I am not familiar with Simplified Calibration Interval Analysis (I'm assuming you're referring to a specific protocol). But one approach I'm familiar with is measuring the Allan Variance curve. Basically, it is total noise vs integration time. For example, you can collect a hundred samples of nominally zero signal at a range of integration times. The shorter times will be dominated by white noise, and the longer times will eventually be dominated by 1/f noise. Usually there is a minimum in between these extrema, which is the optimum integration time.

Referring to the headline of your question:

The calibration interval is mainly a function of the rate of change of your instrument over time vs. the accuracy specified.

• Instruments that do not 'age' significantly (eg. multimeters, oscilloscopes etc.) are usually calibrated ever year or everest 2nd year.
• Faster aging instruments (eg. electrochemical sensor cells) may require shorter, regular calibration intervals.
• Instruments that age even faster (eg. spectrographic instruments with fast aging light sources) may require daily calibration. Or calibration every x hours of operation.
• Thinking of pH meter electrodes, calibration (1 point 'check') prior every single measurement is highly recommended to prevent measurement error.

All in all a wide span.

Hope this helps

For radiometric instruments like Pyranometrers, UV meters, the mean calibration interval suggested by the constructors is around every 6 months.

Spectro-radiometers with PM tube as detectors requires more often verification with a standard source for instance, It is also useful to verify angular response, linearity, straight  light,...

Hope it will help you