cs.CLFeb 17, 2026

Towards Expectation Detection in Language: A Case Study on Treatment Expectations in Reddit

Authors: Aswathy Velutharambath, Amelie Wührl

Organizations: University of Stuttgart, Germany · IT University of Copenhagen, Denmark

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Patients' expectations towards their treatment have a substantial effect on the treatments' success. While primarily studied in clinical settings, online patient platforms like medical subreddits may hold complementary insights: treatment expectations that patients feel unnecessary or uncomfortable to share elsewhere. Despite this, no studies examine what type of expectations users discuss online and how they express them. Presumably this is because expectations have not been studied in natural language processing (NLP) before. Therefore, we introduce the task of Expectation Detection, arguing that expectations are relevant for many applications, including opinion mining and product design. Subsequently, we present a case study for the medical domain, where expectations are particularly crucial to extract. We contribute RedHOTExpect, a corpus of Reddit posts (4.5K posts) to study expectations in this context. We use a large language model (LLM) to silver-label the data and validate its quality manually (label accuracy ~78%). Based on this, we analyze which linguistic patterns characterize expectations and explore what patients expect and why. We find that optimism and proactive framing are more pronounced in posts about physical or treatment-related illnesses compared to mental-health contexts, and that in our dataset, patients mostly discuss benefits rather than negative outcomes. The RedHOTExpect corpus can be obtained from https://www.ims.uni-stuttgart.de/data/RedHOTExpect

Figures & tables

idTexts expressing expectation
1All this talk about the new bus lanes. At least I’ll have a new excuse for being late to work from now on.
2I’m pretty optimistic that the new exam format might finally test how well we think, not how long we can sit still
3This phone should last at least three years
4How is this supposed to have an effect with all the weed I’ve been smoking?
5It’s a stupid reason, but everyone says how the meds can make you feel bloated.
6I bet no one tested this for darker skin tones. Makes me scared to even fill the prescription
Table 1: Example texts expressing expectation across general and medical domain.
Figure 2: Automated pipeline for creating annotated RedHOT-Expectation data.
Class / SubsetCountMeanStdMedianMax
No expectation7,364136.6124.71021,771
Expectation (all)4,433182.9170.81382,944
has TEO2,529212.6198.71582,944
no TEO1,904143.6113.01131,042
Table 2: Token length statistics by class and within the subset of expectation posts, distinguishing those with and without extracted treatment–expectation–outcome (TEO) triplets.
DimensionFeatures
Uncertaintytentat, certitude, cogproc, discrep, insight
Time focusfocuspast, focuspresent, focusfuture
EmotionAffect, tone_pos, tone_neg, emo_pos, emo_neg, emo_anx, emo_anger, emo_sad
Social ref.self_reference, ppron, i, we, you, Social, family, friend
MotivationDrives, achieve, power, affiliation, need, want, lack, fulfill, reward, risk
Healthhealth, illness, wellness, mental, substances
Table 3: LIWC feature groups used in the analysis.
Figure 3: Mean feature difference in Expectation vs. Non-Expectations posts, with non-significant differences in grey.
DimensionΔ\Delta (Exp–Non)Cliff’s δ\deltaFDR- pp
Motivation+0.10840.1780< ⁣10−41<\!10^{-41}
Time focus+0.18980.0912< ⁣10−11<\!10^{-11}
Style+0.18000.0867< ⁣10−9<\!10^{-9}
Social refs.-0.0985-0.03890.0045
Health-0.04650.02730.0482
Emotion-0.01380.02290.0895
Table 4: Group-level LIWC differences (Mann–Whitney with Benjamini–Hochberg FDR). Δ\Delta is mean(Exp) −- mean(Non); positive values indicate higher means in expectation posts.
Figure 4: Mean differences by condition types.
Figure 5: Confusion matrix representing alignment of expectation and outcomes types.
Figure 6: Distribution of expectation basis across medical conditions grouped by condition types.
More in this topic

Clinical NLP

Clinical natural language processing (NLP) develops methods to extract, interpret, and generate information from healthcare text and conversations.

See all 142 papers

Explore similar work

CardsList

  1. LLM-Augmented Therapy Normalization and Aspect-Based Sentiment Analysis for Treatment-Resistant Depression on Reddit

    Mar 12, 2026Yuxin Zhu, Sahithi Lakamana, Masoud Rouhizadeh +3Mental HealthSentiment Analysis

    Department of Biomedical Informatics, Emory University, Atlanta, GA · Department of Pharmaceutical Outcomes and Policy, University of Florida, Gainesville, FL

  2. Predicting Mental Health Help-Seeking Behaviour on Reddit: A Survival Analysis of Linguistic and Behavioural Signals

    Oct 12, 2026Nisar Ahmed, Ammara Kazmi, Irtaza Bashir Raja +1Mental HealthSocial Media Analysis

    Department of Computer and Information Sciences, University of Strathclyde, Glasgow, UK · Department of Computer Science, Sir Syed CASE Institute of Technology, Islamabad, Pakistan · Department of Engineering Management, NUST, Rawalpindi, Pakistan

  3. RedditPersona: A Modular Framework for Community-Conditioned LLM Adaptation from Reddit

    Jun 4, 2026Amirhossein Ghaffari, Ali Goodarzi, Huong Nguyen +3LLM EvaluationLLM Fine-Tuning

    Future Computing Group University of Oulu Oulu, Finland · Centre for Applied Computing University of Oulu Oulu, Finland