If you’re running a self-funded health plan at 100–1,500 employees, there’s a good chance the answer to “who do I call when the plan needs to change?” is your TPA. That’s a problem.
A third-party administrator processes claims, manages eligibility, and handles network access on your behalf. They’re operationally indispensable. They’re also not your benefits strategy. The two get conflated constantly — especially in the mid-market, where employers don’t have a benefits department large enough to push back — and the conflation costs real money.
What a TPA actually does
A TPA’s job, narrowly defined, is to operate the plan you’ve already designed. They route claims through your network. They administer your stop-loss policy. They issue ID cards. They generate the monthly utilization report and (if you’re lucky) a quarterly summary of where your spend went. Their margin model rewards efficiency — high claim throughput, low administrative cost per member — and that’s appropriate for what they do.
What they don’t do, structurally:
- Tell you the plan you’ve designed is wrong. Their incentive is to keep operating the existing plan, not to redesign it. A TPA that recommends you replace 30% of your medical spend with a flat-PMPM onsite clinic is a TPA that just shrunk its own claims-administration revenue.
- Model alternative plan structures against your data. They have your data — every claim, every diagnosis, every dollar — but their reporting tools don’t generate “what if you replaced fee-for-service primary care with onsite primary care” scenarios. That isn’t a feature they sell.
- Negotiate against the carriers and PBMs you also use. Most mid-market TPAs are owned by, partnered with, or financially aligned with the carriers and PBMs that source the network and the pharmacy benefit. Asking your TPA to fight your PBM is asking it to fight its own house.
None of this is venal. It’s the structural shape of what a TPA is. The mistake isn’t in choosing a TPA. The mistake is in expecting your TPA to also be the strategist — the one who decides what your plan should look like five years from now.
Why mid-market employers default to this
Two reasons, both structural.
First, the broker channel mid-market employers rely on is itself often compensated by the carriers and TPAs they recommend. Not always — there are excellent fee-only and consultant-style brokers serving the mid-market — but the average mid-market broker is paid via commission baked into the medical premium or per-member-per-month admin fee. That commission flows from the same vendor whose plan the broker is recommending. The broker has every reason to recommend a plan structure their carrier and TPA partners can administer. They have less reason to recommend a plan structure those partners can’t.
Second, the mid-market doesn’t have the internal benefits muscle to push back. A 5,000-employee enterprise has a dedicated benefits team with a vendor manager, an actuary on retainer, and a procurement function. They can interrogate the TPA’s reporting and demand alternative scenarios. A 250-employee company has one HR generalist who’s also running payroll, recruiting, and onboarding. They take the TPA’s quarterly report at face value because they don’t have time to do otherwise.
Both reasons are explainable. Neither is changeable from where you sit. What is changeable is whether you treat the TPA’s quarterly report as the strategy or as the input to the strategy.
What a strategy layer actually looks like
A real benefits strategy layer asks questions the TPA can’t answer:
- Where in the spend curve is the leverage? Most mid-market plans have a long tail of small claims and a short head of catastrophic claims. The middle — the chronic-disease management spend, the urgent-care and ER diversion spend, the pharmacy spend that isn’t catastrophic but isn’t small either — is where 60% of total cost lives and where 90% of cost reduction can happen. A TPA report shows you the spend; a strategy layer tells you what to do about it.
- What plan structure would change the curve? Replacing fee-for-service primary care with onsite primary care, restructuring the PBM relationship, narrowing the network, adding stop-loss protection, layering in direct contracting with specific specialty providers — these are plan-design moves the TPA implements but does not recommend.
- What does the math look like against your specific population? “18% average savings” is a useful benchmark. The right number for your plan depends on your member demographics, your claims pattern, your geography, and what’s already working.
The honest answer to “who provides the strategy layer” varies by employer. Sometimes it’s a fee-only broker. Sometimes it’s an outside benefits consultant on retainer. Sometimes it’s a direct primary care or onsite clinic vendor that brings the analysis as part of the engagement. What it’s not, structurally, is the TPA.
What to do about it this renewal cycle
Three concrete moves:
-
Ask your TPA to send you the trailing 12 months of detailed claims data — not the summary report. Member-level detail, ICD codes, CPT codes, dollar amounts, place of service. They have it. They’ll tell you it’s hard to extract; it isn’t. If you can’t get the raw data, you can’t do strategy on top of it.
-
Pull the PMPY (per-member-per-year) cost in a few cuts: by service category, by chronic-disease cohort, by ER-or-urgent-care utilization. This is what tells you where the leverage is. If 65% of your spend is going to fee-for-service primary care + urgent care + ER, that’s the reducible portion. If 80% is going to specialty and hospital admits, the reducible portion is smaller and the levers are different.
-
Bring in someone whose job is plan design, not plan administration, and have them model alternatives. If you don’t have a benefits consultant, ask a vendor like Archer to run the model — the analysis is free and we’ll tell you straight whether onsite primary care is a fit. If it’s not, you’ve still got the analysis, which is more strategic input than your TPA has handed you in a year.
Your TPA isn’t the bad guy. They’re the operations layer. Treat them like the operations layer. Find the strategy layer separately.
That’s the real plan-design move available to mid-market self-funded employers right now, and almost nobody is making it.