{"id":"a936002b-6fdc-4761-ada2-37dc747e25a3","arxiv_id":"2501.13308","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"An interview study identifies trigger properties, compulsion patterns, and coping-strategy conflicts that should shape future just-in-time OCD self-management technologies.","lead":"The authors interviewed 10 people with OCD and 7 OCD therapists to learn how real-world triggers and compulsions work and what support is missing. They turn these experiences into design suggestions for smartphone and wearable tools that could offer help at the moment a symptom starts.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central claim that trigger properties 'predominantly' shape distress is not securely grounded: the interview protocol explicitly prompted for fine-grained trigger characteristics, and the claim is inferred from 10 self-selected participants without systematic comparison of trigger types vs.","rationale":"The paper is a careful qualitative need-finding study, and I read it in good faith: it does not claim intervention effectiveness, it is explicit about its sample limitations in Section 6, and it provides a transparent reflexive account in Section 3.4. The reader's CONDITIONAL verdict is appropriate, and my concern does not move the verdict. However, I think the most load-bearing assumption is not only the external transferability to other populations but also the internal grounding of the paper's headline contribution. The abstract asserts that trigger properties 'predominantly' affect user experience, and Section 4.1.2 frames this as a shift from trigger types to trigger properties. This claim is used to motivate the fine-grained symptom formulation and property-based tracking in Section 5.1.2. If that finding were an artifact of the interview protocol's explicit prompting for fine-grained characteristics, the empirical foundation for the design opportunity would be substantially weakened, even though the broader need-finding themes might survive. The concrete test of comparing spontaneous versus prompted mentions is a feasible, low-cost way to assess whether the claim emerged inductively or was elicited. This is a partial agreement with the reader: the reader emphasized external validity (sample generalizability), while my concern is internal validity (prompting and overstatement). Both are real, but my check targets the specific mechanism by which the 'properties over objects' conclusion could be overstated. The paper deserves acceptance with revisions, not rejection: the authors can either provide evidence from the existing transcripts that the property theme was spontaneously raised, or moderate the claim and reframe the design implication as provisional. Neither change is fatal, which is why the reader's CONDITIONAL verdict remains the correct one.","tokens_in":28684,"tokens_out":4362,"duration_ms":44024,"concrete_test":"Re-code the 10 OCD participant interviews using the cleaned transcripts described in Section 3.3, splitting each trigger description into (a) spontaneous unprompted mentions of property-based intensity variation and (b) mentions occurring only after the interviewer's fine-grained follow-up questions described in Section 3.2. Require two independent coders with a pre-registered codebook; then compute, per participant, whether property-based intensity variation was offered before any prompting. If fewer than half of participants spontaneously described property-level determinants, revise the abstract's 'predominantly' claim to 'participants identified' and reframe Section 5.1.2 as a prompted-elicitation insight rather than a demonstrated predominance.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing claim is the headline finding that trigger properties, not trigger types, 'predominantly affect user experience' (Abstract; Section 4.1.2). The design opportunities in Section 5.1.2 (fine-grained symptom formulation, property-based exposure hierarchies) rest on this claim. The evidence, however, is vulnerable to a prompting artifact. Section 3.2 states that the interview protocol asked participants for 'fine-grained details with a goal of identifying the characteristics and rationales of the triggers.' This targeted elicitation could have generated the property-based descriptions, so their prevalence in the data does not establish that properties predominate over trigger types. Section 4.1.2 itself wavers between 'it was often the specific properties' and 'instead of the object type in itself.' With only 10 self-selected participants (Section 3.1.1), many from contamination/responsibility themes (Table A.1), the strength of the 'predominantly' generalization is not supported by a systematic comparison of type-level vs. property-level contributions to trigger intensity. If the property finding is largely an interviewer-driven construction, the recommended shift from object-level to property-level detection loses its empirical foundation. The paper is otherwise transparent and methodical, and the need-finding contribution would survive with a moderated claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a semi-structured interview study with 10 people with mild-to-moderate OCD and 7 OCD therapists to understand triggers, compulsions, coping strategies, and technology needs, and derives design opportunities for just-in-time interventions. It claims that OCD triggers are highly individualized and that specific trigger properties (e.g., fullness of a trash can, lid presence, color contrast, texture, proximity, familiarity) rather than trigger types predominantly shape distress; that coping strategies can conflict with therapist recommendations (e.g., relaxation techniques that may become a new compulsion); and that existing symptom tracking lacks personalization, privacy protection, and support for mental compulsions. The analysis uses a hybrid inductive/deductive thematic approach, with over 700 codes and 20 themes, and the paper includes detailed participant tables, quotes, a reflexivity statement, and an explicit limitation section.","tokens_in":28874,"tokens_out":5783,"duration_ms":55271,"significance":"If the central claims are appropriately moderated, this is a useful need-finding contribution for just-in-time OCD intervention design. The study's strengths include a transparently reported recruitment and analysis procedure, the inclusion of both people with OCD and OCD professionals, explicit attention to researcher positionality, and an honest limitations section. The paper ships no equations or fitted parameters; the design opportunities are presented as extrapolations for future implementation, which is appropriate for a qualitative HCI need-finding study. The most distinctive potential contributions are the shift from object-level to property-level trigger characterization, the identification of mental compulsions as a distinct detection target, and the documented tension between client coping strategies and therapist recommendations, particularly around relaxation techniques.","major_comments":[{"comment":"The headline finding that trigger properties, rather than trigger types, \"predominantly\" affect user experience is not securely grounded because the interview protocol explicitly prompted for fine-grained trigger characteristics. Section 3.2 states that \"for triggers, we specifically asked for fine-grained details with a goal of identifying the characteristics and rationales of the triggers.\" The prevalence of property-based descriptions in the data may therefore reflect the elicitation procedure rather than the relative importance of properties over object types in participants' experience. The Abstract's \"predominantly\" and Section 4.1.2's \"Trigger Intensity: Properties over Objects\" heading state the claim more strongly than the evidence supports; the section text itself wavers to \"it was often the specific properties of the object.\" This issue is load-bearing because the design opportunities in Section 5.1.2 (fine-grained symptom formulation and property-based exposure hierarchies) derive directly from this finding. Please moderate the claim to something like \"participants frequently described specific properties as important to trigger intensity,\" add an explicit discussion of how the targeted prompting may have shaped responses, and, where possible, report whether property descriptions also arose in unprompted parts of the interviews.","section":"Section 3.2, Section 4.1.2, Abstract"},{"comment":"The generalizability of the main findings is constrained by the sample: 10 self-selected, English-speaking, US-based participants with mild-to-moderate OCD, with a concentration of contamination and responsibility themes (six and five participants, respectively, in Table A.1) and only one participant with harm/sexual OCD and two with just-right OCD. While Section 6 appropriately acknowledges the exclusion of severe OCD and less common themes, the paper's broader framing still generalizes from this sample to people with OCD in general. The abstract and Section 4.1.2 should therefore present the property-over-type finding and the derived design opportunities as provisional for the studied sample, not as validated guidelines for the full heterogeneity of OCD. This is not a request for a different study; it is a request to align the strength of the claims with the evidence actually collected.","section":"Section 3.1.1, Table A.1, Section 6"}],"minor_comments":[{"comment":"Please include the full interview protocol and the two codebooks as supplemental material; the analysis procedure is described in detail, but the absence of the actual instruments limits transparency and reproducibility.","section":"Section 3.2, Section 3.3"},{"comment":"The claim of contributing \"the first interview study\" is an overclaim; recommend changing it to \"to our knowledge, the first interview study\" or otherwise softening the absolute phrasing.","section":"Section 1"},{"comment":"There are minor typos: \"axiety disorder\" in Table 2 should be \"anxiety disorder,\" and \"therpy\" in Section 4.4.1 should be \"therapy.\"","section":"Table 2, Section 4.4.1"},{"comment":"The three criteria for distinguishing compulsive behaviors are each attributed to a small subset of therapists (T1; T5-T7; T7); please report the level of consensus more explicitly so readers can gauge how established these criteria are within the professional sample.","section":"Section 4.2.1"},{"comment":"The presentation of trigger properties would benefit from prevalence information; the current text gives illustrative examples from specific participants (P1, P2, P3, P7, P8) without indicating how many participants mentioned each property category, which makes it harder to judge the robustness of each sub-finding.","section":"Section 4.1.2"}],"recommendation":"major_revision","confidential_remarks":"This is a competent qualitative need-finding paper that fits the scope of ASSETS. The main problem is the overstatement of the property-over-type finding, which is plausibly inflated by the protocol's targeted probing for fine-grained trigger characteristics. The core need-finding contribution is valuable and can survive a revision that moderates the headline claim, adds transparency about the prompting effect, and softens the 'first interview study' overclaim. With those changes, I would be supportive of acceptance. I do not see grounds for rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThis is a worthwhile read, with a caveat. The paper is a need-finding interview study (10 people with mild-to-moderate OCD, 7 therapists) that surfaces design opportunities for just-in-time OCD interventions. The two genuinely useful findings are: trigger intensity is often tied to fine-grained properties of the trigger (fullness, contrast, texture, proximity, familiarity) rather than the object type alone, and people's real-world coping strategies—especially relaxation techniques—can conflict with therapist recommendations, since these techniques risk becoming a new compulsion. The five-category compulsion taxonomy tied to triggers is also a useful organizing device for future detection systems.\n\nThe methods are solid for this kind of work. Recruitment criteria, analysis steps, codebook development, and reflexivity are all reported in detail, and the authors explicitly prioritize participant narratives over therapist interpretations. The limitations section is honest: they note the exclusion of severe OCD and rare themes, compensation imbalance, and that this is a need-finding study with unvalidated proposals. Credit where due: this is careful, replicable qualitative research.\n\nThe main soft spot is the headline \"predominantly\" claim. The abstract says trigger properties, not trigger types, \"predominantly affect user experience,\" but the interview protocol specifically prompted participants for fine-grained trigger characteristics. That targeted elicitation makes the prevalence of property-based descriptions less surprising and doesn't establish that properties predominate over types. Section 4.1.2 itself uses softer language (\"it was often...\"), so the abstract is overreaching. Ten self-selected participants, many with contamination/responsibility themes, further limit the generality. This is not a fatal flaw—the property finding is plausible and well-illustrated—but the design opportunities in 5.1.2 lean on it, and the claim needs moderation or a systematic comparison.\n\nOther nits: the \"first interview study\" framing is hard to verify and probably too strong given prior qualitative work on OCD episodes (Knapton 2016 is cited); and the design discussions (AR, LLMs, physiological sensing) are speculative, as the authors admit.\n\nWho is this for? HCI, accessibility, and digital mental health researchers looking for grounded direction on OCD self-management tech. It deserves a serious referee; the empirical core is solid even if the abstract overstates. I'd recommend engaging with the work and asking for a revised abstract and a clearer separation of the prompted finding from an unprompted one.","headline":"A careful, transparent need-finding study on OCD triggers and self-management, with a strong design-oriented contribution that slightly overstates its headline property-vs-type claim.","tokens_in":29424,"tokens_out":2373,"would_cite":true,"duration_ms":32017,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"OCD self-management tech should be built around trigger properties and the risk that coping strategies become new compulsions.","keywords":["OCD","obsessive-compulsive disorder","just-in-time intervention","trigger properties","exposure and response prevention","acceptance and commitment therapy","mental compulsions","self-management technology"],"falsifier":"A pre-registered field study with a larger, more severe, and non-US sample that codes both trigger type and candidate properties (e.g., fullness, color contrast, proximity) and compares their predictive power for individual distress ratings could settle the property-over-type claim: if trigger type alone predicts distress as well as properties do, the framing would be falsified. Likewise, a longitudinal trial of app-delivered relaxation techniques that finds no increase in compulsions would weaken the paper's concern that relaxation can become a new compulsion.","tokens_in":28457,"feed_emoji":"🧠","tokens_out":3943,"duration_ms":37907,"temperature":0.7,"pith_summary":"This paper argues that just-in-time technology for obsessive-compulsive disorder has been built on an incomplete picture of what actually triggers and sustains symptoms. Drawing on interviews with ten people with OCD and seven OCD therapists, it shows that a trigger's type matters less than its properties—such as a trash can's fullness, a stain's color contrast, a knife's proximity, or whether a familiar person is involved. The paper also finds that the coping strategies people with OCD use in daily life can directly conflict with therapist recommendations, notably relaxation techniques that professionals caution can harden into a new compulsion. If this picture is right, future interventions should detect trigger properties rather than generic trigger checklists, and treat mental compulsions as a distinct and critical monitoring target.","feed_headline":"OCD interventions should target trigger properties, not trigger types","feed_subtitle":"Why just-in-time OCD tech must sense properties like fullness and contrast, and watch for coping becoming compulsion.","key_machinery":"The conceptual engine is the distinction between a trigger's type and its properties: the authors argue that properties such as fullness, color contrast, texture, proximity, quantity, pose, and familiarity determine how distressing a trigger is, and that this property-level analysis should drive both detection and intervention design. A second mechanism is the staged view of an OCD episode, in which the authors map coping strategies to different moments—accepting and challenging obsessive thoughts, resisting compulsions through exposure, tapering compulsions, and using distraction or relaxation as a safety net—and then use the observed conflicts between patient practice and therapist guidance to define what feedback a just-in-time system should and should not deliver.","core_discovery":"The central claim is that OCD self-management in the real world is misaligned with current technology because both trigger recognition and intervention feedback are modeled on the wrong units. Based on qualitative analysis of patient and therapist interviews, the authors establish two headline findings: first, trigger intensity is governed by fine-grained properties of triggers—sensory attributes, affordances, spatial attributes, familiarity, and internal state—not by the object category itself; second, the coping strategies people adopt (e.g., breathing, distraction, reassurance logging, exposure practice) sometimes work well but also sometimes contradict professional guidance, and when they do, technology that uncritically reinforces them can make OCD worse. The authors therefore propose design opportunities for personalized, property-aware symptom tracking, just-in-time interventions that support exposure and resistance without becoming new compulsions, and privacy and social features specific to OCD.","pith_inferences":["If the property-over-type finding replicates across larger and more severe samples, it would suggest that clinical exposure hierarchies should be built not only by listing trigger objects but by systematically varying the properties that modulate intensity, and that standard checklist instruments could be complemented by property-level ratings.","The therapist caution about relaxation techniques implies that app-delivered breathing or grounding exercises, popular in general mental health apps, could backfire for a subset of OCD users; a longitudinal experiment could test whether relaxation prompts increase later compulsion urges.","The finding that familiarity can intensify or dampen triggers depending on OCD subtype suggests a testable personalization variable for predictive models: a user-specific familiarity weight for each trigger, which could improve distress prediction.","The proposed wind-down strategy for technology reliance is a testable design principle for any assistive mental-health tool: intervention frequency should taper over time to help users internalize coping skills rather than depend on the device."],"forward_implications":["Just-in-time OCD interventions should sense trigger properties, not just recognize objects, using multimodal data (e.g., object fullness, color contrast, proximity, sound) and should let users specify personalized trigger hierarchies with fine-grained intensity levels.","Intervention feedback should be carefully matched to the stage of an episode and should avoid reinforcing relaxation or distraction as the primary response, since these can become compulsions unless guided with exposure principles.","Mental compulsions (rumination, internal counting, self-reassurance) need to be treated as a distinct detection target because physical compulsion reduction can be compensated by covert mental rituals.","Symptom tracking tools should be personalized, automatic, and privacy-preserving, with obfuscation of sensitive visual and textual data, since fear of data exposure can block adoption.","Technology-supported interventions should include a wind-down feature that gradually reduces reliance on the tool, avoiding the tool itself becoming a safety behavior or an avoidance strategy."],"supporting_citations":[{"why":"Supplies the Y-BOCS symptom categories the paper uses as the clinical baseline and contrasts with its property-focused trigger analysis.","marker":"[41]"},{"why":"Represents prior work that treated OCD triggers as a standardized common checklist, which the paper's individualized-property view extends.","marker":"[72]"},{"why":"Documents the effectiveness of exposure therapy, the core skill participants struggle to apply in daily life.","marker":"[13]"},{"why":"The only existing just-in-time OCD intervention, limited to detecting checking behavior, which this paper generalizes to diverse triggers and compulsions.","marker":"[88]"},{"why":"Defines just-in-time adaptive intervention principles that structure the paper's proposed feedback design.","marker":"[78]"},{"why":"Shows that residual symptoms persist after CBT completion, motivating the need for real-world self-management support.","marker":"[33]"},{"why":"Describes OCD episodes without clearly distinguishing triggers, a gap the paper's fine-grained trigger analysis addresses.","marker":"[59]"}],"fun_headline_variants":["Trigger properties, not types, drive OCD urges","Why OCD tech must sense trigger properties, not categories","When coping becomes compulsion: redesigning OCD tech","Just-in-time OCD aid must target trigger traits, not triggers","OCD self-management fails because tools ignore trigger properties"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The design opportunities assume that the experiences of ten self-selected, English-speaking US participants with mild-to-moderate OCD, plus seven OCD therapists, adequately represent the heterogeneity of OCD triggers, compulsion patterns, and management needs across the broader OCD population.","fun_headline_variants_meta":{"raw":{"variants":["Trigger properties, not types, drive OCD urges","Why OCD tech must sense trigger properties, not categories","When coping becomes compulsion: redesigning OCD tech","Just-in-time OCD aid must target trigger traits, not triggers","OCD self-management fails because tools ignore trigger properties"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000172,"raw_usage":{"total_tokens":1247,"prompt_tokens":890,"completion_tokens":357,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":506,"completion_tokens_details":{"reasoning_tokens":280}},"tokens_in":506,"tokens_out":357,"duration_ms":4381,"temperature":1.0,"reasoning_tokens":280,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T16:15:47.960181+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A pre-registered field study with a larger, more severe, and non-US sample that codes both trigger type and candidate properties (e.g., fullness, color contrast, proximity) and compares their predictive power for individual distress ratings could settle the property-over-type claim: if trigger type alone predicts distress as well as properties do, the framing would be falsified. Likewise, a longitudinal trial of app-delivered relaxation techniques that finds no increase in compulsions would weaken the paper's concern that relaxation can become a new compulsion.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Represents prior work that treated OCD triggers as a standardized common checklist, which the paper's individualized-property view extends."},{"cited_title":"Bi- ological Psychology 127 (2017), 53–63","cited_arxiv_id":null,"evidence_quote":"The only existing just-in-time OCD intervention, limited to detecting checking behavior, which this paper generalizes to diverse triggers and compulsions."}],"review_version":1}