A denial specialist opens the queue on a Tuesday morning. Forty-one denials arrived overnight. She reads the first reason code, CO-197, precertification absent. She finds the authorization number in a scheduling note, adds it to the claim, resubmits, and moves to the next one. By lunch the queue is at twelve. That is follow-up, it is skilled work, and it is essential.
It is also endless. Nothing about fixing those forty-one claims prevents forty-one more from arriving next Tuesday, because the reason the authorization number was in a scheduling note instead of on the claim has not changed. Root-cause analysis is the work of finding that reason and changing it. Most practices we meet do follow-up well and root-cause analysis never, not because they disagree with it but because it is nobody's job.
This article explains the difference between the two, why they have to live in the same team, and how to run root-cause analysis with a spreadsheet and one hour a month.
Key takeaways
- Follow-up answers "what do I do with this claim?" and is measured in recovery rate and days to resolution. Root-cause analysis answers "where was this denial created?" and is measured in denials that stop arriving.
- Every denial was created at a step in the practice: scheduling, registration, eligibility, authorization, documentation, coding, claim setup, or the payer's own policy. Assign each category to a step and an owner.
- Eight denial families cover almost everything. Rank them by dollars, not by count; five denied surgeries matter more than fifty eligibility denials.
- The people doing follow-up see patterns first. When follow-up and root-cause analysis are split across teams, the pattern becomes a quarterly surprise.
Follow-up answers "what do I do with this claim?"
Correct the ID, add the modifier, attach the notes, resubmit or appeal before the deadline. Good follow-up is measured in recovery rate, days to resolution, and the share of denials resolved before the appeal window closes. It requires a queue with deadlines, because payer appeal windows range from 30 to 180 days and a denial that misses its window becomes a write-off no matter how good the appeal would have been.
The trap is that follow-up feels productive. The queue goes down, money comes in, and the monthly report shows dollars recovered. What the report does not show is that the practice paid for the same denial twice: once in the rework and once in the delay. A claim denied and corrected is paid 30 to 60 days later than a claim submitted clean, and the staff time to fix it was not free.
Root-cause analysis answers "where was this denial created?"
Every denial was created at a step in the practice: scheduling, registration, eligibility, authorization, documentation, coding, claim setup, or the payer's own policy. Root-cause analysis assigns each denial category to a step and an owner, then asks what changed there.
An example. A practice sees forty denials in a month for "service not covered under the patient's plan" from one commercial payer. Follow-up bills the patients or writes them off. Root-cause analysis notices that all forty are for a single lab code the payer moved to a non-covered list in January, and that the front desk was never told to collect a waiver. The fix is a five-minute conversation with the front desk and a change to the lab order set. The denials stop.
A second example, from a specialty practice. An orthopedic group noticed that MRI claims from one of its two locations were denied for missing prior authorization at roughly twice the rate of the other. Follow-up appealed each claim with the authorization number, which existed, and won most of them at a cost of about forty minutes per appeal. Root-cause analysis found that the second location's scheduling template did not have the authorization field, so the number was recorded in a free-text note and never reached the claim. A ten-minute template change ended the category. The practice recovered a few thousand dollars through appeals; it saved far more by never producing the denials again.
The eight denial families
Denials arrive with hundreds of possible claim adjustment reason codes, but they sort into a small number of families, and each family is created at a predictable place. This is the categorization we use:
| Family | Typical reason codes | Where it was created | Preventable? |
|---|---|---|---|
| Eligibility and coverage | CO-27, CO-31, CO-109 | Registration, eligibility check | Almost always |
| Authorization and referral | CO-197, CO-15 | Scheduling, front desk | Usually |
| Coding and bundling | CO-4, CO-97, CO-236 | Coding, charge setup | Usually |
| Medical necessity | CO-50, CO-167 | Documentation, order justification | Sometimes |
| Timely filing | CO-29 | Billing workflow | Always |
| Duplicate | CO-18 | Billing workflow, corrected claim process | Always |
| Coordination of benefits | CO-22 | Registration | Usually |
| Payer error or policy | Varies | Payer | No, but appealable |
Two people categorizing the same denials should land in the same family. If they do not, tighten the definitions before you do anything else, because a log with inconsistent categories will point you at the wrong fix.
How to do it without a data team
Start with a denial log. It needs seven columns: date received, payer, provider, CPT code, billed amount, reason and remark codes, and the family you assign. Add three more as the process matures: preventable (yes or no), action taken (corrected, appealed, written off) and outcome. Most practice management systems can export the first seven from remittance data; the last three are the team's work. Keep it in one place that the whole billing team updates, not in personal spreadsheets.
Then, once a month, spend an hour on it:
- Categorize every denial into one of the eight families. If the log is exported from remittances, most of this can be done with a lookup on the reason code.
- Flag preventability. Could the practice have avoided it? Eligibility denials almost always yes; payer errors no.
- Rank by dollars, not count. One family is usually responsible for most of the money. Fifty small eligibility denials may matter less than five denied surgeries.
- Find the concentration. Within the top family, sort by payer, then provider, then CPT code. Root causes are almost always concentrated: one payer, one location, one order set, one provider's documentation habit.
- Assign the top family to a process owner with a specific change and a date to check whether the family shrank. Not "improve eligibility"; rather "batch eligibility runs two days before every visit starting Monday, front desk lead owns it, review on the 15th".
A practice with 2,000 claims a month and a 7 percent denial rate produces about 140 denials. That is a manageable log. If 55 of them are eligibility and 12 are denied procedures worth $1,800 each, the dollar ranking says the authorization family is worth $21,600 and the eligibility family, at an average $140 a claim, is worth $7,700. Work the authorization cause first.
Prevention actions that work quickly
| Family | Fix that usually stops it |
|---|---|
| Eligibility and coverage | Batch eligibility two days before the visit; card scanned at every check-in; script for failures |
| Authorization | Authorization field on the scheduling template; procedure list that flags what needs authorization; number carried to the claim automatically |
| Coding and bundling | Claim scrubber edits updated quarterly; coder feedback loop for the top three edits; modifier education for the providers involved |
| Timely filing | Daily rejection review; hold queue worked to zero weekly; filing-limit report by payer |
| Duplicate | A written corrected-claim process so resubmissions use the right frequency code |
| Coordination of benefits | Secondary insurance question at every registration; Medicare secondary payer questionnaire on file |
Notice that most of these fixes belong to the front desk, the scheduler or the provider, not to the billing team. That is the uncomfortable part of root-cause analysis and the reason it needs a practice manager or physician behind it: the billing team can find the cause, but it usually cannot change it alone.
Measuring whether prevention is working
Track each family's denial count and dollars per 1,000 claims, month by month. A prevention action should show a visible drop in its family within two months. If it does not, the cause was misidentified, the fix was not implemented, or a second cause exists. All three are worth knowing, and none of them are visible from the recovery total alone.
The monthly denial report should therefore show the denial rate and denied dollars with a trend line; denials by family, payer, provider and top CPT code; recovery dollars and write-offs with reasons; appeals due in the next 30 days; and the prevention actions in progress, each with an owner and the family it targets. The last section is the one that distinguishes a denial management program from a denial queue.
Questions we hear
What denial rate should a practice expect?
Industry surveys commonly cite initial denial rates between five and ten percent of claims. Well-run practices get below five percent. Your first monthly log establishes your baseline, and the trend by family matters more than the headline number.
Should every denial be appealed?
No. Some denials are correct, and some appeals cost more than the claim is worth. The decision should be deliberate: correct, appeal, or write off with a reason and a name, and never a write-off by default because the window closed. A written-off denial still belongs in the log, because it still tells you where it was created.
Who should own root-cause analysis?
The same team that does follow-up, with a practice manager or physician who can carry the fixes to the front desk and the providers. The people working the queue see the patterns first. This is why Revelrex offers Denial Management as an add-on to Medical Billing rather than as a separate service: the same team submits the claims, works the denials and reports the causes back to the practice.
What to do this week
- Export last month's denials from your remittance data with payer, provider, CPT code, billed amount and reason codes, and put them in one shared log.
- Categorize each one into the eight families and mark preventable yes or no.
- Rank the families by dollars and, within the top family, find the concentration by payer, location, provider or code.
- Name one process owner for that family, write down the specific change, and set a review date sixty days out.
- Add "denials per 1,000 claims by family" to the monthly report so the change is visible when it happens.
