Using data analytics to improve performance and patient care

the Inland Northwest Washington service area for Providence, a 52-hospital health system operating in five western states.

To find out more about how labs, like the ones at Providence, are compiling and using data to improve performance, Medical Laboratory Observer created a State of the Industry (SOI) survey on clinical data analytics. It is the first of four SOI surveys and associated articles that MLO plans to produce in 2022. (MLO has been producing a series of four SOI reports annually since it launched this survey research in 2020.)

In the 2022 survey, most survey respondents were in administration: 45% were lab managers, administrators, or supervisors; 12% were lab directors; and 6% were section heads or department supervisors. A small number were in other administrative roles, such as in QA/QC, information technology, or point of care.

They were distributed across a range of lab sizes. A total of 22% work in labs with 1-10 employees, 16% work in labs with 11-20 employees, 16% work in labs with 21-50 employees, 21% work in labs with 51-100 employees, and 26% work in labs with more than 100 employees.

Most survey respondents (71%) work in hospital labs, while physician’s office labs and independent labs each accounted for 11% of participants, and group practice labs accounted for 4%.

The LIS

The laboratory information system (LIS) is the hub of labs’ electronic processes, and it is the starting point for collecting and storing data used in analytics.

While an on-premises software deployment is still the most common architecture for an LIS, cloud-based systems are slowly becoming more common at labs, according to the results of MLO’s SOI survey. The percentage of survey participants working at labs using a cloud-based LIS increased to 24% in 2022, up from 17% in 2021. The percentage of labs with an on-premises infrastructure was 76% in the 2022 SOI survey, dropping from 83% in 2021.

In 2022, most labs (62%) used LIS software that is part of an electronic health record (EMR), compared with 38% with a stand-alone LIS. The breakdown was virtually the same in 2021: 61% with an enterprise-wide system and 39% with a stand-alone LIS.

Labs have automated a variety of clinical processes using LIS functionality. For example, nearly all (95%) use the LIS for electronic orders and results, compared with 92% in 2021. The majority of survey participants in 2022 also have integrated their analyzers with the LIS (72%) and rely on the LIS for QA/QC processes (70%). Automated billing and revenue cycle processes also were common in 2022 (60%), compared with 58% in 2021.

Other functions were not as likely to be automated through the LIS. Slightly more than one third (39%) of the SOI respondents in 2022 use a module for point-of-care testing, 29% use it for regulatory compliance and reporting, 26% for scheduling, 27% for customer service, and 15% for inventory control/supply chain management.

David Nichols, MBA, President and Founder of the Nichols Management Group, a consulting firm, said that automating some of these processes is not a top priority for many lab managers.

For example, Nichols said he thinks that labs do not place enough emphasis on nurturing customer relationships, so it is not surprising that few have automated those processes or developed standardized performance measures. “That’s kind of a no brainer in most industries, and we, labs, are significantly behind in that regard,” he said.

Kim Futrell, MT(ASCP), MSHI, Senior Strategic Marketing Manager, Orchard Software, said she expects more labs to adopt the functionality in their LIS for regulatory reporting and compliance. “In my experience both at Orchard and as a laboratory manager, there are a multitude of features in the LIS that support regulatory compliance reporting, so if this functionality is not currently being used by the majority, I would expect it to be the largest area of growth because much of the functionality is already within the top LISs. Features like data mining, QC tracking, audit trails, and rules-based decision support can automate a great deal of regulatory requirements.

“In addition, COVID-19 has brought some urgent reporting needs that the LIS can support from a regulatory stance. COVID-19 is a required state reportable, and manual reporting on a massive scale with severe staffing shortages results in delayed and inaccurate statistics,” Futrell said.

Labs also might be more likely to automate compliance activities in response to federal regulations governing the Promoting Interoperability program, she added. The Centers for Medicare and Medicaid Services (CMS) now places more emphasis in that program on public health reporting than was the case in the past, Futrell added.

Top strategic priorities

When survey participants were asked about their top strategic IT priority for the next three years, the most common answer was infrastructure and platform development — the option selected by 30% in 2022 and 31% in 2021. A total of 26% chose data analytics optimization to support lab management as the top priority in the 2022 survey, compared with 25% in 2021; while 22% chose a new LIS in 2022, compared with 13% in 2021.

Revenue cycle management optimization became a less popular strategic priority, declining to 9% in 2022, compared with 17% in 2021. Very few chose a point-of-care testing product as their top priority in 2022 (3%) or 2021 (6%).

Tapping into data analytics

No matter what their strategic priority is, most labs already are using data analytics to monitor and manage lab operations. In the 2022 SOI survey, 42% of survey respondents said they are “utilizing data analytics for some aspects, and we’re planning more,” while 17% said they are “utilizing data analytics for all aspects of lab management,” and 8% said they are “utilizing data analytics for some aspects, and we are not planning more.”

On the other hand, 16% of respondents in 2022 said they “are not using data analytics for lab management, and we have no plans to start in the near future,” and 16% said they “have not used data analytics yet in any significant way, but we want to start.”

To improve the uptake of data analytics, organizations should push standardized reports or electronic dashboards out to managers, rather than expecting them to build custom reports, Nichols recommends. “It has to be easy and automatic. Managers aren’t going to spend their time querying a database,” he said.

Tim Bickley, VP of Sales for Visiun, said, “Laboratory analytics shoud provide the lab’s leadership with the ability to run a report on demand or automate reports with a scheduler. Not only should the automated reports be set up on a scheduler, laboratories should be able to set outlier criteria, so they are made aware of when they are not meeting a key performance metric.”

Some labs are monitoring performance, but they aren’t using automated tools for data analytics. One example is 26-bed Anson (TX) General Hospital. Laboratory Director William Lee monitors turnaround time and test utilization through the LIS. He tracks cost per test using a spreadsheet in which he figures in the cost of reagents, labor, and non-chargeable items such as test tubes. He updates the spreadsheet every two weeks.

Frequency of refreshing data

Another way to improve the uptake of analytics is to provide current data. Labs refresh their data for analytics with varying amounts of frequency, but, overall, the industry appears to be moving in the direction of real-time data. According to the results of the MLO SOI survey, 41% were using real-time data in 2022, compared with 29% in 2021.

For those not using real-time data, the frequency varied. In the 2022 survey, 24% said they refreshed their data weekly, compared to 33% in 2021; 24% refreshed data daily in 2022, compared to 28% in 2021; 4% refreshed data hourly in 2022, compared with 2% in 2021; and 7% refreshed data in minutes in 2022, compared with 8% in 2021.

To slice and dice the data, many labs rely on electronic tools. In 2022, 65% of respondents used a tool that is integrated with the LIS and 35% used a separate tool. The breakdown was identical in the 2021 survey.

Key performance indicators

Labs also are using data analytics to track a variety of key performance indicators (KPIs). The most common KPI, not surprisingly, is turnaround time (TAT), which 90% of survey respondents in 2022 said they track, up from 80% in the 2021 survey.

Other KPIs that survey respondents said they track include quality improvement initiatives (67% in 2022 and 74% in 2021), cost per test (58% in 2022 and 63% in 2021), billable tests versus performed tests (48% in 2022 and 53% in 2021), staff productivity goals (47% in 2022 and 42% in 2021), medical necessity (34% in 2022 and 30% in 2021), and unnecessary tests (22% in 2022 and 30% in 2021).

As Remillard explains, “Our quality metrics and productivity metrics provide an excellent foundation to know if our current staffing and test performance goals are meeting our patients’ needs.”

One example is appropriate test utilization. Providence actively monitors what tests providers request for hospital patients to make sure the tests are related to inpatient care and that the results will be ready before the patient is discharged. For example, genetic testing should only be requested for newborns. Otherwise, the tests should be ordered on an outpatient basis after the patient is discharged, he said.

These guidelines help ensure that payers will approve the charges and pay the health system for the tests, Remillard said.

Providence’s labs in inland Washington also track utilization of blood products for transfusions. The goals are to ensure that transfusions occur only when medically appropriate and to prevent blood products from being wasted. For example, Providence tracks how many units of blood are ordered for each procedure to discourage providers from ordering more units of blood than necessary.

 The labs also match up orders for blood products with patients’ medical information, such as hemoglobin or platelet levels, to help ensure that providers order transfusions appropriately.

In addition to testing-process metrics, Nichols said labs should use data analytics to manage the supply chain, particularly related to utilization and spoilage. “Supply costs are going up in the industry really fast,” he said.

Forecasting The Future

Of all the resources that go into diagnostic testing, lab managers have the most difficulty forecasting staffing needs.

At least that is the case in the short term, or the next three years, survey participants said.

In both the 2022 and 2021 MLO SOI surveys, majorities of respondents ranked staffing as the most challenging need to forecast. On a scale of 1-5, with 5 being the “least challenging,” 76% of respondents in 2022 ranked staffing with a 1, or “most challenging.” In 2021, 56% ranked staffing with a 1.

These findings are not surprising, given fluctuating testing demands in molecular testing and increasing demand for testing from an aging population. Those challenges are compounded by the lab industry’s ongoing difficulty finding enough employees to fill all shifts.

In its 2020 survey report on staffing, the American Society for Clinical Pathology found that the national vacancy rate was highest, at 12.7%, for chemistry/toxicology and lowest, at 3.9%, for cytology. Other departments with relatively high vacancy rates in the study were immunology (11.2%), phlebotomy (11.1%), blood bank (10.4%), core lab (10.3%), LIS/QA/PI (10.3%), and flow cytometry (10.1%).¹

To forecast staffing needs, Maggie Morrissey, Director of Recruiting at Lighthouse Lab Services, suggests labs measure how many samples a day are run by their techs, so they can see, on average, how many samples one tech can run. “With that information, they should be able to calculate how many techs to hire based on the expected volume of samples,” she said.

Nichols said that tracking overtime is important because this information can help spotlight problems to tackle, such as inadequate compensation or a dirty and crowded space. “In general, significant overtime is a management issue,” he said.

Forecasting resource needs other than staffing is less challenging, survey participants said. For example, only 26% ranked funding challenges with a “1,” or most challenging, on a scale of 1 to 5, and only 7% assigned “1” to forecasting technology needs. A minority of participants chose “1” for forecasting training (13%) or ROI/costs (13%).

Data analytics and COVID-19

COVID-19 also has been an area of intense focus for tracking and analyzing metrics.

Remillard said, “Use of data analytics has been crucial in dealing with the COVID pandemic. We are closely monitoring positivity rates and testing workloads to ensure adequate inventories throughout our system. We have a weekly system wide laboratory command center where we discuss testing needs and strategies in all 5 states where Providence has hospital ministries.”

Another example is the trio of interactive infectious disease dashboards that TriCore Reference Laboratories makes available to providers and the public in New Mexico, based on the company’s diagnostic testing activity. The first dashboard provides a week-at-a-glance of 13 of the most common respiratory infections and the current week’s positive count. The second dashboard is a deeper dive into respiratory disease trends and the third dashboard is a heat map display.2

MLO’s survey participants also are monitoring their pandemic response. When asked what measures they are tracking related to COVID-19 testing, 42% said positive and negative test results, 23% said number of tests performed, 21% said turnaround time, and 8% said supplies used/inventory management. Type of test performed and cost per test were each tracked by 3% of survey respondents.

Nichols said he expects COVID-19 testing volume to decline by 75% in 2022, if no new variants emerge, making inventory management particularly crucial. “I would urge people to monitor their inventories and look at what kind of expiration dates there are for various kits in their labs.” He also says that if labs choose to carry excess inventory because of uncertain demand, they should only do that for one analyzer platform.

Linda Wilson

Senior Editor

[email protected]

References

1. Garcia E, Kundu I, Kelly M, Soles R. The American Society for Clinical Pathology 2020 Vacancy Survey of Medical Laboratories in the United States. Am J Clin Pathol. 2021 Dec 1;aqab197. doi: 10.1093/ajcp/aqab197.
2. TriCore Reference Laboratories unveils its updated New Mexico Infectious Disease Dashboards. TriCore Reference Laboratories. https://www.tricore.org/tricore-reference-laboratories-unveils-its-updated-new-mexico-infectious-disease-dashboards/. Posted April 6, 2021. Accessed January 27, 2022.