Table 3. Estimated Effects of the Million Hearts Model on Primary and Secondary Long-Term Study Outcomes.
| Outcome | Intervention | Control | Difference | Relative change or effect, %a | P value | |
|---|---|---|---|---|---|---|
| Unadjustedb | Adjusted (90% CI) | |||||
| Health events (No. per 1000 person-years) | ||||||
| High- and medium-risk beneficiaries | ||||||
| First-time myocardial infarctions, strokes, and TIAs (from Medicare claims only) (primary outcome) | 15.8 | 17.0 | HR, 0.92 | HR, 0.97 (0.93 to 1.0) | −3.3 | 0.09 |
| First-time myocardial infarctions, strokes, or TIAs, including deaths due to CHD or cerebrovascular disease | 18.7 | 20.3 | HR, 0.92 | HR, 0.96 (0.93 to 0.99) | −4.2 | 0.02 |
| All-cause deaths | 28.0 | 29.7 | HR, 0.94 | HR, 0.96 (0.93 to 0.98) | −4.3 | 0.01 |
| High-risk beneficiaries | ||||||
| First-time myocardial infarctions, strokes, and TIAs (from Medicare claims only) (primary outcome) | 21.3 | 22.7 | HR, 0.94 | HR, 0.99 (0.94 to 1.03) | −1.3 | 0.63 |
| First-time myocardial infarctions, strokes, or TIAs, including deaths due to CHD or cerebrovascular disease | 25.7 | 27.3 | HR, 0.94 | HR, 0.98 (0.94 to 1.02) | −2.0 | 0.45 |
| All-cause deaths | 38.4 | 39.3 | HR, 0.97 | HR, 0.99 (0.95 to 1.03) | −0.8 | 0.72 |
| Medium-risk beneficiaries | ||||||
| First-time myocardial infarctions, strokes, and TIAs (from Medicare claims only) (primary outcome) | 13.3 | 14.6 | HR, 0.92 | HR, 0.95 (0.91 to 0.99) | −4.7 | 0.06 |
| First-time myocardial infarctions, strokes, or TIAs, including deaths due to CHD or cerebrovascular disease | 15.7 | 17.3 | HR, 0.91 | HR, 0.94 (0.91 to 0.98) | −5.6 | <0.001 |
| All-cause deaths | 23.5 | 25.4 | HR, 0.92 | HR, 0.93 (0.90 to 0.96) | −6.8 | 0.02 |
| Medicare spending ($ per beneficiary per mo) | ||||||
| High- and medium-risk beneficiaries | ||||||
| Medicare spending on first-time myocardial infarctions, strokes, and TIAs and 90-d follow-up (primary outcome) | $37.56 | $40.72 | −$3.16 | −$1.83 (−$3.97 to $0.30) | −4.7 | 0.16 |
| Total Medicare Parts A and B spending before model payments | $958.50 | $954.82 | $3.69 | $0.87 (−$17.91 to $19.64) | 0.1 | 0.94 |
| Total Medicare Parts A and B spending including model paymentsc | $959.75 | $954.82 | $4.93 | $2.11 (−$16.66 to 20.89) | 0.2 | 0.85 |
| High-risk beneficiaries | ||||||
| Medicare spending on first-time myocardial infarctions, strokes, and TIAs and 90-d follow-up (primary outcome) | $50.92 | $55.78 | $4.86 | −$3.09 (−7.33 to 1.15) | 5.7 | 0.23 |
| Total Medicare Parts A and B spending before model payments | $1104.50 | $1088.32 | $16.18 | $9.93 (−$18.54 to +$38.40) | 0.9 | 0.57 |
| Total Medicare Parts A and B spending including model paymentsc | NAc | NAc | NAc | NAc | NAc | NAc |
| Medium-risk beneficiaries | ||||||
| Medicare spending on first-time myocardial infarctions, strokes, and TIAs and 90-d follow-up (primary outcome) | $31.36 | $33.84 | $2.48 | −$1.30 (−3.66 to 1.07) | −4.0 | 0.37 |
| Total Medicare Parts A and B spending before model payments | $893.76 | $895.61 | $1.85 | −$3.22 (−21.08 to 14.65) | −0.4 | 0.77 |
| Total Medicare Parts A and B spending including model paymentsc | NAc | NAc | NAc | NAc | NAc | NAc |
| Health care service use (No. per 1000 beneficiaries per year) | ||||||
| High- and medium-risk beneficiaries | ||||||
| All-cause hospitalizations | 255.3 | 251.9 | 3.4 | 9.0 (3.8 to 14.3) | 3.7 | 0.005 |
| Circulatory system-related hospitalizations | 56.0 | 57.0 | −1.1 | 0.8 (−1.0 to 2.6) | 1.4 | 0.48 |
| All-cause ED visits | 386.3 | 382.8 | 3.5 | 8.2 (−1.1 to 17.6) | 2.2 | 0.15 |
| Circulatory system–related ED visits | 31.6 | 33.0 | −1.4 | 0.1 (−1.7 to 1.9) | 0.4 | 0.91 |
| Office visits | 10 544.7 | 10 173.6 | 371.1 | 101.1 (−55.5 to 257.7) | 1.0 | 0.29 |
| High-risk beneficiaries | ||||||
| All-cause hospitalizations | 309.2 | 302.4 | 6.8 | 12.3 (3.5 to 21.1) | 4.1 | 0.02 |
| Circulatory system-related hospitalizations | 75.6 | 75.6 | 0.1 | 2.3 (−0.7 to 5.3) | 3.1 | 0.21 |
| All-cause ED visits | 422.7 | 416.2 | 6.5 | 11.9 (0.2 to 23.6) | 2.9 | 0.09 |
| Circulatory system–related ED visits | 38.4 | 39.3 | 0.9 | 1.2 (−1.2 to 3.5) | 3.1 | 0.41 |
| Office visits | 11 312.4 | 10 875.2 | 437.3 | 138.0 (−57.6 to 333.6) | 1.2 | 0.25 |
| Medium-risk beneficiaries | ||||||
| All-cause hospitalizations | 231.4 | 229.5 | 1.9 | 7.5 (2.3 to 12.7) | 3.4 | 0.02 |
| Circulatory system–related hospitalizations | 47.2 | 48.8 | −1.6 | 0.1 (−1.7 to 2.0) | 0.2 | 0.92 |
| All-cause ED visits | 370.1 | 368.0 | 2.2 | 7.4 (−2.7 to 17.5) | 2.0 | 0.23 |
| Circulatory system–related ED visits | 28.6 | 30.2 | −1.6 | −0.4 (−2.2 to 1.4) |
−1.4 | 0.71 |
| Office visits | 10 204.2 | 9862.5 | 341.7 | 81.6 (−68.8 to 232.0) | 0.8 | 0.37 |
Abbreviations: CHD, coronary heart disease; CMS, Centers for Medicare & Medicaid Services; ED, emergency department; NA, not applicable; TIA, transient ischemic attack.
The relative change is the estimated relative increase or decrease in the outcome associated with the Million Hearts Model, expressed as a percentage. For the health-event outcomes, hazard ratios are already on a relative scale rather than an absolute scale. The relative change as a percentage effect is calculated as (HR − 1) × 100%. For outcomes of Medicare spending or health care service use, the unadjusted and adjusted differences are on the absolute scale, so the relative change as a percentage effect is calculated from the adjusted difference relative to the outcome mean. For example, the percentage effect for Medicare spending on first-time myocardial infarctions, strokes, and TIAs (a primary outcome) among high- and medium-risk beneficiaries is the adjusted effect (−$1.83) divided by the adjusted control group mean, where the adjusted control group mean equals the intervention group mean minus the adjusted difference (−$1.83/[$37.56 – −$1.83] = −0.047 = −4.7%). CIs are not provided for relative change, but can be inferred from the CIs provided for the “Adjusted (90% CI)” column.
Unadjusted differences are calculated before rounding the intervention or control group rates or means. As a result, numbers might not sum across rows exactly.
Model payments included (1) the risk assessment payments CMS paid for intervention group beneficiaries entering the model through December 2018 (all risk groups), as risk assessments were needed to identify high- and medium-risk beneficiaries; (2) cardiovascular care management payments for intervention group high-risk beneficiaries, which CMS paid only in 2017; and (3) risk reduction payments CMS paid through December 2021 for intervention group high-risk beneficiaries who entered the model through December 2018. All together, these payments summed to $7.2 million, and we divided this total payment evenly across high- and medium-risk beneficiaries and their follow-up months to estimate means and effects. We did not estimate effects separately for either the high-risk or the medium-risk beneficiaries, given the risk-assessment costs were incurred by the population as a whole.