Home Health Care · BPO Services
A CFSS agency is the employer: it recruits the support worker, schedules the visit and pays for the hour whether or not the claim behind it survives. So every gap between the visit and the claim - an EVV record that did not capture, a timesheet that runs past the authorisation, a code from the old programme - lands on the agency's own balance sheet.
We took the cycle between the visit and the payment off the agency desk: EVV and timesheet data reconciled before anything becomes a claim, billing prepared and submitted, and every rejection worked back to a resubmission.
Why home care billing breaks where it does
Home care is the part of healthcare where the evidence for the claim is created in somebody's living room, on a phone, by a worker who is mid-shift. Federal law has made that evidence the basis of payment: the 21st Century Cures Act requires electronic visit verification for Medicaid personal care and home health services, and Minnesota's implementation records six things about every visit - the service performed, who received it, who provided it, the date, the location, and the time it began and ended.
Those six fields are also the six ways a claim can fail. Minnesota DHS uses EVV data in a post-payment review, and where a claim is not supported by the EVV record in the aggregator the money can be taken back after it has been paid. A billing process that checks a claim only for format is checking the wrong thing.
The programme change makes it harder again. CFSS replaced Personal Care Assistance and the Consumer Support Grant on 1 October 2024, and under the agency model the lead agency grants service units in fifteen-minute blocks to the provider agency, which is the employer. Every authorisation has a ceiling in units, every visit consumes some of it, and every mismatch between what was worked, what was recorded and what was authorised is a rejection waiting to happen.
None of that stops the payroll run. It surfaces weeks later as a rejection, or later still as a takeback, by which time the person who could have explained the visit has moved on - and the agency doing that reconciliation in-house is spending its office week on matching and chasing rather than on recruiting workers and looking after clients.
How we set the engagement up
We started from the exceptions rather than the happy path. A billing run that only handles clean visits saves an agency very little, because the cost sits in the small share that do not reconcile. So the first thing we defined was what an exception looks like, who decides it, and how quickly it has to come back.
Everything else follows from that. The pre-billing check runs against the records the agency already keeps, so a visit has one version rather than a billing copy that can drift away from the roster. Rejections are coded by reason rather than cleared one at a time, because a reason that repeats is a process fault and not a claim fault. And what goes back to the agency is a reconciled position - billed, paid, open, and why - rather than a task list the office would then have to interpret.
What the billing workflow includes
- A pre-billing reconciliation of every EVV visit record against the caregiver timesheet, so a missing or incomplete visit is found before it becomes a claim
- A units check against the client's current service authorisation, so an over-authorisation is caught in the same cycle rather than in a post-payment review
- Claim preparation and submission each billing cycle, coded to the programme the visit was actually delivered under
- A worked queue for rejections and denials, each one coded by reason so a repeating fault shows up as a pattern
- Correction and resubmission inside the payer's timely filing window
- A reconciliation report back to the agency: what was billed, what was paid, what is still open and why
What changed for the agency
The reconciliation now happens before submission rather than after payment. A visit that did not capture, a timesheet that runs past the authorisation, a code that belongs to the old programme - each is raised as an exception in the same cycle, while the caregiver and the client can still explain it.
It also changed what the office does with its week. The work that used to sit between the schedule and the payment - matching, chasing, resubmitting - now arrives as a reconciled position with the exceptions already separated out, so the agency's own team spends its time on recruitment, client care and compliance rather than on claim admin.
No performance figures are published on this page yet. Volumes, acceptance rates and turnaround are being measured with ATRIO and will be added here once they are confirmed.
What another home care agency should take from this
Fix the record, not the claim. Almost everything that gets rejected was already wrong at the visit - a start time that did not save, a unit count nobody compared against the authorisation. A billing process that begins when the claim is built has started one step too late.
And treat a denial as data. One rejection is an admin task; the same rejection three cycles running is a scheduling or training problem that will keep costing money until somebody names it. Coding every denial by reason is what turns a queue into a fix.
Home care billing questions
1What does an employee billing BPO actually take off our desk?
The cycle between the visit and the payment: reconciling EVV and timesheet data, checking hours against the service authorisation, preparing and submitting the claim, and working every rejection back to a resubmission. Scheduling, clinical decisions and the employment relationship stay with the agency.
2Why does EVV sit at the centre of home care billing?
Because it is the evidence for the claim. Minnesota DHS uses EVV data in a post-payment review, and a claim the EVV record does not support can be recovered after it has been paid. Checking the visit record before submission costs far less than defending it afterwards.
3Do you replace our billing or EVV system?
No. The service is built around the systems the agency already runs, so there is no migration and no second copy of the visit data to keep in step.
4What happens to a denied claim?
It goes into a worked queue rather than a folder. Each denial is coded by reason, corrected and resubmitted inside the timely filing window, and the reason is reported back - so a fault that keeps recurring is visible as a pattern instead of arriving as a surprise at the year end.