Is your software working, or are you just a hostage to a contract?

Plenty of operators know, privately, that their software is not working. And they stay anyway, because the thought of the data migration and the exit fees feels scarier than the daily pain. That fear is not always an accident. In some vendor relationships, friction at the exit is a deliberate retention strategy. In others, it is just a badly negotiated contract nobody has revisited in five years. Either way, the fix is the same: stop guessing and start measuring.

Lock-in is engineered, more often than operators assume

The mechanics show up repeatedly in healthcare software contracts: data-export fees that scale with the volume of your own data, weak or absent contractual obligations to hand over a complete extract at termination, and penalties for using competitors' tools on data you generated (EHR Source; Paubox). The highest financial risk in the relationship usually shows up not at go-live but at renewal and exit, precisely when your leverage is lowest and your census depends on the system staying up.

That does not mean every vendor is engineering a trap. It means the burden of proof sits with your organization. If you cannot say, with a number attached, what staying costs you every month, you cannot fairly judge whether leaving is worth it.

Why the operating stakes keep rising

The provider landscape you are competing in has grown and kept shifting. SAMHSA's 2024 N-SUMHSS counted 15,953 substance-use-treatment facilities, up from 14,620 in 2023, 14,854 in 2022, and 14,101 in 2021, out of 21,205 total behavioral health facilities surveyed nationally (SAMHSA, 2024 N-SUMHSS). A larger, more crowded provider base makes scalable intake, authorization, documentation, and billing workflows a competitive requirement, not a nice-to-have.

Meanwhile the workforce you're trying to protect is under real pressure. As of April 2024, roughly 122 million Americans lived in a federally designated Mental Health Professional Shortage Area, and HHS projects the behavioral health workforce could shrink 28% by 2037 even as demand rises 59%, an estimated 45% gap (HHS, National Strategy for the Behavioral Health Workforce). HRSA's longer-range projections put 2038 shortages at roughly 77,050 addiction counselors, 99,780 mental-health counselors, 99,840 psychologists, 43,810 psychiatrists, and 33,840 marriage and family therapists (HRSA, Health Workforce Projections). Every hour a clinician spends fighting duplicate documentation or a manual payer portal is an hour you cannot get back, and you have fewer clinicians to spare than you did five years ago.

Authorization workflows are only getting more structured

CMS's interoperability and prior-authorization rule requires impacted Medicaid, CHIP, and federally facilitated Marketplace plans to deliver prior-authorization decisions through APIs, with decision-time standards phasing in starting 2026, plus public reporting of authorization metrics (CMS-0057-F). CMS also required Medicare Advantage organizations to publish their internal coverage criteria starting in 2024 (GAO, Medicare Advantage).

One payer's 2025 behavioral health dataset shows what good authorization tracking looks like in practice: 105,020 standard requests, with 104,144 approved and 876 denied, roughly a 99% approval rate, and an average decision time of 1 day, 13 hours, and 41 minutes (Connecticut DSS, CMS Prior Authorization Metrics, March 2026). Of 297 appeals, 182 were approved after appeal, or 61%. Numbers like these are payer-specific, but they show the point: if your system cannot reliably record submission time, response time, expiration date, and appeal outcome, you cannot tell whether a denial is a payer problem or a software problem.

What to measure before you accept "switching is too risky"

A credible decision rests on your own baseline, not a vendor's sales deck. Before your next renewal date, track at least:

• Intake conversion: inquiry-to-assessment, assessment-to-admission, time to first response, and lost referrals by source.
• Authorization performance: first-pass approval rate, average turnaround, resubmission rate, denial and appeal rates, and days of care delivered without authorization.
• Revenue-cycle leakage: clean-claim rate, initial denial rate, days in accounts receivable, unbilled days, underpayments, and authorization-related takebacks.
• Labor burden: staff hours on duplicate entry, payer portals, reconciliation, claim correction, and manual reporting.
• Documentation burden: time per note, late or incomplete records, duplicate forms, clinician hours diverted from care.
• Data portability: whether you can export the complete designated record set, demographics, encounters, assessments, treatment plans, medications, authorizations, claims, remittances, attachments, and audit logs, in usable formats.
• Contract exposure: renewal notice deadlines, auto-renewal clauses, termination assistance, export fees, minimum volumes, and price escalators.

Run this for 90 days before you decide anything. Convert the labor hours into fully loaded cost. Quantify revenue leakage separately from staff time. That single number, cost of staying, is what you compare against a migration estimate, and industry figures put a system change in the tens to hundreds of thousands of dollars with months of disruption. Some of that is unavoidable. Some of it is manufactured. The 90-day baseline is how you tell the difference.

The switch is expensive by design, sometimes

A large share of switching cost is friction built into the contract and the data model on purpose, so leaving hurts enough that you keep paying for a tool you'd otherwise replace. But a vendor confident in its product can make leaving easy: month-to-month terms instead of multi-year lock-in, and data that moves in and out cleanly, with a signed reconciliation report so "did everything come over" has a documented answer instead of a prayer.

Before replacing one system with another, define exit criteria in writing: complete data export, reconciliation of open authorizations and claims, validated interfaces, staff training, and a rollback plan. Test the replacement on one location or one workflow first. Compare authorization turnaround, clean claims, documentation completion, and staff time against your baseline before converting everything. And make sure you're not trading one lock-in for another: require documented export methods, role-based access, audit logs, transparent pricing, and the ability to drop a single module without losing your underlying record.

The questions to ask before you renew

Ask your vendor two things. First: if we decided to leave, what would it cost, in dollars and in days, to get a complete, usable copy of our own data out? Second: what happens at renewal, notice period, auto-renew terms, and price escalators? The answers, combined with your own 90-day baseline, tell you whether you bought software or signed a lease on your own information.

If leaving is the expensive part and you cannot say why in specific numbers, you have not evaluated your vendor. You have just been afraid of them.