In this paper the full data sets {(ti, c(ti)); I = 0, 1, 2, …, k} have been generated by the authors from different pharmacokinetic models and some of the data points for each case were deleted from the above mentioned data sets to form reduced data sets. A monotonic piecewise rational quadratic s(t) was fitted to the above mentioned data sets and the range of integration was divided in to a number of sub-intervals. Then the area under plasma concentration-time curve was computed with four different repeated quadrature rules. The concentration values required to apply quadrature rules were approximated by corresponding values of s(t). The values of the computed integrals with minimum absolute error percentage were found. It was noted that if at the most 30% data points are ignored and the range is divided into appropriate number of subintervals, then the value of the computed integral is very close to the exact value of the integral in most of the cases. Also, these values are close to the respective values of computed integrals corresponding to the full data sets. The exact values of the integrals were calculated by dividing the range of integration into a large number of subintervals and then applying the repeated trapezium rule.