The Use of Web-Based Patient Reviews to Assess Medical Oncologists' Competency: Mixed Methods Sequential Explanatory Study.

BACKGROUND: Patients increasingly use web-based evaluation tools to assess their physicians, health care teams, and overall medical experience. OBJECTIVE: This study aimed to evaluate the extent to which the standardized physician competencies of the CanMEDS Framework are present in web-based patient reviews (WPRs) and to identify patients' perception of important physician qualities in the context of quality cancer care. METHODS: The WPRs of all university-affiliated medical oncologists in midsized cities with medical schools in the province of Ontario (Canada) were collected. Two reviewers (1 communication studies researcher and 1 health care professional) independently assessed the WPRs according to the CanMEDS Framework and identified common themes. Comment scores were then evaluated to identify κ agreement rates between the reviewers, and a descriptive quantitative analysis of the cohort was completed. Following the quantitative analysis, an inductive thematic analysis was performed. RESULTS: This study identified 49 actively practicing university-affiliated medical oncologists in midsized urban areas in Ontario. A total of 473 WPRs reviewing these 49 physicians were identified. Among the CanMEDS competencies, those defining the roles of medical experts, communicators, and professionals were the most prevalent (303/473, 64%; 182/473, 38%; and 129/473, 27%, respectively). Common themes in WPRs include medical skill and knowledge, interpersonal skills, and answering questions (from the patient to the physician). Detailed WPRs tend to include the following elements: experience and connection; discussion and evaluation of the physician's knowledge, professionalism, interpersonal skills, and punctuality; in positive reviews, the expression of feelings of gratitude and a recommendation; and in negative reviews, discouragement from seeking the physician's care. Patients' perception of medical skills is less specific than their perception of interpersonal qualities, although medical skills are the most commented-on element of care in WPRs. Patients' perception of interpersonal skills (listening, compassion, and overall caring demeanor) and other experiential phenomena, such as feeling rushed during appointments, is often specific and detailed. Details about a physician's interpersonal skills or "bedside manner" are highly perceived, valued, and shareable in an WPR context. A small number of WPRs reflected a distinction between the value of medical skills and that of interpersonal skills. The authors of these WPRs claimed that for them, a physician's medical skills and competence are more important than their interpersonal skills. CONCLUSIONS: CanMEDS roles and competencies that are explicitly patient facing (ie, those directly experienced by patients in their interactions with physicians and through the care that physicians provide) are the most likely to be present and reported on in WPRs. The findings demonstrate the opportunity to learn from WPRs, not simply to discern physicians' popularity but to grasp what patients may expect from their physicians. In this context, WPRs can represent a method for the measurement and assessment of patient-facing physician competency.

Identifying adverse drug reactions from patient reviews on social media using natural language processing.

Drugs have the potential of causing adverse reactions or side effects and prior knowledge of these reactions can help prevent hospitalizations and premature deaths. Public databases of common adverse drug reactions (ADRs) depend on individual reports from drug manufacturers and health professionals. However, this passive approach to ADR surveillance has been shown to suffer from severe under-reporting. Social media, such as online health forums where patients across the globe willingly share their drug intake experience, is a viable and rich source for detecting unreported ADRs. In this paper, we design an ADR Detection Framework (ADF) using Natural Language Processing techniques to identify ADRs in drug reviews mined from social media. We demonstrate the applicability of ADF in the domain of Diabetes by identifying ADRs associated with diabetes drugs using data extracted from three online patient-based health forums: askapatient.com, webmd.com, and iodine.com. Next, we analyze and visualize the ADRs identified and present valuable insights including prevalent and less prevalent ADRs, age and gender differences in ADRs detected, as well as the previously unknown ADRs detected by our framework. Our work could promote active (real-time) ADR surveillance and also advance pharmacovigilance research.

Communicating Commitment to Antibiotic Stewardship: an Effective Strategy for Responding to Online Patient Reviews.

BACKGROUND: Public health officials have worked to address the growing threat of antibiotic resistance. To slow the emergence of antimicrobial-resistant bacteria, it is important to improve patients' understanding of antibiotics and adjust their expectations of them. This study explores strategic antibiotic resistance communication between patients and health care providers in an online review platform. METHODS: Based on two experimental studies, we demonstrate the effectiveness of the provider's commitment messaging when dealing with patients' complaints about not receiving requested antibiotics during their visit. RESULTS: The findings from study 1 show that communicating the commitment to antibiotic stewardship makes participants have more favorable feelings toward the provider. Commitment messaging also makes readers perceive the provider as more credible, and they are more willing to visit the clinic in the future. Study 2 findings demonstrate a robustness of commitment messaging in increasing readers' willingness to visit the clinic, while the provider's response exhibits the limited impact of correcting patients' common misunderstandings of antibiotics and adjusting their expectations of antibiotics. CONCLUSIONS: Our findings illustrate some clear benefits of engaging with negative online patient reviews to minimize potential reputational damage and reestablish the credibility of care providers.

How Are Patients Reviewing Spine Surgeons Online? A Sentiment Analysis of Physician Review Website Written Comments.

STUDY DESIGN: A Sentiment Analysis of online reviews of spine surgeons. OBJECTIVES: Physician review websites have significant impact on a patient's provider selection. Written reviews are subjective, but sentiment analysis through machine learning can quantitatively analyze these reviews. This study analyzes online written reviews of spine surgeons and reports biases associated with demographic factors and trends in words utilized. METHODS: Online written and star-reviews of spine surgeons were obtained from healthgrades.com. A sentiment analysis package was used to analyze the written reviews. The relationship of demographic variables to these scores was analyzed with t-tests and word and bigram frequency analyses were performed. Additionally, a multiple regression analysis was performed on key terms. RESULTS: 8357 reviews of 480 surgeons were analyzed. There was a significant difference between the means of sentiment analysis scores and star scores for both gender and age. Younger, male surgeons were rated more highly on average (P < .01). Word frequency analysis indicated that behavioral factors and pain were the main contributing factors to both the best and worst reviewed surgeons. Additionally, several clinically relevant words, when included in a review, affected the odds of a positive review. CONCLUSIONS: The best reviews laud surgeons for their ability to manage pain and for exhibiting positive bedside manner. However, the worst reviews primarily focus on pain and its management, as exhibited by the frequency and multivariate analysis. Pain is a clear contributing factor to reviews, thus emphasizing the importance of establishing proper pain expectations prior to any intervention.

Are Online Patient Reviews Associated With Health Care Outcomes? A Systematic Review of the Literature.

Although online patient reviews (OPRs) are ubiquitous, there has been debate about whether OPRs are associated with measures of health care outcomes. The majority of patients search for online information about a hospital or provider before scheduling an encounter, therefore, the extent to which OPRs are valid information is important. We conducted a systematic review of the literature to examine the extent to which OPRs are associated with measures of patient experience, clinical quality, and other outcomes. We searched Medline, CINAHL, Web of Science, and PsycInfo, which yielded 32 studies. There were consistent positive correlations between OPRs and patient experience at both the organization and individual provider levels of analysis. However, associations between OPRs and quality measures were mixed. Organizational level OPRs may be more reliable than individual provider OPRs. In addition, the strength of association could be affected by the type of encounter setting, specialty, and specific measures.

Understanding patient reviews with minimum supervision.

Understanding patient opinions expressed towards healthcare services in online platforms could allow healthcare professionals to respond to address patients' concerns in a timely manner. Extracting patient opinion towards various aspects of health services is closely related to aspect-based sentiment analysis (ABSA) in which we need to identify both opinion targets and target-specific opinion expressions. The lack of aspect-level annotations however makes it difficult to build such an ABSA system. This paper proposes a joint learning framework for simultaneous unsupervised aspect extraction at the sentence level and supervised sentiment classification at the document level. It achieves 98.2% sentiment classification accuracy when tested on the reviews about healthcare services collected from Yelp, outperforming several strong baselines. Moreover, our model can extract coherent aspects and can automatically infer the distribution of aspects under different polarities without requiring aspect-level annotations for model learning.

Potential Adverse Events Reported With the Janus Kinase Inhibitors Approved for the Treatment of Rheumatoid Arthritis Using Spontaneous Reports and Online Patient Reviews.

The aim of this study was to analyze the potential adverse events (AEs) caused by Janus kinase (JAK) inhibitors, including tofacitinib, baricitinib, and upadacitinib, used to treat rheumatoid arthritis using spontaneous AE reports from the FDA (FAERS) and interpreting them in correlation with those from Korea (KAERS) and an online patient review (WebMD). Potential AEs were identified based on a disproportionality analysis using the proportional reporting ratio (PRR), reporting odds ratio (ROR), and the information component (IC). A total of 23,720 reports were analyzed from FAERS database, of which 91.5% were reports on tofacitinib. Potentially important medical AEs related to infections were reported frequently, as well as thromboembolism-related AEs. The AEs, such as malignancy, interstitial lung diseases, myocardial infarction, and gastrointestinal disorder, also reported. In an online patient review report, the ineffectiveness of the drug and gastrointestinal AEs were frequently reported. Infection with baricitinib and symptoms related to pain or edema due to upadacitinib were the main discomfort experienced by patients. In conclusion, the results of this study highlight the possible safety issues associated with JAK inhibitors. Routine clinical observations and further research using various real-world databases are needed.

Using artificial intelligence to examine online patient reviews.

Healthcare consumers are increasingly turning to online sources such as educational websites, forums and social media platforms to share their experiences with medical services and to demystify the uncertainties associated with undergoing various procedures. This study demonstrates a non-invasive way of understanding the feelings and emotions that consumers share via electronic word of mouth. By using IBM Watson, a content analysis tool that harnesses artificial intelligence, we show how a large amount of unstructured qualitative data can be transformed into quantitative data that can be subsequently analysed to generate novel insights into what patients are sharing about their healthcare experiences online.

Concierge care and patient reviews.

We examine how patient numerical ratings and specific words in written reviews of family physicians and internists in the states of California and Florida differ based upon concierge doctor status. Data are drawn from Healthgrades.com, one of the largest providers of online reviews, and a machine-learning sentiment analysis is used to determine the predictors of concierge status and numerical patient ratings. We find that reviews of concierge doctors are more likely to contain technical words associated with health care, such as "staff" and "office," compared with traditional physicians. In contrast, interpersonal bedside-manner words, like "listen" or "concerns," are most likely in reviews for nonconcierge doctors. We further determine that, whereas interpersonal words exhibit both positive and negative effects on numerical ratings, technical terms seem to primarily correlate negatively with patient scores for all doctors. The present work represents a first step towards understanding the measures of quality of care that relate with the patient experience, and in particular with respect to the growing field of concierge medicine. It is also the first attempt we are aware of that employs sentiment analysis in this context.

Cross-sectional analysis of online patient reviews of infertility care providers.

OBJECTIVE: To observe the effects of practice type, location, and mandated insurance coverage on infertility physician online reviews by patients. DESIGN: Retrospective cohort study. SETTING: Not applicable. PATIENTS: Patient online reviews of fertility specialists from 2016 to 2019. INTERVENTIONSS: None. MAIN OUTCOME MEASURES: The analysis consisted of the average rating out of 5 for each physician published on Vitals, RateMD, and Healthgrades. RESULTS: Data were collected on 1,097 specialists. Physicians practicing in states with versus without mandated insurance coverage received an average rating of 4.093 versus 4.076, respectively. The average rating was 3.964 for physicians affiliated with a university or hospital versus 4.128 for those working in a private practice. Significant differences were found in physician ratings from the four regions. It was revealed that physicians who practiced in the South (n = 354) received significantly higher mean average ratings than those in the Northeast (n = 327) and Midwest (n = 175). Physicians practicing in the West (n = 241) received significantly higher ratings than those in the Midwest (n = 175). CONCLUSIONS: The average online patient rating of infertility specialists was found to be significantly higher for physicians working in a private practice compared with those affiliated with a university or hospital system. No significant difference was found between the average rating in states with versus without mandated insurance coverage for infertility treatment. We propose that qualities other than patient financial responsibility are implicated in the factors used to rate physicians.

Content, Quality, and Assessment Tools of Physician-Rating Websites in 12 Countries: Quantitative Analysis.

BACKGROUND: Websites on which users can rate their physician are becoming increasingly popular, but little is known about the website quality, the information content, and the tools they offer users to assess physicians. This study assesses these aspects on physician-rating websites in German- and English-speaking countries. OBJECTIVE: The objective of this study was to collect information on websites with a physician rating or review tool in 12 countries in terms of metadata, website quality (transparency, privacy and freedom of speech of physicians and patients, check mechanisms for appropriateness and accuracy of reviews, and ease of page navigation), professional information about the physician, rating scales and tools, as well as traffic rank. METHODS: A systematic Web search based on a set of predefined keywords was conducted on Google, Bing, and Yahoo in August 2016. A final sample of 143 physician-rating websites was analyzed and coded for metadata, quality, information content, and the physician-rating tools. RESULTS: The majority of websites were registered in the United States (40/143) or Germany (25/143). The vast majority were commercially owned (120/143, 83.9%), and 69.9% (100/143) displayed some form of physician advertisement. Overall, information content (mean 9.95/25) as well as quality were low (mean 18.67/47). Websites registered in the United Kingdom obtained the highest quality scores (mean 26.50/47), followed by Australian websites (mean 21.50/47). In terms of rating tools, physician-rating websites were most frequently asking users to score overall performance, punctuality, or wait time in practice. CONCLUSIONS: This study evidences that websites that provide physician rating should improve and communicate their quality standards, especially in terms of physician and user protection, as well as transparency. In addition, given that quality standards on physician-rating websites are low overall, the development of transparent guidelines is required. Furthermore, attention should be paid to the financial goals that the majority of physician-rating websites, especially the ones that are commercially owned, pursue.