Why a Tied Maximum Needs Careful Reporting
The observed maximum concentration is often straightforward to identify: find the largest recorded concentration in the profile. The corresponding peak time can be less straightforward. If the same largest concentration is recorded at more than one sampling time, the data identify several times at which the observed maximum occurred; they do not identify one uniquely supported sampling time as Tmax.
This matters because Tmax is a sampling-grid statistic: it is tied to the times at which observations were collected. A set of equal maximum measurements may be consistent with a flat-looking region in the observed profile, but does not establish precisely when the underlying concentration curve reached, or remained at, its true maximum. The appropriate response is to report the tie transparently, rather than manufacture a unique peak time through interpolation or arbitrary rounding.
Here \(T_{\max}^{*}\) denotes the set of tied sampling times, not a claim that the concentration remained at its maximum throughout a continuous time interval. In ordinary prose and tables, Tmax is often used for the reported peak time. When there is a tie, state whether the reported Tmax is a set or range, or give a single value only after naming the convention used to select it.
Identifying a Tie and Describing Its Shape
First identify the maximum among the valid observed concentrations. Then collect every sampling time at which that maximum was recorded. If the tied times are consecutive in the ordered sampling schedule, the observations show a plateau-like maximum: several successive recorded concentrations have the same maximum value. Its observed span runs from the first tied sampling time to the last. If tied times are separated by observations below the maximum, report the tied times as separate observations; equality alone does not show that the profile was flat between them.
“Consecutive” refers to the order of the observations, not necessarily to evenly spaced measurements. For example, equal maximum values measured at 2 and 5 hours with a lower value at 3 hours are not a consecutive observed plateau. They are separate sampling times tied for the observed maximum. Describing the recorded pattern precisely avoids suggesting a continuous shape that was not measured.
The maximum and its tied times should be calculated from the unrounded concentrations available for analysis. Rounding for display is a separate step. If different recorded concentrations become equal only after rounding, they are not necessarily tied at the precision used in the analysis. Conversely, if the assay or data system records concentrations only to a fixed resolution, equal stored values are tied as observed; that does not establish that the underlying concentrations were exactly equal.
Worked Example: A Consecutive Observed Plateau
Suppose a hypothetical profile has the following measured concentrations. The concentrations are in ng/mL and the times are in hours.
| Time (h) | Concentration (ng/mL) |
|---|---|
| 0 | 0.0 |
| 2 | 3.2 |
| 4 | 7.5 |
| 6 | 7.5 |
| 8 | 5.1 |
Find the maximum. Comparing the recorded values gives \(C_{\max}=\max(0.0,3.2,7.5,7.5,5.1)=7.5\) ng/mL.
Find all tied times. The value 7.5 ng/mL occurs at 4 and 6 hours, so \(T_{\max}^{*}=\{4,6\}\) h. These are consecutive observations. The observed plateau-like maximum spans from 4 to 6 hours, a difference of \(6-4=2\) h.
A clear report is: “Observed Cmax was 7.5 ng/mL, tied at 4 and 6 h; the tied observations span 4–6 h.” If a summary requires a single Tmax and the prespecified convention is to use the earliest tied time, report Tmax = 4 h and disclose that the value is the earliest-time convention. The data themselves do not establish a unique peak at 4 h.
What the Observations Do and Do Not Establish
A tied maximum is a property of the recorded profile. Its interpretation is limited by the sampling schedule and by the precision of the recorded concentrations. Even when several successive measurements are equal, the observations do not show the concentration at every time between samples. The actual curve could have remained nearly level, risen or fallen between samples, or varied in a way the schedule did not capture. Thus, “observed plateau” is a description of the sampled values, not proof of a continuous plateau in the underlying concentration-time curve.
The same distinction applies to the reported peak time. If the tied observations occur at 4 and 6 hours, the measured data support those two sampling times as times at which the maximum was observed. They do not support reporting 5 hours as the uniquely determined time of the true peak merely because it is the midpoint. Nor does a tie by itself justify a claim about the exact onset or end of a plateau.
Worked Reporting Choices
Worked Example: Equal Maxima at Separated Times
Consider this hypothetical profile, with time in hours and concentration in mg/L: at 0, 1, 2, 3, 4, 5, and 6 h, the respective concentrations are 0.8, 4.8, 6.2, 5.0, 6.2, 4.9, and 2.2.
Calculate Cmax. The greatest recorded concentration is \(C_{\max}=\max(0.8,4.8,6.2,5.0,6.2,4.9,2.2)=6.2\) mg/L.
Identify the times. The maximum occurs at 2 and 4 h, so \(T_{\max}^{*}=\{2,4\}\) h. These are not consecutive observations because the 3 h concentration, 5.0 mg/L, is lower than 6.2 mg/L.
Report: “Observed Cmax was 6.2 mg/L, tied at 2 and 4 h.” Do not call the 2–4 h span a plateau: the intervening measurement is below the maximum. The equality establishes a tie in the observed values, not a continuous shape or an explanation for why the two measurements match.
Worked Example: Applying an Earliest-Time Convention
Suppose a hypothetical profile has concentrations of 1.1, 5.6, 8.0, 8.0, and 6.3 mg/L at 0, 1, 2, 3, and 4 h, respectively. A study analysis plan specifies that, when the observed maximum is tied, the earliest tied sampling time will be reported as Tmax.
Calculate the maximum and tie. The largest concentration is \(C_{\max}=\max(1.1,5.6,8.0,8.0,6.3)=8.0\) mg/L. It is observed at \(T_{\max}^{*}=\{2,3\}\) h.
Apply the stated rule. The earliest member of the tie set is \(\min\{2,3\}=2\) h. Therefore the single convention-based summary is Tmax = 2 h. The full report should retain the context: “Observed Cmax was 8.0 mg/L at 2 and 3 h; under the earliest-time convention, Tmax was reported as 2 h.”
This calculation follows the specified reporting rule. It does not show that the concentration reached its underlying peak at exactly 2 h rather than 3 h. If no convention has been specified, selecting the earliest time without saying so hides an analytical choice.
Why a Single Time Cannot Be Derived from a Tie
The logic can be stated directly. Let \(M\) be the set of indices for which the observed concentration equals the maximum. By definition, for each \(i\in M\), \(C_i=C_{\max}\). If \(M\) has two or more members, the observations provide no criterion for ranking those tied sampling times by concentration: every one of them has the same recorded value. Selecting the earliest, latest, or midpoint therefore adds a reporting rule rather than extracting a uniquely preferred time from the measurements.
The same argument explains why interpolation does not resolve the issue without additional assumptions. Interpolation supplies values between observed times according to a chosen mathematical rule; it does not create new measurements. When tied observations occur at 4 and 6 h, assigning a peak at 5 h requires a model of the unobserved curve. Unless that model and its purpose are justified, the midpoint is not a more data-supported estimate than either observed time.
Worked Example: Reporting a Tie at Recorded Precision
A hypothetical data system stores concentrations to one decimal place. At 2, 3, and 4 h it records 10.0, 10.0, and 10.0 µg/L; at 5 h it records 8.7 µg/L. Assume these are the values available for the exposure-metric calculation.
Calculate the recorded maximum. \(C_{\max}=\max(10.0,10.0,10.0,8.7)=10.0\) µg/L. The recorded tie set is \(T_{\max}^{*}=\{2,3,4\}\) h, and the consecutive tied observations span \(4-2=2\) h.
A suitable report is: “Recorded Cmax was 10.0 µg/L at 2, 3, and 4 h; the tied observations span 2–4 h.” Because the stored concentrations have one decimal place, this report describes equality at the available recorded precision. It does not claim that unrounded laboratory values were identical, or that the underlying concentration was exactly constant across the span.
Diagnostics, Alternatives, and Limitations
Before reporting a tied peak, verify that the profile has been ordered by actual sampling time and that the concentrations being compared use consistent units and the same analysis basis. Check the available stored precision before rounding for presentation. A tie created only by display rounding should not be confused with a tie in the analysis data. If a value is missing or invalid, do not silently treat it as a measured concentration; establish which observations are eligible for the maximum according to the analysis rules.
A second diagnostic is to inspect the tied times and the observations between them. Consecutive tied maximum values support the description “consecutive observed plateau” or “tied observations spanning [first time]–[last time].” Tied values separated by lower observations support a list of tied times, not a plateau over the full range. In either case, sparse sampling limits what can be concluded about unobserved times.
When a single Tmax is required, common summaries include the earliest tied time, the latest tied time, or the full set or range of tied times. The full set is the most transparent description of what was observed. An earliest- or latest-time convention can be useful for a consistent dataset, but it must be prespecified or clearly stated. A midpoint can be compact, but implies a unique value that the tied measurements did not select. No one convention removes the limitations of the sampling schedule.
| Reporting approach | What it communicates | Main caution |
|---|---|---|
| List every tied sampling time | The observed maximum and all times at which it was recorded | May be less compact in a summary table |
| Report the first-to-last tied-time range | The span of tied observations; especially clear for consecutive ties | For separated ties, the range can misleadingly imply a plateau unless the intermediate values are described |
| Use earliest or latest tied time | A single value chosen under a stated convention | Must not be presented as uniquely determined by the tied data |
| Use the midpoint | A numerical summary between tied times | Can imply unsupported timing precision and a peak at an unobserved time |
The most important limitation is that observed equality is not a direct measurement of the continuous curve between samples. A tie can reflect the shape of the profile, the timing of the sampling grid, or the precision of recorded values. This tutorial addresses how to describe tied observed maxima; it does not establish a mechanistic explanation for them or recover the unobserved peak time.
Common Mistakes
- Reporting only the first time without naming the rule. The earliest time is a possible convention, but a tie means the observed maximum also occurred at the other tied sampling times.
- Calling every tie a plateau. A plateau-like observed pattern requires consecutive tied observations. Equal maxima separated by a lower value are not a continuous observed plateau.
- Reporting the midpoint as the measured Tmax. The midpoint may not be a sampling time, and the observed concentrations do not select it as the true peak time.
- Interpreting rounded equality as exact equality. Equal displayed values can hide differences in the unrounded data, while equality in stored values still supports only a tie at the recorded precision.
- Equating an observed plateau with a known continuous plateau. Measurements at a finite set of times do not establish the concentration at every time between them.
Check Your Understanding
Answer each question using the observed concentrations and state what can and cannot be concluded.
- At 1, 2, 3, and 4 h, concentrations are 2, 6, 6, and 4 mg/L. Find Cmax and the tied-time set. Is the observed maximum consecutive?
- At 0, 1, 2, 3, and 4 h, concentrations are 1, 7, 5, 7, and 2 µg/L. Give an appropriate report and explain why 1–3 h is not an observed plateau.
- A profile has Cmax = 12 ng/mL tied at 3 and 5 h. A report gives Tmax = 4 h without explanation. What is the problem, and how could the report be improved?
- Two values differ in the analysis data but display as 9.4 mg/L after rounding to one decimal place. Should they be treated as tied for analysis? Explain.
Solutions. (1) \(C_{\max}=\max(2,6,6,4)=6\) mg/L and \(T_{\max}^{*}=\{2,3\}\) h. The tied times are consecutive observations, so they form a consecutive observed plateau-like maximum spanning 2–3 h; this does not prove a continuous plateau. (2) The maximum is 7 µg/L at 1 and 3 h, so report “Observed Cmax was 7 µg/L, tied at 1 and 3 h.” The 2 h value is 5 µg/L, below the maximum, so the tied observations are separated and do not form a plateau across 1–3 h. (3) The data identify the tie at 3 and 5 h, not a uniquely observed maximum at the unobserved midpoint 4 h. Report the full tie, or state a prespecified convention such as earliest tied time, in which case Tmax = 3 h under that convention. (4) No. Use the unrounded analysis values to determine whether the maximum is tied; display rounding alone should not create an analysis tie. State the displayed precision separately if needed.