Health App Roundup: What's New March 2026

Calm's biggest feature release of Q1 2026 is a structured 8-week anxiety management program co-developed with licensed clinical psychologists. Unlike the app's standard guided meditations, the new Clinical Anxiety Program follows a CBT (cognitive behavioral therapy) structure: psychoeducation in weeks 1–2, cognitive restructuring exercises in weeks 3–5, and behavioral activation and maintenance in weeks 6–8. Each session runs 15–25 minutes. Calm cites a 78% participant completion rate in the beta cohort of 3,200 users, alongside statistically significant reductions in GAD-7 anxiety scores at completion. This positions Calm's premium tier more squarely in competition with dedicated mental health apps like Woebot and Headspace. Premium required; no change to pricing this month.

Oura's March software update adds persistent temperature trend monitoring that flags multi-day deviations from individual baseline rather than single-night outliers. The previous algorithm could be triggered by a warm room or hormonal cycle fluctuations; the new version uses a rolling 5-day baseline with deviation scoring that better distinguishes genuine immune response from environmental noise. The update also improves integration with the Readiness Score — temperature deviations now explicitly suppress Readiness and add a contextual explanation in the app. Separately, Oura updated its period cycle prediction algorithm with additional mid-cycle accuracy improvements it says reduce phase misclassification by 12%.

WHOOP's March update introduces a weekly Cardiovascular Load (CVL) metric that aggregates total cardiac stress across all tracked activities into a single weekly summary. The CVL score accounts for duration, intensity zone distribution, and recovery between sessions — producing a value that allows comparison of week-over-week training stress regardless of activity type. For multi-sport athletes managing run, bike, and strength training loads, this is a genuinely useful high-level metric that neither WHOOP's existing Strain scores nor individual session data provides cleanly. The update also surfaces CVL trend lines in the Health Monitor, providing context for sustained fatigue or overtraining indicators.

Strava's long-awaited dynamic training plans feature launched in March 2026 for running and cycling. Unlike the static plans Strava has offered since 2023, AI Training Plans now adapt week-to-week based on what you actually completed: if you missed Tuesday's interval session, the plan adjusts Wednesday's easy run and the following week's intensity to account for the gap. The adaptation also incorporates health data from connected wearables — if WHOOP or Apple Watch flags elevated stress or low HRV, the plan will suggest reducing intensity. The feature is available with Strava Premium ($7.99/month). Beta testing across 45,000 runners showed plan adherence 23% higher for adaptive plans versus static equivalents.

Apple's March iOS 19.4 update added automatic Health Trends summarization via Siri, which now proactively generates a monthly health summary from Apple Health data — covering sleep trends, heart rate variability, activity levels, and medication adherence if applicable — and surfaces it as a weekly Siri suggestion. The summaries use on-device processing for privacy and can be reviewed at any time by asking Siri 'How has my health been this month?' The feature is limited to Apple Watch users and requires Health data sharing to be enabled. Early testing suggests the summaries are high-level rather than clinically detailed, but they provide useful month-over-month pattern highlights for users who don't actively monitor their Health app.

PlateLens updated its restaurant menu recognition engine this month, adding expanded coverage of regional and fast-casual chains across the Southeast and Midwest US. The update brings the app's restaurant database to 45,000+ menu items across 380+ chains. Of note to users: the update specifically improved recognition accuracy for bowls and composed dishes from regional chains — historically a weak spot where visual similarity between items reduced accuracy. PlateLens also added a 'recent restaurant' quick-log feature that surfaces previously visited restaurants at the top of the search interface, reducing the steps required to log repeat restaurant visits.