Institute of Tropical Medicine
Collect2Know (C2K): designing a <$2 blood self-sampling device for home use — from first prototype to peer-reviewed results, a patent, and a Red Dot Award
Voxdale partnered with the Institute of Tropical Medicine (ITM) in Antwerp to co-develop the Collect2Know (C2K) device — a low-cost, single-component blood self-sampling tool that enables home-based capillary blood collection — running five iterative prototypes alongside a structured clinical study, filing a patent, publishing the results in a peer-reviewed journal, and winning a Red Dot Award in 2026.
Executive Summary
Collecting blood at home, without a nurse or a lab, sounds simple. The engineering reality is harder: existing methods fail 20–45% of users, they're unhygienic, and they can't reliably collect the volumes that infectious disease and chronic condition testing require.
The Institute of Tropical Medicine came to Voxdale with a clear brief: design a device that anyone can use at home, collects 500 µL or more of fingerprick blood reliably, and costs under $2 to manufacture.
Voxdale embedded its engineers in ITM's clinical research programme — developing five prototypes, testing them with 101 participants across six structured iterations, and using each round of usability and blood-volume data to drive the next redesign.
The critical breakthrough: replacing a multi-component 90-degree clip with a single-component device placed vertically over the fingertip. Blood collection success rose from 30% to 77.5%. The result has been patented, published in a peer-reviewed journal, and won a Red Dot Award.
Client Context
The Institute of Tropical Medicine (ITM) in Antwerp is one of the world's leading institutions for research, education, and clinical care in infectious and tropical diseases. With deep global partnerships across Africa, Asia, and Latin America, ITM's work sits at the intersection of science and public health practice — producing knowledge designed to reach people who need it most.
For a significant group of ITM's priority populations — HIV/STI patients, people on chronic medication, individuals in low-resource or rural settings — routine blood sampling creates a real barrier to care. Venipuncture requires a clinic visit, a trained professional, and laboratory infrastructure. When people need frequent testing, that barrier compounds. When they face stigma, limited mobility, or simply live far from a facility, testing may be abandoned entirely.
The SARS-CoV-2 pandemic underscored the urgency. Decentralised, home-based testing became a public health necessity overnight — and the limitations of existing self-sampling methods became impossible to ignore.
ITM's team, led by researcher Irith De Baetselier, had the scientific insight and the clinical network to validate a new device. What they needed was an engineering partner who could translate that insight into a product that could actually be manufactured, distributed, and used by real people in real settings — including settings where a $15 device is simply not viable.
The Brief
The clinical brief.
-
Collect 500–1000 µL of capillary fingerprick blood in a standardised, reliable, user-friendly way
-
Designed for use at home, without any healthcare worker present
-
Compatible with standard BD lancets and microtainers — no new consumables
-
Manufacturable at under $2 per unit — specifically to be viable in low-resource settings
-
Classified as an accessory to a medical device (not a medical device) under EU MDR, to simplify regulatory approval
-
Performance targets: ≥75% usability and acceptability scores; ≥60% of users collecting ≥500 µL
The harder, less obvious part.
The brief carried constraints that were fundamental and interacting — not engineering details to be optimised later:
-
Volume reliability at home. The existing free-fall fingerprick method succeeds in only 55–80% of users even in online HIV/STI studies with some support. Without any guidance, untrained users need a device that compensates for the wide variation in how people hold their fingers, apply pressure, and respond to a lancet.
-
Ergonomics as a performance variable. Blood flow from a fingerprick is directly affected by hand and arm position, pressure, and movement after puncture. A device that forces an unnatural finger angle — even slightly — will reduce blood yield. The clip geometry and the user's finger position were the same problem.
-
Single component for manufacturability. Early prototypes had multiple components — a cup top, cup bottom, pin, and spring. Each added assembly cost, quality risk, and user confusion. The brief required a single injection-moulded component. That constraint shaped every design decision.
-
Iterative co-design with a clinical study. The device and the evidence base had to be built together. Each prototype version was not just an engineering iteration — it was a formally structured clinical study with ethical approval, informed consent, and published outcomes. Voxdale's engineers had to work within a clinical research framework where design changes were governed by usability data, not just engineering judgement.
What Voxdale Delivered
Voxdale engineered five generations of the C2K device in direct co-development with ITM's clinical research team.
P1 — first prototype
A 3D-printed clip placed at a 90-degree angle on the finger, with the lancet fully integrated — allowing only a single puncture attempt. Tested with 10 participants at ITM. Only 30% collected ≥500 µL. Fully redesigned.
P2 — removable lancet, multiple pricks
A 3D-printed clip, still in the 90-degree orientation, with a removable lancet allowing multiple puncture attempts. Tested with 10 participants. 60% success — meeting the threshold and triggering a move to injection moulding. Usability above 75%.
P3 — injection-moulded P2
The multi-component architecture translates into injection moulding. Tested across three iterations (supervised at ITM, unsupervised at home, and with healthcare worker correction). Blood volume success remained at 54–60% — consistently failing the threshold when unsupervised. User feedback identified the 90-degree position as a recurring difficulty. Full redesign initiated.
P4 — breakthrough design
A completely new architecture: a single-component clip placed vertically over the fingertip, with two pricking locations and multiple prick capability. No assembly required. Designed for injection moulding from the outset. Tested with 40 participants across two iterations. Blood collection success: 77.5% (31/40). Usability: 89.1%. Acceptability: 75.6%. User preference over P3: 82.4%. Design freeze confirmed.
P5 — exploratory modification
A modified P4 with a single pricking location and a shortened blood travel path. Tested in parallel with P4. While P5 showed advantages in blood spillage reduction, P4 was consistently preferred (64.7%) and performed better on volume collection. P5 discontinued.
-
Design for manufacturability. The final P4 design was developed as a single-component injection-moulded part — no assembly, manufacturable at an anticipated unit cost of under $2. Voxdale optimised the geometry for both clinical performance and production constraints simultaneously.
-
Patent filing. The prototypes are patented under International Publication Number WO 2024/089619 A1.
-
VLAIO-funded R&D partnership. The development was funded by the Flemish Government innovation agency VLAIO (grant HBC.2021.0642), supporting the full collaborative research and development programme between Voxdale and ITM.
How We Worked
Voxdale engineers worked directly embedded in ITM's clinical research programme — not as a supplier delivering prototypes for external testing, but as co-authors of the design and the evidence it generated.
-
ITM research team (Irith De Baetselier, lead; Corinne Herrijgers, Ella Van Landeghem, Tom Platteau and colleagues) — clinical study design, participant recruitment, usability and acceptability measurement, blood volume analysis, publication.
-
Voxdale engineering team (Rien Wymeersch, Kristof Sorgeloos, Koen Beyers, Tim Dierickx) — prototype design and manufacture, materials selection, design for manufacturability, IP filing.
Each iteration followed a formal protocol: a minimum of 10 participants, structured usability and acceptability questionnaires (ASQ scale), blood volume measurement, and qualitative observation by study staff. The results — not engineering preference — determined whether to move forward or redesign.
This discipline — letting clinical evidence govern engineering decisions rather than the other way around — is what produced a device that works for the actual users it was designed for: people at home, without guidance, pricking their own fingers.
The Hardest Challenge — and How We Solved It
The challenge.
Three prototypes (P1–P3), each generation more refined than the last, failed to reliably get untrained users to collect ≥500 µL of fingerprick blood independently. Usability was acceptable. Acceptability was acceptable. The blood volume wasn't.
Why it was difficult.
The 90-degree clip architecture seemed mechanically sound. In controlled conditions with healthcare worker support, it performed adequately. Unsupervised, at home, by people who had never used any self-sampling device, it didn't. The device was asking users to hold their arm and hand in a slightly unnatural position — and that slight unnaturalness was enough to reduce blood flow and make the difference between a sufficient sample and an insufficient one.
What would not work.
Refining the P3 geometry further. Adapting the instructions for use (attempted in IT4 — it didn't close the gap). Adding healthcare worker correction as a fallback (moved success from 55% to 55%, not to 60%). The data was clear: the architecture was the problem, not the execution.
What we changed:
-
Rotated the device axis entirely. P4 placed the device vertically over the fingertip — aligned with the natural finger axis, not perpendicular to it. Users could hold their hand in a natural, comfortable position. The ergonomics worked with the body, not against it.
-
Eliminated all assembly. The multi-component architecture of P1–P3 (cup top, cup bottom, pin, spring) was replaced with a single injection-moulded part. Fewer components meant fewer failure modes, less user confusion, and direct compatibility with the sub-$2 unit cost target.
-
Two pricking locations. P4 gave users two possible puncture sites on the same device — reducing the consequence of a missed or insufficient first prick without requiring the user to restart with a new device.
-
Let the data decide design freeze. Rather than engineering a sixth prototype to chase incremental improvements, the team ran P4 against a modified version (P5) in a direct comparison. P4 won on blood volume, acceptability, and user preference. Design freeze was called.
What it proved.
P4: 77.5% of participants collected ≥500 µL. Usability 89.1%. Acceptability 75.6%. Both primary performance thresholds cleared — in a formally structured clinical study, with ethics committee approval, with 40 participants across two independent iterations. Published in Expert Review of Medical Devices, Vol. 22, No. 10, 2025 (DOI: 10.1080/17434440.2025.2553046). Red Dot Award 2026.
Prototypes, Tests, Evidence
-
Five prototypes (P1–P5), six iterations (IT1–IT6), 101 participants. Each iteration conducted at ITM (supervised) or at home (unsupervised), with ethical approval from the ITM Institutional Review Board and the Ethics Committee of the University Hospital of Antwerp.
-
Usability: measured via the After Scenario Questionnaire (ASQ, 7-point scale). P4 final score: 89.1%.
-
Acceptability: measured via structured questionnaire based on HIV self-testing literature (5-point scale). P4 final score: 75.6%.
-
Blood volume: ≥500 µL in 77.5% of participants (31/40) in the final P4 iterations.
-
User preference: 82.4% preferred P4 over P3; 64.7% preferred P4 over P5.
-
Patent: WO 2024/089619 A1
-
Peer-reviewed publication: Expert Review of Medical Devices, 2025, Vol. 22, No. 10, pp. 1127–1134. DOI: 10.1080/17434440.2025.2553046. Published online 25 August 2025. Authors include Voxdale team members Rien Wymeersch, Kristof Sorgeloos, Koen Beyers, and Tim Dierickx alongside ITM researchers.
-
Red Dot Award 2026
-
Conference presentations: STI & HIV 2023 World Congress (poster); AU-EU Innovation Festival, Cape Town, June 2023 (elevator pitch).
-
VLAIO funding: HBC.2021.0642
-
Anticipated unit cost: under $2 per unit
-
Pre-analytical validation (unpublished): diagnostic accuracy maintained for HIV, syphilis, and hepatitis C serology when samples tested within 7 days and stored below 30°C.
Outcome
-
Collect2Know (C2K) brought from concept to a validated, design-frozen prototype ready for injection moulding.
-
Final design (P4): single component, vertically-aligned fingertip clip, two pricking locations, multiple pricks possible, no assembly required.
-
77.5% of participants successfully collected ≥500 µL of capillary blood in the final clinical iteration.
-
Performance targets met: ≥75% usability (89.1%) and acceptability (75.6%); ≥60% blood volume success (77.5%).
-
Patent filed: WO 2024/089619 A1
-
Published in Expert Review of Medical Devices, Vol. 22, No. 10, 2025.
-
Red Dot Award 2026
-
VLAIO-funded development (HBC.2021.0642)
-
Anticipated manufactured unit cost: under $2 — specifically designed to be viable in low-resource and decentralised healthcare settings globally.
-
Applications confirmed or in development: HIV/STI screening, chronic disease monitoring, seroprevalence studies, potential use in outbreak response and low-resource settings.
What This Proves for Similar Teams
C2K is relevant for research institutions, public health organisations, NGOs, and MedTech innovators who have the scientific or clinical insight to define what a product needs to do — but need an engineering partner who can translate that insight into a manufacturable device that actually works for real users in real conditions.
This case shows how Voxdale runs that kind of project: embedded in a clinical research programme, letting evidence govern design decisions, running fast hardware iterations in parallel with structured participant studies, and holding the manufacturability constraint — unit cost, single component, injection moulding — from the very first prototype.
It also shows that the hardest design problems in accessible medical devices are often not technical in the conventional sense. The breakthrough in C2K was not a new material, a new manufacturing process, or a new electronic system. It was rotating the device 90 degrees and removing three components. Getting to that answer took three prototype generations, 60 participant tests, and the discipline to follow the data.
Navigation