Arete Healthcare LLC
Independent Practices | Shared Infrastructure | Clinician-Led Governance
Independent Practices | Shared Infrastructure | Clinician-Led Governance
Arete Analytic Brief
A second provider-level sensitivity analysis of spinal CMT utilization in a tiering appeal
The chart below is the core financial argument of this brief: the intermediary’s prior-authorization remedy costs the practice roughly fourteen times what it could plausibly save the insurer.
Figure 1. Illustration of the financial impact of requiring a chiropractor to participate in a prior authorization based on a determination that the chiropractor’s use of the 98941 spinal CMT code exceeds the intermediary’s “benchmark.” While all 98941 services provided during the measurement period were appropriately documented in the medical record, on submitted claims, and claims were adjudicated and paid, the “excess” 98941 code use according to the intermediary is estimated to have an approximate $475 financial impact, of which the patient who sought the treatment is responsible for over 50% through a copay and/or deductible. With the intermediary’s prior-authorization process applied to all patients — and notably not addressing the visit-specific CMT code level for which the intermediary is concerned — the minimum total administrative cost to the chiropractor’s practice will be approximately $3,425. The chiropractor’s overall utilization (visits per patient and services per visit) is very low. Visualization was human generated.
This case is notable not because it is unique, but because it is not.
A chiropractor in a different practice than a previous analysis, and evaluated by the same intermediary across the same three insurers, was “down-tiered” after 98941 exceeded the same 60% spinal CMT “benchmark” established by the intermediary. The intermediary did not evaluate a single patient treated, did not review medical records and did not review individual claims.
The underlying claims tell a familiar story: a small subset of patients receives substantially more care, those patients are somewhat older and slightly higher-risk, and nearly every 98941 encounter is accompanied by diagnoses spanning at least three spinal regions.
For this provider, 17 (12.4%) of 137 patients account for nearly half (49.5%) of all 98941 use. Removing those 17 patients changes spinal CMT 98941 use of the provider from 74.8% to exactly 60.0%. All 98941 services were supported with appropriate diagnosis codes. The pattern is consistent across all three insurers: 98941 represents 76.3% of spinal CMT for Insurer 1, 72.8% for Insurer 2, and 77.2% for Insurer 3. The result is therefore not attributable to one payer population.
Replicating the analysis used for another provider, and using the same $5 assumed difference between 98940 and 98941 levels of reimbursement, 50% patient responsibility through copays and deductibles, and $25 in practice administrative costs to submit a prior authorization request for every patient treated, the practice will incur a minimum of $3,425 in annual prior authorization related administrative costs. For the 17 patients, there were ~95 “excess” 98941 services for an “excess” cost of ~$475, and ~$235 net cost to the 3 insurance companies after patient copays and deductibles.
$3,425 in annual practice administrative costs to save 3 insurance companies a total of $235. Nearly identical to the $3,500 in annual practice administrative costs to save the same 3 insurance companies a total of $125 identified in a previous analysis (arete.healthcare/research/spend-3-500-to-save-125).
Two chiropractors, $7,000 in annual administrative costs to save 3 insurance companies $350.
Two chiropractors with decades of serving their local communities providing exceptional, personalized care at extremely low cost and nearly identical performance, yet labeled an “outlier” by an intermediary chiropractic network who never saw a patient or evaluated a patient’s medical records – perhaps it is the network that has it wrong.
Extrapolate this to a national scale – does any of this make sense?
This brief does not determine whether the intermediary’s 60% benchmark is methodologically valid. It does show why provider appeals benefit from moving beyond a single crude percentage to examine patient concentration, visit frequency, diagnosis concordance, and denominator composition. A repeated finding in two separate practices suggests that this should be a routine analytic step whenever a tiering decision materially affects a provider and their patients.
| Measure | All patients | Without 17 | 17 patients |
|---|---|---|---|
| Unique patients | 137 | 120 | 17 |
| Patient age, mean | 45.5 | 44.5 | 52.9 |
| HCC RAF, mean | 0.281 | 0.278 | 0.298 |
| Female, % | 58.4% | 59.2% | 52.9% |
| Visits | 651 | 413 | 238 |
| Units | 1,324 | 855 | 469 |
| Spinal CMT units (98940/1/2) | 643 | 405 | 238 |
| 98940 units | 162 | 162 | 0 |
| 98941 units | 481 | 243 | 238 |
| 98942 units | 0 | 0 | 0 |
| Visits per patient | 4.75 | 3.44 | 14.00 |
| Units per patient | 9.66 | 7.13 | 27.59 |
| Units per visit | 2.03 | 2.07 | 1.97 |
| % spinal CMT — 98940 | 25.2% | 40.0% | 0.0% |
| % spinal CMT — 98941 | 74.8% | 60.0% | 100.0% |
| % spinal CMT — 98942 | 0.0% | 0.0% | 0.0% |
| 98941 visits with ≥3 expanded regions on same visit† | 99.6% | 99.2% | 100.0% |
| 98941 visits with ≥3 expanded regions in patient 12-month profile‡ | 99.8% | 99.6% | 100.0% |
Analysis period: September 1, 2025–August 31, 2026, after filtering to Insurer 1, Insurer 2, and Insurer 3. Age and HCC RAF are patient-weighted.
†Expanded spinal-region mapping is an analytic sensitivity heuristic, not a payer coverage rule.
‡Longitudinal concordance counts a 98941 visit when the patient had ICD-10 diagnoses spanning at least three mapped spinal regions anywhere in the same 12-month period; it is not a visit-level medical-necessity determination.
The payer-level analysis shows that the provider’s elevated 98941 distribution is present across all three insurer populations rather than being driven by a single payer.
| Insurer | Spinal CMT units | 98941 units | % 98941 |
|---|---|---|---|
| Insurer 1 | 274 | 209 | 76.3% |
| Insurer 2 | 290 | 211 | 72.8% |
| Insurer 3 | 79 | 61 | 77.2% |
Interpretation. All three insurer populations independently exceed the intermediary’s 60% threshold, with 98941 rates ranging from 72.8% to 77.2%. The sensitivity cohort is also clinically coherent rather than diagnostically anomalous. Every one of its 238 98941 visits had same-visit diagnoses mapping to at least three expanded spinal regions. The retained population remained highly concordant as well: 241 of 243 98941 visits (99.2%) met the same-visit expanded definition, and 242 of 243 (99.6%) met the longitudinal definition.
A chiropractor in a group practice was down-tiered by an intermediary network because 98941 represented more than 60% of spinal CMT. Arete analyzed complete claim-line data across three insurers for September 1, 2025–August 31, 2026, including all 12 ICD-10 diagnosis positions, patient demographics, HCC RAF, utilization, and a minimum-cardinality patient-level sensitivity analysis.
The provider’s baseline 98941 rate was 74.8% across 643 spinal CMT units. Seventeen of 137 patients (12.4%) accounted for 37.0% of spinal CMT and 49.5% of 98941 use; counterfactually removing that cohort reduced the remaining rate to exactly 60.0%. The 17 patients averaged 14.0 visits per patient versus 3.4 among the remaining 120 patients and were modestly older with slightly higher HCC RAF. Diagnosis concordance was high: 99.6% of all 98941 visits had same-visit diagnoses spanning at least three expanded spinal regions and 99.8% met a 12-month longitudinal definition; the 17-patient cohort was 100% concordant under both measures.
This separate provider and practice reproduce a concentration-sensitive pattern previously observed in another provider exposed to the same intermediary, insurer configuration, and market. The finding illustrates the practical value of patient-level sensitivity analysis when a crude provider percentage is used in tiering decisions.
Context. This brief is intentionally limited to provider-specific performance and sensitivity analysis. It does not evaluate the intermediary’s benchmark development, profiling validity, risk-adjustment methodology, or tiering decision process.
De-identified EHR/RCM claim-line data for one chiropractor were filtered to Insurer 1, Insurer 2, and Insurer 3, which use the same intermediary chiropractic network, for September 1, 2025 through August 31, 2026, using all 12 ICD-10 diagnosis positions. Visits were unique provider-patient-date-payer combinations; spinal CMT comprised CPT 98940, 98941, and 98942. For each patient, excess contribution equaled 98941 units minus 60% of total spinal CMT units; patients were removed in descending order of excess contribution until the remaining provider rate reached 60% or lower, identifying a minimum-cardinality sensitivity cohort. Diagnosis concordance used a conservative region-specific mapping for cervical, thoracic, lumbar, sacral, and pelvic diagnoses, assessed at both the same-visit and 12-month patient-profile level.
Single-provider case study. Findings describe one provider and should not be generalized to other providers, practices, markets, or intermediary networks.
Intermediary specifications were unavailable. The exact profiling interval, attribution logic, denominator, exclusions, benchmark-development process, and any risk adjustment used by the intermediary were not provided; this brief therefore does not attempt to reproduce or validate that methodology.
Sensitivity analysis is counterfactual. The 17 patients are not proposed exclusions from profiling; the analysis quantifies how strongly patient concentration influences the aggregate rate. Ties can produce alternative minimum sets.
Claims diagnoses do not establish medical necessity. Same-visit and longitudinal concordance indicate anatomical consistency between diagnoses and multi-region CMT coding, but neither substitutes for medical-record review.
HCC RAF is a general risk measure rather than a musculoskeletal-specific severity measure. Differences in RAF should not be interpreted as complete clinical risk adjustment.
Claims history may differ from the intermediary’s adjudicated dataset because of eligibility, attribution, run-out, product mapping, timing, or other administrative differences.
The observation that this case resembles a prior Arete case is descriptive replication, not proof that the same mechanism explains all provider profiles.
Executive Reviewer Approval: Approved by Dave Elton, Executive Reviewer, September 21, 2026.
Data Source: De-identified operational claims data. Provider, practice, insurer, and patient identifiers are omitted from this publication; the internal reproducibility file retains source-level linkage required for appeal review.
AI Use Disclosure: AI-assisted analysis and drafting were used to recompute source data, structure the sensitivity analysis, generate draft visualizations, and prepare this document. Quantitative findings were verified by the Arete Executive Reviewer prior to release. Figure 1 (In Pictures) was human-generated.
Scope: This brief supports a provider-specific tiering appeal. It intentionally does not assess the intermediary’s profiling methodology, benchmark validity, or decision-making process.
Recommended Citation: Arete Healthcare. Different Practice, Same Pattern: A Second Provider-Level Sensitivity Analysis of Spinal CMT Utilization. Arete Analytic Brief ARETE-AB-2026-002, Version 1.0. September 2026.
1. Arete Healthcare. Athena EHR/RCM provider-profile claim-line extract, complete 12-position ICD-10 coding; analysis period September 1, 2025–August 31, 2026. Internal de-identified analytic source; not publicly distributed.
2. Arete Healthcare. Provider-Level 98941 Sensitivity Analysis, group practice, September 2026. Internal reproducibility workbook; not publicly distributed.
3. Arete Healthcare. Spend $3,500 to “Save” $125?! A Sensitivity Analysis of Spinal CMT Utilization in a Provider Tiering Appeal. Arete Analytic Brief ARETE-AB-2026-001, Version 0.7. September 2026.