I was reviewing PFT reports today and noticed that a patient appeared to have had the wrong gender entered in their demographic information. Specifically, the patient had an unambiguously masculine name but had been entered as female. Just to be sure I checked the patient’s on-line medical record and there he was listed as male. I had noticed from the trend report that the patient had been in the PFT lab numerous times. Since the basic patient demographics (name, date of birth, height, gender, etc.) are automatically forwarded into a new demographics record when a new PFT lab visit is created it struck me as odd that after all this time we had somehow managed to make a mistake with something as basic as gender. For this reason I thought it would be a good idea to see how far back this problem existed and started going back through the patient’s PFT records. About four visits ago the patient’s name suddenly changed to one that was unambiguously feminine.

I was immediately concerned that two different patient’s records had somehow gotten merged. The last time this happened was over 20 years ago and was due to an entry error in the patient ID that was further compounded by how the lab’s software handled new demographic records at the time. Merged records is therefore a symptom of a serious database problem but when I compared the date of birth of the two patients, I was immediately able to see that they were the same. Since this is incredibly unlikely my thought then was that the patient may have had a gender reassignment. When I went back to the patient’s online medical record and searched more carefully, I was able to find that this had occurred over a year ago. This is not the first time we’ve had a transgender patient and so it is an issue we’ve learned how to handle.

So what effect does gender reassignment have on an individual’s pulmonary function test results?

None whatsoever. Gender reassignment by itself does not affect FVC, FEV1, TLC or DLCO. What it does affect is how we interpret the test results and it can also cause some interesting data management problems that are worth noting.

All pulmonary function reference equations differentiate between genders. Although the differences between races and ethnicities is somewhat open to question, there is little doubt about the differences between genders. When individuals with the same height are compared, females universally have lower flow rates, volumes, respiratory muscle strength, gas exchange and oxygen consumption than males. Because lung function is determined during an individual’s childhood and adolescent developmental periods, gender reassignment does not affect lung function and when it is assessed this has to be done using reference equations that are appropriate to an individual’s original gender.

Depending on which way a gender reassignment occurs, results that would be considered normal for a female would likely look reduced for a male, and results that would be considered reduced for a male would likely look normal for a female. The selected gender will therefore make a difference about what an individual’s PFT results look like to a reviewer.

The patient whose gender raised this issue has relatively severe lung disease and is probably not the best example for this, but its what’s in front of me right now.

The male results would be classified as a moderate restrictive ventilatory defect with a coexisting obstructive ventilatory defect and a moderate gas exchange defect. The female results would be classified as a moderate restrictive ventilatory defect with a coexisting obstructive ventilatory defect and a mild gas exchange defect. Not a real big difference, but the percent predicted all of the male results are noticeably lower than the female and I’m certain that for some individual’s whose results are on one threshold or another, this could make a significant difference in how results are interpreted.

Gender really only becomes a problem when it is recorded incorrectly by mistake or omission, or when our data management systems forces us to enter it incorrectly.

Gender can be entered incorrectly every so often (for my lab probably around a half dozen times a year) and for this reason I make a point of checking the patient’s name and gender when I review tests. Many transgender individuals however, are (quite understandably given society’s conflicted and often negative viewpoints on the subject) reluctant to disclose this fact, even to medical staff. Transgenders can also be insistent about identifying themselves as their reassigned gender, again understandably. For these reasons, demographic information can be entered incorrectly because the patient has omitted to provide the correct information.

A more insidious problem however, lies at the intersection between PFT lab databases and a hospital’s information system (HIS). This is a problem my lab had up until the time our HIS interface was last updated, about three years ago. Specifically, when we uploaded test results, the patient’s id number, last name, first name, date of birth and gender were cross-referenced in order to be sure the results were assigned to the correct patient in the hospital information system.

Our problem, therefore, was that if we entered the patient’s original gender in order to get the right predicted values, we weren’t able to upload the test results because the patient’s gender in the PFT lab didn’t match the gender in the hospital information system. If we entered the gender that was in the hospital information system, we didn’t get the right predicted values. The work-around we finally came up with was based on the fact that at that time our interface was based on transferring text files and that the text files were created by one process and uploaded by another. So when it was necessary I would intervene and grab the text file before it was uploaded and manually edit the patient’s gender to match what was in the hospital information system (although to be honest this usually after the interface kicked out an error, not before, but it still got done).

For (some) good and (many) bad reasons gender is not currently an identifying factor with our HIS interface. We are however, in the final planning stages of another revision to our interface with the hospital information system that will (finally!!!) let the physicians electronically sign PFT reports. (Although it is actually more correct to say that we are bystanders in the revision process since almost all of the decisions on how it is going to be implemented are being made by the hospital’s IS department and our equipment/software manufacturer.) At this time it is not clear what identification factors will be used to cross-reference patient records so this may well become a problem again.

There are however, a number of hospital information systems that are very tightly coupled to a PFT lab’s database. In some cases, usually as a function of scheduling, the hospital’s demographic information is used to populate the demographic information in a patient’s PFT database. In these circumstances it’s not clear to me whether gender is an item that can be changed once it has been inserted into the PFT database and if it can be changed, what effect it has on uploading results. I’ve visited labs that have had this kind of a system but this is one item I’ve never thought to check.

It has been estimated that between 2% and 5% of the population have some degree of sexual dysphoria (an unease or dissatisfaction with their gender) but gender reassignment is still a somewhat uncommon process. A related, and very interesting issue however, has to do with individuals born with ambiguous sexual characteristics. Depending on how you want to define it (and there over a dozen different definitions) up to 1 in 100 children are born with ambiguous or intersexual characteristics. Maybe as many as 1 in 1000 children have surgery to “normalize” their genitalia. Many of these individuals do not know they were born this way and many times they never find out. Despite the fact that gender has a significant effect on pulmonary function results I have not found any research that looked at the effect that ambiguous or intersex characteristics may have on the development of lung function. I have to wonder if at least a small part of the variability found when studying pulmonary function in large populations is due to sexual characteristics that are not readily apparent.

The primary problem that we have with managing pulmonary function results for transgender individuals is the dissonance between their “real” and their “assigned” genders. There are physiological reasons for using one, and sociological and psychological reasons for using the other. The real limitation is in our databases and the way we collect demographic information. Gender is always assumed to be only a binary choice (male | female) but as we learn more (and become more open) about human variation (genetically, developmentally and socio-psychologically) a more nuanced approach would probably be more appropriate. One solution would be to have at least two gender choices for an individual, ‘physiological/developmental’ and ‘preferred/apparent’ and this should be applied to both PFT lab and hospital databases.

PFT Blog by Richard Johnston is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License