Table 3.
Associations between high-risk results, provider recommendations, and patient-reported health behavior changes.
| Time point/behavior | Monogenic disease risk (ref: none) OR (95% CI) | p | Cardiometabolic risk prediction (per condition in 80th percentile risk) OR (95% CI) | p |
|---|---|---|---|---|
| Recommendations | ||||
| Any change | 0.59 (0.15–1.84) | 0.391 | 1.68 (1.07–2.72) | 0.027 |
| Health behavior | 0.79 (0.21–2.50) | 0.702 | 1.41 (0.88–2.32) | 0.161 |
| Medication | 1.51 (0.19–8.41) | 0.652 | 1.99 (0.94–4.97) | 0.097 |
| Changes at 6 weeks | ||||
| Any change | 1.72 (0.62–4.79) | 0.292 | 1.49 (1.00–2.23) | 0.049 |
| Exercise | 1.94 (0.66–5.74) | 0.228 | 1.54 (0.99–2.41) | 0.057 |
| Diet | 1.60 (0.55–4.67) | 0.387 | 1.27 (0.83–1.96) | 0.266 |
| Supplements | 1.73 (0.30–9.92) | 0.534 | 1.25 (0.54–2.86) | 0.602 |
| Medications | 0.37 (0.04–3.21) | 0.366 | 1.28 (0.64–2.56) | 0.474 |
| Other | NAa | 0.568 | 0.48 (0.08–2.94) | 0.424 |
| Changes at 6 months | ||||
| Any change | 2.11 (0.74–6.03) | 0.160 | 1.52 (1.04–2.24) | 0.032 |
| Exercise | 2.39 (0.85–6.72) | 0.098 | 1.59 (1.05–2.40) | 0.029 |
| Diet | 1.88 (0.67–5.27) | 0.227 | 1.34 (0.91–1.98) | 0.134 |
| Supplements | 0.79 (0.08–7.39) | 0.832 | 1.23 (0.59–2.57) | 0.571 |
| Medications | 0.29 (0.03–2.49) | 0.258 | 1.50 (0.83–2.72) | 0.175 |
| Other | 2.03 (0.16–25.50) | 0.578 | 1.06 (0.32–3.53) | 0.918 |
High-risk results were defined in two ways: (1) identification of an unexpected monogenic disease risk (analyses exclude two patients who were previously diagnosed with hemochromatosis, who received monogenic disease risk findings on their GS report) and (2) number of polygenic predictions about cardiometabolic traits (range: 0–8) where patients were identified to be in the 80th percentile for genetic risk, per polygenic risk predictions. Means and 95% confidence intervals of odds ratios were estimated using logistic regression or generalized estimating equations using logit linking functions and binomial distributions.
aAn odds ratio could not be estimated, because no participant with an unexpected monogenic disease risk reported a change to this category at 6 weeks post disclosure. The p-value was estimated using a Fisher’s exact test on available data.