Aligning BI skill, practice and healthcare outcomes

ICT skills development has long been lauded as a key input to business productivity, a vehicle for improving labour employability and indeed as the means to enhancing the competiveness of nations. But another ‘truism’ has also long held sway – that Canada suffers from a chronic shortage of skilled ICT professionals. A 2015 ICTC IT labour report sponsored by the Government of Canada’s Sectoral Initiatives Program, for example, has concluded that by 2019, there will be a skills gap of 182,000, with Ontario suffering the greatest shortages of ICT talent. But there are nuances in these numbers, both in terms of specific skills that are in need and the reasons for these shortages. According to the ICTC, A key contributor to demand-supply imbalances is the “skills mismatch,” where people do have IT training and expertise – just not in areas that are hiring.

Dire predictions for a shortfall in the number of data professionals – back in 2011 McKinsey predicted shortages of 140,000 to 190,000 people with deep analytical skills by 2018 in the US alone – has recently come under some scrutiny; however, new data drivers such as IoT, combined with cheap storage costs, argue for ongoing demand for data workers of multiple skill stripes to manage an anticipated explosion of information assets. In healthcare informatics, the Human Resources Outlook, 2014-2019 study prepared for Canada Health Infoway and others has identified business analytics and data analytics roles, as priority occupations where shortages “pose the greatest risk and should be the focus of proactive steps to address likely gaps in supply.” But what mechanisms can be put in place to ensure better alignment of skill demand and human resource?

One approach is to create public-private relationships aimed at equipping workers with the skills that are in high, current demand in the workplace. This is a tall order as it demands advance knowledge of talent needs, the ability to build programming that focuses on relevant technical education, strong support from the business community – and their combination in a way that advantages all stakeholders. But when it works, this type of collaboration can be impressive, as was demonstrated at the Health Information Science Day@Conestoga College, held in Kitchener, Ont. this April. Organized to celebrate the achievements of students in Conestoga’s health information sciences program, the event also reinforced the benefits of cooperation between educational institutions and business, in this case, the purveyors of BI, data integration and analytics solutions, Information Builders.

Launched last year, the Conestoga partnership aimed at developing a pool of talent with relevant healthcare analytics expertise actually consists of three collaborators: Guelph General Hospital, which built hospital datasets for use as a student testing ground and delivered instructional support in Conestoga courses, as well as internship opportunities for select students; Conestoga College, which has created curriculum that features problem solving through the use of analytics in real world healthcare scenarios, and Information Builders, which has provided its WebFOCUS BI software to the college, along with  technical and services support. According to Justin St-Maurice, a professor in Applied Health Information Sciences at Conestoga College, the team has made good progress on three fronts over the past year: it has improved operation of the student BI contest (which was judged at the event), improved teaching of the BI courses, and the setup of technology infrastructure at the school: “This year is the culmination of many lessons, as we learned how to remove the kinks from the program,” he noted, “and what’s going to be good next year is that year four, three and two students will all have WebFOCUS experience so we will be able to continue to build program and contest momentum.”

In addition, this year’s Health Information Science Day itself, offered greater opportunity to demonstrate what St-Maurice described as partnership “synergies” designed to create a “different way of doing business.” While Conestoga and Guelph were able to showcase student talent enrolled in BI courses and participants in a related BI competition, combining this with the school’s Student Awards ceremony, Information Builders was able to welcome current and potential clients, demonstrating the application of its software capabilities to interesting BI use cases, as well as new software features and functions that will be available in the upcoming 8.3 release of WebFOCUS.

Easy access to diagnostic data (Hanna Close, Xiomy Brenes, Corrie Fox)

But the best proof of this approach is in the putting, which four student groups managed admirably in contest presentations animated by the theme “Don’t get your NACRS in a twist.” Working with the CEDIS (Canadian Emergency Department Information Systems) data standard, which features information on diagnoses, and the NACRS (National Ambulatory Care Reporting System) database maintained by the Canadian Institute for Health Information, which contains day surgery, ER and outpatient clinic metadata, the first group focused on presenting complaints when patients enter hospital Triage, with the goal of speeding and tracking diagnoses in order to improve efficiencies and financials around this process. In this project, the students encountered challenges that are common to efforts to derive usable insight from information repositories – looking at patient level data, case mix groups, diagnoses and care processes, the group found that that across the databases, there were issues with standardization, the voluntary nature of reporting and time lags between reporting. In addition, information integration was stymied by formatting issues as NACRS data was not in in usable format for the CEDIS classification system. With WebFOCUS, students were able to manage these issues, and built a dashboard that mapped the top ten presenting complaints to the most common diagnosis, introducing categorization based on presenting complaint and a short description of diagnosis and treatment. Using the new WebFOCUS dashboard, emergency room workers would be able to drill down to link presenting data with diagnosis and the most common treatment, benefiting from a single portal for information that features expedited access via keyword search.

Building predictive capabilities (Victoria Phillips, Jessica Jasper)

The second group presentation, “Skye Analytics: Options as wide as the sky,” focused on developing predictive analytics capabilities to solve issues around finance, workflow (patient wait time) and staff allocation. Specifically, the students used WebFOCUS to analyze historical trends that would identify peaks in the number of patients admitted to hospital – the average number of patients on a monthly basis – to look at wait times for patients to cycle through different processes – on a daily, monthly and quarterly basis – and to develop financial analysis, based on an assessment of the cost associated with patients entering ER for specific problems. The student dashboard, featuring easy to use graphic reporting, also generated a CTAS (Canadian Triage & Acuity Scale) score for prioritizing patient care requirements, and to support better management of emergency room case mix and staff resources.

Managing process flow (Christy Smith, Heidi Evans, Hashim Garsi, Sydney Reeves)

A third group targeted wait times in emergency, compiling an impressive set of statistics that associated a lag in diagnosis with impact on patient outcomes. Designed to explore levels of acuity for patients who have shorter or longer emergency room wait times, the students’ “Go-Me Dashboard in the Emergency Department,” aimed at understanding, managing and improving patient flow in emergency rooms with the ultimately goal of shortening wait times to improve patience care. Based on assessment of CTAS levels for patients with specific diseases, such as COPD, the Go-Me Dashboard is a useful tool that can help hospitals better predict and plan for patients so that they can get where they need to be in hospital without undue delays.

Data-based decisions (Ilkay Sakir, Victoria Macdonald)

A final contestant in the BI competition considered the “Appropriate Use of Primary Care,” exploring the need for more efficient use of primary care services with the ultimate aim of reducing hospital readmissions. Examining rates of readmission within 30 days of a patient’s last admission for pains in the throat and chest from 2010-2014 based on Guelph Hospital outpatient data in the NACR database, the students in this group used IB analytics to correlate readmission with a patient’s relative access to a primary care physician. Interestingly, the group found that lack of access was not necessarily the cause of hospital visits, and was able to suggest, based on insights provided by data analysis, alternative reasons for patient behaviour. And more importantly, they were also able to recommend change management techniques, ranging from patient education on healthy lifestyle choices and available clinic resources, to fees for hospital visits when primary care was not chosen as a first option, to better patient to doctor ratios to improve wait times in primary care provider facilities as a means of decreasing stress load in emergency departments, doctors’ offices and of reducing costs for hospital stay. These conclusions offered a good example of the potential benefits of data-driven decision support.

A key thrust in Information Builders’ software development has been the democratization of business intelligence and analytics tools, and the student presentations and testimony attest to success achieved in putting information in the hands of individuals who can use it. Armed with varying levels of college instruction (students begin to learn WebFOCUS in year two of the Conestoga program), student groups were able, in a relatively short period of time, to ask highly sophisticated questions of healthcare data, and to provide answers with dashboard visualizations designed for easy access to information for front line healthcare workers. As Ilkay Sakir, second year student and winner with Victoria Macdonald of the BI contest for their Primary Care dashboard, explained, even though their particular project uniquely involved SQL writing and Java script to hard code the data, “the tool itself [WebFOCUS] is pretty easy to use. This is our first year using it.”

Even more striking in the presentations was the students’ ability to quickly recognize and articulate critical challenges in healthcare delivery and to build data-driven solutions around these, a competence born of the Conestoga approach which combines academic exploration of health information issues, use of real life datasets, and relevant software skills development. As fourth year student Corrie Fox noted, the Conestoga program offers strong support for all the components required in building solutions: “we talk about problem solving, and from the technical side, about how we can get creative with information systems and their role. But on the other side, we also look at the business challenges. The program offers good balance.” And as Victoria Macdonald explained, asking the right questions of data is a function of the interplay of technology and this business awareness: questions for her group emerged as the students worked with the data to test assumptions. “We knew the issue is that there is a high remit rate across the board for what seem to be non-emergency issues, so we started looking at change management, and thinking, how can we change this process to make it more efficient? This would be beneficial for everyone – for patients, for physicians, for hospitals and other healthcare facilities,” she added. Key to the kind of change needed in healthcare delivery, this skill, knowledge and outcome-based process is also the dream combination for potential employers looking to connect information assets with data value.