Добавил:
kiopkiopkiop18@yandex.ru t.me/Prokururor I Вовсе не секретарь, но почту проверяю Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:
Ординатура / Хирургия / Библиотека им академика М.И. Перельмана / Книга_5441_Библиотеки_им_академика_М_И_Перельмана.pdf
Скачиваний:
0
Добавлен:
10.10.2026
Размер:
9 Мб
Скачать
☆
118

5.3.1 Automated Dispensing System

Solid oral dosage form material handling in the present era relies heavily on auto-
mated dispensing devices. Electronic devices called automated dispensing systems
are used to accurately and consistently deliver medicine (Paul etal. 2023). They
have transformed the way pharmacies and pharmaceutical companies handle phar-
maceuticals, making the procedure quicker, more effective and more precise (Sng
et al. 2019). Automated dispensing systems (ADS) are computer-controlled
machines that store, dispense and track medication inventory (Shin etal. 2023). The
computer system monitors the medicine dispensing process and controls the dis-
pensing device. The computer system may also produce reports on drug consump-
tion, stock levels and patient information (Gharib etal. 2023). For the purpose of
ensuring that prescription orders are appropriately lled and documented, ADS can
also be connected with electronic health records (EHR). Medication error reduction
is one of the main advantages of ADS (Brax etal. 2023). One research found that
one in ve drug dosages given in hospitals are delivered incorrectly, demonstrating
the prevalence of pharmaceutical mistakes in the healthcare industry (Paimard etal.
2023). By ensuring that the right prescription and dosage are administered, auto-
mated dispensing systems may dramatically lower the chance of medication mis-
takes. The capacity of ADS to enhance drug management is another benet. It is
simpler for pharmacies to maintain their stock levels and replenish medication as
necessary, thanks to ADS’s ability to keep pharmaceutical inventory and track con-
sumption (Benoit and Beney 2011). These devices automate and employ robotics to
correctly distribute the right amount of material. By delivering drugs in exact
amounts, ADS can assist decrease waste by lowering the quantity of medication that
is thrown away (Alahmari etal. 2022). On the market, there are several varieties of
automated dispensing systems (ADS). The most typical examples include the
following:
5.3.1.1 Unit Dose Dispensing Systems
These systems distribute medication in premeasured amounts, which makes it sim-
pler for patients to take their prescription as prescribed and lowers the possibility of
medication mistakes (Hänninen etal. 2023). These systems deliver medication in
specic dosages, making them perfect for healthcare facilities like hospitals and
nursing homes (Dubourg etal. 2013).
5.3.1.2 Centralised Dispensing Systems
In hospitals, clinics and other healthcare institutions, centralised dispensing systems
are frequently used to administer solid oral dose forms including tablets, capsules
and powders (Balka and Nutland 2004). Compared to conventional manual dispens-
ing techniques, these systems provide a number of benets, including enhanced
efciency, accuracy and safety (Tsao etal. 2014).
P. Saikiran etal.
119
5.3.1.3 Robotic Dispensing Systems
Robots are used in these systems to take medications out of storage and provide
them to patients (Hänninen etal. 2023). These robots are set up to retrieve pharma-
ceuticals from predetermined storage places and distribute them into bottles or blis-
ter packs of various sizes (Batson etal. 2021).

5.3.2 Vacuum Conveying Systems

Vacuum conveying is one of the most widely used material handling techniques in
solid oral dosage forms (Beaulac etal. 2022). A vacuum generator, a conveying
pipeline and a receiving vessel are the three major parts of vacuum conveying sys-
tems for solid oral dosage forms. The solid dosage form is dragged through the
pipeline and into the receiving vessel by the negative pressure that the vacuum gen-
erator produces in the pipeline (Verstraeten etal. 2017). Stainless steel or similar
substance suited for pharmaceutical uses is often used to make the pipeline (Tu and
Yan 2022).
The utilisation of advanced sensors and controls is among the major develop-
ments in vacuum conveying systems for solid dosage forms (Bhatia 2019). These
systems have the ability to continuously monitor the ow of material and adapt the
vacuum pressure and ow rate as necessary (Hou etal. 2023). This makes it possible
to precisely manage the conveying process, which is necessary when working with
delicate or sensitive materials (Chen et al. 2022). The application of specialist
machinery for handling powders and other ne materials is another development in
vacuum conveying systems for solid oral dosage forms (Klinzing etal. 2010). These
valves are intended to stop substances from escaping into the environment or being
caught in the valve. Additionally, they may be set up to reduce dust production,
which is crucial for maintaining a hygienic and secure working environment (Gomes
de Freitas etal. 2021). The accuracy, dependability, and adaptability of vacuum
conveying systems to a variety of pharmaceutical applications are increasing due to
advancements in sensor technology, specialised machinery, and system design
(Kuang etal. 2020).

5.3.3 Flexible Screw Conveyors

Flexible screw conveyors are a kind of mechanical conveyor that transports goods
through a tube or duct using a revolving helical screw. This approach is very adapt-
able and may be created to meet the particular requirements of a production facility
(Chongchitpaisan and Sudsawat 2022). Since they can handle a variety of materials
and offer exibility of conveying distance, angle and routing, these conveyors are
enormously benecial in the manufacturing of pharmaceuticals. Additionally, ex-
ible screw conveyors are kind to the goods and can reduce break or damages (Zhang
etal. 2021a).
5 Advances inPharmaceutical Oral Solid Dosage Forms
120
Flexible screw conveyors’ capacity to handle a variety of materials is one of its
main benets. This is crucial in the production of pharmaceuticals since various
materials may need to be carried at various points along the manufacturing process
(Zhang etal. 2021b). Flexible screw conveyors are an adaptable option for pharma-
ceutical manufacturing since they can easily handle powders, granules, capsules,
tablets and other products. Additionally, exible screw conveyors are simple to
clean and maintain, which is crucial in the pharmaceutical business where cleanli-
ness and hygienic conditions are of the utmost importance (Dhaval etal. 2022).
Furthermore, these conveyers are easy to clean and are made of materials that are
resistant to corrosion and degradation (Dahlgren etal. 2019).
5.4 Advantages ofNew-Age Material Handling Techniques

5.4.1 Automation

Modern material handling techniques are characterised by automation, which pro-
vides organisations with a number of advantages (Sng etal. 2019). Automated solu-
tions may boost productivity, cut labour expenses, increase accuracy and lower the
risk of accidents and injuries from human handling (Rao and Pathak 2022).

5.4.2 Enhanced Safety

Modern techniques for material handling can also increase workplace security.
Automated machinery and systems can lower the chance of disasters and injuries,
which is crucial in elds that deal with big or bulky objects (Denis etal. 2020).

5.4.3 Higher Productivity

Modern approaches to material management can boost output in a number of ways
(Ansari and Gangil 2022). Automated systems reduce bottlenecks and increase ef-
ciency by handling a greater number of materials in a shorter amount of time
(Björnsson etal. 2018).

5.4.4 Enhanced Accuracy

Additionally, automated devices can increase the precision of material handling
operations. These systems are made to carry out repeated activities precisely, lower-
ing the possibility of mistakes and enhancing data accuracy (Arshad etal. 2021).
P. Saikiran etal.
121

5.4.5 Reduced Costs

Modern material handling techniques can also aid in cost-cutting for companies
(Zaman etal. 2022). Because they require less manual effort to complete tasks,
automated devices can lower labour expenses (Arvind and Gunasekaran 2014).
5.5 Limitations ofNew-Age Material Handling
Even though material handling advancements have many advantages, they also have
signicant limitations (Björnsson etal. 2018). The necessary initial investment, the
possibility of technology failure and the disappearance of the need for labour are
only a few of the drawbacks (Denis etal. 2020).
5.6 Utilisation ofArtificial Intelligence inSolid Oral
Dosage Forms
The practice of simulating human intellect with computers is known as articial
intelligence (AI). In 1956, Marvin Minsky and John McCarthy hosted a symposium
where the idea was rst put out (Hassanzadeh etal. 2019). Technology adoption
may reduce costs and save time while also improving understanding of formulation
and process characteristics (Agrawal 2018). Previously, AI was only used in the
eld of engineering, but more recently AI has become crucial in a number of
pharmacy- related elds, along with drug discovery, the development of drug deliv-
ery formulations, marketing, management, quality management, hospital pharmacy,
etc. (Grof and Štěpánek 2021). In the creation of drug delivery formulations, differ-
ent articial neural networks (ANNs), including deep neural networks (DNNs) and
recurrent neural networks (RNNs), are used (Elbadawi etal. 2021a). Then again, de
novo design encourages the development of much newer medicinal compounds that
possess the desirable characteristics (Das etal. 2021). The pharmaceutical market is
dominated by solid dosage forms (Shaikh etal. 2018). Tablets are thought to make
up the majority of solid dosage forms and account for more than two thirds of the
worldwide market (Xiouras etal. 2022). AI may be used to create improved formu-
lations, which would be extremely benecial to the pharmaceutical sector (Lou
etal. 2019). Four phases make up a typical AI workow: data gathering and prepa-
ration, AI modelling, simulation, testing and implementation (McCarthy et al.
2006). Implementing algorithms and seeing patterns in data to aid decision-making
is known as machine learning, a branch of AI (Catania 2021), articial neural net-
works (ANNs). Compared to traditional machine learning algorithms, ANNs (arti-
cial neural networks) drew inspiration from the biological neuronal structure of
human brains that provide superior computational and predictive power (Wang
etal. 2021b). Additionally, deep learning has been extensively employed for several
applications, including image classication, object recognition, image segmenta-
tion, natural language processing (NLP) and medical image analysis (Elbadawi
5 Advances inPharmaceutical Oral Solid Dosage Forms
122
etal. 2021b). Since it has been used so frequently in the pharmaceutical sector,
AI-based drug development is viewed as a potentially effective alternative to the
traditional approach (Bannigan etal. 2021). The framework known as ‘Pharma 4.0’
aims to address several persistent problems in the pharmaceutical manufacturing
industry by using modern digital methodologies (Nagy etal. 2019). In 2021, the US
FDA released the ‘Articial Intelligence/Machine Learning (AI/ML)-Based
Software as a Medical Device (SaMD) Action Plan’ with the goal of adjusting regu-
latory oversight and enabling the enhancement of patient lives (Wang etal. 2022).
Solid dosage forms are the most signicant form of dosage forms in pharma eld,
and over 50% of new molecular entities (NMEs) are being accounted repeatedly
according to the Food and Drug Administration Center for Drug Evaluation and
Research (FDA CDER), because of its several advantages, which include shelf sta-
bility, patient adherence, simplicity of transportation and exact dosing
(Suryadinata 2017).

5.6.1 Widely Used Databases

Obtaining a database is the initial stage in executing an AI-based study. A high-
quality database must rst be created in order to properly create a workable formu-
lation development model (Liu etal. 2020). There are various traditional methods
for creating a modelling database with analytical techniques like the Design of
Experiment (DOE) tool (Zhao etal. 2019). Scientists can identify the connection
between many elements and reactions using the organised approach known as
Design of Experiments (DOE) (Lee etal. 2022). They may also ascertain how vari-
ous components interact and maximise the response (Palo etal. 2021).
5.6.2 Methods forProcessing Data
Before creating the models, researchers must process the raw data they have
obtained from public resources or internal experimental ndings (Dong etal. 2021).
To adapt and then evaluate the data, it is important to employ several extensively
used techniques, such as data cleaning, dimension reduction, unbalanced data solu-
tions and data splitting (Han et al. 2019). The removal of data points and their
replacement with mean and median values are two ways to clean up data (Hesse
etal. 2021). One more essential step in data processing is data partitioning (Mak
and Pichika 2019). The entire data set will normally be separated into three sub-
groups using this methodology: training, validation and evaluation. Consequently,
prior to performing modelling tasks, data processing and splitting procedures are
required (Vamathevan etal. 2019).
P. Saikiran etal.
123
5.6.3 Development ofSolid Dosage Forms Using AI Algorithms
In recent times, a variety of AI-based approaches have been effectively used in the
creation of pharmacological solid dosage forms. Articial intelligence (AI) is a
combination of information technology, data analysis and mathematics (Shah etal.
2019). Combining computer science, data analytics and mathematics generates arti-
cial intelligence (Paul etal. 2021). ML is a branch of AI that is often divided into
three categories: supervised learning, unsupervised learning and reinforcement
learning (Aksu etal. 2012). A method known as ‘deep learning’ comprises of out-
put/target variables that will be estimated from a collection of input variables
(Castro etal. 2021). Throughout the training process, a relation of the input vs.
intended output will be developed, resulting in the accuracy level that is wanted
(Shaheen 2021). Solid dose formulations have been generated using a variety of
deep learning techniques, including decision tree, logistic regression, linear regres-
sion, k-nearest neighbors (KNN), random forest, XGBoost, LightGBM and support
vector (Ma etal. 2020). Unsupervised machine learning is an algorithm that con-
trols just the input variables and uses feature-nding and grouping techniques (Patel
and Shah 2022). A branch of machine learning called deep learning (DL) uses
cutting- edge algorithms like convolutional neural networks to learn from a signi-
cant quantity of experimental data (van der Lee and Swen 2023). In order to make
predictions, deep learning algorithms introduce a very complex model structure. In
recent decades, deep learning (DL) algorithims, aresuccessfully being used in phar-
maceutical eld for different purposes indevelopment ofa solid oral dosage formu-
lation (Fig.5.2).
Fig. 5.2 AI algorithms used for process optimisation of oral dosage forms
5 Advances inPharmaceutical Oral Solid Dosage Forms
124
5.6.4 Evaluation andExplainability ofModel
Predictive Performance
Assessment of the models’ predicting abilities is required following the machine
learning modelling procedure (Shaheen 2021). We may categorise the measures
used to evaluate the predicted accuracy into regression metrics and categorisation
metrics (Khanna etal. 2020). The coefcient of determination (R
2
), mean squared
error (MSE), root mean squared error (RMSE) and mean absolute error (MAE) are
often used metrics for evaluation of regression modelling assignments (Moingeon
etal. 2022). A confusion matrix will be used to construct various metrics, such as
accuracy, precision and recall, for classication modelling work initially. However,
in the process of unbalanced classication modelling, the outcomes of several clas-
sication metrics, such as accuracy, are deceptive (Floresta etal. 2022). As a result,
new assessment measures including Cohen’s kappa, receiving operating character-
istic (ROC) and area under the curve (AUC) are used to evaluate models (Lowe
etal. 2022).
5.6.5 Applications ofArtificial Intelligence (AI) inSolid
Dosage Forms
5.6.5.1 Tablets
The most important oral solid dose form is a tablet. A tablet is often created by
compression or moulding and contains a combination of APIs and excipients (Lou
etal. 2019). Excipients are substances added to tablets to promote tableting perfor-
mance. Examples include lubricants, glidants, binders, diluents, sweeteners, food
colouring and drug release modiers which all serve to improve aesthetics and dis-
integrants and sustain release polymer coating (Lamberti etal. 2019).
5.6.5.2 Predicting Drug Release
During the development of a product, two of the most important preclinical trials
are drug release studies, including invitro and invivo tests (Dropka and Holena
2020). Important material properties and critical processing variables have an
impact on medication release patterns. For instance, little adjustments to the com-
paction parameters, such as pressure and tablet shape, or other elements, like drug
loading, may have a big impact on dissolving rates (Elbadawi etal. 2021b). In addi-
tion, specialised apparatuses such as UV-visible spectrophotometers and USP-
approved vessels are needed for a typical invitro drug release investigation (Wirtz
etal. 2019).
5.6.5.3 Developing 3D-Printed Tablets Using Artificial
Intelligence (AI)
One of the most cutting-edge methods for customised medicine is three- dimensional
(3D) printing, which has the capacity to create tablets taking patients’ physiology,
genetic proles and pharmacological responses into account (Boobier etal. 2020).
P. Saikiran etal.
125
Individualised 3D-printed tablets have been created using a variety of techniques,
including fused lament manufacturing, binder jetting and selective laser sintering
(SLS) (Duch etal. 2007). AI technologies have enormous potential to be imple-
mented into this approach and discover the design window in order to improve the
3D printing process and decrease the test effort with many variables.
5.6.5.4 Detecting Tablet Defects
During the production process, it is normal for tablets to exhibit aws including
cracking, capping, binding and sticking. These damaged tablets often need to be
ltered out manually, which calls for a sizable workforce and is difcult to scale up.
X-ray computed tomography (XRCT) (Doerr and Florence 2020), for instance, can
be utilised to analyse the interior structure of tablets in order to x this problem.
Researchers have effectively detected tablet faults by combining XRCT with the
deep learning approach to broaden the use of this technology.
5.6.5.5 Granules
Another pharmaceutical solid dose formulation made up of aggregation of powder
particles containing medicines and excipients is called granules. Granules are more
shelf-stable than liquid formulations and provide variable dosing for people who
have problems inswallowing pills or capsules (Zhao etal. 2021).Articial intelli-
gence techniques were used to assess and anticipate the medicine contents in sugar-
free granules (Jiang etal. 2022). In this work, near-infrared (NIR) spectroscopy rst
showed that it was possible to determine the amount of medication in granules. The
drug remaining was then predicted using several machine learning techniques using
the near-infrared (NIR) spectrums. This proves the drug content in granules may be
quantied using AI models, which are acceptable tools for doing so (Grof and
Štěpánek 2021).

5.7 Continuous Manufacturing Technology

The technique of making medicinal items constantly, without breaks or stops in the
assembly line, is known as continuous manufacturing technology of solid oral dos-
age forms (Rogers etal. 2013). This strategy is distinct from conventional batch
manufacturing, which involves creating medicinal items in large batches and then
placing them through additional processing procedures (Lee etal. 2015).
The continuous manufacturing (CM) approach, in contrast, has been effectively
utilised for many years in the petrochemical, food, consumer goods and automotive
industries to increase production productivity and save costs (O’Connor etal. 2016).
In a batch mode, raw materials are loaded into a single-system operation after which
the process is performed at approved parameters until the predetermined endpoint is
reached (Hock etal. 2021). Until all standards of quality are satised and the mate-
rial may be moved to the following unit process, these intermediates are momen-
tarily kept in the warehouse (Myerson etal. 2015). For steady-state processing, a
constant hold-up mass must be present in the system, and this must be done while
5 Advances inPharmaceutical Oral Solid Dosage Forms
126
the same ow rate is constantly used to remove the nal goods from the process (Yu
et al. 2014). A traditional manufacturing process for pharmaceutical products is
made up of several independent unit operations such as high shear granulation,
tableting, drying and blending (Haimhoffer etal. 2021). The unit activities are all
combined into a single production train in an integrated continuous manufacturing
process, without breaks or starts between units of work (Fig.5.3) (Rathore etal.
2015). Recalls due to poor product quality have led to medicine shortages (Nambiar
etal. 2022). These quality problems were frequently brought on by out-of-date pro-
duction techniques and tools as well as a shortage of efcient quality control proce-
dures (Burcham et al. 2018). In addition to advancing technology, including
high-quality medications into products throughout development will strengthen
process capabilities (Poechlauer etal. 2013). The ICH (International Conference on
Harmonisation) Q8 denes QBD (quality by design) as a systematic method of
product development that starts with predetermined goals and places a signicant
emphasis on product/process understanding and control systems as a solid scientic
basis for quality risk management (Welch etal. 2017). Establishing clinically appli-
cable standards is a key goal of QBD (quality by design), i.e. relevant standards for
product quality that are focused on ensuring clinical efcacy (Balogh etal. 2018).
Secondly, a more capable process will result in less variation in the nal output,
which will lower the prevalence of aws (Thabet etal. 2018).
5.7.1 Overcoming Obstacles toContinuous Manufacturing
When it comes to combining unit activities into a single continuous process train,
the pharmaceutical sector has been somewhat cautious. A number of obstacles still
need to be overcome before continuous manufacturing (CM) becomes a preferred
and widely utilised technology platform throughout the industry (Medendorp
etal. 2020).
Fig. 5.3 Overview of batch and continuous manufacturing technology of solid oral dosage form
P. Saikiran etal.
127
5.7.1.1 Regulatory Uncertainties
For many years, the regulatory setting tended to restrict any modications made to
medicinal products after their approval (Srai etal. 2015). This contributed to a
mindset whereby batch procedures were still seen as the only viable option, despite
missed possibilities to accelerate the manufacture and release cycle (Srai et al.
2020). The FDA’s innovative technology team is willing to support the development
of this ground-breaking technology by engaging in early and participatory talks
with the organisation while creating a continuous manufacturing (CM) process
(Nasr etal. 2017). Process validation is the gathering and assessment of data from
the method design stage to commercial operations that creates scientic proof that
a process is capable of reliably producing high-quality products (Matsuda 2018).
5.7.1.2 Process Automation Technologies (PAT)
Most process automation technology (PAT) applications used nowadays in tablet
secondary manufacture are based on NIR (near-infrared) or Raman spectroscopy,
since these methods do not require sample preparation and provide quick, nonde-
structive assessments of physical and chemical characteristics (Fig.5.4) (O’Connor
et al. 2016). Both NIR (near-infrared) and Raman spectroscopies are molecular
spectrometric methods, yet they complement one another since they detect func-
tional groups in molecules differently (Allison etal. 2015). NIR (near infrared)
is more sensitive to polar bonds and asymmetric vibrations, whereas Raman
spectrometry is especially effective for examining nonpolar bonds and symmetric
vibrations (Esmonde-White etal. 2022). Raman spectroscopy has a lower limit-
of-detection than NIR spectroscopy. The most common problem for in-line
monitoring for NIR (near-infrared) and Raman spectroscopy is low detection
sensitivity (De Beer etal. 2011). Nevertheless, it should be considered that NIR
(near-infrared) and Raman spectrometry detection sensitivities are extremely
formulation dependent.
Fig. 5.4 Utilisation of PAT tools for optimisation of pharmaceutical solid dosage forms
5 Advances inPharmaceutical Oral Solid Dosage Forms