PharmaDossier: Mapping Coverage Gaps to Trial Treatment Interruptions

When routine-care or protocol-required background therapy in an ongoing clinical trial encounters an access delay, prior-authorization dispute, or benefit-routing rejection, trial sites face an immediate documentation dilemma: does an administrative payer denial constitute an intercurrent event (ICE), does it create missing endpoint data, or is it merely an operational friction that leaves clinical exposure intact? The answer is foundational to trial integrity: trial teams must record actual drug exposure and administrative coverage status as two completely independent source streams before any biostatistician or medical monitor classifies an intercurrent event under the prespecified estimand.

Administrative interruptions also occur when a product moves between pharmacy and medical benefit structures. Adjacent access-operations reading in PharmaDossier's article on pharmacy-benefit lockouts when route of administration shifts a drug between medical and pharmacy benefits, a route-of-administration or site-of-care change can move a drug between pharmacy and medical benefits and reset access operations. PharmaDossier is a publication covering those benefit-channel operations—it is not a payer, a specialty pharmacy, a contract research organization (CRO), a trial biostatistician, or a source of coverage. That related page is not this trial source-record worksheet, and a payer explanation of benefits is not a verified administration record.

Confusing an administrative insurance rejection with physical treatment discontinuation introduces severe classification bias into trial datasets. A trial participant whose specialty pharmacy delivery is delayed may continue taking existing home supply, borrow emergency doses from a health system clinic, or transition to manufacturer patient-assistance supply, maintaining full compliance with the protocol regimen despite a formal insurance denial letter. Conversely, a participant may discontinue background medication entirely while insurance authorization remains fully active and unflagged. If data management or clinical research associates (CRAs) conflate payer status with patient exposure, they risk mischaracterizing unobserved endpoints as intercurrent events or treating unverified administrative hurdles as treatment failures.

This guide provides an operational framework for trial sites, data managers, and biostatisticians to construct a standardized access-event chronology. Rather than recasting foundational estimand theory—such as our prior analyses on ICH E9(R1) estimand construction and intercurrent event strategies and distinguishing missing data from intercurrent events in sensitivity analyses—this article establishes the front-line evidentiary boundaries that govern how site records must capture therapy identity, accountability chains, actual dosage, and endpoint timing before any statistical classification occurs.

ICH E9(R1) defines intercurrent events as events occurring after treatment initiation that affect either the interpretation or the existence of the measurements associated with the clinical question of interest. Missing data are data that would be meaningful for the analysis of a given estimand but were not collected. They should be distinguished from data that do not exist or that are not considered meaningful because of an intercurrent event. The addendum is a guideline, not a statute.

ICH E9(R1) section A.3.1 defines intercurrent events as events occurring after treatment initiation that affect either the interpretation or the existence of the measurements associated with the clinical question of interest. The addendum's examples include discontinuation of assigned treatment, use of an additional or alternative treatment, and terminal events such as death. Events such as discontinuation of treatment, switching between treatments, or use of an additional medication do not, by themselves, mean that the variable cannot be measured thereafter, though the measures may not be relevant. A participant whose background therapy is interrupted can still attend a later scheduled visit and complete the protocol variable. Whether those later measurements remain meaningful is a question for the locked estimand, not for a coverage letter.

Whether a change in background therapy is even an intercurrent event depends on how the protocol specifies the treatment of interest, not on whether a claim was paid. In ICH E9(R1) section A.3.4, referring to the treatment attribute in A.3.3, the addendum gives this example: it is possible to specify treatment as intervention A added to background therapy B, dosed as required; in that case, changes to the dose of background therapy B would not need to be considered as an intercurrent event, whereas use of an additional therapy would. If treatment is specified as intervention A only, then both changes in background therapy and use of additional therapy would be addressed as intercurrent events. The addendum also notes that the description of intercurrent events might in theory reflect very specific details such as a single missed dose of a chronic treatment; where such specific criteria are not expected to affect interpretation of the variable, they would not need to be addressed as intercurrent events.

The addendum states that a change to the estimand should usually be reflected through amendment to the protocol. When a clinical operations team discovers widespread coverage denials for background therapies across trial sites, the team cannot unilaterally reclassify those events as missing data, nor can biostatisticians alter the intercurrent event handling rules post-hoc. The trial must capture precise exposure facts, leaving the locked protocol and prespecified statistical analysis plan (SAP) endpoint rules to govern the analysis.

To prevent operational confusion during monitoring and audits, clinical trial investigators and sponsors must recognize the distinct legal and regulatory standards governing investigational accountability versus routine medical history. In the United States, Food and Drug Administration (FDA) regulations under 21 CFR 312.62 divide investigator recordkeeping into two separate subsections with fundamentally different evidentiary burdens:

• Investigational Drug Disposition (21 CFR 312.62(a)): An investigator is required to maintain adequate records of the disposition of the drug, including dates, quantity, and use by subjects. In this IND context that sentence is investigational-product accountability. A prior-authorization flag, a pharmacy-benefit rejection, or a medical-versus-pharmacy routing status is not that disposition record.

• Case Histories and Pertinent Observations (21 CFR 312.62(b)): The investigator must prepare and maintain adequate and accurate case histories that record all observations and other data pertinent to the investigation on each individual administered the investigational drug or employed as a control. When a protocol mandates routine-care background therapy (such as a standard-of-care oral antineoplastic or baseline inhaler), actual participant compliance, missed doses, and the clinical rationale for omission belong in these pertinent case observations. Administrative billing status may be recorded as contextual history, but it cannot substitute for factual observations of patient use.

These expectations are reinforced by the consolidated ICH E6(R3) Guideline for Good Clinical Practice (16 June 2026). Section 2.12.2 states that the investigator/institution should maintain adequate source records that include pertinent observations on each of the trial participants under their responsibility. Source records should be attributable, legible, contemporaneous, original, accurate, and complete. Changes to source records should be traceable, should not obscure the original entry, and should be explained if necessary via an audit trail. Section 2.12.6 states that data reported to the sponsor should be consistent with the source records or the discrepancies explained.

Regarding drug supply, ICH E6(R3) section 2.10.4 requires comprehensive documentation of investigational product delivery, inventory, use by each participant (confirming participants received protocol-specified doses), and return or destruction. While section 2.10.4 acknowledges that for authorised medicinal products used in a trial, alternative approaches to inventory records may be considered in accordance with local regulations, those allowances do not relieve the site of documenting actual exposure. Placing an insurer's 'Claim Denied' notification in a subject's binder without recording patient-reported pill counts or administration dates fails both 21 CFR 312.62(b) and ICH E6(R3) 2.12.2.

A frequent source of confusion among clinical investigators is the assumption that federal coverage mandates guarantee third-party reimbursement for all treatments administered within a clinical trial. When a claim is rejected, site personnel sometimes assume a regulatory breach has occurred or that the patient must immediately discontinue therapy. Examining the primary statutes and Medicare determinations reveals why insurance denials are often the legally expected outcome rather than an operational failure.

In the public payer sphere, CMS National Coverage Determination 310.1, Routine Costs in Clinical Trials, is effective for items and services furnished on or after 9 July 2007 and was last reviewed in July 2007. Transmittal 12590 / CR13597, issued in April 2024 and effective and implemented 27 May 2024, announces a technical change to Pub. 100-03, Chapter 1, Part 4, section 310.1. It is not a new clinical-trial coverage policy.

Under NCD 310.1, Medicare covers the routine costs of qualifying clinical trials, as well as reasonable and necessary items and services used to diagnose and treat complications arising from trial participation. Routine costs include items and services otherwise generally available to Medicare beneficiaries outside of a trial. However, NCD 310.1 explicitly carves out three major exclusions:

1. The investigational item or service itself, unless otherwise covered outside of the clinical trial.

2. Items and services provided solely to satisfy data collection and analysis needs that are not used in the direct clinical management of the patient (e.g., protocol-mandated pharmacokinetic blood draws or specialized research imaging).

3. Items and services customarily provided by the research sponsors free of charge for any enrollee in the trial.

NCD 310.1 states clearly: if an item or service is the focus of a qualifying clinical trial and is not covered by a national non-coverage policy, the routine costs of the trial are covered, but the non-covered item itself will not be. Therefore, when a Medicare administrative contractor denies reimbursement for an investigational agent, that denial is the standard statutory result. It is not an unexpected barrier, and it provides zero evidence that routine background therapy was compromised.

In commercial and employer-sponsored insurance, a parallel protection exists under federal law at 42 U.S.C. 300gg-8, codified from Section 2709 of the Public Health Service (PHS) Act as added by the Patient Protection and Affordable Care Act (ACA). Under 42 U.S.C. 300gg-8, a group health plan or health insurance issuer offering group or individual coverage may not deny a qualified individual participation in an approved clinical trial, may not deny or limit coverage of routine patient costs, and may not discriminate against the individual. However, the statutory scope contains rigorous boundaries:

• Disease Indication Scope: An 'approved clinical trial' is strictly defined under 42 U.S.C. 300gg-8(d) as a Phase I, II, III, or IV trial conducted in relation to the prevention, detection, or treatment of cancer or other life-threatening disease or condition. Trials in non-life-threatening chronic indications are not covered by this statutory mandate.

• Exclusions Parallel to Medicare: Routine patient costs do not include the investigational item, device, or service itself; research-only data collection procedures; or services clearly inconsistent with widely accepted and established standards of care.

• Network Limitations: Subsection (c) explicitly clarifies that the statute does not require a plan to provide out-of-network routine patient care benefits unless the plan otherwise provides out-of-network coverage.

The original decision asset is an access-event chronology with seven fields. The table that follows is labeled hypothetical. The rows are internally consistent worked examples; they are not an actual trial, coverage determination, missed-dose finding, or ICE classification. Leave unknown fields unknown. Hand the completed chronology to the protocol or SAP owner named in the protocol. The worksheet does not invent a new estimand. Record these seven fields for each access disruption:

1. Therapy Identity & Setting: Exact compound name and its protocol-defined role (investigational product vs protocol-mandated background therapy vs supportive routine care).

2. Supply Responsibility: Identifies the responsible supply chain (sponsor-provided clinical trial material under IND accountability vs commercial pharmacy distribution billed to participant insurance).

3. Authorization / Routing Status: The administrative payer status (e.g., prior authorization pending, pharmacy benefit lockout, claim denied under NCD 310.1, grandfathered plan denial, or unknown).

4. Evidence of Actual Exposure: Contemporaneous clinical observations (subject dosing diary, pill counts, nursing infusion records, clinic supply bridge, or explicitly unknown).

5. Scheduled Endpoint Timing: Status of the corresponding protocol assessment (assessment collected on schedule, visit missed, or assessment not yet due).

6. Remaining Data Gaps: Explicit enumeration of missing clinical facts that require site query resolution (e.g., unreturned blister packs or unverified pharmacy refill dates).

7. Protocol / SAP Action: Named protocol or SAP owner who applies the already-locked ICE or missing-data rule. The worksheet itself does not classify the event.

The four labeled hypothetical scenarios below match the table rows. They cannot be mistaken for an actual trial, a coverage determination, a missed-dose finding, or an ICE classification.

Hypothetical Scenario A: Prior-Authorization Lag with Continuous Home Supply. A labeled hypothetical: a participant receiving protocol-required oral background therapy has a commercial prior-authorization review pending, and the specialty pharmacy holds the next refill. Site source review shows remaining home supply from prior fills and contemporaneous evidence that daily doses continued. The scheduled endpoint visit is completed in window. Authorization status is recorded as pending. Actual-dose evidence is recorded separately. Remaining unknown: the date any re-authorization, if it occurs, will be confirmed. The chronology is handed to the named protocol or SAP owner. The pending authorization is not itself a missed dose and is not itself an intercurrent event.

Hypothetical Scenario B: Pharmacy-to-Medical Benefit Routing Lockout with Omitted Doses. A labeled hypothetical: protocol-required background biologic therapy is routed from the pharmacy benefit to the medical benefit, and specialty-pharmacy fills stop while a medical-benefit authorization is pending. Adjacent benefit-channel background is in PharmaDossier's pharmacy-benefit lockout article on medical-versus-pharmacy channel shifts, a route change can lock a product out of the pharmacy channel while medical-benefit review proceeds. In this worked example, the participant reports missed background doses during that lockout, and the scheduled primary-endpoint visit is completed. The site records authorization status as a pharmacy-benefit lockout, records the participant-reported missed doses as actual-exposure evidence, and records that the endpoint variable was collected. Remaining unknown: the date any medical-benefit bridge, if arranged, actually started. The chronology is handed to the named protocol or SAP owner to apply the locked treatment-attribute rule. Because the endpoint visit exists, the default is not to treat the observation as missing data. The lockout is not, by itself, one intercurrent event.

Hypothetical Scenario C: Investigational Product Claim Rejection Under NCD 310.1. A labeled hypothetical: a hospital billing office submits a Medicare claim for the investigational product. The contractor denies the claim under NCD 310.1 because the investigational item itself is not a routine cost unless otherwise covered outside the trial. Sponsor accountability and site pharmacy dispensing plus administration records document that the protocol-specified investigational doses were provided. The scheduled visit procedures are collected. The denial is recorded as an administrative coverage field. Actual IP use is recorded from 21 CFR 312.62(a) disposition and ICH E6(R3) 2.10.4 use records, not from the claim status. The chronology is handed to the named protocol or SAP owner. The expected non-coverage of the investigational item is not proof that investigational doses were missed.

Hypothetical Scenario D: Both Exposure Evidence and the Endpoint Visit Remain Unknown. A labeled hypothetical: a commercial claim for concomitant standard-of-care therapy is denied, and the plan is recorded only as possibly grandfathered under 45 CFR 147.140. The participant does not bring a dosing diary, pharmacy-fill history is not in the source file, and the scheduled endpoint visit is not completed. Actual-dose evidence is unknown. Endpoint timing is unknown. Remaining unknowns stay unknown. The site issues a source query and hands the incomplete chronology to the named protocol or SAP owner. The team does not invent missed-dose dates, does not classify an intercurrent event from the denial, and does not fill the blank visit with a post-hoc estimator. Adjacent SAP-locking work is described in what endpoint definitions the SAP must lock before unblinding; that article is not this source-field worksheet.

The following diagram is a record-keeping path, not an ICE classifier. It separates therapy identity, actual-dose evidence, and authorization status, then hands the chronology to the named protocol or SAP owner.

flowchart TD
    A["Access disruption noted as an administrative fact"] --> B{"Record therapy identity"}
    B -->|"Investigational product"| C["Record IP accountability separately from any claim status"]
    B -->|"Background or routine care"| D["Record actual-dose evidence separately from authorization status"]
    B -->|"Unknown class"| E["Leave therapy class unknown"]
    C --> F["Leave unknown dose or visit fields unknown"]
    D --> F
    E --> F
    F --> G["Hand the chronology to the named protocol or SAP owner"]
    G --> H["Do not invent an ICE, a missed-dose finding, or a new estimand from the coverage decision"]

Operational execution requires strict demarcation of functional roles to maintain audit readiness under ICH E6(R3) clinical quality and audit trail standards:

• Site Clinical Research Coordinator (CRC) & Principal Investigator (PI): Responsible for conducting contemporaneous participant interviews, collecting paper/electronic dosing diaries, performing physical pill counts of returned background bottles, and documenting factual exposure in the primary source record. Never record 'treatment discontinued' based solely on a pharmacy fax.

• Clinical Research Associate (CRA / Monitor): Performs source data verification (SDV) between participant dosing logs and CRF entries. Ensures that administrative denial correspondence in regulatory binders is not inappropriately transcribed as clinical dose omissions.

• Clinical Data Management (CDM): Configures electronic CRFs to capture exposure dates, dose modifications, and reason categories separately from reimbursement flags. Issues formal queries when discrepancies exist between visit attendance and drug administration logs.

• Trial Biostatistician & Protocol Steering Committee: Applies the locked statistical analysis plan rules to evaluate whether verified background interruptions meet the protocol-specified criteria for an intercurrent event. Preserves blind integrity and ensures sensitivity analyses align with the pre-unblinding SAP.

Trial teams should keep four methodological boundaries in view:

1. Never Infer Clinical Exposure from Administrative Notices. A payer denial letter, a pharmacy benefit rejection code, or an automated prior-authorization expiration notice reflects administrative billing eligibility, not physical drug consumption. Factual exposure must be established through direct clinical source evidence, including pill counts, subject logs, and verified clinic administrations.

2. No Post-Hoc Estimand Modifications. If access barriers affect multiple participants, the team cannot retroactively alter the estimand or invent a new ICE strategy after seeing denials. ICH E9(R1) states that a change to the estimand should usually be reflected through amendment to the protocol. A different job—reviewing endpoint observations after analyzer service—is covered in MedDeviceRepair: assessing endpoint data after analyzer service; it is not this access-event chronology.

3. Respect Statutory Coverage Limits. Do not tell sites or participants that a payer must pay for investigational therapy. NCD 310.1 and 42 U.S.C. 300gg-8 exclude the investigational item from specified routine-cost protections. 45 CFR 147.140 states that PHS Act 2709 does not apply to grandfathered health plan coverage. Those non-coverage results are not a reason to skip investigational-product accountability, and they are not missed-dose findings.

4. Preserve traceable source-record discrepancies. When a participant's self-reported dosing differs from pharmacy refill records, record both statements contemporaneously and explain the discrepancy without obscuring the original entry. Do not rewrite source to match a billing statement. ICH E6(R3) 2.12.2 and 2.12.6 require traceable source records and that reported data match source or that discrepancies be explained.