Barry Brewster, Head of Team's Project Management Group, writes:
So, you’ve developed a brilliantly clever diagnostic technology. You might be thinking the hard work is now done and all you have to do is make it point-of-care and launch a killer product…
Before you do – take a minute to consider our six rules for developing a new diagnostic device. We have put these rules together to help you in your journey to a commercially successful device, drawing on over 20 years of experience developing products ranging from home pregnancy test sticks to state-of-the-art lab on chip diagnostics for intensive care units.
1 Planning a successful product
You may have a brilliant technology, but that alone doesn’t guarantee commercial success. You need to be clear on the key attributes that will allow your product to differentiate itself in a competitive market. Define the intended use of your product, as well as where and to whom it will be sold. Being clear on the intended use and intended user is core to defining the regulations you need to meet. It will also help you lay out your regulatory strategy, which you should establish as soon as possible.
Be prepared to spend some time doing market sensing and early user engagement to hone your technology and product vision into a customer requirements specification. This will help steer you on the right course to avoid the costly mistake of developing the wrong product with your brilliant technology.
2 Research – what do you need to prove your product
At some point, before launching your product, you are going to have to prove your technology to the regulators and win over some customers. Take time to research the type of testing you will need to do to prove the performance of your product. This applies to standards and regulatory compliance as well as the benchmarking that might be required to secure reimbursement or approvals from bodies such as NICE. In addition to the harmonised standards – soon to be common specifications associated with regulatory compliance – look at product literature for competitor devices to determine what sort of performance testing they did.
“At some point, before launching your product, you are going to have to prove your technology to the regulators and win over some customers.”
Research similar products on the MHRA and FDA websites and look to get a good feel for the type and volume of data they will expect you to submit. Remember that the FDA offers a free IDE pre-submission process where they will review your plans and advise what they consider is necessary to gain approval for your product.
Now you have done your research you can build a realistic budget and timeline. How many devices are you going to need to cover all this testing? How many batches will you need? How long might it take to complete? What volume of reagents will you need? Who will do this work? What facilities might you need e.g. a micro lab, specialist cabinets etc.? Plan your clinical study and think about how you will access patient samples; consider the sample numbers you need for a trial and the time required to recruit patients, particularly if prevalence of the disease/analyte is low. Consider what other clinical evidence might be required.
Thinking about the end game in this way will help you understand the volume of testing you need to undertake to ultimately prove the performance of your device. In the short to medium term it will also force you to scale up appropriately. Scaling up at an early stage will speed up your product development by allowing you to run (and repeat) the right experiments faster and at the required scale.
“Scaling up at an early stage will speed up your product development by allowing you to run (and repeat) the right experiments faster and at the required scale.”
3 Tackle shelf-life early
Evaluate your shelf-life early on to flush out any issues. Diagnostics assays can often be a mixture of biological reagents (e.g. antibodies, enzymes, nucleic acid) and the plastics, metals and electronics that form the ‘cartridge’. Biological reagents do not have the same robustness as inorganic compounds and are degraded by more than just time and heat. Volatiles from plastics, glues, electronics and inks can all affect the activity of your biological reagents and ultimately degrade your assay performance. Setting up regular shelf-life studies can help establish whether any of the components are resulting in shelf-life issues. Your final shelf-life study must be set-up using the “cartridge” in its final embodiment and primary packaging. It is also a good idea to set-up interim studies before the final embodiment is fixed, using representative samples of the other components so as to uncover any problems early (see later point about quality control).
4 Understand the demands of production
Going from bench to full-scale production is likely to be an increase in scale of at least 2 orders of magnitude. Reagents will disappear at 100x the rate they did when you were working in the lab. How robust is your assay to changes in batches of raw materials? If you change batch of raw materials part way through a batch of your assay, will the change affect its performance? In theory, all batches of raw materials are high grade and high purity but in practice you may find that some batches perform better than others. You need to know which materials affect the performance of your assay and which leave it unchanged. Those that leave the performance unchanged can be mixed within a batch; those that change performance can’t be mixed within a batch. Be aware that biological reagents may be sold on the basis of their purity – not their activity or immunoreactivity – so performance of these in your assay can vary from batch-to-batch.
5 Establish robust quality control (QC) procedures
You need to produce consistent product and you need to be able to monitor how consistent it is, so that you can be sure that each batch will meet the claimed performance. How are you going to monitor this performance from batch to batch? What are the key parameters you are going to measure? It’s highly likely you are going to need some standard solutions of your analyte. Ideally these will be referenced to a gold standard. If your test is visually read by the user then you may still need a reader system that will give you a quantitative output so you can monitor performance from batch to batch and see if it is trending up or down. Again, think about how you will evaluate the performance of each batch and how many devices this will need so you plan for the correct scale, and be aware that QC can use up a lot of devices so plan your batch size accordingly and build this into your cost model.
6 Determine how to steer your immunoassay
If you are making a visually read test then you will need to be able to steer the assay up and down (increase or decrease sensitivity) to ensure performance is maintained. What levers (if any) do you have for controlling the sensitivity of your assay? How big is your acceptable window of sensitivity?
If your visually read test becomes too sensitive then you may start to get false positives and if the sensitivity drops then your assay might not be able to meet the claims on the pack insert and false negatives will be the result.
A similar set of considerations apply for a test read by an instrument, although calibration can correct some of these issues. Be aware though that even for an instrument-read test there will be an optimum sensitivity for your assay to meet its performance claims. Too low a sensitivity can still result in false negatives and too high could result in false positives, a loss of linearity at high analyte values, or a loss of precision due to increased “noise”.
I hope these six rules help you in your diagnostic product development journey. Planning ahead to an appropriate level of detail is key to managing risk and avoiding nasty surprises late in your journey. There will always be more problems to solve and challenges to overcome but it’s a rewarding trip. Get in touch if you would like to discuss any of these points.