Impact of IoT on the Healthcare Producers: Epitomizing Pharmaceutical Drug Discovery Process

Abstract

Internet of things (IoT) emerged as a promising technology in the last decade and predicted to be ascendant in the next. Its application in the producer side of the healthcare industry is still in the nascent stage but expected to increase manifold in the near future. The purpose of this chapter is twofold; first, illuminate on the IoT applications on the pharmaceutical manufacturing and supply chain practices with real examples, and second elaborate the wide avenue of the opportunity of IoT it has in the drug discovery. Where most of the previous works argue the prospect of IoT in the conceptual or theoretical manner, we, however, intend to show the utility of automatic information processing in the context of computational drug design, which is an integral part of the drug discovery process. We integrate quantitative structure relationships with activity (QSAR), property (QSPR), and toxicity (QSTR) by utilizing an optimization technique to come up with a combined decision model. Numerical analysis has been performed with the developed optimization model considering three different cases using a simple chemical structure to test the model. Results suggest that the developed mathematical model can successfully be able to integrate QSAR, QSPR, and QSTR parameters which in terms of aid in automatic information and data capturing and lessen human efforts. This automatization can help in generating “optimal” drug candidates by considering all necessary facets. The present chapter also discusses other aspects of the healthcare producers where IoT can be proven beneficial in the near future.

Appendix: Tutorial for Building Optimization Model in Excel Solver

We have provided a step-by-step procedure to prepare the excel solver input–output analysis for solving the linear optimization model. After each step, readers are advised to check their formulated model in the excel file with the figures provided for the proper reproducibility. However, it is a better approach to build own model first and then cross-verify with the tutorial figures for a complete understanding.

• STEP 1: Writing the decision variable and parameters row-wise so as to ensure that multiplication and addition between variables and parameters turn out to be easy when we formulate the objective function. For all the six positions in the structure, we prepare the excel for ADME property, therapeutic value, toxicity value, and the decision variable line, i.e., choice of ith element in the Nth position. We then provide the value for the parameters and set the value of the cells under decision variables to zero. Figure 1 is the screenshot for further understanding.

  Fig. 1

  

  Preparing the excel for decision variables and parameters input for solver analysis

• STEP 2: Next, we prepare and formulate the objective function with the help of STEP 1 where we prepared the entire input matrix of decision variables and parameters. Therefore, we first prepare the multiplicative operations, i.e., ADME Property * (Therapeutic Value – Toxicity Value) * Decision Variable. We then add all such cells for each of the cells prepared from the input matrix. Please refer Fig. 2a, b for further clarity in this regard.

  Fig. 2

  

  a Formulating the objective function in line with the input matrix formed from decision variables and parameters. b Formulating the objective function in line with the input matrix formed from decision variables and parameters

• STEP 3: Now we prepare the constraints in the excel sheet. The first constraint is to ensure that all the values as entered in STEP 2 are greater than or equal to zero. In the absence of this constraint, the excel solver may assume negative values for a particular cell. The next set of constraints (one constraint for each position) is to ensure that only one element is fixed to a position. We show the formula for each constraint in separate screenshots as shown in Fig. 3a–g for the purpose of clarity to the students.

  Fig. 3

  

  a Screenshot for Constraint 1 in the optimization model. b Screenshot for Constraint 2 in the optimization model. c Screenshot for Constraint 3 in the optimization model. d Screenshot for Constraint 4 in the optimization model. e Screenshot for Constraint 5 in the optimization model. f Screenshot for Constraint 6 in the optimization model. g Screenshot for Constraint 7 in the optimization model

• STEP 4: Next we prepare the optimization algorithm by initiating the solver function in excel. Students should first go to the “Data” tab and at the right-hand most corners, EXCEL SOLVER will be present. Students are advised to first click the button. A pop-up screen will appear. Students will just need to select the (A) objective function, (B) maximization or minimization option, (C) select the decision variables, and (D) select the constraints. Please note that in addition to the constraints explained in STEP 3, we have also incorporated the binary constraint for the decision variable. Then, we have to click the options button and then click “Assume Non-Negative” and “Assume Linear Model” to ensure that the optimization model is a linear programming problem and all the decision variables are non-negative. Then, we click solve to get our answer. Refer Fig. 4a, b for further understanding.

  Fig. 4

  

  a Initiating the solver function in the Excel and incorporating the constraints, decision variables, and objective function. b Incorporating linearity and non-negativity in the optimization model