Table 3.
Multimorbidity interaction sensitivity model, MEPS 2015, 2016, 2018, and 2022
| Variable | Additive model | Interaction model | ||||
|---|---|---|---|---|---|---|
| β | (SE) | p | β | (SE) | p | |
| Intercept | 0.8671 | (0.0026) | < 0.001 | 0.8666 | (0.0026) | < 0.001 |
| Age group (ref: 18–29 years) | ||||||
| 30–39 years | − 0.0140 | (0.0015) | < 0.001 | − 0.0140 | (0.0015) | < 0.001 |
| 40–49 years | − 0.0230 | (0.0016) | < 0.001 | − 0.0228 | (0.0016) | < 0.001 |
| 50–59 years | − 0.0370 | (0.0017) | < 0.001 | − 0.0368 | (0.0017) | < 0.001 |
| 60–69 years | − 0.0428 | (0.0019) | < 0.001 | − 0.0428 | (0.0019) | < 0.001 |
| 70–79 years | − 0.0483 | (0.0022) | < 0.001 | − 0.0483 | (0.0022) | < 0.001 |
| ≥80 years | − 0.0894 | (0.0031) | < 0.001 | − 0.0895 | (0.0031) | < 0.001 |
| Sex (ref: Female) | ||||||
| Male | 0.0172 | (0.0009) | < 0.001 | 0.0172 | (0.0009) | < 0.001 |
| Race/Ethnicity (ref: non-hispanic white) | ||||||
| Non-Hispanic Black | 0.0150 | (0.0016) | < 0.001 | 0.0149 | (0.0016) | < 0.001 |
| Hispanic | 0.0245 | (0.0015) | < 0.001 | 0.0244 | (0.0015) | < 0.001 |
| Non-Hispanic Asian/Pacific Islander | 0.0078 | (0.0021) | < 0.001 | 0.0079 | (0.0021) | < 0.001 |
| Non-Hispanic Other | − 0.0082 | (0.0031) | 0.008 | − 0.0081 | (0.0030) | 0.008 |
| Education (ref: No degree) | ||||||
| High school diploma / GED | 0.0147 | (0.0016) | < 0.001 | 0.0146 | (0.0016) | < 0.001 |
| Other post-secondary degree | 0.0207 | (0.0026) | < 0.001 | 0.0206 | (0.0026) | < 0.001 |
| Bachelor’s degree | 0.0341 | (0.0019) | < 0.001 | 0.0341 | (0.0019) | < 0.001 |
| Master’s or Doctorate degree | 0.0442 | (0.0021) | < 0.001 | 0.0443 | (0.0021) | < 0.001 |
| Income (ref: Poor, < 100% FPL) | ||||||
| Near poor (100–124%) | 0.0207 | (0.0034) | < 0.001 | 0.0207 | (0.0034) | < 0.001 |
| Low income (125–199%) | 0.0342 | (0.0025) | < 0.001 | 0.0343 | (0.0025) | < 0.001 |
| Middle income (200–399%) | 0.0540 | (0.0021) | < 0.001 | 0.0540 | (0.0021) | < 0.001 |
| High income (≥ 400%) | 0.0703 | (0.0021) | < 0.001 | 0.0704 | (0.0021) | < 0.001 |
| Cardiometabolic conditions | ||||||
| Type 2 diabetes | − 0.0411 | (0.0022) | < 0.001 | − 0.0439 | (0.0041) | < 0.001 |
| Heart disease | − 0.0500 | (0.0018) | < 0.001 | − 0.0436 | (0.0024) | < 0.001 |
| High blood pressure | − 0.0313 | (0.0014) | < 0.001 | − 0.0312 | (0.0021) | < 0.001 |
| High cholesterol | − 0.0232 | (0.0013) | < 0.001 | − 0.0256 | (0.0016) | < 0.001 |
| Stroke/CVA | − 0.0710 | (0.0035) | < 0.001 | − 0.0709 | (0.0035) | < 0.001 |
| Obesity (BMI ≥ 30) | − 0.0143 | (0.0018) | < 0.001 | − 0.0120 | (0.0019) | < 0.001 |
| Body mass index | ||||||
| BMI − 25 kg/m2 (linear) | − 0.0014 | (0.0002) | < 0.001 | − 0.0014 | (0.0002) | < 0.001 |
| (BMI − 25)2 (quadratic) | − 0.000017 | (0.000014) | 0.233 | − 0.000017 | (0.000014) | 0.236 |
| Two-way interactions (interaction model only) | ||||||
| High blood pressure × High cholesterol | – | – | – | 0.0051 | (0.0027) | 0.060 |
| High blood pressure × Obesity | – | – | – | − 0.0035 | (0.0023) | 0.127 |
| Heart disease × High blood pressure | – | – | – | − 0.0112 | (0.0033) | < 0.001 |
| Type 2 diabetes × high blood pressure | – | – | – | 0.0092 | (0.0046) | 0.044 |
| Type 2 diabetes × Obesity | – | – | – | − 0.0071 | (0.0041) | 0.083 |
Survey-weighted linear regression (svyglm, Gaussian family) pooling MEPS 2015, 2016, 2018, and 2022. Sampling weights divided by 4. Additive model: Model 4 specification (sociodemographic covariates + six cardiometabolic conditions + BMI + BMI2) with no interaction terms. Interaction model: same specification plus five two-way interaction terms for the most prevalent co-occurring condition pairs. Positive interaction coefficient, sub-additive harm (co-occurrence less detrimental than sum of individual effects); negative, super-additive harm; β, regression coefficient (adjusted EQ-5D utility difference); SE, Standard error; FPL, Federal poverty level