Why Interrupted Dosing Needs Explicit Records
A dose history is not simply a copy of the prescribed schedule. During treatment, a dose may be withheld, reduced, or resumed after an interruption. If a dataset retains only the original schedule, it can imply exposure that did not occur. If it records only a new dose amount without indicating when the change began, it can misstate the regimen in a different way. Population pharmacokinetic analyses need the actual administration history represented separately from the plan.
This distinction matters even when the planned and actual regimens differ for only a short period. A held dose contributes no administered amount at that opportunity; a reduced dose contributes the amount actually given; a restart is a new administration after the interruption, not evidence that intervening doses were given. The record should preserve what is known and make uncertainty visible rather than filling gaps from the prescription.
Representing Opportunities and Events
For clarity, index planned opportunities by \(j\). Let \(\tau_j\) be the planned time and \(P_j\) the amount prescribed for that opportunity. Record its status as administered, held, or unknown. If it was administered, record the actual amount \(A_j\) and the actual administration time \(t_j\). If it was held, record the hold status and its time or relevant interval, but do not define an administered amount \(A_j\) for it. If the status is unknown, leave it unknown until supporting records resolve it.
The prescribed amount \(P_j\) and actual amount \(A_j\) answer different questions. A record that says “100 mg prescribed” does not by itself establish that 100 mg was administered. A reduction should therefore retain both the prescribed plan, when available, and the actual amount given. Similarly, a hold is not an administration of a zero-milligram dose. A dataset may use a zero-valued row as a software-specific bookkeeping device, but it must identify that row as a held opportunity, not as a dose event.
A practical event history can retain the opportunity or event identifier, event status, actual time when known, actual amount when administered, and a source or comment that supports the status. Keep prescription changes as documented changes to the plan, rather than using them to overwrite earlier administrations. When the evidence does not establish whether a dose was given, the correct status is unknown—not held, and not administered at the planned amount.
This cumulative amount sums only over documented administered opportunities. A held opportunity is excluded because no \(A_j\) is defined for it. If a confirmed administration was partial, the sum uses the partial amount actually administered. This total is a bookkeeping summary over the specified period; it does not by itself describe the timing of doses or the concentration-time profile.
Worked Example: One Held Dose in a Repeated Schedule
Suppose a hypothetical participant is prescribed 100 mg every 12 hours, with planned opportunities at 0, 12, 24, and 36 hours. The medication record confirms administrations at 0, 12, and 36 hours, and a nursing record documents that the 24-hour dose was held. Reconstruct the events and total administered amount through 36 hours.
The opportunities at 0, 12, and 36 hours are administered events, each with an actual amount of 100 mg. The 24-hour opportunity is recorded as held, with no administered amount assigned to it.
The planned total over the four opportunities would have been \(4(100)=400\) mg, but that is not the administered total. The difference is the held 100 mg opportunity. It would be incorrect to calculate an amount for the held opportunity by multiplying an indicator of zero by an undefined \(A_j\); the sum is instead defined only over administered opportunities.
Reductions and Restarts
A reduction changes the amount actually delivered, not the historical dose record. For each reduced administration, preserve its actual amount and time. Where the record supports it, also retain the prescribed amount that would otherwise have applied and the documented instruction that changed the dose. Do not infer the effective time of a reduction from the first later dose if the change order has a different documented effective time.
A restart marks a return to administration after an interruption. It does not establish that the former dose was resumed unchanged. The restart amount must come from the administration record. A restarted dose may equal the former dose, be reduced, or follow a separately documented regimen. If the restart instruction describes a percentage but no actual administration is documented, the instruction alone is not proof that a dose was given.
Worked Example: A Reduced Regimen Followed by a Hold and Restart
In a hypothetical record, a participant is initially prescribed 300 mg every 24 hours. A documented order reduces the dose to 180 mg beginning on day 1. The administration record confirms 300 mg on day 0, 180 mg on day 1, and 180 mg on day 2. The day 3 dose is documented as held. The participant restarts with 180 mg on day 4. Find the total administered amount through day 4 and summarize the sequence.
The administered events are: day 0, 300 mg; day 1, 180 mg; day 2, 180 mg; and day 4, 180 mg. Day 3 is a held opportunity and contributes no administered amount. The order changing the plan helps explain the reduction, but the amounts in the sum come from the documented administrations.
Five daily opportunities from day 0 through day 4 would have totaled 1,500 mg at the original prescribed amount, but this hypothetical participant did not receive that planned total. The calculated 840 mg reflects the confirmed administrations and preserves the day 3 interruption. It does not imply that all doses after day 1 were reduced solely because of the order; the record of each administration remains essential.
Reconstruction and Quality Checks
Reconstructing a regimen means arranging documented events into a coherent timeline, not filling every expected opportunity with a dose. A useful review compares the medication administration record, dose-change orders, and relevant clinical documentation. These sources may answer different questions: an order can establish what was prescribed, while an administration record can establish whether and how much was given. When records conflict, keep the discrepancy visible for resolution instead of silently selecting the version that makes the schedule regular.
Use the documented regimen to identify expected dose times and prescribed amounts. Mark opportunities only where the schedule and period of interest are supported.
Use the available evidence to classify it as administered, held, or unknown. Do not turn an absent record into a confirmed hold unless the source establishes that a hold occurred.
For every confirmed administration, record its actual amount and time. Link reductions and restarts to their corresponding administrations and retain the supporting documentation.
Check whether changes have plausible effective times, whether the status agrees across sources, and whether any amount has been copied from the prescription rather than confirmed administration.
The central accounting identity follows directly from what “administered total” means: it is the sum of amounts delivered in administered events. If the administered-event set is \(J_{\mathrm{admin}}\), then by definition each \(j\) in that set has a defined actual amount \(A_j\), and the total is \(\sum_{j\in J_{\mathrm{admin}}}A_j\). A held or unknown opportunity is not in \(J_{\mathrm{admin}}\), so it contributes no term. This is why a status indicator multiplied by an undefined amount is not a valid substitute for defining the sum over actual administrations.
Worked Example: Distinguishing a Restart Instruction From a Given Dose
A hypothetical participant has a documented 80 mg dose at hour 0, followed by records confirming holds at hours 12 and 24. A note at hour 30 says “restart at 50% of prior dose.” The administration record confirms 40 mg at hour 36 and 40 mg at hour 48. Reconstruct the confirmed dose history through hour 48.
The note indicates an instruction to restart at half of 80 mg, which is \(0.50(80)=40\) mg. The hour 36 administration record confirms that amount was given; the hour 48 record confirms another 40 mg. The instruction by itself would not establish an administration, but here the administration records do.
The history is therefore 80 mg at hour 0, holds at hours 12 and 24, and administrations of 40 mg at hours 36 and 48. The two held opportunities are not counted as doses. If the hour 36 administration record were absent, the restart note alone would not justify adding 40 mg to the administered total.
Diagnostics, Alternatives, and Limitations
Several checks can reveal a reconstruction that is going wrong. Compare each actual amount with the prescribed amount that applied at that time; an amount below the plan may be a reduction, a partial administration, a data-entry issue, or a unit problem requiring investigation. Check that a documented hold is not simultaneously treated as an administered event. Review abrupt changes in dose amount against the recorded change order, and look for apparent gaps that were classified as holds without supporting evidence. A restart should have an actual dose amount and time if it is included as an administration.
One alternative is to model only the nominal prescribed schedule. That can be useful for describing the intended regimen, but it is not a substitute for actual dose history when the analysis requires exposure to doses received. Another alternative is to encode every planned opportunity in a table and add a status field. This can make interruptions easy to audit, provided held opportunities remain distinguishable from administered events and no undefined amount is used in calculations. For modeling input, the appropriate event representation depends on the analysis system, but the underlying distinction between plan, hold, and actual administration must be preserved.
These records have limits. A medication record may be incomplete, times or amounts may conflict across sources, and documentation of a plan may not establish what the participant received. A reconstructed history is only as reliable as its evidence. Do not use the apparent regularity of a regimen to fill missing doses, and do not interpret the cumulative amount alone as a complete account of exposure: timing, route, and formulation also matter.
Common Mistakes
- Counting prescribed doses as given. A schedule is evidence of the plan, not proof of administration. Use confirmed actual administrations in the administered total.
- Encoding a hold as an ordinary zero-amount dose. A zero may be a software bookkeeping value, but preserve a distinct hold status so it cannot be mistaken for an administration.
- Assuming a restart restores the old dose. Use the amount documented for the restart administration; do not infer it from the pre-hold regimen.
- Treating missing documentation as a confirmed hold. Missing status is uncertainty. Resolve it from supporting records where possible, and retain it as unknown if it remains unresolved.
- Using an amount that is undefined for a held opportunity. Sum over administered events only; do not multiply an indicator by an undefined actual amount.
Check Your Understanding
Answer each question using the distinction between a planned opportunity and a confirmed administration.
- A participant is prescribed 50 mg every 12 hours for three opportunities. Records confirm 50 mg at hours 0 and 24, and a hold at hour 12. What is the administered total?
- A 200 mg dose is prescribed, but the administration record documents that only 150 mg was given. What amount belongs in the administered total, and why?
- A note says “restart tomorrow,” but there is no record confirming a dose the next day. Should that dose be counted as administered?
- Why is it not generally valid to write a total as a sum of an administration indicator multiplied by an amount that is undefined for held opportunities?
- A participant receives 100 mg at hour 0, has a confirmed hold at hour 12, and receives 60 mg at hour 24 after a documented reduction. What is the administered total through hour 24?
Solutions. (1) \(50+50=100\) mg; the held hour 12 opportunity is excluded. (2) Use 150 mg, because the total records the amount actually administered, not the prescribed 200 mg. (3) No. The note documents a plan, not a confirmed administration; the status remains unknown unless another source confirms it. (4) Multiplication does not make an undefined amount valid when the indicator is zero. Define the sum only over administered opportunities, or use an explicitly defined bookkeeping value for holds while keeping those records distinct from administered events. (5) \(100+60=160\) mg; the hold contributes no administered amount.