DBS reference specimens were prepared on collection cards used by the Newborn Screening Quality Assurance Program (NSQAP) at the US Centers for Disease Control and Prevention (CDC). DBS simulating acceptable newborn specimens were prepared by applying 75uL of normal adult blood with hematocrit adjusted to 50% [2]. DBS containing an excessive blood volume were prepared by applying two aliquots of 75 ul to the same spot, and DBS with a deficient blood volume were prepared by applying 10 uL per spot. DBS containing a deficient red blood cell content were made by applying 75 uL of normal adult blood with the hematocrit reduced to 27%. Misshapen DBS were made by moving the pipette tip while dispensing 75ul of normal adult blood. DBS were also made with 75uL of normal blood lysed by freezing. All DBS were allowed to dry at room temperature for a minimum of three hours before evaluation.

The study was conducted on newborn DBS specimen cards submitted to the Newborn Screening Laboratory at the Georgia Department of Public Health (GDPH) for routine screening. Each DBS specimen card contained either 5 or 6 heelstick blood spots.

For routine operations, all newborn screening DBS specimen cards are initially evaluated in the Accessioning Section of the GDPH Laboratory. Any specimen cards deemed to be questionable are then reviewed by the Newborn Screening Laboratory manager or a Newborn Screening Unit supervisor, who (1) makes the final decision whether to grade the overall card as acceptable, and (2) determines whether there are preferred or unacceptable blood spots on the card. In this program, a total of nine 3-mm punches are needed for initial testing. Additional punches are required for any necessary repeat or reflex testing. For this reason, a specimen is considered unsatisfactory if it does not have at least two acceptable blood spots each of which can produce 5–6 punches.

For this study, 77 DBS specimen cards that were initially deemed questionable by one ELP in the Accessioning Section were independently reviewed by three other ELP in the Newborn Screening Laboratory. Each spot on each card was rated as 0 (unacceptable), 1 (marginal), or 2 (acceptable), and the entire specimen card was then graded as either satisfactory or unsatisfactory. An additional 100 DBS specimen cards that were initially graded as satisfactory in the Accessioning Section were also subjected to OSI assessment.

An OSI (CardScan (BSD Robotics,Brisbane, Australia)) provided for evaluation was first installed at CDC by one of the authors (DG), and parameters were adjusted using DBS reference specimens. Blank NSQAP DBS collection cards were used to create a template map by which the instrument located the individual blood spots on each card. Instrument settings were adjusted to identify at least 20 locations among all DBS on a specimen card from which suitable 3mm punches could be obtained. OSI evaluation parameters were optimized to discriminate acceptable, marginal, and unacceptable DBS and to identify acceptable punch areas within acceptable amd marginal DBS. The OSI was then transferred to the GDPH laboratory, where new template maps were created for the GDPH DBS collection cards. Some of the OSI parameters originally set at CDC were refined by GDPH for evaluating actual newborn DBS (Table 1).

Prior to each use, an instrument system calibration was completed using the CardScan Calibration Card. The calibration process adjusted the optical source light-emitting diodes (LED) to achieve pre-set values for the Transmission Target and the Compensation Target. The OSI evaluation algorithm (Figure 1) includes area, circularity, convexity, and consistency.

A clear plastic sleeve was placed over each DBS specimen card to ensure the card was flat while being scanned. Each scan required about three seconds. The results appeared on the computer screen with a scanned image of the DBS card showing colored circles superimposed on acceptable locations for punching 3mm discs (Figure 2). Specimen cards with at least 20 such locations were considered satisfactory. About 100 bloodspot cards could be scanned in an hour.

All specimen cards containing DBS made with the 75 uL of normal blood were evaluated as satisfactory by OSI, with a distribution of acceptable punch locations on each DBS. All of the specimen cards made with either 150 uL or 10 uL of blood were evaluated as unsatisfactory, with no suitable punch locations identified on any DBS.

Of the 100 DBS specimen cards initially assessed as satisfactory by a single ELP in the Accessioning Section, all cards (100%) were found to be satisfactory by OSI evaluation.

Of the 77 cards initially assessed as questionable in the Accessioning Section, 45 (58%) were subsequently re-assessed as satisfactory: 33 (43%) by all three ELP, and 12 (16%) by two of three ELP. Among the 32 that were re-assessed as unsatisfactory, 19 (25%) were rejected by all three ELP, and 13 (17%) by two out of three ELP (Figure 3). Overall, complete ELP consensus was reached on 52 (68%) of the questionable specimen cards, and only partial agreement on 25 (32%).

OSI evaluation of these 77 questionable cards found 48 (62%) to be satisfactory, 19 (25%) to be unsatisfactory, and 10 (13%) to be questionable,. When these OSI-questionable cards were grouped with those evaluated as unsatisfactory, the overall agreement between OSI and ELP was 52/77 (67%). Among all 177 newborn DBS specimen cards that were evaluated by OSI, 152 (86%) agreed with ELP assessment.

The categorical agreement between ELP and OSI assessment depended on the extent of ELP consensus (Figure 4). The overall agreement between ELP assessment and OSI evaluation was higher on specimens with full consensus by ELP (36/52, 69%) than on specimens with partial consensus by ELP (11/25, 44%).