Diagnostics as a Service migration, or how to change the laboratory behind your product

Most brands that ask us about Diagnostics as a Service already have a laboratory partner, and a fair number of them do not know which laboratory it is. A woman who runs a health brand told me that she did not know which laboratory analysed her customers' blood. She had a contract and she had an account manager, but she had never asked what was behind either of them.

She found out in the end by reading her own privacy policy, because somebody on her team had listed the sub-processors properly years earlier, and the answer had been published on her own website ever since.

She is neither unusual nor careless, because when testing works nobody looks underneath it. People start looking when something goes wrong, which means they are learning how the whole arrangement fits together at the same moment they are trying to repair it.

This article is written for the version where you look first, and it follows the order in which the work actually has to happen. You establish what you already have, then decide why you are moving, then work out what will go wrong, and only after that do you move.

First, establish what your current diagnostics arrangement is

Before you speak to anybody new, write down your current arrangement, because most teams cannot answer the following from memory.

• Which laboratory performs each analysis by name, and which accreditation that laboratory holds.
• Who takes the samples, and under what legal basis in each country you operate.
• Which sub-processors handle customer data, and which country each of them is in.
• What your contract says about notice, and what it says about your data on exit.
• Whether you can export your customers' historical results in a usable format, and who owns them.

Your own privacy policy and sub-processor list will answer several of those questions, because European data protection law requires processors to be documented and requires you to authorise sub-processors.¹ If your own published documents cannot answer them, that is the first problem to repair, and it needs repairing whether or not you change partners.

Second, be specific about why you are moving

You cannot negotiate well until you know which problem you are solving, because each of the reasons below leads to a different set of commitments.

Samples fail too often, which means customers are asked to give blood again more than they should be. This is the most common reason for moving and the most damaging to your brand, and it is worth knowing that some failure is unavoidable. A meta-analysis of 26 studies covering more than 16 million requests found a pooled rejection rate of about two in a hundred, with clotting, breakdown of the sample and insufficient volume causing most of it.² So the question is not whether your provider has failures, but whether yours are above what the work requires and who pays for them.

Results take longer than your product promised, which means your support team spends its day absorbing the difference between the two.

Your panel cannot grow, because you want to add markers or modalities that your current setup cannot run, and your roadmap is limited by your supplier rather than by your customers.

You are entering a market your current partner does not serve. Before you assume that a new partner solves that on its own, remember that the rules change at every European border.

You are too small to matter to them, which shows itself as slow answers rather than as broken service, and it is the reason people give last and feel first.

Third, do not choose a DaaS partner on price alone

With the reason established, the temptation is to line up quotes and compare them. Price is the easiest thing to compare and the weakest basis for a move, because any saving disappears the first time a batch of samples fails or a result comes back late enough that support has to explain it.

The things that really determine whether a partner is better are harder to compare, and they are worth the extra effort.

• What proportion of samples produce a usable result, and what happens when one does not.
• Who carries the cost of a repeat test when a sample has failed.
• What is committed in writing, as against what is described on a website.
• How quickly somebody who can actually fix a problem answers you.
• Whether adding a marker, a panel or a country is a change request or a new project.

Ask for all of that in writing, because a partner willing to put a number into a contract is telling you something quite different from a partner publishing the same number in marketing. Our procurement guide goes through the full question set in order.

Fourth, understand what actually goes wrong when you move

This is the part most teams discover afterwards, and it is the reason this article exists at all.

Laboratories do not use the same reference intervals for the same marker. A 2023 review in Clinical Chemistry states it directly, saying that significant and unwarranted variation in reference intervals continues to exist across laboratories, including between laboratories using the same analytical method. The published consequence is an increased risk of misinterpreting results.³

So when you change laboratory, a customer's value can appear to move even though nothing about their health has changed.

If you sell a single test then that is a footnote in your documentation, but if you sell repeat testing or tracking or anything a customer reads as a trend, then it is your product. The customers affected worst are the ones with the longest history with you, and those are the customers you can least afford to confuse.

There are three things to settle before you sign anything.

• Whether the new laboratory's intervals differ from the old ones, marker by marker, across the markers you actually sell.
• How you will explain any change to customers who already have a history with you, because saying nothing damages trust more than explaining it does.
• Whether your new partner curates its own intervals across laboratories, since that is what keeps a customer's history comparable when the analysis moves.

Fifth, run both partners at the same time

Do not move everybody at once, and run a parallel pilot instead, because that is the only way to see the answers to the questions above rather than being told them.

1. Pick one panel to test with, and pick the one you sell most of.
2. Send a defined set of samples through the new partner while the old one keeps serving everybody else.
3. Compare what comes back, which means the usable result rate, the time to result, how the report reads and how values compare against the intervals you have been using.
4. Put a deliberately awkward case through it, such as a sample that arrives late, a result badly out of range, or a customer who wants to speak to a person.
5. Only then agree a date for moving everybody else across.

Step four is the important one, because every provider handles the easy path well. What you are really buying is how they handle a bad day, and the only way to see that is to create one on purpose while you still have an alternative running.

Sixth, negotiate your exit on the way in

The best moment to agree how you leave is before you arrive, and asking early tells you a great deal about who you are dealing with.

• Whether you can export your customers' full result history, in a documented format, whenever you want it.
• What notice is required from each side, and whether anything renews automatically.
• What happens to samples in transit and results not yet delivered on the day you leave.
• Whether sub-processors are listed, and whether you are notified before any of them change.
• Whether the data agreement is a standard one your own lawyer can read today.

A partner who answers all of that plainly is not worried about you leaving, which is the answer you want. It is also the answer that makes your next move easier than this one.

Last, the move itself is mostly communication

The technical work of moving is real but bounded, and what usually goes wrong is what customers are told about it.

• Tell your customers before they notice anything, explaining what changed behind the scenes, what stays the same and what may look different.
• Brief your support team on the real reason a value might read differently, so nobody is guessing on a live chat.
• Do not run the move during your busiest month of the year.
• Keep the old arrangement available for rather longer than you think you need it.

None of those four things is difficult on its own, but all of them are skipped when a migration is running late, which is exactly the moment they matter most.

How Aniva handles a Diagnostics as a Service migration

Most companies we speak to already have a laboratory partner, which is the normal situation, and we treat it as the starting point rather than as a complication.

• One contract covers the laboratory, the collection and the logistics, together with the software and the compliance paperwork, so a move consolidates rather than adds.
• Aniva owns its blood draw network across Germany, so collection is ours to answer for and not a subcontractor's.
• We curate reference intervals alongside the laboratory-reported ones, which keeps a customer's history readable when the analysis moves.
• Results come back in one format across markets, so a second country is a configuration rather than a rebuild.
• Diagnostics as a Service, or DaaS, is available to partners across Europe, and the result format is the same in every market.
• A parallel pilot on a single panel is a normal way to begin, and we would rather you ran one.

Bring the panel you sell most of and the thing your current setup does badly. Book a 30-minute demo, or send your engineers to the API documentation to see the result format before anybody discusses commercials. If you are weighing this against bringing the whole operation in-house, we have set out what that involves separately.

A summary of the main points

• Start by writing down what you already have, because your own privacy policy will tell you more than you expect.
• Be specific about why you are moving, because the reason changes what you should ask for.
• Do not move on price alone, and move instead on the failure rate, on what is committed in writing and on how fast you get an answer.
• Reference intervals are not harmonised between laboratories, so a customer's value can move when their health has not, and that has to be settled before you sign.
• Run both partners in parallel on one panel, and put a deliberately difficult case through the new one.
• Agree your exit terms on the way in, because the answer tells you who you are dealing with.

Questions we are asked most often

How do I find out which laboratory is behind my current provider?

Read your own privacy policy and sub-processor list first, because European data protection law requires those to be documented and the answer is often published there already.¹ If it is not there, ask your provider in writing which laboratory performs each analysis and which accreditation it holds. A provider who will not answer that question has told you something useful.

What is the biggest risk when changing laboratory?

The biggest risk is to your customers' result history, because reference intervals are not harmonised between laboratories, even between laboratories using the same analytical method, and the published consequence is an increased risk of misinterpreting a result.³ If you sell repeat testing, agree how intervals and comparability are handled before you sign anything.

Should I change partner to save money?

Price is the weakest of all the reasons to move, because a saving disappears the first time a batch of samples fails or results come back late enough that support has to explain them. Compare the usable result rate, who pays for repeat tests, what is committed in writing, and how quickly you reach somebody who can act.

How much sample failure is normal?

A meta-analysis of 26 studies covering more than 16 million requests found a pooled rejection rate of about two in a hundred, with regional subgroups from 0.55 percent to 2.82 percent.² Some failure is unavoidable, so the real question is whether yours is above what the work requires and who bears the cost of it.

How should I test a new laboratory partner before committing?

Run a parallel pilot on your highest volume panel while your current partner keeps serving everybody else, and compare the usable result rate, the time to result and how the report reads. Then put a deliberately awkward case through it, because how a provider handles a bad day is the thing you are actually buying.

What should the contract say about leaving?

It should give you full export of your customers' result history in a documented format, clear notice terms with no automatic renewal, and a stated position on samples in transit on your last day. It should also list the sub-processors with notification before any of them change, and the data agreement should be one your lawyer can read before you sign.

Notes and sources

1. Regulation (EU) 2016/679, the General Data Protection Regulation, Article 28 on processors and the authorisation and documentation of sub-processors. EUR-Lex
2. Getawa S, Aynalem M, Melku M, Adane T. Blood specimen rejection rate in clinical laboratory: a systematic review and meta-analysis. Practical Laboratory Medicine 2022;33:e00303. Pooled rejection rate 1.99 percent across 26 studies and 16,118,499 requests. Leading causes: clotted specimen 32.2 percent, haemolysis 22.9 percent, insufficient volume 22.8 percent, labelling errors 7.3 percent. Regional subgroups from 0.55 percent in the Americas to 2.82 percent in Asia. doi:10.1016/j.plabm.2022.e00303
3. Bohn MK, Bailey D, Balion C, Cembrowski G, Collier C, De Guire V, Higgins V, Jung B, Mohammed Ali Z, Seccombe D, Taher J, Tsui AKY, Venner A, Adeli K. Reference interval harmonization: harnessing the power of big data analytics to derive common reference intervals across populations and testing platforms. Clinical Chemistry 2023;69(9):991 to 1008. doi:10.1093/clinchem/hvad099
4. Lin Y, Spies NC, Zohner K, McCoy D, Zaydman MA, Farnsworth CW. Pre-analytical phase errors constitute the vast majority of errors in clinical laboratory testing. Clinical Chemistry and Laboratory Medicine 2025;63(9):1709 to 1715. doi:10.1515/cclm-2025-0190

This article is general information for companies reviewing their diagnostics supplier. It is not legal advice. Laboratory accreditation and reference intervals belong to the analysing laboratory, and interpretation of any individual result remains with a treating clinician.