Blood glucose meters, ketone meters and continuous glucose monitors have a range within which they can measure levels. When a reading falls outside of this range, we need a mechanism to show that we have received an out of range reading. This is done by creating a CBG, SMBG or blood ketone reading with a value that is 1 unit over or below the range threshold, and then annotating the reading.
For example, for a low CGM reading for a device that displays in mg/dL with a low threshold of 40 mg/dL:
{
"type": "cbg",
"units": "mg/dL",
"value": 39,
"trendRate": -2.2,
"trendRateUnits": "mg/dL/minute",
"sampleInterval": 300000,
"clockDriftOffset": 0,
"conversionOffset": 0,
"deviceId": "DevId0987654321",
"deviceTime": "2016-06-13T19:05:43",
"guid": "cf64b6f6-54af-4e8b-a432-08d53c2484d4",
"id": "da34f6f278984ae385c91cac6b91eff7",
"time": "2016-06-14T02:05:43.959Z",
"timezoneOffset": -420,
"uploadId": "SampleUploadId",
"annotations": [
{
"code": "bg/out-of-range",
"value": "low",
"threshold": 40,
}
]
}Unlike all other data model types, the out of range annotation does not currently include a field for units. Because of this, when annotating CBG or SMBG readings, the threshold should always be in mg/dL, even when the device displays readings in mmol/L. Although an unfortunate oversight, the reason for this is that all devices (that Tidepool knows of) store readings in mg/dL, even if they display in mmol/L.
For blood ketone readings, the threshold should be a value measured in mmol/L. We acknowledge that this is inconvenient and inconsistent with the rest of our data model, and hope to address it in the future by adding a units field, similar to the field used in the CBG, SMBG and blood ketone data types.
For example, for a high CGM reading for a device that displays in mmol/L with a high threshold of 23.0 mmol/L:
{
"type": "cbg",
"units": "mmol/L",
"value": 23.035604162838963,
"trendRate": -0.4,
"trendRateUnits": "mmol/L/minute",
"sampleInterval": 300000,
"clockDriftOffset": 0,
"conversionOffset": 0,
"deviceId": "DevId0987654321",
"deviceTime": "2016-06-13T19:05:43",
"guid": "cf64b6f6-54af-4e8b-a432-08d53c2484d4",
"id": "da34f6f278984ae385c91cac6b91eff7",
"time": "2016-06-14T02:05:43.959Z",
"timezoneOffset": -420,
"uploadId": "SampleUploadId",
"annotations": [
{
"code": "bg/out-of-range",
"value": "high",
"threshold": 414,
}
]
}If the device (or manufacturer) does not provide the exact threshold levels, then the threshold field should not be provided, and an additional [datatype]/unknown-value annotation should be provided. This is the case in our current Abbott Precision Xtra driver for ketone readings:
{
"type": "bloodKetone",
"units": "mmol/L",
"value": 10.0,
"clockDriftOffset": 0,
"conversionOffset": 0,
"deviceId": "DevId0987654321",
"deviceTime": "2016-06-13T19:05:43",
"guid": "ae40651d-b8e7-428e-840f-bbb3e1132569",
"id": "55346384504b4b76b24c60a236448012",
"time": "2016-06-14T02:05:43.500Z",
"timezoneOffset": -420,
"uploadId": "SampleUploadId",
"annotations": [
{
"code": "ketone/out-of-range",
"value": "high"
},
{
"code": "ketone/unknown-value"
}
]
}In this instance, because the threshold was unknown, we used a value that we knew was out of the meter's range.