Automation and Human Expertise Equates to Big Gains
Jun 10, 2025
Key Takeaways:
• Problem: Client’s manual process for verifying Medicaid enrollment among ordering providers wasn’t efficient, leading to a rise in denials
• Solution: Wuscott built an automated process to determine every ordering provider’s Medicaid enrollment by searching the NM Medicaid portal and updating the EMR
• Coverage: Over 19,000 NPIs are reviewed monthly with zero disruption to workflow
• Accuracy: QA review showed 99.0% accuracy with only 4 errors in a stratified 385-record sample
• Outcome: Fewer denials, improved compliance, and FTE capacity redirected towards more meaningful work.
Background:
Medicaid is a major federal and state expenditure, with $310 billion in federal outlays in FY2014 alone. That same year, the Centers for Medicare & Medicaid Services (CMS) reported an estimated $17.5 billion in potentially improper Medicaid payments [*].
To address this, CMS implemented stricter screening requirements, mandating that ordering, referring, rendering, and prescribing providers enroll in Medicaid as of 2018. The goal: ensure that only properly licensed, non-excluded providers are involved in Medicaid services.
For ancillary providers, however, fulfilling this requirement is uniquely challenging. While they can register themselves, confirming the enrollment status of each ordering provider is an ongoing burden.
One of Wuscott’s clients faced this exact challenge—with a single FTE manually checking enrollment part-time. As a result, many NPIs went unreviewed, and denials were rising.
Method:
To solve the problem, Wuscott developed a custom bot that automatically verifies the Medicaid enrollment status of ordering providers in New Mexico.
The bot:
• Extracts NPIs from the client EMR’s
• Checks NM Medicaid public enrollment
• Categorizes results into one of five outcomes:
• Enrolled
• Not Enrolled
• MCO Only
• Deceased
• Unknown
The bot then updates the EMR in real-time to assist our client in identifying any provider who cannot bill NM Medicaid—enabling registration and billing teams a chance to intervene before a denial occurs.
To validate accuracy, Wuscott also conducted a statistically valid QA study. From a population of ~19,000 NPIs, we randomly sampled 385 records for manual review—providing 95% confidence with a 5% margin of error.
Result:
Wuscott’s bot now reviews over 19,000 providers every month. While the average check takes 10–12 seconds, real-world batch processing completes in ~15 days due to expected internet hiccups like captchas, site slowness, and service interruptions.
April 2025 | May 2025 | |
|---|---|---|
No. of Providers | 18,977 | 19,315 |
Enrolled | 46.1% | 49.0% |
Not Enrolled | 45.2% | 49.3% |
Needs Checking* | 7.2% | 0.1% |
MCO Only | 1.3% | 1.4% |
Deceased | 0.2% | 0.2% |
Average Time | 12 seconds | 10.5 seconds |
In both April and May 2025, enrollment was nearly a 50/50 split between "Enrolled" and "Not Enrolled" statuses—making accurate classification even more critical. Wuscott grouped 20+ website result variations into the five categories above and manually reviewed a stratified sample.
Count | % of Total | Random Sample Size | Result | Errors | |
|---|---|---|---|---|---|
Not Enrolled | 9,250 | 49.6% | 191 | Pass (N=188) | 2 Enrolled, 1 MCO Only |
Enrolled | 9,420 | 48.8% | 188 | Pass (N=187) | 1 Terminated |
MCO Only | 259 | 1.4% | 5 | Pass (N=5) | - |
Deceased | 45 | 0.2% | 1 | Pass (N=1) | - |
TOTAL | 18,977 | 385 | 99.0% (N=381) |
Conclusion:
Wuscott’s automation delivers a reliable, scalable solution for validating Medicaid enrollment—far exceeding the reach of manual checks.
Our client now benefits from consistent monthly reviews, fewer preventable denials, and freed-up staff enabling them to focus on higher-value work.
If your organization is still relying on spreadsheets or post-denial appeals, it’s time to let automation do the heavy lifting.
*United States Government Accountability Office. Additional Actions to Help Improve Provider and Beneficiary Fraud Controls. May 2015. https://www.gao.gov/assets/gao-15-313.pdf
Key Takeaways:
• Problem: Client’s manual process for verifying Medicaid enrollment among ordering providers wasn’t efficient, leading to a rise in denials
• Solution: Wuscott built an automated process to determine every ordering provider’s Medicaid enrollment by searching the NM Medicaid portal and updating the EMR
• Coverage: Over 19,000 NPIs are reviewed monthly with zero disruption to workflow
• Accuracy: QA review showed 99.0% accuracy with only 4 errors in a stratified 385-record sample
• Outcome: Fewer denials, improved compliance, and FTE capacity redirected towards more meaningful work.
Background:
Medicaid is a major federal and state expenditure, with $310 billion in federal outlays in FY2014 alone. That same year, the Centers for Medicare & Medicaid Services (CMS) reported an estimated $17.5 billion in potentially improper Medicaid payments [*].
To address this, CMS implemented stricter screening requirements, mandating that ordering, referring, rendering, and prescribing providers enroll in Medicaid as of 2018. The goal: ensure that only properly licensed, non-excluded providers are involved in Medicaid services.
For ancillary providers, however, fulfilling this requirement is uniquely challenging. While they can register themselves, confirming the enrollment status of each ordering provider is an ongoing burden.
One of Wuscott’s clients faced this exact challenge—with a single FTE manually checking enrollment part-time. As a result, many NPIs went unreviewed, and denials were rising.
Method:
To solve the problem, Wuscott developed a custom bot that automatically verifies the Medicaid enrollment status of ordering providers in New Mexico.
The bot:
• Extracts NPIs from the client EMR’s
• Checks NM Medicaid public enrollment
• Categorizes results into one of five outcomes:
• Enrolled
• Not Enrolled
• MCO Only
• Deceased
• Unknown
The bot then updates the EMR in real-time to assist our client in identifying any provider who cannot bill NM Medicaid—enabling registration and billing teams a chance to intervene before a denial occurs.
To validate accuracy, Wuscott also conducted a statistically valid QA study. From a population of ~19,000 NPIs, we randomly sampled 385 records for manual review—providing 95% confidence with a 5% margin of error.
Result:
Wuscott’s bot now reviews over 19,000 providers every month. While the average check takes 10–12 seconds, real-world batch processing completes in ~15 days due to expected internet hiccups like captchas, site slowness, and service interruptions.
April 2025 | May 2025 | |
|---|---|---|
No. of Providers | 18,977 | 19,315 |
Enrolled | 46.1% | 49.0% |
Not Enrolled | 45.2% | 49.3% |
Needs Checking* | 7.2% | 0.1% |
MCO Only | 1.3% | 1.4% |
Deceased | 0.2% | 0.2% |
Average Time | 12 seconds | 10.5 seconds |
In both April and May 2025, enrollment was nearly a 50/50 split between "Enrolled" and "Not Enrolled" statuses—making accurate classification even more critical. Wuscott grouped 20+ website result variations into the five categories above and manually reviewed a stratified sample.
Count | % of Total | Random Sample Size | Result | Errors | |
|---|---|---|---|---|---|
Not Enrolled | 9,250 | 49.6% | 191 | Pass (N=188) | 2 Enrolled, 1 MCO Only |
Enrolled | 9,420 | 48.8% | 188 | Pass (N=187) | 1 Terminated |
MCO Only | 259 | 1.4% | 5 | Pass (N=5) | - |
Deceased | 45 | 0.2% | 1 | Pass (N=1) | - |
TOTAL | 18,977 | 385 | 99.0% (N=381) |
Conclusion:
Wuscott’s automation delivers a reliable, scalable solution for validating Medicaid enrollment—far exceeding the reach of manual checks.
Our client now benefits from consistent monthly reviews, fewer preventable denials, and freed-up staff enabling them to focus on higher-value work.
If your organization is still relying on spreadsheets or post-denial appeals, it’s time to let automation do the heavy lifting.
*United States Government Accountability Office. Additional Actions to Help Improve Provider and Beneficiary Fraud Controls. May 2015. https://www.gao.gov/assets/gao-15-313.pdf