Can healthcare learn from software’s shift-left approach?

In software engineering, there’s a powerful principle called shift left that has transformed how we build and deliver technology. Could this same concept revolutionize healthcare delivery? The parallels are striking, yet the challenges run deeper than you might expect.

Shift left refers to finding and fixing problems as early as possible in the development process. This approach is grounded in Boehm’s Law, which states that the cost of repairing software defects grows exponentially the later they’re discovered in the development cycle. What starts as a simple fix during coding becomes a catastrophic expense in production.

The concept traces back to William Edwards Deming’s management principles from the 1980s and 1990s, inspired by the success of Japanese automotive manufacturing. Adam Kolawa, founder of Parasoft, brought these ideas to software development in the early 2000s, emphasizing early defect detection and prevention measures like unit testing. Today, these concepts form the backbone of modern agile development, DevOps practices and continuous integration/continuous deployment (CI/CD) pipelines.

Healthcare has witnessed remarkable technological advancements. Regional hospitals now routinely use sophisticated diagnostic tools, including ultrasound, X-ray, CT, MRI and PET-CT scanners. These powerful technologies mirror the comprehensive testing tools used in software development, complete with specialized teams and departments. More intriguingly, we’re seeing the rise of point-of-care technology (POCT), which consists of simpler, localized tests that can be performed where care is delivered rather than in centralized labs. Covid-19 home tests represent the ultimate evolution: true ‘DIY’ healthcare diagnostics. This mirrors software engineering’s shift toward developer-friendly tools like code analysis and unit testing that individual programmers can use independently.

If healthcare has all the technological ingredients for a shift-left approach, moving from cure to prevention, why hasn’t it happened? Several barriers emerge, beginning with the fear of overtreatment. Early detection might reveal conditions that would never have caused problems, leading to unnecessary interventions. This parallels software engineering’s struggle with static analysis false positives – warnings that developers often ignore because they seem unlikely to cause real-world issues.

However, the fundamental challenge is what’s known as the wrong pockets problem. Unlike software development, where one organization controls the entire budget, healthcare involves multiple independent parties governed by different regulations. Investing in prevention increases costs for one provider while delivering savings to another. When prevention succeeds, hospitals lose patients and revenue, creating perverse incentives against the very outcomes we want.

A compelling experiment in the Dutch village of Afferden demonstrated both the promise and peril of healthcare’s shift left. A general practitioner implemented longer, proactive consultations with patients, focusing on prevention rather than reactive treatment. The results were encouraging, but the nearby hospital faced financial difficulties as fewer patients required expensive interventions. The risk of hospital bankruptcy ultimately ended the experiment.

The software industry underwent a massive transformation, eliminating entire testing departments in favor of integrated, continuous quality practices. Healthcare needs a similar systematic overhaul, but it requires thinking beyond organizational silos. True healthcare shift left demands a patient-level focus rather than applying principles at the macro level, with operational applications centered on individual patient journeys. It requires integrated incentives through financial structures that reward prevention across the entire care ecosystem.

Unlike software engineering, healthcare lacks well-developed evaluation and implementation frameworks for prevention-focused approaches, necessitating significant methodological development. Finally, it demands regulatory alignment by coordinating fragmented legislation across different care domains, from basic health insurance to long-term care and social support.

Healthcare’s fragmentation makes shift left both highly valuable and practically challenging to implement. Success requires moving beyond technical solutions to address systemic issues such as aligning incentives, determining which methodological principles work best in highly unpredictable and fragmented environments, and learning from software engineering’s more mature implementation approaches. The ingredients for transformation exist, but assembling them into a coherent system remains healthcare’s greatest challenge. Perhaps it’s time to stop thinking in silos and start building truly continuous healthcare delivery, one patient journey at a time.