Tosoh HLC-723GX LIS Integration: GX Form vs RP Mode

Tosoh HLC-723GX LIS interface: GX FORM vs RP MODE explained. What each format carries, how results are matched to patients, and which setting to choose.
Tosoh HLC-723GX LIS Integration: GX Form vs RP Mode
The Tosoh HLC-723GX can send its HbA1c results to a host computer in more than one format, and the format you pick decides what your LIS interface can reliably do. Most labs never choose deliberately. They connect the cable, see data arrive, and discover months later that patient matching is unreliable or that a setting was changed during a service visit.
This post explains the two formats LISBridge supports for the Tosoh HLC-723GX LIS interface, GX FORM and RP MODE, what each one carries, where each one stops, and which one to use. If you want the broader introduction to this analyzer first, start with our HLC-723GX integration overview.
One Analyzer, Two Wire Formats
Neither format is HL7, and neither is ASTM E1394. Both are fixed-width ASCII records with no pipes, no carets and nothing to split on. Fields are defined by byte position, which means a parser has to know the layout in advance. That is why many middleware products skip this analyzer, and why a lab ends up reading HbA1c off the screen and typing it into the LIS. We cover that wider pattern in analyzers that speak neither HL7 nor ASTM.
Both formats travel inside the same style of framing: STX, the record, ETX, then a one-byte BCC block check that is an XOR across the frame. The difference is in what sits between those markers and in how well the vendor documents it.
GX FORM: The Documented Host Connection
GX FORM is the format described in the vendor's host connection specification, and it is the analyzer's factory default (the TRANS MODE setting reads GX FORM). It is the right choice for a host connection for three reasons.
- It is documented. Every field position comes from a published specification, not from inference.
- It uses the HI-LEVEL handshake. The host acknowledges each record with ACK or rejects it with NAK, and the analyzer sends EOT after the last record of a data set.
- It separates specimens from calibrators. The record carries a specimen-type marker, so calibration runs are not mistaken for patient results.
From a GX FORM record LISBridge reads total A1%, SA1c on the JDS, NGSP and Mono S scales, and SA1c (IFCC) in mmol/mol. That is the full set a lab needs to report HbA1c on whichever scale its clinicians expect. It is also the mode in which LISBridge can tell a calibrator from a specimen without guessing.
RP MODE: What It Is and Why It Is Different
RP MODE is what the analyzer transmits when RP MODE = YES. Tosoh reserves it for its separate data-management program, the 501RP. The host connection specification does not document it at all. It points the reader to the 501RP manual and tells you to set RP MODE to NO for a host connection.
That matters because it changes how a parser can be built. With no vendor specification, the RP MODE parser in LISBridge is derived from real wire captures rather than from a document. The fixed structure of the stream is what makes that workable. A run is a series of blocks, each introduced by a type digit and each of a fixed length: a sample header, a sample number, a results block, method information, a peak table, an end marker, the chromatogram points, and calibration coefficients.
What a Fixed Length and a Checksum Buy You
Because every block type has a fixed length and ends with a verifiable BCC, the stream can be walked deterministically. You do not need to find a delimiter. You read a block of the known length and check that the checksum agrees. If it does, the block is real.
The same property rescues unframed captures. Some capture tooling strips the STX and ETX bytes, leaving the payloads run together, and occasionally mangles a checksum. LISBridge accepts both framed and unframed RP MODE streams, and it recovers the unframed one by length and checksum rather than by guesswork.
The results block carries twelve right-aligned values. The ones LISBridge identifies and reports are total A1%, A1a, A1b, F, LA1c, SA1c and A0. SA1c is the reportable HbA1c. RP MODE also includes the chromatogram, so LISBridge can render the trace with the integration windows marked, which GX FORM does not provide.
Verified at Runtime, Not Trusted Offline
An undocumented format deserves suspicion, so LISBridge cross-checks it on every message. Each named peak in the peak table has to match the value the positional decode produced, and total A1% is only reported when it agrees with the sum of its components. If a firmware version laid the fields out differently, the outcome is no result rather than a wrong one. For a number that drives a diabetes decision, a missing result is a recoverable problem and a wrong result is not.
How a Result Finds Its Patient
This is where the two formats differ most for day-to-day operation.
In GX FORM the record is defined by the vendor specification, including its sample identification, so a result can be tied to the order your LIS already holds. In RP MODE the identifier is a sample number: the month and day plus a four-digit number. LISBridge reports the four-digit sample number and drops the date. There is no barcode in the part of the stream that has been confirmed.
One field in the RP MODE sample header looks like it might hold the bar-coded sample ID. LISBridge does not use it. No capture has yet carried a known barcode that would confirm where the field boundary sits, and guessing would risk binding a result to the wrong patient order. We would rather leave a field unused than match a patient on a hunch.
The practical consequence is simple. If your workflow depends on barcode-driven matching, GX FORM is the safer choice. If you run RP MODE, your sample numbering discipline matters: numbers must be unique within the period your LIS looks back over, and the person loading the rack must enter them consistently.
What Is Not Supported
LISBridge does not auto-route the other TRANS MODE formats the analyzer offers: G8, G9, G7 and GHbV. We refuse to guess at them. Without a specification or verified captures, a guessed layout is how a wrong HbA1c reaches a chart.
RP MODE has one further limit. No calibrator or QC marker has been observed, so every RP MODE run is classified as a result. GX FORM does not have this problem. If your lab runs calibrators and controls through the analyzer regularly, that alone is a reason to prefer GX FORM.
Which Setting to Choose
- Default to GX FORM. It is documented, it is the factory default, and it separates specimens from calibrators.
- Use RP MODE only when you have a reason. The usual reason is an existing 501RP workflow that you cannot change, or a need for the chromatogram image alongside the result.
- Never switch formats mid-shift. Changing TRANS MODE changes the byte layout. Settle it once, record it, and make sure whoever services the analyzer knows not to reset it.
- Do not use G8, G9, G7 or GHbV for a LISBridge connection.
Setting It Up
Connect the analyzer by RS-232 serial or TCP/IP, set TRANS MODE to GX FORM, and add the instrument in LISBridge. Map the analyzer's parameter names to your LIS test codes yourself in the mapping screen, so no custom programming or vendor ticket is involved. Raw instrument data is stored next to every parsed result, which means you can always compare the exact bytes the analyzer sent against what was recorded. If the link drops, LISBridge reconnects with exponential backoff and jitter, and crash-safe local storage keeps results through a power cut.
The instrument page lists the current details: Tosoh HLC-723GX on LISBridge. For the protocol background, see HL7 and ASTM support and how data parsing works.
Frequently Asked Questions
Which format does the HLC-723GX use by default?
GX FORM. The TRANS MODE setting reads GX FORM from the factory, and that is the format the vendor documents for a host connection.
Can LISBridge read RP MODE?
Yes, including captures where the STX and ETX bytes were stripped, as long as the fixed block lengths and checksums verify.
Why is RP MODE not recommended for patient matching?
It carries a month-day plus four-digit sample number rather than a confirmed barcode, so matching depends on consistent sample numbering.
Are calibrators and QC separated in RP MODE?
No marker has been observed for RP MODE, so every run is classified as a result. GX FORM separates specimens from calibrators.
What about G8, G9, G7 and GHbV formats?
They are not auto-routed. Use GX FORM.
Start your free trial — no credit card required. lisbridge.com/pricing
Comments (0)
No comments yet. Be the first to leave one!


