Risk Engine

POST https://models-nl.dacadoo.com/risk/1.4.0

Request

The use of data values can by any combination of imputation (used to derive other values), imputed (can be derived from other values), confidence (confidence intervals are provided), and estimation (used to estimate some risk) as given by the Use column in the tables below.

Metric Health Model

Biometric data. The eth value must only be used in compliance with applicable regulation. ICD-10 codes are provided where applicable.

   ethnicity codes:

      0   =>	Not specified
      3   =>	African American
      4   =>	American Hispanic
      5   =>	Asian
      10  =>	Southern European
      12  =>	Middle-eastern, non-Arab

Smoking

Smoking history and status.

Movement

Energy transfer from physical activity, represented as time series up to 60 days with array index 0 being today. Use null for gaps. A decay is applied with more weight assigned to more recent entries. To avoid the time decay or to supply an aggregate value, fill in index 0 only.

Sleep

Sleep, represented as time series up to 60 days with array index 0 being today. Use null for gaps. A decay is applied with more weight assigned to more recent entries. To avoid the time decay or to supply an aggregate value, fill in index 0 only.

Nutrition

Nutrition questions. These questions are meant to be answered with a slider ranging from never (0) to all the time (1). The questions denoted as Starter must all be answered for the input to be considered.

Quality of Life

Psychometric questions. These questions are meant to be answered with a slider ranging from not at all (0) to very much so (1).

Confidence

Confidence parameters, and intervals (on output). Confidence intervals are provided as per the Metric Health Model table above where the Use column includes confidence. Each confidence interval is represented as an array with its lower and upper bounds, in that order.

Interview Guidance

Interview guidance endpoint choice, and uncertainties of estimation (on output). Uncertainties are provided as per the table below for groups of inputs as well as overall. Each uncertainty is represented as an array with the lowest and highest expected index value of the chosen endppoint, in that order. For continuous variables, these values are based on the confidence intervals of imputations of missing inputs, at the specified confidence level (see Confidence above). If an input group is fully provided, there is no uncertainty, and the two values are identical and correspond to the estimated index value. Otherwise, there may be a difference between the two values. An interview implementation may consider that difference as the remaining uncertainty for the input group, and prefer input groups with a large difference.

   endpoint codes:

      0  =>  All-cause mortality. Excludes death from accident, infectuous disease and self inflicted death.
      1  =>  Cancer.
      2  =>  Cardiovascular disease.
      3  =>  Hypertension.
      4  =>  Type 2 diabetes mellitus.
      5  =>  Metabolic dysfunction.
      6  =>  COVID-19 susceptibility, i.e. risk if contracted.

Response

Risks

Condition-based mortality and morbidity risks, specifically:

• all-cause. All cause mortality. Excludes death from accidents, infectuous diseases, and self-inflicted death.
• cancer. Cancer.
• cardiovascular-disease. Cadiovascular disease.
• hypertension. Hypertension.
• type-2-diabetes. Type 2 diabetes mellitus.
• metabolic-dysfunction. Metabolic dysfunction.
• covid-19. COVID-19 susceptibility, i.e. risk if contracted. The model is based on a meta-analysis and was not validated against independent data.

Some fields are based on a reference population and related stratification. The reference population is prepared individually using customer-supplied data on a sufficient number of lives, and using an agreed stratification for the business case. In evaluation mode, a proxy reference population stratified by age and sex is used.

Examples

Request Body

{
  "mhm": {
    "age": 45,
    "hgt": 178,
    "wgt": 97,
    "sex": 0    
  },
  "smk": {
    "now": 0,
    "evr": 0
  },
  "mvm": {
    "nrg": [ 6.5 ]
  },
  "slp": {
    "bed": [ 8.5 ],
    "slp": [ 7.5 ],
    "awk": [ 2 ]
  },
  "nut": {
    "nqs": [ 0.5, 0.5, null, 0.5 ]
  },
  "qlm": {
    "q01": 0.8,
    "q02": 0.9,
    "q04": 0.0
  },
  "conf": {
    "level": 0.95,
    "onesided": 0
  },
  "clip": false
}

Response Body

{
  "risks": [
    {
      "type": "mortality",
      "condition": "all-cause",
      "index": 4.83,
      "ratio": 0.616,
      "zscore": -0.886,
      "percentile": 18.77
    },
    {
      "type": "mortality",
      "condition": "cancer",
      "index": 2.07,
      "ratio": 0.693,
      "zscore": -0.672,
      "percentile": 25.08
    },
    {
      "type": "mortality",
      "condition": "cardiovascular-disease",
      "index": 0.59,
      "ratio": 0.479,
      "zscore": -0.694,
      "percentile": 24.37
    },
    {
      "type": "morbidity",
      "condition": "hypertension",
      "index": 0.98,
      "ratio": 0.723,
      "zscore": -0.715,
      "percentile": 23.74
    },
    {
      "type": "morbidity",
      "condition": "type-2-diabetes",
      "index": 0.99,
      "ratio": 0.72,
      "zscore": -0.73,
      "percentile": 23.27
    },
    {
      "type": "morbidity",
      "condition": "metabolic-dysfunction",
      "index": 6.05,
      "ratio": 0.689,
      "zscore": -1.003,
      "percentile": 15.78
    },
    {
      "type": "mortality",
      "condition": "covid-19",
      "index": 0.58,
      "ratio": 0.57,
      "zscore": -0.464,
      "percentile": 32.13
    }
  ],
  "mhm": {
    "sbp": 120,
    "dbp": 74.7,
    "fbg": 5.3,
    "note": "67 values redacted in evaluation mode"
  },
  "conf": {
    "level": 0.95,
    "onesided": 0,
    "sbp": [90.2, 149.1],
    "dbp": [54.4, 95.7],
    "fbg": [3.1, 7.7],
    "note": "38 values redacted in evaluation mode"
  },
  "clip": false
}