The methodological approach behind the study deserves closer attention, because it illustrates how text mining has matured into a practical instrument for health services research. The researchers gathered reviews about several top-tier large general Grade-A tertiary hospitals from Dazhongdianping, a widely used domestic review platform whose public postings resemble a hybrid of consumer commentary and personal narrative. Rather than relying on a single analytical technique, the team combined the text-analysis software ROST CM 6.0 with custom Python programming to carry out a sequence of complementary procedures: data collection, high-frequency word analysis, semantic network analysis, sentiment analysis, and trend analysis. Each of these steps answers a different kind of question. High-frequency word analysis reveals which terms dominate the conversation, semantic network analysis shows how those terms relate to one another in patients’ mental maps of a hospital visit, sentiment analysis classifies the emotional valence of the reviews, and trend analysis tracks how attention and feeling shift over time.

The decision to layer these techniques matters because any one of them, used alone, can mislead. A word that appears frequently may do so in both praise and complaint, and a sentiment classifier applied without contextual grounding can mistake clinical terminology for emotional language. By triangulating across methods, the researchers could distinguish, for example, between the sheer prominence of a topic and the emotional charge attached to it. This is particularly important in the medical setting, where the same word such as an appointment system can anchor positive sentiment when it works smoothly and negative sentiment when it fails. The study’s design therefore reflects a broader trend in which computational text analysis is treated not as a single tool but as a toolkit whose components check and correct one another.

The headline finding, that patients paid the most attention to medical staff’s service attitude and to medical convenience, aligns with a substantial body of patient-experience research conducted through surveys, yet it carries a distinctive weight because it emerges from spontaneous, unprompted commentary. Survey respondents answer the questions they are asked; review writers choose their own topics. When thousands of people independently converge on attitude and convenience without being steered toward them, the signal is arguably stronger than what structured instruments capture. It suggests that these two dimensions function as the primary lenses through which patients interpret an entire hospital encounter, with judgments about medical service quality diverging around them rather than around clinical outcomes.

The emotional geography of the reviews is equally instructive. Positive sentiment clustered around appointments and facilities, meaning that patients felt good about the hospital when scheduling was manageable and the physical environment met their expectations. Negative sentiment, by contrast, concentrated on process order and doctor-patient communication. The asymmetry is telling: the things patients praise tend to be structural and impersonal, while the things that provoke frustration tend to involve sequence, fairness, and human interaction. A patient can forgive an aging waiting room more easily than a chaotic queue or a physician who explains nothing. Communication failures appear to be especially corrosive because they occur at the point of maximum vulnerability, when a patient is seeking not just treatment but understanding.

The roughly even split between positive and negative sentiment across the corpus is itself a finding worth pausing over. It indicates that online reviews of even elite hospitals are not dominated by either cheerleading or grievance, but instead form a genuinely mixed record. For hospital administrators, this cuts both ways. There is no large reservoir of goodwill that excuses systemic problems, and there is no pattern of relentless complaint that suggests the institutions are failing outright. The feedback stream is best read as a running balance sheet in which specific, fixable aspects of the experience determine which side of the ledger a given visit lands on.

The temporal patterns the authors identified add another dimension that cross-sectional surveys routinely miss. Patients’ concerns and sentiments were not static; they shifted over time, which implies that the drivers of patient experience evolve with changes in hospital operations, platform culture, and perhaps broader expectations about digital services. A hospital that surveys its patients once a year may be measuring a moving target. Continuous monitoring of review streams, as this study demonstrates, can reveal when a particular concern is rising before it hardens into a reputational problem, giving administrators a form of early warning that traditional feedback mechanisms cannot provide.

Perhaps the most consequential observation in the study is the low attention paid to technical factors. Reviews concentrated on non-technical elements such as attitude, process, and waiting time, while clinical skill, the very thing hospitals most pride themselves on, received comparatively little mention. The authors interpret this as evidence of the substantial impact non-technical factors have on the overall medical experience. There are several plausible readings of this pattern. One is that patients lack the expertise to evaluate technical quality directly and therefore use proxy signals, such as how respectfully they were treated, to infer it. Another is that technical competence is assumed at top-tier institutions, so it simply does not register as noteworthy, while lapses in courtesy or efficiency do. Either way, the practical implication is the same: a hospital’s reputation among the public is built largely in the waiting room and at the registration desk, not only in the operating theater.

The study’s institutional context is worth noting. All of the affiliated authors are based at Xijing Hospital of the Fourth Military Medical University in Xi’an, spanning departments of medical insurance, information, and gastroenterology, and the work was supported by the hospital’s Booster Plan. This mix of administrative, technical, and clinical perspectives is fitting for a project that sits at the intersection of data science and patient care. The authors state that the study does not involve ethical issues and required no ethical approval, and they declare no competing interests. The data were publicly accessible, drawn from a platform whose postings users had already chosen to share openly, which places this kind of research in a comparatively uncomplicated ethical position relative to studies that collect private health information.

The publication timeline also reflects the current pace of health services research. The manuscript was received in December 2025, accepted in September 2026, and published on 23 September 2026 as an open access article under a Creative Commons Attribution-NonCommercial-NoDerivatives license. Readers encountering the article now are seeing a version shared early to provide faster access to peer-reviewed, accepted research; it is citable and carries a permanent DOI, but it remains subject to further edits before the final Version of Record replaces it automatically. For a study whose subject is the speed and transparency of patient feedback, there is a certain symmetry in a publication model designed to shorten the distance between acceptance and availability.

For health systems outside China, the findings travel with some caveats but also with considerable relevance. Review platforms, appointment systems, and hospital hierarchies differ across countries, yet the underlying psychology of the patient as a person navigating an unfamiliar, anxiety-inducing environment is close to universal. The study validates online review analysis as a feasible research approach for medical services, and its conclusion is directed squarely at top large medical institutions: improve service attitude, strengthen communication, streamline processes, and enhance the overall experience, all while continuing to focus on professional technology. In effect, the patients writing these reviews have delivered a consistent message that technical excellence is the price of entry, but the remembered quality of care is written in the tone of a conversation, the length of a wait, and the ease of finding one’s way through a complicated system. Hospitals that listen to this unfiltered stream of voices gain access to a continuous, candid audit of exactly those dimensions that patients weigh most heavily, and that no annual survey can fully replicate.

Subject of Research: Text mining of online patient reviews to identify what shapes patient experience and sentiment in large Chinese tertiary hospitals

Article Title: Decoding patients’ voices: how social media reviews reveal the true impact of medical service

Article References: Liu, X., Zhao, L., Liu, W., Hu, D., & Chen, M. (2026). Decoding patients’ voices: how social media reviews reveal the true impact of medical service. BMC Health Services Research. https://doi.org/10.1186/s12913-026-15674-w

Image Credits: AI Generated

DOI: 10.1186/s12913-026-15674-w

Keywords: patient experience, online reviews, social media, text mining, sentiment analysis, hospital care, health services research, service attitude, doctor-patient communication, tertiary hospitals, China, patient satisfaction

The core difficulty the researchers set out to solve lies in a family of techniques known as Multi-Attribute Decision-Making, or MADM. These methods are designed to rank alternatives—here, scenic spots—against a set of weighted criteria. They work elegantly when the number of attributes is small and the data are homogeneous. But modern tourism evaluation data, harvested at scale from platforms such as Ctrip, violate both assumptions. Hundreds of attributes may be involved, many of them correlated, some redundant, and some outright contradictory in how they rank the same destinations. Feeding such high-dimensional data directly into a decision model can distort results, because redundant or conflicting attributes exert an outsized influence on the final score.

The team’s solution unfolds in three stages. The first stage is a consistency-driven dynamic clustering algorithm. Instead of clustering the scenic spots themselves, the method clusters the evaluation attributes, grouping together those criteria that produce similar ranking patterns across the destinations being compared. In other words, if two attributes—say, visitor ratings of cleanliness and ratings of facility maintenance—consistently order the scenic spots in nearly the same way, they are placed in the same cluster. The word dynamic is important: the algorithm adjusts cluster assignments as it processes the data rather than fixing them in advance, and it uses measures of rank consistency to decide when attributes genuinely belong together. This step reduces the dimensionality of the problem in a principled way, without discarding any information outright.

The second stage applies a well-established decision tool within each cluster: TOPSIS, the Technique for Order Preference by Similarity to Ideal Solution. TOPSIS, introduced in the operations research literature in the 1980s and refined through the 1990s, ranks alternatives by measuring how close each one is to an ideal positive solution and how far it is from a negative ideal solution. Within a single cluster of similar attributes, this is a reasonable and well-conditioned calculation, because the attributes behave coherently. Running TOPSIS separately inside each cluster yields a set of local rankings—one ordered list of scenic spots per cluster—each reflecting the perspective of a coherent family of evaluation criteria.

The third stage is where the method borrows from an unexpected field: voting theory. The researchers use the Borda count, an aggregation procedure devised in the eighteenth century by Jean-Charles de Borda, to fuse the local rankings into a single global ranking. Under the Borda count, each local ranking awards points to each scenic spot according to its position, and the spot with the greatest accumulated total across all clusters wins the top place. This is the same arithmetic that determines winners in some ranked-ballot elections, and its appeal here is the same as in politics: it aggregates many partial, imperfect orderings into one socially robust verdict while dampening the influence of any single anomalous cluster. The authors note that prior theoretical work has argued the Borda count is in an important sense the optimal ranking method, which makes it a natural choice for fusion.

The pipeline was tested on evaluation data for China’s 5A scenic spots, collected from publicly accessible tourism review information on the Ctrip platform. The 5A designation is the pinnacle of China’s tourist rating system, so the destinations involved are already elite; the question was whether a data-driven ranking built from raw, high-dimensional visitor evaluations could reproduce and refine the official assessments. The empirical results showed a high degree of consistency between the method’s output rankings and existing tourism ratings, suggesting that the clustering-and-fusion pipeline captures the same underlying quality signals that expert evaluators respond to, but derives them transparently from visitor data rather than from a fixed administrative rubric.

Why not simply cluster the destinations instead of the attributes? The authors’ design choice reflects the structure of the problem. Tourists evaluate the same scenic spot along many dimensions, and the statistical relationships among those dimensions—how they correlate, which ones move together across destinations—carry the real information. By grouping attributes with similar ranking behaviour, the method effectively discovers latent evaluation perspectives: bundles of criteria that visitors treat as a single experiential quality. Each cluster then functions as a coherent judge, and the Borda count lets these judges vote. This architecture mirrors recent work on large-scale group decision-making, in which clustering and consensus techniques are used to manage hundreds or thousands of participants, and it connects to a growing literature on ranking fusion for massive online review datasets.

The technical payoff is robustness against heterogeneity. Real-world evaluation data are rarely generated by a single consistent process: different visitors weight different concerns, different attributes suffer different amounts of noise, and some criteria may be systematically inflated or deflated on particular platforms. A monolithic MADM model must average all of these effects together, which can wash out genuine distinctions between destinations. The cluster-then-rank-then-fuse design isolates each coherent evaluation channel, ranks destinations within that channel, and only then combines the verdicts. Conflicting attributes are no longer forced into a single weighted sum; they are allowed to speak in their own cluster and then reconciled at the voting stage. The authors position the framework as particularly valuable for scenarios involving high-dimensional and heterogeneous evaluation data, where conventional approaches struggle.

The practical implications extend beyond holiday planning. Tourism management authorities could use such a framework to monitor how the data-driven ranking of attractions evolves as visitor sentiment shifts, identifying destinations whose reputations are improving or deteriorating and tracing those changes back to specific clusters of attributes. Travel platforms could integrate the method into recommendation engines that must reconcile millions of reviews across many criteria. And the underlying algorithmic idea—cluster the criteria, rank within clusters, fuse by voting—is domain-agnostic, potentially transferable to any multi-attribute decision problem drowning in high-dimensional data, from product selection to service quality assessment.

The study also offers a candid account of its data foundations. The datasets were drawn from publicly accessible tourism evaluation information on Ctrip, and the full raw dataset cannot be redistributed because of third-party platform terms of use and intellectual-property restrictions. To support reproducibility, the published paper documents the data source, structure, preprocessing, variable construction, clustering procedure, parameter settings and experimental design in detail, with derived statistical information and non-restricted processed data available from the corresponding author on reasonable request. The research was supported by the Hunan Provincial Natural Science Foundation of China and the Scientific Research Fund of the Hunan Provincial Education Department. As online reviews continue to swell into one of the largest bodies of human judgement ever collected, methods like this one—part statistics, part social choice theory—point toward a future in which the wisdom of crowds is not merely accumulated, but properly counted.

Subject of Research: A dynamic clustering and ranking fusion method for multi-attribute evaluation of 5A scenic spots using tourist review data

Article Title: Dynamic clustering and ranking fusion for tourism recommendations: A data-driven approach to 5A scenic spot evaluation

Article References: Li, W., Xiao, Z., Yu, B., & Xu, Z. (2026). Dynamic clustering and ranking fusion for tourism recommendations: A data-driven approach to 5A scenic spot evaluation. Applied Intelligence, 56(15), Article 433. https://doi.org/10.1007/s10489-026-07488-4

Image Credits: AI Generated

DOI: 10.1007/s10489-026-07488-4

Keywords: tourism recommendation, dynamic clustering, ranking fusion, TOPSIS, Borda count, multi-attribute decision-making, high-dimensional data, 5A scenic spots, online reviews, Ctrip, consistency-driven clustering, Applied Intelligence