Principal Findings
Our findings revealed that internet search trends in Baidu Index could comprehensively reflect the actual concerns of Chinese populations regarding health issues. Previous studies have confirmed that users' information-seeking behavior profiles from big data platforms are adeptly proficient in forecasting disease outbreaks, identifying interest at the population level, monitoring health education campaigns, and tracking the trends of patients' preferences [14,15,24,25]. As a primary health care information source, patients are more inclined to initiate a preliminary consult on the internet before seeing a doctor [34]. Hence, investigations of the infodemiology and infoveillance features of LUTS-related terms are capable of revealing a massive amount of users’ behavioral information and reflecting actual prevalence trends to some extent [24,25]. This feature is particularly crucial when there are a lack of related epidemiology data and real-time information.
By using Baidu Index, we found that the popularity of each entry varied greatly. The most inquired term, urinary frequency, was searched 25 times more often than the least inquired term, nocturnal enuresis. Intuitively, this result shows different degrees of attention for LUTS and therefore reflects the relative prevalence of symptoms in the population. With regard to trends, we illustrated that search volumes for each LUTS-related term had maintained a certain order of magnitude. A sustained growing trend was observed for the terms frequency, incomplete emptying, nocturia, and incontinence. The trends for other phrases were relatively flat and steady. These trends indicate that while the prevalence of LUTS may remain stable, public awareness and internet penetration may be on the rise, and people’s habits of seeking medical advice are changing. Further, 75.31% (50,183,464/66,635,247) of the search requests were conducted on mobile devices. This fact may have resulted from the development of portable technology and mobile internet. Some clinics use smartphones to track older adults with frailty and their quality of life [35]. Additionally, 43.5% of individuals have acknowledged that they use cellphones while on the toilet; therefore, it is no trouble for these users to conduct searches while “on the go,” especially when they are experiencing a LUTS attack [36].
We noticed a spike in search volume for the term incontinence, which surged to over 3400 from 1100 on May 6, 2013. However, just 15 days prior, an earthquake (7.0 on the Richter scale) struck Ya'an, Sichuan Province, resulting in 196 deaths, 21 missing individuals, and 11,470 injured individuals. We believe that the spike in search volume for incontinence is related to this event, as most incontinence symptoms would have appeared after the injuries from the earthquake [37]. Additionally, we noticed an abnormal fluctuation that was above the average search volume for the term split stream. The total increase in this term’s search volume originated from term 4 (尿尿分叉; ie, Wee Wee split). We suspect that this may have been caused by the bidding manipulation of specific keywords by some individuals or institutions because the popularity of other terms has not increased significantly [25]. However, due to the openness of the Baidu Index platform, this abnormal fluctuation can only affect the search popularity of a specific term for a certain period of time [23]. Nevertheless, it will not affect the popularity of other synonyms. Additionally, despite the fact that the competitive ranking promotion mechanism is well known in Baidu's product line, this mechanism is limited to ranking search results instead of guiding or changing users' search preferences by shifting their needs. Hence, this abnormal search volume phenomenon only occurs occasionally. The relatively consistent search trend for other terms suggests that the BSI can still be a good indicator of public attention toward specific LUTS if keywords are sufficiently included in the search.
After analyzing the regional differences in search volumes, we found that the search trends of each LUTS-related term were in line with those of other infodemiology research [24,25]. The search volume for each LUTS-related term was highest in the East China region, followed by those in North, South, and Central China. In Northwest China, search volumes were lowest. However, the trends revealed by the BSI contradict previous investigations on the prevalence of LUTS and BPH. In earlier investigations, the prevalence of LUTS was highest in the northwest and northeast regions and lowest in the southwest areas [12,38]. However, at the provincial level, the prevalence of LUTS in Guangdong, Shanghai, and Beijing was ranked highest compared to those of other provinces [38]. These three provinces are representative of the South, East, and North China regions. Further, the data sets of previous investigations pooled individuals aged over 40 years. As a result of this, the rankings from these investigations were not much different from ours. Moreover, medical infrastructures, internet access, and public health awareness are considered better in these higher income regions. Coupled with the fact that Baidu's data sources involve people of all ages from across the whole country, the geographic patterns in BSIs may reflect the socioeconomic and population rankings of regions in mainland China.
In Groutz et al’s [39] and Heylen et al’s [40] investigations, the prevalence of voiding difficulty in the female population was vastly underestimated. The prevalence of voiding difficulty significantly increases with age and the degree of pelvic organ prolapse [39,40]. Additionally, voiding symptoms such as weak streams, hesitancy, and strained urination are especially prevalent in females when other comorbidities are present [41]. Specifically, the pooled incidence rate of voiding difficulty is 21.3% and 15.5% in males and females, respectively [42]. Thus symptom differences based on sex are not significant. Further, when comparing LUTS prevalence rates, Apostolidis et al [43] recognized that they failed to investigate incidents of incontinence in both sexes. This was because certain symptoms, such as incontinence, are not listed in the IPSS or National Institutes of Health Chronic Prostatitis Symptom Index assessments, which are the most commonly applied assessments and checklists for men with LUTS and chronic pelvic pain syndrome symptoms [42]. The IPSS is universally used, even when evaluating the efficacy of prostate surgery—a known leading cause of male urinary incontinence [44,45]. The evaluations for male incontinence is listed in the “expanded” index for “prostate cancer patients” [46]. This is probably because of the belief that male incontinence has almost always resulted from prolonged bladder outlet obstruction [47]. Nevertheless, it should be noted that other causes, such as sphincter injury following prostatic surgery, polypharmacy, detrusor degeneration, and sustained urinary tract infection, have also been identified as common factors of male incontinence onset [48]. Therefore, prostate problems may be an essential factor of the systematic neglect of primary incontinence symptoms in the male population. Similarly, the diagnostic criteria for female bladder obstruction have been misestimated, and it is recommended that female bladder obstruction should be confirmed with urodynamic examinations [39,43]. Despite these issues, there are differences in the distribution of LUTS between men and women. It is worth pointing out that the real factor that determines whether patients seek medical treatment is not the characteristics of symptoms but the severity of symptoms [49]. Additionally, as internet users with web-based health issue inquiries are 60% more likely to seek and use health care services, the ratio of BSIs for reported symptoms could be used to remind practitioners to avoid ignoring “uncommon” symptoms in each gender population [50].
Baidu Index lists the top 10 related theme words for each search term. In this platform, the trends for correlated search terms are calculated and updated weekly. This system enables practitioners to gain insight into the most concerning problems from patients, confirm individuals’ main intentions, and examine the most exposed information to users. In our study, we found that the related theme words were not limited to the field of disease diagnosis, cause, and treatment, as previously described [25]. Instead, users had an extensive range of questions and concerns about LUTS (eg, the best hospital, treatment efficacies, and disease prognosis), and Baidu Index includes a lot of content that has nothing to do with health counseling. The differences in the distribution of the theme words varied vastly for each LUTS-related term. In terms of the relevance ratio, the term dysuria had the highest number of relevant term entries and search frequencies. However, although the term urinary incontinence was included in 85.67% (8134/9495) of the number of valid search entries, this term had the lowest search volume (1,726,303/283,932,352, 6.08%). Further, we noticed that the terms related to symptoms, etiology, and treatment were the terms that were the most inquired by Baidu platform users. With regard to the term incontinence, a large number of inquiries related to diagnosis mainly included the term incontinence, which is both a symptom description and a diagnostic. Hence it is clear that Baidu platform users are more inclined to make inquiries by describing their symptoms and seek treatment. Although the diagnosis and prognosis of LUTS are also of concern, the rate is relatively low.
We also noticed that there were numerous inquiries that were formatted as “what are the symptoms of...?” This form of question implies that Baidu platform users may have a definite awareness of the diagnoses related to their symptoms and that they have needed to confirm or exclude a particular diagnosis. This is a typical internet self-assessment pattern. Further, for the treatment category, we found that most users focused on inquiries such as “the most effective medication” or “the fastest efficacy.” These facts reveal that Baidu platform users’ views of LUTS and related diagnoses are one-sided. Their intention is to control LUTS as soon as possible instead of consulting a doctor to obtain a definite diagnosis and treatment regimen. It is worth pointing out that despite the fact that diseases indicated by LUTS are often not life-threatening, completing routine urine tests or ultrasonic examinations as part of standardized diagnosis and treatment procedures can at least reduce the likelihood of misdiagnosing malignancies [51,52]. Furthermore, some patients believe that examinations with negative results are conducted with the purpose of making a profit and are therefore meaningless and a waste of money [53]. These beliefs may worsen when these patients come across information that claims people can cure LUTS with free, “do-it-yourself” methods. Such information will strengthen their prejudice against hospitals and increase their likeliness of using “do-it-yourself” methods to heal their LUTS [54]. Patients’ limited understanding of health issues predisposes them to misdiagnoses, results in illness delays, and poses considerable health risks.
Limitations
Several limitations in this study should be addressed. One is that Baidu Index only analyzes search data from Baidu and does not analyze search data from social media platforms. Therefore, the data generated by these platforms could not be assessed. Further, the types of Baidu platform users could not be determined due to confidentiality (ie, users’ privacy protection). Consequently, the analysis of demographic data regarding information-seeking preferences and behaviors could only be based on age, gender, and regions. Information such as socioeconomic status, ethnicity, and educational background were not obtainable. Furthermore, LUTS and their associated diseases, such as prostatitis, BPH, and cystitis, are not covered by national surveillance systems and lack corresponding epidemiological data. Additionally, instead of real search frequencies, the BSI is just a weighted index derivative; we could only speculate the relative popularity of LUTS-related search terms and their development trends.
Notwithstanding these limitations, this is the first study that investigates public concerns about LUTS, to the best of our knowledge. We chose to examine the infodemiology characteristics of LUTS because the clinical manifestations of LUTS are obvious to patients. As such, LUTS are easy to compare when analyzing clinical diagnoses and diseases. Therefore, research on the search volumes of LUTS-related keywords can directly mirror existing problems from the patient perspective.
Finally, we believe that infodemiology research based on the BSI reveals people's behaviors when they search for related keywords and that such data can be used as an important reference for understanding the population’s needs. In terms of medical and health-related information, Baidu should seek government health management departments' help to revise and standardize medical-related keywords and consultations, so that their platform can better serve consumers, potential patients, health management departments, and medical practitioners. Since the real-time data on search volumes and information-seeking behaviors on the platform are renewed over time, these data can potentially be used to provide supplementary references and up-to-date information for improving care practice standards and making policies in a timelier manner.