Want Better Results? Keep Decision-Making Close to the Operation

A new pharmaceutical industry benchmarking study found that better manufacturing performance results when operations staff make key decisions. Giving the quality assurance manager, instead of the head of QA for manufacturing, responsibility for deviation management and lot failure corresponded to lower numbers of failed batches, higher yields, and lower cycle times, the study found.

Jeffrey Macher of Georgetown University and Jackson Nickerson of Washington University in St. Louis issued the final benchmarking report of their Pharmaceutical Manufacturing Research Project, launched in 2002. The project investigated the effects of product, process, manufacturing site location, and organizational structure on manufacturing performance. In this second phase of the project, the researchers collected data from 42 manufacturing facilities owned by 19 manufacturers. Their report provides raw benchmarking data, as well as statistical analysis, on factors correlated with cycle time yield performance, deviation management outcomes, product unavailability, and process development. The first phase of the project involved a study of Food and Drug Administration inspection data, and those results were reported to FDA in January 2005.

The team identified several major factors that correlate to higher or lower manufacturing performance, though the researchers are careful to note that the correlations do not imply causation.

Positive Factors: Local Decision-Making and Effective Use of IT

Two factors linked to superior performance were effectively using information technology to automatically report and track deviations, and keeping decision-making, particularly on deviation management and lot failure, closer to the operation.

“The locus of decision rights impacts performance,” says Macher. “Placing individuals who are closest to the phenomenon or the problem tends to have a superior performance effect than having upper-level management make those kinds of decisions.” Before beginning the study, the researchers had preliminary information that showed wide variations in the way companies organized their approaches to solving deviations, and who was responsible.  So we wanted to see if there was one approach that mattered,” says Macher. “And one thing that did matter was how far away management is from the actual problem itself. What it suggests is that engineers have a lot of good ideas on how to handle deviations and they can be made more use of.”

Information technology was another important factor, but Macher empasizes that making big investments in IT is not enough; it's using IT effectively that counts. “Firms that automate the handling of information-by electronically and automatically tracking and reporting deviations-tend to do better,” he says. By providing engineers and quality managers with information on deviations and manufacturing problems, they can track down the causes more easily.

Benefits of PAT Unclear

One surprising find of the study was that using process analytic technology (PAT) tools corresponded to lower performance levels. “This [result] could be because facilities with manufacturing problems are looking for tools like PAT to help them solve problems better,” Macher says. “Most likely PAT tools are good, but because we're studying PAT tools at the beginning of their implementation in the industry, naturally they're going to show a negative performance. If we extended the study over time, you might see that firms that have PAT tools implemented show increasing performance going forward.”

Employee Experience Improves Manufacturing Performance

The study also found that more employee experience is associated with fewer batches failed. Although this is not surprising, Macher says they also saw this result in their FDA study, in the first part of the pharmaceutical research project, and in an earlier study on the semiconductor industry.

In the FDA study, Macher and Nickerson found that firms tended to have an increase in current good manufacturing practices (cGMP) violations two or three years after a merger. Macher postulates that by that time, the acquiring company has laid off employees who had a lot of experience in that manufacturing site. “Then a problem occurs, and the people with the most experience at that site are no longer available to solve it,” he theorizes. “There are a lot of nuances in any facility.”

Macher and Nickerson's study was nonpartisan, without any funding from either FDA or industry. They plan to publish additional papers based on a more in-depth analysis of the data. The full report is available at: http://faculty.msb.edu/jtm4/PMRP%20Results