This to track discussion and decisions made about user story #3 on "Automatically trigger 'needs attention' status/flag/indicator on D. Remote Temperature Monitoring (RTM).
For the automatically triggering of "Needs attention" status/flag/indicator, we talked about how we need some sort of "formula" where the RTM data can trigger the flag. However, the parameters will be set up in ColdTrace, not in OpenLMIS.
Current CCE inventory Screen:
Insert the Order Fulfillment page where the flag needs to be visable
The ColdTrace API
The current design of the API is outlined here: https://docs.google.com/document/d/1cwdTIOtPbuf1O2_s5BPgEKEAps9HyU4A9TEbM6Owmk0/edit
will provide the following:
- alert_id [string] - opaque identifier (UUID etc)
- device_id [string] - identifier defined by partner system (FHIR device id, serial number, etc).
- start_ts [int] - microseconds since epoch
- end_ts [nullable] - microseconds since epoch
- status [map[IETF_lang_tag (i.e. “en-US”): string]] - Translated - Human readable description of the alert in the configured languages for the integration
- alert_type [string] - predefined set of alert types [temp out of range, uptime out of range, etc...]
"en-US": "Equipment needs attention: too hot"
ColdTrace sets the Conditions for all the alert_types.
Balsamiq Mock Ups (editable)
|1||What is the 'status' flag in OpenLMIS? Is it a checkbox? What is the visual element which will be either turned on or off based on the message from ColdTrace.||Discussing a 'bell' icon which will have two states initially (discuss snoozing in the future). Active (filled in) and deactive (outline)||Closed|
|2||How to calculate "last 25 days." I was assuming that we would calculate the 25 days by just counting 25 days back from whatever today’s date is. So the 25 day-range for today starts on Sep 30, the 25 day range for tomorrow starts Oct 1, etc.||Will be addressed in ColdTrace, not in OpenLMIS. The alert conditions will be defined in ColdTrace.||Closed|
Should "needs attention" flag/indicator also be automatically un-checked once fridge is back to normal functionality?. Ok, so using that assumption, here’s the scenario I was trying to describe about how to figure out when a “needs attention” checkbox should be un-checked.
Scenario: The fridge has been in temperature range for 79% of the time for 25 days, and so the RTM sends the alert to OpenLMIS to automatically trigger a “needs attention” status. Two days later, the fridge has been doing better and has been in temperature range for 81% of the last 25 days.
|The dismissing of alerts will initially be done by alert versus by device. The second iteration we can address dismissing all alerts for the device.||In progress|
|4||Does the needs attention checkbox remain unchecked until a DIVO unchecks it? Or should the RTM system send a separate “uncheck” call to OpenLMIS to automatically un-check once the fridge is back within that 80% range? Or maybe the system just runs the formula every night, so that if tomorrow the fridge is doing better the box won’t get checked? I think probably the last option makes the most sense, but there are some other questions that it brings up as well. Such as: (see next open question)||Current working assumption is that a user would manually need to override (still determining what override means) the Needs Attention flag.|
|5||What if the RTM sent that flag but the DIVO wanted to manually uncheck it? If the fridge is out of range tomorrow also, would the system show that flag again? Or do we want to give the DIVO the option to “snooze” the flag for a certain period of time? Keep in mind that what I’m imagining here isn’t like a pop up notification or something that would be really in-your-face. It would be like a little icon or checkbox within the CCE inventory list next to the functionality status.||For 3.3 we will work to build a dismiss but not a snooze or more complicated process. Essentially it is an "off" until it comes again.||In progress|
|6||In what scenario would a DIVO want that flag/checkbox to go away? Is there any reason they’d want it to go away while the fridge is still having issues?|
Will send the alert at the start and end.
We could put in a automatic dismiss when the end alert comes.
|7||How much historical data do we want? Do we want this to feed into the DISC alert indicator?|
|8||How do we want to display the start/end timestamp for an alert? Just the date? Date and time? "minutes/hours/days ago"?|
Currently, we have not displayed an actual time in OpenLMIS.
|See the example, addressed. ||Closed|
Based on a conversation with Ana Costache, we'd like to start with these parameters: Normal fridge function: Fridge is in the 2-8 degree range 80% of the time or more over a period of 25 days Needs Attention / abnormal function: Fridge is within 2-8 degrees LESS than 80% of the time, over a period of 25 days We'd like these to be configurable, but these represent a good starting point.