Getting to grips with "fast-track" development challenges
by Dr John Attridge, for Medical Device Technology Magazine

The current financial climate puts considerable pressure on companies to reduce timescales for the development of new products. In the past, many organizations would undertake a sizeable portion of this work in-house, recruiting people with relevant expertise across the range of skills required. Now, however, it is more common for companies to concentrate on their core skills, assay development for example, while finding a ‘development partner’ to tackle the instrumentation side of the programme.

As an instrumentation company we have had to be responsive to these changes. We have invested heavily in the techniques and skills necessary for the rapid development of our clients’ products and can take a product from concept right through to final manufacture.

Many companies might consider using a small consultancy firm for the initial development work, seeing this as a low-cost route to getting a new idea off-the-ground. What is not often appreciated, however, is the considerable additional cost and time incurred in transferring prototype designs into a second organisation for manufacture. By having an in-house production capability we ensure that production requirements are integrated into the design, right from the outset.

Our specialist approach is illustrated by the recent rapid development of a hand held photometer for our clients, Provalis PLC (formally known as Cortecs Diagnostics). This is one of a range of point-of-care diagnostic and therapeutic monitoring products that are presently driving growth in the medical device market.

This was Provalis’ first instrument based test so it was important for them to choose a development and manufacturing partner with the experience to understand the complex issues involved in assay development. To have any hope of meeting their development deadlines, the partner would have to have in-house expertise in technical and scientific evaluation, industrial and mechanical design, through to software and electronics engineering.

When Provalis first approached us they had already developed a visual immunochromatographic test device for the diagnosis of osteoporosis, or ’brittle bone disease’. They now wished to develop a hand held reader that could provide a more objective quantitative result.

To perform the test a moulded sample collection ‘comb’ is simply dipped into a urine sample and inserted into the test device. This action transfers a metered volume of sample to a filter pad impregnated with coloured reagent. Capillary action ‘draws’ the sample and dissolved reagent through the filter onto a thin nitrocellulose membrane printed with two zones of immobilised antibody. A test zone identifies whether a target marker molecule for osteoporosis is present in the sample, while a second control zone indicates that the assay has run to completion. Depending upon the type of assay, either a strong or weak line develops in the test zone when the marker molecule is present.

For many products on the market, e.g. pregnancy tests, a simple "yes" or "no" result is all that is required, and it is usually possible to discriminate unambiguously between these two possibilities. A problem arises when a quantitative result is required, as the interpretation of line density is highly subjective. Our challenge was to develop a simple, low cost, hand held device capable of recognizing at least five levels of line density.

At an initial meeting Provalis outlined a challenging 10-month development schedule along with a low cost target for the reader. A project proposal was prepared in four days outlining the following project phases:

• Phase I: scientific assessment, proof of principle, concept illustrations and appearance model

• Phase II: detailed design, production engineering, prototyping

• Phase III: pre-production manufacture and validation

• Phase IV: serial manufacture

A fixed price was provided for Phase I together with estimates for the later phases. This gave a budgetary estimate for the entire project but did not commit either party to the complete development programme before any of the technical issues had been addressed.

The first activity was to review all the technical issues and to produce a draft requirement specification. Almost immediately a choice had to be made between absorbance or reflectance measurements of line density. Absorbance was preferred because, when calculated correctly, it is essentially a self-calibrating measurement; an important issue for a small hand held instrument.

Experimental work started immediately. By adapting tried and tested electronics from one of our marine fluorimeters we were able to acquire absorbency scans of assay devices within two days of starting the project. The feasibility of the measurement was quickly established, and emphasis shifted towards optimising the assay for an instrument-based test rather than visual interpretation. This meant that Provalis now required a laboratory reader to begin work on the assessment of key features of the assay, such as: response shape and position, device to device reproducibility and signal to background. This instrument was designed and delivered within two weeks, enabling rapid progress to be made with the chemistry.

In parallel with the experimental effort, work was in progress to produce an instrument design concept that would be suitable for cost effective, high volume manufacture. This was developed in-house and concept drawings and an appearance model were provided, together with a detailed technical report, in time for the scheduled completion of Phase I, one month into the programme.

Phase II began immediately with the aim of detailing the reader design and producing a fully functional prototype within three months. The prototype was to be capable of measuring the absorbance of the test strip in four positions to provide an assay, control and two background readings.

At this stage it became clear that, to generate the volume of data required for clinical trials, it would be necessary to develop additional readers capable of downloading test data to a PC. Though these readers would not require the final moulded casework, for clinical validity, their performance would still have to be a true representation of the final design. The development of this additional reader was run in parallel with the main programme.

The delivery of a working lab reader was dependent upon the early definition of the interface between the test device and the reader. The feasibility work had demonstrated that the reported absorbance was critically dependent upon the optical geometry. It was also important to incorporate features that would warn users if a test device had been inserted incorrectly. A design for the interfacing components was developed and visualised using ProEng, a 3D CAD package. The part files generated by this programme were then e-mailed for stereolithographic reproduction in epoxy resin, enabling us to receive solid models with a few days of completing the design.

Once the reader interface had been defined, the external ergonomics of the instrument could be developed. ProEng allowed the complex geometry of the appearance model to be described in a way that would have been extremely time consuming using conventional 2D drawings.

The reader was destined for use all over the world, so, to avoid language difficulties, it was proposed that the user would be prompted by an icon based liquid crystal display. Because of the long lead times involved in custom LCD manufacture, the highest priority had to be given to finalising the display’s design. A typical test protocol was defined and a ‘storyboard’ prepared for evaluation by Provalis’ marketing group. Once agreed, the final LCD design was generated in-house and sent to Singapore for tooling and prototyping.

A key feature of the reader was that it should not require battery replacement over the predicted three-year life of the instrument. This presented a considerable challenge for the electronic design team. Key to meeting this requirement was the identification of a ultra low-power microprocessor that incorporated all the functions required by the instrument. This also meant that very little external circuitry was needed which helped to keep the component costs down. Early calculations showed that this processor, in combination with a high energy density battery, would provide the solution.

The measurement of line density also presented a number of unique challenges. Wet nitrocellulose has an optical density of 1, so all the useful assay information is contained in only 10% of the available signal. To keep final manufacturing costs within target it was necessary to avoid any set-up or calibration of the reader in production. This meant that the signal to noise characteristics of the analogue circuitry had to be high enough to compensate not only for the demanding requirements of the measurement, but also a significant variability in LED output. A low-noise design for the analogue electronics was developed and analysed in detail to optimise both measurement reproducibility and battery lifetime. Circuit schematics were prepared and printed circuit board layouts completed and fabricated on a three-day turnaround into order to meet the agreed reader delivery schedule.

The prototypes used a re-programmable EPROM version of the microprocessor to enable rapid code development. A microprocessor development system was chosen so that the embedded software could be developed without the need for any of the final hardware. This meant that a tested version of code was available as soon as the circuit boards became available.

A development reader and a fully functional prototype were delivered at the end of this three-month phase of work in time for successful clinical trials and product announcement.

In this phase the design of the casework and internal moulded components were finalised so that injection mould tools could be fabricated. A toolmaker was identified that could handle the 3D design files directly, thus avoiding laborious detailing in 2D. A complete production documentation package was then compiled in accordance with our ISO9001 accredited quality procedures in preparation for a transfer to manufacture.

During this phase of work 10 pre-production instruments were required to validate the reader design and assay performance. As the lead-time for injection moulding is typically 12 weeks, we had to find an alternative to produce the number of casework and internal components required for these readers. Stereolithographic models were used to produce soft mould tools allowing components to be vacuum cast in polyurethane. This process would enable up to 20 components to be produced per tool. Using this approach we were able to deliver these pre-production readers seven months into the programme.

The electronics design went through a further iteration with the EPROM version of the microprocessor being replaced with the ‘one time programmable’ device to be used in production. Production documentation was finalised and long lead items identified and ordered in readiness for full-scale volume production.

From the initial contact with the client, rapid prototyping techniques have allowed us to develop an idea into a fully validated design, ready for volume production in less than ten months.

(top)
..