Across the globe, nurses are often on the front lines for health care delivery in places where physicians are scarce or nonexistent, such as in rural areas or urban slums [2]. For example, the constitution of Brazil guarantees free public healthcare to every citizen, however, many doctors and specialists are concentrated in the large coastal cities of the country, with populations in the rural interiors drastically underserved by health care professionals [3]. The Health Ministry of Brazil implemented the Programa Telessaúde Brasil Redes to combat this issue by providing health support, advice from specialists, and permanent health education via telemedicine. Telemedicine can connect any member of the health care team across the continuum of care to enhance outcomes. This connection can entail submission of asynchronous questions via online systems for expert consultation.
Rede NUTES is an extension of the Programa Telessaúde Brasil Redes based in the northeastern state of Pernambuco in Brazil. Rede NUTES offers telehealth services to its 80 municipalities by offering web seminars, telecare service, and second opinions on submitted questions via their online submission system for other health care professionals [7]. Data from the Rede NUTES submitted questions were analyzed and explored using big data methods to gain insights into how and for what Brazilian nurses utilized the system, as well as gain new understanding from the questions that nurses answered.
Methods:
Step 1 was to explore and analyze the data, questions submitted by nurses, nurse technicians, and the two categories combined were identified and isolated from the complete data set to understand ratios of how often nurses and nurse technicians were using the Rede NUTES system in comparison to their peers. Questions that were answered by nurses at Rede NUTES were also queried from the general data to review the frequency of responses given by this group of users.
In step 2, the questions were reviewed by calendar months. This visualized usage trends over the course of a year for questions both submitted and responded to. A histogram was created to visualize this trend.
Step 3 entailed determination of question themes per calendar each month. For example, a spike in gynecology questions dealing with pregnancy was typically followed by a corresponding increase in obstetrics questions nine months later. By visualizing question topics and clustering those topics in a timeline, it was possible to understand trends in the data for nurses. These question topic clusters were compared to the questions submitted and answered by the other medical professionals utilizing the online question system.
In the final step, the data were reviewed for lexical relationships using Linguistic Inquiry and Word Count (LIWC) program in Portuguese [6]. LIWC returns values on dimensions such as number of self-references, level of positivity, and number of cognitive words.
Results:
As stated in [2], the nurses and nurse technicians in Pernambuco are indeed on the frontlines of serving their communities. Nurses and nurse technicians submitted the most questions to Rede NUTES out of any of the health practitioner categories for two out of the three years reviewed (38% in 2010 and 32% in 2011), and in all three years combined (36%). Furthermore, in the majority of the months in the years reviewed, nurses and nurse technicians submitted the largest percent of the questions from any health care practitioner category, submitting 92% of all questions in December of 2012. The top question topics across submitted concerned cardiology, gynecology, obstetrics, pediatrics, and dentistry for all three years. In 2012 and 2013, the top question theme dealt with electrocardiograms, while
Another notable result was that nurses also answered the most questions during the three years reviewed. Questions that were responded to followed a similar thread as to those that were submitted, with specialized nurses responding to the majority of questions dealing with obstetrics, gynecology, and pediatrics, with the top question theme dealing with hypertension.
Conclusion:
Exploratory data analysis is the first step in using big data for future experiments [8]. It allows researchers to detect mistakes in the data that could skew machine learning results, while also checking intuitions about relationships in the data. Future work on this data set will explore relationships in the data for other medical fields.
While asynchronous online telemedicine applications to obtain expert counsel such as Rede NUTES offers positive features, likewise negative features must be acknowledged. There are legitimate causes for concern over privacy and protection of data. Likewise, as many people fear receiving a break-up text, so should bad news of a diagnosis be conveyed in person. Future research should work to migitate these concerns while enhancing positive outcomes.