# Diastolic Blood Pressure Task Card Task id: `dbp` Evidence status: grounded Evidence reviewed: 2026-07-19 Benchmark status: complete ## Target Diastolic Blood Pressure as a continuous participant-visit measurement. Implementation facts: - bodily system: cardiovascular/lipid - target route: `blood_pressure.sitting_blood_pressure_diastolic` - unit: mmHg - device/acquisition: OMRON HEM-RML31; backup/manual source is not row-labelled - construction/QC: canonical first sitting reading after rest; participant-stage mean; documented BP QC - valid range: 10-180 mmHg - default stages: `00_00_visit` and `02_00_visit` - paper fidelity: exact HealthFormer endpoint ## Why This Matters DBP reflects arterial pressure during cardiac relaxation and adds information distinct from SBP; it carries independent cardiovascular prognostic value particularly in middle-aged adults, while systolic pressure dominates risk in older adults (Flint et al. 2019). Prediction does not diagnose hypertension or establish treatment need. ## Related HPP Measurements SBP, pulse pressure, MAP, pulse rate, orthostatic readings, medications, BMI, and vascular measures. ## Clinically Meaningful Variants Continuous prediction is primary. Guideline diastolic thresholds differ across bodies (ACC/AHA 2017 places stage 1 hypertension at DBP 80-89 mmHg and stage 2 at >=90 mmHg; ESH 2023 and ESC 2024 place hypertension at DBP >=90 mmHg), and a hypertension diagnosis generally rests on more than a single office reading -- typically repeated readings and often out-of-office (home or ambulatory) confirmation. This single-visit participant-mean DBP is therefore not a hypertension label, and diagnostic bins are not implied by this Task. ## Baselines And Ceilings For V1→V2 evaluation, compare age/sex/BMI demographic floor, prior-value LOCF, OLS on prior value plus demographics, and GBDT on the same inputs. Use one cohort and split for every comparison. ## Caveats - Mask the full same-visit blood-pressure family, including repeat/contralateral DBP, SBP, MAP, pulse pressure, and derived hypertension states. - Cuff size, posture, rest, medication, acute state, and unlabelled backup/manual measurements affect DBP. - A good score supports prediction of this measurement only; it does not establish diagnosis, causality, or intervention response. ## Metrics Primary R2. Also report RMSE, MAE, Pearson r, Spearman rho, split Ns, target quantiles, and same-cohort deltas versus LOCF. ## References - Lutsker et al. (2026), HealthFormer, . - Leiherer et al. (2024), earlier-life SBP/DBP and cardiovascular risk, . - Whelton et al. (2017), ACC/AHA high-blood-pressure guideline (DBP thresholds), . - Mancia et al. (2023), ESH guidelines for the management of arterial hypertension, . - McEvoy et al. (2024), ESC guidelines for elevated blood pressure and hypertension, . - Flint et al. (2019), systolic and diastolic BP and cardiovascular outcomes, . - HPP documentation `blood_pressure` dataset documentation and QC inventory. ## Execution Evidence - protocol: paired HPP V1→V2; every baseline uses the same eligible participants - common cohort: n=5901; train=4077, validation=928, test=896 | Baseline | Strategy | Test R2 | RMSE | Pearson r | | --- | --- | ---: | ---: | ---: | | Demographic | `demographic_floor` | 0.133 | 9.961 | 0.366 | | Linear | `linear` | 0.401 | 8.280 | 0.634 | | Trees | `gbdt_delta_r2` | 0.400 | 8.290 | 0.633 | | LOCF | `last_observation_carried_forward` | 0.135 | 9.948 | 0.622 | ## Benchmark-Track Evidence Benchmark-Track Evidence reviewed: 2026-07-29. HPP and literature review supports V1 sleep and fundus morphology beyond age/sex/BMI. The full same-visit BP and vascular family must remain excluded. See `04-cardiovascular-lipid.md`. ## v0 Final-Test Status The classical release is complete: Ridge/logistic and GBDT are remotely verified on every approved exact cohort and rendered in the Figure 1 task deep dives. ## Open Questions - Domain-expert review remains outstanding; evidence grounding and real-run gates are complete.