Building a better self-injection solution
PA's Kevin Deane, a medical device expert, is quoted extensively in an article in Pharmaceutical Technology, which looks at self-administration becoming a fast trend in healthcare today. Kevin discusses how the pharma industry is looking towards autoinjectors as a drug-delivery solution that can offer improved patient experience and compliance.
Kevin says that as a result, the market for autoinjectors is expanding rapidly, driven primarily by two factors: “the proliferation of biologic compounds, which are filling pharma’s pipelines, and the need to move drug administration away from clinics and hospitals into home use and self-administration.”
Kevin notes that Humira (adalimumab) and Enbrel (etanercept), which are among the top five best-selling drugs, are self-administered by patients using autoinjectors.
The article goes on to discuss the different aspects involved in developing an autoinjector, including the importance of human factor engineering.
Kevin talks about the challenges of developing an autoinjector and says: “The key challenges in developing an autoinjector are cost of goods, drug viscosity, and freedom within a crowded IP landscape. The requirement to minimize the cost of goods drives the designer to minimize the number of parts in the device. This is most effectively achieved by placing multiple functions on each part, but this adds to the complexity and increases the importance of design for manufacture and assembly activities.
“Drug viscosity can prove a challenge both at high and low viscosities. At high viscosity, a high power source is required to ensure the injection time is acceptably short for the user. Retaining such high loads throughout the shelf life of the device and ensuring the device still functions, are challenging and require extensive modeling and accelerated aging to simulate high stress and long life. At low viscosity, the challenge is lengthening the injection time to avoid pain, but at the same time, avoiding device stall as friction loads dominate the injection. Both of these aspects can be partially solved by modification of the needle size, but this change is not available to the designer.
“The IP landscape for autoinjectors is crowded, particularly in the space offering the simplest and most reliable device architecture. A sleeve actuated autoinjector, which is becoming the standard embodiment, is not a complex device. The function can be fulfilled with a small number of parts, and thus, there are a limited number of mechanisms that will provide the functionality. As a result, there are limited opportunities to innovate without infringing existing IP. To overcome this design constraint and avoid the need to license, the designer is driven to move to less elegant design solutions, typically introducing additional parts or suboptimal mechanisms, with a knock-on impact on molding and assembly complexity as well as the inevitable impact on cost of goods.”
The article goes on to look at how prefilled syringes and cartridges be effectively integrated into an autoinjector.
Kevin explains: “The prefilled syringe was never designed to be mounted within an automated delivery device such as an autoinjector. As a result, the geometry and emptying force are poorly defined and controlled. Unlike a person, an autoinjector cannot adapt to this variability and thus, the user is exposed to inconsistent injection depths and variable injection times. Moreover, there is a risk of damaging the syringe, presenting a safety risk to the user. The designer can solve some of these problems with complex mechanisms (e.g., electromechanical closed-loop feedback systems), but to effectively integrate the syringe requires extensive testing and characterization. The characterization is complicated by the lack of access to multiple syringe manufacturing batches (both unfilled and filled) and the difficulty of capturing the force to expel the drug.
“Access to multiple batches can be impossible if the drug is in development, and filling is not representative of the production fill. Even in later stages, getting hold of multiple batches to assess process variability is difficult due to the mismatch in syringe and autoinjector production volumes. This issue necessitates that characterization activities continue throughout the development.
“Capturing the force to expel the drug is challenging as the typical autoinjector power source, a spring, generates a decaying force during the delivery of the dose. Some theoretical modeling is possible, but limited in its validity. Empirical testing presents its own challenge as standard test equipment does not exist. Dynamic force testers are available but have limited functionality. Thus, typically a combination of constant force, constant velocity, and device power pack testing is used to measure the force to expel the drug.
“Cartridges require the same force characterization, but the geometry specification is much tighter. The additional challenge of the cartridge is designing the interface with the needle. In its simplest form, this design can be user applied (which has associated usability challenges and user safety risks). In the more complex form, the device must automatically couple the needle to the cartridge, piercing the septum as part of the injection process.”
Innovation As Unusual: How organisations can stop wasting their best ideas
The article goes on to look at regulatory guidelines in the industry and how human factor engineering can be integrated in the development of an autoinjector.
On this, Kevin says: “Human factors work is most efficient and productive when planned from the start as part of an iterative development program. Initial research should be carried out to understand the intended users and how their capability might affect their interaction with an autoinjector; what the context of use would be and how that might affect the interaction with the device; and what training would be expected.
“For example, an autoinjector for rheumatoid arthritis may have two distinct user groups: healthcare professionals and lay users. The design of the device would need to consider their physical impairment (reduced hand strength and mobility) and the size of device (larger proportion of female patients, generally adults). Because people with rheumatoid arthritis are generally older adults, consideration would also be needed for age-related decline in visual acuity and hearing.
“The next stage is to incorporate human factors into risk management to identify foreseeable misuse scenarios and to identify which use steps are related to patient safety and the ability to deliver medication effectively. These so-called critical tasks help inform what usability testing should be done. During the design and development phase, formative usability testing helps inform and iterate the design as well as identifying unforeseeable misuse scenarios that would need to be tested during the final simulated use human factors validation study (previously known as a summative study).
“Formative testing during the design phase is generally with small groups (5-8 participants) and may be aimed to look at one aspect of the design, an early sketch model, or high fidelity prototype. These experiments are not preference studies but are aimed to uncover key usability and safety issues related to the design. Human factor validation studies need to be run like a repeatable scientific study, although they are generally not clinical studies. FDA recommends using a minimum of 15 participants per distinct user group. The tasks carried out in the human factors validation study should be linked to the tasks identified in the human factors risk assessment and be comprehensive in scope to represent generalized use.
“FDA has provided guidance on which aspects of an autoinjector should be tested, including:
- The ability of users to read and understand the autoinjector instructions
- The ability to set up the autoinjectors
- The ability to perform an injection
- The ability to dispose of the injector.”
To conclude, Kevin talks about verifying and validating the autoinjector and the tests that need to be carried out. He says: “Verification of performance against functional specifications is a critical part of the regulatory approval process for autoinjectors. Verification planning begins early, in parallel with the development of functional specifications, and continues throughout the design process. Development test data are used to understand key risk areas for the design and highlight potential measurement challenges. A design verification matrix establishes full traceability between functional specifications and the planned set of verification tests, defining the scope of the verification test program.
“Wherever possible, test methods are based on automated measurement processes to effectively remove the operator contribution to gauge variability. A wide range of measurements, generally including delivered dose, needle depth, needle protection override force, and audible dose indicator performance, are all integrated into a single measurement sequence, streamlining the method validation and test processes.
“Acceptance criteria are derived in accordance with ISO 11608, combining accuracy (process mean and specification limits) and precision (process standard deviation) into a single, target tolerance-limit factor that defines a minimum required pass rate at a given confidence level. Analysis methodologies are fully defined before commencement of testing and are focused on assessment against the predefined acceptance criteria. Subtleties include taking into account the underlying distribution of the measurement data; for example, injection times for mechanical autoinjectors generally follow a log-normal distribution.
“Design validation uses inputs from regulatory guidance, risk analysis, formative user studies, client requirements, and verification results to design a user study with a representative end-user population. This design validation study comprehensively addresses the high-level product requirements identified with user needs and intended product uses, in addition to foreseeable misuse scenarios.”