Skip to content

Estimated Timestamps

Some devices, most commonly blood pressure monitors and glucometers, rely on internal power to maintain their date and time. Under certain conditions, the device clock can be temporarily reset.

When this happens, the device may still capture a valid measurement, but the recorded time on the device may not accurately reflect when the reading was taken. Passing that incorrect time through unchanged would create misleading datapoint since the timestsamp would be incorrect.

To prevent this, Tenovi applies timestamp estimation when it detects that a device’s internal time cannot be trusted.

What Triggers Timestamp Estimation?

A common trigger is temporary loss of power, such as when batteries are removed from a device. When batteries are taken out:

  • The device may lose its internal clock
  • The next reading may be taken before the device has reconnected to the gateway
  • Until that reconnection happens, the device does not yet have a reliable real-world time reference

In these cases, readings are still captured correctly, but their timestamps must be estimated. Once the device reconnects to the gateway, accurate time is restored automatically and subsequent readings use normal timestamps.

Common Patient Use Patterns

These are patterns we commonly see in everyday use. They are not errors, and they do not indicate anything wrong with the device or the reading itself.

Battery removal between uses

Some users remove batteries between readings, often as part of their routine. If a reading is taken after the batteries are reinserted but before the device reconnects to the gateway, that reading will receive an estimated timestamp.

Readings Taken Away From The Gateway

Readings may be taken in another room, while traveling, or while the gateway is unplugged or temporarily offline. In these situations:

  • The device stores the reading
  • Time synchronization cannot occur immediately
  • When the reading is later uploaded, Tenovi estimates the timestamp

Repeated Power Resets

If batteries are removed and reinserted frequently, and readings are taken before reconnection each time, multiple readings in a row may be flagged as estimated.

Combined Behaviors Over Time

When temporary power loss and offline readings happen together over a period of time, all readings during that window may require timestamp estimation until a successful gateway connection re-establishes accurate time.

How The Estimation Works

When a timestamp needs to be estimated, Tenovi looks for a reliable reference point in the following order:

  1. The last valid measurement timestamp
  2. The delivery date
  3. The shipped date plus estimated transit time
  4. The date the device was first registered in Tenovi

Using that reference, Tenovi calculates elapsed time and applies it to produce the most accurate estimate possible.

Estimated timestamps mean:Estimated timestamps do not mean:
✅ Tenovi detected a loss of reliable device time❌ The measurement value is incorrect
✅ A protective correction was applied❌ The device malfunctioned
✅ The reading remains valid and clinically useful❌ Data integrity was compromised

How Estimated Readings Are Identified

Each reading that has an estimated timestamp includes a flag:

"estimated_timestamp": true

This provides transparency and allows downstream systems to handle these readings intentionally.

[
{
"metric": "string",
"device_name": "string",
"hwi_device_id": "string",
"patient_id": "string",
"hardware_uuid": "string",
"sensor_code": "string",
"value_1": "string",
"value_2": "string",
"created": "2019-08-24T14:16:18Z",
"timestamp": "2019-08-24T14:15:22Z",
"timezone_offset": -2147483648,
"estimated_timestamp": true, // denotes an estimated timestamp
"filter_params": {}
}
]

Recommendations When Working With Estimated Timestamps

  • Use the estimated_timestamp field on metrics to branch logic
  • Consider falling back to created timestamp (the time the metric was saved on Tenovi servers) for estimated readings
  • Avoid automatically excluding estimated data, as this can unintentionally remove valid measurements
  • Treat metrics with estimated timestamps visually different to help clinicians and care managers know they are working with a metric that has an estimated timestamp