Differences

This shows you the differences between two versions of the page.

--- guides:sharpinsight [2026/05/12 18:42] – Move from private:guides:sharpinsight (going public) brad.sharp
+++ guides:sharpinsight [2026/09/17 14:41] (current) – Add Preventable Revenue Loss report (Coding & Compliance) and the Patient Responsibility Codes setting brad.sharp
@@ Line 226: / Line 226: @@
   * Payers with win rates above 50% are strong candidates for consistent appeal efforts.
   * If appeal rates are low but win rates are high, you may be leaving recoverable revenue on the table.
+==== AI Rule Suggestions ====
+Groups your denials into patterns by reason code, payer, and procedure code, then lets you ask AI to analyze any pattern and suggest how to resolve and prevent it. For each pattern the AI explains what is likely happening, lists possible causes and recommended actions, and where appropriate recommends an automated rule to stop the denial from recurring.
+**Key Metrics:**
+^ Metric ^ Description ^
+| Denial Patterns Found | The number of distinct reason-code + payer + procedure denial patterns in the selected period. |
+| Total Denial Amount | The combined dollar amount of all denials analyzed. |
+| Total Service Lines | The number of denied service lines behind the patterns. |
+| Denial Code | The CARC group and reason code (e.g., CO-197) shown with its description. |
+| Top CPT Codes | The procedure codes most frequently denied within each pattern. |
+**Tips:**
+  * Click **Analyze** on a pattern to generate AI suggestions. These are a starting point — confirm against the claim detail or with the payer before acting.
+  * When the AI recommends an automated rule, it shows a confidence score and pre-fills values from the denial data so you can prevent the pattern going forward.
+  * Start with the highest Total Amount patterns for the biggest recovery and prevention impact.
 ----
@@ Line 285: / Line 302: @@
 ===== Coding & Compliance =====
-Reports for tracking procedure code changes, CARC trends, profitability, and contracted rates.
+Reports for tracking procedure code changes, CARC trends, preventable patient balances, profitability, and contracted rates.
 ==== Procedure Code Changes ====
@@ Line 314: / Line 331: @@
   * Watch for CARC codes that are trending upward month-over-month — they may indicate new payer rules or process breakdowns.
   * Cross-reference with the **Denial Root Cause** report for a complete picture of adjustment impacts.
+==== Preventable Revenue Loss ====
+//New in September 2026.//
+Finds patient balances that are worth a second look before anyone collects them, groups them by the **process failure that produced them**, and shows whether each of those processes is getting better or worse month over month.
+Some of what it finds is still recoverable from the payer. The rest is money to stop losing next month. That is the point of the report: a single total is an invoice, but a category trending the wrong way is a reason to change something.
+<WRAP center round info 90%>
+**Legitimate cost sharing is never flagged.** Deductible (PR-1), coinsurance (PR-2) and copay (PR-3) are shown alongside the findings for scale only. None of it is a problem.
+</WRAP>
+=== Running the report ===
+Choose a **Start Date** and **End Date**, optionally enter a **Payer Number**, and click **Search**. When a payer filter is applied a banner appears at the top reminding you that every figure below is for that payer alone, with a **Show all payers** button to clear it.
+=== The headline numbers ===
+^ Metric ^ Description ^
+| **Preventable** | The slice your own process could have stopped: missed authorizations, late filing, claim data errors, duplicates, unverified eligibility. This is the number to manage against. |
+| **Still Recoverable** | Balances that can still be appealed or corrected and resubmitted, before the window closes. |
+| **Already Paid** | The payer settled this on an earlier remittance. Not a loss. Only shown when there is some. |
+| **Payer Caused** | Not your process. Escalate these with the payer. Only shown when there is some. |
+| **Total Flagged** | Everything the report flagged, and the number of claims behind it. Includes amounts that are not losses, so this is the widest number, not the actionable one. |
+| **Legitimate Cost Sharing** | Deductible, coinsurance and copay for the same period. Context, not a finding. |
+=== "What was checked" ===
+Directly under the headline numbers is a card showing how many ERAs were scanned, how many carried patient responsibility, and which reason codes were applied.
+**Read this before concluding you have no problem.** An empty report has three very different causes, and this card tells you which one you are looking at:
+  * No ERAs were received in the date range, so there was nothing to check
+  * No codes are configured for your account, so nothing could be flagged
+  * There genuinely were no findings
+The card also lists anything that could **not** be evaluated, so you know what is missing from the totals, and warns you if the detail list at the bottom was shortened (the totals always cover everything).
+=== What to fix, and whether it is improving ===
+Each finding is assigned a **prevention category**: the process that failed. Every category comes with the change that stops the next one, plus a month-by-month bar strip so you can see the direction of travel.
+^ Category ^ What to change ^
+| Authorization not obtained | Verify and record authorization before the service is rendered. |
+| Filed too late | Shorten the gap between service and submission, and watch the oldest unbilled claims. |
+| Claim data errors | Tighten claim scrubbing at entry so these are caught before submission. |
+| Coding | Review coding and documentation for these services with the coding team. |
+| Duplicate submission | Check submission controls. These claims went out more than once. |
+| Eligibility not verified | Run eligibility before billing so protected patients are flagged up front. |
+| Payer error (not yours) | Not a process failure on your side. Escalate these with the payer. |
+| Already paid (not a loss) | Confirm the earlier payment posted before deciding how to handle the balance. |
+=== Actions ===
+Every flagged balance carries an **Action** badge telling the biller what to look at next.
+^ Action ^ Meaning ^
+| **Appeal** | Worth appealing to the payer. |
+| **Correct & resubmit** | Fix the claim and send it again. |
+| **Write off** | Not collectable and not recoverable. |
+| **Review** | Needs a human look before it is categorised. |
+| **Verify** | Stop and confirm the situation before deciding how to handle the balance. |
+A smaller note beside the action says **why** the row was flagged: //Not patient liability//, //Medicaid cost share//, or //QMB protected patient//.
+=== QMB patients needing review ===
+Where eligibility reports a patient as a Qualified Medicare Beneficiary but the check could not be confirmed for the service dates, those patients are **listed individually** in their own section with the member ID, payer, reported coverage status, why it needs review, when the check ran, and the coverage period.
+These rows are **excluded from every total above**. Confirm coverage before deciding how to handle them.
+<WRAP center round important 90%>
+QMB findings only appear for accounts that run eligibility through Sharp Revenue. If the "What was checked" card says QMB was not evaluated, that is the reason. See the [[guides:patient-coverage-on-claims|Patient Coverage on Claims]] guide.
+</WRAP>
+=== The tables ===
+  * **By Reason Code** — each code with its meaning, claim and adjustment counts, action, prevention category, and total
+  * **By Payer** — grouped by the payer number on the remittance, with recoverable and total amounts, and a **Filter** button to drill the whole report into one payer. A payer that sends more than one payer number appears on more than one row.
+  * **Affected Claims** — the individual balances: patient account number, patient, payer, code, action, amount, and ERA received date. Sortable and paged.
+**Tips:**
+  * **Watch the trend per category, not the single total.** A category going the wrong way is the signal worth acting on.
+  * **Start with Preventable, not Total Flagged.** Total Flagged deliberately includes amounts that are not losses.
+  * **Configure your codes.** The code list and the action assigned to each one are per account, under **Analytics Settings**. Adjust them if a contract genuinely permits patient responsibility where the standard discourages it.
+  * **Totals will not tie out against CARC Trends.** That report reads service-line adjustments only, while this one reads both claim-level and line-level amounts.
 ==== Procedure Profitability ====
@@ Line 342: / Line 446: @@
   * Keep contracted rates up to date to ensure accurate underpayment detection.
   * If a payer uses a Medicare multiplier, you only need one entry rather than individual procedure rates.
+==== Diagnosis Ordering Impact ====
+Finds cases where the same payer reimburses the same CPT code differently depending on the order the diagnosis codes are listed. When claims share a payer, a CPT code, and the same set of diagnosis codes but list them in a different order, the payer may pay them differently. This report surfaces those groups and shows which ordering pays best.
+**Key Metrics:**
+^ Metric ^ Description ^
+| Groups with Impact | The number of payer + CPT + diagnosis-set groups where the ordering changed the payment. |
+| Max Payment Variance | The largest average payment difference between the best and worst ordering across all groups. |
+| Estimated Recoverable | The total dollar opportunity if every claim had used the best ordering. |
+| Best / Worst Avg Pay | The average payment for the highest- and lowest-paying ordering in a group. |
+| Variance % | The percentage payment difference between the best and worst ordering. |
+**Tips:**
+  * Expand any group to compare orderings side by side. The row marked **Best** is the ordering to train billers to use for that payer + CPT combination — list that primary diagnosis first.
+  * Use **Min Claims per Ordering** to require more claims before an ordering appears (higher = more reliable), and **Min Variance %** to hide differences too small to act on.
+  * Click an ordering to drill into the underlying claims, then open the claim editor or ERA directly.
+  * Use **Create Rule** to turn a best-ordering finding into an automated rule.
 ----
@@ Line 390: / Line 512: @@
   * Use this to research appeal strategy — find similar claims that were successfully paid after initial denial.
   * High similarity scores (90%+) provide the most reliable comparison data.
+==== Worked & Resubmitted ====
+Shows claims that were submitted or resubmitted within a date window and whether that work led to payment, denial, or rejection. Every revision of a claim is grouped into a single "claim family" so you can see the full submit-and-rework path and its final outcome. A direct way to measure whether follow-up effort is actually converting to paid claims.
+**Key Metrics:**
+^ Metric ^ Description ^
+| Claims Worked | The number of distinct claims submitted or resubmitted in the selected window. |
+| Resubmissions | The number of send events that were resubmissions of a claim sent previously. |
+| Paid / Denied / Rejected | The count and percentage of worked claims that ended in each outcome. |
+| Awaiting | Worked claims that have not yet reached a final outcome. |
+| Dollars Paid | Total payments received on claims worked in the window. |
+**Tips:**
+  * Use the By Payer table to see which payers reward resubmission effort and which keep denying.
+  * Click any claim family to expand its full revision path — received, sent, status, payer acceptance, advice, and outcome for each revision.
+  * A high resubmission count with a low paid percentage points to rework that is not fixing the underlying denial reason — cross-reference the **Denial Root Cause** report.
+  * Export to CSV to build follow-up worklists.
 ----
@@ Line 492: / Line 632: @@
   * **Similarity settings** — Set the minimum similarity score and maximum results for the Claim Similarity report
   * **CARC exclusions** — Exclude specific CARC codes from trending and analysis
+  * **Patient responsibility codes & actions** — The reason codes that the **Preventable Revenue Loss** report flags, and the action assigned to each one. Legitimate cost sharing (PR-1 deductible, PR-2 coinsurance, PR-3 copay) is never flagged. Use **Restore Defaults** to load the standard set, then add, remove, or re-assign actions to match your contracts. With no codes configured, the report flags nothing.
   * **Denial settings** — Configure write-off threshold days and appeal detection behavior
 ----
-//Need help? Contact ClaimRev support at [[mailto:support@claimrev.com|support@claimrev.com]] or call 918-942-9564.//
+//Need help? Contact ClaimRev support at [[mailto:help@claimrev.com|help&#64;claimrev.com]] or call 918-842-9564.//