Visit ten provider organizations running value-based care programs and you will see ten TCOC reports. Nine of them will be useless.
They will be useless not because the underlying arithmetic is wrong — total cost of care is not a hard calculation — but because the report does nothing. It sits in a monthly deck, gets glanced at by the CMO, and never produces a decision. The organization knows its TCOC. It does not know what to do about it.
TCOC is the common currency of VBC. It is the number the contract is written against, the number the board asks about, and the number the medical director is supposedly managing. A TCOC analytic capability that does not translate into operating decisions is not an analytic capability. It is a scorecard.
Over many engagements, we have seen a small set of principles separate useful TCOC analytics from decorative ones.
Principle one: Decompose, don't just report
A headline TCOC number — PMPM or PBPM — tells leadership almost nothing. The same number can conceal very different operating realities. An organization at a $1,050 PMPM with stable inpatient spend and runaway post-acute spend faces a fundamentally different problem than an organization at the same PMPM with runaway specialty pharmacy spend.
A useful TCOC analytic decomposes into at least four layers:
- Category of service — inpatient, outpatient, professional, SNF, home health, pharmacy
- Utilization vs. unit cost vs. mix — the classic decomposition of change
- Cohort — attributed population broken by risk tier, chronic condition, or program enrollment
- Provider — whose patients are consuming what
Only at the intersection of those four views does a number become a decision. "TCOC is up three percent" is not a decision. "TCOC is up three percent because SNF days per thousand rose eight percent in the top-two-risk cohort, driven by three discharging hospitalists" is a decision.
Principle two: Benchmark to action, not to peers
Most TCOC reports benchmark against a regional average or a payer-provided reference. These benchmarks have two problems. First, they are usually lagged — by six to twelve months — which is useless for operating decisions. Second, they tell leadership where they stand, not where they could be.
A more useful benchmark is internal: the top-performing quartile of the organization's own providers, the top-performing care program, the top-performing geography. An internal benchmark is always current, always actionable, and always defensible. If your top-quartile PCPs are running an inpatient PMPM fifteen percent below the bottom quartile, the question is no longer "where do we stand" — it is "what are the top quartile doing that we can spread."
Principle three: Pair every TCOC view with a lever
A finding without a lever is a complaint. A useful TCOC analytic always surfaces alongside the operational lever that moves it.
If SNF days per thousand are high, the levers are discharge planning, SNF network performance, and hospitalist behavior. If specialty pharmacy PMPM is rising, the levers are site-of-service, utilization management, and biosimilar substitution. If inpatient admissions for CHF exacerbation are up, the levers are care management enrollment, medication reconciliation, and access to same-day primary care.
Mature TCOC analytics make these pairings explicit. Every category of spend has a named owner, a named lever, and a target. Less mature analytics produce reports without owners. Nothing moves.
Principle four: Report on cadence that matches the decision
Monthly is too slow for some decisions and too fast for others. High-risk cohort enrollment should be reviewed weekly. Hospitalist discharge patterns should be reviewed monthly. Network strategy should be reviewed quarterly. Benchmark resetting should be reviewed annually. The same TCOC data supports all of these — but a single monthly dashboard collapses them into a rhythm that fits none.
Mature organizations produce tiered TCOC analytics: a weekly view for care management operations, a monthly view for clinical leadership, a quarterly view for executive steering, an annual view for contract negotiation. Each is derived from the same underlying data. Each answers a different question.
Principle five: Model the counterfactual
The most important TCOC question is not "what is our TCOC" but "what would it have been without the interventions we made." Most organizations cannot answer this question. They confuse the number with the counterfactual — and congratulate themselves for trends that would have happened anyway.
Counterfactual modeling — what risk-adjusted spend would have been in the absence of program enrollment, care management, network steering — is the highest-order TCOC analytic. It is what turns a TCOC report into a case for continued investment. It is also what the CFO will ask for, sooner or later, when the program needs to be re-funded.
What good looks like
The best TCOC analytics we have seen in operator settings share four attributes. They are decomposed to the action level — provider, cohort, category, and driver. They are benchmarked internally, not to lagged external averages. They are paired with named levers and named owners. And they are reported on a cadence that matches the decision, not the calendar.
Built that way, TCOC stops being a number on a scorecard and becomes the operating rhythm of a risk-bearing organization. It becomes the reason a hospitalist changes discharge behavior in March and the medical director redesigns the SNF network in April. It becomes, in short, useful.
If you are building or improving your TCOC analytic capability, get in touch. We help clients translate the data they already have into the operating answers they need.