Honest Standard Monthly Check

Use this skill when a practice running a Medicare between-visit program (chronic care management, remote monitoring, transitional care, APCM, behavioral health integration, and the like) wants to check each month that it is billing for real, needed,…

This skill is Labor as Code™ for the integrity method in Getting Paid to Do the Right Thing. Labor as Code is a skill that encapsulates a person's knowledge, judgment, taste, and skill in durable, inspectable, versioned form. The know-how is holding a between-visit program to a standard stricter than eligibility, every month, from inside the practice: eligibility tells you a service can be billed, and the Honest Standard asks whether it should be delivered, whether it was, and whether it helped.

The method lives in "Don't Turn Good Reimbursement Into Bad Medicine": the honest behavior and the two drifts behind each payment rule (Table 15.1), Summit Endocrine's fourteen months, the seven tests with the evidence to check for each (Table 15.2), Riverbend's monthly self-check, consent as a conversation about money, the three documentation drift signals, vendor contracts, and the four reasons to stop a service. The monthly worksheet is "Honest Standard monthly check" in "Field Worksheets." Read them for the argument and the cases. This skill runs the same check on the user's own program and never treats Summit's or Riverbend's findings as the user's.

This is an operating discipline, not legal advice. The book says it plainly and so does this skill: whether a specific arrangement, contract, or billing practice is lawful in the user's state and with their payers is a question for a qualified healthcare attorney or compliance adviser. The skill never says a program "is compliant," never says a claim must or needn't be repaid, and routes every legal question to that adviser with the evidence attached.

Clinical calls stay with clinicians. Whether a patient needs the program is a clinical judgment. The skill lays out the evidence in the record and what's missing; a clinician decides.

The skill takes de-identified records only. Names, dates of birth, record numbers, addresses, and phone numbers removed, dates replaced with day offsets, or a tool covered by the practice's business associate agreement used under its rules. Patient-understanding calls are made by practice staff, not by the skill.

Before running

1. Name the program and the month. One check per program per month.
2. Name who runs it. The book's pattern is a care-management lead and a physician together, about thirty minutes. A small group names one compliance lead across sites and runs the check at every site on the same schedule; a large organization samples every team and includes clinicians among the reviewers.
3. Pull the sample. Ten enrolled patient-months at random, not the ten the team knows best. Ask the user how they randomized.
4. Gather what sits outside the records. Phone-system or message timestamps for the sampled months, the list of patients disenrolled this month with reasons, complaints about the program, the consent script in use, any vendor contract, and last month's findings.

Step 1: Run the seven tests on the ten records

For each sampled patient-month, answer yes or no to each test, with the evidence behind the answer and what's missing:

1. For medical necessity, report what the record states and leave the judgment to the clinician. Strike any sentence in the skill's own output that says a patient "needs" or "doesn't need" the program.
2. For real service, compare documented minutes with the timestamps the user supplied. Note the direction of each difference. Documented time a little shorter than the phone log is the direction you want an error to run; documented time longer than any record of the work is a finding.
3. For patient understanding, the user's staff call about three of the ten and ask two questions: what is this program, and do you get a bill for it? The skill records the answers the user reports. An answer that names another office is a sign another practitioner may be billing the same service for the same month.
4. For measured outcome, count how many sampled patients have a recent value for the program's chosen measure. A patient with no value is a patient nobody can say the program helps.

Step 2: Check the three drift signals across the sample

These are cheap, countable, and the ones a compliance reviewer looks for:

• Time clustering. Real work produces a spread of times. Count the sampled documented times that sit within two minutes above a billing threshold. If the user can supply counts per time band for the last three months across the whole program, check the distribution, not just the sample. A pile just above the threshold means somebody is writing to it.
• Identical text. List any phrase repeated word for word across different patients' notes, beyond headings and standard labels.
• Frozen plans. Notes that say the care plan was reviewed when the plan hasn't changed in months, for patients whose conditions have. Also check for medication lists copied forward from last month.

If the program uses AI drafting or ambient documentation, apply the same rule the book gives: the person who signs the note attests that it describes what happened. A draft that says more than happened is wrong, however well it reads.

Step 3: Read the rules for drift

For each requirement the program checks before billing (a monthly time threshold, consent, a care plan, device data days, a contact deadline), write three lines: the honest behavior, the bureaucratic drift, and the fraudulent drift. Use the book's Table 15.1 as the pattern: for a monthly time threshold, the honest behavior is doing the work the patient needs and recording the time it took; the bureaucratic drift is stretching calls or calling patients who don't need a call so the month counts; the fraudulent drift is recording twenty minutes for a four-minute call.

For each requirement, name one early warning sign the practice can count in its own records. Drop any sign it can't count ("staff morale"). The bureaucratic drift is the real risk for an ordinary practice, because it happens gradually and looks like diligence, and it is the road to the fraud column.

Then check what staff and vendors are measured and paid on. Whatever the practice measures its staff on, they will produce. Paying anyone per enrollment or per billed minute is a finding on its own.

Step 4: Check the consent script

If the user supplies the script and the list of required consent elements from the payer's current guidance or their compliance adviser:

1. Quote the sentence that covers each required element, or mark it missing.
2. Check that a patient hearing it on the phone would know what the program is, what it will cost each month in plain dollars (after the practice checks secondary coverage), that only one office can provide it at a time, and that they can stop at any time.
3. Check that it asks whether another office already calls the patient every month, and that it offers time to think.
4. Flag any script that calls the program "free," promises to waive or cover the patient's cost, or pressures a decision on the call. Whether a practice can reduce or waive cost sharing is a compliance question, never something said on the phone.

The book's test for consent: if the patient is surprised by the first bill, the consent didn't happen, whatever the chart says.

Step 5: Read any vendor contract for what it rewards

If a vendor touches the patients or the claims, list every way the vendor is paid and the behavior each payment term rewards and doesn't reward. Then check for the book's six questions, quoting the clause or marking it missing:

• What is the vendor paid for: enrollments, billed months, a share of collections, or a flat fee for a defined scope of work?
• Who owns consent: whose script, whose recording, whose sign-off?
• How does the vendor coordinate with the practice's own clinicians, and who stops a vendor call when the practice's nurse or educator already manages the patient?
• What can the practice audit, and how fast: any note, call record, or claim on demand?
• What happens when an audit fails: does the vendor share repayment and rework, or keep its fees?
• How does the practice leave: patient notification and record ownership?

Every claim goes out under the billing practitioner's name, whoever did the work. Treat anything about legality as a question for the attorney, and hand over the missing-clause list.

Step 6: List who should stop

From the sample and the program's disenrollment list, list patients the clinician should consider for stopping, by the book's four reasons:

• The need resolved (for example, controlled for months on a stable regimen, with calls that have become social).
• Someone else is doing it (another practice, a specialist, or a health plan's program).
• The patient doesn't want it (said yes to be polite, doesn't answer).
• It isn't helping (months of calls, no change in anything measured, and no clinical reason to expect one).

A program that disenrolls nobody for six months is telling you nobody is looking.

Step 7: Conclude with actions, owners, and what goes to the adviser

Close with a short statement:

• Results by test: passes and fails across the ten, and which tests repeat a failure from last month. Two months in a row with the same "no" means the process, not the person, needs changing; name the process.
• Each "no" with a named owner and a fix before next month's claims go out. Where a sampled month fails real service, true documentation, or right supervision, recommend holding that claim (or reviewing it, if it already went out) and taking it to the compliance adviser. The skill doesn't decide repayment.
• The drift reading: whether the program shows bureaucratic drift (time clustering, identical text, frozen plans, incentives that pay for the proxy), with the counted evidence.
• Consent and contract gaps, each with who fixes it.
• The stop list for the clinician to decide on.
• A program-level call: keep running and check again next month; hold new enrollment until named failures have owners; or bring the program to its quarterly stop, hold, or scale review early because it failed an Honest Standard test in a way the process can't fix.
• Compliance review. If the program is about to scale, or its staffing or payment has changed, say it needs a qualified adviser's review first, and assemble the packet the book recommends: the consent script, the written workflow, the last three months of Honest Standard findings, the distribution of documented times, any vendor contract, and the stopping rules. List the evidence the practice has and lacks for each of the seven tests, and the questions it can't answer yet.

Where the evidence supports two readings (a time pile at the threshold that could be writing to the threshold or could be a program whose standard call genuinely runs that long), present both, with the check that would settle it (phone-log match, a wider sample). Where the sample is too small or the records too thin to judge a test, say so and name the evidence that would resolve it. Never round a thin sample up to a clean bill of health.

Checks before calling the work done

• The sample was random, and its ten patient-months are de-identified.
• Every test has a yes or no with evidence, or "can't judge" with what's missing.
• No sentence says a patient needs or doesn't need the program, and none says the program is compliant.
• Time clustering and identical text were counted, not described.
• Every "no" has an owner and a date before the next claims run.
• Legal questions are listed for the adviser, not answered.

What this skill does not do

It doesn't give legal, billing, coding, or compliance advice, and it doesn't decide repayment. It doesn't decide medical necessity or disenroll anyone; clinicians do. It doesn't call patients. It doesn't accept identifiable records outside a tool the practice has a business associate agreement for. It doesn't treat eligibility as the standard: every patient Summit enrolled was eligible. And it doesn't replace the monthly human read; software makes the sample cheap to check, and people decide what the findings mean.