The August 2026 issue of The Self-Insurer leads with a feature forecasting medical costs — the shift from analytics that explain what already happened to models that tell plan sponsors where spend is heading and why. TPN.health CEO Trevor Colhoun was featured in the section on behavioral health, making the case that the category most often left out of the forecast is the one quietly inflating claims across the rest of the plan.
The problem: forecasting that stops at the medical claim
Data fragmentation defeats prediction. Clinical, claims, pharmacy and treatment history sit in separate systems, so a claim can tell you a test was ordered but not what it found, or that a member escalated to a second line of therapy but not whether that escalation was clinically warranted. Without that context, cost forecasting stays reactive.
Behavioral health is where the gap is widest. The piece notes that researchers — including a project funded by the Advanced Research Projects Agency for Health (ARPA-H) — are working to build predictive models that link continuous behavioral data from smartphones and wearables to large-scale electronic health records. In the meantime, most plans have no forward-looking view of behavioral health at all, even though the downstream costs land squarely on the medical side of the ledger.
Colhoun put a number on what that blind spot costs:
"Organizations that invest in proactive behavioral healthcare navigation – getting members matched to appropriate care quickly and keeping them engaged – are addressing one of the most overlooked drivers of high-cost claims. Untreated mental health and substance use conditions drive medical costs two to three times higher across the entire system, generating higher ER utilization, more preventable admissions, and worsening comorbidities that compound over time."
TPN.health's approach: accurate matching as a cost-containment strategy
The forecast only matters if there's something to do with it. Colhoun's argument is that the intervention behavioral health actually needs is not another risk score — it's getting the member to the right provider on the first attempt, before the crisis that generates the claim.
The article points to TPN.health's 95.2% match proficiency rate as evidence of what human-guided care navigation achieves when the network data underneath it is accurate. Both halves are load-bearing: a licensed Care Navigator gathers clinical and personal context and routes the member, working from real-time visibility into which providers are active, what they treat, where they're licensed and whether they're accepting new patients.
"Getting people matched to the right behavioral health provider the first time is where the savings begin, and where behavioral health stops being a blind spot and starts driving financial strategy."
What that looks like on the plan
Across the TPN.health network:
Members receive a response to their text within 10 minutes on average, are placed within 1 business day, and have a first visit within 7 days
Time to first appointment improves 92%
Treatment dropout falls 65%
Employers realize 15% savings on claims tied to eliminating unnecessary, historically billed visits
The mechanism is the same one the feature describes for oncology and chronic disease: intervene earlier, and you manage the cost instead of absorbing it. Behavioral health has simply been slower to get the same treatment.
Read the full feature
Read the complete piece in the August 2026 issue of The Self-Insurer.