I try never to forget that in most cases, the simplest solution is the best. This applies to software as well. The ideal situation is to have one interface to the blood bank software.
In many laboratory softwares, there is a middleware to interpret the raw data as it comes out of the equipment which may reformat it and interpret it. This approach means that you have to have one interface then middleware and then finally the software.
Inevitably, both the middleware and the laboratory software will need updating. Can you assume that they will still work after either or both are updated? Will there be regression errors?
A good example of regression errors is the Microsoft Windows Feature Update (sometimes referred to as the Update from Hell!). Previously working functionality gets broken, data can be lost, etc.
In Medinfo, you store a truth table of possible results and interpretations. Medinfo will directly read the machine interface data, interpret it according to the rules you make. The data can be numeric or alphanumeric or both. Each test, each equipment can have its own unique rules if necessary.
In my opinion, it is best to avoid middleware if your blood bank software can perform this function directly. When you upgrade it, it is much less likely to show errors—and you only have to deal with one vendor. Can you be certain that the middleware vendor and the blood bank software vendor will work well together to resolve any issues?
The following are some examples of truth tables, i.e. the rules for interpretation and disposition of interface results actually used at HMC Doha during my tenure there:
Example 1: Blood Component Production Truth Table
Production can only proceed if the volumes are within the specified ranges. This is very important if you are going to perform pathogen-inactivation since the ultraviolet illuminator requires a specific range, even more so if you are adding Mirasol (riboflavin) and PAS (platelet additive solution).
Example 2: ABO/D Antigen Typing:
One manufacturer’s requirements:
Example 3: Complicated D Typing Algorithm:
The acceptable range for automatic interpretation is much more complicated for D typing with the Ortho Vision MAX and uses 3 different monoclonal cocktails:
Example 4: Direct Antiglobulin Test Algorithm:
Both alphanumeric and numeric results are used.
Example 5: Four-Cell Antibody Screen:
Truth table for interpretation requires both alphanumeric and numeric results:
If you are fortunate to have a dedicated blood bank software, you may not need middleware. Otherwise, you may need to use it for linking to general laboratory systems. Hopefully, your vendors will cooperate with each other.
To Be Continued: