According to the CDC, an estimated 20.9% of the US adult population experienced chronic pain in 2021, while 6.9% were affected by high-impact chronic pain which limited their daily activities.[i] The need for novel analgesic drugs with superior efficacy, reduced adverse effects, and lower abuse potential is strong, and GlobalData’s Trials Intelligence platform reveals that more than 2,500 clinical trials were initiated for pain within the first nine months of 2025, while a total of just over 2,900 commenced during 2024.[ii]
However, drugs targeting pain have some of the lowest success rates. Indeed, a 2023 analysis from industry group Biotechnology Innovation Organization discovered that just 0.7% of new pain drugs progress from Phase I clinical trials to FDA approval, in comparison to 6.5% across all therapeutic areas.[iii] According to GlobalData’s analysts, the lack of objective pain measurements, safety issues, and varying tolerances are some of the main factors hindering pain clinical trials and subsequently the development of novel analgesic drugs. Furthermore, GlobalData’s Drugs Intelligence platform reports that of the pain drugs that successfully made it to market, 16% were withdrawn.[iv]
Placebo response and retrospective data in pain trials
Another key issue is placebo response, which might account for as high as 75% of the overall analgesic effect in pain trials[v], masking true treatment effects and leading to failed studies.
According to Marco Calabresi, M.D, PhD, Senior Medical Director at Fortrea, the industry has implemented several strategies to combat these effects, but methods have often fallen short. “Protocol design elements such as placebo run-in periods and efforts to recruit ‘non-placebo responders’ are costly, wasteful, and have not been shown to be effective in general,” Dr. Calabresi explains.
When it comes to outcome measures, standardized pain scales such as the Visual Analog Scale and Numeric Rating Scale have been introduced, but the challenge is that pain is highly subjective and strongly influenced by psychological factors. Patient-specific life events, site effects, and expectations all play a part in influencing the patient’s perception of their symptoms on a given day, and these are incredibly difficult to predict or control. The subjectivity of pain itself is “intractable”, says Dr. Calabresi. “What we can (and must) do is improve the accuracy and reliability of the measurements,” he adds.
Meanwhile, retrospective self-report data lacks routine error control, and traditional monitoring cannot tackle deeper measurement errors that contribute to placebo response. In general, collecting data in real-time via an approach known as ecological momentary assessment (EMA) is best practice in pain trials. This is because recall introduces bias during site visit assessments, with the patients’ current mood or pain levels influencing their memories of previous experiences.
Conducting EMA through electronic devices supports real time centralised statistical monitoring and the detection of quality issues that could play a part in inflated placebo responses. Electronic formats also open the doors to integration with wearables and real-time patient monitoring tools.
How digital health innovation can support better pain data collection and monitoring
Potential applications for AI in the neuroscience field are growing and strengthening, as explored in the latest whitepaper from leading clinical research organisation, Fortrea. In pain-related clinical trials, integrating ecological momentary assessment with AI offers a transformative approach to data collection and analysis.
EMA’s timeliness – beyond mere speed – ensures that patient-reported outcomes are captured in context, while its high-frequency transmission yields a rich, continuously updated stream of information. However, this volume quickly overwhelms traditional human review processes, making manual interpretation a bottleneck. AI can alleviate this constraint by continuously analysing incoming data, identifying patterns, and flagging anomalies in real time, thereby streamlining human oversight and enhancing the precision and responsiveness of trial monitoring.
In 2024, 8.2% of new pain trials in GlobalData’s Clinical Trials database used an element of mobile/digital technology such as smartphones or remote patient monitoring, representing an increase from the 4.5% of pain trials that used these technologies in 2018.[vi] As the use of digital technologies continues to grow in the field, the option to combine them with AI presents a golden opportunity.
In addition, AI could enable better and faster trend analysis and data monitoring in pain trials, helping to appropriately encourage patient compliance and protocol adherence without compromising the blind. Under this approach, algorithms provide blinded analytical output to facilitate human decision-making and judgement.
Can AI help tackle placebo response?
AI could also play an important role in understanding the impact of placebo effects in pain trials. There are several ways in which a patient’s propensity to experience a placebo response can be predicted using both psychometric data – personality traits that correlate with placebo response – and biomarkers such as brain activity and genetics. This has paved the way for AI models to harness such data during screening and enrollment of pain trials, however, Dr. Calabresi warns that there’s a right and a wrong way to use such models in pain research, and that the end goal should always be to enhance the reliability and interpretability of clinical trial outcomes. As one example, AI models can be used within “propensity weighting”, helping to randomise patients in a way that ensures a balance of placebo responders across each arm.[vii]
“Rather than selecting for a specific type of reporter in a way that could bias results, a better approach focuses on identifying patterns of consistent and meaningful reporting,” Dr. Calabresi explains. “This enables real-time enrichment of study populations by leveraging historical and behavioral data to account for placebo susceptibility, not to exclude or skew, but to ensure that variability is understood and transparently incorporated into analysis. The ultimate aim is to improve signal detection while preserving the integrity and representativeness of the data.”
AI will play an important role in the future of neuroscience research but implementing it responsibly and ethically is central. Discover more possibilities and considerations by downloading Fortrea’s whitepaper below.
[i] https://www.cdc.gov/mmwr/volumes/72/wr/mm7215a1.htm [ii] GlobalData, Pharmaceutical Intelligence Center, Clinical Trials database. Accessed 30 September 2025. [iii] https://go.bio.org/rs/490-EHZ-999/images/BIO_The_State_of_Innovation_in_Pain_and_Addiction_2017_2022.pdf [iv] https://pharma.globaldata.com/Analysis/TableOfContents/2023-Starts-Off-with-a-Surge-in-Clinical-Trials-Tackling-Pain?Viewpoint=1 [v] Zou, K., Wong, J., Abdullah, N., Smith, T., Doherty, M., Zhang, W. Examination of overall treatment effect and the proportion attributable to contextual effect in osteoarthritis: meta-analysis of randomised controlled trials. Annals of the Rheumatic Diseases, Volume 75, Issue 11, 1964 – 1970 (2016). [vi] GlobalData, Pharmaceutical Intelligence Center, Clinical Trials database. Accessed 30 September 2025. [vii] Gomeni, R., Bressolle-Gomeni, F. & Fava, M. Artificial intelligence approach for the analysis of placebo-controlled clinical trials in major depressive disorders accounting for individual propensity to respond to placebo. Transl Psychiatry 13, 141 (2023). https://doi.org/10.1038/s41398-023-02443-0
