Big Data and AI: How They Changed the Rules of the Game and Created the Pharma Industry's New Gold Mine

Toma Tomov is an expert with over 20 years of professional experience in developing and managing innovation policies and projects in the fields of healthcare, the public sector, and regional development. His career spans leading expert and management positions in the National Assembly, the Council of Ministers, and the Ministry of Health, where he was responsible for changes regarding the legislative framework and operational programs in the field of healthcare, including the implementation of E-health projects at the Ministry of Health. Over the last decade, he has led and consulted on projects funded by European programs such as HRDOP, RDOP, ACAP, and INTERREG, focusing on e-health, the digital transformation of public services, and strategic planning. He has an active international profile with participation in projects in North Macedonia, Greece, and Southeast Europe. He consults on strategic transformations, the introduction of innovations, and e-governance in the public sector. In Healthcare Magazine, Toma Tomov shares an expert perspective on topics such as health innovation, e-health, and digital solutions for more sustainable health systems.

Big Data analytics and Artificial Intelligence (AI) are fundamentally changing the way medical therapy outcomes and clinical trial processes are monitored and evaluated—from participant selection and protocol design to monitoring, results analysis, and success prediction. Big Data offers a deeper, more dynamic, and more precise approach—with continuous collection, analysis, and adaptation of therapy based on the patient’s real life. By gathering data from electronic health records (EHR), laboratory systems, wearable devices, mobile applications, and patient-reported outcomes, rich, multimodal datasets are created. This data is not just statistics—AI, machine learning, deep neural networks, and predictive algorithms can discover patterns that are invisible to the human eye. For example, such models can predict which patients are at risk of discontinuing therapy, forecast adverse reactions, or optimize dosage based on individual response.

A report by Fortune Business Insights estimates that the global market for AI solutions in clinical trials was valued at approximately USD 2.76 billion in 2024 and is expected to grow to approximately USD 54.81 billion by 2032, with a compound annual growth rate (CAGR) of about 46.4%. (Fortune Business Insights)

WHAT ARE THE KEY ROLES THAT AI AND BIG DATA PLAY IN THIS TRANSFORMATIONAL PROCESS?

Participant Selection and Recruitment

One of the biggest obstacles to clinical trials has always been appropriate participant recruitment—this process is expensive, time-consuming, and often fails. Now, AI models working on large databases of electronic health records, demographic profiles, and genomic data can predict which patients will meet the criteria, when they might drop out, and the risk of non-participation. This accelerates recruitment, reduces costs, and improves efficiency.

Optimization of Trial Design and Predicting Outcomes

AI is used to simulate various protocols—changing inclusion/exclusion criteria, dosages, and endpoints—even before the trial begins. This leads to smarter, more flexible designs that can reduce the number of late-phase failures. These technologies also predict outcomes: which participants are unlikely to respond to therapy, who might experience adverse reactions, and what the risks and timelines for results are.

Real-Time Monitoring and Post-Marketing Surveillance

By collecting data from wearable devices, mobile applications, digital health platforms, and clinical systems, AI models can monitor the health status of participants in “virtual” or hybrid trials. This enables early detection of complications, reduced response times, and better safety and efficacy of the research.

Economic Efficiency and Accelerated Drug Development

One of the primary reasons for the growing adoption of AI in clinical trials is economic—pharmaceutical companies and research organizations are looking for ways to shorten drug development time, reduce costs, and increase the probability of success. Big Data enables the creation of cost-benefit prediction models to justify investment.

Despite the vast opportunities, integrating AI and Big Data into clinical trials also brings several risks. These include the need for high-quality and representative data, protection of personal health data, transparency and explainability of models, regulatory certainty, and the question of whether AI-based approaches will truly lead to higher trial success rates or merely process optimization. Simultaneously, over-reliance on AI could lead to an undervaluation of the human factor and clinical expertise.

AI and Big Data are no longer “extras” or experiments in clinical trials—they are a key component of modern drug development infrastructure. The Fortune Business Insights report shows a clear trajectory of massive growth, turning this segment into a strategic investment for the pharma industry. For companies that manage to integrate AI technologies in time, this means faster scientific engineering, lower risk, and higher profitability. For health systems and society, the potential lies in more rapidly accessible therapies and better control of chronic diseases. However, success will depend on how data management, support, ethics, and regulations are handled.

From Data to Decisions: Bulgaria’s Big Data Potential in Healthcare

Although Bulgaria is still in the early phase of integrating Big Data and AI in healthcare, the country is developing a stable and dynamic ecosystem of companies, hospitals, and technological platforms that are implementing world-class digital solutions. The accumulated data from electronic healthcare, the presence of a strong IT sector, and the gradual introduction of international standards position Bulgaria as a natural candidate for a regional innovation hub in health technologies.

At the center of this ecosystem stands the National Health Information System (NHIS)—an integrated platform that unites electronic prescriptions, examinations, laboratory results, and health records. Despite the significant volume of hundreds of millions of documents, the system remains far from its potential to function as an effective Big Data infrastructure. Key challenges include limited interoperability between hospital information systems, leading to duplicate records and inconsistencies; the absence of mandatory implementation of international standards such as SNOMED CT, LOINC, and HL7 FHIR, which leaves much of the data unstructured; as well as unstable architecture, frequent technical interruptions, and variable quality of entered data. To become a strategic asset, the NHIS requires modernized architecture, consistent standardization, reliable quality control, and secure access to anonymized data for scientific and analytical purposes—conditions without which national registries, real-world data analysis, and AI models remain limited. Furthermore, a real opening of the system to competition and the integration of existing solutions and applications is needed, as the private sector demonstrates significant capacity for innovation alongside state infrastructure. The expertise of teams and the competitiveness of companies under real market mechanisms accelerate the implementation of new technologies, improve the quality of solutions, and stimulate the development of innovative products.

The Bulgarian private sector already offers solutions fully comparable to international practices—digital health applications providing personal health diaries, chronic disease tracking, integration with medical centers and online consultations, telemedicine stations, cloud platforms for electronic records, and remote monitoring solutions. Their approach, combining devices, software, and centralized data collection, is the first step toward implementing “virtual hospitals” and continuous patient tracking—a paradigm widely adopted in modern health systems.

At the same time, more and more healthcare organizations—hospitals, clinics, laboratories, and digital platforms—are introducing electronic documentation, predictive triage algorithms, and telemedicine services for chronic patients. Although initiatives are fragmented, they demonstrate the readiness of the medical environment to transition to the active use of data.

Obstacles such as missing interoperability, inconsistent data quality, and a limited regulatory framework for AI do not change the fact that Bulgaria possesses the key elements to build a modern, data-oriented healthcare model. With strategic integration, standardization, and coordination between the state, healthcare organizations, industry, and technology companies, the country can develop a sustainable national Big Data platform that will provide a stable analytical foundation for improving clinical decisions and the efficiency of the healthcare system.

BEST PRACTICES

International experience clearly shows what becomes possible when medical data is used as a strategic resource. A specific example is the Johns Hopkins platform for heart transplant patients. It uses a machine learning algorithm that analyzes laboratory results, physiological indicators, and medication history to predict the likelihood of organ rejection weeks in advance. As a result, clinicians manage to prevent complications instead of reacting after the fact. This reduces hospitalizations and the need for aggressive interventions. A similar approach is used by the Mayo Clinic in monitoring patients with heart failure. A machine learning algorithm analyzes more than 20 continuous clinical parameters and can predict with 78–83% accuracy which patients will be admitted for emergencies within the next 30 days. This enables early interventions—therapy adjustments, additional medical contact, or home monitoring. Another impressive example comes from oncology. At the MD Anderson Cancer Center, an AI system has been implemented that analyzes a combination of imaging studies (CT, MRI), genetic data, and chemotherapy response. The system predicts the effect of treatment 30–50% more accurately even before the second cycle, allowing the patient to be transferred to an alternative and more suitable therapy earlier—without losing time and health on ineffective treatment.

These examples show that the role of AI and Big Data is no longer an addition to medicine but a central element of it—an early warning system, a personalization mechanism, a tool for proving efficacy, and ultimately, a way to treat faster, better, and more economically. This is a change that was almost unthinkable just 10 years ago, but today it shapes a new standard of medical care and accountability.

Data from wearable devices is also beginning to play a key role. Stanford Medicine conducted a study with over 2,000 patients using smartwatches that constantly measure heart rate, activity, sleep, and heart rate variability. Analysis shows that these parameters can predict the worsening of chronic diseases (e.g., pulmonary and cardiac) up to 2 weeks in advance, allowing doctors to intervene in time.

Regarding diabetes, an example is the Livongo program (now part of Teladoc). The platform collects glucose measurements, nutrition, and behavioral data and sends automatically generated recommendations. The clinical results are impressive: users experience 19–30% fewer hypo- and hyperglycemic events, and hospitalization costs are significantly reduced.

THE GLP-1 MODEL AND HOW THE ECONOMY OF METABOLIC DISEASES WAS BORN WITH THE HELP OF NEW TECHNOLOGIES?

GLP-1 therapies—the most current medications for diabetes and obesity—also benefit from this model. Companies offering digital tracking in combination with treatment report that patients using a monitoring app plus AI-personalized recommendations achieve between 15% and 20% greater weight loss over 12 months compared to patients receiving only medication and standard follow-up exams. This proves that data enhances not only the treatment but also influences patient behavior.

The greatest success of incretin therapies is not only in their medical efficacy but in the fact that they created an entirely new category of treatment—one that requires months and years of therapy, monitoring, digital tools, and analysis of patient behavior and response. This model is much closer to the software business than to traditional drug sales. For companies, it means not a one-time sale, but a sustainable, recurring revenue stream that grows with every month the patient continues treatment.

According to analysis by Fortune Business Insights, the global market for GLP-1 agonists reached a value of USD 52.08 billion in 2024 and is expected to grow to approximately USD 186.64 billion by 2032. This represents an impressive compound annual growth rate of about 16.8%—a figure rarely seen in mature therapeutic categories.

But behind these numbers lies another revolution—that of data. GLP-1 drugs collect vast volumes of real-world information: from electronic records and hospital systems to laboratory results, wearable devices, behavioral data from mobile apps, and telehealth platforms. This data allows for something that pharma could not previously do on such a scale: showing in real-time what is happening to patients and proving the efficacy of treatment not just in clinical trials, but in actual medical practice. It is this real-world efficacy that is becoming the strongest weapon in negotiations with insurers and government payers.

Few pharmaceutical products in recent decades have managed to simultaneously change medical practice, corporate strategies, and investment logic as GLP-1 medications have. What began as a therapy for type 2 diabetes has today become a global business phenomenon capable of transforming entire sectors—from healthcare and insurance to corporate employee care and public finance.

GLP-1 therapies are already among the highest-margin pharma products in the world. They are sold in massive volumes at record unit prices and generate large-scale sustainable revenue. Market forecasts indicate that the segment could exceed USD 100 billion before the end of the decade, placing it alongside the most profitable pharma categories in history. GLP-1 medications are no longer just a medical breakthrough—they have become a financial asset with a central place in the business strategy of leading companies. Net income from these products is used not only for research and development but also for share buybacks, increasing dividends, and boosting market capitalization. In this way, GLP-1 determines not only revenue but also the stock market dynamics and investment attractiveness of corporations.

Market expansion is also fueled by the broadening of medical indications. In addition to diabetes and obesity, GLP-1 medications are being actively studied for non-alcoholic fatty liver disease, chronic kidney disease, and cardiovascular risks. This increases the number of potential patients, extends the product lifecycle, and provides grounds for maintaining premium pricing positions in the long term.

At the same time, innovation risk remains real. The entry of new competitors, the development of second- and third-generation molecules, and the future emergence of generic versions could put pressure on prices. In response, companies are building complex patent structures and “patent thickets” covering molecules, formulations, devices, and methods of administration—a strategy aimed at extending market protection and maximizing returns.

At the heart of this new business model, however, lies another factor—data. Big Data and analytics are becoming a strategic tool through which therapy efficacy is proven, pricing is optimized, and risk is managed. Real-world clinical data from electronic records, insurance registries, and long-term tracking allow companies to demonstrate economic value: fewer hospitalizations, lower risk of complications, and reduced costs in the long run.

Post-marketing monitoring is now supported by artificial intelligence, which analyzes vast datasets in real-time and detects early signals of adverse reactions. This reduces regulatory risks and contributes to a more sustainable image before the market and investors. At the same time, analytics allow for patient segmentation and targeting, prediction of therapeutic response, and personalization of doses, which increases the efficacy and profitability of treatment.

Big Data also enables precision medicine. By segmenting patients by clinical and genetic profiles, predicting therapeutic response, and creating personalized dosing schedules, companies increase the efficacy and return on treatment at an individual level. Monitoring behavior and adherence through apps and wearable devices allows for early identification of patients at risk of discontinuation and the integration of coaching programs and personalized interventions.

Large datasets also support the development of health-economic models proving that GLP-1 therapies can reduce long-term healthcare costs. This provides a basis for introducing innovative contracts where payment depends on actual results achieved—for example, the percentage of weight loss after a certain period.

On this basis, new strategic models are emerging. Increasingly, manufacturers are negotiating direct partnerships with large employers, providing GLP-1 therapy in a package with digital tracking and behavioral support. For companies, this means a healthier and more productive workforce, fewer absences, and lower health insurance costs. Parallel to this, “pay-for-performance” contracts are becoming established, where payment is tied to actual clinical indicators, such as the percentage of weight lost.

Despite impressive growth, the model remains vulnerable to several factors—limited access, public outcry over high prices, the risk of therapy discontinuation, and increasing regulatory pressure. All this turns GLP-1 not only into a business opportunity but also into a test of the sustainability of the modern health-economic model.

Ultimately, GLP-1 is not just a drug—it is a new industrial paradigm. It combines pharmacy, technology, finance, and behavioral economics into a single platform that generates massive revenue and redraws the way value is created in healthcare. For business, this is not just a trend but a signal: the future belongs to companies that can combine medication, data, and strategy into a comprehensive economic model.

Insurers are increasingly demanding a “pay-for-performance” model—payment only if the patient achieves measurable clinical goals. This further reinforces the importance of continuous collection and analysis of real-world medical data.

In some countries, reimbursement contracts already exist where payment for the medication depends on results—for example, the percentage of weight lost over a certain period. This is unthinkable without easy access to large datasets and algorithms that can evaluate efficacy at both individual and population levels. Thus, for the first time, the business model of a drug begins to resemble that of a digital service—one that depends on digitally measurable value, not just clinical evidence.

There is also another side to the coin. High prices raise several social and political questions: who has access to treatment, who can fund it, how long the system can pay, and what happens if patients discontinue therapy. Health systems worldwide are trying to balance the promises of reduced complications and costs against the risk of a single product consuming a disproportionately large share of the pharma budget.

The situation in Bulgaria today reflects part of this global dynamic. Semaglutide is available in the pharmacy network, but coverage by the NHIF is limited. Patients using GLP-1 medications outside of diabetes indications or under private medical schemes often pay the full price. This limits access to therapy, especially when used for treating obesity—one of the most common chronic conditions in our country and the world. At the same time, Bulgarian regulators and professional organizations warn of a growing gray market for illegitimate or unregistered medications offered online—a risk that is a direct result of the rapid increase in demand.

The country has the potential to build a strong model based on data—both for real-world efficacy assessment and for the economic justification of funding. Bulgarian e-health systems, if integrated more effectively, could serve as the basis for national registries, outcome tracking, and even AI models to determine which patients are most likely to succeed or are at risk of adverse reactions. However, this requires a strategic decision: to use collected data to create value without compromising patient rights and security.

And here lies the biggest question for the coming years: who owns this data? Pharma companies have it to prove efficacy. Insurers have it to select optimal patients and funding models. Hospitals and doctors have it to track their results. The state has it to regulate costs. And the patient has it to be sure that the data is used in their interest. The battle for ownership and control over medical information has already begun, and it involves not just money, but principles.

Incretin therapies have shown that a single drug can change not only the approach to treating chronic diseases but also the way business, value, access, and data are thought of in healthcare. Technology, artificial intelligence, and real-world medical practice are beginning to intertwine in a way that seemed like science fiction a decade ago. And health systems—including the Bulgarian one—are faced with a choice: whether to remain observers or learn to manage the new reality and benefit from it.

One thing is certain—the era of drugs as a standalone product is ending. The era of medications as data, algorithms, and platforms is beginning. And GLP-1 is the first major example of this.