Are ICTs disrupting the health knowledge economy?
A senior official of a donor agency recently asked one of the authors whether their agency should take digital health into account in the provision of support for health system strengthening. There is no simple answer to this question. As discussed above, most donor-funded investments in digital health have not yielded big health benefits. However, the studies in Bangladesh and China found evidence of a variety of emergent organisations, inter-organizational relationships and business models, suggesting the possibility of future changes at scale. Elsewhere too, the growing tendency of large American internet companies to offer health-related services, in the form of fitness-related devices, systems to support people with chronic illnesses and so forth, indicates their recognition of potentially big business opportunities.
We suggest that the relatively slow development of digital health reflects the following special characteristics of the health knowledge economy. First, a useful health service in a country like China or Bangladesh needs to include a combination of diagnosis, expert advice and drugs. This is likely to involve relationships between organisations with different roles, responsibilities and values in the health and communication sectors. Second, health systems often combine public funding of some services with out-of-pocket payment for others, so that business models need to combine funding from multiple sources. Third, well-functioning health systems must be embedded in institutional arrangements for accountability and efforts to reshape markets and associated regulatory frameworks are complex and politically charged. The need to achieve progress on each of the above, has reduced the speed with which new technological possibilities have been translated into large-scale changes in the way health services are organised [1].
Debates about the desirability of further investment in digital health reflect the different time horizons and attitudes to risk of large ICT companies, venture capital funds, donor agencies and governments. Some large companies take a long view aimed at creating a niche in a rapidly changing economy and are not afraid to incur significant losses in the short term. For example, Google invested heavily in creating detailed city maps, to produce a digital infrastructure. Other companies have experienced losses, whilst building an enormous customer base, which eventually generated profits. A number of companies are investing to establish future positions in the health knowledge economy. Donor agencies and governments, on the other hand, have a shorter time horizon, preferring low-risk investments likely to yield immediate benefits. Recent efforts to coordinate donor investments in digital health reflect the beginnings of a strategic shift, with mechanisms such as the Health Data Collaborative (HDC) and USAID’s Digital Health Initiative emerging to support country-level capacity development [44]. However, these initiatives have largely focused on public sector developments. We argue that governments and donor agencies need to adopt a longer-term perspective in their interventions in the health knowledge economy to support developments in the public and private sectors that build local capacity and meet the needs of the poor [45].
Regulatory challenges in the health knowledge economy
The developments described above pose big regulatory challenges and governments need to modify and strengthen institutional arrangements to address them. This will involve changes to regulations that block the development of potentially beneficial services. For example, there may be rules that reserve the right to provide advice on the use of many drugs to a licensed doctor, even though large numbers of people buy these products from local drug sellers without a prescription. It will also involve additions to the regulatory framework.
One challenge arises from the increasing use of treatment guidelines. This builds on previous investments by the governments of several advanced market economies to support the generation and synthesis of knowledge to inform evidence-based medical care. This knowledge has been incorporated into diagnosis and treatment guidelines in text books and manuals used by providers of medical care and increasingly by the general public. The use of guidelines for a growing number of treatment decisions is diminishing the role of clinical judgement in the management of many common health problems [29]. This is likely to reduce the capacity of doctors to control the use of drugs for these conditions. Other approaches will be needed to encourage people to use them appropriately.
The translation of treatment guidelines into computerized decision-making algorithms is increasing their influence. Algorithms that provide advice on the basis of answers to simple questions and diagnostic information, such as blood pressure, temperature and blood sugar can substantially increase the ability of people to make use of treatment guidelines. Their incorporation into easy-to-use smartphone apps will further simplify their use.
The producers of algorithms can influence the decisions of many people. Their underlying assumptions, cultural understandings and financial interests are likely to affect the content of the algorithms [46]. For example, if the questions an algorithm asks and the data it uses focus on the relationship between pharmaceutical treatment and health, it will provide advice on the choice and dosage of drugs, but not on other factors that influence health, such as diet, lifestyle and exposure to environmental toxins. This could lead to an excessive reliance on drugs to control risk factors for non-communicable diseases. Another example is the establishment of a new diagnosis. This could have major financial implications if it were to justify a particular treatment regime [47, 48]. In some cases, pharmaceutical companies have attempted to influence the definition of a diagnosis as a strategy for creating markets for their products [49]. This applies especially to the growing number of “lifestyle”-related uses of pharmaceuticals for altering moods, increasing libido, building or losing weight, increasing athletic performance and so forth. The boundary between “lifestyle” and “medical” decision-making algorithms is becoming increasingly difficult to define in the face of the rising burden of chronic non-communicable diseases, including mental illness, and of measures to reduce symptoms and control risk-factors, such as hypertension.
The public domain has tended to lead the development of treatment guidelines, with a substantial involvement of organised medical professions and training and research institutions. The character of many digital health applications, in contrast, is largely opaque. A study of the top-rated medical and health applications available through Apple’s App store reveals that most make unsubstantiated claims to medical authority, leaving both expert and lay users in the dark about where the information and advice they provide is derived from and whether the producers of the applications have links with companies that sell pharmaceuticals or diagnostic equipment [50, 51].
Two developments are likely to increase the importance of health-related algorithms and the governance challenges associated with them. The first concerns the increasing availability of low-cost diagnostic technologies in the form of smartphone and computer attachments or standalone devices meant to be used by patients, front line medical staff or other suppliers of drugs and health-related commodities. The second concerns automated processes for updating the content of original algorithms on the basis of incoming data (machine learning). The continuous collection of data that links indicators, such as blood pressure and blood sugar to specific treatments creates the possibility of collecting a large body of data that could guide future treatment regimens. This raises issues regarding the accuracy of the data collected and also the specific “research” questions that drive the data collection. The organisations that own the data and use them to revise algorithms will accrue increasing influence.
The availability of health-related algorithms could become especially important in countries with pluralistic health systems, where people take a lot of responsibility for their own health care and access to trustworthy treatment guidelines could be particularly useful [28, 52]. However, inappropriate guidelines or apps aimed at generating new markets for pharmaceuticals or diagnostic equipment pose a particular risk in these countries due to poor regulation and lack of consumer protection. Lewis and Wyatt [53] present a framework that uses a combination of usage scenarios, contextual factors and app complexity to assess the risk of harm from a so-called health app.
There are a variety of potential regulatory approaches for addressing these issues. Some health problems may be relatively minor and regulators could follow a strategy of “buyer beware” by strengthening consumer protection and focusing regulations on informing the public about the contents of any products and preventing false and misleading claims. In other cases, major deleterious consequences could arise from the provision of misinformation and inappropriate treatment, making information asymmetry an important consideration. Governments and organised professional and expert bodies have an important role to play in these cases. This may involve the development of treatment guidelines and algorithms as national or global public goods to be made available for use by both public and private health service providers.
Alternatively, it may be appropriate to regulate the production and use of treatment algorithms. Mechanisms are needed to ensure that the organised professions and the medical and nursing schools associated with them, pharmaceutical companies and newer entrants in the health knowledge economy are accountable for the advice they provide. This will require strong strategic leadership by national governments and international organisations.
In China and Bangladesh, the links between providers of medical advice and treatment and suppliers of pharmaceuticals are complex. In Bangladesh, informal village doctors and drug sellers, working outside the regulatory framework, are important sources of both advice and treatment, especially for the poor. In that country, there have been several attempts to link a digital health company with village doctors, as a means of improving the practice of the latter. In China, where the formal retail sector developed rapidly since the beginning of the transition to a market economy in the 1980s, several online shopping companies are now making large investments to establish links with local shops and create a capacity to make next-day deliveries. These examples illustrate how companies can use information technologies to enable people to link to a network of small businesses, which can supply goods and services. This kind of network that links local suppliers of pharmaceuticals to a source of treatment algorithms and a means for monitoring drug quality could provide a low-cost way to meet health care needs of the relatively poor. It might involve an internet platform and pharmaceutical wholesalers and retailers. But, there are governance risks concerning the management of conflicts of interest between those of clients, who are seeking evidence-based and cost-effective ways to deal with health problems, and businesses, whose revenues depend on the volume of sales of pharmaceuticals and diagnostic devices. Governments will need to play new and challenging stewardship roles as these changes go to scale.
Strengthening public health in the interests of the poor: Supporting transitions in the health knowledge economy
The Bangladesh and China cases illustrate the different kinds of organisation that are becoming involved in the health knowledge economy and the variety of new kinds of partnership that are being established. It is possible to envisage quite different pathways of development as new possibilities for organising access to health services have impact at scale. Some pathways could result in big increases in access to appropriate advice and effective treatment. Others, however, could reflect the interests of powerful stakeholders, such as the producers and distributors of diagnostic tests and drugs, and encourage unnecessary use of these products. The “choice” of development pathways will be strongly influenced by political factors and the actions of government and other stakeholders to establish new kinds of partnership and reform the regulatory framework.
Developments in the broader knowledge economy are likely to influence the health knowledge economy. In the past 20 years a number of information companies have grown very quickly to take advantage of short-term monopoly positions, becoming very large corporations that combine provision of access to the internet, production of knowledge content and the diffusion of content through the mass media [54, 55]. Other companies like Google, Facebook, Amazon and Alibaba have become actors in political negotiations about the shape of knowledge industry markets [56–58]. These companies are seeking niches in the health knowledge economy and will eventually establish relationships with organisations that deliver health services and supply health-related commodities. This opens up the possibility of the emergence new types of powerful organisation that seek to influence the regulatory framework.
In China, pharmaceutical producers, retail pharmacy chains and large online shopping platforms are making substantial investments to develop digital health services. They are engaging with the organised health system and health regulatory agencies to influence the institutional arrangements. In Bangladesh, retail chains are building their presence in a market place that is changing more slowly, certainly in terms of consumer demand. Much of the investment in Bangladesh has been in the form of grants by donor agencies to relatively small ICT companies and by mobile phone operators seeking added-value lines of business. As Bangladesh’s consumer base increases with economic growth, these increased market opportunities are likely to attract larger players in the knowledge industry markets. One particularly important trend is the increasing role of digital platforms as intermediaries between different stakeholders. In China and Bangladesh, health sector intermediaries include websites of existing brick-and-mortar suppliers of drugs and health-related products, telemedicine platforms and online platforms that offer a wide variety of goods and services. These intermediaries are likely to gain influence as gatekeepers to information and advice.