### Recovery

Recovery studies involve the addition of a known amount of analyte to a sample and then determining what percent of the amount added is detected. A patient sample can be spiked with varying amounts of a pure standard to give concentrations at medical decision levels (usually the upper and lower reference limits). A similar amount of diluent is added to an aliquot of the same sample and run to give the baseline level. The volume of the addition should be <10% of the sample volume so that the matrix is minimally disrupted. Each level should be run in triplicate and the results averaged. The amount and percent of analyte recovered are calculated as follows:

Concentration recovered = test sample - baseline sample

**% Recovery = (concentration recovered/concentration added) x 100**

Recovery experiments assess the degree of proportional error, which is defined as 100 - % recovery. For example, if the percent recovery is 90%, then proportional error is 100 - 90 = 10%. Acceptable error levels vary by analyte, but generally should be less than total allowable error (see Appendix B).

The following example using calcium illustrates how to determine if a calculated percent recovery is acceptable. If %recovery is 92.5%, then proportional error is 7.5% (100 - 92.5% = 7.5%). Proportional error is converted to a decimal by dividing by 100 (7.5%/100 = 0.075). This number is multiplied by the medical decision level of calcium (e.g. 10.5 mg/dL) to obtain the amount of non-recovered standard (0.075 x 10.5 mg/dL = 0.79 mg/dL). The answer is compared to the total allowable analytical error, which is 0.5 mg/dL for calcium. Since 0.79 mg/dL is greater than 0.5 mg/dL, recovery is not acceptable. Recovery is acceptable, only if proportional error is less than total allowable error.