1. Salantera, Sanna PhD, RN

Article Content

Electronic documentation is gaining ground in both public and private health services globally. Many health organizations, health systems, and whole countries already have 100% of their health records in electronic form. Governments and healthcare services have taken action toward harmonizing documentation systems and providing standard vocabulary, language, and structures for documenting and storing information. Electronic healthcare systems are expensive, and a large portion of the health sector's money is invested in the systems. It is important that we continue to develop these systems to their full capacity in patient care and in research.


Nursing notes are a critical part of the electronic health records. These notes are based on continuous monitoring of a patient's condition and are usually composed partly of structured responses and partly of free text. Nursing notes reflect each patient's individual health, health problems, and care. They also reflect the reasoning processes of the nurse and actions taken based on the patient's health problems.


Research is necessary in many aspects of nursing documentation for us to be prepared for the big data analysis and active use of nursing information for the good and safety of the patients as well as for the decision support of nurses and nurse leaders. The newest push for data analytics of large sets challenges health professionals to gather, process, and analyze massive amounts of data in meaningful ways for improving patient care. Nursing researchers need to be involved in this now.


Data mining-that means analyzing data from several different angles and summarizing it to meaningful information-of structured data alone is not enough. Not all nursing can be turned into numbers and structured data. A patient's condition, emotions, and daily care can sometimes be best described with freely chosen words, and this should be allowed also in the future. However, our capacity to handle this nuanced rich freely written text and recognize patterns and trends in it is still limited. Text mining has great potential in detecting and understanding nursing observations concerning symptoms and emotions. With analyzing large amounts of free text nursing notes, we could, for example, see patterns in how certain types of symptoms are described, how patients react to different care interventions, and how effective the nursing interventions are.


Free text nursing notes are not structured and often contain inconsistent syntax and semantics, abbreviations, Latin, and minor mistakes (e.g. "Pleura drain l.dx. secretes abundantly, breating heavy, does not co-operate"). Usually these do not make the text difficult for a professional to understand but are challenging for the computer. Text mining research gives one option for analyzing nursing documentation and written text. Text mining is a process where the free text is semantically and lexically analyzed with methods of natural language processing (e.g., Hyun, Johnson, & Bakken, 2009). And notably, text mining research in the field of nursing will not evolve to its full potential if nurses do not get involved in its development. It is not easy for a linguist or an engineer to recognize the relevant research questions that help in solving nursing problems. Hence, nurse researchers should be leaders or part of the natural language processing research team. Transdisciplinary research teams are wanted. Completely new translational innovations that are not discipline-specific can be created in transdisciplinary research teams.


Big data analysis could compare nursing information with the newest research on the topic. By developing analysis systems, we could create alerts and predictors of recovery process into electronic health records. For instance, if a nurse writes about a patient's reactions to pain medication, this information is instantly compared with current research on side effects, and the program alerts the nurse that an adverse event might occur and gives a suggestion for relevant care. Perhaps we could even create automatic summaries of a patient's condition and care.


Obtaining large amounts of daily nursing notes for research purposes has some obstacles. Laws in many countries prohibit the use of medical and nursing documents for research where there are no clearly defined research questions and no clearly set sample. Anonymized data can be used in some cases, but even then, health texts are sensitive. The sensitivity of the data requires high-quality security systems and clear understanding of the nature of the data. Also the data itself has to be available for long periods of time, for validity and reliability purposes. The individual patient's rights have to be protected. Nurses are needed in designing big data research architectures, so that it is possible for our research questions to be answered.


Future research using nursing documentation combined with information about individual's genes, physiology, information from scientific literature and social media, as well as general information about the environment seems very exciting, and we should be part of this development to ensure that patients and professionals can benefit from the knowledge stored in nursing documentation systems.




Hyun S., Johnson S. B., Bakken S. ( 2009). Exploring the ability of natural language processing to extract data from nursing narratives. CIN:Computers, Informatics, Nursing, 27, 215-223. doi:10.1097/ NCN.0b013e3181a91b58 [Context Link]