VARDR Program Review
Missouri

Synthetic data. Practice names are invented. Several practices carry planted billing patterns, and several benign look-alikes are planted to show what the analysis declines to flag. Cases, people and outcomes are invented too.

Program integrity · Medicaid · Missouri

Signal origins

Where the review list comes from. Every signal starts in a data source, passes through a measure and a rule, and counts once for its evidence family. This screen shows that path for the whole program, and for any single signal.

Run RUN-e13e368456792cc8, computed 2026-11-16 by engine billing-pattern-5 on input 0004c6ed954b…. 109,202 claim lines, 42 practices, 6 quarters. Period shown: 2026Q3.

Inputs by evidence family

Each cell says what an input contributes to a family, with the number of practices that carry that family this quarter. Cells overlap: a practice can carry several families and several inputs feed one family, so the numbers are never added up.

Inputs by evidence family: what each input contributes, and how many practices carry the family
InputCoding5 practicesVisits2 practicesSpend5 practicesGrowth1 practiceBilled time2 practicesLinkage structure7 practices
Paid claim lines109,202 lines over 6 quarters; 18,592 in 2026Q3. Built-in synthetic extract.
5codes, units, levels, modifiers
2visits and patients
5dollars paid
1change over the window
2timed units per clinician-day
·
Provider enrollment file42 practices with specialty and county. Sets each practice's peer group.
5peer group by specialty
2peer group by specialty
5peer group; sole-provider rule by county
1peer group by specialty
··
Eligible member-months by countyUsed so that a practice growing with its county's eligible population is not read as growing by itself.···
1county growth, to adjust the panel
··
Payment and enrollment links956 observed links (deposit account, owner, address, phone, clearinghouse). Synthetic.·····
7who is connected to whom
Federal exclusion list8 current entries as of 2026-10-01. Synthetic. 2 possible matches, 2 still unverified.
Context on a practice. Not an evidence family: a possible match changes no tier, and stays unverified until a person decides.
Death and incarceration recordsNot connected for providers. Recipient matches are on the Recipients screen and stay unverified until a person confirms them.
Not connected. Shown so its absence is visible; it is not a zero.
National provider registry41 entries as of 2026-10-01. Synthetic. 3 of 42 practices carry a note: 1 not listed, 1 shown deactivated, 1 with a different primary taxonomy.
Names for providers in an uploaded extract, and notes where the registry and the claims disagree. Not an evidence family: a note changes no tier.
Cases citing each family when opened:Coding 4 opened, 3 resolvedVisits 1 opened, 0 resolvedSpend 4 opened, 2 resolvedGrowth 1 opened, 0 resolvedBilled time 2 opened, 1 resolved

What each measure did

Latest quarter, 2026Q3. "Counting" means the measure is the one that counts for its family on a practice, after persistence and the false-discovery control.

MeasureFamilyComparedNot comparedOutside the ruleCounting
Codes per visitClaim lines ÷ the lines expected for the practice's own visits at its peers' median lines per visit.Claim linesEnrollment fileCoding42022
Units per code lineUnits billed ÷ the units expected, code by code, at its peers' median units per line for the same code.Claim linesEnrollment fileCoding172511
Level 4–5 E/M share (age-standardised)Level 4–5 office visits ÷ the number expected for the practice's own mix of patient ages at peer rates.Claim linesEnrollment fileCoding142811
Modifier 25 on E/MOffice-visit lines carrying modifier 25 ÷ the number expected at the pooled peer rate.Claim linesEnrollment fileCoding142811
Visits per patientVisits ÷ the visits expected for the practice's own patients at its peers' median visits per patient.Claim linesEnrollment fileVisits42022
Paid per patientDollars paid per patient ÷ the median among peers, judged on a robust log scale. Set aside for the only practice of its specialty in a county.Claim linesEnrollment fileSpend42055
Paid per visitDollars paid per visit ÷ the median among peers, judged on a robust log scale. Set aside for the only practice of its specialty in a county.Claim linesEnrollment fileSpend42020
Intensity growth on a flat panelChange in billing per patient from the first two quarters to the last two, against peers' change, where the patient panel did not grow faster than the county's eligibles.Claim linesEnrollment fileEligibility countsGrowth42011
Billed time over 16 hours in a dayClinician-days in the last two quarters with more than 16 hours of timed services billed for one clinician. Counts at 3 or more.Claim linesBilled time42022

“Not compared” covers practices with too little volume or too few peers, and measures that do not apply to a specialty. It is a statement about what could be judged, not a clean result.

Reference checks

Practices compared with: Federal exclusion list (synthetic), 8 current entries as of 2026-10-01. A match is a candidate: found by a number or a name, and unverified until a person says what they checked.

2 possible matches: 2 unverified, 0 confirmed by a person, 0 found to be a different party.

  • Caring Hands Home CarePossible match · unverified

    Listed as Caring Hands Home Care, LLC, MO, from 2026-01-15, under section 1128(b)(4).

    The list entry carries this practice's NPI. A strong candidate; confirm the entry is current and the number is not a typing error.

    Record what you checkedRecorded with your name. Changes no tierOpen

    For example: the enrollment file's owner and tax number against the list entry.

  • Lakeview MedicalPossible match · unverified

    Listed as LAKEVIEW MEDICAL INC, MO, from 2019-06-20, under section 1128(a)(1).

    The list entry has the same business name in this State and carries no NPI. A weak candidate: names are shared, and this is often a different party.

    Record what you checkedRecorded with your name. Changes no tierOpen

    For example: the enrollment file's owner and tax number against the list entry.

Signals that count: 15

One row per practice and family. Select one to trace it from the source to the case.

Trace: Visits per patient, Clearwater Counseling

From the source to the case, in the order it happened. Each step is something that can be checked.

Clear
  1. 1

    Source

    0004c6ed954bca6e

    Paid claim lines: Built-in synthetic extract.

    3,897 of this practice's lines over 6 quarters; 658 in 2026Q3.

    No quarter held back: the built-in extract is complete.

  2. 2

    Measure

    Visits per patient

    Visits ÷ the visits expected for the practice's own patients at its peers' median visits per patient.

    Reads: Claim lines, Enrollment file.

  3. 3

    Peers

    10 other mental health counseling practices statewide, with this practice left out of its own reference.

  4. 4

    Result

    Observed 363 against 162 expected: a ratio of 2.24×.

    After shrinkage toward the peer group: 2.23×, 90% interval 2.04× to 2.42×.

    Probability the true ratio exceeds 1.5×: 100.0%. False-discovery q-value: 0.000.

  5. 5

    Rule

    billing-pattern-5 · be6f8bfe67db

    A measure counts only above 1.5× peers with 95% probability, after a 5% false-discovery control, in 2 of the last 3 quarters, with at least 30 visits and 11 patients.

    Counts toward review: flagged in 3 of the last 3 quarters.

  6. 6

    Signal

    Counts as the Visits family for this practice. A family counts once, by its strongest measure.

    Sees each patient 2.24× as often as peers in the same specialty.

  7. 7

    Review

    Priority review: 3 independent families (Coding, Visits, Spend).

    A review priority, not a probability and not a finding.

  8. 8

    Case

    No case has been opened for this practice. Opening one is a person's decision.