Quick answer
A nursing facility should investigate the cause of its workflow problems before deciding to replace its EHR. Compare a defined improvement to the current system with a tested alternative using the same tasks, staffing assumptions and safety requirements. Replacement becomes a stronger option when material problems remain after feasible improvements, the alternative demonstrates a meaningful benefit, and the facility can support a safe transition.
Start with one expensive failure, not a satisfaction score
A director of nursing says staff cannot find the right information. The business office says charges arrive late. Both complaints may be accurate, but neither identifies the cause. Buying a new platform before tracing the work can carry the same broken handoff into a different screen.
The ONC Health IT Playbook treats selection, implementation and optimization as related activities. The decision method below is an original exercise for a nursing facility considering replacement. It evaluates three options: continue with explicit mitigations, improve the current setup, or replace it. It does not assume a vendor or a purchasing outcome.
Choose a workflow with a named owner and a visible consequence. “Improve usability” is too broad. “An evening admission requires three staff members to locate the current referral packet” can be observed. Include an ordinary case and a difficult case; an option that works only for complete referrals has not solved a missing-document problem.
Separate a missing capability from an unused one
Complete this worksheet using recent, appropriately protected examples. Do not send identifiable resident records to prospective vendors to make the complaint more convincing.
| Observed problem | Possible explanation | Evidence needed before deciding |
|---|---|---|
| Staff enter the same information twice | Missing interface, mapping failure, or local policy | Follow one field from its source through both destinations |
| An order is hard to locate | Search behavior, permissions, or record placement | Test the same task under the actual nursing role |
| A report is inconsistent across sites | Different definitions or different product behavior | Recalculate one metric with a common population |
| A ticket remains unresolved | Incomplete reproduction, unclear ownership, or a product defect | Reproduce the issue and obtain a dated response |
| A requested function is unavailable | Unlicensed module, unsupported workflow, or planned feature | Written scope and a demonstration of the proposed arrangement |
More than one cause can apply. For example, an interface may exist but require a module the facility does not license, while local staff also lack a reliable exception process. Price and test the complete proposed solution. An assurance that “the system can do that” is not an operating result.
Run a bounded improvement trial
Set a review date, a responsible person, the exact change and the evidence that would support continuing. Define safety protections with the clinical lead. An administrative evaluation must not delay required care or replace established clinical escalation.
In a fictional facility, missing referral attachments trigger repeated calls. The team trials a revised intake checklist and a configured shared queue. It records 30 eligible referrals before the change and 30 afterward, separating complete packets from incomplete ones. The coordinator tracks active follow-up minutes, unresolved items at shift change and corrections discovered after arrival.
The first result is ambiguous: time decreases, but the second period contains fewer incomplete referrals. The team compares the incomplete-referral subgroup and extends observation across another weekend. This is more useful than declaring success from the overall average. A small operational trial can identify problems without establishing a general causal effect or a promised return on investment.
Compare the alternative on the same work
Give a prospective replacement the same fictional case, with the same missing information and role restrictions. Record the help its demonstrator provides. A presenter who silently completes a difficult step has transferred work out of the measured workflow.
For each option, document five items: demonstrated result, unresolved dependency, recurring staff work, incremental cash requirement and transition burden. Use “unknown” where evidence is absent. Do not convert missing evidence into a zero-cost assumption. The GAO cost guide supports making scope and assumptions explicit; the decision worksheet is not a government rating system.
A fictional comparison might find that current-system improvements reduce routine work adequately but leave an essential cross-facility workflow unresolved. Replacement can then be evaluated against that remaining problem. It should not receive credit for solving the portion already corrected inexpensively.
Decide what would change the recommendation
Write the decision as a short, revisable record:
- The problem and population affected.
- What the current-system trial demonstrated.
- What an alternative demonstrated under comparable conditions.
- Costs and dependencies still unpriced or untested.
- The decision owner, review date and evidence that would reverse the choice.
An independent facility may choose a smaller improvement because its implementation capacity is constrained. A group may need a shared capability that cannot be reproduced efficiently at every site. Neither result follows automatically from facility count. Both require evidence about the actual organization.
If replacement remains justified, carry the unresolved workflow into the SNF requirements workbook. If optimization wins, retain the measurements and review date. A useful evaluation can conclude that staying is the better decision today.
Questions & answers
Does dissatisfaction with the current EHR justify a replacement project?
It justifies investigation. Record which tasks fail, their consequences and the cause. A configuration problem, unclear ownership and an unavailable product capability require different responses.
What if the current vendor cannot demonstrate the proposed fix?
Keep the fix unverified, request a bounded test and record the dependency. Include the cost of continued workarounds while evaluating alternatives; do not assume either proposal works until it has evidence.
Primary sources
Use the cited source and its applicable version when making a facility-specific decision. These guides explain operational evaluation methods; they are not individualized clinical, coding, or legal advice.