Developed Countries

Health Tech

The reality check: Making health tech work in real healthcare settings.

What is it that we like about hos­pi­tals? We have been putting sick peo­ple togeth­er in one build­ing since the days of ancient Greece. And hos­pi­tals were the test­ing grounds for 20th-cen­tu­ry med­ical advances, such as X‑rays, insulin and antibi­otics. But let’s con­sid­er some of the com­pelling argu­ments against them.

Hos­pi­tals can be hard to get to, espe­cial­ly for those most dis­ad­van­taged. And they kill peo­ple. Burnout, which has been found in a sur­vey to affect up to half of all doc­tors in Ger­many, Spain, Por­tu­gal, the UK and US, is a major cause of med­ical errors. Wards and oper­at­ing the­aters are also breed­ing grounds for dead­ly “super­bugs”.

Once a hos­pi­tal has been built, it must be filled with patients to be eco­nom­i­cal­ly viable, even in pub­licly fund­ed health­care sys­tems. So hos­pi­tals pro­vide an incen­tive for unnec­es­sary treat­ment and a dis­in­cen­tive for upstream pre­ven­tive health­care. Despite these facts, hos­pi­tals remain the bedrock of the world’s health­care sys­tems.

Tra­di­tion­al hos­pi­tals were cen­ters for expen­sive machines and hard-to-acquire med­ical spe­cialisms, but the log­ic of today’s fast evolv­ing, data-enabled tech­nolo­gies means that med­i­cine will increas­ing­ly move away from ded­i­cat­ed build­ings and towards patients.


Niels van Namen, glob­al health­care leader of Swiss-based CEVA Logis­tics, points out that a tsuna­mi of med­ical inno­va­tion has hit the devel­oped world in the last two decades. Cell and gene ther­a­pies sup­port tar­get­ed, pre­ci­sion med­i­cine; drugs are becom­ing smarter. Mean­while, AI has accel­er­at­ed diag­no­sis, and tech­nolo­gies such as drones, vir­tu­al real­i­ty, robot­ics and 3D print­ing are increas­ing­ly being used for med­ical appli­ca­tions.

To these fac­tors should be added the com­mu­ni­ca­tion rev­o­lu­tion of both smart­phones and wear­able sen­sors and med­ica­tion sys­tems, con­nect­ed by the “inter­net of healthy things”. For many peo­ple, van Namen argues, dig­i­tal con­nec­tiv­i­ty could remove the need for human con­tact in health­care. And yet, in the “advanced” med­ical sys­tems of most OECD coun­tries, we still trav­el miles to a hos­pi­tal for a rou­tine test. It makes no sense.

There would still be hos­pi­tals, in van Namen’s ide­al world, but they would be used only for pedi­atrics, inten­sive care, surgery and imag­ing. They would be com­ple­ment­ed by agile, mobile health cen­ters, which would also deliv­er some acute care. For the major­i­ty of peo­ple, includ­ing some with acute con­di­tions, there would be no need to go to the hos­pi­tal any­more – they would receive most of their care in their own homes.


Dri­ven by hos­pi­tal-based invest­ment, the cost of health­care is ris­ing. On aver­age, since 2013, annu­al per capi­ta health spend­ing across the OECD has grown by 2.4 per­cent. In the US it has increased six­fold over the last four decades, from $1,832 per capi­ta in 1970 to $11,172 in 2018.

Van Namen argues that a home-based health­care sys­tem could be up to ten times cheap­er than one deliv­ered from tra­di­tion­al hos­pi­tals with their high fixed costs. The short-term cycles of polit­i­cal deci­sion-mak­ing are the fac­tor that leads, para­dox­i­cal­ly, to inflex­i­ble, long-term invest­ment in hos­pi­tals.

In the future we’ll be ready to deliver more flexible care.

Rachel Dunscombe
CEO, NHS Digital Academy, UK

Admit­ted­ly, some progress has been made. Ger­many has recent­ly passed the world’s first leg­is­la­tion allow­ing doc­tors to pre­scribe health apps. We have pills con­tain­ing microchips and smart­watch­es that can mon­i­tor heart rate to a diag­nos­tic stan­dard using high­ly accu­rate elec­tro­car­dio­gram (ECG) track­ers. But we have a long way to go. The US Food and Drug Admin­is­tra­tion has only recent­ly approved wear­able med­ical devices.

Dr. Greg Parston is a vis­it­ing pro­fes­sor at the Insti­tute of Glob­al Health Inno­va­tion, based at Impe­r­i­al Col­lege, Lon­don. He notes that, before Covid-19, the aver­age imple­men­ta­tion peri­od between an innovation’s arrival and its appli­ca­tion in med­ical set­tings was com­mon­ly cit­ed as 17 years. One prob­lem, he says, is the time required by clin­i­cal tri­als: “It can be argued that they take far too long – we already have the next inno­va­tion before we’ve com­plet­ed the last tri­al.”


A study by Parston for the World Inno­va­tion Sum­mit for Health looked at fac­tors which slow down the adop­tion of med­ical inno­va­tion. Its inter­na­tion­al case stud­ies includ­ed pro­grams to pro­mote road safe­ty, vit­a­min use and vac­ci­na­tion and to diag­nose infant HIV; a health insur­ance sys­tem; and the adop­tion of dig­i­tal imag­ing for X‑rays. The main fac­tors slow­ing inno­va­tion, it found, were clin­i­cal bias, delays in the pub­li­ca­tion of research find­ings, and resis­tance to change.

“One issue is that the clin­i­cal lead­er­ship in most insti­tu­tions is old men. If you begin focus­ing on younger clin­i­cians, I think you’ll find a quick­er pace of change,” says Parston. “Also, doc­tors are taught to doubt. That is an impor­tant part of diag­no­sis. And doubt plays a big role in their sus­pi­cion of new ways of work­ing.”

qXR is a chest X-ray screening tool developed by Mumbai start-up It uses deep learning to automate and speed up the process.QURE.AI

Rachel Dun­scombe is CEO of the Nation­al Health Service’s (NHS) Dig­i­tal Acad­e­my in the UK and a mem­ber of a small group of pro­fes­sion­als advis­ing the UK gov­ern­ment on dig­i­tal tech­nol­o­gy. As Arch Col­lab­o­ra­tive lead for research com­pa­ny KLAS on refin­ing elec­tron­ic health records, she is in a good posi­tion to assess what has slowed down the dig­i­tal health­care rev­o­lu­tion. “It’s not skills. The dig­i­tal pro­fes­sion­als in health­care are in many cas­es where they need to be,” she says.

“In the UK, pro­cure­ment has been a drag fac­tor, but we are start­ing to improve that with new dynam­ic frame­works.” It is also not the tech industry’s lack of will­ing­ness. The main prob­lem, says Dun­scombe, is the lega­cy IT sys­tems that lock in data and can­not talk to each oth­er. “The aver­age orga­ni­za­tion has some­thing like 700 dif­fer­ent sys­tems,” she says. Parston iden­ti­fied four key fac­tors need­ed to speed up change: vision and strat­e­gy, a spe­cif­ic agency to pro­mote dif­fu­sion, ded­i­cat­ed fund­ing, and effec­tive com­mu­ni­ca­tion. When all are in place, the results can be impres­sive.


The NHS has invest­ed in tech­nol­o­gy through its high-tech off­shoot, NHSx, its Dig­i­tal Acad­e­my and the NHS Inno­va­tion Accel­er­a­tor. The gov­ern­ment has devot­ed almost $6.5 bil­lion to its dig­i­ti­za­tion strat­e­gy in the last five years. The NHS data-shar­ing plat­form, The Spine, is used dai­ly by half a mil­lion health­care pro­fes­sion­als, sup­port­ing up to 47 mil­lion trans­ac­tions. But this invest­ment is dwarfed in the US where, last year alone, pri­vate investors poured more than $8 bil­lion dol­lars into dig­i­tal health­care start-ups.

Investment in hospitals has driven a rise in the cost of healthcare in OECD countries. In the US, per capita health spending has increased sixfold over the last four decades.


The hos­pi­tals of the future, pre­dicts Dun­scombe, will be sup­ple­ment­ed by high-tech clin­i­cal back offices that will receive and aggre­gate data from mul­ti­ple sources, enable per­son­al­ized med­i­cine, and deploy human resources as required – whether a nurse, a social work­er or health coach. It has been called an Opo­do mod­el of health­care.

Cen­tral to this vision of inte­grat­ed care are “smart rout­ing tools” and “clin­i­cal field force man­age­ment”. Health­care is ulti­mate­ly about logis­tics – get­ting the right ser­vice to the per­son at the right Invest­ment in hos­pi­tals has dri­ven a rise in the cost of health­care in OECD coun­tries. In the US, per capi­ta health spend­ing has increased six­fold over the last four decades. time, enabled by the right data. The grow­ing num­ber of dig­i­tal­ly enabled patients will rarely require face-to-face con­tact. Those who do need it will get it. How long before this vision arrives? Thir­ty years to ful­ly real­ize, Dun­scombe antic­i­pates.

What about data con­fi­den­tial­i­ty? Con­cern about it has been a major drag on health-tech adop­tion. The need, says Dun­scombe, is for secure, audit­ed, encrypt­ed data that can only be used for the pur­pose of direct care, unless oth­er use is con­sent­ed. The answer? Dis­trib­uted ledger sys­tems, like blockchain. In Esto­nia, 1.3 mil­lion cit­i­zens have the holy grail of a “sin­gle uni­fied iden­ti­fi­er” and a dis­trib­uted record. They can all access their med­ical records online. They also use the sys­tem for vot­ing and shop­ping.


The Covid-19 pan­dem­ic will have a last­ing lega­cy for health­care. In the US, for exam­ple, health­care providers rapid­ly scaled tele­health offer­ings, see­ing 50 to 175 times the num­ber of patients com­pared to pre-Covid. Dun­scombe esti­mates that it has accel­er­at­ed tech­ni­cal evo­lu­tion in some parts of the UK health­care sys­tem by up to a decade.

In some cas­es, tech­nolo­gies pre­vi­ous­ly regard­ed as improb­a­ble ideas have moved straight to imple­men­ta­tion. Dur­ing the lock­down, for exam­ple, it is esti­mat­ed that up to 93 per­cent of GP con­sul­ta­tions were car­ried out vir­tu­al­ly. The NHS in Bolton worked at great speed with Mum­bai-based med­ical tech com­pa­ny to imple­ment an AI-based sys­tem to scan X‑rays for Covid-19 symp­toms. It func­tions thou­sands of times more quick­ly than human eyes.

Oth­er exam­ples include 3D print­ing used to mass-pro­duce screens for per­son­al pro­tec­tive equip­ment (PPE), head-mount­ed cam­eras allow­ing senior sur­geons to super­vise oper­a­tions, and US mil­i­tary robots, used at three NHS sites, to min­i­mize phys­i­cal con­tact with high­ly infec­tious patients.

A robot is used to distribute hand sanitizer at a shopping mall in Bangkok. Mladen Antonov/AFP Via Getty Images

There are an esti­mat­ed 2 mil­lion new med­ical stud­ies a year and med­ical knowl­edge dou­bles every two months. Ideas and inno­va­tions pro­lif­er­ate more quick­ly than any human could pos­si­bly keep up with. But, although human and tech­ni­cal fac­tors may be hold­ing us back, the pan­dem­ic has sig­nif­i­cant­ly accel­er­at­ed both tech­nol­o­gy and its adop­tion. The New York-based Macy Foun­da­tion is set to bring togeth­er aca­d­e­mics and clin­i­cians to learn lessons for med­ical train­ing, as it did after World War Two. And the US Defense Advanced Research Projects Agency (DARPA), begun in the Cold War to accel­er­ate US space and defense tech­nol­o­gy and with a $3.4 bil­lion annu­al bud­get, is research­ing tech­ni­cal and clin­i­cal solu­tions.

“Look at the speed with which we built the Nightin­gale Hos­pi­tal in Lon­don for Covid-19 patients. It was set up in nine days. The pair­ing of mil­i­tary-style gov­er­nance and deci­sion-mak­ing with the best of the NHS and tech­nol­o­gy was incred­i­ble,” says Dun­scombe. “Add to that the abil­i­ty to com­mu­ni­cate dig­i­tal­ly with every cit­i­zen, and in the future we’ll be ready to deliv­er more flex­i­ble care when the next cri­sis hits.”

1 To speed up the adoption of innovations, healthcare systems must have a vision and strategy, an agency to promote diffusion, dedicated funding and effective communication. 2 Legacy IT systems that lock in data and cannot talk to each other are one of the biggest obstacles that must be overcome. 3 Covid-19 has shown that technologies that previously had the status of improbable ideas can be moved straight to implementation if the will is there.
Read the next topicPreparing for the future of healthcare