many experiments use cyberspace instead of typical physical assets [29]. New care designs and business models in telemedicine, such as “e-health” (e-referral, e-consult services) and “telehealth” (tele-ICU and tele-stroke services) are forming early diversifications [30]. The same Internet, digitization, and social media drivers that produced new business models in other industries will do so in health care. Early diversifications interact, so it is not surprising to see early diversifications appearing in flexible, wireless, mobile, biocompatible, and biodegradable “bionics,” accelerating medical decentralization and cyberspace applications [27]. The interaction allows new combinations in care design, which will amplify prephysician care by expanding virtual office or pre-emergency room care [31]. These early diversifications incorporating mobility, the home, wireless monitoring, and care by nonphysician professionals will accelerate a deprofessionalization trend [28,32]. These, like other early diversifications, face decimation. The survivors can be either sustaining or disruptive to current health care models. Disruptive survivors are likely to congregate around prephysician care involving 346
nontraditional clinical sites, paramedics, and midlevel professionals, some of whom will come from outside the medical profession. Strategic and tactical analysis is needed to deal with expanding deprofessionalism [32]. Radiology needs to be alert, as nonphysicians will increasingly need to order imaging tests appropriately. It is beyond the scope of this paper to spell out how radiology should respond to decentralization and deprofessionalization.
EARLY DIVERSIFICATION SECONDARY EFFECTS Unusual and unexpected value chain components may materialize in early diversification. Taxicab businesses did not expect competing business models, such as Uber and Lyft, to pair smart phones with a new component, private cars [33]. Budget hotels did not expect their customers to be linked via the Internet to a new value chain component, privately owned rooms, until Airbnb appeared [34]. The new “shared economy” can recruit decentralized components to new value chains to compete against longlived, staid entities. This represents the democratization of physical assets. Who cares? Late diversification of Uber could include moving things (packages, medical supplies) in addition to or with people. Although it is pure speculation, Uber and Airbnb could add personalized services to personalized medicine. RELEVANT EARLY DIVERSIFICATIONS Prodigious, new molecular diagnosis designs driven by molecular medicine and big data herald early diversification. Genomic and molecular diagnostic testing is undergoing “explosive” early diversification in cancer detection, prognosis, prediction, and monitoring. Hidden from many radiologists is dramatic progress in using genomics, epigenomics, proteomics, circulating tumor cells, circulating tumor DNA and circulating microribonucleic acid (miRNA) to achieve nonradiologic diagnosis and monitoring of treatment. Some sophisticated, successful fluid- and cell-based diagnostic and monitoring designs, will survive decimation. Radiology must attend to such survivors that could sustain or disrupt its value chain. Molecular diagnostics, when fully digitized, will spawn new competitive designs in “telediagnostics,” just as image information became more competitive when distributed via teleradiology. University of California, Los Angeles, telepathology is up and running in China. Monitoring molecular diagnostics decimation to identify surviving models will be critical in assessing risk to, and impact on, diagnostic radiology. This should stimulate strategic thinking because potential disruption Journal of the American College of Radiology Volume 12 n Number 4 n April 2015
depends on the “host’s response,” a point emphasized by Christensen.
RADIOLOGY’S CHALLENGES Radiology must digest Christensen’s innovator’s dilemma [11]. Although integrating sustaining technologies is relatively straightforward, integrating potentially disruptive ones is not. One reason is that sustaining technologies tend to increase profitability, whereas disruptive technologies often decrease it. A disruptive survivor might well reduce profit from radiology’s current transactional professional service model by challenging one of its value chain components. Whole new diagnostic value chains might be constructed outside of radiology [35]. The issue may come to radiology’s willingness to “disrupt” its value chain or wait for others to do so. Incorporating new elements into an existing value chain has its challenges, but IBM showed that it can be done when it expanded its hardware business model to encompass analytic software [36]. IBM chose to painfully disrupt itself internally. The company is still called IBM, but its business model is different. Incorporating disruptive technology often means a reformulation of the business model value chain, but that is preferable to its extinction. Moving from “volume to value” certainly entails change in radiology’s high-volume, transactional business model. By the way, “volume and value” are not mutually exclusive concepts. We briefly consider 3 potential disruptive scenarios. One radiology approach would be to change only tactically and remain focused on high-volume transactional imaging. This risks not only commoditization but also substantial outside disruption by molecular diagnosis. The risk is that blood tests and cell-based test components of new value chains will diagnose and monitor cancer (breast, prostate, lung, etc) or Alzheimer’s disease and substantially replace commodity imaging [37]. Diagnostic value will move out of radiology if it limits itself to isolated, transactional, diagnostic imaging. A second approach is to evolve into an information business, using late diversification. This requires some modification, but not reinvention of the current value chain to create robust, longitudinal, integrated, imagingbased diagnostic reports. Decision support fits in by delivering this value through pretest diagnostic information. However, image-based information businesses could easily take root outside of radiology, in part because other subspecialties may not face the innovator’s dilemma of decreased profit in assembling new value chains [19]. Radiology should be attentive to early diversifications in other clinical, image-based information businesses. It Journal of the American College of Radiology Enzmann n The Risks of Innovation in Health Care
would do well to push its own late diversification while it has capital to do so. Consolidation in radiology will facilitate this and the following model by providing organizational skills and capital. A third possibility is self-disruption by expanding the imaging information business to encompass nonimaging, molecular diagnostic information in comprehensive diagnostic business services and reports that more broadly facilitates clinical decision making, while adding value with time-saving, diagnostic accuracy [38]. This requires significant modification to the current value chain, probably to the point of a new business model, to integrate new molecular diagnostic information with analytic imaging, while putting both into clinical, scientific, and social determinants’ context. In such an endeavor, radiology assumes some higher risk inherent in early diversifications. Even the eminently successful Google has moved from its famed search (diagnostic) business to an information business via its Knowledge Graph [39]. This means that radiology would explore internal disruption à la IBM and maybe learn lessons from pharma’s relationships with biotechnology. This business model could go further in more closely aligning radiologic and pathologic diagnostic services by linking “tele” and physically located integrated diagnostic services. This does not require merger or loss of identity. This widens the design range of business models in early diversifications of integrated diagnostic information businesses. Attaching interventional radiology adds even more potential.
DIVERSIFICATION AND RISK All diversification experiments represent optionality, which is critical to all markets and economies [4,40]. Radiology must track all medical early and late diversifications. Understanding the temporal relationship and sequence of design experiments helps categorize them and their risks. Neither early nor late diversifications are controllable at the system level, but radiology’s reaction to them depends on whether you are a participant or observer. The highest risk is to participants in early diversifications and to observers of late diversifications in disruptive survivors. Early and late diversifications pose risks and offer opportunities to radiology. Radiology must participate in its own late diversification, because exploring the adjacent possible generates low-risk business model variants. Radiology should not only monitor molecular diagnoses but consider internal disruption to explore new business models, perhaps triggering some early diversification experiments. Should radiology choose 347
to disrupt itself, its ingenuity, expertise, and cleverness will need to be exercised.
TAKE-HOME POINTS n
n
n
n
n
Health care “adaptations” and “innovations” are equivalent to experiments that can be qualitatively categorized. An evolution analogy forms the basis of this categorization into early and late diversifications. Early diversifications exhibit prolific experiments, which carry high risk. Late diversifications exhibit a much more limited range of experiments, which carry lower risk. Radiology will be exposed to both early and late diversifications, so identifying them is informative because the former harbor the potential for disruption and the latter offer potential for growth.
ACKNOWLEDGMENTS I would like to thank Alex Bui, PhD, Lawrence R. Muroff, MD, FACR, and David T. Nelson for helping improve this article. REFERENCES 1. Enzmann DR, Feinberg DT. The nature of change. J Am Coll Radiol 2014;11:464-70. 2. Morrison I. The innovative imperative. H&HN Daily July 1, 2014. 3. Hofstadter D, Sander E. The analogical animal: the key to human cognition may well be the ability to compare one thing to another. The Wall Street Journal May 4-5, 2013:C3. 4. Taleb NN. Antifragile: things that gain from disorder. New York: Random House; 2012:177. 181. 5. Kaufman L. AtlantiCare and Geisinger health systems make merger plans official. Available at: http://www.healthexo.com/topics/finance/ atlanticare-and-geisinger-health-systems-make-merger-plans-official. Accessed September 29, 2014. 6. Demos T. Companies rush to join IPO surge. The Wall Street Journal March 7, 2014:A1. 7. Gould SJ. Wonderful life: the Burgess shale and the nature of history. New York: W. W. Norton; 1989:304-7. 8. Wagner A. Neutralism and selectionism. Nat Rev Genet 2008;9: 965-74. 9. Bak P. How nature works. New York: Springer-Verlag; 1996:1-7. 131, 146, 152. 10. Herman B. Consolidation could be next for academic medical centers. Available at: http://www.modernhealthcare.com/article/20140705/ MAGAZINE/307059964/consolidation-could-be-next-for-academicmedical-centers&template¼mobile. Accessed September 29, 2014. 11. Christensen CM. The innovator’s dilemma. Boston, Massachusetts: Harvard Business School Press; 1997:46-7. 90-1, 172-5.
348
12. Comstock J. Revisiting how Christensen’s “disruptive innovation” in healthcare means decentralization. Available at: http://www. healthcareinnovationssummit.com. Accessed September 29, 2014. 13. Thiel P. Zero to One: notes on startups, or how to build the future. Random House/Crown Business, NY; 2014:57-8. 14. Tellis GJ. The Columbus Effect in Business. Wall Street Journal October 13, 2014:A17. 15. E-commerce: same-day dreamers. The Economist May 3, 2014:61. 16. Barr A. Google’s new moonshot project: the human body. Available at: http://online.wsj.com/articles/google-to-collect-data-to-define-healthyhuman-1406246214. Accessed September 29, 2014. 17. Winter C. How three Germans are cloning the web. Bloomberg Businessweek February 29, 2012. 18. Silva E, Breslau J, Barr RM, et al. ACR white paper on teleradiology practice. J Am Coll Radiol 2013;10:575-85. 19. Praire E. vRad to acquire NightHawk. Available at: http://www.health careitnews.com/news/vrad-acquire-nighthawk. Accessed September 29, 2014. 20. Seong NJ, Kim B, Lee S, et al. Off-site smartphone reading of CT images for patients with inconclusive diagnoses of appendicitis from on-call radiologists. AJR Am J Roentgenol 2014;203:3-9. 21. Rockoff JD, Whalen J, Falconi M. Global drug industry about-face. The Wall Street Journal April 23, 2014:A1. 22. Pharmaceuticals. The Economist April 26, 2014:65. 23. Gottlieb S. A biotech lesson for big pharma mergers. The Wall Street Journal April 30, 2014:A15. 24. The new rules of attraction. Economist November 15, 2014:66-7. 25. Cassels: why some Pioneer ACOs are dropping out. Available at: http://www.advisory.com/daily-briefing/2013/07/01/cassels-on-pioneeracos. Accessed September 29, 2014. 26. Vasko C. Radiology and the ACO. Available at: http://www.radiology business.com/topics/policy/radiology-and-aco-ea. Accessed September 29, 2014. 27. Hotz RL. A future where bionics track your health. The Wall Street Journal April 22, 2014:D2. 28. Holms L. Should nurse practitioners be able to treat patients without physician oversight? The Wall Street Journal June 13, 2013. 29. Reddy S. Drugstores play doctor. The Wall Street Journal April 8, 2014:D1. 30. Beck M. Where does it hurt? The Wall Street Journal May 9, 2014: B1. 31. Zamosky L. The virtual doctor visit. The Los Angeles Times July 13, 2014:B3. 32. Crisp N, Chen L. Global supply of health professionals. N Engl J Med 2014;370:950-7. 33. MacMillan D. The fiercest rivalry in tech. The Wall Street Journal August 12, 2014:B1. 34. MacMillan D, Karmin C. New York’s supreme court hosts Airbnb case. The Wall Street Journal April 22, 2014:B5. 35. Enzmann DR. Radiology’s value chain. Radiology 2012;263:243-52. 36. Ante SE. Reboot: IBM turns to software. The Wall Street Journal June 13, 2010:B1. 37. New blood test could give early warning of Alzheimer’s disease. Available at: http://www.cambridge-news.co.uk/Cambridge/New-blood-test-couldgive-early-warning-of-Alzheimers-disease-20140708120221.htm#ixzz 376nsGOP3. Accessed September 29, 2014. 38. Lee CI, Enzmann DR. Measuring radiology’s value in time saved. J Am Coll Radiol 2012;9:713-7. 39. Winkler R. Google pushes its own content. The Wall Street Journal August 19, 2014:B1. 40. Spence M. The next convergence. New York: Farrar, Straus; 2011:72.
Journal of the American College of Radiology Volume 12 n Number 4 n April 2015
The Risks of Innovation in Health Care Dieter R. Enzmann, MDa Abstract Innovation in health care creates risks that are unevenly distributed. An evolutionary analogy using species to represent business models helps categorize innovation experiments and their risks. This classification reveals two qualitative categories: early and late diversification experiments. Early diversification has prolific innovations with high risk because they encounter a “decimation” stage, during which most experiments disappear. Participants face high risk. The few decimation survivors can be sustaining or disruptive according to Christensen’s criteria. Survivors enter late diversification, during which they again expand, but within a design range limited to variations of the previous surviving designs. Late diversifications carry lower risk. The exception is when disruptive survivors “diversify,” which amplifies their disruption. Health care and radiology will experience both early and late diversifications, often simultaneously. Although oversimplifying Christensen’s concepts, early diversifications are likely to deliver disruptive innovation, whereas late diversifications tend to produce sustaining innovations. Current health care consolidation is a manifestation of late diversification. Early diversifications will appear outside traditional care models and physical health care sites, as well as with new science such as molecular diagnostics. They warrant attention because decimation survivors will present both disruptive and sustaining opportunities to radiology. Radiology must participate in late diversification by incorporating sustaining innovations to its value chain. Given the likelihood of disruptive survivors, radiology should seriously consider disrupting itself rather than waiting for others to do so. Disruption entails significant modifications of its value chain, hence, its business model, for which lessons may become available from the pharmaceutical industry’s current simultaneous experience with early and late diversifications. Key Words: Health care, change, innovation, risk, experimentation, evolution J Am Coll Radiol 2015;12:342-348. 2015 Published by Elsevier Inc. on behalf of American College of Radiology
EXPERIMENTS AND RISKS In facing change, a previous recommendation was for radiology to experiment in the realm of the “adjacent possible” [1]. The easiest adjacent possible experiments are variations of current practices and business models. An evolution analogy reveals two different kinds of experiments, early and late diversifications, which have different risk profiles. It behooves radiology to understand, look for, and distinguish them, certainly within the imaging domain, but also inside and outside health care. We shall see in radiology that variants of current business models are late diversifications with low risk, whereas new business models are early diversifications with high risk. This article suggests considering higher risk experiments. The clarion call to organizations, professional groups, and individuals for “adaptation and innovation on a a
Department of Radiological Sciences, David Geffen School of Medicine at UCLA, Los Angeles, California. Corresponding author and reprints: Dieter R. Enzmann, MD, David Geffen School of Medicine at UCLA, Department of Radiological Sciences, 924 Westwood Blvd, Suite 805, Los Angeles, CA 90024; e-mail: [email protected].
massive scale” in response to health care changes entails risk [2]. From a system point of view, whether you call it “adaptation” or “innovation,” it is fundamentally experimentation [1]. The experiments are just that— experiments whether one perceives them as precipitated by an event, the Patient Protection and Accountable Care Act, or driven by scientific development, such as deoxyribonucleic acid (DNA) sequencing. Evolution continually experiments to create new species or subspecies, which incur risk when facing natural selection. That determines survival, replication, or extinction. Adaptation or innovation experiments in health care, including new organizational structures, new business models, new care models, new IT solutions, new technologies, and so on, incur risk when facing market, patient, and professional selection forces, which determine their fate. I tap into evolution using species and subspecies as an analogy to new and variant business models respectively [3]. All business models must generate cash, and they do so by using different components to construct a value chain. Different business models are defined by significant differences in value chain components. If radiologists ª 2015 Published by Elsevier Inc. on behalf of American College of Radiology
342
1546-1440/15/$36.00
n
http://dx.doi.org/10.1016/j.jacr.2014.09.026
work in capitated organizations as a cost center, in private practices as a profit center, or as salaried employees of universities, they are in different business models even though their clinical work is similar. Cash is generated differently in each. Our interest is in experiments that are capable of altering business models. Some innovations, such as tracking inventory by using radiofrequency identifiers and MR fingerprinting, just change business or clinical practices without altering the business model. Some innovations sound like new business models, but are not. As spectacular as CT and MR technologies were at their introduction, they did not disrupt radiology’s business models. They produced value chain variants, but cash was still generated in the same way, albeit in higher amounts. The variants were subspecies rather than new species, because the acquisition device component was only a variant; it still generated images, only differently. Sometimes experiments string together components into an entirely new value chain. In the past, a business model built around an Internet search engine did not exist. Many experiments were run; some failed (Netscape), and some survived (Google). Some survived by completely reassembling the advertising value chain, going from print area to clicks. These were new species, new business models. In medicine, a new business model example would be concierge private practice, in which real cash is generated by monthly retainers independent of service volume. Human experiments are attempts to change in response to or in anticipation of change. Experiments carry different risks depending on their intrinsic characteristics and on whether one runs or merely observes them. This article concentrates on the intrinsic nature and provenance of experiments. Paradoxically, experimentation generates not only risks but also options to counterbalance them [1,4]. Health care systems increasingly experiment to create options. Anticipating the need to manage population health risk, some experiment in care delivery using accountable care organizations (ACOs), while others experiment in consolidation to increase size and scale [5]. Health care entrepreneurs experiment with new companies evidenced by a recent spike in health care and biotechnology initial public offerings [6]. Radiology should ask, What is the risk profile of experiments surrounding us? Are we participants or observers? Evolution reveals two fundamental types of experiments, “early diversification” and “late diversification,” which differ intrinsically in character and in risk. The evolution analogy reveals important distinctions between them. Some detail is necessary because their temporal Journal of the American College of Radiology Enzmann n The Risks of Innovation in Health Care
sequence relative to selection forces is important in assessing risk (Fig. 1). Early and late refer to the relationship of the two types of experiments to each other, not to any time interval from what may be perceived as a triggering event. Assigning causality in complex systems is problematic, and any set of conditions can precede either or both diversifications [1].
MULTISCALE EVOLUTIONARY ANALOGY Using varying terminology, temporal sequences of design change in the evolution of complex systems have been identified [7,8]. The order of experiments matters. One sequence is based on speciation, nature’s way to innovate. It starts with early diversification, during which a remarkably large number of “experiments” generate a wide range of new species (Fig. 1). Another sequence occurs at the molecular level, with the creation of many new molecules, termed neutral diversity expansion [8]. Dramatic early design expansions inevitably face selection pressures, which unfortunately end on a down note, with most new designs disappearing. Species experience a “decimation” stage, during which an equally remarkable burst of extinction occurs, reflecting the high risk to participants (Fig. 1). At the molecular level, this selection is termed selective diversity contraction [8]. By appearing at multiple scales, this temporal sequence reveals its generality. Species designs surviving decimation form the “early standardization” stage and ultimately transition to late diversification, during which additional experimentation is limited to variations of surviving designs [7]. The design range is much narrower than early diversification. Late diversification, not being followed by decimation and having a low extinction rate, poses lower risk to participants and to observers, with one important exception noted below (Fig. 1). Transitions between diversifications and extinctions form steps in the punctuated equilibrium (PE) pattern of system change, meaning that they are unpredictable, more sudden than expected, and usually of indeterminate causality [1]. Many lineages in complex systems, be they animal species, cars, computers, TVs, medical services, and so on, trace out this sequence in a PE pattern [9]. LATE DIVERSIFICATIONS Let us work backward because late diversifications are easier to understand. These experiments carry lower risk to participants because the designs, rather than being untested new ones, are ones already tested by natural selection, ie, decimation. They have early diversification 343
Fig 1. This diagram shows the evolutionary sequence of experiments (new species) starting from early diversification through decimation to early standardization and finally to late diversification. Early diversification features prolific experiments, which decrease substantially during decimation to result in a few survivors that constitute early standardization. The color and number coding emphasize the large number of early diversification experiments, decreasing to a small number of survivors, which are both disruptive and sustaining types. Late diversification consists of surviving experiments expanding within a narrower design range (decimal numbers). The sequence highlights the relative changes in the number, type, and fate of experiments, which are analogous to experiments in health care. The icons are meant to reflect environmental factors related to, but not causal of, diversifications.
precedents and become variants, which tend to expand current business models. Radiology, for example, has created such low-risk diagnostic and treatment variants by incorporating CT and MR technologies into imaging subspecialties, by developing musculoskeletal ultrasound, and by performing image-guided tumor ablation. Although CT and MR technologies were low-risk, late diversification for radiology participants, they posed higher risks to observers, neurology, and general surgery. A striking example of late diversification is consolidation. Although it includes a mild form of extinction, the resulting business models are just larger variants of the consolidating entities. Consolidation occurs on many scales: a large health system merging with smaller hospitals, academic medical centers merging into large health systems, or smaller radiology groups merging with larger ones [5,10]. Consolidation, be it in airlines, hospitals, companies, or radiology groups (or societies), is classic late diversification because the underlying business models remain stable, the variation being primarily increased size and scope. When a small radiology group is absorbed by a larger group, but the cash-generating profit 344
center model remains dollars per examination, it may feel “disruptive.” It is not in Christensen’s sense, because the cash-generating business model is unchanged. If a small group were absorbed into a capitation, cost-center, salaried-radiologist model, it would be “disrupted” because cash would be generated in ways other than dollars-perexamination “piecework.” At the system level, this is still late diversification because the group has simply joined an existing radiology model variant, which once was disruptive. Observers in late diversification can face high risk. Radiology’s cottage industry is waning.
EARLY DIVERSIFICATIONS Early diversifications are more complicated and of higher risk for two reasons: they have a high mortality rate because of follow-on decimation and because they have a higher likelihood of generating disruptive new designs. The rapid appearance of multiple new designs in any health care segment heralds early diversifications. It is prudent to be alert to their development because they are harbingers of potentially uncomfortable change by germinating truly disruptive business models. Journal of the American College of Radiology Volume 12 n Number 4 n April 2015
There is a simplified relationship between these diversifications and Christensen’s innovation concepts. Early diversifications are more likely to produce “disruptive innovations,” whereas late diversifications tend to generate “sustaining innovations” [11,12]. Disruptive innovations are new, different species that disrupt rather than fit into existing business models, just as new species do not stay within their ancestor species. They add considerable risk to the environment. Sustaining innovations, by being incorporated into existing business models, add less risk [11,12]. Christensen’s disruption and sustaining terminology refers primarily to the innovation’s effect on business models, its value chain components, and how cash is generated, not on the technologies per se and not on disruption of individual careers. Another feature of this early-late diversification sequence merits attention. Survivors of early diversification are likely to be a mixture of “sustaining” and “disruptive” experiments (Fig. 1). Sustaining ones can be incorporated into existing business models and, hence, are of low risk. Disruptive early diversification survivors are higher risk because they cannot be incorporated. When disruptive survivors enter late diversification, their disruptive effects are amplified. It is often the “later mover” that is highly successful as a presumably wiser variant of the “first mover” in early diversification. [13] This is also known as the “Columbus Effect,” one worth knowing [14]. As noted earlier, mere observers can face higher risk. Identifying and monitoring the temporal relationships of experiments is central to assessing their risks.
DIVERSIFICATION SEQUENCE AND TIMING The well-publicized dot-com era illustrates the diversification sequence. An initial burst of a wild number of creative dot-com companies, most with untested business models, formed the early diversification. This “boom” was followed by decimation, the dot-com “bust.” Some survivors were disruptive—Google, Amazon, Expedia, Facebook, and others. Late diversification appeared when dotcom businesses again expanded as variations on successful surviving designs. Late diversification variants, such as Amazon Fresh and Google Shopping Express, are still appearing [15]. Late diversification of disruptive models can bring new entrants into health care. Google, in exploring its adjacent possible (grand-scale computing), is attracting vast, new resources into health care by assembling knowledge generally residing outside traditional medicine [16]. Late diversification of disruptive survivors intensifies their disruption (see corporate teleradiology). Journal of the American College of Radiology Enzmann n The Risks of Innovation in Health Care
Late diversifications can be quite successful and profitable. A German group simply copies successful American dot-com companies and expands them around the globe [17]. This is classic late diversification, during which companies essentially copy decimation survivors. Pharmaceutical companies (pharma) have done well by expanding “me too” drugs, drugs that have survived decimation. Late diversifications are constantly playing out in different arenas, ranging from travel to pharmaceuticals to health care. While they may be of lower risk to participants, they can be high risk to observers. Radiology faced this sequence on a smaller scale when it encountered early diversification in teleradiology. Decimation has waned, leaving survivors as a mix of disruptive and sustaining models [18,19]. Sustaining models, having been incorporated into radiology practices (night coverage), are already entering late diversification with mobile variants [20]. A disruptive, corporate venturebacked form of teleradiology (VRad, Radiology Partners, etc) has survived. Although it seems to be an adjacent possible, its provenance reveals it to have arisen outside of radiology, and it has a different value chain that allows hospitals to essentially outsource radiology. This model, as with other disruptors upon entering late diversification, as evinced by consolidation efforts, could increase its disruption. Variants are appearing, which bundle radiology with other subspecialties, expanding the scope of outsourcing and clearly exhibiting a different value chain and disruptive business model. Complaining about disruptive survivors and hoping they go away is unrealistic. A compelling industry-wide, simultaneous interplay between early and late diversifications is unfolding in pharma [21,22]. Early diversification has formed in biotechnology, driven by the fundamental shift to the molecular redefinition of diseases and derivative-targeted therapies. Tiring of shrinking profits from late diversification me-too drugs, pharma is availing itself of proliferating biotechnology companies to rejuvenate lagging innovation in their own research and development [23]. Forming various relationships with biotechnology companies allows pharma to parry some early diversification high risk. This early diversification, based on fundamental changes in science and technology, creates innovation opportunities in drugs, biologics, and cell therapy for pharma, which is simultaneously engaged in a classic late diversification activity, consolidation, buying innovative companies and “asset-swapping” to increase economies of scale in more focused, specialized disease areas [21-24]. To what extent early diversification biotechnology survivors will sustain or disrupt pharma is not yet clear. 345
Health care sectors experience early and late diversifications at different times, sometimes concurrently. Consolidation of health care systems is late diversification, whereas ACO experiments are early diversifications [5,25]. The recent “explosion” of ACOs (>600) indicates early diversification in these business and patient care models. The risk is high, as a large number will fail during decimation [25]. The PE change pattern means timing of failure is problematic, but health care systems and radiology groups would be wise to track both sustaining and disruptive ACO survivors [26].
SPOTTING EARLY DIVERSIFICATION Where do other early diversifications lurk? Christensen suggests that new experiments will proliferate as increasingly sophisticated care migrates from expensive to less expensive venues; from hospitals to clinics, to offices, to homes, and eventually to mobile sites, spurring decentralization and deprofessionalization of medical care [27]. Technology, medical skills, and medical knowledge will move centrifugally from long-lived standard models. Radiology has experienced this with handheld ultrasound. Early diversification will become more apparent as new patient care designs appear outside traditional venues [28,29]. Retail medical clinics in drug stores, “hospitals at home,” and “home palliative” care are early examples. Retail medicine is disruptive. Many experiments use cyberspace instead of typical physical assets [29]. New care designs and business models in telemedicine, such as “e-health” (e-referral, e-consult services) and “telehealth” (tele-ICU and tele-stroke services) are forming early diversifications [30]. The same Internet, digitization, and social media drivers that produced new business models in other industries will do so in health care. Early diversifications interact, so it is not surprising to see early diversifications appearing in flexible, wireless, mobile, biocompatible, and biodegradable “bionics,” accelerating medical decentralization and cyberspace applications [27]. The interaction allows new combinations in care design, which will amplify prephysician care by expanding virtual office or pre-emergency room care [31]. These early diversifications incorporating mobility, the home, wireless monitoring, and care by nonphysician professionals will accelerate a deprofessionalization trend [28,32]. These, like other early diversifications, face decimation. The survivors can be either sustaining or disruptive to current health care models. Disruptive survivors are likely to congregate around prephysician care involving 346
nontraditional clinical sites, paramedics, and midlevel professionals, some of whom will come from outside the medical profession. Strategic and tactical analysis is needed to deal with expanding deprofessionalism [32]. Radiology needs to be alert, as nonphysicians will increasingly need to order imaging tests appropriately. It is beyond the scope of this paper to spell out how radiology should respond to decentralization and deprofessionalization.
EARLY DIVERSIFICATION SECONDARY EFFECTS Unusual and unexpected value chain components may materialize in early diversification. Taxicab businesses did not expect competing business models, such as Uber and Lyft, to pair smart phones with a new component, private cars [33]. Budget hotels did not expect their customers to be linked via the Internet to a new value chain component, privately owned rooms, until Airbnb appeared [34]. The new “shared economy” can recruit decentralized components to new value chains to compete against longlived, staid entities. This represents the democratization of physical assets. Who cares? Late diversification of Uber could include moving things (packages, medical supplies) in addition to or with people. Although it is pure speculation, Uber and Airbnb could add personalized services to personalized medicine. RELEVANT EARLY DIVERSIFICATIONS Prodigious, new molecular diagnosis designs driven by molecular medicine and big data herald early diversification. Genomic and molecular diagnostic testing is undergoing “explosive” early diversification in cancer detection, prognosis, prediction, and monitoring. Hidden from many radiologists is dramatic progress in using genomics, epigenomics, proteomics, circulating tumor cells, circulating tumor DNA and circulating microribonucleic acid (miRNA) to achieve nonradiologic diagnosis and monitoring of treatment. Some sophisticated, successful fluid- and cell-based diagnostic and monitoring designs, will survive decimation. Radiology must attend to such survivors that could sustain or disrupt its value chain. Molecular diagnostics, when fully digitized, will spawn new competitive designs in “telediagnostics,” just as image information became more competitive when distributed via teleradiology. University of California, Los Angeles, telepathology is up and running in China. Monitoring molecular diagnostics decimation to identify surviving models will be critical in assessing risk to, and impact on, diagnostic radiology. This should stimulate strategic thinking because potential disruption Journal of the American College of Radiology Volume 12 n Number 4 n April 2015
depends on the “host’s response,” a point emphasized by Christensen.
RADIOLOGY’S CHALLENGES Radiology must digest Christensen’s innovator’s dilemma [11]. Although integrating sustaining technologies is relatively straightforward, integrating potentially disruptive ones is not. One reason is that sustaining technologies tend to increase profitability, whereas disruptive technologies often decrease it. A disruptive survivor might well reduce profit from radiology’s current transactional professional service model by challenging one of its value chain components. Whole new diagnostic value chains might be constructed outside of radiology [35]. The issue may come to radiology’s willingness to “disrupt” its value chain or wait for others to do so. Incorporating new elements into an existing value chain has its challenges, but IBM showed that it can be done when it expanded its hardware business model to encompass analytic software [36]. IBM chose to painfully disrupt itself internally. The company is still called IBM, but its business model is different. Incorporating disruptive technology often means a reformulation of the business model value chain, but that is preferable to its extinction. Moving from “volume to value” certainly entails change in radiology’s high-volume, transactional business model. By the way, “volume and value” are not mutually exclusive concepts. We briefly consider 3 potential disruptive scenarios. One radiology approach would be to change only tactically and remain focused on high-volume transactional imaging. This risks not only commoditization but also substantial outside disruption by molecular diagnosis. The risk is that blood tests and cell-based test components of new value chains will diagnose and monitor cancer (breast, prostate, lung, etc) or Alzheimer’s disease and substantially replace commodity imaging [37]. Diagnostic value will move out of radiology if it limits itself to isolated, transactional, diagnostic imaging. A second approach is to evolve into an information business, using late diversification. This requires some modification, but not reinvention of the current value chain to create robust, longitudinal, integrated, imagingbased diagnostic reports. Decision support fits in by delivering this value through pretest diagnostic information. However, image-based information businesses could easily take root outside of radiology, in part because other subspecialties may not face the innovator’s dilemma of decreased profit in assembling new value chains [19]. Radiology should be attentive to early diversifications in other clinical, image-based information businesses. It Journal of the American College of Radiology Enzmann n The Risks of Innovation in Health Care
would do well to push its own late diversification while it has capital to do so. Consolidation in radiology will facilitate this and the following model by providing organizational skills and capital. A third possibility is self-disruption by expanding the imaging information business to encompass nonimaging, molecular diagnostic information in comprehensive diagnostic business services and reports that more broadly facilitates clinical decision making, while adding value with time-saving, diagnostic accuracy [38]. This requires significant modification to the current value chain, probably to the point of a new business model, to integrate new molecular diagnostic information with analytic imaging, while putting both into clinical, scientific, and social determinants’ context. In such an endeavor, radiology assumes some higher risk inherent in early diversifications. Even the eminently successful Google has moved from its famed search (diagnostic) business to an information business via its Knowledge Graph [39]. This means that radiology would explore internal disruption à la IBM and maybe learn lessons from pharma’s relationships with biotechnology. This business model could go further in more closely aligning radiologic and pathologic diagnostic services by linking “tele” and physically located integrated diagnostic services. This does not require merger or loss of identity. This widens the design range of business models in early diversifications of integrated diagnostic information businesses. Attaching interventional radiology adds even more potential.
DIVERSIFICATION AND RISK All diversification experiments represent optionality, which is critical to all markets and economies [4,40]. Radiology must track all medical early and late diversifications. Understanding the temporal relationship and sequence of design experiments helps categorize them and their risks. Neither early nor late diversifications are controllable at the system level, but radiology’s reaction to them depends on whether you are a participant or observer. The highest risk is to participants in early diversifications and to observers of late diversifications in disruptive survivors. Early and late diversifications pose risks and offer opportunities to radiology. Radiology must participate in its own late diversification, because exploring the adjacent possible generates low-risk business model variants. Radiology should not only monitor molecular diagnoses but consider internal disruption to explore new business models, perhaps triggering some early diversification experiments. Should radiology choose 347
to disrupt itself, its ingenuity, expertise, and cleverness will need to be exercised.
TAKE-HOME POINTS n
n
n
n
n
Health care “adaptations” and “innovations” are equivalent to experiments that can be qualitatively categorized. An evolution analogy forms the basis of this categorization into early and late diversifications. Early diversifications exhibit prolific experiments, which carry high risk. Late diversifications exhibit a much more limited range of experiments, which carry lower risk. Radiology will be exposed to both early and late diversifications, so identifying them is informative because the former harbor the potential for disruption and the latter offer potential for growth.
ACKNOWLEDGMENTS I would like to thank Alex Bui, PhD, Lawrence R. Muroff, MD, FACR, and David T. Nelson for helping improve this article. REFERENCES 1. Enzmann DR, Feinberg DT. The nature of change. J Am Coll Radiol 2014;11:464-70. 2. Morrison I. The innovative imperative. H&HN Daily July 1, 2014. 3. Hofstadter D, Sander E. The analogical animal: the key to human cognition may well be the ability to compare one thing to another. The Wall Street Journal May 4-5, 2013:C3. 4. Taleb NN. Antifragile: things that gain from disorder. New York: Random House; 2012:177. 181. 5. Kaufman L. AtlantiCare and Geisinger health systems make merger plans official. Available at: http://www.healthexo.com/topics/finance/ atlanticare-and-geisinger-health-systems-make-merger-plans-official. Accessed September 29, 2014. 6. Demos T. Companies rush to join IPO surge. The Wall Street Journal March 7, 2014:A1. 7. Gould SJ. Wonderful life: the Burgess shale and the nature of history. New York: W. W. Norton; 1989:304-7. 8. Wagner A. Neutralism and selectionism. Nat Rev Genet 2008;9: 965-74. 9. Bak P. How nature works. New York: Springer-Verlag; 1996:1-7. 131, 146, 152. 10. Herman B. Consolidation could be next for academic medical centers. Available at: http://www.modernhealthcare.com/article/20140705/ MAGAZINE/307059964/consolidation-could-be-next-for-academicmedical-centers&template¼mobile. Accessed September 29, 2014. 11. Christensen CM. The innovator’s dilemma. Boston, Massachusetts: Harvard Business School Press; 1997:46-7. 90-1, 172-5.
348
12. Comstock J. Revisiting how Christensen’s “disruptive innovation” in healthcare means decentralization. Available at: http://www. healthcareinnovationssummit.com. Accessed September 29, 2014. 13. Thiel P. Zero to One: notes on startups, or how to build the future. Random House/Crown Business, NY; 2014:57-8. 14. Tellis GJ. The Columbus Effect in Business. Wall Street Journal October 13, 2014:A17. 15. E-commerce: same-day dreamers. The Economist May 3, 2014:61. 16. Barr A. Google’s new moonshot project: the human body. Available at: http://online.wsj.com/articles/google-to-collect-data-to-define-healthyhuman-1406246214. Accessed September 29, 2014. 17. Winter C. How three Germans are cloning the web. Bloomberg Businessweek February 29, 2012. 18. Silva E, Breslau J, Barr RM, et al. ACR white paper on teleradiology practice. J Am Coll Radiol 2013;10:575-85. 19. Praire E. vRad to acquire NightHawk. Available at: http://www.health careitnews.com/news/vrad-acquire-nighthawk. Accessed September 29, 2014. 20. Seong NJ, Kim B, Lee S, et al. Off-site smartphone reading of CT images for patients with inconclusive diagnoses of appendicitis from on-call radiologists. AJR Am J Roentgenol 2014;203:3-9. 21. Rockoff JD, Whalen J, Falconi M. Global drug industry about-face. The Wall Street Journal April 23, 2014:A1. 22. Pharmaceuticals. The Economist April 26, 2014:65. 23. Gottlieb S. A biotech lesson for big pharma mergers. The Wall Street Journal April 30, 2014:A15. 24. The new rules of attraction. Economist November 15, 2014:66-7. 25. Cassels: why some Pioneer ACOs are dropping out. Available at: http://www.advisory.com/daily-briefing/2013/07/01/cassels-on-pioneeracos. Accessed September 29, 2014. 26. Vasko C. Radiology and the ACO. Available at: http://www.radiology business.com/topics/policy/radiology-and-aco-ea. Accessed September 29, 2014. 27. Hotz RL. A future where bionics track your health. The Wall Street Journal April 22, 2014:D2. 28. Holms L. Should nurse practitioners be able to treat patients without physician oversight? The Wall Street Journal June 13, 2013. 29. Reddy S. Drugstores play doctor. The Wall Street Journal April 8, 2014:D1. 30. Beck M. Where does it hurt? The Wall Street Journal May 9, 2014: B1. 31. Zamosky L. The virtual doctor visit. The Los Angeles Times July 13, 2014:B3. 32. Crisp N, Chen L. Global supply of health professionals. N Engl J Med 2014;370:950-7. 33. MacMillan D. The fiercest rivalry in tech. The Wall Street Journal August 12, 2014:B1. 34. MacMillan D, Karmin C. New York’s supreme court hosts Airbnb case. The Wall Street Journal April 22, 2014:B5. 35. Enzmann DR. Radiology’s value chain. Radiology 2012;263:243-52. 36. Ante SE. Reboot: IBM turns to software. The Wall Street Journal June 13, 2010:B1. 37. New blood test could give early warning of Alzheimer’s disease. Available at: http://www.cambridge-news.co.uk/Cambridge/New-blood-test-couldgive-early-warning-of-Alzheimers-disease-20140708120221.htm#ixzz 376nsGOP3. Accessed September 29, 2014. 38. Lee CI, Enzmann DR. Measuring radiology’s value in time saved. J Am Coll Radiol 2012;9:713-7. 39. Winkler R. Google pushes its own content. The Wall Street Journal August 19, 2014:B1. 40. Spence M. The next convergence. New York: Farrar, Straus; 2011:72.
Journal of the American College of Radiology Volume 12 n Number 4 n April 2015
