Next Article in Journal
Why Do Consumers Abandon the E-Carts?
Previous Article in Journal
Digital Advertising and Customer Movement Analysis Using BLE Beacon Technology and Smart Shopping Carts in Retail
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
JTAER
Volume 20
Issue 2
10.3390/jtaer20020056
jtaer-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

How to Self-Disclose? The Impact of Patients’ Linguistic Features on Doctors’ Service Quality in Online Health Communities

by
Mengyuan Peng
1,†,
Kaixuan Zhu
1,†,
Yadi Gu
2,†,
Xuejie Yang
3,*,
Kaixiang Su
1,† and
Dongxiao Gu
1,4,*
1
School of Management, Hefei University of Technology, Hefei 230009, China
2
Mental Health Education Center, University of Shanghai for Science and Technology, Shanghai 200093, China
3
School of Humanity and Law, Hefei University of Technology, Hefei 230009, China
4
Laboratory of Data Science and Smart Society Governance of the Ministry of Education, Hefei University of Technology, Hefei 230009, China
*
Authors to whom correspondence should be addressed.
†
These authors contributed equally to this work.
J. Theor. Appl. Electron. Commer. Res. 2025, 20(2), 56; https://doi.org/10.3390/jtaer20020056
Submission received: 25 January 2025 / Revised: 13 March 2025 / Accepted: 21 March 2025 / Published: 25 March 2025
(This article belongs to the Topic Data Science and Intelligent Management)
Browse Figure
Versions Notes

Abstract

:
In online medical consultations, patients convey their medical condition through self-disclosure, and the linguistic features of this disclosure, as signals, may significantly impact doctors’ diagnostic behavior and service quality. Based on signaling theory, this paper collects consultation data from a large online medical platform in China, employs text mining and classification techniques to extract relevant variables, and applies econometric models to empirically examine the effect of patients’ self-disclosure linguistic features on the quality of online medical services. The results indicate that the completeness and readability of patients’ self-disclosure have a significant positive impact on the quality of doctors’ services, while the expertise and positive sentiment of the disclosure have a significant negative effect. From the perspective of signaling theory, this study reveals the mechanism through which patients’ self-disclosure linguistic features influence doctors’ online consultation behavior, providing an important theoretical foundation for promoting online doctor–patient interaction and enhancing patient well-being.

1. Introduction

In recent years, online medical consultation platforms have gradually become a crucial channel for patients to access healthcare services, overcoming the temporal and spatial constraints of traditional offline healthcare and greatly facilitating doctor–patient communication [1]. This not only optimizes the allocation of medical resources but also enhances the accessibility and equity of healthcare services, which is particularly significant for patients in remote areas [2]. In China, an emerging economy characterized by significant urban–rural disparities and cultural diversity, the structural imbalance in healthcare resources combined with the influence of traditional cultural beliefs creates a unique context for the development of online medical consultations [3,4]. Specifically, although tertiary-A hospitals (the highest-level medical institutions) account for only 7.6% of all hospitals nationwide, they handle nearly 47% of outpatient visits. The absence of an effective hierarchical diagnosis and treatment system has resulted in the concentration of high-quality healthcare services in major cities, where these resources are persistently overburdened [5,6]. Given these challenges, online medical consultations have emerged as a vital means of mitigating the structural imbalance in healthcare resources. However, unlike offline consultations, online consultations lack direct, in-person interaction, where patients and doctors can convey information through non-verbal signals such as eye contact, facial expressions, and body language [7]. Furthermore, traditional diagnostic methods like palpation and auscultation cannot be performed in the online environment. In this context, language itself becomes the primary medium of communication between doctors and patients. Therefore, the linguistic expression of patients plays a pivotal role in facilitating doctor–patient communication and enhancing doctors’ engagement in online consultations.
The primary goal of patients in seeking medical consultations is to obtain accurate diagnoses and effective treatment recommendations, as well as a satisfactory healthcare experience. Doctors’ online engagement and service quality are critical variables that determine patients’ trust and satisfaction, playing a significant role in facilitating doctor–patient communication and enhancing patient well-being [8,9]. Existing studies typically assess the quality of offline healthcare services across three dimensions: environmental quality, outcome quality, and interaction quality [10]. However, the computer-mediated communication (CMC) nature of online healthcare introduces two structural differences. First, the absence of a physical environment renders traditional assessments of environmental quality inapplicable [11]. Second, online consultations primarily focus on information exchange and lack physical treatment interventions, making outcome quality more difficult to measure [12]. Given these constraints, interaction quality has become the core dimension for evaluating service quality in online healthcare. Current research generally uses two types of observable indicators to measure interaction quality in online consultations. The first type involves behavioral metrics such as doctors’ response speed, and frequency of interaction, which are used to quantify doctors’ service effort [13,14,15]. The second focuses on content analysis, assessing the informational support and emotional support provided by doctors based on the textual content of their communication [16]. In addition, subjective evaluations from patients, such as doctor ratings and online reviews, are commonly used to measure perceived service quality, although these assessments typically lag behind objective behavioral and content-based indicators [17,18].
How can doctors be encouraged to provide higher-quality services in online consultations? Given the importance of doctors’ engagement, existing research has explored the motivation and influencing factors behind doctors’ contributions in online consultations, from perspectives such as platform incentives, doctors’ traits, and reputation [19,20,21]. However, even for patients with similar conditions, the quality of diagnosis and service provided by the same doctor often varies. Therefore, exploring the influence of patient-related factors on doctors’ consultation behavior from the patient’s perspective holds significant theoretical and practical implications. Existing research has considered various factors from the patient’s perspective that influence doctors’ behavior, such as monetary and non-monetary incentives. However, the linguistic expression of patients during consultations, particularly the manner and characteristics of their self-disclosure, has not received sufficient attention [14,22]. While a few studies have examined the impact of patients’ self-disclosure on doctors’ consultation behavior, most have focused on the depth or content of the disclosure, lacking a detailed and in-depth analysis from a linguistic perspective [23].
Language shapes not only the content of communication (what is said) but also its form and style (how it is said) [24]. From the perspective of signaling theory, patients can adjust their linguistic signals to influence doctors’ responses and obtain higher-quality services and medical advice. Although individuals are often unaware of consciously controlling their language style, these linguistic features can significantly affect how information recipients perceive and respond to the message [25]. Therefore, based on signaling theory, this study constructs an empirical model using real consultation data from the platform, Haodf.com (accessed on 15 October 2024), to examine the impact of the linguistic characteristics of patients’ self-disclosure on the quality of doctors’ services in online consultations. Specifically, the study focuses on four core linguistic features of patient self-disclosure: (1) completeness, (2) readability, (3) expertise, and (4) sentiment valence. These features were selected for two main reasons. First, they collectively involve the four key dimensions of textual linguistic analysis: lexical, syntactic, semantic, and pragmatic [26]. The use of medical terminology reflects the patient’s medical expertise and pertains to the lexical dimension [27]; readability primarily involves syntactic structure [28]; sentiment valence belongs to the semantic dimension [29]; and completeness reflects the extent and adequacy of the information disclosed by patients, involving pragmatic as well as other linguistic dimensions [30]. Second, these linguistic features are closely related to the context of doctor–patient communication. From the perspective of signaling theory, these features enhance the observability of patients’ signals, which can influence the perceived psychological distance between doctors and patients. Specifically, completeness enables doctors to fully understand the patient’s condition; high readability facilitates quick and accurate comprehension of patient concerns; expertise reflects patients’ health literacy and may reduce communication barriers; and emotional expression directly affects emotional resonance and the quality of interaction between doctors and patients [31,32].
In summary, this study takes a patient-centered perspective to further advance research in the field of online doctor–patient interactions. By examining how linguistic features of patient disclosure influence doctors’ quality, the study offers theoretical contributions and practical implications for improving online healthcare services.

2. Theoretical Foundation and Literature Review

2.1. Signaling Theory

Signaling theory was first proposed by Spence in 1973, primarily to explain the communication and transmission of information between information seekers and providers in situations of information asymmetry [33]. Initially, the theory was used to explain how sellers send signals to buyers to promote consumer purchasing behavior in the context of information asymmetry [34,35]. Nowadays, it has been widely applied in studies on online social platforms. For example, Choi et al. employed signaling theory as a theoretical model to examine the impact of product quality cues on the sales of digital video games [36]. Benlian et al. found that information technology features serve as important signals affecting user trust and engagement in online communities [37]. In the context of online consultation platforms, signaling theory provides a framework to understand how patients use external cues (signals) to convey their health conditions and other relevant information to doctors, thereby facilitating doctor–patient interaction and communication [4]. Patients and doctors are two parties with a high degree of information asymmetry: patients possess private information about their health condition but lack medical expertise, while doctors have medical knowledge and skills but are unaware of the specific details of the patient’s condition [38]. In such an information asymmetry context, patients may not only provide specific details of their health condition to doctors but may also adjust their method of self-disclosure, that is, linguistic signals, to receive more attention and service from the doctor [39]. In patient-to-patient communities, signaling theory has been applied to examine how the linguistic characteristics of posts influence peer attention and support [30,40]. Similarly, in online doctor–patient interactions, language serves not only as a medium for conveying medical information but also as a critical signal shaping the doctors’ responses. Khurana et al. identified doctors’ online responses as signals that significantly impact patient satisfaction [41]. Moreover, Zhang et al. demonstrated that patient-generated information serves as an essential linguistic signal, influencing subsequent patients’ information-seeking behavior and healthcare choices [42]. Thus, this study applies signaling theory to investigate how patients’ linguistic features serve as signals that shape doctors’ service quality and doctor–patient interactions.

2.2. Linguistic Characteristics of Self-Disclosure

Self-disclosure refers to the intentional sharing of personal information with others and is an important factor in the development of interpersonal relationships [43]. On social media platforms, individuals enhance familiarity and a sense of connection through self-disclosure, thereby gaining more attention and interaction [44]. Furthermore, self-disclosure helps to build trust and facilitates cooperation and the achievement of transactions [45]. Previous studies have explored various aspects of self-disclosure in social media. For example, Quirdi investigated the predictors of inappropriate and career-oriented self-disclosure [46]. Bae et al. examined the relationship between self-disclosure of different types of background information and social support [47]. However, most existing studies have focused primarily on the level and content of self-disclosure, with limited attention paid to the systematic analysis of its linguistic features [43,48].
Linguistic features serve not only as a means of conveying the communicator’s intentions but also have the potential to evoke behavioral changes in the recipient through various performative functions [49,50]. The linguistic characteristics of a text typically cover four dimensions: vocabulary, syntax, semantics, and pragmatics [26]. The vocabulary dimension focuses on the use of words and phrases, such as vocabulary richness and the use of technical terms. In the healthcare domain, the use of health-related terminology reflects the patient’s level of expertise [51,52]. The syntax dimension examines sentence structure complexity and the coherence between sentences, which directly affect the readability of the text [28]. The semantic dimension investigates the meanings conveyed by the language, with sentiment valence being a key semantic feature [29,53]. The pragmatic dimension emphasizes the social functions of language, where the length of the text reflects the level of completeness in the expression [54]. In the field of doctor–patient interaction, existing studies have primarily focused on the linguistic characteristics of doctors, such as emotions, topic diversity, and vocabulary richness, exploring their impact on patient feedback [55,56]. In contrast, research on the linguistic characteristics of patients mainly focuses on patient communities, with relatively few studies on doctor–patient interaction scenarios [30,40]. Moreover, existing research often emphasizes common, single features such as emotions or text length, lacking a systematic and fine-grained analysis specific to the context of online consultations [23,57,58]. Therefore, this study, grounded in the context of online consultations and the four dimensions of linguistic features, explores the completeness, readability, expertise, and sentiment polarity of patients’ self-disclosure, and their potential impact on the quality of doctor services.

2.3. Doctor’s Consultation Behavior and Service Quality

In online consultations, doctors’ behaviors influence the patient’s healthcare experience as well as subsequent treatment decisions [59]. Social support from doctors is a crucial indicator of the quality of doctor–patient consultations, primarily in the forms of informational support and emotional support. Informational support helps patients better understand their condition, reduce cognitive uncertainty, and make more informed treatment choices [60,61], while emotional support alleviates patients’ psychological burden, enhances their confidence in treatment, and thereby facilitates recovery [16,62]. To date, research on doctors’ online consultation behaviors has primarily focused on their impact on patient satisfaction and subsequent decision-making [18,63,64]. For instance, Yang et al. investigated how doctors’ interaction rates and response times in online consultations influence patient satisfaction [13], while Khurana et al. found that doctors’ responses to patients’ inquiries positively affect patients’ recommendations [41]. In contrast, studies exploring the motivations and influencing factors behind doctors’ online behaviors remain limited, with most focusing on doctors’ personal characteristics, status, incentives, and online reviews [4,19,20]. Communication in online consultations primarily occurs through language, which serves as the primary means by which patients convey to doctors. However, existing research has overlooked the impact of patients’ linguistic expressions on doctors’ consultation behaviors, focusing instead on other aspects of communication. Therefore, this study aims to examine doctors’ consultation behaviors and service quality from the perspective of patients’ self-disclosure linguistic features.

3. Research Hypotheses and Model

3.1. Research Hypotheses

This study focuses on the context of online doctor–patient interactions, exploring the impact of patients’ self-disclosure on doctors’ consultation behaviors and service quality based on four linguistic dimensions (vocabulary, syntax, semantics, and pragmatics). Specifically, it examines how the completeness, readability, expertise, and sentiment polarity of patients’ self-disclosure influence doctors’ behaviors and service quality.

3.1.1. Completeness

Completeness refers to the level of detail in patients’ self-disclosure, reflecting the adequacy of the information provided. Generally, longer texts contain more self-disclosure, thus conveying a richer amount of information. Previous studies have shown that individuals who disclose more information are more likely to receive social support [23]. In the context of electronic healthcare, detailed information helps doctors gain a more comprehensive understanding of the patient’s condition, thereby enhancing doctor–patient interaction and response [65]. Specifically, a thorough and complete description of the patient’s health status not only facilitates more accurate diagnoses and treatment decisions but also conveys the patient’s concern about their condition and expectation for assistance, prompting more attention and positive responses from the doctor. Thus, we propose the following hypotheses:
H1a. 
The completeness of patients’ self-disclosure positively influences the informational support provided by doctors in online consultations.
H1b. 
The completeness of patients’ self-disclosure positively influences the emotional support provided by doctors in online consultations.

3.1.2. Readability

Readability refers to the ease with which a text can be understood, also known as comprehensibility [66]. Readability is typically negatively correlated with text perplexity: higher perplexity usually indicates more complex language structures, grammar, and vocabulary choices, resulting in lower readability [40]. In online consultations, the readability of patients’ self-disclosure directly impacts how well doctors can understand and respond to the information. Patients’ communication abilities vary, and those with stronger expression skills tend to describe their health condition and needs more clearly and concisely, enabling doctors to quickly comprehend the situation and respond accurately. In contrast, texts with poor readability increase the cognitive load and comprehension cost for doctors, reducing communication efficiency and limiting their ability to provide assistance. Therefore, we propose the following hypotheses:
H2a. 
The readability of patients’ self-disclosure positively influences the informational support provided by doctors in online consultations.
H2b. 
The readability of patients’ self-disclosure positively influences the emotional support provided by doctors in online consultations.

3.1.3. Expertise

Expertise refers to the knowledge and skills demonstrated by patients in the healthcare field, often closely associated with health literacy. It can be understood as an individual’s ability to acquire, comprehend, and apply health information, which is crucial for making informed health decisions and effectively managing healthcare needs [67]. Expertise is a key factor influencing doctor–patient communication and health outcomes [68]. During consultations, patients who use specialized terms can convey their understanding of medical knowledge and their condition more clearly, reducing communication barriers and misunderstandings. Furthermore, according to communication accommodation theory, using vocabulary and language that matches the communication partner’s style significantly enhances communication effectiveness and acceptance [26]. This professional expression style aligns more closely with the doctor’s communication, encouraging more interaction. Therefore, we propose the following hypotheses:
H3a. 
The expertise of patients’ self-disclosure positively influences the informational support provided by doctors in online consultations.
H3b. 
The expertise of patients’ self-disclosure positively influences the emotional support provided by doctors in online consultations.

3.1.4. Sentiment Valence

Self-disclosure often involves sentiment valence [69]. Previous studies have shown that negative emotions significantly impact communication, with recipients typically responding more strongly to negative information [70,71]. In online social media, negative emotions are often more contagious and attract more attention than positive ones [72]. In online consultations, the sentiment polarity conveyed by patients sends important signals to doctors. Negative emotions tend to draw more attention from doctors, promoting more support and assistance, especially when emotional responses are strong. In such cases, doctors may perceive that the patient needs more care and intervention. Thus, we hypothesize the following:
H4a. 
Positive sentiment valence in patients’ self-disclosure negatively influences the informational support provided by doctors in online consultations.
H4b. 
Positive sentiment valence in patients’ self-disclosure negatively influences the emotional support provided by doctors in online consultations.

3.2. Research Model

In summary, based on signal theory, this study analyzes and tests the impact of the linguistic features of patients’ self-disclosure on doctors’ service quality. The overall research model is shown in Figure 1.

4. Research Methodology

4.1. Research Context and Data Collection

China, the world’s most populous country, faces persistent shortages and structural disparities in healthcare resources [73]. The absence of a strict hierarchical medical system, coupled with weak primary care and low patient trust, has driven excessive reliance on specialist care, even for minor conditions [74]. This overconcentration of resources has intensified healthcare access difficulties, making it a major public health challenge [75]. In this context, the rapid development of internet-based healthcare has created new pathways for restructuring doctor–patient interactions. This study examines how patients’ language expression strategies in the Chinese-language context influence the quality of online medical consultations. The findings not only inform efforts to improve online healthcare but also provide insights for cross-cultural digital health research.
This study employs a Python web scraping tool (Python version 3.7.13) to collect data from Haodf.com, a major online healthcare platform in China, which provides extensive doctor profiles and doctor–patient consultation records, making it an ideal context for this research [76]. We collected all consultation records and relevant doctor profiles for two diseases—gastritis and liver cancer—between June 2014 and September 2023. A total of 67,871 doctor–patient consultation records were collected. After filtering for text and image-based consultation data and performing data cleaning, 52,730 valid records were obtained from 4871 doctors, including 26,654 records related to gastritis and 26,076 records related to liver cancer.

4.2. Variables

4.2.1. Dependent Variables

Information Support: The amount of information provided by the doctor during a round of doctor–patient interaction, measured by the number of content words. Content words refer to words that have concrete meanings, can independently serve as sentence components, and carry lexical and grammatical significance, including nouns, verbs, adjectives, adverbs, numerals, classifiers, pronouns, state words, and differentiating words [28]. In this study, text segmentation and part-of-speech tagging were performed using the Jieba segmentation tool, with analysis based on a medical terminology dictionary. The number of content words was then calculated.
Emotional Support: The total amount of emotional support provided by the doctor during a round of doctor–patient interaction [29]. A BERT classifier was used to train a text classification model to determine whether emotional support was provided in each doctor’s response (labeled as 0 for no emotional support and 1 for emotional support) [77]. The sum of emotional support provided during a round of interaction was then calculated. The BERT classifier achieved a 96% accuracy on the test set. The detailed text classification process is provided in Appendix A.1.

4.2.2. Independent Variables

Completeness: The degree of detail in patients’ self-disclosure during online consultations, measured by the total length of the patient’s text in a round of doctor–patient interaction.
Readability: The ease of understanding the text. Based on existing research, readability is measured by 1 minus the normalized perplexity of the sentence [40]. Perplexity is an indicator used to assess a language model’s ability to predict text, representing the model’s ‘uncertainty’ or ‘complexity’ in relation to the text. The lower the perplexity of the sentence, the simpler the text structure and the more natural the grammar, making it easier to understand. This study used a Chinese pre-trained GPT-2 model to calculate the perplexity of the patient’s text.
Expertise: The average number of medical terms used by the patient during a round of doctor–patient consultation, reflecting the patient’s expertise. This study used the medical terminology dictionary from Tsinghua University Open Chinese Thesaurus and supplemented it with terms related to stomach and liver diseasesbased on the content of the doctor–patient dialogue. The average number of medical terms in the patient’s text was calculated using a Python program [78].
Sentiment Valence: The average emotional tendency expressed by the patient during a round of doctor–patient interaction. A BERT model was used for text classification, categorizing each patient’s consultation text into three sentiment categories: negative (−1), neutral (0), and positive (1) [71]. The average sentiment valence for the patient’s text in a round of interaction was then calculated. A value closer to 1 indicates more positive sentiment, while a value closer to −1 indicates more negative sentiment. The text classification method used is the same as that for emotional support, with a classification accuracy of 90% on the test dataset. The detailed text classification process is also shown in Appendix A.1.

4.2.3. Control Variables

To control for potential confounding factors, we introduced a series of control variables, including patient gender, patient age, doctor gender, doctor’s employment grade, doctor’s recommendation heat, hospital level, and communication time. The descriptions of all variables are shown in Table 1.

4.3. Model Estimation

We employ Ordinary Least Squares (OLS) regression to estimate the data. The regression models are represented as follows:
The effect of linguistic features of patients’ self-disclosure on doctors’ informational support:
D I n f o S i = β 0 + β 1 T h o r o u g h i + β 2 R e a d a b i l i t y i + β 3 P r o f e s s i o n a l i + β 4 P S e n t i m e n t i + β 5 C o n t r o l i + ε i
The effect of linguistic features of patients’ self-disclosure on doctors’ informational support:
D E m o S i = β 0 + β 1 T h o r o u g h i + β 2 R e a d a b i l i t y i + β 3 P r o f e s s i o n a l i + β 4 P S e n t i m e n t i + β 5 C o n t r o l i + ε i

5. Results

5.1. Descriptive Statistics and Correlation Analysis

We used Stata 17 to analyze the data. Table 2 and Table 3 present the descriptive statistics and correlation matrix for the variables, respectively. Given the different scales of the variables, the independent and control variables were standardized, and a linear model was used to check the Variance Inflation Factor (VIF). As shown in Table 2, the VIF values are all less than 2, indicating that there was no significant multicollinearity among these variables.

5.2. Regression Results Analysis

Table 4 presents the ordinary least squares (OLS) regression results with robust standard errors. Model 1 examines the relationship between the linguistic features in patients’ self-disclosure and doctors’ provision of informational support (DInfoS). The results show that completeness (Completeness, β = 11.063, p < 0.001) and readability (Readability, β = 1.544, p < 0.001) have significant positive effects on doctors’ informational support. This suggests that the more comprehensive and readable the patient’s inquiry, the more informational support the doctor provides, supporting H1a and H2a. However, the coefficient of expertise (Expertise, β = −1.055, p < 0.01) is negative and statistically significant, suggesting that patients using more medical terminology and demonstrating higher medical expertise tend to receive less informational support from doctors. As a result, H3a is rejected. Additionally, positive sentiment (PSentiment, β = −1.655, p < 0.001) is negatively correlated with doctors’ informational support, implying that the more positive a patient’s sentiment, the less informational support they receive. Conversely, patients with more negative emotions receive more informational support, supporting H4a.
Model 2 investigates the effects of the linguistic features in patients’ self-disclosure on doctors’ provision of emotional support (DEmoS). The results reveal that completeness (Completeness, β = 0.047, p < 0.001) and readability (Readability, β = 0.022, p < 0.001) are positively associated with emotional support, indicating that patients who provide more comprehensive and easily understandable inquiries receive greater emotional support from doctors. Thus, H1b and H2b are supported. However, the coefficient of expertise (Expertise, β = −0.074, p < 0.001) is negatively significant, suggesting that patients displaying a higher level of medical expertise tend to receive less emotional support from doctors, leading to the rejection of H3b. Similarly, positive sentiment (PSentiment, β = −0.077, p < 0.001) is negatively correlated with emotional support, suggesting that the more positive a patient’s sentiment, the less emotional support they receive, while patients with more negative emotions receive more emotional support, supporting H4b. Table 5 summarizes the hypothesis testing results.

5.3. Robustness Checks

To ensure the robustness of our findings, we employed two different robustness testing methods, as summarized in Table 6.
First, given that both informational support (DInfoS) and emotional support (DEmoS) are count variables, we applied a negative binomial regression model instead of OLS to account for data overdispersion [79,80]. Model 3 presents the negative binomial regression results for informational support, while Model 4 reports the results for emotional support. The estimates remain consistent with those in Table 5, supporting the robustness of our findings.
Second, to verify the consistency of our results under alternative measures of doctors’ service quality, we used text length per interaction and total number of responses as substitute indicators [14,81]. Model 5 reports the OLS regression results using text length as the dependent variable, which aligns with the main conclusions, reinforcing the robustness of our model. Model 6, which examines the effect of linguistic features on the number of responses, reveals a directional change in the coefficients for readability and sentiment polarity. Specifically, higher readability of patient text is associated with fewer doctor responses, but this reduction in response frequency does not indicate a decrease in information provision. Instead, the length of doctors’ replies increases, suggesting that when patient texts are more readable, doctors may respond less frequently but provide more detailed replies. On the other hand, more positive patient sentiment correlates with an increase in the number of doctor responses. One possible explanation is that doctors tend to be more cautious when addressing negative emotions, resulting in fewer but more extensive responses.

6. Discussion, Implications and Limitations

6.1. Key Findings

Drawing on signaling theory, this study examines the mechanisms through which linguistic features of patient self-disclosure influence doctors’ service quality in an e-health context. The results reveal that completeness in patient self-disclosure has a significant positive effect on both informational and emotional support provided by doctors, aligning with previous research [23]. Additionally, readability in patient self-disclosure also has a positive effect on doctors’ service quality, possibly because more detailed and clearly expressed information enhances doctors’ understanding of patients’ conditions and needs, enabling them to provide more precise support. Existing research has found that on short-term rental platforms, the readability of hosts’ self-descriptions is positively associated with guests’ perceived trust, thereby promoting booking behavior [28]. Building upon this, our study further explores the relationship between text readability and service quality in the context of doctor–patient interactions. In contrast, expertise in patient self-disclosure negatively impacts doctors’ informational and emotional support, which contradicts our initial hypotheses. A possible explanation is that the excessive use of medical terminology reduces text readability, increasing the cognitive burden on doctors and ultimately affecting their service quality [82]. Moreover, patients’ positive sentiment also negatively affects doctors’ service quality. According to the negativity bias theory, doctors tend to allocate more attention and assistance to patients who exhibit negative emotions, whereas those displaying positive emotions may be perceived as requiring less intervention, leading to a reduction in doctors’ service engagement. A prior study on offline clinical settings found that when patients express negative emotions or concerns, doctors tend to avoid engaging in emotional discussions, especially when patients explicitly express their concerns [32]. However, given the substantial differences between face-to-face and online consultations, doctors’ responses to patient emotions may vary in digital settings. Effective face-to-face doctor–patient communication is shown to not only improve medical decision-making quality but also enhance patients’ emotional well-being and other health-related outcomes [83]. Building on this foundation, this study extends the exploration of doctor–patient communication to the online consultation setting, examining how patients can adjust their communication strategies to enhance doctors’ service quality and improve their overall consultation experience.

6.2. Implications

6.2.1. Theoretical Implications

This study offers several theoretical contributions.
First, by introducing signaling theory, this study explores how patients’ linguistic signals influence doctors’ online engagement in teleconsultations. The findings reveal that text completeness and readability facilitate higher-quality medical services, while expertise and positive sentiment negatively affect service quality. These results extend the application of signaling theory in doctor–patient interactions, demonstrating that patients’ linguistic expressions serve not only as a means of information transmission but also as crucial signals influencing doctors’ responses. Furthermore, this study provides a new theoretical perspective on how optimizing patients’ language expression can enhance healthcare service quality.
Second, existing research on online consultations has primarily focused on patient satisfaction and subsequent decision-making behaviors, with limited attention to patients’ self-disclosure and linguistic features [16,64,84]. This study addresses this gap by examining completeness, readability, expertise, and sentiment polarity in patient self-disclosure from lexical, syntactic, semantic, and pragmatic dimensions, enriching the literature on patients’ disclosure behavior and linguistic features.
Third, previous studies on doctor–patient interactions have mainly focused on how doctors’ service quality influences patients, with limited exploration of the factors affecting doctors’ service quality [13,85]. By focusing on patients’ agency and proactivity, this study delves into doctor–patient dialogues to investigate how patients’ linguistic expressions impact doctors’ service quality, offering a novel perspective on online doctor–patient interactions.
Finally, employing real online consultation data, this study applies text mining and text classification techniques to accurately extract and quantify patients’ linguistic features as well as doctors’ informational and emotional support. This approach expands the application of natural language processing (NLP) in online healthcare research and provides a methodological reference for future studies on doctor–patient interactions.

6.2.2. Practical Implications

From a practical perspective, the findings of this study provide valuable insights for patients, doctors, and online consultation platforms in optimizing medical interactions.
First, this study reveals how linguistic features of patient self-disclosure influence doctors’ service quality, offering insights into how patients can optimize their text expression in online consultations. The results indicate that a more comprehensive and readable description leads to greater social support from doctors. Therefore, patients should provide complete, well-structured descriptions of their medical conditions, avoiding missing or unclear information that may lead to misinterpretation or inefficient responses from doctors. Moreover, the excessive use of medical terms negatively impacts social support from doctors, suggesting that patients should use clear and accessible language to describe their symptoms. Additionally, emotional expression also influences doctors’ responses—patients displaying highly positive emotions may receive less emotional support. Thus, patients should balance their emotional expression, appropriately conveying concerns and needs to ensure doctors fully understand their situation and provide adequate responses.
Second, the findings provide new insights for doctors, emphasizing the importance of recognizing variations in patients’ linguistic expressions when delivering online medical services. Doctors should be more sensitive to patients’ linguistic features, as differences in communication styles may unintentionally influence their judgment and responses. More importantly, they should prioritize patients’ actual needs rather than relying solely on linguistic cues when determining the level of informational and emotional support required, ensuring medical fairness in online consultations.
Finally, the findings also provide valuable guidance for the optimization of online consultation platforms through technological and interface enhancements to facilitate more effective communication [86]. One possible improvement is the implementation of structured text input tools, which can help guide patients through step-by-step descriptions of their medical history, symptoms, onset time, and severity, allowing doctors to quickly grasp key information and improve consultation efficiency. Additionally, AI-assisted language optimization tools can be integrated to automatically refine patient input, such as simplifying complex sentences, replacing obscure terms, and suggesting missing but essential information, thereby enhancing readability and completeness [87]. Another potential advancement involves emotion recognition technology, which could be used to analyze patients’ emotional states and provide real-time indicators on the doctor’s interface, encouraging them to offer more targeted emotional support when necessary. By incorporating these optimizations, online consultation platforms can enhance communication efficiency, improve user experience, and ultimately advance the quality of telemedicine services.

6.3. Limitations and Future Research Directions

This study has several limitations.
First, the analysis is based on data from a single online consultation platform in China and examines the impact of patients’ self-disclosure on doctors’ service quality in a Chinese-language context. However, linguistic and cultural differences may significantly influence self-disclosure and its effects on physician behavior. For example, Chinese and English differ significantly in language structure, information organization, and emotional expression, which could affect the patterns of self-disclosure and their impact on medical interactions. Consequently, our findings may primarily apply to Chinese-language consultations, and their generalizability to other linguistic contexts remains uncertain. Future research could replicate this study using data from multiple countries and platforms to enhance the robustness of the conclusions. Additionally, cross-cultural and cross-linguistic comparisons could provide valuable insights into how language and cultural backgrounds shape patient communication and doctor–patient interactions, offering a broader perspective for global online healthcare research.
Second, this study focuses exclusively on text-based online consultations, without considering telephone or other telemedicine modalities. In settings involving voice communication, paralinguistic cues such as tone and speech rate may impact doctors’ responses and engagement. Future studies could expand to audio and video-based consultations to investigate the influence of vocal and visual signals in doctor–patient interactions, contributing to a more comprehensive understanding of communication mechanisms in online healthcare settings.
Third, this study has methodological limitations related to variable measurement. Doctors’ service quality is evaluated based on informational and emotional support. However, due to dataset limitations, we can not incorporate indicators of how well patient needs were met, such as patient ratings and satisfaction scores. Additionally, potential biases in text data analysis should be acknowledged, as variable extraction may be influenced by interpretation bias or algorithmic constraints. Future research could employ more refined measurement approaches to improve the accuracy of variable assessment, thereby enhancing the reliability and external validity of the findings.
Finally, although this study focuses on gastritis and liver cancer based on disease severity, different diseases may involve varying patient information needs and communication styles. Future research could incorporate a wider range of diseases to strengthen the generalizability of the findings and further explore how patient linguistic features influence doctors’ service quality across different medical conditions.

Author Contributions

Conceptualization, M.P. and X.Y.; methodology, M.P. and K.Z.; formal analysis, K.Z., Y.G. and K.S.; data curation, K.S.; writing—original draft preparation, M.P.; writing—review and editing, D.G. and X.Y.; funding acquisition, D.G. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Natural Science Foundation of China (Grant No. 72131006, 72271082, 72071063) and the Natural Science Foundation of Anhui Province (Grant No. 2408085J041).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable. We would like to clarify that our study is based on publicly available online consultation data obtained from a widely used Chinese online healthcare platform, with all data processing conducted in compliance with the platform’s terms of service and relevant ethical guidelines. Therefore, informed consent was not required.

Data Availability Statement

The data presented in this study are available upon request from the corresponding author.

Conflicts of Interest

The authors declare no conflicts of interest.

Appendix A

Appendix A.1. Text Classification Process for Sentiment Polarity and Emotional Support

The text classification process consists of four main steps: defining annotation rules and sample labeling, text preprocessing, feature extraction and vectorization, and classifier construction and evaluation. The specific workflow is as follows:
  • Defining Annotation Rules and Sample Labeling
We first established annotation rules for identifying patients’ sentiment polarity and doctors’ emotional support based on unstructured text data from doctor–patient interactions. Sentimet polarity was categorized into negative, neutral, and positive, while doctors’ emotional support was characterized by expressions of reassurance, encouragement, or empathy, such as “Don’t worry”, “I understand you”, and “Stay strong”. Specific annotation rules are shown in Table A1.
We recruited three medical graduate students for the annotation process: First, the annotators compiled an initial lexicon based on the specified annotation rules and the context of the consultations. Then, each student randomly selected 1000 patient texts and 1000 doctor texts for annotation. Cross-validation was conducted to assess annotation consistency, and any disagreements were discussed to refine the rules. Finally, after finalizing the annotation rules, we randomly selected 3000 doctor texts and 3000 patient texts for formal annotation. Each annotator labeled 2000 texts, and to ensure reliability, all annotations were cross-checked so that each text was labeled by two annotators.
2.
Text Preprocessing
We used the jieba package in Python to segment the text data and removed stop words based on a stop word dictionary, resulting in segmented text.
3.
Feature Extraction and Vectorization
To represent textual features effectively, different vectorization methods were applied based on the classification model. (a). Support Vector Machine (SVM), used Term Frequency-Inverse Document Frequency (TF-IDF) to compute word weights for text representation. (b). Text Convolutional Neural Network (TextCNN), trained a SkipGram word vector model and utilized the trained Word2Vec model for word embedding, converting natural language into word vectors. (c). Bidirectional Encoder Representations from Transformers (BERT), loaded the pre-trained BERT model, which directly transformed text into word vector through its embedding layer.
4.
Classifier Construction and Evaluation
The labeled dataset was split into 80% training data and 20% testing data. We implemented three classification models. We used three models for text classification: SVM, TextCNN, and BERT. The models were trained using the training set and evaluated on the test set. As shown in Table A2, BERT outperformed SVM and TextCNN across all key performance metrics, including accuracy, precision, recall, and F1-score. Therefore, BERT was selected as the final classifier for automatically labeling the unlabeled dataset.
Table A1. The labeling rules of sentiment polarity and emotional support.
Table A1. The labeling rules of sentiment polarity and emotional support.
VariablesClassificationKeywordsExampleTags
Patient’s Sentiment PolarityPositivePositive words like happy, peaceful, grateful, love, etc.
Negative words like worry, sad, depressed, etc.		“I have great faith in the future.”1
Neutral“Which kinds of medicine to take?”0
Negative“I’m afraid, please help me.”−1
Doctor’s Emotional Support ProvisionYesdon’t give up, understand, sympathize, come on, be optimistic, etc.“Let’s face it together”1
No“Well, keep the fluids flowing and give your intestines a rest”0
Table A2. Classifier performance evaluation results.
Table A2. Classifier performance evaluation results.
ClassifiersSentiment PolarityEmotional Support
AccuracyPrecisionRecallF1-ScoreAccuracyPrecisionRecallF1-Score
SVM0.810.780.380.390.870.890.840.86
TextCNN0.820.610.430.460.780.360.460.40
BERT0.900.820.740.770.960.950.950.94

References

  1. Yang, X.; Xi, N.; Gu, D.; Liang, C.; Liu, H.; Tang, H.; Hamari, J. Medical Practice in Gamified Online Communities: Longitudinal Effects of Gamification on Doctor Engagement. Inf. Manag. 2023, 61, 103906. [Google Scholar] [CrossRef]
  2. Fan, W.; Jiang, Y.; Pei, J.; Yan, P.; Qiu, L. The Impact of Medical Insurance Payment Systems on Patient Choice, Provider Behavior, and Out-of-pocket Rate: Fee-for-service versus Diagnosis-related Groups. Decis. Sci. 2024, 55, 245–261. [Google Scholar] [CrossRef]
  3. Liu, Y.; Kong, Q.; Wang, S.; Zhong, L.; van de Klundert, J. The Impact of Hospital Attributes on Patient Choice for First Visit: Evidence from a Discrete Choice Experiment in Shanghai, China. Health Policy Plan. 2020, 35, 267–278. [Google Scholar]
  4. Li, J.; Tang, J.; Jiang, L.; Yen, D.C.; Liu, X. Economic Success of Physicians in the Online Consultation Market: A Signaling Theory Perspective. Int. J. Electron. Commer. 2019, 23, 244–271. [Google Scholar] [CrossRef]
  5. Wu, Q.; Xie, X.; Liu, W.; Wu, Y. Implementation Efficiency of the Hierarchical Diagnosis and Treatment System in China: A Case Study of Primary Medical and Health Institutions in Fujian Province. Int. J. Health Plann. Manag. 2022, 37, 214–227. [Google Scholar]
  6. Zhang, W.; Zhou, F.; Fei, Y. Repetitions in Online Doctor-Patient Communication: Frequency, Functions, and Reasons. Patient Educ. Couns. 2023, 107, 107565. [Google Scholar]
  7. Shreve, E.G.; Harrigan, J.A.; Kues, J.R.; Kagas, D.K. Nonverbal Expressions of Anxiety in Physician-Patient Interactions. Psychiatry 1988, 51, 378–384. [Google Scholar] [CrossRef]
  8. Tan, H.; Yan, M. Physician-User Interaction and Users’ Perceived Service Quality: Evidence from Chinese Mobile Healthcare Consultation. Inf. Technol. People 2020, 33, 1403–1426. [Google Scholar] [CrossRef]
  9. Gu, D.; Li, T.; Wang, X.; Yang, X.; Yu, Z. Visualizing the Intellectual Structure and Evolution of Electronic Health and Telemedicine Research. Int. J. Med. Inform. 2019, 130, 103947. [Google Scholar] [CrossRef]
  10. Dagger, T.S.; Sweeney, J.C.; Johnson, L.W. A Hierarchical Model of Health Service Quality: Scale Development and Investigation of an Integrated Model. J. Serv. Res. 2007, 10, 123–142. [Google Scholar]
  11. Akter, S.; D’ambra, J.; Ray, P. Development and Validation of an Instrument to Measure User Perceived Service Quality of mHealth. Inf. Manag. 2013, 50, 181–195. [Google Scholar] [CrossRef]
  12. Yang, Y.; Zhang, X.; Lee, P.K. Improving the Effectiveness of Online Healthcare Platforms: An Empirical Study with Multi-Period Patient-Doctor Consultation Data. Int. J. Prod. Econ. 2019, 207, 70–80. [Google Scholar] [CrossRef]
  13. Yang, H.; Guo, X.; Wu, T. Exploring the Influence of the Online Physician Service Delivery Process on Patient Satisfaction. Decis. Support Syst. 2015, 78, 113–121. [Google Scholar] [CrossRef]
  14. Wang, Q.; Qiu, L.; Xu, W. Informal Payments and Doctor Engagement in an Online Health Community: An Empirical Investigation Using Generalized Synthetic Control. Inf. Syst. Res. 2023, 35, 706–726. [Google Scholar] [CrossRef]
  15. Lu, N.; Wu, H. Exploring the Impact of Word-of-Mouth about Physicians’ Service Quality on Patient Choice Based on Online Health Communities. BMC Med. Inform. Decis. Mak. 2016, 16, 151. [Google Scholar] [CrossRef]
  16. Chen, S.; Guo, X.; Wu, T.; Ju, X. Exploring the Online Doctor-Patient Interaction on Patient Satisfaction Based on Text Mining and Empirical Analysis. Inf. Process. Manag. 2020, 57, 102253. [Google Scholar] [CrossRef]
  17. He, Y.; Guo, X.; Wu, T.; Vogel, D. The Effect of Interactive Factors on Online Health Consultation Review Deviation: An Empirical Investigation. Int. J. Med. Inform. 2022, 163, 104781. [Google Scholar] [CrossRef]
  18. Cao, X.; Liu, Y.; Zhu, Z.; Hu, J.; Chen, X. Online Selection of a Physician by Patients: Empirical Study from Elaboration Likelihood Perspective. Comput. Hum. Behav. 2017, 73, 403–412. [Google Scholar] [CrossRef]
  19. Zhang, X.; Guo, F.; Xu, T.; Li, Y. What Motivates Physicians to Share Free Health Information on Online Health Platforms? Inf. Process. Manag. 2020, 57, 102166. [Google Scholar] [CrossRef]
  20. Yang, H.; Du, H.S.; He, W.; Qiao, H. Understanding the Motivators Affecting Doctors’ Contributions in Online Healthcare Communities: Professional Status as a Moderator. Behav. Inf. Technol. 2021, 40, 146–160. [Google Scholar] [CrossRef]
  21. Guo, S.; Guo, X.; Fang, Y.; Vogel, D. How Doctors Gain Social and Economic Returns in Online Health-Care Communities: A Professional Capital Perspective. J. Manag. Inf. Syst. 2017, 34, 487–519. [Google Scholar] [CrossRef]
  22. Wang, Y.; Wu, H.; Xia, C.; Lu, N. Impact of the Price of Gifts From Patients on Physicians’ Service Quality in Online Consultations: Empirical Study Based on Social Exchange Theory. J. Med. Internet Res. 2020, 22, e15685. [Google Scholar] [CrossRef]
  23. Liu, J.; He, J.; He, S.; Li, C.; Yu, C.; Li, Q. Patients’ Self-Disclosure Positively Influences the Establishment of Patients’ Trust in Physicians: An Empirical Study of Computer-Mediated Communication in an Online Health Community. Front. Public Health 2022, 10, 823692. [Google Scholar]
  24. Chiu, C.-M.; Hsu, M.-H.; Wang, E.T.G. Understanding Knowledge Sharing in Virtual Communities: An Integration of Social Capital and Social Cognitive Theories. Decis. Support Syst. 2006, 42, 1872–1888. [Google Scholar]
  25. Parhankangas, A.; Renko, M. Linguistic Style and Crowdfunding Success among Social and Commercial Entrepreneurs. J. Bus. Ventur. 2017, 32, 215–236. [Google Scholar]
  26. Johnson, S.L.; Safadi, H.; Faraj, S. The Emergence of Online Community Leadership. Inf. Syst. Res. 2015, 26, 165–187. [Google Scholar] [CrossRef]
  27. Hartzler, A.; Pratt, W. Managing the Personal Side of Health: How Patient Expertise Differs from the Expertise of Clinicians. J. Med. Internet Res. 2011, 13, e1728. [Google Scholar] [CrossRef]
  28. Zhang, L.; Yan, Q.; Zhang, L. A Text Analytics Framework for Understanding the Relationships among Host Self-Description, Trust Perception and Purchase Behavior on Airbnb. Decis. Support Syst. 2020, 133, 113288. [Google Scholar] [CrossRef]
  29. Zhou, J.; Zhang, Q.; Zhou, S.; Li, X.; Michael Zhang, X. Unintended Emotional Effects of Online Health Communities: A Text Mining-Supported Empirical Study. MIS Q. 2023, 47, 195–226. [Google Scholar] [CrossRef]
  30. Chen, L.; Baird, A.; Straub, D. A Linguistic Signaling Model of Social Support Exchange in Online Health Communities. Decis. Support Syst. 2020, 130, 113233. [Google Scholar] [CrossRef]
  31. Al Salman, A.; Kim, A.; Mercado, A.; Ring, D.; Doornberg, J.; Fatehi, A.; Crijns, T.J. Are Patient Linguistic Tones Associated with Mental Health and Perceived Clinician Empathy? J. Bone Jt. Surg. 2021, 103, 2181. [Google Scholar] [CrossRef]
  32. Mjaaland, T.A.; Finset, A.; Jensen, B.F.; Gulbrandsen, P. Physicians’ Responses to Patients’ Expressions of Negative Emotions in Hospital Consultations: A Video-Based Observational Study. Patient Educ. Couns. 2011, 84, 332–337. [Google Scholar] [CrossRef]
  33. Spence, M. Job Market Signaling. Q. J. Econ. 1973, 87, 355–374. [Google Scholar] [CrossRef]
  34. Cheung, C.M.K.; Xiao, B.S.; Liu, I.L.B. Do Actions Speak Louder than Voices? The Signaling Role of Social Information Cues in Influencing Consumer Purchase Decisions. Decis. Support Syst. 2014, 65, 50–58. [Google Scholar] [CrossRef]
  35. Wells, J.D.; Valacich, J.S.; Hess, T.J. What Signal Are You Sending? How Website Quality Influences Perceptions of Product Quality and Purchase Intentions. MIS Q. 2011, 35, 373. [Google Scholar] [CrossRef]
  36. Choi, H.S.; Ko, M.S.; Medlin, D.; Chen, C. The Effect of Intrinsic and Extrinsic Quality Cues of Digital Video Games on Sales: An Empirical Investigation. Decis. Support Syst. 2018, 106, 86–96. [Google Scholar] [CrossRef]
  37. Benlian, A.; Hess, T. The Signaling Role of IT Features in Influencing Trust and Participation in Online Communities. Int. J. Electron. Commer. 2011, 15, 7–56. [Google Scholar] [CrossRef]
  38. Kiyohara, L.Y.; Kayano, L.K.; Kobayashi, M.L.T.; Alessi, M.S.; Yamamoto, M.U.; Yunes-Filho, P.R.M.; Pessoa, R.R.; Mandelbaum, R.; Okubo, S.T.; Watanuki, T.; et al. The Patient-Physician Interactions as Seen by Undergraduate Medical Students. Sao Paulo Med. J. 2001, 119, 97–100. [Google Scholar] [CrossRef]
  39. Connelly, B.L.; Certo, S.T.; Ireland, R.D.; Reutzel, C.R. Signaling Theory: A Review and Assessment. J. Manag. 2011, 37, 39–67. [Google Scholar] [CrossRef]
  40. Jiang, S.; Liu, X.; Chi, X. Effect of Writing Style on Social Support in Online Health Communities: A Theoretical Linguistic Analysis Framework. Inf. Manag. 2022, 59, 103683. [Google Scholar] [CrossRef]
  41. Khurana, S.; Qiu, L.; Kumar, S. When a Doctor Knows, It Shows: An Empirical Analysis of Doctors’ Responses in a Q&A Forum of an Online Healthcare Portal. Inf. Syst. Res. 2019, 30, 872–891. [Google Scholar] [CrossRef]
  42. Yang, H.; Guo, X.; Wu, T.; Ju, X. Exploring the Effects of Patient-Generated and System-Generated Information on Patients’ Online Search, Evaluation and Decision. Electron. Commer. Res. Appl. 2015, 14, 192–203. [Google Scholar] [CrossRef]
  43. Ma, J.; Gao, S.; Wang, P.; Liu, Y. High Level of Self-Disclosure on SNSs Facilitates Cooperation: A Serial Mediation Model of Psychological Distance and Trust. Comput. Hum. Behav. 2024, 150, 107976. [Google Scholar] [CrossRef]
  44. Finkel, E.J.; Norton, M.I.; Reis, H.T.; Ariely, D.; Caprariello, P.A.; Eastwick, P.W.; Frost, J.H.; Maniaci, M.R. When Does Familiarity Promote Versus Undermine Interpersonal Attraction? A Proposed Integrative Model From Erstwhile Adversaries. Perspect. Psychol. Sci. 2015, 10, 3–19. [Google Scholar] [CrossRef] [PubMed]
  45. Utz, S. The Function of Self-Disclosure on Social Network Sites: Not Only Intimate, but Also Positive and Entertaining Self-Disclosures Increase the Feeling of Connection. Comput. Hum. Behav. 2015, 45, 1–10. [Google Scholar] [CrossRef]
  46. El Ouirdi, M.; Segers, J.; El Ouirdi, A.; Pais, I. Predictors of Job Seekers’ Self-Disclosure on Social Media. Comput. Hum. Behav. 2015, 53, 1–12. [Google Scholar] [CrossRef]
  47. Bae, S.; Jang, J.; Kim, J. Good Samaritans on Social Network Services: Effects of Shared Context Information on Social Supports for Strangers. Int. J. Hum.-Comput. Stud. 2013, 71, 900–918. [Google Scholar] [CrossRef]
  48. Krasnova, H.; Veltri, N.F.; Günther, O. Self-Disclosure and Privacy Calculus on Social Networking Sites: The Role of Culture. Bus. Inf. Syst. Eng. 2012, 4, 127–135. [Google Scholar] [CrossRef]
  49. Villarroel Ordenes, F.; Grewal, D.; Ludwig, S.; Ruyter, K.D.; Mahr, D.; Wetzels, M. Cutting through Content Clutter: How Speech and Image Acts Drive Consumer Sharing of Social Media Brand Messages. J. Consum. Res. 2019, 45, 988–1012. [Google Scholar] [CrossRef]
  50. Ansari, S.; Gupta, S. Customer Perception of the Deceptiveness of Online Product Reviews: A Speech Act Theory Perspective. Int. J. Inf. Manag. 2021, 57, 102286. [Google Scholar] [CrossRef]
  51. Tweedie, F.J.; Baayen, R.H. How Variable May a Constant Be? Measures of Lexical Richness in Perspective. Comput. Humanit. 1998, 32, 323–352. [Google Scholar] [CrossRef]
  52. Castro, C.M.; Wilson, C.; Wang, F.; Schillinger, D. Babel Babble: Physicians’ Use of Unclarified Medical Jargon with Patients. Am. J. Health Behav. 2007, 31, S85–S95. [Google Scholar] [PubMed]
  53. Yin, D.; Bond, S.D.; Zhang, H. Anxious or Angry? Effects of Discrete Emotions on the Perceived Helpfulness of Online Reviews. MIS Q. 2014, 38, 539–560. [Google Scholar] [CrossRef]
  54. Ludwig, S.; de Ruyter, K.; Mahr, D.; Wetzels, M.; Brüggen, E.; de Ruyck, T. Take Their Word for It: The Symbolic Role of Linguistic Style Matches in User Communities. MIS Q. 2014, 38, 1201–1218. [Google Scholar]
  55. Ouyang, P.; Wang, J.-J.; Jasmine Chang, A.-C. Patients Need Emotional Support: Managing Physician Disclosure Information to Attract More Patients. Int. J. Med. Inform. 2022, 158, 104674. [Google Scholar] [CrossRef]
  56. Kasimtseva, L.; Kiseleva, L.; Dzhabrailova, S. Doctor-Patient Communication as a Linguistic Model. In Proceedings of the 1st International Scientific Practical Conference “The Individual and Society in the Modern Geopolitical Environment” (ISMGE 2019), Volgograd, Russian, 23–29 May 2019; Atlantis Press: Dordrecht, The Netherlands, 2019. [Google Scholar] [CrossRef]
  57. Bientzle, M.; Griewatz, J.; Kimmerle, J.; Küppers, J.; Cress, U.; Lammerding-Koeppel, M. Impact of Scientific Versus Emotional Wording of Patient Questions on Doctor-Patient Communication in an Internet Forum: A Randomized Controlled Experiment with Medical Students. J. Med. Internet Res. 2015, 17, e268. [Google Scholar] [CrossRef]
  58. Greaves, F.; Ramirez-Cano, D.; Millett, C.; Darzi, A.; Donaldson, L. Use of Sentiment Analysis for Capturing Patient Experience From Free-Text Comments Posted Online. J. Med. Internet Res. 2013, 15, e239. [Google Scholar] [CrossRef]
  59. Gu, D.; Li, M.; Yang, X.; Gu, Y.; Zhao, Y.; Liang, C.; Liu, H. An Analysis of Cognitive Change in Online Mental Health Communities: A Textual Data Analysis Based on Post Replies of Support Seekers. Inf. Process. Manag. 2023, 60, 103192. [Google Scholar] [CrossRef]
  60. Wang, X.; High, A.; Wang, X.; Zhao, K. Predicting Users’ Continued Engagement in Online Health Communities from the Quantity and Quality of Received Support. J. Assoc. Inf. Sci. Technol. 2021, 72, 710–722. [Google Scholar] [CrossRef]
  61. Liu, X.; Chi, X.; Li, J.; Zhou, S.; Cheng, Y. Doctors’ Self-Presentation Strategies and the Effects on Patient Selection in Psychiatric Department from an Online Medical Platform: A Combined Perspective of Impression Management and Information Integration. J. Theor. Appl. Electron. Commer. Res. 2025, 20, 13. [Google Scholar] [CrossRef]
  62. Yao, T.; Zheng, Q.; Fan, X. The Impact of Online Social Support on Patients’ Quality of Life and the Moderating Role of Social Exclusion. J. Serv. Res. 2015, 18, 369–383. [Google Scholar] [CrossRef]
  63. Liu, S.; Si, G.; Gao, B. Which Voice Are You Satisfied with? Understanding the Physician-Patient Voice Interactions on Online Health Platforms. Decis. Support Syst. 2022, 157, 113754. [Google Scholar] [CrossRef]
  64. Zhang, X.; Guo, X.; Lai, K.; Yi, W. How Does Online Interactional Unfairness Matter for Patient-Doctor Relationship Quality in Online Health Consultation? The Contingencies of Professional Seniority and Disease Severity. Eur. J. Inf. Syst. 2019, 28, 336–354. [Google Scholar] [CrossRef]
  65. Park, S.; Chung, N. Mediating Roles of Self-Presentation Desire in Online Game Community Commitment and Trust Behavior of Massive Multiplayer Online Role-Playing Games. Comput. Hum. Behav. 2011, 27, 2372–2379. [Google Scholar] [CrossRef]
  66. Lu, X.; Jiang, J.; Head, M.; Yang, J. The Impact of Linguistic Complexity on Leadership in Online Q&A Communities: Comparing Knowledge Shaping and Knowledge Adding. Inf. Manag. 2022, 59, 103675. [Google Scholar] [CrossRef]
  67. Bernhardt, J.M.; Brownfield, E.D.; Parker, R.M.; Schwartzberg, J.G.; VanGeest, J.B.; Wang, C.C. Understanding Health Literacy: Implications for Medicine and Public Health. Chic. Am. Med. Assoc. 2005, 98, 980–981. [Google Scholar]
  68. Mackert, M.; Mabry-Flynn, A.; Champlin, S.; Donovan, E.E.; Pounders, K. Health Literacy and Health Information Technology Adoption: The Potential for a New Digital Divide. J. Med. Internet Res. 2016, 18, e6349. [Google Scholar] [CrossRef] [PubMed]
  69. Howard, D.J.; Gengler, C. Emotional Contagion Effects on Product Attitudes: Figure 1. J. Consum. Res. 2001, 28, 189–201. [Google Scholar] [CrossRef]
  70. Park, C.S. Applying “Negativity Bias” to Twitter: Negative News on Twitter, Emotions, and Political Learning. J. Inf. Technol. Polit. 2015, 12, 342–359. [Google Scholar] [CrossRef]
  71. Naing, S.Z.S.; Udomwong, P. Public Opinions on ChatGPT: An Analysis of Reddit Discussions by Using Sentiment Analysis, Topic Modeling, and SWOT Analysis. Data Intell. 2024, 6, 344–374. [Google Scholar] [CrossRef]
  72. Eslami, S.P.; Ghasemaghaei, M.; Hassanein, K. Which Online Reviews Do Consumers Find Most Helpful? A Multi-Method Investigation. Decis. Support Syst. 2018, 113, 32–42. [Google Scholar] [CrossRef]
  73. Wu, H.; Lu, N. Online Written Consultation, Telephone Consultation and Offline Appointment: An Examination of the Channel Effect in Online Health Communities. Int. J. Med. Inform. 2017, 107, 107–119. [Google Scholar] [CrossRef]
  74. Liu, Q.; Wang, B.; Kong, Y.; Cheng, K.K. China’s Primary Health-Care Reform. Lancet 2011, 377, 2064–2066. [Google Scholar] [CrossRef]
  75. Hu, Y.; Zhang, Z. Skilled Doctors in Tertiary Hospitals Are Already Overworked in China. Lancet Glob. Health 2015, 3, e737. [Google Scholar] [CrossRef] [PubMed]
  76. Fan, W.; Zhou, Q.; Qiu, L.; Kumar, S. Should Doctors Open Online Consultation Services? An Empirical Investigation of Their Impact on Offline Appointments. Inf. Syst. Res. 2023, 34, 629–651. [Google Scholar] [CrossRef]
  77. Sawicki, J.; Ganzha, M.; Paprzycki, M. The State of the Art of Natural Language Processing—A Systematic Automated Review of NLP Literature Using NLP Techniques. Data Intell. 2023, 5, 707–749. [Google Scholar] [CrossRef]
  78. Deng, Z.; Deng, Z.; Fan, G.; Wang, B.; Fan, W.; Liu, S. More Is Better? Understanding the Effects of Online Interactions on Patients Health Anxiety. J. Assoc. Inf. Sci. Technol. 2023, 74, 1243–1264. [Google Scholar] [CrossRef]
  79. Akkas, A.; Gaur, V.; Simchi-Levi, D. Drivers of Product Expiration in Consumer Packaged Goods Retailing. Manag. Sci. 2019, 65, 2179–2195. [Google Scholar] [CrossRef]
  80. Li, Y.; Zhao, L. Collaborating with Bounty Hunters: How to Encourage White Hat Hackers’ Participation in Vulnerability Crowdsourcing Programs through Formal and Relational Governance. Inf. Manag. 2022, 59, 103648. [Google Scholar] [CrossRef]
  81. Jing, D.; Jin, Y.; Liu, J. The Impact of Monetary Incentives on Physician Prosocial Behavior in Online Medical Consulting Platforms: Evidence From China. J. Med. Internet Res. 2019, 21, e14685. [Google Scholar] [CrossRef]
  82. Liu, X.; Wang, G.A.; Fan, W.; Zhang, Z. Finding Useful Solutions in Online Knowledge Communities: A Theory-Driven Design and Multilevel Analysis. Inf. Syst. Res. 2020, 31, 731–752. [Google Scholar] [CrossRef]
  83. Street, R.L.; Makoul, G.; Arora, N.K.; Epstein, R.M. How Does Communication Heal? Pathways Linking Clinician–Patient Communication to Health Outcomes. Patient Educ. Couns. 2009, 74, 295–301. [Google Scholar] [CrossRef] [PubMed]
  84. Liu, S.; Wang, H.; Gao, B.; Deng, Z. Doctors’ Provision of Online Health Consultation Service and Patient Review Valence: Evidence from a Quasi-Experiment. Inf. Manag. 2022, 59, 103360. [Google Scholar] [CrossRef]
  85. Liu, S.; Zhang, M.; Gao, B.; Jiang, G. Physician Voice Characteristics and Patient Satisfaction in Online Health Consultation. Inf. Manag. 2020, 57, 103233. [Google Scholar] [CrossRef]
  86. Liang, C.; Wang, X.; Gu, D.; Li, P.; Chen, H.; Xu, Z. Smart Management Information Systems (Smis): Concept, Evolution, Research Hotspots and Applications. Data Intell. 2023, 5, 857–884. [Google Scholar] [CrossRef]
  87. Zhao, W.; Gu, D.; Yang, X.; Jia, M.; Liang, C.; Wang, X.; Zolotarev, O. MedT2T: An Adaptive Pointer Constrain Generating Method for a New Medical Text-to-Table Task. Future Gener. Comput. Syst. 2024, 161, 586–600. [Google Scholar] [CrossRef]
Jtaer 20 00056 g001
Figure 1. Research model.
Figure 1. Research model.
Jtaer 20 00056 g001
Table 1. Description of variables.
Table 1. Description of variables.
Variable TypeVariableDescription
Dependent VariablesEmotional Support
(DEmoS)		The total amount of emotional support expressed by doctors in a round of doctor–patient interaction.
Informational Support
(DInfoS)		The total amount of information support expressed by the doctor in a round of doctor–patient interaction.
Independent VariableCompletenessThe degree of detail in patients’ self-disclosure is measured by the total length of the patient’s text in a round of doctor–patient interaction.
ReadabilityThe comprehensibility of patient text in a round of doctor–patient interaction.
ExpertiseThe average number of medical terms used by patients in a round of doctor–patient interactions.
Sentiment Polarity
(Psentiment)		Average patient emotional polarity during a doctor–patient interaction. The closer it is to 1, the more positive the emotion, and the closer it is to −1, the more negative the emotion.
Control VariablesPatient Gender
(PGender)		0 represents male, and 1 represents female.
Patient Age
(PAge)		The actual age of the patient, excluding patients with age 0.
Doctor Gender
(DGender)		0 represents male, and 1 represents female.
Employment Grade
(DEmGrade)		Categorical variable, doctor’s professional title: 1 for physician, 2 for attending physician, 3 for associate chief physician, and 4 for chief physician.
Recommendation Heat
(DReHeat)		Continuous variable, doctor’s online recommendation popularity: 0–5 points.
Hospital Level
(HosLevel)		Categorical variable, representing the hospital tier where the doctor practices: 0 for unranked hospitals, 1 for Tier 1, 2 for Tier 1A, 3 for Tier 2, 4 for Tier 2A, 5 for Tier 3, and 6 for Tier 3A.
Communication Time
(CommuTime)		The total number of doctor–patient exchanges in one round of doctor–patient interaction.
Table 2. Descriptive statistics.
Table 2. Descriptive statistics.
VariablesNMeanSDMinMaxVIF
DInfoS52,73088.1296.2902963
DEmoS52,7300.7651.206025
Completeness52,730354.4321.4274811.590
Readability52,7300.9920.0100.5310.9991.040
Expertise52,7300.0750.0300.0020.5001.070
PSentiment52,730−0.1540.206−111.070
PGender52,7300.6370.481011.030
PAge52,73049.1214.791901.100
DGender52,7300.8080.394011.070
DReHeat52,7303.9160.5331.70051.480
DEmGrade52,7303.4250.756141.380
HosLevel52,7305.8840.525061.050
CommuTime52,73021.5121.0416441.590
Table 3. Correlations of variables.
Table 3. Correlations of variables.
Variables12345678910111213
1 DInfoS1
2 DEmoS0.512 ***1
3 Completeness0.294 ***0.228 ***1
4 Readability0.002000.011 ***0.065 ***1
5 AHLiteracy−0.098 ***−0.120 ***−0.110 ***0.007 *1
6 PSentiment−0.00600−0.047 ***−0.136 ***−0.106 ***−0.040 ***1
7 PGender−0.046 ***−0.010 **−0.00200−0.003000.069 ***−0.008 **1
8 PAge−0.099 ***−0.083 ***0.011 ***0.024 ***0.180 ***−0.024 ***0.108 ***1
9 DGender−0.037 ***0.074 ***−0.040 ***0.011 ***0.035 ***−0.039 ***0.107 ***0.118 ***1
10 DReHeat0.015 ***0.108 ***0.084 ***0.072 ***0.090 ***−0.063 ***0.089 ***0.216 ***0.181 ***1
11 DEmGrade0.004000.044 ***0.103 ***0.083 ***0.061 ***−0.060 ***0.034 ***0.125 ***0.017 ***0.501 ***1
12 HosLevel−0.085 ***−0.086 ***−0.046 ***0.024 ***0.053 ***0.00200−0.014 ***0.071 ***−0.059 ***0.169 ***0.106 ***1
13 CommuTime0.368 ***0.319 ***0.553 ***−0.075 ***−0.172 ***0.086 ***0.010 **−0.031 ***−0.015 ***0.013 ***−0.066 ***−0.086 ***1
Notes: *** p < 0.01, ** p < 0.05, * p < 0.1.
Table 4. Ordinary least squares (OLS) regression results.
Table 4. Ordinary least squares (OLS) regression results.
ModelModel 1Model 2
VariablesDInfoSDEmoS
Completeness11.063 ***0.047 ***
(9.01)(3.59)
Readability1.544 ***0.022 ***
(3.64)(4.40)
Expertise−1.055 **−0.074 ***
(−2.71)(−14.91)
PSentiment−1.655 ***−0.077 ***
(−3.58)(−13.98)
PGender−8.205 ***−0.043 ***
(−9.70)(−4.20)
PAge−8.383 ***−0.112 ***
(−19.72)(−22.01)
DGender−6.159 ***0.198 ***
(−5.94)(17.26)
DReHeat2.685 ***0.134 ***
(6.08)(21.20)
DEmGrade1.382 **0.025 ***
(3.29)(4.38)
HosLevel−5.135 ***−0.078 ***
(−10.37)(−11.72)
CommuTime27.872 ***0.342 ***
(20.47)(20.78)
Constant98.319 ***0.633 ***
(94.99)(55.66)
Observations52,73052,730
R-squared0.1550.139
F215.9281.9
Notes: Robust t-statistics in parentheses; ** p < 0.01, *** p < 0.001.
Table 5. Summary of hypotheses testing.
Table 5. Summary of hypotheses testing.
No.HypothesesResults
H1aThe completeness of patient self-disclosure positively influences doctors’ informational support.Supported
H1bThe completeness of patient self-disclosure positively influences doctors’ emotional support.Supported
H2aThe readability of patient self-disclosure positively influences doctors’ informational support.Supported
H2aThe readability of patient self-disclosure positively influences doctors’ emotional support.Supported
H3aThe expertise of patient self-disclosure positively influences doctors’ informational support.Not supported
H3bThe expertise of patient self-disclosure positively influences doctors’ emotional support.Not supported
H4aThe positive sentiment of patient self-disclosure negatively influences doctors’ informational support.Supported
H4bThe positive sentiment of patient self-disclosure negatively influences doctors’ emotional support.Supported
Table 6. Robustness checks.
Table 6. Robustness checks.
Checks 1Checks 2
ModelModel 3Model 4Model 5Model 6
VariablesDInfoSDEmoSDTextLenDResNum
Completeness0.058 ***0.030 ***29.609 ***0.436 ***
(9.41)(3.64)(9.99)(3.84)
Readability0.018 ***0.038 ***3.496 ***−0.088 ***
(3.50)(4.06)(3.70)(−3.64)
Expertise−0.010 *−0.115 ***−2.678 **−0.197 ***
(−2.04)(−14.83)(−3.00)(−9.10)
PSentiment−0.023 ***−0.097 ***−2.395 *0.185 ***
(−4.87)(−15.00)(−2.28)(6.08)
PGender−0.098 ***−0.056 ***−20.091 ***−0.194 ***
(−11.03)(−4.33)(−10.31)(−5.65)
PAge−0.098 ***−0.141 ***−20.765 ***−0.215 ***
(−22.34)(−21.71)(−21.29)(−12.40)
DGender−0.049 ***0.274 ***−14.741 ***−0.526 ***
(−4.68)(15.25)(−6.10)(−11.94)
DReHeat0.043 ***0.179 ***8.034 ***0.230 ***
(8.50)(24.01)(7.69)(10.45)
DEmGrade0.031 ***0.046 ***0.849−0.548 ***
(6.87)(5.91)(0.85)(−17.49)
HosLevel−0.067 ***−0.081 ***−13.440 ***−0.235 ***
(−11.81)(−13.52)(−11.57)(−8.59)
CommuTime0.333 ***0.340 ***70.784 ***5.758 ***
(42.60)(32.16)(20.17)(27.56)
lnalpha−0.312 ***−0.419 ***--
(−44.28)(−18.06)
Constant4.508 ***−0.578 ***237.985 ***9.564 ***
(431.08)(−32.27)(99.16)(221.02)
Observations52,73052,73052,73052,730
Log likelihood−284,046−60,380--
R-squared--0.1680.698
Pseudo R-squared0.01650.0521--
F--250.71719
Notes: Robust t-statistics in parentheses in Models 3 and 4; Robust z-statistics in parentheses in Models 5 and 6; * p < 0.05, ** p < 0.01, *** p < 0.001.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Peng, M.; Zhu, K.; Gu, Y.; Yang, X.; Su, K.; Gu, D. How to Self-Disclose? The Impact of Patients’ Linguistic Features on Doctors’ Service Quality in Online Health Communities. J. Theor. Appl. Electron. Commer. Res. 2025, 20, 56. https://doi.org/10.3390/jtaer20020056

AMA Style

Peng M, Zhu K, Gu Y, Yang X, Su K, Gu D. How to Self-Disclose? The Impact of Patients’ Linguistic Features on Doctors’ Service Quality in Online Health Communities. Journal of Theoretical and Applied Electronic Commerce Research. 2025; 20(2):56. https://doi.org/10.3390/jtaer20020056

Chicago/Turabian Style

Peng, Mengyuan, Kaixuan Zhu, Yadi Gu, Xuejie Yang, Kaixiang Su, and Dongxiao Gu. 2025. "How to Self-Disclose? The Impact of Patients’ Linguistic Features on Doctors’ Service Quality in Online Health Communities" Journal of Theoretical and Applied Electronic Commerce Research 20, no. 2: 56. https://doi.org/10.3390/jtaer20020056

APA Style

Peng, M., Zhu, K., Gu, Y., Yang, X., Su, K., & Gu, D. (2025). How to Self-Disclose? The Impact of Patients’ Linguistic Features on Doctors’ Service Quality in Online Health Communities. Journal of Theoretical and Applied Electronic Commerce Research, 20(2), 56. https://doi.org/10.3390/jtaer20020056

Article Metrics

Back to TopTop