Why Therapeutic Drug Monitoring Is a Workflow
A measured drug concentration is not, by itself, a dosing recommendation. Its meaning depends on why it was measured, what target is relevant to that question, and whether the sample was collected under the conditions for which the target is intended. Therapeutic drug monitoring (TDM) is the coordinated process of using drug concentration measurements, clinical information, and a treatment plan to support decisions about an individual patient's therapy.
TDM is useful when a concentration can help resolve a practical clinical concern. Examples include inadequate response despite a prescribed regimen, suspected concentration-related toxicity, or an unexpected change in a patient's response. Monitoring is not automatically useful for every drug or every patient: a result matters when it can be interpreted and can inform a decision. The workflow therefore begins with the clinical question, not with ordering a level.
A useful sequence is: state the concern; identify the target that applies to it; arrange a sample that can be compared with that target; verify the dose and sampling history; interpret the result with the patient's response and safety in view; then document and follow up the decision. Skipping any link can turn a precise laboratory value into a misleading conclusion.
Connect the Question to the Measurement
The clinical question determines what the result is meant to help assess. An efficacy concern asks whether exposure may be insufficient for the desired response. A toxicity concern asks whether exposure may contribute to an adverse effect or an unacceptable safety risk. A change in response may also prompt a review of the regimen, dose-taking history, clinical status, and measurement conditions. These questions are related, but they are not interchangeable: a concentration that is informative for one purpose may not answer another.
Before sampling, the care team needs a target appropriate to the question and a clear description of the conditions under which that target should be interpreted. The target might apply at a specified point in the dosing interval or under another defined sampling condition. The sample plan must match that condition closely enough to make the comparison meaningful. The later interpretation also requires knowing what was actually given and when, rather than relying only on what was prescribed.
For example: Is inadequate response plausibly related to insufficient exposure, or could an adverse effect be related to excessive exposure?
Record which target or range is being used and what sampling conditions it assumes. Do not compare a result to a target chosen for a different question or condition.
Record the dose history and the sample time relative to dosing. Check that the sample condition matches the target's intended condition.
Consider the concentration, response, adverse effects, dose-taking history, and relevant changes in the patient's condition together.
Document the reasoning and the plan, including whether or when monitoring will be repeated. A concentration alone does not dictate a dose change.
Why Sample Conditions Matter
A concentration changes over time after a dose. Let \(C(t_s)\) denote the patient's concentration at sample time \(t_s\), and let \(C_{\mathrm{obs}}\) denote the reported measurement. A simple measurement model is
Here, \(\varepsilon\) represents measurement error and other small discrepancies between the reported value and the concentration at that time. The equation emphasizes that a reported value belongs to a particular time and dosing history; it is not a time-free summary of the patient's exposure.
Suppose the target \(C^*\) is intended for a defined sample condition at time \(t^*\). The observed difference from that target can be written as
The first bracket is the difference caused by comparing concentrations from different times or conditions. If \(t_s\) does not match the target condition, this component may be substantial; it is not measurement noise and cannot be corrected merely by reading the assay more precisely. The comparison is most directly interpretable when the sample condition matches the target condition. If it does not, the team needs a justified way to account for the mismatch or should treat the result as unsuitable for that comparison.
This argument assumes that the sample is correctly identified, the assay result is valid, the relevant dose history is known, and the target's intended condition is understood. It does not assume that concentration alone determines response. A measured value can be accurate and still fail to answer the clinical question if the sample or target is mismatched.
Worked Example: Investigating an Efficacy Concern
Hypothetical case. A patient taking a prescribed medicine at 100 mg every 12 hours has had inadequate symptom control. For this example, the clinical team has already specified a trough target range of 5 to 10 mg/L. The sample is collected immediately before the next dose, after the patient and team confirm that the recent doses were taken as scheduled. The result is 4.0 mg/L.
Check the comparison. The result is from the sample condition specified for this hypothetical target, so a direct comparison is appropriate under the stated assumptions. The lower-bound difference is \(4.0\ \text{mg/L} - 5.0\ \text{mg/L} = -1.0\ \text{mg/L}\). The measured value is 1.0 mg/L below the target range's lower boundary. Relative to that boundary, it is \((5.0-4.0)/5.0 = 0.20\), or 20% lower.
Interpretation. The result raises the possibility that exposure is insufficient, which is relevant to the efficacy concern. It does not prove that low concentration caused the poor symptom control. The team should also consider the clinical response, dose-taking history, changes in the patient's condition, and whether there are other explanations for inadequate benefit. The percentage calculation describes the difference from the boundary; it is not, by itself, a dose-adjustment rule.
Worked Example: Assessing a Possible Toxicity Concern
Hypothetical case. A patient has a new adverse effect. For this example, a protocol has specified that a particular postdose sample condition should be compared with an upper target limit of 12 mg/L. A sample collected under that specified condition measures 14.4 mg/L. The patient's dose history and sample condition have been checked, and the result is assumed to be analytically valid.
Check the comparison. The result exceeds the upper limit by \(14.4\ \text{mg/L} - 12.0\ \text{mg/L} = 2.4\ \text{mg/L}\). The ratio to the upper limit is \(14.4/12.0 = 1.20\), so the result is 20% above that limit.
Interpretation. The result is a safety signal that should be considered alongside the adverse effect and the patient's full clinical status. It does not establish that the concentration caused the effect, nor does it specify a particular dose change. A clinician must assess the seriousness of the symptoms, relevant alternative explanations, the regimen, and the appropriate next step. If the sample condition had not matched the one specified for the limit, the numerical exceedance would not support the same direct comparison.
Worked Example: An Apparently Low Value From a Mismatched Sample
Hypothetical case. A patient takes a medicine at 08:00 and 20:00. A trough target range of 5 to 10 mg/L has been specified for a sample taken immediately before a dose. Instead, a sample is drawn at 18:30 and measures 4.6 mg/L. Assume the recorded dose times are correct and no other doses were taken.
Check the timing. The sample was collected \(18{:}30-08{:}00=10.5\) hours after the 08:00 dose. The next dose was due at 20:00, so the sample was taken \(20{:}00-18{:}30=1.5\) hours before that dose. It was not collected immediately before the dose, as specified for the trough comparison.
Interpretation. Although \(4.6\ \text{mg/L}\) is numerically below the target's lower boundary, directly classifying it as a below-target trough is not justified by the stated sample condition. The workflow has identified a timing mismatch, not demonstrated inadequate exposure. If the clinical question remains important, the team can plan a sample under the specified condition or use another justified interpretation. The data given do not support calculating what the concentration would have been at 20:00.
Interpretation, Alternatives, and Failure Modes
TDM is one input to a clinical decision, not a replacement for clinical assessment. A concentration within a target does not guarantee benefit or rule out toxicity. A concentration outside a target does not prove treatment failure or drug-caused harm. Targets describe a basis for comparison; individual response and safety still matter. Similarly, a result that seems unexpected should prompt a check of the chain that produced it before it prompts a change in treatment.
The main alternative is to guide treatment using clinical outcomes without routine concentration measurement. That can be appropriate when the response is directly observable and concentration results are not needed to make a decision. Fixed-dose treatment is another approach when the regimen is adequate for the relevant population and routine individual measurements add little value. In contrast, TDM can add information when measured concentration helps distinguish plausible explanations or supports a safer, more individualized decision. It does not make clinical response monitoring unnecessary.
Model-informed interpretation may also help relate a concentration to a patient's dose history and sampling conditions, particularly when a measurement does not match the usual target condition. Such an approach depends on appropriate data and a suitable model; it should not be used to conceal missing or unreliable dose and sample information. The practical choice is not “concentration versus model” in every case: the clinical question and available evidence determine whether direct comparison, model-supported interpretation, or clinical assessment without a concentration is most useful.
Common workflow failures include:
- Ordering a level without a decision question. A number with no defined purpose may not change management. State the concern and the decision the result could inform before sampling.
- Comparing unlike conditions. A result collected at a different point or under different conditions from those assumed by the target may not be directly comparable. Verify the sample condition or qualify the interpretation.
- Using the prescribed regimen as if it were the administered regimen. Missed, delayed, or extra doses can alter what the measurement represents. Confirm what was actually taken and when.
- Treating a target boundary as a diagnosis. A value above or below a target is a reason to investigate in context, not proof of toxicity, lack of efficacy, or causation.
- Failing to close the loop. If the team makes a change, records the rationale and follow-up plan. If no change is made, document why the result does not justify one or what further information is needed.
A reliable monitoring record therefore includes the clinical question, the target and its intended conditions, the reported concentration and units, dose and sample timing information, the patient's response and adverse effects, and the resulting plan. These details make the interpretation auditable and help the next measurement answer a defined question rather than repeat an ambiguous one.
Check Your Understanding
Answer each question, then compare your reasoning with the solutions.
- A patient has poor clinical response, but the sample was collected under conditions different from those specified for the target. What is the first interpretive problem to resolve?
- A result is 9 mg/L and the specified upper limit is 8 mg/L. Calculate the absolute and relative amount by which the result exceeds the limit. What can and cannot be concluded?
- Why does a target comparison require dose and sample timing information, even when the laboratory assay is accurate?
- Name two pieces of clinical information, apart from the concentration, that should inform interpretation when toxicity is suspected.
Solutions. (1) Resolve the mismatch between the actual sample condition and the condition assumed by the target; the value cannot yet be treated as a directly comparable target measurement. (2) The absolute excess is \(9-8=1\ \text{mg/L}\). Relative to the limit, it is \(1/8=0.125\), or 12.5% above. This establishes that the reported value exceeds the stated limit if the sample and target conditions match; it does not establish that the drug caused harm or dictate a dose change. (3) Concentration depends on time and dosing history, so a precise measurement may represent a different point in the concentration-time profile than the target describes. (4) Examples include the nature and severity of the adverse effect, the patient's overall clinical status, the actual doses taken, and other plausible explanations for the symptoms.