In my article about the Patient Engagement Open Forum in Brussels this September, I shared with you the discussions around measuring the impact of patient engagement – something essential to actual patient-centricity. There is a list of things that professionals who attended confirmed they can measure and one of them is Recruitment Rate.
Recruitment Rate is a metric that doesn’t come up only in relation to patient engagement. Actually, it is far more important for feasibility and clinical operation experts at CROs and pharma companies who plan clinical trials, as this is one of the main metrics predicting the speed of completing a study.
At FindMeCure we also used to rely on it to build better strategies for patient recruitment (when this was the focus of our company), among other real-time data we use. We have done extensive research on the ways Recruitment Rates can be estimated and what the main challenges are.
In this article, I will outline some of the basic and “must-know” things about Recruitment Rate with the hope to help you use this metric as reasonably as possible.
1. How much time does it take to calculate Recruitment Rate (RR)?
The Recruitment Rate (RR) of a clinical trial is about how many patients were recruited on average for one month in one site. RR is calculated individually for each site in order to measure the site performance.
Calculating RR per study is probably the most argued about process among feasibility and clinical research experts.
There are two main ways to do the calculation. Both of them have pros and cons:
- Average Recruitment Rate (ARR) of the study is when you sum up all recruited patients and divide by the number of sites. This is pretty popular among feasibility experts as a way to see the overall performance of the sites. The reason why some professionals frown upon this technique and label it “inaccurate” is that this way you calculate the performance of high-performing sites and low-performing sites together. The result is an Average Recruitment Rate that maybe few or even none of the sites actually reached – some were with higher RR and others with lower. The pro about ARR is that you can calculate it with some level of confidence for trials for which you don’t know the actual recruitment for each site. The resulting number, while not perfect should serve as a good benchmark.
- Median Recruitment Rate (MRR) of the study is when you list all RR of sites and find the number in the middle. This way you have a real patients/month/site number and again is some way average for the study. The obvious drawback of this method is that you have to know the study details. Unless the study was conducted by/for your organisation, this is a hard task.
2. What data do you need to calculate Recruitment Rate?
If you are from a company like Pfizer, IQVIA, Novartis, Covance, etc. you will benefit from a lot of in-house data including the speed of recruitment for a particular study per site. Then they additionally build algorithms around ARR and MRR per study and indications they have previous experience with. This can also be applied to sites and their historical performance whose methodology is used by TransCelerate where many of the biggest organizations share their historical data in order to improve their site and investigator selection.
Clinical Research Organisations that are focused on concrete Therapeutic Areas/indications and regions (for example Dermatology, Germany) are also quite useful in this case as they also have a lot of historical data.
What do you do when you don’t have historical data?
When CROs are working in a new TA or indication or just don’t have enough historical data to build a benchmark for RR, they rely on two other sources to help them predict their patients/site/month:
- Site surveys: They go and ask sites about their previous performance and historical data. The challenge here is that you rely on the site to track this and to be transparent with all their previous clinical trials.
- Publicly available clinical trial registers: Registers like EudraCT and CT.gov contain a lot of data including the study timelines, patients recruited, etc. Thanks to this data you can calculate ARR for the study and try to use it as a comparison metric when looking at the site’s reported previous experience.
3. How to mitigate the risk when using Average Recruitment Rate?
Like I mentioned in point 1, there are a lot of discussions about the accuracy of ARR. Our TrialHub team has spent a lot of time applying different statistical methodologies in order to find a way to mitigate risk around calculating ARR. If you are manually calculating ARR, there are two things that you need to be doing:
- Make sure you calculate all available study ARRs and identify the so-called “outliers” – the ARRs that are too high or too low in comparison to the rest. You might want to “ignore” them when calculating your final ARR.
- Something that can help you get closer to statistical accuracy is to answer the question: Based on how many studies’ ARRs is my final ARR calculated? If it is below 10 studies’ ARRs it might not be so reliable. If it is above 10 studies then you can rely on statistical accuracy.
4. What else should we calculate besides ARR per study?
Screen-Failure Rate and Average Dropout Rate are two metrics that are not very popular for planning a clinical trial. However, they are essential for predicting successful recruitment. Like RR and ARR they can be based on historical data and publicly available data. In both cases having these metrics can help you prevent challenges caused by unreasonable (in the patient perspective) screening procedures and by losing participants even when you have high RR.
Do you want to know more about How to measure patients’ interest in clinical trials? You can jump on a quick call with us and see how TrialHub helps you do that quickly and efficiently.
First published on LinkedIn