While searching you’re likely thinking about all the time you’re wasting and all the other things you could and should be doing.
If you are a clinician, scientist or other type of knowledge worker, you likely spend multiple hours every week searching for the latest information on evidence-based therapies, on drug side effects or on other critical details that help you perform your job better. In fact, the market research firm IDC estimates that knowledge workers spend almost nine hours each week searching for information. Even when using health- or pharma-specific search tools, clinicians and scientists usually must wade through mountains of documents – many of which are not highly relevant – to find the precise insights needed for quality decision-making.
The proliferation of data
The proliferation of data in recent decades has impacted search efforts. The amount of available life science and healthcare information has grown dramatically, paralleling the widespread adoption of EHRs, the introduction of sophisticated digital technologies, and the growth of online journals. Both life sciences and healthcare sectors encourage the sharing of scientific knowledge and collaboration among researchers, and every year, millions of new documents are published in the form of academic research, patent applications, clinical trial findings, and more. The sheer volume of available data can be overwhelming, even for doctors and scientists with very narrow fields of study.
Physicians and scientists often struggle to weed through the vast amount of available data to find the answers they need. Adding to the challenge is the fact that as much as 80 percent of information is stored in unstructured formats that are difficult to search and analyze using traditional manual methods. In order to effectively and efficiently take advantage of the wealth of available data, knowledge workers must leverage new technologies.
Consider, for example, a researcher wanting to compare drugs for hypertension. A search might produce long list of drugs – yet fail to categorize which drugs are for the treatment of hypertension and which are known to cause hypertension. To correctly classify the drugs, the user may have to perform individual searches of each of the identified drugs – a very inefficient use of time.
Medical literature searches inform clinical decision making at Georgetown University Medical Center
Similarly, a physician might want to search for a possible linkage between a patient’s multiple diseases in order to identify the most appropriate therapies. At Georgetown University Medical Center (GUMC) in Washington, D.C., for example, Clinical Informaticist Jonathan Hartmann is regularly asked to search medical databases to help physicians make clinical decisions.
“Once when I was helping a physician in our Pediatric Intensive Care Unit (PICU), the doctor needed information about a child who had three different co-occurring conditions and wanted to know about the relationship between those diseases, which could impact the treatment given,” explained Hartmann.
If Hartmann had conducted a standard search of the three conditions, it’s unlikely he would have found any relevant articles revealing a known causal relationship because most search engines lack comprehensive ontologies for key healthcare and life science concepts. Without a relevant ontology, it’s difficult to effectively match significant relationships between the concepts. Rather than producing a clear answer that facilitates treatment, search results would likely include a long list of hyperlinks and few details about what’s contained in each link. To find germane articles, a researcher would have to open and read each document behind every individual link, requiring many hours – and even then, a clear answer may not be obvious.
NLP enhances traditional search options
Retail, banking, travel and other sectors are leveraging AI technologies such as natural language processing (NLP) to help users more-easily search unstructured text sources to find high value, relevant results. At GUMC, Hartmann utilizes NLP tools to wade through huge amounts of text quickly to identify the information doctors need. He notes that NLP, “allows us to get to that information that we otherwise wouldn’t be able to at all or in a timely manner.” For the PICU patient, Hartmann used NLP tools to quickly identify two relevant journal articles that helped the physician to make an immediate treatment decision.
Despite the proven benefits of AI-based search tools, healthcare and the life sciences have generally been slow to adopt such solutions, in part because NLP searches typically require the expertise of technical users to build the queries and extract data insights. NLP-based queries can yield excellent results, but most organizations have too few technical experts on staff to quickly address all their users’ search needs. End-users often resort to searching on their own utilizing standard search engines – which again, generally lack domain-specific ontologies and the required matching tools to easily identify causal relationships.
Enabling better searches with self-service tools
Organizations that equip users with better search tools have the opportunity to empower physicians and researchers to effectively and efficiently perform their own searches. Such technologies must provide context around concepts, and not only key words. For example, while searching for details on the side effects of breast cancer treatments, a user should not have to type all the words that could be substituted for side-effects (complications, adverse events, consequences, etc.) nor input multiple terms for breast cancer (ductal carcinoma, intraductal carcinoma, ductal carcinoma in situ, etc.) Instead, the search technology should be sophisticated enough to match similar concepts based on the context.
Making search tools more context-specific is essential for organizations that want to empower end-users to conduct their own searches. Standard search engines rarely produce results that are precise enough to meet the needs of healthcare and life science users. If a search engine does include healthcare and life ontologies, it’s likely to lack user-friendly interfaces, meaning users need assistance from technical experts. To enable better searches that end-users can perform on their own, organizations need solutions that include intuitive interfaces and enable context-specific searches that can produce deep insights from a single search.
The inefficiencies of today’s search processes limit the ability of healthcare providers and life science researches to quickly identify appropriate therapies or expedite the time-to-market for new drugs. They waste too many hours a week searching, rather than performing the jobs they were trained to do. By empowering end-users with easy-to-use self-service AI-powered search tools, physicians and researchers can spend less time searching and more time developing effective therapeutics and enabling better patient care.
Simon Beaulah is Linguamatics’ senior director of healthcare and is responsible for the company’s healthcare products and solutions, including applications for clinical risk models, population health and medical research. Jane Reed is Linguamatics’ head of life science strategy and responsible for developing the strategic vision for Linguamatics’ growing product portfolio and business development in the life science market.
