HealthcareAnalytics.info

Is it a Metric or a Key Performance Indicator (KPI)?

Sarianne Gruber — Wed, 15 Feb 2012 18:23:51 +0000

[Editor's Note: This guest post is written by Sarianne Gruber of SG Analytics LLC. We are happy to welcome Sarianne as a contributor to HealthcareAnalytics.info!]

Beyond the numbers – measuring the right things

Indentifying appropriate Key Performance Indicators (KPIs) in healthcare is not a trivial task. And when it relates to improving healthcare quality and operational performance, few tasks are as important.

KPIs are often confused with general metrics just because you can measure them – although there is a very important distinction. I like to consider metrics as what information you use to manage a particular service or as a measure that represents the workings of your organization in relation to one or more components.

A Key Performance Indicator (KPI) goes beyond this, however, by tying into the goals of an organization. In other words, a KPI should:

be quantifiable
reflect critical success factors
show progress towards (or, away from) organizational goals, and
evaluate performance over a designated time period.

What distinguishes a KPI from a “standard” metric is that a KPI is tied to a definable and agreed upon target. Components for a KPI can include volume, cost, timeliness, quality, or satisfaction. A KPI will show how far a metric is above or below a pre-determined target. KPIs are usually shown as a ratio of actual to target.

A KPI example

Consider a Healthcare Organization (HCO) that chooses to monitor the number of hospital readmissions; this is too weak to be considered a KPI as there is no sense of magnitude nor duration. Further defining the indicator as the rate of unplanned hospital readmissions within 30 days for a specific chronic condition provides further clarity. Using both variables: (1) number of hospital readmissions and (2) a 30 day time period, provides the necessary context and results in a much better performance monitoring tool.

Introducing KPIs to an organization

All key stakeholders in the performance of an HCO should be identified and asked to participate in the identification of KPIs. When introducing KPIs, it is important for stakeholders to understand:

A KPI is quantifiable metric tied to an organizational goal with a set target.
KPI are used to determine whether the practice, hospital or Accountable Care Organization is meeting predefined targets, and to compare performance results over a specified time period.
Appropriate benchmarks are necessary to determine how an HCOs performance compares against it’s own desired goals and objectives, and against other (perhaps competing) HCOs.

Fundamentally, a KPI should be a recognizable and inherently understood by all involved. In addition, KPIs must be collected, recorded, measured, and interpreted in the same way by all those involved. The KPI serves to provide insight to where change may be needed or where areas are excelling. Technology allows KPIs to be viewed on dashboards as well as provides quick updates to the decision makers. Results can be reported in monthly, weekly or even real time.

Summary

Appropriate and well-defined Key Performance Indicators (KPIs) are a necessary component of an organization’s quest to improve quality and operational performance. KPIs provide monitoring and insight into the critical components of process and policy that need to be improved to achieve the overall goals of healthcare improvement.

About the author

Sarianne Gruber started SG Analytics LLC this year specializing in Healthcare Analytics and Marketing. SG Analytics LLC focuses on healthcare-related outcomes research for tracking, reporting and decision making. She has worked for top tier market research firms, medical institutions and pharmaceutical companies.

She holds a M.S. in Management Science from New York University, a M.P.H. in Biostatistics from Yale University and a B.A. in Bio-Computing from Goucher College.

Integrating Analytics and Healthcare Quality Improvement Initiatives

Trevor Strome — Mon, 13 Feb 2012 16:17:37 +0000

Analytics alone isn’t the answer to improving healthcare

In my recent Health IT Exchange article, I make the assertion that analytics alone cannot improve healthcare. Dashboards, scorecards, reports, and other information tools by themselves cannot change people’s practices. In order for real change to occur, it is necessary that analytic and Business Intelligence (BI) tools are applied in a strategic combination with quality improvement initiatives.

As I mention in the article, Lean and Six Sigma are two of the more common quality improvement approaches in health care, and are often used in concert with each other. Lean is a systematic approach to improving quality and value to the patient by eliminating the waste that is present in almost every health care process. Six Sigma goes a step further, and is a rigorous data-driven approach to reduce process variation and eliminate defects in the delivery of health care.

Quality improvement methodologies provide context for appropriate decision-making

A quality improvement methodology or framework helps provide the necessary context for analytics to help drive appropriate decision making. Quality improvement methodologies identify the most pressing quality issues based on inputs such as patient safety and quality needs (which are the most important factors), the overall goals of the HCO, national standards and legislative requirements.

In times when health care is under quality, financial, and legislative pressure, it is important to provide management and quality improvement teams with the key pieces of information they need to focus on the most important problems, and to make appropriate, timely decisions. The critical indicators derived from approaches such as Lean and Six Sigma can guide decision making on the part of the HCO and result in real health care improvement.

Further reading

Read my Health IT Exchange article for more on how better integration of analytics and quality improvement methodologies can lead to better metrics, better decision-making, and better healthcare.

Creating Great Healthcare Analytics Teams

Trevor Strome — Fri, 03 Feb 2012 19:06:01 +0000

My friend and colleague Scott Wanless recently posted an article entitled “Create a Great Healthcare Analytics Team.” In the article he breaks down the composition of particularly well-functioning healthcare analytics teams he has worked with, and describes what he feels has made them successful. One of the features I liked about Scott’s article is that he describes the team members as individuals, not as roles or resources, and makes the point that it’s the people that make a team successful.

Some of the success factors listed by Scott in his article include:

“Decision support is part of the organization’s quality and information services group,… [and] decision support forms part of the organization’s system of measurement.”
The team has strong member from diverse backgrounds so that “taken together, these people know, or know somebody who knows, the answer to virtually any question the group is asked.”
A strong leader whose job it is to “champion and sponsor the group’s activities, including promoting it, funding it, protecting it and funnelling the never-ending stream of requests appropriately.”

Of course,even well functioning teams are immune from problems. Scott outlines three major risks to well-functioning teams:

Arrogance – which forms as a result of building a world-class analytics system, but may lead to self-imposed isolation by viewing non-technical people as somehow inferior. This can be addressed by including team members on non-technical projects and creating a “a resource for the entire organization instead of an insular, selfish group.”
Infighting – which may occur if the team becomes bored and turns to internal politicking. Scott recommends “maintaining focus on the mission of the group, which is to serve the organization as its system of measurement.” I have found that trying out new tools, approaches, and “playing/innovating” helps fight boredom, and keeps the team fresh.
Burnout – which occurs when faced with a constant barrage of information and development requests, will reduce the overall productivity of the team. Scott recommends finding outets for the team to ensure consistent productivity.

Scott’s article is a must-read for any healthcare manager working with analytics development teams. His insights are particilarly relevant as analytics teams are coming of age in healthcare, and as the need for analytics by a healthcare system under duress only continue to grow.

Five Project Management Strategies To Successfully Leverage Analytics for Healthcare Improvement

Trevor Strome — Fri, 27 Jan 2012 16:35:16 +0000

Implementing business intelligence and analytics projects within a Healthcare Organization (HCO) are challenging due to regulatory and financial pressures facing healthcare, the complexity of data involved, the multitude of stakeholders, and relative newness of BI technologies to healthcare. Below are some strategies that can help your HCO implement BI and analytics projects that become integral to healthcare improvement and innovation efforts.

1. Get all stakeholders involved early and frequently. The IT specialists, program analysts, and clinical subject mater experts (SMEs) may be worlds apart in terms of aptitude, skills, and knowledge, but must work closely to ensure a successful BI implementation. Ensure key stakeholders (such as program leadership, project sponsors, clinicians, analysts, and technical specialists) are engaged early and communicated-with well throughout the project.

2. Understand the needs of the users. The variety of ways in which information is utilized within an HCO can make it challenging to meet everyone’s needs. Develop a data utilization strategy in collaboration with end-users that outlines what information is required, how often it is required, what format it’s required in, and how they will access the information.

3. Focus on comprehensive documentation. Not having clear, concise, and comprehensive documentation of available data will hinder ongoing maintenance and development of the analytics tools. Mappings to source systems and associated workflows provide the necessary context to the information. Furthermore, as workflows and source systems change, they need to be documented and communicated to all stakeholders. This will ensure a consistent understanding and interpretation of information derived from the analytics tools.

4. Keep the data relevant to the end-users. Strategic or executive dashboards are undoubtedly the “sexy” BI application. Healthcare professionals involved in quality improvement projects, however, will require detailed data based on metrics relevant to the processes being changed and evaluated. Many current process improvement methodologies used in Healthcare (such as Lean and Six Sigma) rely on frequent performance audits and detailed performance data to coach staff on new workflows and processes. The sooner issues can be spotted and brought to the attention of process owners, the easier it is to provide coaching and to manage overall process performance. Likewise, once a process has stabilized, quality teams will move on to other issues, and the analytics need to be flexible to support these changes in focus.

5. Set realistic expectations. Expectations naturally run very high after a BI implementation. It is important that all end-users understand that the analytics and BI tools will not result in an immediate resolution to all their information and/or quality improvement needs. Keep end-users informed as to the capabilities of the tools, including: what reports, dashboards, and other information tools are in-scope (and out-of-scope), what data is available for ad-hoc reporting, and what expected turn-around times are for new development and ad-hoc data requests.

Summary
Implementation of healthcare BI projects can be complex. These strategies will help set up a project for success, and more importantly, help healthcare organizations realize the quality and process improvement benefits that well-implemented BI systems can support.

Putting Analytics to Work for your Healthcare Improvement Initiatives

Trevor Strome — Tue, 17 Jan 2012 22:35:31 +0000

The number of healthcare analytics options are increasing.

Some HCOs have the benefit of a large Business Intelligence (BI) or analytics platform. These stand-alone BI systems are somewhat general-purpose in nature, and operate in many industries other than healthcare. Many of these platforms are powerful indeed, with many analytics functions and features available.

In addition to the traditional BI platforms, there are some special-purpose healthcare analytics tools; these boutique tools tend to combine standard BI and analytics functionality with healthcare-specific context built-in.

Finally, there is a growing trend for Healthcare Information Technology (HIT) vendors to offer a set of analytics tools as part of, or in addition to, their main-stream” clinical systems such as electronic patient record (EPR) systems.

Success factors for healthcare improvement analytics

Despite what vender literature might claim, the ability to do meaningful analytics work that can truly impact healthcare improvement does not necessarily depend on having extensive (and expensive) analytic and business intelligence platforms.The critical factors for successfully applying analytics to healthcare improvement projects are:

well defined and relevant performance indicators,
timely performance data,
appropriate analysis, and
dissemination of results to pertinent stakeholders.

Note that these success factors do not depend on having large BI/analytics platforms available. A common phrase in Lean Transformation initiatives is “creativity over capital”. Healthcare improvement initiatives with the benefit of extensive technology support (i.e., big BI, flashy dashboards, etc) can still fail whereas initiatives with more “modest” but innovative analytics approaches and tools can be likely to succeed. The difference, of course, is not which analytic tools where utilized (and how expensive they were), but how they were utilized, how well analytics were integrated into the initiative, and whether the results were acted upon.

Note: The above article is a sample from our special report, Overcome Common Barriers to Effective Implementation of Healthcare Analytics, available for download in the resources section.

Ensuring health care analytics create value by understanding the requirements of information end-users

Trevor Strome — Thu, 12 Jan 2012 19:18:06 +0000

There is a proliferation of “real-time” dashboards and other information displays throughout healthcare organizations. In a recent article on information overload, I mentioned that information on reports, dashboards, and analytic tools that is not up-to date, relevant, and accurate will merely contribute to information overload. This in turn decreases the value, and use, of analytics within an organization, which in the long term hurts performance and quality improvement efforts.To help ensure that end-users are getting the most value out of analytics and business intelligence, below are some questions that should be asked when developing and deploying analytics. These questions will help analytics developers understand what the end-user requirements are, what the limitations of the source data are, and how best to bridge any gaps between information needs and availability.

Who is the information end-user? Is the dashboard, report, or other analytics tool being prepared for a quality improvement team, department manager, or chief executive? Is the user likely to have a statistical background, or not?
What is the information being used for? Is the information being used for real-time or predictive purposes, or is it being used to analyze past performance? Does the user need to focus on individuals (or small groups of individuals), or will the information be used in aggregate? Will the information be used for strategy development and planning, or evaluating outcomes of projects?
How will the information be accessed? Do the users need the information in real-time dashboards, or printed reports? Will the information be presented “statically”, or will users be able to browse through the results?
What source data is available? What system(s) does the analytics draw information from? Is the information required by end-users even available? How often is the source system updated? How is the source data accessed?
How accurate is the source data? What data quality issues exist with the source data? Can data quality be overcome through the use of statistical or other analytic approaches? Do the analytics end-users understand any limitations of the source data?

Quality and performance improvement initiatives can benefit greatly from information that is easily accessed,but only if it’s relevant, accurate and up to date. These questions are certainly no replacement for a thorough requirements analysis. But these questions will help keep in mind that the value of analytics, ultimately, are determined by how well they meet the requirements of the end-users in improving quality and performance within healthcare organizations.

The Case for Self-Serve Healthcare Business Intelligence and Analytics

Trevor Strome — Sun, 27 Nov 2011 23:03:58 +0000

The need for convenient access to information is growing.

The use of data within Healthcare Organizations (HCOs) is growing at an astronomical rate. Efforts to guide decision-making and quality improvement initiatives with the best evidence possible mandate convenient access to up-to-date and accurate healthcare performance data. For example, Dashboards ranging from advanced “flight-board” schedule rosters to real-time Emergency Department wait time displays are now nearly ubiquitous throughout HCOs. Management and researchers alike are adopting advanced resource planning approaches that use simulation and predictive modeling to determine which patients will likely be requiring care, how long they’ll stay once they arrive at the HCO, and what resources they’ll need to expedite their discharge.

The analyst as “gatekeeper” actually hinders access to information.

Due to the specialist knowledge and skills of healthcare business intelligence professionals (which often include both technical and domain expertise), they are often in a position to be the “gatekeepers” of the data. Because most managers and others within an HCO who need information are often too busy (or disinclined) to obtain it by themselves, however, analysts become the de-facto “go-to” people for healthcare operations data.

This “gatekeeper” function analytics professionals often find themselves in, however, actually is counterproductive for both the organization as a whole and the analyst as an individual. Experience at top-performing organizations demonstrate that it is most efficient when executives, managers, and others who need information for decision-making and planning are able to access it for themselves. This “self-serve” access best occurs through well-developed analytics portals and report repositories that are easy to access, use, and navigate.

Healthcare leaders and administrators who need direct access to information are most likely looking for a quick answer to an operational question (such as, “how many Emergency Department visits did we have this weekend”). With the ever-increasing number of data requests, it is easy for analyst “gatekeepers” to become overwhelmed with daily requests for “routine” operational data, if preparing these manually. It is these requests that are better left to “self-serve” systems.

When being “busy” is not being “productive”.

Not only does going through an analyst for routine information delay getting information into the hands of the people who need it, there is very little intrinsic value (in terms of professional development) or challenge for the analyst in handling such requests. When the routine operational data requests are handled through a self-serve analytics portal, the analytics professional will be freed for more true analytics work, including more time for innovation, research, and collaboration activities. For example, more time will be freed up to improve existing analytics tools, develop new predictive models, and be more responsive to requests for information that actually require the skills of an analytics professional.

Healthcare organizations should implement self-serve business intelligence and analytics portals that allow routine operations information to be accessed directly by leaders and administrators. This will enable analytic professionals to look more deeply into the data and apply the tools to obtain more insight and improve return on investment in healthcare data.

Healthcare Analytics – A Strategic Resource for Healthcare Quality Improvement

Trevor Strome — Tue, 01 Nov 2011 22:53:29 +0000

Analytics is changing how health care organizations (HCOs) improve clinical quality, operational efficiency and patient safety. Unfortunately, many HCOs may not be fully benefiting from the potential goldmine of useful information made possible by computerized records and analytics resources (including specialists and software) within their organizations. In my recent article posted on the Health IT Exchange, I discussed how to ensure that analytics are a strategic resource for healthcare transformation and operations excellence, and not merely a “report repository’.

The following are three of the lessons learned by HCOs seeking to leverage analytics on Quality Improvement initiatives (check out the article for the full list):

Consider analytics from the start. Avoid analytics becoming an “afterthought,” – consider the information and analysis needs for a QI initiative from the time planning of the initiative starts, not when it is well underway.
Focus on information needs of end-users. When developing analytical applications and BI tools, know which stakeholders require what information, at what frequency, and in what format.
Innovate. Analytics and BI offer so much more than reporting – be sure to use available tools to their maximum potential. This can range from making information tools more intelligent by highlighting where and alerting when certain defined conditions occur, through to the use of simulation modeling to test new workflows and processes prior to implementation.

The bottom line is that analytics professionals and quality improvement practitioners must work more closely together than in the past to improve healthcare quality and operations excellence. Even as the volume of healthcare data continues to grow by virtue of an increasing array of computerized record systems available, healthcare organizations don’t require more “reports”. They require targeted, intelligent “information tools” that identify the issues most pressing to an HCO and highlight possible mitigation strategies. This level of tool can only be developed once analytics becomes a strategic organizational resource and becomes fully part of ongoing transformation initiatives.

2011 Healthcare BI Summit a Great Success

Trevor Strome — Fri, 14 Oct 2011 19:40:59 +0000

I would like to thank the folks at Lancet Software for hosting the 2011 Healthcare Business Intelligence Summit held in Minneapolis MN and for inviting me to speak at the event. The event was very well attended, with many healthcare BI stakeholder groups (including clinicians, executives, and analysts) represented.

There were many excellent speakers at the event, so it was unfortunate I couldn’t see them all – but click here to see the full list of speakers. I would, however, like to give a special shout-out to the speakers I did have the good fortune to see. Mark Hagwarth (Healthcare Informatics magazine) gave an excellent keynote speech that really kicked off the day. Scott Wanless (Resource Management Professionals) and Tom Ludwig (Forward Healthcare Solutions) presented on the five key critical success factors for BI to be successful in healthcare organizations. Laura Madsen (Lancet Software) and Lisa Gordon (Microstrategy) spoke about the current state of Healthcare BI and what we can expect in the future. Allana Cummings (Northeast Georgia Health System, Inc.) wrapped up the summit with excellent examples of how business intelligence can transform healthcare organizations, and what needs to be in place to make that positive change happen. I’ll be writing about some of these presentations in more detail in the near future.

My presentation at the summit was entitled “From the waiting room to the treatment room – clinical analytics in action”, and I’ll be posting a link to the presentation soon for those who wish to download it.

If you are a developer or user of healthcare business intelligence and analytics applications, I would strongly recommend you attend the event next year. It is truly an outstanding opportunity for learning, and networking. Congratulations to the entire Healthcare BI Summit team for hosting a great event!