Data management software has a good reputation for tracking the mountains of information engineering and medical companies generate. The systems don't store the data. They store meta data — data about data — so information remains distributed in dozens of servers, and the management software monitors its comings and goings. That the software keeps order is almost common knowledge, and what all companies need.

What engineering and business staffs are less likely to know is that data management software can make events such as audits less painful and time consuming. In addition, several quality management systems (QMS) and corrective and preventive action (CAPA) endeavors are automatically initiated and tracked. One ERP system allows slicing and dicing information for insight into manufacturing and business trends, and still others are taking active steps to mitigate risk. A few companies tell what's new and why.

Risky business

Risk assessment is a big topic in medical circles. ISO 13485 has guidelines for managing risk from design through manufacturing. And although the idea of risk management is not new, a few ideas are surfacing that make it less labor intensive and more consistent. “Consider, for example, the client that gets over 10,000 complaints a month,” say Tim Lozier, spokesman for EtQ Inc., Farmingdale, N.Y. (etq.com). “A complaint can often be loosely defined and spans comments that, for instance, a bandage did not easily come off its paper, all the way to life threatening events,” he says. But which pose the most risk to the company and so should be handled first?

“Previously, the oldest complaints were handled first,” says Lozier. “But many of those are no more serious than, say, scratched labels. A simple fix and they're done.”

Assessing a large number of complaints for the risk they pose can be daunting. “Low-risk items can be fixed quickly and need no CAPA. But there is a point where a complaint requires initiating a CAPA plan and reporting it to the FDA,” he adds. Risk assessment filters out these low-risk events so teams can pay more attention to the high-risk situations.

“The risk-management module in our QMS software helps this way: A team at a medical company pulls up the five-by-five risk matrix or filter in the software. One matrix axis could plot probability or frequency, and the other, severity. Companies could select other conditions.

A severity and probability of 1 on the matrix is not a serious event, and a complaint that earns a five in each case should initiate a CAPA. But what about those in between? What should be considered significant risk?

Many companies hire a consulting firm to help identify the risk associated with an incident. “But a team at the company should decide how to portion the matrix. There are uneven probabilities, and some companies write algorithms into the matrix,” says Lozier. “For example, a severity of five and a low probability of one may happen once in 20 years. That may not need a CAPA. For guidance in these uncertain areas, regulations such as, ISO 14971, describe reducing risk to levels ‘As Low As Reasonably Practicable’”.

Once established, test the matrix. “Put in real world examples from two and three years ago and compare suggested results to what actually transpired. The idea is to not let it produce flexible opinions. You want reliable, consistent decisions from it,” he says.

From a product perspective, set up the risk matrix using company-appropriate terms. “A complaint-handling system could include a risk matrix. Each grid selected might initiate another action. For example, a form might appear asking for more details so the software can send the complaint to an appropriate individual. If a CAPA is needed, the software starts one.

Who runs the risk matrix depends on the company. “One firm sends many complaints to the quality department, and its manager investigates. If it signals a CAPA, they do the risk assessment and their quality group works through the risk matrix to determine whether or not a CAPA should be initiated,” he says.

You can build a knowledge base from this input information, says Lozier. “For example, in several years, the assessing software can provide links to previous events on similar problems. It would be possible to jump start the CAPA because a similar risk assessment is already done.”

Traceability

The FDA may want to know almost everything about your medical devices. That means your data-management software will have to provide some degree of traceability, and in this case it's more than just where the materials came from. “A regulated amount of traceability is expected to be part of an overall management system,” says Jim Thompson, director of healthcare industry solutions consulting for Siemens PLM Software, Plano, Texas (siemens.com/plm). “So when manufacturing medical products, expectations from regulatory agencies are that companies have a good definition of traceability as it relates to their product.”

Conventional traceability refers to storing information from manufacturing through clinical trials. “In fact, it should start all the way back in requirements or needs derived by doctors from patient populations that justify features in the product that address particular medical conditions,” says Thompson.

“Most companies do a good job of maintaining traceability, usually in a document-centric way,” he says. “Traceability also means generating documents at key stages or milestones of a product lifecycle that are reviewed as product ideas develop. In the case of a research project, traceability includes the rationale for moving from one phase to another. That is general practice. Regulatory agencies tend to suggest using a document-centric approach, which has pros and cons. The big advantage is that we know what it means to put ideas and thoughts on paper, or electronically. That's a good thing. You can dig out documents when a problem arises, or if you want to enhance a product, you can read through related documents to understand as much as possible how the device got to where it is today,” he says.

But software technology introduced over the last several years moves away from a document-centric approach. “Documents will not go away. But we'd like companies to think in a more data-centric manner so information in documents link elements of a product's lifecycle to each other. Rather than have a document that captures requirements for a medical device and letting another describe the product spec and manufacturing, we want that information linked so if there is a therapeutic benefit, it is traced to particular product features and manufacturing controls.”

A discussion of paper verses electronic documents may lead one to believe that the electronic version is the end game. “It is not. It's a stepping stone on the way to a more integrated systematic environment supported by an IT system,” says Thompson.

Here's where traceability pays off first: In audits. “You regularly hear that audits can be extremely labor intensive whether imposed by a regulatory agency, or by the company. Audits tend to put entire manufacturing organizations and key individuals in a completely reactive state. The activity means they put all normal work aside to retrieve documents to support requests,” he says.

But an audit need not be disruptive or painful. “It is possible with the adoption of modern software to significantly decrease the labor and pain a company tolerates supporting an audit by simply relying on its PLM system to bring past events into the present, along with interconnected data and documents that make up the product's history,” says Thompson.

PLM, meet QMS software

There is no one all encompassing PLM-QMS software, but links in the software let programmers swap information from one to another. There are good reasons for doing so. For one, as companies grow, departments take on their own functions and before they know it, communication between them subsides.

“PLMs generally focus on R&D through manufacturing,” says Nikki Willett, VP Marketing & Regulatory Products at Pilgrim Software, Tampa, Fla. (pilgrimsoftware.com). “The software provides a range of functions focused on design collaboration and engineering project management.”

“The issues in life sciences concern quality and the FDA, often around CAPA. Some companies have tried to address these issues within the engineering process. Some effort is coming from the regulation for medical devices, part 820 and design controls. Despite the development, PLM and QMS (quality management systems) are still isolated environments. How does information from the PLM get where it's needed in the quality system? When organizations look at quality, they must track the process, data, people, and product from the beginning of an idea through customer and vendors,” says Willett.

QMS provides companies with compliance and quality systems aimed at life sciences. “CAPA is one area that must be in compliance with regulation standards. Software is now capable of tracking nonconformance, CAPA, compliance handling, internal and external auditing, control of standard operating procedures, employee training, and equipment calibration,” says Willett.

However it happens, the systems become isolated and the value of collaborating around quality gets lost. “QMS software, however, captures a lot of nonconformances at their source,” she adds.

In addition to PLMs, companies might have isolated systems for supplier tracking, for dealing with out-of-spec equipment, or tracking complaints taken by a call center. An enterprise-wide quality and compliance platform, such as Pilgrim's, could capture all that in one source of information and provide it to the PLM program,” says Willett.

Take for example, finding the root causes of problems. “These deviations or nonconformities must be investigated, and you need to find the root cause and plan action to correct the error and prevent it from happening again. The analysis of root cause, FMEA, (failure modes effects analysis, a search for root causes ) needs data coming from QMS system. If we have the source information of the incidents coming in, we can provide that to PLM as well as the data or report that assists with risk assessment for the product or new products,” she says.

Let's say nonconformance is coming from the design. This calls for a change. It may call for a new material or circuit board. But if a manufacturing leader says ‘we need a change,’ the comment could be lost because of the disconnect between engineering and manufacturing. So the effort to let PLM and QMS communicate would be well spent.

Better tracking prevents bad blood

Texas-based Carter Bloodcare serves some 300 medical facilities in 54 Texas counties. The company recently decided to implement TrackWise software, enterprise quality and compliance management software from Sparta Systems Inc., Holmdel, N.J. (spartasystems.com), to help it comply with FDA guidelines for quality management. The company says the software will let it standardize and automate previous paper-based processes used among the facilities. The software was chosen for its ease of use, and ability to track and trend across departments, while reducing IT administration.

A business-rules engine for workflow will help manage company preferences for meeting FDA regulations and provide insight into ways of streamlining and improving how employees manage their activities. In addition, the software will help track employee training, changes to procedures, auditing for compliance to procedures, and increase employee productivity by improving management of center-wide training policies and requirements.

Analytics give more insight to company performance

Some analysis of manufacturing efficiency comes from standard reports on sales and backlogs, but more insight into manufacturing operations might come from slicing and dicing information to a company's preference. Syspro Inc., Costa Mesa, Calif., (syspro.com) for instance, says it has a method for doing so in its Analytics module that is part of its ERP (enterprise resource planning) system.

“One purpose for the Analytics module is to turn 2D tables and data into information that is more revealing and appears 3D,” says Syspro product manager Rene Inzana. “The software transforms data into cubes which let users create multi-dimensional views of sales and operational data to find patterns and underlying trends and gauge performance. ERP systems are information hubs, but the information becomes meaningless if it cannot be analyzed and used to make informed decisions. Information is power and the need to be able to really explore it and drill down to underlying sources has never been more apparent than in today's fast-paced global environment. This is where Syspro Analytics comes in because it manages the task of extracting and transforming ERP data, eliminating the need for knowledge of internal data structures and technical OLAP know-how,” says Inzana.

“Although a few key performance indicators or KPIs are built into the system, users can devise their own. A viewer in the software lets users present data in several layers if appropriate, and that gives the 3D quality,” she says.

To simplify use and startup, users select from predefined templates of metrics and KPIs to provide information on key aspects of their organization's operations.

“You might start off reviewing sales by salesperson and while analyzing this information, you realize that you'd like to know who the sales people are selling to. Normally, you'd have to produce another report or go to another query. With Analytics, you just drag the customer column onto you view and instantly see what you need. You then might want to see exactly which items are being sold — same action applies. After looking further at the data, you might decide you want to see the sales by month. Instead of having to go to another report, you expand the YTD figure to months, or even weeks or days. You could even decide that you just wanted to compare “this month” to “this month last year.” Analysis screens or Cubes simplify the work. “Managers could drag that information into a cube and software finds relationships. It's easy to find more data, change views, and compare sales records for specific quarters,” she says.

“The need to understand where information comes from is done away with through a series of analysis objects and metric templates,” says Inzana. This lets manufacturers reach decisions promptly.

The software is said to incorporate the best practices from Microsoft using MS Analysis Services with graphing software as the analysis interface. Data from the Analytics viewer can be exported to Excel and PowerPoint for presentations.