Unaffected Comparison Groups Separate Program-Specific Provider Growth From System-Wide Enrollment Surges
Comparing treated and comparison care lines to separate modeled policy shifts from broad enrollment waves.
Medicaid agencies evaluate new service rules during periods when statewide enrollment and renewal policies also shift. Between 2019 and 2024, continuous coverage rules lifted patient volume across almost all Medicaid care lines at once. Reviewing one treated service line alone mixes those system-wide waves with the program effect.
When evaluating a Medi-Cal service rollout across 84 months (2018 to 2024), comparing telehealth-eligible care against other Medicaid care isolates a modeled net +5.62 percentage point provider gain through 2023 and +4.60 percentage points through 2024.[1] A single-group pre-post check credits all +24.33% growth to telehealth without subtracting the +18.71% enrollment wave that lifted other care at the same time.[2] Pairing the two-group check with the +1.21-point pre-2020 base gap bounds the budget forecast.[3]
Table 1. Two-Group Medicaid Provider Panel Across 4 Study Windows (2018 to 2024)
Yearly means of monthly billing providers from HHS Medicaid claims with modeled two-group gaps and base-trend checks.
| Two-Group Measure |
Pre-2020 Check
Base 2018 to 2019
|
Peak Window
2019 to 2023
|
Full Post Window
2019 to 2024
|
Renewal Year
2023 to 2024
|
|---|---|---|---|---|
| Video Care Monthly Providers | 84,422 -> 91,399 | 91,399 -> 113,638 | 91,399 -> 101,978 | 113,638 -> 101,978 |
| Video Care Provider Change | +6,977+8.26% | +22,239+24.33% | +10,579+11.57% | -11,660-10.26% |
| Other Care Monthly Providers | 101,905 -> 109,087 | 109,087 -> 129,492 | 109,087 -> 116,690 | 129,492 -> 116,690 |
| Other Care Change | +7,182+7.05% | +20,405+18.71% | +7,603+6.97% | -12,802-9.89% |
| Net Two-Group Gap* | +1.21 ppbase gap | +5.62 pp | +4.60 pp | -0.37 pp |
| Gap Minus Pre-2020 Slope* | Base ref0.00 | +0.78 ppafter 4 yrs | -1.45 ppafter 5 yrs | -1.58 ppafter 1 yr |
| Medicaid Paid: Video vs. Other | $19.23B->$23.56B$85.47B->$99.06B | $23.56B->$40.29B$99.06B->$153.37B | $23.56B->$34.93B$99.06B->$145.72B | $40.29B->$34.93B$153.37B->$145.72B |
Table 1 Notes & * Modeled Estimates
Expand notes
Table 1 Notes & * Modeled Estimates. Provider counts and paid dollars come from the HHS Medicaid Provider Spending file (84 months, Jan 2018 to Dec 2024). Mode A rows subtract the other-care % change from the video-care % change and test extending the +1.21 pp 2018-to-2019 gap.
Column Scope. Base 2018 to 2019. Before 2020 Pay Rules. 2019 to 2023. Stay-Enrolled Years. 2019 to 2024. Includes 2024 Renewal Year. 2023 to 2024. Renewal Restart Window.
Two-Ruler Benchmark Summary. Mode A Two-Group Check on HHS Medicaid Provider Spending (84 Months, Jan 2018 to Dec 2024). Ruler 1: Single-Group Pre-Post Change (Includes Shared Medicaid Waves): +22,239 Providers (+24.33%). Video care in Medicaid averaged 91,399 monthly billing providers in 2019 and 113,638 in 2023 (+22,239, +24.33%). That raw jump mixes broad enrollment growth, stay-enrolled rules, and any video pay shift into one count. Ruler 2: Mode A Two-Group Gap vs. Other Medicaid Care: +5.62 pp (2023); +4.60 pp (2024). In the Mode A two-group check, subtracting the other-care gain (+18.71%) leaves +5.62 pp (24.33 − 18.71) through 2023 and +4.60 pp (11.57 − 6.97) through 2024. The 2018 to 2019 growth gap was already +1.21 pp (8.26 − 7.05). Part of the post-2019 gap may continue that earlier trend.
Method & Data Audit (TooEarlyToSay.com): When Difference-in-Differences Fits → · Medicaid Provider Spending File Guide →
Table 2. Shared Shocks, Base Trends & Budget Fit
How subtracting shared enrollment waves changes how much growth budget teams credit to policy.
| Choice Factor | Base 2018 to 2019 | 2019 to 2023 | 2019 to 2024 | 2023 to 2024 |
|---|---|---|---|---|
| Shared Shift in Other Care | Both care lines grew before 2020, staying within 1.21 points. | Other care growth (+18.71%) matches 77% of the +24.33% gain, leaving +5.62 pp. | Through 2024, other care growth matches 60% of the +11.57% gain. | Both groups dropped by ~10% in 2024 from shared claim lags. |
| Pre-2020 Base Trend Check | Video care grew 1.21 points faster before 2020, so the two lines were not flat. | Extending the 1.21-point yearly gap for 4 years leaves +0.78 pp of the +5.62 pp gap. | Extending the gap for 5 years turns the +4.60 pp gap into -1.45 pp in 2024. | One pre-2020 year gives just two yearly means, so monthly leads give a fuller check. |
| Budget Tradeoff | Budget teams can show the net gap under both zero and 1.21-point pre-2020 drift. | Funding expansion on +24.33% alone credits video rules with broad Medicaid growth. | Paid dollars rose in lockstep, leaving a +1.16 pp paid gap. | Reading the 2024 dip as a video drop-off cuts a service line whose net gap is -0.37 pp. |
| Best-Fit Policy Rule | Add Earlier Base Months. Include pre-2018 months or monthly leads to test if the 1.21-point gap stays steady. | Show Net Gap Next to Raw Gain. Present the +5.62 pp net gap beside the +24.33% raw change in budget briefs. | Bound the 2024 Net Estimate. Report the 2024 net gap under flat trends and extended drift. | Wait for Late-2024 Claims. Refresh Nov and Dec 2024 after claims clear before reading the 2024 dip as real. |
Table 2 Notes: Shared Shock Timing & Two-Group Math
Expand notes
• Mode A Two-Group Math (2019 to 2023, HHS Medicaid Provider Spending yearly means): ΔVideo = (113,638 − 91,399) / 91,399 = +24.33%. ΔOther = (129,492 − 109,087) / 109,087 = +18.71%. Net gap = 24.33 − 18.71 = +5.62 pp. Through 2024: (+11.57%) − (+6.97%) = +4.60 pp. Pre-2020 (2018 to 2019): (+8.26%) − (+7.05%) = +1.21 pp.
• Late-2024 Claim Lags: In the HHS Medicaid Provider Spending series, billing providers drop in Nov and Dec 2024 for both groups (video 88,364 and 47,562; other 101,476 and 57,878). Jan to Oct 2024 ranged from 101,691 to 119,937 and 118,399 to 130,486. Using Jan to Oct 2024 means (108,781 video, +19.02% vs. 2019; 124,092 other, +13.76%) yields a 2019-to-2024 net gap of +5.26 pp.
• Extended Cell Context. Shared Shift in Other Care (Base 2018 to 2019). Both care lines grew before 2020 (+8.26% vs. +7.05%), staying within 1.21 points. · Shared Shift in Other Care (2019 to 2024). Through 2024, other care growth (+6.97%) matches 60% of the +11.57% gain (+4.60 pp). · Shared Shift in Other Care (2023 to 2024). Both groups dropped by ~10% in 2024 (-10.26% vs. -9.89%) from shared claim lags. · Budget Tradeoff (2019 to 2024). Paid dollars rose in lockstep (+48.26% vs. +47.10%), leaving a +1.16 pp paid gap. · Best-Fit Policy Rule (2019 to 2024). Report the 2024 net gap under flat trends (+4.60 pp) and extended drift (-1.45 pp).
How Two-Group Comparisons Guide Budget Choices
During broad shifts like the 2020 to 2024 enrollment wave and Medicaid redetermination, almost every Medicaid count moves at once. If an agency credits telehealth policy with the raw +24.33% rise in billing providers, it cannot tell if that gain came from new billing rules or a system-wide enrollment surge.
A two-group comparison (Side-by-Side Group Comparison, or DiD) uses a second care line on the same timeline, here other Medicaid care (+18.71%), to subtract the shared shift. That step changes the budget takeaway. The +24.33% raw gain narrows to +5.62 pp through 2023 and +4.60 pp through 2024. Extending the +1.21-point pre-2020 gap trims it to +0.78 pp and -1.45 pp. The result rests on how closely the two groups tracked before the rule change.
Data Sources, Model Inputs, Notes & Endnotes
(Numbered endnotes [1]–[5], state data sources & math rules)
Expand details
Case Study Focus: Program Impact · Two-Group Checks & Shared Shocks (Mode A on HHS Medicaid Provider Spending)
Numbered Endnotes & Main Sources
- [1]Two-Group Medicaid Provider Spending Panel (2018-2024): U.S. HHS Medicaid Provider Spending File monthly counts (84 months). Billing providers for telehealth-eligible care grew +24.33% (91,399 to 113,638). Other Medicaid care grew +18.71% (109,087 to 129,492), leaving a +5.62 pp net gap through 2023 and +4.60 pp through 2024.
- [2]Pre-2020 Base Trend Check (2018-2019): Billing providers for video care grew +8.26% (84,422 to 91,399) versus +7.05% (101,905 to 109,087) for other care, a +1.21 pp yearly pre-2020 gap.
- [3]Medicaid Continuous Coverage & Telehealth Rules: CMS T-MSIS Analytic Files Documentation and California DHCS Telehealth Policy Guidance.
- [4]Medicaid.gov Open Data
- [5]California DHCS Telehealth Policy Guidance
Mode Classification (Mode A Two-Group Check on Observed Series): Billing provider counts and paid dollars come from the HHS Medicaid Provider Spending file (telehealth_month_summary.parquet, Jan 2018 to Dec 2024), split by whether a code allows telehealth billing. Net two-group gaps and pre-trend bounds are Mode A calculations on these observed series. Unit of analysis: care-group-month (2 groups × 84 months = 168 rows).
Primary Data Sources: U.S. HHS Medicaid Provider Spending File, Medicaid.gov Open Data, CMS T-MSIS Analytic Files Documentation, and California DHCS Telehealth Policy Guidance.
Companion Method & Code Guides (TooEarlyToSay.com)
- Using difference-in-differences in practice. When a comparison group can stand in for the treated group, which assumptions matter, and worked cases labeled by whether their materials are public or article-reported.
- What 238 Million Medicaid Billing Rows Can Support. Row grain of the HHS Medicaid provider-spending file (aggregated billing cells by provider, procedure code, and month) and its suppression rules for multi-period comparisons.
Common Questions & Method Notes
- How does a two-group comparison (DiD) separate policy shifts from a shared wave?
- It subtracts the before-and-after change in a comparison group hit by the same wave (other Medicaid care, +18.71% from 2019 to 2023) from the change in the treated group (video care, +24.33%). That leaves a +5.62 percentage-point net gap.
- Why does the 2019-to-2024 net gap (+4.60 pp) differ from the 2019-to-2023 gap (+5.62 pp)?
- From 2023 to 2024, video-care providers fell 10.26% and other-care providers fell 9.89%. Both drops reflect Medicaid redetermination and unpaid Nov and Dec 2024 claims.
- What must hold for other Medicaid care to work as a valid comparison group?
- Both care groups must follow parallel paths when no new video rule is enacted. In 2018 to 2019, the treated group grew 1.21 points faster. Extending that pre-2020 gap trims the 2019-to-2023 net gap to +0.78 pp.