The age of the electronic medical record (EMR) is here. Virtual private network, Web portals, and Web-based EMRs have made it possible to access charts virtually anywhere there is a computer. EMRs provide higher levels of confidentiality, eliminate the physical wear and tear that plague paper charts, and greatly reduce the cost and space of maintaining thousands of files. The benefits are apparent and well documented.

Yet all these improvements are just the tip of the iceberg. To truly gain return on investment and fully benefit from an EMR, an organization must adopt the automated workflow technology embedded in the EMR. You must eliminate as much of the paper processes as possible; otherwise those same paper bottlenecks will drag down even the best EMR.

This approach goes beyond the mere conversion of paper to electronic format. Automated workflow allows facilities to challenge traditional processes and improve them, even reinvent them. Compliance/deficiency improvement tools, time management and workflow tools, auditing tools, and productivity and management tools are all products of automated workflow.

The EMR is the outer shell and contains the data, documents, images, and other parts of the record. However, automated workflow is the backbone that supports all the meaningful new processes that can come out of an EMR product.

What Is Workflow?
Workflow is the progression of tasks and activities that make up a business process. Every business has workflow. Processes are written and personnel familiarize themselves with the process and perform the required activities. This concept has been around a long time and provides guidance to workers and managers. So what’s the big deal about automated workflow?

Automated workflow is the system’s ability to qualify and route work instantaneously, with or without human intervention. Once predefined criteria are met, the work flows on to the next stage. Paper reports, once printed and manually reviewed, are now automatically run and placed in applicable electronic work areas. Automated workflow not only ensures that work is presented to the proper party, it also promotes completion in a timely manner by providing audit reports and supervisory notifications. How can automated workflow enhance a hospital’s HIM processes?

Improving the Processes
The typical approach to process improvement starts by thoroughly documenting current processes. This is a critical requirement before implementing automated workflow. To be able to improve performance, you must fully understand the current processes.

Recording the data in a flow diagram is an effective way to visualize a process. Generally, processes with frequent delays and bottlenecks will reflect a complex and cumbersome flow chart. Additionally, you should record the specific measures of performance that require attention. Although automated workflow is a powerful tool with multiple applications, we will focus on the most prevalent utilization—deficiency management. Targets could include general measures, such as the following:

• number or percentage of delinquent records;

• days from discharge to completion; and

• days from transcription to signature.

Or, they could include specific measures, such as the presence of the following:

• physician signature;

• history & physical;

• op report;

• consult;

• PKU; and

• discharge summary.

The list is quite long. You must have measurable values to set goals and objectives. As one of the most regulated departments under JCAHO, myriad standards already exist to set goals and objectives.

Let’s assume the targets and goals have been established within the department or as requirements for JCAHO credentialing. How can workflow solve the complex time and schedule constraints, interdepartmental communication processes, and physician expectations in resolving deficiency issues?

One great improvement provided by automated workflow is the ability to present individuals with charts ready for their viewing and completion. Under traditional paper processes, reports are created and worked based on multiple criteria—patient type, admission area, services performed, types of deficiencies, and so on. The individual working on that report must look through all charts, recording complete charts, missing documentation, missing physician signatures, noncompliant records, and other items too numerous to mention.

If the documentation is missing or incomplete, the individual continues to review until the record is complete and updated in the system. In a paper world, it is not uncommon for an individual to spend 30% to 50% of his or her time just confirming whether records are complete.

Automated workflow alleviates redundancy by delivering only the records ready to be processed by that individual. The same criteria generating paper reports can be built into a workflow—checking for the applicable documentation, signatures, completion level, and responsibility. As a result, individuals spend time only on issues requiring their attention.

Another improvement achieved through automation is the ability to simultaneously process multiple work items. Coders no longer have to wait for analysts and/or physicians to complete a chart prior to coding and visa versa. They can begin coding as soon as the documentation is in the system and presented to them electronically. Normally the documentation is scanned within 24 hours of discharge and the chart is coded the day after. At the same time, workflow can monitor for chart completion. This capability to process multiple work items is attributed to workflow’s rule-based programming.

One of the most painful processes in medical records can be acquiring a signature from the elusive physician who rarely visits the HIM department. It may become necessary for the paper chart to leave health informatics and be presented to the physician. Once in the physician’s hands, accurately tracking the chart is impossible. This process not only requires more human intervention and time but has the potential to end with a missing chart.

Workflow improves chart security and shortens the cycle of obtaining signatures. The ability to create work queues and route work, including the involvement of oversight committees and physician leadership, can help stem delinquency significantly. With the help of multiple stage delinquency queues, it is possible to significantly reduce physician suspension days. More importantly, the workflows start from the date the record or signature is absent. This could even be on the day of admission.

The goal is to create a computerized model for all forms of deficiencies, starting as soon as the deficiency is present, and eliminating as many manual interventions as possible.

The more complex workflow programming stems from the incorporation of Optical Character Recognition (OCR), which allows programs to create new forms and reports based on the existence of specific data elements within another document. For example, based on a specific word or phrase in a pathology report, a workflow program utilizing OCR can create a cancer screening form, route it to a queue for the physician to complete, then on to transcription, and back for a physician’s electronic signature. Eventually, it will generate the appropriate reporting requirements.

While some are uncomfortable with the reliability of OCR, others embrace the added functionality. With OCR, the only area left for human intervention is the written documentation. While Intelligent Character Recognition (ICR) is available to perform this function, we all know how difficult it can be to read a physician’s handwriting. Computer software has the same difficulty and is, for the most part, not considered an option.

In addition to document-based criteria, workflow can incorporate HL-7 values. Any information contained in the HL-7 interface can trigger workflow. This can be in the form of specific HL-7 values or a combination of HL-7 codes and documents. Any information contained in the EMR database can be used to trigger the workflow thread. The options are numerous and only limited to the ingenuity and creativity of the users and analysts.

As if the ability to automate nearly every aspect of deficiency management and shorten the timeline to complete records is not enough, automated workflow can also significantly improve the overall management and auditing of the compliance process. In the case study below, automated workflow programs were used to monitor and measure American Joint Committee on Cancer (AJCC) compliance records. Before automated workflow, the facility randomly audited approximately 12% of the charts for compliance. The compliance level of the audited charts was 61%. After automated workflow, the facility was able to validate 100% of the charts each month. During the same transition, the compliance level rose from 61% to an average of 95%, with 100% compliance achieved multiple times.

A Successful Case Study
Let us take a look at the implementation of a successful automated workflow for completion of AJCC staging for the cancer data management department in the Gibbs Regional Cancer Center (GRCC), a part of the Spartanburg (S.C.) Regional Healthcare System (SRHS).

The GRCC, which opened in June 1999, provides comprehensive cancer services for upstate South Carolina residents. It is adjacent to the main hospital campus and offers patients a variety of treatment options and care services. The SRHS has been an American College of Surgeons (ACoS) Commission on Cancer (CoC) Community Hospital Comprehensive Cancer Program since 1959. In 2004, there were 1,339 cancer cases diagnosed and treated through the SRHS. The Cancer Care Committee provides oversight for all cancer services provided throughout the SRHS.

The cancer data management department or cancer registry serves as a repository for the collection, management, and analysis of data on persons diagnosed with a malignant or neoplastic disease in upstate South Carolina. The basic source document is the patient’s health record, from which pertinent information is abstracted and utilized for research activities, education, and lifetime patient follow-up.

The ACoS CoC implemented a requirement for all analytic cases—those who are diagnosed and/or treated in an approved cancer program—to be staged by the managing or treating physician using the AJCC system, which was established in 1995 and is a tool for documenting tumor size, nodal involvement of the cancer, and any metastatic disease. This information is imperative for planning treatment and offering the best options in fighting the patient’s cancer.

In 2001, a retrospective chart review for completion of AJCC staging on a sampling of analytic cancer cases was completed. The sample included cases from August 2000 to July 2001 (n = 140). Newly diagnosed cancer cases were reviewed for meeting the completion of staging requirement. The data demonstrated 61% compliance and was presented to the Cancer Care Committee at GRCC.

A work team, consisting of key personnel, was commissioned by the Cancer Care Committee to improve the process for completion of the tumor, node, metastases (TNM) staging document in the EMR. Members included staff from multiple departments such as health informatics (medical records), information services, cancer data management, oncology quality services, and a physician champion. The team’s goals were to create a user-friendly process for completion of AJCC staging within the EMR and to improve AJCC staging compliance.

The team reviewed the current workflow and set out to create an automated workflow process that would ensure improvement. They found that when a pathology report was positive for cancer, a staging report was provided by the pathology secretary. If the patient was currently an inpatient, the report and staging document was placed on the medical record for physician completion. If the patient had been discharged, the pathology report and staging document were sent to the health informatics department for scanning into the patient’s EMR.

The health informatics department would then review the staging document for completeness. If the staging form was incomplete, a dictation deficiency was assigned to the managing physician. The physician would then complete the staging form by either completing a hard-copy form that would be scanned into the medical record when completed or by dictating a staging summary. If dictated, the dictation summary was cold fed into the patient’s EMR by health informatics and then assigned a signature deficiency. The physician would then resolve the signature deficiency in the EMR.

The paper process was not consistent. The staging forms were not always found within the EMR. Therefore, no physician deficiency was assigned to the managing physician, which contributed to the unacceptable 61% compliance rate.

The team tackled the process to create a workflow system for identifying the EMR for review, assigning the documentation deficiency for the managing physician, and reanalysis to ensure completion of the staging document.

The cancer data management department took ownership of the process developed by the work team. Information Services developed a work queue with logic—if a patient had a diagnosis of cancer and a pathology report, the patient encounter information would route to a cancer staging work queue. This queue was then reviewed by the cancer data management department to determine whether the AJCC staging document for physician completion was attached.

If a completed staging form was on the medical record, no delinquencies were posted for the physician. If the form was incomplete, the cancer data management department electronically pulled the appropriate form into the patient record. Health informatics then assigned a missing text deficiency (HIM staging form analysis queue) to be completed by the managing physician.

The cancer data management department would then review the document for completeness (HIM cancer registry reanalysis queue), routing incomplete forms to the managing physician with an explanation. If the cancer data management department was unable to resolve the documentation deficiency after the second submission to the managing physician, it would contact the work team physician liaison, ACoS liaison, or the chairman of the Cancer Care Committee to intervene with the managing physician.

The work team chose five physicians with varying degrees of computer skills to pilot the new process during November. On completion of the pilot, each physician had completed more than four AJCC staging forms using the new process. Only one of the physicians stated it was somewhat difficult to complete the form using the new process. Four out of five physicians stated that it required less than five minutes to complete the form. All the physicians believed completing the electronic form in five minutes was acceptable.

During the pilot, other opportunities to assist the physicians with completion were discovered. The following actions were instituted after surveying the pilot physicians:

• access to the EMR via hospital intranet for physicians;

• initial physician training on the new process with easy access to health informatics and cancer data management personnel for questions following initial education;

• easier access to necessary information to complete the staging form (ie, pathology report, radiology reports, operative notes, and progress notes); and

• an upgrade to the electronic patient folder allowed text editing on the computerized document.

The new process was implemented in January 2002 for all physicians. The monthly compliance percentages for 2002 ranged from 90% to 100% with a 12-month average of 95%. In 2003, the GRCC received a three-year accreditation as a Comprehensive Community Cancer Center from the ACoS CoC.

The automated workflow system accomplished the Cancer Care Committee’s goals to create a user-friendly system and capture AJCC staging on analytic cancer cases. This automated method also removed the ineffective paper process and allowed 100% review of AJCC staging on all analytic cases.

User-defined fields have been set up in the cancer data management’s software system as a mechanism for collecting and identifying all eligible analytic cases that have not been staged properly or accurately using the AJCC-TNM Cancer Staging System. The fields are completed at the time the patient’s cancer information is abstracted from the EMR, which is never greater than six months from the date of first contact with the patient.

Reports are then generated from the IMPAC Medical Systems software to determine the rate of completion and accuracy. The results are monitored and reported at least annually to the Cancer Care Committee. CoC standards for cancer program approval require that 90% or greater of cases be staged appropriately. That performance level has been exceeded since the implementation of the EMR workflow process.

Beyond the Basic EMR
The cancer center case study highlights the benefits that can be achieved through the adoption of EMR automated workflow. As this facility concludes its third year of accreditation, it also brings to light that this technology has been available for some time. Workflow programming available in today’s EMRs is quite powerful, yet all too often underutilized.

The workflow solutions are not that complicated—all it requires is the dedication of HIM and a reliable IT partner. Put together a team, assess your situation, have an IT resource educate the team on your current systems’ capabilities, and move forward. Use the 80/20 rule—attack the problem areas responsible for the greatest compliance deficiencies. If your current system does not have automated workflow capability, use that same assessment to determine whether investing in that technology is justified.

In the GRCC case study, the facility successfully utilized Horizon Patient Folder and Image Link Engine as tools, part of the facility’s McKesson suite of products. However, the same solutions are available within most healthcare IT systems that incorporate a document management product.

— David Rodrigues, who provided the technical workflow programming and system management, is a consultant with HealthTek Solutions Inc, a healthcare IT consulting company.