Should individualised benefit-harm assessment be a core part of real-world research?

Benefit-risk (of harm) assessment is the cornerstone of decision making in medical care [1]. Aggregated clinical trial data plays a critical role in both bringing treatments to market (by informing licensing and reimbursement) and in informing treatment decisions made by health care professionals and patients in clinical practice. That is, being able to show the efficacy and the safety profile of a treatment allows regulators, payers/HTA agencies, doctors and patients to make a judgement about whether – on average – the benefits outweigh the risks of harm.

Why averages do not tell the whole story

Averages are important in medical research. They make it possible to statistically compare different treatments, patient groups and outcomes across different settings, and without them it would be difficult to evaluate whether one treatment works better than another. It is well known that averages can hide individual variation, be distorted by outliners and obscure subgroup effects and modern statistical techniques and data presentations can easily account for this, helping people interpret the averages in the context of the whole dataset. What these statistical techniques cannot do however is change the inevitable “one size fits all” thinking that comes from reviewing averages [2].

Patients don’t all define “worthwhile” in the same way

If a medicine demonstrates a favourable benefit-risk profile, should we therefore assume that:
(a) most patients objectively experience more benefit than harm?
(b) most patients personally perceive the treatment as worthwhile?

The answer to both is “no”.

The relative importance individuals assign to the benefits and risks/harms associated with a treatment varies greatly [1]. A treatment experience that one person perceives to be broadly positive may be perceived as negative by another person depending on their preferences, priorities and needs. For example, one person may be willing to accept some disabling side-effects of treatment if it increases the amount of time they live. Another person may be more concerned about living without side-effects even if this limits the duration of their life. Both perspectives are understandable and valid. Such preferences and priorities do not only differ between people but can also change within people over time. A person’s propensity to prioritize an extension of life over treatment side-effects is likely to differ when their children are young vs grown, and again when (if) their children bless them with grandchildren, let alone depending on how many different treatment options they have tried and their associated disease fatigue/burnout.

What individualised benefit-harm assessment can add to real-world research

I have been fortunate to be involved in a number of initiatives in this space in recent years – from defining personalised health frameworks [3] to assessing individualised benefit-harm [1,4], tolerability [5,6], and medication acceptability [7-9] through questionnaires, and understanding people’s treatment experiences through qualitative research [10]. I am a huge advocate of understanding how people perceive their treatments and how and why they are making decisions about whether the treatment is “worth it”.  The vast majority of this research has taken place in the context of clinical trials where treatment adherence is encouraged and the benefits of participation (and potential harm of withdrawing) extend beyond the medication effects to improved monitoring, enhanced care, and access to novel treatment. Restricting our observations to clinical trials then may mean that we are not getting the full picture on the decisions that people make around medicines in clinical care.

A huge amount of research over the years has explored the reasons for non-adherence and non-persistence to medication in a real-world setting. A wealth of data supports the premise that suboptimal adherence and persistence are in part a function of clinical teams making treatment decisions that insufficiently consider patients’ preferences, experiences, and psychosocial concerns [11,12]. Such research has led the charge to a shared decision-making model in healthcare where the patient and physician partner in making treatment decisions [13] which is a hugely positive step, albeit a slowly moving one. As anticipated, when patients are asked to make decisions about treatment, it is generally based on some combination of the anticipated or realised benefits “the good stuff”, “the bad stuff”, and the trade-off between them [4].

Measuring individualised benefit-harm

The PQAT-RW [14] – a mixed-methods patient-reported outcome measure (PROM) – was developed to measure exactly this. It asks patients to describe (via free-text) and rate (via likert scales) their experience of benefits and disadvantages of treatment, as well as their willingness to continue treatment. Patients are also asked to make a judgement on whether and how much the disadvantages outweigh the benefits of treatment, or the benefits outweigh the disadvantages (responses are given on a 7-point Likert scale with a neutral mid-point of equal benefits and disadvantages). Responses are not restricted to efficacy and safety parameters, but rather patients are invited to think holistically about the treatment, including mode, frequency and convenience of treatment, etc which may appear on both the benefits and disadvantages list. Free-text responses are especially informative when evaluating treatments in heterogenous populations where varied outcomes or experiences are expected, and valued differently, by individual patients.

What could we learn?

Incorporating the PQAT-RW (or similar instruments) routinely into observational studies and prospective registries alongside effectiveness and safety measures will allow for some important and as-yet unanswered questions to be considered, including:

• Is perceived benefit-harm predictive of adherence and persistence of treatment?

• Can perceived benefit-harm provide an early indicator of likely non-adherence?

• Are perceived benefits or disadvantages better predictors of adherence/persistence?

• Is perceived benefit-harm predictive of switching behaviour?

• Are patients with similar clinical experiences (safety, effectiveness) reporting different perceptions of benefit-harm trade-offs? Do these patients differ in other ways (e.g. socio-demographic, diagnostic journey, culture)?

• Can benefit-harm measures improve shared decision-making?

Real-world evidence has transformed our understanding of treatment effectiveness and safety in routine clinical practice. The next evolution should be understanding whether patients themselves believe the benefits of treatment justify its disadvantages. Personally, I believe that individualised benefit-harm assessment should become a core part of real-world research, and I would genuinely love to discuss further with other like-minded people who are interested in moving this conversation forward!

References

[1] Reaney, M., Bush, E., New, M., Paty, J., Roborel de Climens, A., Skovlund, S. E., Nelsen, L., Flood, E., & Gater, A. (2019). The Potential Role of Individual-Level Benefit-Risk Assessment in Treatment Decision Making: A DIA Study Endpoints Community Workstream. Therapeutic Innovation & Regulatory Science, 53(5), 630–638.

[2] National Institutes of Health. (2025, January 21). The promise of precision medicine. U.S. Department of Health and Human Services. Available at: https://www.nih.gov/about-nih/nih-turning-discovery-into-health/promise-precision-medicine. Accessed 10 July 2026.

[3] Reaney, M., Stassek, L., Martin, M., McCarrier, K., Slagle, A., Shields, A., & Gwaltney, C. J. (2019). Creating a personalized evaluation framework for patient-reported outcomes: An illustration using the EQ-5D visual analogue scale. Expert Review of Pharmacoeconomics & Outcomes Research, 19(1), 97–104.

[4] Gater, A., Reaney, M., Findley, A., Brun-Strang, C., Burrows, K., Nguyên-Pascal, M.-L., & Roborel de Climens, A. (2020). Development and first use of the Patient’s Qualitative Assessment of Treatment (PQAT) questionnaire in type 2 diabetes mellitus to explore individualised benefit-harm of drugs received during clinical studies. Drug Safety, 43(2), 119–134.

[5] Buzaglo, J., Roborel de Climens, A., & Reaney, M. (2025). Defining and collecting patient-reported treatment tolerability to inform drug development. Available at: https://www.iqvia.com/-/media/iqvia/pdfs/library/white-papers/2025/defining-and-collecting-patient-reported-tolderability.pdf. Accessed 10 July 2026.

[6] Reaney, M., Chassany, O., Dhillon, T., & Dietrich, E. S. (2021). The PRO-CTCAE to understand symptomatic toxicities with cancer medication: An American dream or a global opportunity [Conference presentation]. ISPOR Europe 2021. Available at: https://www.ispor.org/docs/default-source/euro2021/europe21reaney.pdf?sfvrsn=8c3af423_0. Accessed 10 July 2026.

[7] Reaney, M., Bruce, R., Kelly, K., & Hughes, L. (2024). Expanding conceptual model of disease into conceptual model of patient experience for designing, interpreting medication effects. Applied Clinical Trials (June 28 2024). Available at: https://www.appliedclinicaltrialsonline.com/view/expanding-model-disease-conceptual-patient-experience-interpreting-medication-effects. Accessed 10 July 2026.

[8] Turner-Bowker, D. M., An Haack, K., Krohe, M., Yaworsky, A., Vivas, N., Kelly, M., Chatterjee, G., Chaston, E., Mann, E., & Reaney, M. (2020). Development and content validation of the Pediatric Oral Medicines Acceptability Questionnaires (P-OMAQ): Patient-reported and caregiver-reported outcome measures. Journal of Patient-Reported Outcomes, 4(1), 80.

[9] Bailey, J. R., Fonseca, E., Borsa, A., Hawryluk, E., Gubernick, S. I., de la Motte, A., Karantzoulis, S., Reaney, M., & Saretsky, T. L. (2024). Three novel patient-reported outcome measures to assess the patient experience with daily and weekly HIV oral antiretroviral therapy. Journal of Acquired Immune Deficiency Syndromes, 97(3), 286–295.

[10] Ervin, C., Joish, V. N., Evans, E., DiBenedetti, D., Reaney, M., Preblick, R., Castro, R., Danne, T., Buse, J. B., & Lapuerta, P. (2019). Insights into patients’ experience with type 1 diabetes: Exit interviews from Phase III studies of sotagliflozin. Clinical Therapeutics, 41(11), 2219–2230.e6.

[11] Reaney, M. (2015). The need for a tool to assist health care professionals and patients in making medication treatment decisions in the clinical management of type 2 diabetes. Diabetes Spectrum, 28(4), 227–229.

[12] Reaney, M., McHorney, C. A., Curtis, B., Rydén, A., Chassany, O., & Gwaltney, C. (2017). Using individual experiences with experimental medications to predict medication-taking behavior postauthorization: A DIA Study Endpoints Workstream. Therapeutic Innovation & Regulatory Science, 51(4), 404–415.

[13] New, M., Paty, J., & Reaney, M. (2022). Harnessing patient-centered science to improve health outcomes and commercial pharma success. Applied Clinical Trials (November 1 2022). Available at: https://www.appliedclinicaltrialsonline.com/view/harnessing-patient-centered-science-to-improve-health-outcomes-and-commercial-pharma-success. Accessed 10 July 2026.

[14] Roborel de Climens, A., Findley, A., Bury, D. P., Brady, K. J. S., Reaney, M., & Gater, A. (2024). Development and content validation of the Patient’s Qualitative Assessment of Treatment–Real-World (PQAT-RW): An instrument to evaluate benefits and disadvantages of treatments in real-world settings. Patient Related Outcome Measures, 15, 255–269.