In its sixth wave, the Survey of Health, Ageing and Retirement in Europe (SHARE), implemented the collection of dried blood spot (DBS) samples in twelve of the SHARE countries as an innovative method to gather objective health data. Approximately 27,000 blood samples have been collected from respondents in those countries by trained interviewers. DBS collection is an efficient and feasible way to gather biomarker information in a large international population‐representative survey like SHARE. The blood is taken by a simple prick into the respondents’ fingertips, dropped on a filter card to create a blood spot and, after drying, is sent by standard mail services to a biobank for storage. In the lab, small discs are punched from these spots for subsequent biomarker analyses.
It has to be taken into account that the samples are not collected under controlled laboratory conditions, but during the survey in the home of the respondents. They are inevitably exposed to varying fieldwork conditions such as outside temperature and shipment time. In addition, the sample quality may be affected by shortened drying times, missing humidity protection during shipment, or failure to collect optimal blood volumes, the latter leading to small spots. It is known that these factors, environmental as well as collection-caused, influence the quality of the DBS and the herein measured biomarker levels. In SHARE, we collected information on these factors and DBS quality, so we can use them to adjust the measured raw biomarker values.
This project investigates the impact that the blood volume of each SHARE-collected and analyzed DBS has on the levels of different biomarkers. While we can observe and document many of the above described quality factors, it was not possible for us to measure the blood volume applied to the filter card at the time of the collection. Previously described precise approaches measured the blood volume by weighing a punched disc or applying radio-isotopic methods. There are further approaches that use the spot size as a proxy for the blood volume. Hereby, a size estimate is gained by measuring the diameter by hand or dividing the spots into different size categories by eyesight. Yet, neither of these methods is suitable for the volume or spot size determination of huge amounts of field-collected DBS, as they are available for SHARE.
We took advantage of the fact that all of the SHARE DBS samples were photographed during the punching process. We used the photographs to establish a new algorithm that precisely measures the blood-covered area of a spot for each DBS in an automated way suitable for large amounts of field-collected DBS. We have no knowledge of any other automated spot size measurement for such a large amount of DBS samples.
We show that the derived spot sizes as well as the other environmental and collection-caused factors explain part of the variability in the raw biomarker results (we control for respondents characteristics that influence these levels for biomedical reasons). We also compared the exact measures made by the new algorithm to mere size estimation, where the spots are divided into different groups depending on their sizes. Based on our findings, we state that (i) spot size measures have to be taken into account when working with biomarker data derived from fieldwork collected DBS samples and (ii) an exact measurement of spot size is better than a mere estimation of the size.