🧠 Intelligence Layers & Health Scores
Transform raw biomarkers into actionable insights with scores, archetypes, trends, and comparisons
Transform raw biomarkers into actionable insights with scores, archetypes, trends, and comparisons
Raw biomarkers tell you what happened. Steps taken, hours slept, heart rate measured—these are valuable data points, but they're just the beginning.
Intelligence layers tell you what it means. Is this user's sleep quality improving or declining? Are they a "Consistent Early Riser" or "Highly Irregular Sleeper"? How does their activity compare to their demographic group?
By applying advanced analytics and pattern recognition to biomarker data, intelligence layers provide:
This transforms health data from passive monitoring into proactive insights that drive engagement, personalization, and outcomes.
See the comprehensive breakdown of all activity, body, vitals, sleep, and reproductive biomarkers.
📊 View Complete Biomarker Table →Behavioral archetypes and insights represent a fundamentally different kind of intelligence than scores or raw metrics. While a score tells you how much (0.7 sleep quality, 8,247 steps), behavioral intelligence tells you who someone is and how they behave over time.
People don't identify with numbers, they identify with behaviors.
Tell a user "Your sleep score is 0.68" and they nod but don't internalize it. Tell them "You're a Consistent Early Riser—your routine is your superpower" and they recognize themselves. This is behavioral identity: a label that captures lifestyle patterns and creates meaning. When users see their behavioral archetype, they understand not just their current state, but their typical patterns and tendencies.
Prediction requires understanding patterns, not just data points.
A single night's sleep score doesn't predict much. But "Highly Irregular Sleeper"—someone whose sleep varies wildly from night to night—carries meaningful information. This person is more vulnerable to burnout, mental health fluctuations, and disengagement. The archetype captures weeks or months of behavior compressed into a single meaningful label that enables proactive intervention.
Personalization at scale requires automated segmentation.
Without behavioral archetypes, you face a choice: send generic messages to everyone (low engagement) or manually segment users (doesn't scale). Archetypes solve this by automatically categorizing behavioral patterns. "Sedentary" users need gentle encouragement. "Weekend Warriors" need injury prevention guidance. "Inconsistent" users need routine-building support. The archetype becomes the basis for intelligent, automated personalization across your entire user base.
Classification complexity goes beyond simple thresholds.
You can't just say "sleep duration < 6 hours = poor sleeper." Meaningful behavioral archetypes require analyzing regularity (week-to-week consistency), continuity (nightly interruptions), debt (cumulative sleep deficit), circadian alignment (biological clock synchronization), and recovery patterns (physical and mental restoration). Each archetype represents a cluster in this multi-dimensional behavioral space, where boundaries are fuzzy and context-dependent.
Validation demands longitudinal data and clinical expertise.
Creating archetypes that are statistically accurate is straightforward. Creating archetypes that are clinically meaningful and predictively valuable is exceptionally difficult. You need longitudinal data showing how behaviors correlate with outcomes over months or years. You need clinical expertise to ensure archetypes align with established behavioral science. You need validation studies proving that your classifications actually predict what you claim they predict. This is why Sahha's archetypes were developed in partnership with the University of Otago using thousands of participants tracked over extended periods.
Behavioral patterns evolve and models degrade.
A behavioral model trained on pre-pandemic data may not apply to post-pandemic behavior patterns. Seasonal changes affect activity and sleep. Demographic shifts alter what "typical" looks like. Population behavior evolves. Maintaining accurate behavioral intelligence means continuous model updates, retraining on new data, incorporating latest research, and validating that classifications remain predictive. This is an ongoing commitment, not a one-time development effort.
When behavioral archetypes are properly developed and validated, they unlock capabilities that raw biomarkers and simple scores cannot provide:
Behavioral intelligence transforms health data from measurement into understanding. It's the difference between knowing someone walked 8,247 steps today and knowing they're a "Weekend Warrior" who needs weekday consistency coaching. This is the foundation for truly personalized health applications.
360° measures of health and wellness • Normalized 0.0-1.0 range • No wearable required
| Score Name | Range | Key Factors | Description | Wearable? |
|---|---|---|---|---|
| Wellbeing | 0.0-1.0 | steps, active_hours, active_calories, intense_activity_duration, extended_inactivity, floors_climbed, sleep_duration, sleep_regularity, sleep_continuity, sleep_debt, circadian_alignment, physical_recovery, mental_recovery | Holistic measure combining physical, mental, and behavioral health data | 🟢 No |
| Activity | 0.0-1.0 | steps, active_hours, active_calories, intense_activity_duration, extended_inactivity, floors_climbed | Evaluates daily physical activity levels and intensity | 🟢 No |
| Sleep | 0.0-1.0 | sleep_duration, sleep_regularity, sleep_continuity, sleep_debt, circadian_alignment, physical_recovery, mental_recovery | Assesses sleep quality, duration, regularity, and stages | 🟢 No |
| Mental Wellbeing | 0.0-1.0 | steps, active_hours, extended_inactivity, activity_regularity, sleep_regularity, circadian_alignment | Measures mental wellness through behavioral pattern analysis | 🟢 No |
| Readiness | 0.0-1.0 | sleep_duration, physical_recovery, mental_recovery, sleep_debt, walking_strain_capacity, exercise_strain_capacity, resting_heart_rate, heart_rate_variability | Gauges daily readiness and recovery metrics | 🟢 No |
Labels that capture persona and lifestyle • Weekly/Monthly periodicity • Smartphone-only data
| Archetype | Type | Periodicity | Possible Values | Description | Wearable? |
|---|---|---|---|---|---|
| activity_level | Ordinal | Weekly, Monthly | sedentary, lightly_active, moderately_active, highly_active | Overall level of physical activity | 🟢 No |
| exercise_frequency | Ordinal | Weekly, Monthly | rare_exerciser, occasional_exerciser, regular_exerciser, frequent_exerciser | How often the individual exercises | 🟢 No |
| mental_wellness | Ordinal | Weekly, Monthly | poor_mental_wellness, fair_mental_wellness, good_mental_wellness, optimal_mental_wellness | Mental wellness and resiliency | 🟢 No |
| overall_wellness | Ordinal | Weekly, Monthly | poor_wellness, fair_wellness, good_wellness, optimal_wellness | Overall wellbeing across all health aspects | 🟢 No |
| primary_exercise | Categorical | Weekly, Monthly | See possible exercise types | Most frequently performed exercise | 🟢 No |
| primary_exercise_type | Categorical | Weekly, Monthly | strength_oriented, cardio_oriented, mind_body_oriented, hybrid_oriented, sport_oriented, outdoor_oriented | Categorizes primary exercise type | 🟢 No |
| secondary_exercise | Categorical | Weekly, Monthly | See possible exercise types | Second most frequently performed exercise | 🟢 No |
| sleep_duration | Ordinal | Weekly, Monthly | very_short_sleeper, short_sleeper, average_sleeper, long_sleeper | Typical sleep duration relative to norms | 🟢 No |
| sleep_efficiency | Ordinal | Weekly, Monthly | highly_inefficient_sleeper, inefficient_sleeper, efficient_sleeper, highly_efficient_sleeper | Sleep maintenance effectiveness | 🟡 Yes |
| sleep_pattern | Categorical | Weekly, Monthly | consistent_early_riser, inconsistent_early_riser, consistent_late_sleeper, inconsistent_late_sleeper, early_morning_sleeper, chronic_short_sleeper, inconsistent_short_sleeper | Overall sleep behavior patterns | 🟢 No |
| sleep_quality | Ordinal | Weekly, Monthly | poor_sleep_quality, fair_sleep_quality, good_sleep_quality, optimal_sleep_quality | Long-term sleep quality assessment | 🟢 No |
| sleep_regularity | Ordinal | Weekly, Monthly | highly_irregular_sleeper, irregular_sleeper, regular_sleeper, highly_regular_sleeper | Consistency in sleep timings | 🟢 No |
| bed_schedule | Ordinal | Weekly, Monthly | very_early_sleeper, early_sleeper, late_sleeper, very_late_sleeper | Typical bedtime patterns | 🟢 No |
| wake_schedule | Ordinal | Weekly, Monthly | very_early_riser, early_riser, late_riser, very_late_riser | Typical wake-up time patterns | 🟢 No |
Detect directional change over time • Increasing/Decreasing/Stable states • 5 scores + 17 factors
Trends analyze the last 4 complete weeks of data on a rolling basis. For each metric, the system computes weekly averages, compares them to previous weeks, and classifies directional movement as:
Each trend includes: rolling 4-week time window, percent change calculations, state label, and range metadata.
| Category | Name | Description | Unit | Higher = Better? |
|---|---|---|---|---|
| SCORES (5 total) | ||||
| score | sleep | Overall sleep quality | index | ✅ Yes |
| score | activity | Physical activity and movement levels | index | ✅ Yes |
| score | readiness | Body's recovery state and preparedness for exertion | index | ✅ Yes |
| score | wellbeing | Holistic health combining sleep and activity | index | ✅ Yes |
| score | mental_wellbeing | Mental wellbeing state based on behavioral patterns | index | ✅ Yes |
| FACTORS (17 total) | ||||
| factor | sleep_duration | Total time spent asleep | index | ✅ Yes |
| factor | sleep_regularity | Consistency of sleep schedule | index | ✅ Yes |
| factor | sleep_continuity | Uninterrupted sleep with minimal awakenings | index | ✅ Yes |
| factor | sleep_debt | Accumulated sleep deficit | index | ✅ Yes |
| factor | circadian_alignment | Alignment with natural sleep-wake cycle | index | ✅ Yes |
| factor | physical_recovery | Deep sleep phase duration | index | ✅ Yes |
| factor | mental_recovery | REM sleep phase duration | index | ✅ Yes |
| factor | steps | Daily step count | index | ✅ Yes |
| factor | active_hours | Hours with significant physical activity | index | ✅ Yes |
| factor | active_calories | Calories burned during activity | index | ✅ Yes |
| factor | intense_activity_duration | Time spent in high-intensity activity | index | ✅ Yes |
| factor | extended_inactivity | Prolonged sedentary periods | index | ✅ Yes |
| factor | floors_climbed | Vertical movement measurement | index | ✅ Yes |
| factor | activity_regularity | Consistency of daily activity patterns | index | ✅ Yes |
| factor | walking_strain_capacity | Capacity to do low-intensity activities | index | — |
| factor | exercise_strain_capacity | Capacity to do high-intensity exercises | index | — |
| factor | resting_heart_rate | Heart rate during rest | index | ✅ Yes |
| factor | heart_rate_variability | Variation in time between heartbeats | index | ✅ Yes |
Contextualize metrics with 3 reference points • Percentile rankings • 5 scores + 6 biomarkers
Comparisons provide three distinct reference points to contextualize metric values:
Each comparison includes: reference group's average value, percentile position, absolute and percentage differences, and descriptive state label (very_low, low, average, high, very_high).
| Category | Name | Description | Unit | Higher = Better? |
|---|---|---|---|---|
| SCORES (5 total) | ||||
| score | sleep | Overall sleep quality | index | ✅ Yes |
| score | activity | Physical activity and movement levels | index | ✅ Yes |
| score | readiness | Body's recovery state and preparedness for exertion | index | ✅ Yes |
| score | wellbeing | Holistic health combining sleep and activity | index | ✅ Yes |
| score | mental_wellbeing | Mental wellbeing state based on behavioral patterns | index | ✅ Yes |
| BIOMARKERS (6 total) | ||||
| biomarker | steps | Daily step count | count | ✅ Yes |
| biomarker | sleep_duration | Total time spent asleep | minute | ✅ Yes |
| biomarker | heart_rate_resting | Resting heart rate | bpm | ❌ No (lower is better) |
| biomarker | heart_rate_variability_sdnn | HRV measured as SDNN | ms | ✅ Yes |
| biomarker | heart_rate_variability_rmssd | HRV measured as RMSSD | ms | ✅ Yes |
| biomarker | vo2_max | Maximum oxygen uptake | mL/kg/min | ✅ Yes |
Explore the comprehensive breakdown of all 54 biomarkers across activity, body, vitals, sleep, and reproductive categories.
📊 View Complete Biomarker Table →All intelligence layer data (scores, archetypes, trends, comparisons) is sourced from Sahha API Documentation. Sahha is a leading health data intelligence platform that transforms biomarkers into actionable insights for health applications.