Drug development is inherently risky, with some analysis showing that just 13.8% of products that enter industry-sponsored Phase I trials go on to obtain FDA approval.1 Of course, the exact figure varies across therapy areas and development stages, but it is an expensive problem no matter how you slice the data.
So, what can Sponsors and Contract Research Organizations (CROs) do to increase their chances of success? We take a look at some of the leading causes of study failure and explain how improving medication adherence could be the key to risk mitigation.
Why do Clinical Trials Fail?
Clinical trials are incredibly complex, containing multiple opportunities for study failure. Common reasons include a shortage of funding for completion, failure to maintain good manufacturing protocols or to follow FDA guidelines, and problems related to patient recruitment, enrolment, and retention.
However, the primary source of trial failure has been and remains an inability to demonstrate efficacy, according to a 2018 review of data from the previous 30 years.2
The authors pointed to a study by Hwang et al which assessed 640 Phase III trials of novel therapeutics. It found that 54% had failed in clinical development and that 57% of these failures were attributed to inadequate efficacy 3
Yet failure to demonstrate efficacy during a clinical trial does not always translate to mean the drug does not work. Such a finding could be the result of uncertainties related to dose selection or the study being underpowered, for example2
Safety is another leading reason for study collapse, with Hwang et al finding it to be the root cause of failure in 17% of the Phase III trials examined.3
Again, a study failing due to safety concerns does not necessarily mean the drug is unsafe. Incorrect dosing, for instance, can result in adverse events that do not accurately reflect the effect of the product.4
Ultimately, the reasons for study failure are as varied and complex as the trials themselves. Yet there is a common thread that runs through the most common factors (i.e. efficacy, safety, dose selection), and that is poor medication adherence.
The Adherence Confounder
Poor medication adherence may be a key confounder of the root causes of study failure, but it is not a new problem. It is a well-documented issue and one the industry has been grappling with for some time.
Studies have shown that each Phase III trial participant is responsible for an average of $42,000 in costs5, yet 30% are non-adherent by day 100. 6 Across all clinical trial phases, 50% of participants admit to not following the dosing regimen set out in the protocol.7
And when study participants do not take their medication as directed, it can lead to underestimation of drug efficacy, overestimated dosing requirements, and raise concerns about safety – the key reasons for delay and denial of regulatory approval.
By decreasing effect size and increasing variability, poor medication adherence drains study power. The exponential relationship between nonadherence and sample size means that any decrease in adherence must be met with a risky increase in dose and/or expensive increase in study participants to maintain power and avoid study failure.
Poor medication adherence can thus have a concerning impact on patient safety. It generates poor data, leading to improper calculations of the correct therapeutic dose, and resulting in avoidable adverse events. Participant-driven sporadic dosing patterns, such as taking medication holidays, for example, can also lead to problems such as unnecessary withdrawal symptoms or side effects.
In essence, poor medication adherence gives a distorted picture of how a drug works in each patient population, compounding the very issues that can lead to study failure.
Imperfect Measurement Methods
The factors that lead to poor medication adherence are multifaceted and individual to each participant. And while they fit broadly into two categories, intentional and unintentional, in truth there is a degree of crossover between the two.
Intentional non-adherence is deliberate, and largely associated with patient motivation, whereas unintentional non-adherence is largely driven by a lack of capacity or resources to take medications.
“However, it is important to acknowledge that the reasons underlying intentional and unintentional non-adherence are not entirely independent in those certain types of unintentional non-adherence e.g. forgetting, are logically more likely when the motivation for medication is low,” writes Molloy et al. 8
Traditional tools for the measurement of adherence, such as pill counting, self-reporting, and biomarkers, are simply not sensitive enough to identify non-adherence in this rich tapestry of factors drivers.
Counting returned tablets is the most commonly used adherence measure in clinical trials. However, it is easily censored by participants and only provides a summary of medication-taking behavior between visits.9
The self-reporting approach, whereby participants record their doses in a patient diary, is, in theory, a sound medication adherence assessment method. The reliability of such diaries, however, can be poor, as they are intrinsically vulnerable to inaccuracies and bias.
Even those with advanced digital applications, such as mobile apps, reminders, or even the ability to film the medication being taken, can be troublesome. They require increased patient involvement and are intrusive. This can lead to lower patient acceptability and has been associated with missing data and poor reliability.10
Likewise, monitoring drug or drug metabolites in blood, urine, or hair only provides a snapshot of behaviors. So-called “white coat adherence”, where participants only take the investigational product the day before the visit, is a particular vulnerability. In addition, the approach is invasive, places an additional burden on participants and staff, and is largely restricted to use in the active arms of a trial.
Breckenridge et al highlighted the inadequacies of traditional methods in a 2017 paper on the consequences and solutions of poor medication adherence in clinical research.11
They conclude: “Methods for measuring and encouraging adherence are essential components of clinical trials and every effort should be made to incorporate adherence measures in phase II and phase III trials to avoid important errors in interpreting benefits and harms.”
Digital Adherence Monitoring
Traditional methods are imperfect at best. But digital adherence monitoring, which utilizes connected packaging and powerful data analytics, is different, writes Breckenridge et al.11 It is objective, qualitative, and able to guide informed decisions and interventions.
Connected pre-filled syringes, for example, can collect essential information, such as whether the injection has been completed, time and date of dose, type of drug, batch number, and expiration date.
These data analytics are then transmitted to a cloud-based platform for sophisticated analysis of medication-taking behaviors. Visualizations enable study teams to spot patterns that might indicate “at-risk” participants, allowing risk stratification and guiding individualized interventions.
The approach can even be integrated into third-party applications, such as patient-facing apps designed to continually encourage adherence and engagement with the protocol.
This is no longer an emerging technology. Studies have shown that smart package monitoring is 97% accurate, compared to 60% for pill counting, 50% for healthcare professional rating, and just 27% for self-report.12
Crucially, it provides a complete understanding of the patient adherence behaviors and the risk indicators that matter most for study success.13
Mitigate the Risk of Failure
Drug development is costly and the stakes, both for industry and the patients waiting for life-changing medications, are high. Organizations, then, must do everything they can to maximize their chances of success.
While we know that poor medication adherence can contribute to two of the biggest reasons for study failure – a lack of drug efficacy and safety concerns – addressing the problem has proved challenging.
Yet, with the emergence of digital adherence solutions sponsors and CROs now have the tools they need to improve medication-taking behaviors, thus mitigating the risk of study failure and increasing return on investment.
- Heem Wong, C., Wei Siah, K et al. Estimation of clinical trial success rates and related parameters. (2018). https://academic.oup.com/biostatistics/article/20/2/273/4817524
- Fogel, D. Factors associated with clinical trials that fail and opportunities for improving the likelihood of success: A review. (2018). https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6092479/
- Hwang, T., Carpenter., D. et al. Failure of Investigational Drugs in Late-Stage Clinical Development and Publication of Trial Results. (2016) https://pubmed.ncbi.nlm.nih.gov/27723879/
- Berlin, J., Glasser, S. et al. Adverse event detection in drug development: recommendations and obligations beyond Phase 3. (2008). https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2446471/
- Sertkaya, A., Birkenbach, et al. Examination of clinical trial costs and barriers for drug development. (2014). https://aspe.hhs.gov/system/files/pdf/77166/rpt_erg.pdf
- Blaschke, T., Osterberg, L., et al. Adherence to medications: insights arising from studies on the unreliable link between prescribed and actual drug dosing histories. (2012). https://pubmed.ncbi.nlm.nih.gov/21942628/
- Eliasson, L., Clifford, S., et al. How the EMERGE guideline on medication adherence can improve the quality of clinical trials. (2020). https://bpspubs.onlinelibrary.wiley.com/doi/full/10.1111/bcp.14240
- Molloy, G., Messerli-Burgy, N. et al. Intentional and unintentional non-adherence to medications following an acute coronary syndrome: A longitudinal study. (2014).https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4005033/
- Larsen KG, Areberg J, Åström DO. Are self-reported and self-monitored adherence good proxies for reaching relevant plasma concentrations?: Experiences from a study of anti-depressants in healthy volunteers. (2021) 9
- Breckenridge, A., Aronson, J., et al. Poor medication adherence in clinical trials: consequences and solutions. (2017). https://www.nature.com/articles/nrd.2017.1?WT.ec_id=NRD-201703&spMailingID=53536275&spUserID=ODkwMTM2NjI2OQS2&spJobID=1120295377&spReportId=MTEyMDI5NTM3NwS2
- Alili, M., Vrijens, B., et al. A scoping review of studies comparing the medication event monitoring system (MEMS) with alternative methods for measuring medication adherence. (2016). https://pubmed.ncbi.nlm.nih.gov/27005306/
- Vrijens, B., Urquhart, J. Methods for measuring, enhancing, and accounting for medication adherence in clinical trials. (2014). https://pubmed.ncbi.nlm.nih.gov/24739446