Most cost overruns in medical device trials are not budget surprises. They are costs that were preventable. Two cautionary tales, one structural diagnosis of why CRO quotes never tell you what a trial really costs, and one combination of software plus expert judgment that prevents all of it.
Global MedTech CROOne synopsis on the left. Three forecasts on the right, stacked in parallel: cost on top, recruitment timeline in the middle, site network at the bottom. Every slider you move on the left reflows all three forecasts at once, so you see immediately how a sample size change ripples into recruitment time, site count and final budget. You will see the same screen during the walkthrough, with your real protocol on it.
Milo TF runs three parallel forecasts on every protocol you bring us. Costs, site partnering, recruitment. Each forecast contains three scenarios. Swift, base, conservative. Nine scenario outputs in total, anchored to seven years of clinical trial delivery and the Notified Body costs precedents our experts defend them against. The numbers move as you adjust the inputs, on screen, with you in the room.
P75 site recruitment rates, smaller site network, lighter monitoring. The fastest defendable path through MDR Article 62 or 74. Useful when the budget is tight and the methodology can survive scrutiny on minimum viable design.
Median historical site rates, standard MDR pivotal or PMCF footprint. The scenario most sponsors actually live in, with the contingency room a CFO can defend to the board. Anchored to the protocol design our experts approve.
P25 site rates, maximum monitoring intensity, worst case IRB lag, slowest site activation, longest follow-up. The number you bring to the board so nobody can say later "we didn't see this coming". The contingency envelope that protects the programme.
The setup. A mid-size cardiovascular sponsor obtained the CE mark for a coronary device under EU MDR. As part of the regulatory dossier, the clinical and regulatory teams submitted a Post Market Clinical Follow-up Plan to the Notified Body, a binding commitment to collect real-world evidence after CE mark. The plan specified an RCT with a sizable patient cohort, several years of follow-up, with defined endpoints and a defined site network. It was elegant. It was defensible. The Notified Body signed off.
The envelope. The internal financial envelope had been put together by a small team using rules of thumb and prior project memory. Nobody had actually costed the plan.
The first CRO quote. Months later, when procurement went out to CROs to execute the PMCF, the first quote came back well above the envelope, for the CRO scope alone. Site grants were quoted as a separate line, "to be negotiated directly with the sites". By the time the site contracts were signed and the pass-through expenses surfaced, the real number was several times the original envelope.
And then the cascade. The original site network was undersized for the recruitment target. The defined sites could not enrol fast enough to hit the timeline committed in the PMCF Plan. Additional sites had to be activated late, at premium activation fees, in countries with higher per-patient grants the sponsor had not budgeted for. Cost, sites and recruitment all moved against the sponsor at the same time.
The regulatory trap. The sponsor considered going back to the Notified Body to propose a leaner plan. The regulatory consultants advised against it. Reopening the post market commitment after CE mark would have triggered a fresh review of the supporting evidence. The cheapest path forward was the most expensive plan.
The cost of not forecasting. The programme ultimately ran, but it consumed the budget reserved for a second indication. That second device launched late, and a competitor took the market window.
A PMCF Plan submitted to a Notified Body is a financial commitment, an operational commitment, and a recruitment commitment, all at once. Cost, sites and recruitment have to be forecast together, before submission. Forecasting one and assuming the other two will follow is how trials end up several times over budget.
The setup. A wound care company launched a clinical trial in diabetic foot ulcer healing. The protocol assumed twenty four months of enrolment across twelve sites. The principal investigators were enthusiastic. Two of them had run similar studies before. They told the sponsor confidently that hitting the target was realistic.
It took ten years.
What actually happened. Site activation took longer than expected. The eligible patient population was narrower than the PIs had estimated, because half of the candidates had comorbidities that excluded them. Three sites under-enrolled and were closed. Two replacement sites were added late. Recruitment stalled for a full year and a half during the pandemic.
The cost of optimism. The programme cost the sponsor several times the planned budget. The launch was missed. A competitor with a less effective device but a more realistic operational plan captured the market segment. By the time the trial published, the sponsor's commercial window had closed.
The forecast that would have saved them. It is not complicated. It uses historical recruitment rates from comparable trials in the same indication, applied across the actual site network proposed, with population multipliers for the eligibility criteria. The result is three scenarios. Swift at the P75 site rate. Base at the median. Conservative at the P25.
The decision that never happened. If they had run that forecast at the design stage, the pessimist scenario would have produced a recruitment estimate of seven to nine years. The board would have demanded a redesign before launch. The trial that actually happened would never have started.
PI optimism is not a character flaw, it is the default. Site recruitment forecasting has to be anchored to historical data, not to clinic estimates. Three scenarios, with the pessimist scenario taken seriously, would have changed every decision this sponsor made.
The two stories above are not random misfortunes. They are the predictable outcome of how CRO budgeting and forecasting actually work today. Site grants are not in the quote. Hourly rates are hidden. Recruitment timelines come from PIs, not historical data. You spend four to eight weeks reshaping scenarios just to find one that fits your envelope. Here is the contrast.
The two sponsors did not lose money because they were careless. They lost money because the forecast they needed did not exist as a fast, integrated tool. And the CRO market today still operates with the opacity described above. Milo TF is the software half. our Expert team are the human half. The combination is what makes the promise real, fifteen minutes for a live booth walkthrough, under thirty minutes for the full async forecast.
You drop a PMCF Plan or synopsis. The engine parses the design, applies our FMV anchored European rate card built from seven years of trial delivery, and runs three parallel forecasts. Cost, sites, patient recruitment. Each with three scenarios. All in under thirty minutes.
Software gives you a number. Experts tell you whether the number is defendable. Expert 1 judges the trial design against what a Notified Body will actually approve. Expert 2 judges whether the site network and recruitment plan will hold up under real operational conditions.
No account managers, no junior staff. The two people who can answer the design and operational questions that decide whether your forecast is defendable are at the booth with the software open.
We are at EuroPCR all four days. Bring a synopsis, a PMCF plan, or just a one-pager concept. We will run all three forecasts in front of you and our experts will defend or challenge the result in real time. No booking required, walk-ins welcome.
