One month in, the practice kept
a third more of what it billed.
An independent US oncology and infusion practice, about 30 patients a day, handed us a revenue cycle that was losing roughly seventy cents of every dollar billed. July 2026 was the first full month under CareCapital. This is what changed, how, and which part of it we expect to hold.
“CareCapital has been an excellent partner since taking over our billing in July. They are responsive and reliable. Claims that used to sit for weeks are worked in days, and posting is reconciled daily.
They have great expertise in infusion and drug billing, and I have real confidence our billing is in good hands. Their proactive approach and attention to detail have made a noticeable difference in our Collections.”
Of everything billed, how much actually landed in the bank. January to June against July. Higher is better.
How long a claim takes to convert to cash. Lower is better, so the gold dot sits to the left.
Unchanged across the comparison. This was a collections change, not more patients, and it is the figure that answers the first question anybody sensible asks about the two above.
The difference was not effort. It was that somebody was looking.
None of this needed new software or another person at the practice. It needed someone whose job is the claim that is not moving, and a process that surfaces that claim instead of waiting for a payer to mention it.
Claims that were held, not rejected
BeforeA held claim appears on no rejection report. Nothing surfaced them and nothing was chasing them.
NowA held-claim queue that gets worked to a schedule. Nothing sits waiting on nobody.
Denials
BeforeWorked one claim at a time, where they were worked at all.
NowGrouped by cause, with the cause corrected upstream, so next month does not produce the same batch.
Coding and eligibility errors
BeforeFound by the payer, at the cost of a denial and a month.
NowFound at eligibility and authorization review, before the claim goes out the door.
Work that could not bill
BeforeMissing signatures and unentered chart values had nothing chasing them.
NowTracked, returned to the provider, and billed once resolved.
Aged claims
BeforeAged.
NowWorked as a standing queue, largest and oldest first, until the backlog is gone.
The denials were not random. They were five patterns.
Grouped by cause rather than worked one claim at a time, most of a month of denials in an infusion practice comes down to a short list. Fixing the pattern is what stops next month's version of it.
- Drug waste modifiers
- Infusion and injection lines denied as incomplete because JW or JZ was absent. The same claims were often billed at the wrong unit count, so the two were corrected together.
- Clinical modifiers on administration codes
- Denials on drug administration for a missing EA, EB or EC modifier. A one-character omission that stops the whole line.
- Office visits billed alongside a procedure
- Visits billed on the same day as an administration or drug code need modifier 25, and read better with a distinct primary diagnosis where the record supports one.
- Codes that will not pay alongside anything
- Port irrigation bundles when it is billed with other services on the same day and accepts no modifier to separate it. It bills alone or it is adjusted after the denial. Knowing which is the difference between a write-off and an appeal that was never going to win.
- Coordination of benefits
- Claims submitted to the wrong primary payer. The fault was upstream, in the order insurances were attached when the patient chart was created, so that is where we fixed it.
Payers are described generically here. Which plan denied which code for which practice stays between us and the practice.
July was the practice's strongest collection month of the year, on the same number of patients as the month before.
It came in roughly 76% above what they had been averaging each month across the first half of the year. That figure sits here rather than at the top of the page because it is the one least likely to repeat. The two at the top are the ones to hold us to.
What repeats, and what does not.
Part of that first month was catch-up. We would rather say so now than have month two read as a step backwards.
Repeatable
- A higher first-pass payment rate on new claims
- Fewer denials caused by coding and eligibility errors
- Faster submission, so cash arrives sooner each cycle
- Consistent follow-up on every unpaid claim
One-time
- Recovery of aged claims that had built up before we started
- That backlog is finite. It gets smaller every month it is worked
- This part of the lift tapers across the next two quarters
The measure of month two is not whether it beats July's total. It is whether the collection rate holds.
- The baseline
- The practice's own January to June 2026 actuals, averaged. Not an industry benchmark and not our own estimate.
- What was controlled
- Patient volume was flat across the comparison, so the change is not more patients. We have not adjusted for drug mix, payer mix, fee schedule changes or seasonality.
- What was one-time
- Some of month one was recovery of claims that already existed when we took over. That money is real and it is finite. We have not published a split between recovery and run rate because we do not yet have one we would stand behind.
- The sample
- One practice, one month, published with the practice's permission. The results of a single engagement are not a forecast of yours, and anyone who tells you otherwise is selling.
Figures are the practice's own, published with permission. The practice is not named, at its request and ours.