Minimum Dietary Diversity for Women: Partitioning Misclassifications by Proxy Data Collection Methods using Weighed Food Records as the Reference in Ethiopia.

Background: Nonquantitative list-based or open 24-h recalls (24-HRs) have been shown to overestimate the prevalence of Minimum Dietary Diversity for Women (MDD-W), as compared with direct quantitative observations. However, the main sources of error are unknown. Objectives: To assess the measurement agreement of proxy data collection methods for MDD-W, as compared with weighed food records (WFRs). Methods: Applying a noninferiority design, data were collected from 431 nonpregnant females in Ethiopia. MDD-W estimates from both proxy data collection methods were compared with the WFR prevalence by McNemar's chi-square tests, Cohen's Kappa, and receiver operator characteristic analyses. Ten-point food group diversity scores (FGDS) were compared by Bland-Altman plots, Wilcoxon matched-pairs tests, and weighted Kappa. Food group misclassifications were partitioned into errors related to respondent biases or the questionnaire development. Results: List-based and open 24-HRs overreported MDD-W by 8 and 4 percentage points, respectively, as compared with WFR (objective MDD-W prevalence: 8%). Furthermore, list-based 24-HRs overestimated FGDS by 0.4 points (limits of agreement [LOA]: -1.1, 2.0), whereas open 24-HRs led to a 0.3 point (LOA: -1.2, 1.7) overestimate. Food groups most likely to be misreported using proxy data collection methods were "pulses," "nuts and seeds," "dairy products," and "other fruits." Underreporting of consumption occurred among <4% of females for all food groups. Furthermore, respondent biases were the predominant cause of food group overreporting, except for the "pulses" and "other vegetables" food groups, where food items incorrectly included on the food list were the main source of errors. Conclusions: Food group consumption misclassifications by proxy data collection methods were mainly attributable to females overreporting consumption because of respondent biases or the criterion for foods to be counted, rather than the suboptimal development of the food list in Ethiopia. To obtain precise and accurate MDD-W estimates at the (sub)national level, rigorous context-specific food list development, questionnaire pilot testing, and enumerator training are recommended to mitigate identified biases.

A review of the peak-end rule in mental health contexts.

The peak-end rule, a memory heuristic in which the most emotionally salient part of an experience (i.e., peak) and conclusion of an experience (i.e., end) are weighted more heavily in summary evaluations, has been understudied in mental health contexts. The recent growth of intensive longitudinal methods has provided new opportunities for examining the peak-end rule in the retrospective recall of mental health symptoms, including measures often used in measurement-based care initiatives. Additionally, principles of the peak-end rule have significant potential to be applied to exposure-based therapy procedures. Additional research is needed to better understand the contexts in which, and persons for whom, the peak-end rule presents a greater risk of bias, to ultimately improve assessment strategies and clinical care.

Statistical Analysis of Self-Reported Health Conditions in Cohort Studies: Handling of Missing Onset Age.

OBJECTIVE: This paper discusses methodological challenges in epidemiological association analysis of a time-to-event outcome and hypothesized risk factors, where age/time at the onset of the outcome may be missing in some cases, a condition commonly encountered when the outcome is self-reported. STUDY DESIGN AND SETTING: A cohort study with long-term follow-up for outcome ascer- tainment such as the Childhood Cancer Survivor Study (CCSS), a large cohort study of 5-year survivors of childhood cancer diagnosed in 1970-1999 in which occurrences and age at onset of various chronic health conditions (CHCs) are self-reported in surveys. Simple methods for handling missing onset age and their potential bias in the exposure-outcome association infer- ence are discussed. The interval-censored method is discussed as a remedy for handling this problem. The finite sample performance of these approaches is compared through Monte Carlo simulations. Examples from the CCSS include four CHCs (diabetes, myocardial infarction, osteoporosis/osteopenia, and growth hormone deficiency). RESULTS: The interval-censored method is usable in practice using the standard statisti- cal software. The simulation study showed that the regression coefficient estimates from the 'Interval censored' method consistently displayed reduced bias and, in most cases, smaller stan- dard deviations, resulting in smaller mean square errors, compared to those from the simple approaches, regardless of the proportion of subjects with an event of interest, the proportion of missing onset age, and the sample size. CONCLUSION: The interval-censored method is a statistically valid and practical approach to the association analysis of self-reported time-to-event data when onset age may be missing. While the simpler approaches that force such data into complete data may enable the standard analytic methods to be applicable, there is considerable loss in both accuracy and precision relative to the interval-censored method.

Differential recall bias in estimating treatment effects in observational studies.

Observational studies are frequently used to estimate the effect of an exposure or treatment on an outcome. To obtain an unbiased estimate of the treatment effect, it is crucial to measure the exposure accurately. A common type of exposure misclassification is recall bias, which occurs in retrospective cohort studies when study subjects may inaccurately recall their past exposure. Particularly challenging is differential recall bias in the context of self-reported binary exposures, where the bias may be directional rather than random and its extent varies according to the outcomes experienced. This paper makes several contributions: (1) it establishes bounds for the average treatment effect even when a validation study is not available; (2) it proposes multiple estimation methods across various strategies predicated on different assumptions; and (3) it suggests a sensitivity analysis technique to assess the robustness of the causal conclusion, incorporating insights from prior research. The effectiveness of these methods is demonstrated through simulation studies that explore various model misspecification scenarios. These approaches are then applied to investigate the effect of childhood physical abuse on mental health in adulthood.

Validation of mobile phone use recall in the multinational MOBI-kids study.

Potential differential and non-differential recall error in mobile phone use (MPU) in the multinational MOBI-Kids case-control study were evaluated. We compared self-reported MPU with network operator billing record data up to 3 months, 1 year, and 2 years before the interview date from 702 subjects aged between 10 and 24 years in eight countries. Spearman rank correlations, Kappa coefficients and geometric mean ratios (GMRs) were used. No material differences in MPU recall estimates between cases and controls were observed. The Spearman rank correlation coefficients between self-reported and recorded MPU in the most recent 3 months were 0.57 and 0.59 for call number and for call duration, respectively. The number of calls was on average underestimated by the participants (GMR = 0.69), while the duration of calls was overestimated (GMR = 1.59). Country, years since start of using a mobile phone, age at time of interview, and sex did not appear to influence recall accuracy for either call number or call duration. A trend in recall error was seen with level of self-reported MPU, with underestimation of use at lower levels and overestimation of use at higher levels for both number and duration of calls. Although both systematic and random errors in self-reported MPU among participants were observed, there was no evidence of differential recall error between cases and controls. Nonetheless, these sources of exposure measurement error warrant consideration in interpretation of the MOBI-Kids case-control study results on the association between children's use of mobile phones and potential brain cancer risk.

Do Patients Accurately Recall Pain Levels Following Sacroiliac Joint Steroid Injection? A Cohort Study of Recall Bias in Patient-reported Outcomes.

BACKGROUND: Sacroiliac joint (SIJ) injections are crucial in the diagnostic toolkit for evaluating SIJ pathology. Recall bias is an important component in patient-reported outcomes that has not been well studied in SIJ injection. OBJECTIVE: The purpose of this study was to characterize the accuracy, direction, and magnitude of pain level recall bias following SIJ steroid injection and study the factors that affect patient recollection. STUDY DESIGN: Prospective cohort study. SETTING: Level 1 academic medical center. METHODS: Using standardized questionnaires, baseline Numeric Rating Scale (NRS-11) scores were recorded for patients undergoing SIJ steroid injections at preinjection, at 4 hours postinjection, and at 24 hours postinjection. At a minimum of 2 weeks postinjection, patients were asked to recall their preinjection, 4-hour, and 24-hour postinjection NRS-11 scores. Actual and recalled NRS-11 scores were compared using paired t tests for each time interval. Multivariable linear regression was used to identify factors that correlated with consistent recall. RESULTS: Sixty patients with a mean age of 66 years (65% women) were included. Compared to their preinjection pain score, patients showed considerable improvement at both 4 hours (mean difference [MD] = 3.28; 95% CI, 2.68 - 3.89), and 24 hours (MD = 3.23; 95% CI, 2.44 - 4.03) postinjection. Patient recollection of preinjection symptoms was more severe than actual (MD = 0.65; 95% CI, 0.31 - 0.99). Patient recollection of symptoms was also more severe than actual at 4 hours (MD = 0.50; 95% CI .04 - 1.04) as well as at 24 hours postinjection (MD = 0.80; 95% CI, 0.16 - 1.44). The magnitude of recall bias was mild and did not exceed the minimal clinically important difference. There was a moderate correlation between actual and recalled pain levels when comparing preinjection with the 4-hour postinjection NRS-11 score (correlation coefficient [r] =0.64; P < 0.001) and moderate correlation when comparing preinjection with the 24-hour postinjection NRS-11 score (r = 0.62; P < 0.001). Linear regression models showed that at preinjection, patients with a lower body mass index and the presence of coexisting psychiatric diagnoses were better at recalling their pain (P < 0.05). Patients with a higher body mass index also experienced less pain relief when comparing preinjection with the 4-hour postinjection NRS-11 score (P < 0.05). LIMITATIONS: Recall pain scores were obtained via telephone surveys, which can lead to interview bias. One patient died, and 3 were lost to follow-up. We did not control for patient use of adjunctive pain relief modalities, which may modulate the overall response to injection. SIJ injections can also be diagnostic, so some patients may not have shared the same indication for injection or pain-generating diagnosis. CONCLUSIONS: Patients had favorable pain level responses to their SIJ steroid injection for both actual and recall surveys. Although patients demonstrated poor recall of absolute pain scores at preinjection, 4-hour postinjection, and 24-hour postinjection, they demonstrated robust recall of their net pain score improvement at both 4- and 24-hours postinjection. These findings suggest that there is utility in using patient recollection to describe the magnitude of pain relief following treatment for sacroiliac joint dysfunction.

Evaluation of measurement errors in the Patient-Oriented Eczema Measure (POEM) outcome.

BACKGROUND: The Patient-Oriented Eczema Measure (POEM) is the recommended core outcome instrument for atopic dermatitis (AD) symptoms. POEM is reported by recalling the presence/absence of seven symptoms in the last 7 days. OBJECTIVE: To evaluate measurement errors in POEM recordings due to imperfect recall. METHODS: Using data from a clinical trial of 247 AD patients aged 12-65 years, we analysed the reported POEM score (r-POEM) and the POEM derived from the corresponding daily scores for the same seven symptoms without weekly recall (d-POEM). We quantified recall error by comparing the r-POEM and d-POEM for 777 patient-weeks collected from 207 patients, and estimated two components of recall error: (1) recall bias due to systematic errors in measurements and (2) recall noise due to random errors in measurements, using a bespoke statistical model. RESULTS: POEM scores have a relatively low recall bias, but a high recall noise. Recall bias was estimated at 1.2 points lower for the r-POEM on average than the d-POEM, with a recall noise of 5.7 points. For example, a patient with a recall-free POEM of 11 (moderate) could report their POEM score anywhere from 5 to 14 (with 95% probability) because of recall error. Model estimates suggested that patients tend to recall itch and dryness more often than experienced (positive bias of less than 1 day), but less often for the other symptoms (bleeding, cracking, flaking, oozing/weeping and sleep disturbance; negative bias ranging 1-4 days). CONCLUSIONS: In this clinical trial data set, we found that patients tended to slightly underestimate their symptoms when reporting POEM, with significant variation in how well they were able to recall the frequency of their symptoms every time they reported POEM. A large recall noise should be taken into consideration when interpreting POEM scores.

Association analysis of self-reported outcomes with a validated subset.

In health-science research, outcomes ascertained through surveys and interviews are subject to potential bias with respect to the true outcome status, which is only ascertainable with clinical and laboratory assessment. This measurement error may lead to biased inference when evaluating associations between exposures and outcomes of interest. Here, we consider a cohort study in which the outcome of interest is ascertained via questionnaire, subject to imperfect ascertainment, but where a subset of participants also have a clinically assessed, validated outcome available. This presents a methodological opportunity to address potential bias. Specifically, we constructed the likelihood in two parts, one using the validated subset and the other using a subset without validation. This work expands on that proposed by Pepe and enables inference with standard statistical software. Weighted generalized linear model estimates for our method and maximum likelihood estimates (MLE) for Pepe's method were computed, and the statistical inference was based on the standard large-sample likelihood theory. We compare the finite sample performance of two approaches through Monte Carlo simulations. This methodological work was motivated by a large cohort study of long-term childhood cancer survivors, allowing us to provide a relevant application example where we examined the association between clinical factors and chronic health conditions.

Recall Bias in the Assessment of Cough for Patients Discharged from Lung Surgery.

Purpose: This study aimed to evaluate the presence of recall bias when patients retrospectively report cough scores. Patients and Methods: Patients who underwent lung surgery between July 2021 and November 2021 were recruited for this study. We retrospectively assessed the severity of cough within the past 24 hours and the past 7 days using a 0-10 numerical rating scale. Recall bias was defined as the difference between the scores reported on the two assessments. Patients were grouped based on the longitudinal change in cough scores from pre-operation to 4 weeks after discharge using group-based trajectory models. Using generalized estimating equation to explore the factors influencing recall bias. Results: Overall, 199 patients were analyzed and demonstrated the three distinct trajectories of post-discharge cough: high (21.1%), medium (58.3%), and low (20.6%). Significant recall bias was found in week 2 for the high-trajectory patients (6.26 vs 5.10, P<0.01) and in week 3 for the medium-trajectory patients (2.88 vs 2.60, P=0.01). Among all recall bias, 41.8% were of underestimation, and 21.7% of overestimation. The high trajectory group (ß=1.14, P<0.01) and measurement interval (ß=0.36, P<0.01) were risk factors for underestimation, while post-discharge time (ß=-0.57, P<0.01) and measurement interval (ß=-0.13, P=0.02) were protective factors for overestimation. Conclusion: Retrospective assessment of post-discharge cough in patients who underwent lung surgery will introduce recall bias, with a tendency of underestimation. The high-trajectory group, interval time and post-discharge time are influencing factors of recall bias. For patients with severe cough at discharge, a shorter recall periods should be employed for monitoring, due to the large bias that results from a longer recall period.

Wide Awake Local Anesthesia No Tourniquet Surgery of Carpal Tunnel Syndrome: Patients' Experience and Recall Bias in a Day-Care Setting.

Background and Objective: Wide-Awake Local Anesthesia No Tourniquet (WALANT) is a technique of local anesthesia commonly used in the surgical treatment of a wide variety of conditions affecting the upper extremity, including Carpal Tunnel Syndrome (CTS). The recent retrospective studies investigated patient experiences in a wide variety of hand disorder-related cases. The aim of our study is to evaluate patient satisfaction regarding open surgical treatment for CTS using the WALANT technique. Material and Methods: we enrolled 82 patients with CTS without medical record of surgical treatment for CTS. For WALANT, a hand surgeon used a combination of 1:200,000 epinephrine, 1% lidocaine, and 1 mL 8.4% sodium bicarbonate solution without tourniquet application and sedating the patient. All patients were treated in a day-care setting. For assessment of patient experience, Lalonde's questionnaire was adapted. Participants completed survey twice: one month and six months after the surgical treatment was performed. Results: the median pre-operative pain score for all patients was 4 (range 0-8) after one month and 3 (range 1-8) after six months. The median intraoperative pain score for all patients was 1 (range 0-8) after one month and 1 (range 1-7) after six months. The median post-operative pain score for all patients was 3 (range 0-9) after one month and 1 (range 0-8) after six months. More than half (61% after one month and 73% after six months) of the patients responded by stating that their real experience of WALANT was better than their initial expectations. An absolute majority of patients (95% after one month and 90% after six months) would recommend WALANT treatment to their relatives. Conclusions: overall, patient satisfaction with treatment for CTS using WALANT is high. Furthermore, complications related to the performed treatment and persistent post-operative pain could be associated with more reliable patient recall of this healthcare intervention. A longer period of time between intervention and assessment of patient experience could possibly be a reason for recall bias.

Do Patients Accurately Recall Their Preoperative Symptoms After Elective Orthopedic Procedures?

Patient-reported outcome measures are a frequent tool used to assess orthopedic surgical outcomes. However, recall bias is a potential limitation of these tools when used retrospectively, as they rely on patients to accurately recall their preoperative symptoms. A database search of Cochrane Library, PubMed, Medline Ovid, and Scopus until May 2021 was completed in duplicate by two reviewers. Studies considered eligible for inclusion were those which reported on patient recall bias associated with orthopedic surgery. The primary outcome of interest investigated was the accuracy of patient recollection of preoperative health status. Any factors that were identified as affecting patient recall were secondary outcomes of interest. Of the 4,065 studies initially screened, 20 studies with 3,454 patients were included in the final analysis. Overall, there were 2,371 (69%) knee and hip patients, 422 (12%) shoulder patients, 370 (11%) spine patients, 208 (6%) other upper extremity patients, and 83 (2%) foot and ankle patients. Out of the eight studies that evaluated patient recall within three months postoperatively, seven studies concluded that patient recall is accurate. Out of the 13 studies that evaluated patient recall beyond three months postoperatively, nine studies concluded that patient recall is inaccurate. The accuracy of patient recall of preoperative symptoms after elective orthopedic procedures is not reliable beyond three months postoperatively.

More than blood: app-tracking reveals variability in heavy menstrual bleeding construct.

BACKGROUND: Heavy menstrual bleeding (HMB) is associated with impaired quality of life and may signal serious health problems. Unresolved challenges in measuring menstrual bleeding and identifying HMB have hampered research and clinical care. Self-reported bleeding histories are commonly used but these may be influenced by recall bias, personal beliefs regarding "normal" flow volume, and the experience of other physical symptoms or disruptions to daily life. The potential usefulness of menstrual-tracking mobile applications, which allow real-time user-entered data recording, for assessing HMB has not been studied. We evaluated recall bias in reported period duration, the relationship of tracked period duration and daily flow volume to subsequently reported period heaviness, variation in quality of life associated with increasing period heaviness, and the advantages and limitations of using app-tracked data for clinical and research purposes. METHODS: An online questionnaire was distributed to current users of Clue, a commercially available menstrual health tracking app, asking them to characterize their last period. We compared responses to the user's corresponding Clue app-tracked data. The study sample comprised 6546 U.S.-based users (aged 18-45 years). RESULTS: Increasing reported heaviness was associated with increasing app-tracked period length and days of heavy flow, impaired quality-of-life (especially body pain severity), and disrupted activities. Of those reporting having had a heavy/very heavy period, ~ 18% had not tracked any heavy flow, but their period length and quality-of-life indicators were similar to those who had tracked heavy flow. Sexual/romantic activities were the most affected across all flow volumes. Compared to app-tracked data, 44% recalled their exact period length; 83% recalled within ± 1 day. Overestimation was more common than underestimation. However, those with longer app-tracked periods were more likely to underestimate period length by ≥ 2 days, a pattern which could contribute to under-diagnosis of HMB. CONCLUSION: Period heaviness is a complex construct that encapsulates flow volume and, for many, several other bleeding-associated experiences (period length, bodily impairments, disruptions of daily activities). Even very precise flow volume assessments cannot capture the multi-faceted nature of HMB as experienced by the individual. Real-time app-tracking facilitates quick daily recording of several aspects of bleeding-associated experiences. This more reliable and detailed characterization of bleeding patterns and experiences can potentially increase understanding of menstrual bleeding variability and, if needed, help to guide treatment.

Social value orientation modulates behavioral and neural responses to social influence.

Substantial studies have investigated the social influence effect; however, how individuals with different social value orientations (SVOs), prosocials and proselfs, respond to different social influences remains unknown. This study examines the impact of positive and negative social information on the responses of people with different SVOs. A face-attractiveness assessment task was employed to investigate the relationships between influence probability, memory, and event-related potentials of social influence. A significant interactional effect suggested that prosocials and proselfs reacted differently to positive (group rating was more attractive) and negative (group rating was less attractive) social influences. Specifically, proselfs demonstrated significantly higher influence probability, marginally better recall performance, smaller N400, and larger late positive potential on receiving negative influence information than on receiving positive influence information, while prosocials showed no significant differences. Overall, correlations between N400/LPP, influence probability, and recall performance were significant. The above results indicate the modulating role of SVO when responding to social influence. These findings have important implications for understanding how people conform and how prosocial behavior occurs.

Recall bias of students' affective experiences in adolescence: The role of personality and internalizing behavior.

INTRODUCTION: Adolescence is characterized by multiple biopsychosocial changes, associated with a high intraindividual variability of emotional experiences. Previous findings suggest that this intraindividual variability is reflected in a recall bias of adolescents' emotion reports. However, corresponding findings are scarce and inconclusive. Studies on predictors of recall bias in adulthood indicate that personality traits, especially neuroticism and extraversion, as well as specific internalizing disorders might affect recall bias of emotion reports. METHODS: The sample consists of 118 Swiss adolescent students in grade 8 and 9 (Mage = 15.15, SDage = 0.89). The students' momentary affective experience was recorded using smartphones over seven consecutive days in situ at 42 randomly generated occasions (six per day), with a total of 1059 protocols on current events. At the end of the experience-sampling phase, students filled out an online questionnaire, providing information about their personality and typical behavior as well as their retrospective affective experience. In addition, the students' behavior was evaluated by their teachers. We applied two-level structural equation modeling with latent difference variables. RESULTS: Adolescents high in extraversion showed retrospective overestimation of positive affective experiences and underestimation of negative affective experiences. Adolescents with high neuroticism tended to overestimate negative affect retrospectively, showing no significant effects for positive affect. However, internalizing behavior did not predict a negative recall bias in adolescents' affective experience. CONCLUSIONS: Retrospective self-reports about adolescents' affective experience are biased by relatively stable individual factors, whereas less stable individual factors did not seem to have any influence.

Oncofertility research pitfall? Recall bias in young adult cancer survivors.

Objective: To assess recall bias by evaluating how well female cancer survivors remember details regarding their cancer diagnosis, treatment, and fertility preservation (FP) counseling.Oncofertility literature cites recall bias as a pitfall of retrospective surveys, but limited data exist to quantify this bias. Design: Retrospective secondary analysis of cross-sectional survey data. Setting: Single academic medical center. Patients: Female oncology patients of reproductive age, 18-44 years old, at least 6 months past their last chemotherapy treatment. Interventions: Not applicable. Main Outcome Measures: Recall of details surrounding cancer diagnosis and chemotherapy regimens, recall of FP counseling and ovarian reserve testing, and rates of chart-documented FP counseling. Results: In total, 117 patients completed the survey, with 112 verified via chart review. When asked to report the chemotherapy regimen, 57% (64 of the 112) marked "I don't know/prefer not to say." Regarding FP, 80% (90 of the 112) denied being offered counseling. Of the 37 (33%) who had documented FP conversations, 13 (35%) did not recall mention of fertility. Only 2 of 8 patients with ovarian reserve testing recalled this being performed at their initial visit. Multivariable logistic regression revealed older age was significantly associated with not being offered FP (odds ratio [OR] 0.87). Conclusions: Our results confirm that the accuracy of oncology patients' reporting is limited by a poor recall, particularly regarding their specific chemotherapy regimen. More than 1 in 3 patients documented to have been offered FP counseling do not recall this discussion. Importantly, only one-third of cancer survivors had chart-documented FP counseling. Increased efforts are needed to ensure adequate follow-up beyond the initial visit.

Has the COVID-19 pandemic changed household food management and food waste behavior? A natural experiment using propensity score matching.

Household food management behavior changed considerably during the COVID-19 pandemic. A growing body of work has quantified the impact of lockdowns on household food waste. Yet, previous studies used a retrospective study design which undermines the accuracy of the causal effect on household food waste. This paper investigates the causal impact of the COVID-19 pandemic on household food management and food waste using a natural experiment approach. Using two large national-scale longitudinal data sets (n = 8157), this study quantifies the impact of COVID-19 on food waste and food behavior of Australian households. Propensity score matching (PSM) was carried out to address potential endogeneity issues and to select control and treatment groups for analysis. Findings reveal that Australian households reduced food waste by 9% on average in 2020 (during COVID-19) compared to the pre-pandemic (2019) level. The use of a grocery list, discount purchases, and 'just-in-case' purchases, and food refrigeration have recorded a marked increase during the pandemic compared to pre-pandemic times. The changes to food management and food discard behavior during the pandemic offer important insights for behavior change campaigns to reduce household food waste. Interventions to sustain good food planning and storage practices and involving food retailers are promising entry points in addressing household food waste. The study also highlights the considerable challenge in achieving SDG 12.3 target by 2030.

Recall Bias Affects Pain Assessment in Knee Osteoarthritis: A Pilot Study.

OBJECTIVE: The objective of this study was to evaluate the recall bias of symptoms evaluation in knee osteoarthritis (OA). DESIGN: In this multicentric pilot study, 50 patients with knee OA used a mobile App (Ecological Momentary Assessment [EMA]) to collect pain and function on two 0 to 10 numerical rating scales (NRS) 2 times a day for 2 months. At the 1-month and at the 2-month follow-up visits, patients retrospectively evaluated the mean level of pain/function of the last month. Recall bias was computed as the difference between the mean level of pain/function reported using the App and the level reported with the retrospective assessment. The correlation between the recall bias and patients' characteristics, as well as pain/function trajectories, was analyzed. RESULTS: A statistically significant recall bias was documented with higher pain reported at 1-month with the retrospective assessment (P < 0.001). These results were confirmed also at the 2-month follow-up (P = 0.002). For function, no significant recall bias was documented. During the first and second months, 47 and 31 patients showed pain peaks, respectively. The number of pain peaks during the first month was correlated with the magnitude of the recall bias (P = 0.02). CONCLUSIONS: The recall bias influences the retrospective self-assessment of pain at the follow-up visits and the presence of pain peaks, a common event in the patients with OA, increases the magnitude of recall bias. The EMA performed with a mobile App is a useful tool to limit the influence of recall bias in the clinical and research setting evaluation of knee OA.