The Pica, ARFID, and Rumination Disorder Interview – ARFID Questionnaire (PARDI-AR-Q) is a 29-item measure used by clinicians to screen for avoidant/restrictive food intake disorder (ARFID) and to identify what is driving a person’s restricted eating. It comes in two forms: a self-report version for young people and adults aged 14 and over, and an informant version for parents and carers reporting on anyone aged 4 and over.
Avoidant/restrictive food intake disorder (ARFID) was introduced as a diagnosis in the DSM-5 (American Psychiatric Association, 2013) and sits within the feeding and eating disorders category. It describes restricted or avoidant eating serious enough to cause significant weight loss or faltering growth, nutritional deficiency, dependence on supplements or enteral feeding, or marked interference with psychosocial functioning. What sets it apart from anorexia nervosa and bulimia nervosa is the absence of weight and shape concern: a person with ARFID is not avoiding food to change their body, but because of how food tastes or feels, because they have little interest in eating, or because they fear something bad will happen if they do.
ARFID presents across the lifespan and frequently co-occurs with anxiety disorders, obsessive-compulsive disorder, and neurodevelopmental conditions — particularly autism, ADHD, and intellectual disability (American Psychiatric Association, 2022). Because those three drivers of avoidance call for different clinical responses, establishing which one predominates often matters as much as establishing the diagnosis.
The Pica, ARFID, and Rumination Disorder Interview – ARFID Questionnaire (PARDI-AR-Q; Bryant-Waugh et al., 2022) was developed to answer exactly that question: it screens for ARFID against DSM-5 criteria and characterises which driver is most prominent for a given person. It is derived from the longer Pica, ARFID, and Rumination Disorder Interview (PARDI; Bryant-Waugh et al., 2019). Although it takes its name from that parent interview, which assesses all three (pica, ARFID, and rumination disorder), the PARDI-AR-Q itself focuses only on ARFID. When the PARDI-AR-Q was developed, few brief self-report measures existed for identifying ARFID (Bryant-Waugh et al., 2022), and relatively few remain validated in adult and community samples.
The PARDI-AR-Q is available in two formats:
The PARDI-AR-Q produces three profile subscales, capturing the three recognised drivers of avoidant and restrictive eating, alongside a separate index of how much the eating difficulty affects daily life:
The PARDI-AR-Q is suited to general mental health practice, paediatric and child health services, eating disorder services, dietetics and nutrition services, autism and ADHD assessment, and primary care settings. Short and low in respondent burden, it is well suited to routine intake: it is typically completed before or at the start of treatment, where it can guide follow-up questions and help determine whether a fuller assessment or referral is warranted. Results can also be shared with the client or family to support feedback and psychoeducation. It is a screening and characterisation tool, not a progress-monitoring measure: its temporal stability has not been established, so changes between administrations should be read descriptively rather than as reliable change.
A distinctive clinical contribution of the PARDI-AR-Q is its subscale structure. Rather than only flagging whether ARFID may be present, it indicates whether a specific subscale predominates, which can directly inform formulation and treatment focus. A predominantly sensory-based presentation points toward graded sensory and food-exposure work; a presentation driven by lack of interest points toward appetite, routine, and nutritional rehabilitation approaches; and a presentation driven by fear of aversive consequences points toward anxiety-focused and exposure-based work.
The PARDI-AR-Q comprises 29 respondent-completed items (items 4 to 32 of the original instrument) that fall into three blocks:
Recall periods are not fixed and vary by item, spanning current, past-month, and past-3-month timeframes.
The PARDI-AR-Q does not produce a single total score. Instead it produces a categorical screening result, three subscale scores, a severity-of-impact score, and body mass index with growth percentiles as clinical context. Higher subscale and severity scores indicate greater symptom intensity.
The screening result (positive/negative possible ARFID) is generated by an algorithm applied to the diagnostic items (6 to 21). A positive screen for possible ARFID requires both:
Because several different combinations of items can produce a positive result, two people who both screen positive may have endorsed quite different items.
A positive screen indicates possible ARFID; it is not a diagnosis. The PARDI-AR-Q does not evaluate the exclusion criteria for ARFID, and does not assess whether the eating difficulty is better accounted for by another condition, such as anorexia nervosa or bulimia nervosa, or by a concurrent medical condition or other mental disorder.
The subscale scores and the severity-of-impact score are calculated as the average of their respective items, yielding scores that range from 0 to 6:
The three subscale and severity scores are reported on their original 0 to 6 metric and interpreted directly against the response scale, which runs from the absence of the feature (0) to its most extreme expression (6).
| Score | 0 (min) | 6 (max) |
|---|---|---|
| Sensory-based avoidance | no particular sensitivity | extremely negative effect on eating |
| Lack of interest in eating or food | never | always |
| Concern about aversive consequences | never | always |
| Severity of impact | no difficulty | extreme difficulty |
In the interpretive text section, each of the three subscales is given an explanatory paragraph when its score reaches the midpoint of this scale (3 of 6) or higher; lower-scoring subscales are reported as a number only, and severity of impact is always described.
Differentiating ARFID from paediatric feeding disorder (PFD). Because the parent/carer form extends to children as young as four, interpret elevated scores in young children with particular caution. ARFID and PFD overlap and can co-occur, but PFD encompasses feeding difficulties arising from medical, nutritional, oral-motor/swallowing, feeding-skill, or psychosocial factors, whereas ARFID refers to avoidant or restrictive eating that exceeds what would be expected from these factors or warrants separate clinical attention. The PARDI-AR-Q does not comprehensively assess these underlying contributors or determine whether they better account for the presentation. Elevated scores should therefore be interpreted alongside medical, developmental, oral-motor and swallowing, nutritional, feeding-skill, and psychosocial assessment (Goday et al., 2019).
For respondents aged 14 to 19 on the self-report form, body mass index is reported as a percentile and z-score for age and sex; for those aged 20 and over, body mass index is reported against adult categories. On the informant form the same percentile and z-score reporting applies across its full 4 to 19 age range. Growth percentiles are derived using the United States Centers for Disease Control and Prevention growth reference
(Kuczmarski et al., 2002; Flegal & Cole, 2013). These values provide clinical context regarding weight status and growth patterns that may be relevant to ARFID-related consequences such as significant weight loss, failure to achieve expected weight gain, or faltering growth.
On first administration, the three subscales are displayed together on a horizontal bar graph.
Additionally, severity of impact is shown on its own horizontal bar graph, divided into five labelled descriptor bands based on the response options.
On a repeated administration a line graph with the three subscales is shown.
Alongside a line graph displaying the severity of impact.
The PARDI-AR-Q exists as two questionnaire forms: a self-report form for people aged 14 and over, and an informant (parent or carer) form for ages 4 and over. Both were derived from the original PARDI, a clinician-administered structured interview that assesses pica, ARFID and rumination disorder across the lifespan (Bryant-Waugh et al., 2019). It was created to provide a brief questionnaire focused specifically on ARFID, capturing similar information with less burden than the full interview. Its preliminary validation studied 71 adolescents and adults aged 14 to 40 (42 with ARFID and 29 healthy controls), a subset of whom also completed the PARDI interview (Bryant-Waugh et al., 2022). Unless stated otherwise, the evidence below is for the self-report form in adolescents and adults, which carries almost all of the available psychometric evidence. Evidence for the parent or carer form comes mainly from Ortiz et al. (2025) and is labelled where it applies.
Convergent validity. In the preliminary validation of the self-report form, the subscales correlated as expected with established measures of related constructs (Bryant-Waugh et al., 2022). Sensory-based avoidance correlated strongly with the picky-eating subscale of the Nine-Item ARFID Screen (NIAS; r = .70) and with the Food Neophobia Scale (r = .75); lack of interest in eating or food correlated with the NIAS low-appetite subscale (r = .76); concern about aversive consequences correlated with the NIAS fear subscale (r = .65); and severity of impact correlated with the Clinical Impairment Assessment (r = .75).
Discriminant validity. No PARDI-AR-Q subscales correlated significantly with the global score of the Eating Disorder Examination Questionnaire, indicating that the measure captures avoidant/restrictive eating distinct from the weight and shape concerns central to other eating disorders (Bryant-Waugh et al., 2022).
The PARDI-AR-Q discriminates ARFID from clinically overlapping presentations. In adult inpatients completing the self-report form, concern about aversive consequences was higher in ARFID than in restricting-type anorexia nervosa (Cohen’s d = .92), while severity of impact was higher in anorexia nervosa, and the concern and severity profiles each discriminated the two groups with an area under the curve of .73 (Manwaring et al., 2025).
Screening accuracy. In the preliminary validation of the PARDI-AR-Q, the diagnostic screening algorithm classified 90% of the ARFID group as screening positive and 93% of healthy controls as screening negative (Bryant-Waugh et al., 2022). Positive predictive value for probable ARFID was 55% for adult self-report and, for the parent or carer form, 76% in parents of children aged 7 to 17 (Ortiz et al., 2025). Bryant-Waugh and colleagues (2022) note that formal sensitivity and specificity were not established in that study and recommend that a positive screen be confirmed by clinical interview.
Internal consistency of the three subscales was strong in the ARFID group of the preliminary validation: sensory-based avoidance α = .89, concern about aversive consequences α = .93, and lack of interest in eating or food α = .83 (Bryant-Waugh et al., 2022; N = 42 with ARFID). Subsequent independent samples have generally supported adequate internal consistency, while showing that it varies by sample and subscale. In a large community sample of adults screening positive for probable ARFID on both the NIAS and the PARDI-AR-Q (Qi et al., 2025; N = 3,299), alphas were .87 (concern), .78 (sensory-based avoidance and severity of impact), and .71 (lack of interest in eating or food). In an adult inpatient eating disorder sample (Manwaring et al., 2025; N = 78), subscale alphas ranged from .73 to .85.
The exploratory factor analysis in the preliminary PARDI-AR-Q validation supported a three-factor solution corresponding to sensory-based avoidance, lack of interest in eating or food, and concern about aversive consequences (Bryant-Waugh et al., 2022). A supplementary analysis that added the severity-of-impact items produced a four-factor solution in which severity of impact loaded as a separate factor, consistent with its treatment as a distinct index alongside the three subscales. This three-subscale structure mirrors the phenotypes described for ARFID in DSM-5 and operationalised in the PARDI interview. This structure rests on a single exploratory analysis in a small ARFID sample (N = 42).
Because ARFID can present with significant weight loss or faltering growth, the PARDI-AR-Q records height and weight so that body mass index and growth status can be reported as clinical context, interpreted against an external growth reference rather than the measure’s own data. For respondents aged 14 to 19 on the self-report form and 4 to 19 on the informant version, body mass index is expressed as a percentile and z-score for age and sex using the 2000 United States Centers for Disease Control and Prevention (CDC) growth charts (Kuczmarski et al., 2002) and their published LMS parameters (Flegal & Cole, 2013), with the CDC 2022 extended method used to estimate percentiles above the 95th percentile. For respondents aged 20 and over, body mass index is reported against the standard adult categories. CDC charts were selected because they provide a single continuous reference from childhood to adulthood and are the conventional choice for this age range in Australian practice.
Most widely used eating disorder questionnaires focus on the weight and shape concerns central to anorexia and bulimia and can miss avoidant or restrictive eating that is unrelated to body-image concerns. The PARDI-AR-Q is built specifically for ARFID, where eating is restricted because of sensory sensitivity, low appetite or interest, or fear of an aversive consequence such as choking. In the validation work, the questionnaire’s subscales were unrelated to the global score of a standard eating disorder measure, which is what you would want from a tool designed to capture a different kind of eating difficulty.
The screening result tells you whether ARFID may be present; the three subscales tell you why a person may be restricting, which is what shapes treatment. A presentation driven mainly by sensory sensitivity may be approached differently in treatment from one driven by low appetite or by fear of choking or vomiting. Because the subscales often overlap, the pattern across the three subscales is usually more useful than any single score. If one subscale is clearly most elevated, it points toward the most relevant focus for further assessment and intervention.
The three ARFID subscales are distinct and can occur in any combination, so combining them into a single score would obscure the distinction the measure exists to draw. A person can score high on sensory-based avoidance and low on the other two, and that subscale means something different from a person with the reverse pattern. For that reason the subscales are reported and read separately rather than combined.
A positive screen means that further assessment is likely warranted, not that ARFID is diagnosed. The recommended next step is typically a clinical interview to confirm the diagnosis, including review of the exclusion criteria for ARFID (such as another medical or mental health condition that better accounts for the eating difficulty), which the questionnaire does not assess. The subscale scores and the recorded growth information can help direct that assessment toward the most relevant areas.
Yes. ARFID commonly co-occurs with diagnoses of autism and ADHD, and the PARDI-AR-Q is frequently used in these populations. Sensory-based avoidance in particular is common in autistic people, and lack of interest in eating can accompany ADHD. The measure can help distinguish whether restricted eating reflects an ARFID presentation that warrants its own attention, which is useful when eating difficulties are easy to attribute solely to the co-occurring condition.
Picky or “fussy” eating is common, usually mild, and typically causes no lasting harm. ARFID is set apart by its impact (restricted or avoidant eating that leads to significant weight loss or faltering growth, nutritional deficiency, dependence on supplements or tube feeding, or marked interference with daily and social functioning) and by which driver predominates (sensory sensitivity, low interest or appetite, or fear of an aversive consequence). The PARDI-AR-Q screens for that clinically significant level, which ordinary picky eating would not reach.
Not for measuring change in a statistical sense. The PARDI-AR-Q has no published test-retest reliability and no reliable-change or minimal-important-difference benchmark, so a difference between two administrations cannot be read as reliable change. It can be re-administered to see whether the screen result shifts and to see the subscale scores descriptively over time.
Bryant-Waugh R, Stern CM, Dreier MJ, Micali N, Cooke LJ, Kuhnle MC, Burton Murray H, Wang SB, Breithaupt L, Becker KR, Misra M, Lawson EA, Eddy KT, Thomas JJ. Preliminary validation of the pica, ARFID and rumination disorder interview ARFID questionnaire (PARDI-AR-Q). J Eat Disord. 2022 Nov 22;10(1):179. doi: 10.1186/s40337-022-00706-7. PMID: 36419081; PMCID: PMC9682666.
© Rachel Bryant-Waugh, Kamryn Eddy, Nadia Micali, Lucy Cooke, Jennifer J. Thomas. Licence: CC-BY 4.0.
Featured resources
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). American Psychiatric Publishing. https://doi.org/10.1176/appi.books.9780890425596
American Psychiatric Association. (2022). Diagnostic and statistical manual of mental disorders (5th ed., text rev.). https://doi.org/10.1176/appi.books.9780890425787
Bryant-Waugh, R., Stern, C. M., Dreier, M. J., Micali, N., Cooke, L. J., Kuhnle, M. C., Burton Murray, H., Wang, S. B., Breithaupt, L., Becker, K. R., Misra, M., Lawson, E. A., Eddy, K. T., & Thomas, J. J. (2022). Preliminary validation of the Pica, ARFID and Rumination Disorder Interview ARFID Questionnaire (PARDI-AR-Q). Journal of Eating Disorders, 10(1), 179. https://doi.org/10.1186/s40337-022-00706-7
Bryant-Waugh, R., Micali, N., Cooke, L., Lawson, E. A., Eddy, K. T., & Thomas, J. J. (2019). Development of the Pica, ARFID, and Rumination Disorder Interview, a multi-informant, semi-structured interview of feeding disorders across the lifespan: A pilot study for ages 10-22. International Journal of Eating Disorders, 52(4), 378-387. https://doi.org/10.1002/eat.22958
Centers for Disease Control and Prevention. (2022). Extended BMI-for-age growth charts. National Center for Health Statistics. https://www.cdc.gov/growthcharts/extended-bmi-data-files.htm
Flegal, K. M., & Cole, T. J. (2013). Construction of LMS parameters for the Centers for Disease Control and Prevention 2000 growth charts. National Health Statistics Reports, (63).
Goday, P. S., Huh, S. Y., Silverman, A., Lukens, C. T., Dodrill, P., Cohen, S. S., Delaney, A. L., Feuling, M. B., Noel, R. J., Gisel, E., Kenzer, A., Kessler, D. B., Kraus de Camargo, O., Browne, J., & Phalen, J. A. (2019). Pediatric feeding disorder: Consensus definition and conceptual framework. Journal of Pediatric Gastroenterology and Nutrition, 68(1), 124-129. https://doi.org/10.1097/MPG.0000000000002188
Kuczmarski, R. J., Ogden, C. L., Guo, S. S., Grummer-Strawn, L. M., Flegal, K. M., Mei, Z., Wei, R., Curtin, L. R., Roche, A. F., & Johnson, C. L. (2002). 2000 CDC growth charts for the United States: Methods and development. Vital and Health Statistics, 11(246).
Manwaring, J. L., Cass, K., Prostko, S., Buros Stein, A., Mehler, P. S., Joiner, T., & Rienecke, R. D. (2025). How do adult inpatients with ARFID or AN-R compare on self-report eating disorder assessments? Eating Disorders. Advance online publication. https://doi.org/10.1080/10640266.2025.2608345
Ortiz, S. N., White, J. P., MacDermod, C. M., Dinkler, L., Thornton, L. M., Johnson, J., Guintivano, J. D., Baker, J. H., Bulik, C. M., Micali, N., & Pisetsky, E. M. (2025). Validating online parent- and self-report screening methods for avoidant/restrictive food intake disorder. International Journal of Eating Disorders, 58(5), 878-889. https://doi.org/10.1002/eat.24376
Qi, B., Kalantzis, M. A., Thornton, L. M., White, J. P., MacDermod, C. M., Ortiz, S. N., Pisetsky, E. M., Dinkler, L., Guintivano, J. D., Johnson, J. S., Micali, N., & Bulik, C. M. (2025). Suicidal ideation and avoidant/restrictive food intake disorder: Findings from the ARFID-GEN study. Psychiatry Research, 348, 116471. https://doi.org/10.1016/j.psychres.2025.116471
