Healthcare Technology Investment: Follow the Constraint and the Cash

A workflow can contain many inefficient tasks while only one limits useful output. I would identify that constraint before comparing AI vendors. If clinician availability limits completed visits, faster intake may increase the waiting list without increasing delivered care. The additional administrative productivity remains real, but its economic consequence is narrower.

The operating case should separate cash released, capacity made available and quality improved. Staff minutes become cash only through a specific staffing or purchasing decision. They become capacity only when the organization can redeploy them to demand it can serve. Neither conversion should be assumed from a time-saving estimate.

For procurement, I would compare the proposed intervention with a process redesign using the existing system. The relevant return is the incremental improvement over that alternative, after review, integration and exception costs. This prevents a technology decision from claiming benefits that simpler operating changes could deliver.

An access intervention can increase a provider’s activity, reduce a patient’s burden and change a payer’s spending in different directions. Those effects need separate accounts. Earlier appropriate care can increase near-term utilization; improved access does not by itself establish medical savings.

I would map the entire receiving pathway: coverage, appointment capacity, practical access, clinical responsibility and payment. A referral-management investment should be funded against a documented failure in that sequence. If the service is unavailable, the next dollar may belong in clinical capacity or a network agreement.

The ACCESS Model and payer participation initiatives identify payment and partnership questions worth investigating. They do not establish the return for an individual organization. That return depends on the covered population, service obligations, payment terms and cost of delivering the contracted result.

Interoperability reduces the cost of moving information; the investment case depends on what happens after receipt. A usable record must arrive in time, match the right person and carry sufficient context for an authorized team to act. Exchange volume measures none of those conditions by itself.

I would assign a business owner to each proposed connection and require an explicit change in work: a redundant assessment removed, an authorization assembled with fewer corrections, or a referral accepted with the necessary information. Technical conformance and operating effectiveness need separate acceptance tests.

The investment sequence matters. Connect the data needed for the highest-value decision first, establish the exception process, then expand the interface portfolio. Broad connectivity without receiving capacity can increase reconciliation costs faster than it improves service.

Value-based contracts introduce a further distinction between social benefit, medical spending and the participant’s cash. Attribution, risk adjustment, quality conditions, settlement timing and retained savings determine whether a clinical improvement finances the organization that produced it.

I would underwrite the service without treating the maximum shared-savings payment as a recurring revenue line. The base case should reflect eligible lives, deliverable capacity and collectible payment; uncertain savings belong in explicit scenarios with a credible comparison and an identified settlement route.

The allocation decision is therefore conditional: fund the pathway when its operating change is feasible and its financing survives conservative assumptions. Where public or patient value exceeds the buyer’s retained return, identify an explicit subsidy or different purchaser instead of concealing the gap in an optimistic ROI.