The Translation Problem

A clinical pathway designed in London works beautifully. Patient flow improves. Wait times drop. Staff satisfaction increases. The designers publish their results. A hospital in Mumbai reads the paper and implements the pathway exactly. Wait times do not improve. Staff struggle to follow it. Patients experience worse outcomes. The hospital concludes the pathway does not work for their context. The designers conclude their work did not translate. Both are wrong. What actually happened is that the pathway was never translated at all—it was simply copied.

This is the translation problem. It affects healthcare innovation more severely than any other domain, and it is almost never named explicitly.

Innovation moves through the world at incredible speed now. A research paper published in Boston is read in Bangalore the same week. A clinical protocol designed in Amsterdam is implemented in Auckland within months. Digital health platforms are built in one market and deployed globally. The speed of copying is unprecedented. The speed of translation is near zero.

Translation is different from copying. Copying means reproducing the same design in a new location. Translation means understanding what made the design work, identifying what is different in the new context, and rebuilding the design for the new context while preserving its essence. Translation is slower. It requires deeper understanding. It is often seen as expensive. It is actually the only approach that works.

The reason requires understanding how contexts differ. Most people think context is geography. It is not. Geography matters, but it is downstream from four deeper dimensions.

The first dimension is system architecture. A clinical pathway designed for the UK National Health Service operates within a system where referral is managed centrally, waiting lists are invisible to patients, and consultant time is allocated by administrators. The same pathway in India operates within a system where patients self-refer, choose their consultants directly, and time is allocated by market preference. The workflow looks the same. The incentives are completely different. What succeeds under central allocation may fail under market selection because clinicians and patients behave differently.

Economist Friedrich Hayek called this the “problem of knowledge.” Centrally planned systems and market systems generate and distribute knowledge differently. Information that is transparent in one system is invisible in the other. What seems like poor adoption in a new context is often a signal that the system architecture has changed the nature of the problem.

The second dimension is cultural logic. This is not stereotyping—it is recognizing that different cultures have different epistemologies, different ways of knowing what is true. In some contexts, authority is trusted. In others, evidence is trusted. In still others, consensus is trusted. A clinical pathway that works by concentrating decision-making in the hands of expert consultants will fail in a culture that expects decisions to be made by consensus, even if the evidence is identical.

This is not a cultural defect in one group or another. It is a difference in how knowledge is legitimated. Ignoring it does not make it disappear. It ensures failure through invisible mechanisms—staff who do not follow protocols they do not perceive as legitimate, patients who do not trust diagnoses that were not made through culturally legitimate processes, systems that look like they are implemented but are actually being subverted.

The third dimension is economic incentive. A pathway designed where clinicians are salaried operates under different incentives than a pathway designed where clinicians are paid per procedure. In one context, the pathway that saves time is preferred. In the other, the pathway that increases volume is preferred. The same work, different incentives, different outcomes. A consultancy that ignores this creates a pathway that is technically sound but economically irrational for the people who must follow it.

The fourth dimension is infrastructure capacity. This is the most visible but often least understood. A pathway designed for a hospital with 24-hour laboratory capacity operates differently from a pathway designed for a hospital with laboratory service only during business hours. A pathway designed for a system with universal data connectivity operates differently from a pathway designed for a system with unreliable internet. These are not implementation challenges to overcome with better training. They are structural constraints that make certain workflows simply impossible.

These four dimensions are not independent. They interact. A system with centralized allocation and culturally-trusted authority and salaried clinicians and excellent infrastructure creates conditions where certain innovations thrive. Move to a system with market allocation and culturally-preferred consensus and pay-per-procedure incentives and fragmented infrastructure, and the same innovation dies even if it is perfectly implemented.

This is why the same clinical pathway succeeds in London and fails in Mumbai. Not because one city is more capable. Because the contexts are different across all four dimensions, and those differences are structural, not fixable with training or better change management.

Most innovation transfer assumes the opposite. Most assume that context is local color—cosmetic differences that do not affect the core logic. The design is treated as universal truth that simply needs implementation. When implementation fails, the blame falls on local capacity or commitment rather than on the actual structure of context difference.

Tarun Khanna and Krishna Palepu, studying emerging market innovation, found that successful transfer happens when companies understand what they call “institutional voids.” These are gaps where formal institutions are weak or absent. A company that tries to operate in an institutional void using business models designed for contexts with strong institutions fails predictably. A company that recognizes the void and redesigns its model for it succeeds. The insight is that context is not a constraint to overcome. It is the actual structure within which innovation operates.

Vijay Govindarajan has documented the opposite problem: innovations from emerging markets that fail in developed markets. Innovations designed for low-cost simplicity and high-volume delivery often fail when translated to wealthy markets with different quality expectations, different regulatory structures, and different economic models. Reverse innovation requires the same translation discipline. What works because it is simple may fail because the wealthy market expects elegance. What succeeds through high-volume economics may fail through low-volume economics.

Understanding this leads to three principles for actual innovation translation.

First, context is not geography. A hospital in suburban London and a hospital in central London operate in different contexts. A rural hospital and an urban hospital operate in different contexts. Mapping context requires analyzing the four dimensions explicitly for each specific location, not generalizing from national stereotypes.

Second, innovation degrades in translation. When you move an innovation across contexts, something is lost. The question is what is lost and whether it matters. A clinical pathway that worked because it centralized decision-making loses that advantage in a system that makes decisions by consensus. The pathway itself does not fail. The reason it succeeded is gone. Effective translation means identifying what made it work and finding what will make the translated version work in the new context.

Third, the most valuable learning flows backward. When a hospital in Mumbai redesigns a London pathway for its own context, it often discovers insights that improve the original design. What seems like constraints in Mumbai—resource limitations, different patient expectations, different infrastructure—often reveal options in London that were never considered because they seemed unnecessary. The constraint forced creativity. The creativity improved the innovation. This is why genuine translation is bidirectional. The design improves both directions.

Healthcare organizations that manage cross-context innovation well do not treat it as a rollout problem. They treat it as a learning problem. They spend time understanding the four dimensions in each context. They involve local leadership in redesign, not just implementation. They measure success differently in each context because success means different things. They expect the design to change. They build feedback loops to learn from the change.

Organizations that treat innovation as universal and context as local detail do what they have always done: copy brilliantly from one place and fail mysteriously in another. Then they blame the new context for being backward. The context is not backward. The translation is incomplete.

In healthcare, where the stakes involve human welfare and where context shapes outcomes as much as clinical science does, the translation problem is not optional. It is foundational. Solving it requires patience, local expertise, structural thinking, and humility about what will change. It requires treating innovation not as something to be deployed but as something to be grown in new soil.