Last updated: 2026-09-08
Claims hiring goes wrong when a job description lists software instead of judgement, so every one of these skills arrives here with the test that proves it. Reading a remittance advice and an explanation of benefits sits underneath the rest, because anyone who can't separate a contractual write-off from a denial from a patient balance will misdirect money in week one. Denial-code literacy builds on that, where a reason code and its remark code read together point at one named fix instead of a resubmission sent in hope. Pre-submission checks are the cheapest hours in the cycle and the easiest thing to interview badly for, since eligibility that lapsed before the date of service costs a minute to catch and a month to unpick. Deadline discipline forgives the least, because a filing window that shuts takes the balance with it and no later effort reopens it. Appeal work gets judged on whether a second phone call goes anywhere, which depends on the reference numbers and named contacts somebody logged during the first one. Then the boundary that never moves, since choosing a code and writing the clinical argument in a letter of medical necessity belong to your credentialed coder and your provider. Where these claims skill facts were checked closes the page.
A claims assistant reads a remittance advice by working it line by line rather than total by total, and the first decision on every line is which of three things that line has become. One is a contractual adjustment the practice absorbs. Another is a denial somebody has to work through. The third is a balance the patient owes, and all three can sit inside one claim while looking almost identical on a posting screen. Sorting them correctly is the competency, and mis-sorting a single line sends money somewhere nobody intended.
Group codes do that sorting. CO marks a contractual obligation, so the practice absorbs the amount and can't bill the patient for it. PR marks patient responsibility, which turns into a statement. PI covers a reduction the payer took on its own initiative rather than under your contract, and OA pushes a balance sideways for some other stated reason. Read a reason code without its group code and a write-off gets billed to a patient, or a deductible gets absorbed as though the contract required it. Neither mistake announces itself. Both of them look like an ordinary posted line.
Arithmetic is the other half of the reading, and it catches the quiet losses. On any single line, the billed charge minus the contractual adjustment minus patient responsibility should land on the amount paid. When the subtraction leaves a remainder, something on that line was denied inside a remittance whose header total looked ordinary. Lines like that never raise a flag. They sit in a balance nobody works, because the claim reads as paid.
Placement matters as much as the codes do. An adjustment attached to one service line says the payer took issue with that service, while the same code sitting at claim level applies to everything on the file. Somebody who posts claim-level and line-level adjustments interchangeably produces a ledger that balances to the penny and points at the wrong service. The Centers for Medicare and Medicaid Services publishes Medicare claims and billing guidance covering where reason and remark codes belong inside an electronic remittance, and that placement rule repays a second reading. How a posting screen displays those segments varies by platform, and our roundup of medical billing tools and software covers which systems surface the detail.
An explanation of benefits is the same adjudication written for a different reader, so reconciling the two is its own small skill. The patient's copy carries the plan's own wording for a decision and rarely carries the code, which is why somebody arrives at the front desk convinced a service was refused when the remittance shows a routine deductible line. Holding both documents beside each other, matching them claim by claim, and telling the patient which number they owe is the fix. Getting that wrong twice with one family costs more goodwill than the balance was worth.
Test this competency with paper instead of with questions. Hand a candidate one redacted remittance page carrying a mixed set of lines, such as a contractual adjustment, a deductible, a line-level denial and one line where the subtraction leaves a remainder, then ask them to label every line and say what happens to it next. Strong candidates find the line that doesn't balance without being pointed at it. Weaker ones total the page, agree with the total, and hand it back.
Denial-code literacy is the ability to move from a claim adjustment reason code and its remittance advice remark code to one named fix, and it separates a claims hire from a data-entry hire. Reason codes give the category of problem the payer found. Remark codes name the specific item inside that category that's missing or wrong. Either one read alone supports several contradictory next steps, which is how a queue turns into guesswork.
Reason code 16 is the standard demonstration. It reports that a claim lacks information or carries a submission error, and on its own that describes half the reasons a claim ever fails. Beside it sits the remark code, naming an item such as an absent referring provider identifier, a date that contradicts the service, or a document the payer wants attached. Anyone working a queue of 16s without opening the remark codes is guessing at all of them.
Some remarks call for nothing, and knowing which ones saves hours. Informational remarks explain how a payer arrived at an amount and ask for no action, so an assistant who treats every remark as a task manufactures busywork out of a clean payment. Reason code 45 deserves the same restraint. That code reports a charge above the allowed amount under your own contract, which is the fee schedule doing its job rather than a decision worth arguing with, and chasing it burns weeks against a rate you already agreed to.
The literacy pays off as routing. Every worked denial ends at one of four destinations, and naming the destination is the output of the skill.
Blind resubmission is the failure this skill exists to prevent. Sending an identical file back into adjudication earns a duplicate-claim rejection, adds a second entry to the aging report, and spends days of a filing window that was running before anybody noticed. A queue full of duplicate denials points at training rather than at the payer, and our walk-through of how a virtual assistant works denials and appeals shows what the notes look like once the habit holds.
Two tests work here, and they run a month apart. Give a candidate a reason code paired with a remark code they've probably never met, then listen for whether they go to the payer's portal text and then to the phone, or whether they resubmit and hope. Later, read a week of their claim notes rather than their throughput number. Every closed denial should carry a cause written in the practice's own words, since a queue that never feeds a cause back into registration and coding stays exactly the size it was.
A claim needs four checks before it goes out, and three of them are administrative verification skills while the fourth ends in a flag rather than a fix. Each one costs a minute at this point and days once a payer has looked at the file. That gap is the whole economic case for putting a careful person in front of submission instead of only behind it.
Eligibility as of the date of service is the first, and the date is the part people get wrong. Coverage confirmed when an appointment was booked says nothing about coverage on the day the patient was seen, since plans terminate mid-month. Retroactive terminations, a new plan year, a switch to a Medicare Advantage product and a change of employer all arrive later as denials that look clinical and are nothing of the kind. Whoever holds the front-end version of this work, such as an insurance verification specialist, owns the prevention, and a claims assistant needs enough of the same knowledge to recognize when a denial belongs back at the front desk.
Identity fields are the second, and they get checked against the card rather than against the chart. The member identifier including any alpha prefix, the patient's name spelled as the plan carries it rather than as the patient prefers it, the date of birth, and the split between subscriber and patient whenever the patient isn't the policyholder. Pediatric claims and claims for adult dependents fail on that split constantly. Such a failure reads as a coverage problem right up until somebody looks at which loop the names are sitting in.
Payer routing is the third. Claims travel on the payer identifier your clearinghouse recognizes, which isn't always the number printed on the card, and a file routed to the wrong entity inside a large payer's family gets rejected without ever reaching adjudication. Coordination of benefits belongs here too, since a secondary claim needs the primary payer's adjudication attached and assembling it afterwards turns one piece of work into two. Catching these before submission is the cheapest hour in the cycle, and our guide on how to reduce claim denials covers the upstream half of it.
The fourth check is a completeness sweep that stops at the flag. An assistant can see that a procedure and a diagnosis don't sit together comfortably, that a modifier looks absent where the place of service implies one, that an authorization number never made it onto the file, or that a referring provider identifier is missing. Every one of those goes to the coder as a named question with the claim attached. None of them gets fixed by the assistant, because adding a modifier to clear a scrubber edit is a coding decision wearing administrative clothing, and the credential behind that decision belongs to your coder.
Test the fourth check first, since it's the one where a helpful answer is the wrong answer. Ask a candidate what they do when a scrubber stops a claim for a missing modifier and the coder is out for the afternoon. You want an answer that names the flag, the queue it waits in, and who gets told. Any answer reaching for the modifier itself is a training problem at best. For the other three checks, seed a small batch with known defects, such as a stale member identifier and a claim routed to the wrong payer entity, then count how many the candidate catches before submission rather than after.
A denied claim gets appealed by dating everything backwards from the day the window shuts, which is why the calendar is the skill and the paperwork is the easy part. Two separate clocks run on any claim. Timely filing runs from the date of service and decides whether the original claim counts as filed at all, while the appeal window runs from the remittance date and decides whether a payer will look again. Missing the first one can't be undone in most cases, and a reason code reporting an expired filing limit is the payer saying so in one line.
Windows differ by payer and by contract, so the calendar has to be a written artifact rather than a memory. Medicare fee-for-service publishes its filing limit in regulation. Commercial windows come out of the agreement your practice signed, and each plan names its own stages differently, with Medicare calling its first appeal level a redetermination while a commercial plan may want a reconsideration request before it accepts a formal appeal. Reading your own payer agreements is the only reliable route to that table, since no national summary knows what you negotiated.
Proof is what makes a filing dispute survivable. When a payer says a claim arrived late and the practice believes otherwise, the thing that settles it is evidence of the original transmission, which lives in the clearinghouse acknowledgement report from the day the batch went out. Saving those reports by date gives an assistant an argument to make. Relying on a practice management system's submitted status leaves you with memory against a payer's records, and that argument loses.
Packet assembly is deliberately narrow, and it's never a copy of the whole chart. Six items go in, such as the payer's current appeal form, the remittance page showing the disputed adjustment, the claim detail exactly as submitted, the paragraph of the payer's own medical policy covering the service, proof of timely filing, and the specific chart pages a provider names. Everything on that list is retrievable by an administrative person. Deciding which chart pages matter is not, so the provider names them.
Tracking is where appeals get quietly lost, and the fix is a record with fixed columns rather than a reminder. One row per appeal carries the submission date, the channel used, the confirmation or reference number the portal returned, the name of whoever answered the phone, the payer's stated decision window, and the date of the next check. Ring back without the reference number and the representative's name, and the conversation starts over from nothing, which is how one appeal turns into six weeks of repeated explanation. Practices that would rather buy the whole function than build it can compare firms in our ranking of denials and appeals specialist companies.
Two tests separate deadline discipline from good intentions. Ask a candidate to show you the tracker they kept in their last position, redacted, and look at whether it has a next-check column at all. In the first month, pick three open appeals at random and ask what the reference number and the next call date are for each of them. Somebody who answers without opening a search box is holding the calendar. Anyone reconstructing it from portal screens is holding a list of things they meant to do.
No, and that line stays fixed however much faster crossing it would make a queue move. Choosing a procedure or diagnosis code, changing one already on a claim, adding a modifier to clear a scrubber edit, and writing the clinical argument inside a letter of medical necessity all belong to your practice's credentialed coder and your provider. An administrative assistant gathers, requests, assembles, tracks, calendars and escalates. Coding judgement and clinical reasoning stay where the credential and the liability already sit.
Credentials themselves come from bodies outside any staffing arrangement. AAPC issues the coding credentials most physician practices hire against, and AHIMA issues the health information credentials sitting alongside them. The US Bureau of Labor Statistics keeps an occupational profile for medical records specialists that describes the education and certification picture for that adjacent occupation. Honest Taskers claims no AAPC or AHIMA credential for its staff and says so plainly, because your team owns the coding and the billing outcome either way.
Where credentialed coding is the thing you need, it gets bought from firms that publish the credential, and this market states its own claims clearly enough to read directly. Transcure lists more than 1,100 certified billers and coders, prices at 3% to 5% of monthly collections, and reports ISO 27001 alongside AAPC certification (company-reported). Plutus Health references AAPC-certified audits against a 99.2% coding accuracy threshold and reports HIPAA and SOC 2 Type II (company-reported). AGS Health, whose service list runs from coding and claims through denials to provider enrollment, describes holding a "Cybersecurity Transparent Designation 2023" rather than a named security certification (Source: AGS Health, company-reported, read 24 August 2026). Health Information Associates sells coding review and documentation improvement and states 10 years of average coder experience, though no individual coding credential appears on that page. Reading four such pages side by side teaches more about what a credential claim covers than any table does. What the coder owns is set out on our page covering medical coder skills.
Honest Taskers works the administrative layer around claims rather than inside the coding, and that scope belongs in a job description before anybody starts. Request intake, payer portal work, queue preparation, documentation chasing, deadline tracking and appeal packet assembly sit inside it. Code selection, coding audit and clinical documentation judgement sit outside it. The talent pool includes licensed nurses and physicians, which describes recruiting rather than the scope of any placement, so ask one candidate about their own claims history instead of reading the pool as a capability. Honest Taskers staff do administrative and clinically adjacent work and never give clinical advice or make clinical decisions.
On terms, rates run $10.00 to $12.65 an hour depending on the role, candidate background, schedule and location. Recruiting runs in the Philippines, Latin America, India and Pakistan, and professionals work the client's US time zone and approved schedule. Staff are HIPAA-trained under a dedicated compliance officer with quarterly HIPAA and data privacy training, a Business Associate Agreement gets signed when a professional will access protected health information, and the company describes its own security environment as SOC 2 audit ready. New clients may receive a two-week working trial with their first selected professional, subject to current service terms, and that trial stays separate from the unlimited replacement support, where a performance-related replacement can qualify for a credit covering the incoming professional's first two weeks. Every client works with a dedicated Customer Success Advocate.
Honest Taskers reports 99.6% average monthly retention, and on a claims desk the reason behind that figure matters more than the figure. Competitive pay, healthcare coverage for eligible team members, interest-free employee loans, wellness support and yearly performance-based raises sit behind the role, and the payoff is knowledge that only accumulates. Which payer buries its remark text several screens deep, which appeal fax number answers, which plan's portal times out at the file upload. None of that transfers in a handover document, so every month a competent claims assistant stays is a month your practice doesn't spend relearning its own payers.
Honest Taskers rates, recruiting geography, trial terms, retention figure, scope boundary and compliance posture come from the company's own published rate card and service terms. Competitor facts were read at each firm's own website on 24 August 2026 and are labelled company-reported wherever a company states them about itself, covering the Transcure certified-coder count, pricing and certifications, the Plutus Health audit credential and accuracy threshold, the AGS Health cybersecurity designation, and the Health Information Associates experience figures. Claim adjustment reason codes, remittance advice remark codes and group codes are national code sets rather than anybody's proprietary list, and the Centers for Medicare and Medicaid Services publishes the Medicare claims and billing guidance describing where those codes belong inside an electronic remittance. Coding and health information credentials are issued by AAPC and AHIMA, and the occupational profile for medical records specialists is published by the US Bureau of Labor Statistics. Filing and appeal windows are set in regulation for Medicare fee-for-service and by contract for commercial plans, so the only accurate table is the one built from your own signed agreements. No denial rate, clean-claim rate, days-in-accounts-receivable figure, appeal success rate, collection percentage or savings percentage appears anywhere above, because your payer mix, specialty and contracts decide every one of them and your own remittance files hold the answer.
Practices that have settled the skill list and want to compare providers rather than candidates can start with our ranking of medical insurance claims virtual assistant companies.
Request claims candidates who can read a remittance and hold an appeal calendar.
