How FDA Evaluates Recall Periods for Patient-Reported Symptom Outcomes: 24-Hour, 7-Day, and Daily Diary Designs

How FDA evaluates recall periods (24-hour, 7-day, daily diary) for patient-reported symptom outcomes

FDA has no default recall period for patient-reported outcome (PRO) symptom measures. Across guidance, clinical outcome assessment (COA) reviews and approved labeling, the agency asks the same few questions every time. Can patients accurately recall the concept over the stated window? Does the window fit how the symptom fluctuates and how the drug is dosed? Do diary procedures make each entry actually cover the intended interval? And does the recall period of the symptom measure match the recall period of the anchors used to interpret it? For daily-varying symptoms, short recall (past 24 hours, or "today") collected in a daily diary is the most common outcome. Reviewers have still accepted longer windows where evidence supported them, and have criticized 24-hour designs that were badly implemented.

Key takeaways

• Recall period is a content-validity question. FDA expects sponsors to justify it based on patients' ability to recall validly, the instrument's purpose, the concept, the condition and the treatment ³⁰⁹³¹⁰.
• Short recall is generally preferred for fluctuating symptoms. Questions that ask patients to remember long periods, compare against an earlier state, or average experience over time are likely to undermine content validity ¹⁷.
• A 7-day recall is not categorically ruled out. FDA has, however, asked sponsors for qualitative evidence justifying 7-day over 24-hour recall ¹⁰⁶⁴⁸.
• A daily diary does not guarantee good data. Reviewers look at completion windows, lockouts, retrospective entry, and how many days are needed to compute a weekly score ^12151211.
• Recall alignment between a PRO and its anchor affects labeling. Mismatches have limited the interpretation of meaningful-change thresholds and moved claims toward descriptive language ^8486117.

The guidance framework

The 2009 PRO guidance

The 2009 guidance Patient-Reported Outcome Measures: Use in Medical Product Development to Support Labeling Claims defines the recall period as the time patients are asked to consider when answering an item. It can be momentary (real time) or retrospective. FDA says it will look for steps taken to make sure patients understood the recall period ³⁰⁹³¹⁰. The guidance generally favors current or recent state over extended recall. It warns that recall-based answers may be influenced by how the patient feels at the moment of recall, and that differential recall can add noise and obscure treatment effects ¹⁷. When detailed retrospective reporting is needed, FDA suggests asking about the worst or best experience during the period, or using a diary ³¹¹. For unsupervised diaries, FDA will check that the protocol makes patients complete entries as the design requires, and not retrospectively just before a clinic visit ¹².

PFDD Guidance 3 (fit-for-purpose COAs)

The Patient-Focused Drug Development guidance Selecting, Developing, or Modifying Fit-for-Purpose Clinical Outcome Assessments repeats that no single recall period applies to every PRO. The period should be clearly specified, short enough to limit recall-related error or bias, and still manageable for respondents. It should also suit how frequent and memorable the events are ¹⁰⁵¹⁰⁶. FDA recommends discussing the planned period with the agency. Patient interviews can show how accurately patients recall different windows, and sponsors may need evidence comparing longer and shorter recall periods ¹⁰⁵¹⁰⁶.

Indication-specific guidances

Many disease-specific guidances turn these principles into concrete recommendations:

Two distinctions in these documents matter for endpoint construction. First, the IBS guidance's "worst abdominal pain in the past 24 hours" is the daily recall question, while the responder definition uses the weekly average of those daily scores. Averaging does not turn a daily worst-pain rating into an average-severity rating ¹²⁵. Second, the allergic rhinitis guidance separates reflective scores (severity over a predefined interval, such as 12 hours) from instantaneous scores (severity just before the next dose) and links each to a different efficacy question ¹²⁶¹²⁷.

How COA reviewers have evaluated 24-hour recall

Reviewers have not treated "past 24 hours" as automatically acceptable. They look for qualitative evidence that patients can use the window, and they check whether the administration schedule produces comparable 24-hour intervals.

Supported by patient interviews:

• Tremfya (guselkumab), Psoriasis Symptom and Sign Diary. In cognitive interviews, 18 of 19 participants felt they could recall symptoms accurately over 24 hours. Participants who compared versions said daily reporting would be more accurate than a 7-day version, though some preferred less frequent completion. The 24-hour version supported a major secondary endpoint ²²⁸¹⁴⁹.
• Omvoh (mirikizumab), Urgency NRS. Interviewees reported no difficulty with the past-24-hour window. A follow-up group with lower education levels also endorsed it, and the reviewer concluded the instrument appeared content valid for its proposed use ^229146147.
• Bimzelx (bimekizumab), HS Symptom Daily Diary. All 20 interviewees said they considered the 24-hour period and could recall it accurately, although two also thought about broader periods such as "bad days" ¹⁴⁴.
• Lucemyra (lofexidine), SOWS-Gossop. Interviews generally supported 24 hours as optimal for acute withdrawal severity, but FDA noted that the modified instructions, including the recall wording, lacked clarity ¹⁵⁰.
• Cosentyx (secukinumab), Psoriasis Symptom Diary. FDA agreed with the selected items and their 24-hour recall ²¹¹.

Implementation and design concerns:

• Palsonify (paltusotine), Acromegaly Symptom Diary. Each item asked about the past 24 hours, but the diary could be completed from 3 p.m. until 9 a.m. the next morning. FDA was concerned that entries made at different times would not represent the adjacent, non-overlapping 24-hour intervals the analysis assumed, which could reduce validity ¹⁵¹¹⁵².
• Ojjaara (momelotinib), MPN-SAF TSS. FDA recommended daily electronic reporting with past-24-hour recall. It preferred this to an undefined comparison with the period before disease onset, which could be hard to recall after longstanding disease ^154155156.
• Eohilia (budesonide), DSQ v5. Items that had referred to the past 24 hours were reworded to "today" after FDA feedback ¹⁵³.
• Emgality (galcanezumab), MSQ v2.1. The reviewer found a 4-week recall potentially appropriate for episodic migraine but suggested a shorter window, such as 24 hours, might be more meaningful in chronic migraine ¹⁶³.

How reviewers have evaluated 7-day and longer recall

Week-long and longer windows get the most scrutiny when the concept fluctuates, when patients must count events, or when patients must compare against a remembered baseline.

• Skyrizi (risankizumab), CSS. FDA asked for qualitative research in the target population to justify a 7-day recall instead of past 24 hours. It also recommended consistent diary completion times and an e-diary lockout to preserve the intended window ⁴⁸¹¹⁶.
• Addyi (flibanserin), FSFI desire domain. Reviewers concluded that neither 7-day nor 28-day recall was appropriate for sexual desire, which can change daily. The Division favored daily electronic capture with an entry window of no more than about 3 days ^45180181182. A separate review noted continued uncertainty: most participants preferred 1 to 4 weeks, while about one in five preferred 24 hours ¹¹⁵.
• Cabometyx (cabozantinib), EORTC QLQ-C30 and QLQ-BN20. FDA said it lacked evidence that asking patients to remember and average one week was appropriate for each concept, particularly fluctuating symptoms such as pain ¹⁸⁶.
• Amitiza (lubiprostone), weekly global IBS relief question. The main concern was the comparison with remembered baseline symptoms from as much as three months earlier, rather than the 7-day window itself. The reviewer said sequential, non-comparative 7-day severity ratings could have avoided this ⁵³.
• Dexilant (dexlansoprazole). The reviewer treated direct daily eDiary reports as the main efficacy source and less subject to bias. Seven-day recall could introduce bias, particularly in pediatric patients ²²¹²²².
• Varithena (polidocanol), VVSymQ. After FDA raised concerns about the original 7-day recall, the revised daily diary used past-24-hour worst-severity items and "last night" items for nighttime symptoms ²²³²²⁴.
• Emend (aprepitant), FLIE. FDA questioned whether extending recall to 5 days validly summarized a changing post-chemotherapy experience, because responses might reflect the patient's current state. It preferred a daily 24-hour diary, contingent on that version's validity being confirmed ^158159160219.
• Movantik (naloxegol), BLSRQ. A 2-week recall was considered too long for a questionnaire that asks patients to count bowel movements, occurrences and laxative use. Reviewers pointed to instruments using daily to prior-week recall ¹⁸⁵.
• Tysabri (natalizumab), IBDQ and SF-36. Two-week averaging and comparative items could increase error. The reviewer advised asking about current state, and said the SF-36's 1-week recall may jeopardize validity ¹⁸⁷¹⁸⁸.
• Otezla (apremilast), BSAS. A 4-week window may be too long for patients to remember the number of new ulcers or lesions ¹⁸³.
• Chantix (varenicline). Monthly recollection of smoking behavior could give a false sense of precision. FDA suggested daily diaries or IVRS ¹⁹¹.

Missing or inconsistent recall instructions are also a deficiency on their own. For Ajovy (fremanezumab), HIT-6 items 1 to 3 stated no period while items 4 to 6 used the last 4 weeks. FDA recommended a clearly specified, standardized recall period ¹⁹⁰.

Daily diaries: timing, lockouts and missing data

FDA reviews diary operations as part of measurement validity:

• Entry windows and lockouts. In Addyi, FDA limited desire entries to a 24-hour window because a mental state is more vulnerable to recall bias than an event. Satisfying-event entries could be made up to 3 days later ²⁰⁷²⁰⁸. A reviewer did not accept "diary fatigue" as a reason to replace daily capture with 28-day recall, noting that about 99% of captured events were entered within 72 hours ²⁰⁹. Guidance also suggests reasonable lockout windows to preserve the intended recall period ¹⁸¹⁹.
• Electronic features. FDA recommends reminders and time-stamped completion to support compliance and reduce missing data, plus a backup plan where applicable ¹⁴¹⁵¹⁶.
• Paired assessments. For Livmarli (maralixibat), FDA checked whether caregivers completed morning and evening ItchRO(Obs) entries together. Only three pairs were less than 6 hours apart, and the effect on data quality was judged minimal ²¹⁸.
• Forced-response design. In Palsonify, FDA warned that requiring all nine items before saving could discard the seven responses needed for the total score ²⁰⁵²⁴⁰.
• Minimum data rules for weekly scores. Cosentyx, Tremfya, Nemluvio and Anzupgo all required at least 4 of 7 daily entries to compute a weekly score. Weeks with fewer entries were treated as missing ^{211212213214215216217}. In Vyleesi, substantial missingness in the daily EDQ (over 31% and 36% of patients in the two studies had no more than three baseline-week entries) led reviewers to exclude weekly averages based on fewer than four observations ²⁴²²¹⁰.
• Daily data used to check longer recall. In Vyleesi, correlations between daily and 4-week desire items gave some support for the longer-recall desire endpoint. They could not show whether 4-week recall was appropriate for distress ¹⁶¹²¹⁰.

Recall alignment between PROs and anchors

A recurring issue in recent reviews is a mismatch between the recall period of the target PRO and that of the global impression anchor used to derive meaningful-change thresholds:

• Dupixent (dupilumab), eosinophilic esophagitis. The PGIS asked about the past week, but the DSQ total covered 14 days. FDA said this complicated anchor-based analyses, ran supplementary analyses on matching 7-day periods, and recommended that future PGIS recall match the DSQ period (for example, "past 14 days") ^11711851.
• Eohilia (budesonide). The 14-day DSQ assessment period did not match the 7-day severity anchor, leaving the proposed meaningful-change threshold unclear. FDA recommended a matching 14-day anchor in future studies ⁸⁶.
• Lynavoy (linerixibat). A past-24-hour PGI-S did not align with the one-week period of the WIS and WSS scores, which reduced its usefulness as a reference measure ¹⁵⁷. For a 7-day PBC-40 variant, the COA reviewer did not review the evidence dossier, so no FDA conclusion on that version is available ⁵⁴.
• Sofdra (sofpironium). The review identified a past-week PGI-S paired with an HDSM-Ax-7 assessment recalling "since waking yesterday" ⁵².

How recall periods appear in labeling

Recall period affects whether a PRO endpoint is interpretable enough to support a claim. In approved labels, the recall question and the aggregation period are described separately:

• Trulance (plecanatide), IBS-C. A responder is defined by "a decrease in the weekly average of worst abdominal pain in the past 24 hours score (measured daily) of at least 30% compared with baseline weekly average," captured in a daily electronic phone diary, and met in at least 6 of 12 weeks ²⁹⁷.
• Viberzi (eluxadoline), IBS-D. Eligibility and the primary composite endpoint are based on daily worst abdominal pain in the past 24 hours compared against a baseline weekly average, from electronic daily diaries ²⁹⁵²⁹⁶.
• Dupixent (dupilumab), prurigo nodularis. Patients rated "the intensity of their worst pruritus (itch) over the past 24 hours" on the WI-NRS. The endpoint was the proportion with a 4-point or greater reduction at Week 24 ²⁹³²⁹⁴.
• Amitiza (lubiprostone). The label describes a weekly global relief question covering the past week compared with before study entry ²⁶³. Addyi describes FSDS-R Question 13 with a 7-day recall as a secondary endpoint ²⁶².

Recall problems have also limited claims. For Nocdurna (desmopressin), FDA considered the NQoL's 2-week recall problematic for a fluctuating condition. That prompted development of the NI Diary, but neither instrument's results supported labeling claims, partly because the analyses were exploratory ⁸³. For Bimzelx in plaque psoriasis, the Patient Symptom Diary used previous-24-hour recall, but the anchor's recall did not match the endpoint assessment period. That limited threshold interpretation, and FDA supported a descriptive claim for itching, pain and scaling instead ⁸⁴⁵⁸⁸⁵. For Emend, FDA asked for evidence supporting the 5-day FLIE recall and cautioned that broad labeling wording about nausea and vomiting impact could mislead without supporting nausea data ¹⁵⁸¹⁵⁹.

Practical implications for sponsors

1. Justify the window with qualitative data. Interview evidence that patients can accurately use the window is what reviewers cite when they accept it (Tremfya, Omvoh, Bimzelx). Requests for qualitative justification follow when it is missing (Skyrizi) ^22822914448.
2. Match recall to symptom dynamics and dosing. Daily or 24-hour recall suits fluctuating symptoms and as-needed dosing. Event logs suit discrete, frequent episodes ^26792111.
3. Specify the aggregation separately. Define the daily recall question, the weekly averaging rule, and the minimum number of daily entries in advance ^125211213.
4. Engineer the diary to protect the recall period. Use consistent completion times, lockouts, reminders and time stamps. Avoid completion windows that create overlapping intervals ^4815114.
5. Align anchor recall periods with the target PRO. Mismatches have repeatedly reduced the interpretability of meaningful-change thresholds and the specificity of labeling claims ^1178684.