Biomedical Engineering Reference
In-Depth Information
13.3 PROCESS ANALYTICAL TECHNOLOGY (PAT)
PAT requires the use of online, in-line, or at-line sensors to monitor continuous or
semicontinuous operations, and it uses the information to make real-time or near
real-time control decisions about critical process parameters and/or product quality. It
was originally developed for telemetry or remote sensing via wire, cable, telephone, or
wireless technology. Telemetry has been used by NASA for over 45 years for a wide
range of applications, monitoring both machines and men [16].
The development of microprocessor-based programmable logic controllers during
the 1970s was an important step for PAT as it now allowed control schemes using
feedback loops to monitor and respond to conditions based on the received signal.
Distributed control system (DCS) and supervisory control and data acquisition (SCADA)
system led to the integrated sensor and control systems now widespread in process
industries such as petrochemical, pulp, paper, and nuclear power plants. Process
analytical instrumentation was introduced in the 1980s and total organic carbon
analyzers and conductivity/resistivity sensors became standards in the semiconductor
industry 10-15 years before their acceptance in the pharmaceutical industry [16].
The pharmaceutical industry has been relatively slow to adopt this approach but the
last 5 years have seen a dramatic increase in interest, with multiple conferences and
magazine articles addressing the adoption of both PATand QbD appearing in the United
States and overseas. There are multiple reasons for the increase in interest, beginning
with the desire for improvements in manufacturing technology and quality initiated by
regulators, more recently shifting to the increasing pressure on manufacturers to pursue
opportunities for continuous improvements in process economics and reductions in the
cost of quality. A2006 survey suggested that the pharmaceutical industrymay bewasting
as much as $50 billion out of $200 billion spent annually on drug manufacturing [17].
The pharmaceutical industry is responding by engaging lean manufacturing prac-
tices, which seek to optimize production processes by the elimination of wasteful
practices and by streamlining processes through a focus on flow. Traditionally the
industry has used batch and queue processes withmultiple hold points. PAT supports lean
manufacturing both directly and indirectly by focusing on several of the seven wastes of
lean; specifically defects, waiting, overprocessing, and inventory (the remainder being
overproduction, transport, and unnecessary motion). It has been proposed that PAT can
improve quality management and reduce costs in quality appraisal and failures, and that
investment in prevention, such as QbD, continuous improvement, and PAT-enabled
controls reduce the total cost of quality by improving process capability [18]. This study
proposed that the major economic impact of PAT lay in its ability to reduce queues by
enabling real-time release of process intermediates and products, thereby reducing cycle
times and working capital (carrying costs of 20-25% were assumed for inventory).
Companies from biotech, branded pharmaceutical, and nonpharmaceutical industries
were benchmarked, and it is of note that biotech had the poorest turnover rates of the
group, suggesting that there are major opportunities for improvement. The extent to
which real-time release may be achievable for biotech products, where sterility is a major
issue, is debatable, but there is little doubt that in-process holds for intermediates can be
shortened with greater adoption of PAT principles.
