In the 17th century, a Dutch draper named Anton van Leeuwenhoek used a small handmade microscope to peer into a world previously unseen by the human eye. Thus he discovered microorganisms and gave rise to the field of microbiology. It offered solutions to challenges in healthcare that until then had seemed intractable.

Today, we face a new set of complex problems in healthcare that seem more intractable than others before for their inherent complexity and the constraints they threaten to impose on resources.

P versus NP

It so happens that an unsolved problem in computer science, simply called the P versus NP problem, could hold the key to these modern-day conundra. While it may sound like a cryptic puzzle reserved for computer science mavens, its implications stretch beyond algorithms and data structures, rippling through diverse fields including healthcare. But what exactly is this puzzle, and how could its resolution unlock a new era in medical science?

Let’s start with a simple arithmetic example. Say you’re asked to multiply 17 with 19. With some time, you’d arrive at the answer: 323. This is a ‘P’ problem: you can solve it reasonably quickly. (‘P’ stands for polynomial time.) Suppose you’re presented with 323 and asked to identify the two prime numbers multiplied to get this. In this case, you will have to take the trial and error route until you arrive at 17 and 19. This is an ‘NP’ problem: it takes longer to solve, but once you have the solution, you can verify it quickly. (‘NP’ here is nondeterministic polynomial time.)

Healthcare is filled with complex problems. Consider scheduling in a hospital: assigning doctors and nurses to shifts, booking operating theatres for surgeries, and organising patient appointments. It is an intricate puzzle that requires considering various factors — staff availability, urgency of medical cases, etc. — and potential changes such as emergency cases and cancellations.

The P vs NP question is this: could there be a shortcut to solve ‘NP’ problems as quickly as ‘P’ problems? Because the implication is that if P equals NP, we could quickly find the optimal solution to these scheduling problems, thus significantly improving patient care.

The implications of resolving this question are profound and wide-reaching, including for healthcare.

Implications for healthcare

The P vs NP question is a problem in mathematics and computer science, but that does not mean it will be confined there. If an existing problem can be given a faithful mathematical representation and is found to be an ‘NP’ problem, the shortcut in question could help by turning it into a ‘P’ problem.

For example, antibiotic resistance is a significant global health concern. If P equals NP, we may have a way to quickly analyse bacterial genomes and predict their resistance patterns, helping doctors prescribe the most effective antibiotics. This would improve patient outcomes and help combat antibiotic resistance, including new antibiotics discoveries for emerging diseases. Of course, patients’ adherence will still matter.

Cancer is a complex disease with myriad mutations. Deciding the best treatment plan is an NP problem because it involves considering all possible combinations of drugs and therapies. If P equals NP, we may have an opportunity to swiftly identify the optimal treatment for each individual cancer patient and potentially save many lives. The catch here is that we will still need a large volume of data.

Insurance companies grapple with NP problems when they have to determine premiums and packages based on considering numerous variables like age, health status, lifestyle, and medical history. Having a shortcut to crack the P vs NP problem could help these companies optimise their decision-making and pave the way to fairer and more accurate premiums and conditions. Further, government spending on healthcare can also be utilised with minimal leakage while programmes like Ayushman Bharath can contribute more effectively to achieving universal health coverage.

By solving these complex problems more efficiently, we could potentially dramatically reduce resource constraints and improve health outcomes.

Surprising sources of progress

While the P vs NP problem is a topic of ongoing study in computer science, the consensus among most experts is that P probably does not equal NP, implying that some problems will remain very difficult to crack, even if a solution — once it is found — will be easier to verify. But this has not deterred researchers from exploring this question, and in the pursuit of which they have unearthed improvements to algorithms and new approaches to dealing with complex problems.

Throughout history, there have been many instances of seemingly insurmountable problems being overcome with innovative thinking. Before the discovery of electricity, for example, candlemakers lit our world. Yet most of them may never have foreseen the revolutionary consequences of Thomas Edison’s incandescent bulb, which brought light to more people and for longer hours.

Similarly, following the invention of calculus and expanding the binomial theorem to negative integers and fractions, Isaac Newton considerably improved our understanding of the irrational number pi. Why, the technology giant Apple has been transforming our expectations of what a watch can be expected to do in ways that Swiss watchmakers may never have anticipated.

Not all will be winners

This said, one potential drawback of P being equal to NP, if ever that outcome comes to pass, lies in the realm of cryptography. Many encryption schemes and algorithms rely on problems that are currently hard to solve, believed to be in the set of ‘NP’, not ‘P’ problems. That is, these schemes protect secrets by hiding them behind a problem that is very hard to solve but easy to verify. If P equals NP, these problems will become easy to solve, rendering these encryption schemes vulnerable to attacks and compromising digital security.

This said, healthcare isn’t the sole beneficiary of this problem-solving. The barrier that the P vs NP problem stands for encompasses every field where the solution to a problem is blocked by the availability of significant computational resources. So these fields include logistics, finance, and even climate modelling, all of which could experience paradigm shifts if the P vs NP problem is solved in favour of the P = NP outcome.

The Clay Mathematics Institute in Colorado continues to offer a million dollars to anyone who can definitively solve the P vs NP problem. But for anyone who does, a million dollars will pale in comparison to the rewards they stand to collect by revolutionising various human enterprises, potentially driving human progress in unimaginable ways.

As we look to the future, let us remember that problems that seem insurmountable today might not be so tomorrow. As with the candlemakers, the watchmakers, and even Anton van Leeuwenhoek, the solution often comes from where we least expect it. Today’s brightest minds grappling with the P vs NP problem may be on the brink of a breakthrough that could redefine healthcare as we know it.

Dr C. Aravinda is a public health physician and student at IIT Madras pursuing a BS degree in data science.

Indonesian researchers have tested a cheaper way of monitoring antimicrobial resistance that could be a key tool for developing countries. Image for Representation.
| Photo Credit: Getty Images/iStockphoto

Indonesian researchers have tested a cheaper way of monitoring antimicrobial resistance that could be a key tool for developing countries.

Indonesian researchers have tested a cost-effective, relatively quick method of measuring changes in antimicrobial resistance that could help developing countries fight a problem seen as a global threat.

Instead of sticking to an approach that emphasises the need for intensive laboratory testing that might not be practical in many countries, they assessed lot quality assurance sampling — a system where a population’s antimicrobial resistance can be assessed using smaller sample sizes.

That’s good news for the G20 after health ministers committed to tackling the threat comprehensively at their meeting last month. Antimicrobial resistance is the ability of microbes to block the effect of drugs that are meant to kill them. This makes infections harder to treat and can lead to longer hospital stays, more expensive care and increased risk of death.

The high level of antimicrobial resistance in Indonesia is now becoming a silent pandemic.

Explained | Is air pollution driving the rise of antibiotic resistance? 

The United Nations has warned the rise of ‘superbugs’ could kill 10 million people a year globally by 2050 and be a drain on the world economy. Like in many other countries, the improper use of antimicrobial drugs and other factors that trigger resistance, such as poor sanitation and air pollution, are prevalent in Indonesia. In 2015, the 68th World Health Assembly adopted the Global Action Plan on antimicrobial resistance.

It emphasises the importance of enhanced global surveillance, particularly in low- and middle-income countries where it is a major concern. One of the key pillars of the Global Action Plan is the support for national strategies through improved global surveillance.

The proposed global surveillance system aims to estimate the prevalence of resistance by using laboratory testing of clinical samples. However, this approach is not always practical in developing countries because of their limited access to quality microbiology diagnostics.

Population-based surveillance is a preferred strategy, but it is also time-consuming, labour-intensive and costly. That means many regions need a rapid, feasible and affordable surveillance strategy.

Explained | The challenge of antimicrobial resistance

Indonesian researchers found an alternative approach: they tried lot quality assurance sampling. This method, which was originally developed in the manufacturing industry to assess batch quality, involves classifying a population as having a high or low prevalence of antimicrobial resistance based on a small sample size.

It has proved to be more practical and cost-effective than conventional population-based surveillance. The Indonesian research applied the lot-based approach to assess the prevalence of antimicrobial resistance in patients with suspected urinary tract infections.

The researchers wanted to estimate the test characteristics for identifying populations with a high prevalence of resistance in urinary tract pathogens, provide lot-sampling classifications for 15 antibiotics in 11 different settings and estimate the cost of implementing lot sampling in a single health facility.

The testing was done in the Indonesian cities of Medan and Bandung, and the exercise was repeated 1,000 times for each of the 13 lots.

Also Read | Putting minds and funds together to find ways of tackling drug resistance 

They found lot testing was 98 percent effective in correctly identifying populations with a high prevalence of antimicrobial resistance.

Overall, the researchers were able to show that lot quality assurance sampling is a promising approach to efficiently estimate antimicrobial resistance prevalence and guide treatment decisions, especially in resource-limited settings.

By significantly reducing sample size requirements and increasing efficiency, lot-based surveillance could significantly contribute to public health efforts to combat antimicrobial resistance worldwide. The findings are particularly significant in the context of the growing importance of surveillance as a crucial tool for antimicrobial stewardship.

Previous studies on drug resistance have shown lot-based surveys are effective in identifying local variations in drug-resistant tuberculosis and assessing the prevalence of transmitted drug resistance in HIV.

This style of survey could be implemented at sentinel sites, enabling regular assessments of changing trends, intervention impacts, or early detection of resistance development after introducing new drugs. This utility is particularly beneficial in settings with limited microbiology capacity or where empirical treatment is common, such as primary care settings globally.

The cost of the lot-based surveys, including 15 antibiotics, ranged between US USD 403 and USD 514 in the 11 sites studied — relatively cheaper than conventional testing regimes. Despite some limitations, including the need for careful site selection and ensuring proper laboratory accreditation and quality control, lot quality assurance sampling shows promise in providing valuable information on antimicrobial resistance for both clinicians and policymakers.