Validity checks

Method Validation The signal path from detector up to the hard copy output of the final results must be perceived as a chain of error-prone components there are errors due to conception, construction, installation, calibration, and (mis-)use. Method validation checks into these aspects. 

Tab. 14.5 Validity check of substructure-based calculation approaches.

DiFelice, R., Rapagna, S., and Foscolo, P. U., Dynamic Similarity Rules Validity Check for Bubbling and Slugging Fluidized Beds, Powder Technol., 71 281 (1992a)... 

Placebo-controlled trial of a new active medicinal product, if positive, means that the trial was capable of detecting a difference, and that the test treatment is, at least, more efficacious than placebo. This achieves two outcomes provision of an internal validity check of the trial methods, and provision to regulatory authorities of a basis on which to judge the difference between a statistically significant but clinically inadequate effect that would probably lead to the drug not being licensed. 

To determine if a process unit is at steady state, a program monitors key plant measurements (e.g., compositions, product rates, feed rates, and so on) and determines if the plant is close enough to steady state to start the sequence. Only when all the key measurements are within the allowable tolerances is the plant considered steady and the optimization sequence started. Tolerances for each measurement can be tuned separately. Measured data are then collected by the optimization computer. The optimization system runs a program to screen the measurements for unreasonable data (gross error detection). This validity checking automatically modifies the model updating calculation to reflect any bad data or when equipment is taken out of service. Data validation and reconciliation (online or offline) is an extremely critical part of any optimization system. 

Tray-to-Tray Rigorous Program Validation Check... 

In this chapter the aim is to get a better insight in the actual state-of-the-art in process industry (and some other industrial sectors, for purposes of comparison). At the same time it will serve as a (face-) validity check of the ideas on the design and implementation of NMMS s. 

MNV i,k) less than MNV i,j) but distance [i,k) greater than or equal to distance then i is not a valid neighbor of /, and / is not a valid neighbor of i. The neighbor validity check can result in many small clusters, but these clusters can be merged afterward by relaxing the reciprocal nature of the check. 

Data Processing derived values, data conversions, scaling, frequencies, periods, calculations, algorithms, validity checks, error correction... 

Data maintenance is required throughout the data life cycle (see Figure 12.2, based on GERM ). Data may be captured by a manual or automated input. User procedures are required for manual data input and their effectiveness should be audited. Software supporting automated data input such as that used for data acquisition by instrumentation or data migration tools requires validation. Checks should include confirming any necessary calibration has been conducted and if interfaces are working correctly as validated. 

Calculation of formulas an formats with ASCII-import ASCII-Export HTML-Import/Export Entry form Validity check/rules Filters... 

General Preliminary Execution Plan Validity Check Project Specific Durations Questions/Decisions Master Project Schedule Firm Execution Plan Presentation Compressing the Schedule Project Coordination Procedure... 

The validity check shows a construction duration of 10 months, one month longer than anticipated in the Initial Plan of Action in Chapter 5. Since an 18-month total project duration is important due to business considerations this delay could have a serious impact on the project economics and corrective action must be taken. 

The validity check (Section 9.4) showed an increase in construction hours with a corresponding schedule extension of one month. 

Fig. 9.2 shows that the field staff would peak at 115 workers instead of the anticipated 133 estimated with the validity check. It also shows the peak staffing for each discipline thus providing an excellent tool for an intelligent survey of local labor availability and contractors capabilities. 

This standardization saves considerable resources, not only in the study set-up process but also in training data entry personnel, coding validation checks, producing monitor reports, and for analysis/reporting. The continued usage of these standards can also allow for constant streamlining and improvement based on experience, with appropriate maintenance and controls. 

The completed validation checks should be run against test data to ensure they are written correctly. As the data is received and validated by these procedures, it is important to review the output and add or delete edit checks as appropriate. 

Conduct an inventory and account for study medications/devices and arrange for extra supplies, including other items, such as CRFs, blank forms and so on, if necessary. Resolve discrepancies between inventory and accountability records, and medication/device use, as recorded in the CRFs. If a pharmacy is involved in the study, the pharmacy and pharmacist must be visited. Check that the medication/device is being dispensed in accordance with the protocol. Check that the medication/device is being stored under appropriate environmental conditions and that the expiry dates are still valid. Check that the medication/device is securely stored in a separate area that is not accessible to individuals not involved in the study. Check that any supplies shipped to the site since the last visit were received in good condition and are properly stored. 

Since DCSs are based on digital controller calculations, a wide variety of special control options are available in self-contained modular form. These can be easily selected by click and drag action on some DCSs. In this manner, complex control configurations can be conveniently assembled, interfaced, and implemented. In addition, a variety of signal-conditioning techniques can be applied to process measurements, including filtering and validity checks. 

As can be seen, validation is a holistic approach where the validity of the model is based on many difference aspects. The guidelines presented by McLeod (1982) can be generalized into the following validity checks ... 

The more validity checks a model passes the more valid the model will be perceived. 

Sometimes when model validation is not needed, or to supplement already discussed validity checks, model stability is assessed. Model stability determines how robust the model parameters are to slight changes in the data and how robust the model predictions are to changes in the either the model parameters or input data. Collectively these tests are referred to as sensitivity analysis. If model outputs are particularly sensitive to some parameter then greater accuracy needs to be obtained in estimating that parameter. A stable model is one that is relatively robust to changes in the model parameters and the input data. 

---