Timing Guidelines

A schedule that specifies an event that may occur multiple times. There are many healthcare activities that
occur on a schedule. The Timing datatype captures the details of a scheduled event. The Timing datatype is
used to specify the frequency and duration that a regular activity is to occur. For example, a medication
dosage may specify that it be taken once a day for 30 days. This datatype is used to capture the 'once a day' as
a frequency, and '30 days' as the Occurrence duration.

Name Description Data Type Value Set Cardinality Notes

The duration
that the timing
BoundsDuration TimeSpan 1
schedule is in
effect.

The upper
value of the
range of time
BoundsRangeHigh TimeSpan 1
that the timing
schedule is in
effect.

The lower value

of the range of
time that the
BoundsRangeLow TimeSpan 1
timing
schedule is in
effect.

Only applicable if
The encoded encompasses
abbreviation of entire timing
the timing, term, not just a
CodeConceptId Integer 1
such as 'BID', portion of it.
'qHS', 'q4H', Examples, BID,
etc. TID, AM, Q1H,
Q12H

The total
number of
Count Integer 1
times the event
occurs in the
 Name Description Data Type Value Set Cardinality Notes

specified
duration of
time.

The upper limit

of the number
CountMax of times the Integer 1
event occurs in
total.

The days of the

mon, tues, wed,
DaysOfWeek week that the Integer N
thurs, fri, sat, sun
event occurs.

The numerical
value of the
Duration TimeSpan 1
duration that
the event lasts.

The upper limit

of the length of
DurationMax TimeSpan 1
time that each
event lasts.

The number of
times this event
Frequency occurs in a Integer 1
specified
period.

The maximum
number of
times this event
FrequencyMax Integer 1
occurs in a
specified
period.

The number of
minutes before
Offset or after the Integer 1
event that the
timing
 Name Description Data Type Value Set Cardinality Notes

schedule is to
occur.

The length of
time over
Period which the TimeSpan 1
frequency of
events applies.

The maximum
length of time
PeriodMax over which the TimeSpan 1
frequency of
events applies.

The times of
TimeOfDay day that the TimeSpan N
event occurs.

The code that Examples: before

specifies when breakfast, late
WhenConceptId Integer N
the event afternoon,
occurs. bedtime, morning

TimeSpan
At Truveta, timespan is represented by days.hr:min:sec.

• Examples:
o 1 day = 1.00:00:00
o 5 hours = 0.05:00:00
o 6 min = 0.00:06:00
o 2 sec = 0.00:00:02
o 3 days, 4 hours, 2 min, and 1 sec = 3.04.02.01
o 1 month = 30.00:00:00
o 1 year = 365.00:00:00
o 10 years = 3650.00:00:00
Mapping to TimingAbbreviation codes:
The first goal is to map to timing components, such as period and frequency. If a timing concept cannot be
captured with components, then a TimingAbbreviation can be used if found.

• Examples:
o BID = frequency: 2, period: 1.00:00:00, do NOT map to SNOMED CT 229799001 | Twice a day
(qualifier value).
o Continuous = map to T.4111604 Continuous because this cannot otherwise be represented as
timing fields

Constraints:
• Duration SHALL be a non-negative value
• Period SHALL be a non-negative value
• If there is an offset, there must be a when
• If there is a timeOfDay, there cannot be a when
• If there is a when, there cannot be a timeOfDay
• BoundsDuration must be a non-negative timespan
• Offset must be a non-negative integer

Truveta Timing Assumptions:

• If a year only is stated, we assume January 1st of that year with BoundsDuration of 365 days.
• If a year and month are stated, we assume 1st of that month with BoundsDuration of 30 days.
• If a season is stated, we use the calendar date of that season with BoundsDuration of 30 days.
• If “second half of the year” is stated, we assume July 1st of that year with BoundsDuration of 182 days.
• If states “couple,” we normalized to “2.”
• If states “few,” we normalized to “3.”
• If states “multiple,” we normalized to a range of 2-10.
• If states “several,” we normalized to a range of 3-10.
• "A number" or "some" or "many" were not normalized.
• “Hundreds” was normalized to count of 100 given likely clinically equivalence.
• If says <1 minute or month or day, we have used duration of 0 to durationMax of 1.
• If states “up to 30 per month” we capture frequency of 0, frequencyMax 30, period 30 days.
• If uses a decimal like 2.5, we have rounded to 3.
• If a specific month is given, we use the number of days in that month – e.g., Oct = 31 days. If simply “3
months” is used, we use 30 days per month.
 • If we were able to normalize a partial timing, for example a count or a duration, but nothing else, we
did so.
• If a relative time term is used and there is no ObservationEffectiveDate, we did not capture the timing,
for example, “a week after surgery” or “a week ago” or “a few days ago.”
• If there is no BoundsDuration, then we are indicating that we could not calculate it, or it was not
reported over what duration this count occurred. E.g., “early in 2019” is normalized to 01/01/2019 but
without a BoundsDuration.

Examples by Timing Type:

1. Frequency: frequency is used to describe the number of times something occurs during a specific
period. It uses the two fields frequency and period.

• Frequency = the number of times this event occurs in a specified period

• Period = the length of time over which the frequency of events applies
• Examples:
o Every 8 hours = frequency: 1, period: 0.08:00:00
Translates to: [one time] in [8 hours]
o Every 7 days = frequency: 1, period: 7.00:00:00
Translates to: [one time] in [7 days]
o 3 times a day = frequency: 3, period: 1.00:00:00
Translates to: [3 times] in [1 day]
o 3-4 times a day = frequency: 3, frequencyMax: 4, period: 1.00:00:00
Translates to: [3 to 4 times] in [1 day]

2. Frequency + Duration: add duration to frequency to capture the length of the event

• Duration = the numerical value of the duration that the event lasts
• Examples:
o Every 21 days for 1 hour = frequency: 1, period: 21.00:00:00, duration: 0.01:00:00
Translation: [one time] in [21 days] for [1 hour]
o Three times a week for ½ hour = frequency: 3, period: 7.00:00:00, duration 0.00.30.00
Translation: [3 times] in [7 days] for [30 min]

3. When: when concepts are used to describe specific descriptive times of days. The value set is named
TimingEvent and contains concepts such as before dinner, before meals, afternoon, bedtime, etc.

• WhenConcept: the code that specifies when the event occurs

• Examples:
 o Before breakfast = whenConceptId: [3056761]
Translation: [before breakfast]

4. When + Offset: use offsets when the timing occurs before or after an event. Whether it occurs before
or after will be described by the concept. Offset can only be a non-negative integer.

• Offset: the number of minutes before or after the event that the timing schedule is to occur
• Examples:
o 10 minutes before meals = whenConceptId: [3056759], offset: 10
Translation: [10 minutes] [before meals]
o 1 hour after meals = whenConceptId: [3056779], offset: 60
Translation: [1 hour] [after meal]

5. Frequency + BoundsDuration: add boundsDuration to capture the length of time the schedule is in
effect.

• BoundsDuration: the duration that the timing schedule is in effect

• Examples:
o TID for 14 days = frequency: 3, period: 1.00:00:00, boundsDuration: 14.00:00:00
Translation: [3 times] in [1 day] for [14 days]

6. Days of the Week: use days of the week to capture events that recur on specific calendar days

• DaysOfWeekConceptId: the days of the week that the event occurs

• Examples:
o Monday, Wednesday, Friday morning = DaysOfWeekConceptId: [4069301,4069303,4069305],
whenConceptId: [3056774]
Translation: [Mon, Wed, Fri] [morning]

7. Using Count:

• Count: the number of times the event occurs in the specified duration of time
• Examples:
o Once = count: 1
Translation: [one time]