Taking care of 2014 ITOP gestational age

Introduction - The VitalPro pregnancy module includes birth, abortion, and fetal death data. To be analyzed, any fields used must agree between the three data sets, or use a lower common denominator of agreeement. We must compare "apples with apples".

Prior to 2014 data, there was a "gestational age" (GEST) field for abortion data. That worked fine. Starting in 2014, apparently related to SB 1 legislation, passed in 2013, there is a "post-fertility age" (PFA) field.

Apparently, most of the SB 1 legislation was struck down as unconstitutional by the Supreme Court, because it puts an undue burden on women's access to abortion. Furthermore, it does not seem SB 1 requires DSHS to report PFA or GEST. Nevertheless, DSHS has ended up collecting PFA data, so we need to take that into account as best we can.

GEST is the number of weeks since the last menstrual period. PFA is the number of weeks since fertilization. The two fields are related. Assuming the data are entered correctly (it seems they are not, see below), one may be converted to the other. PFA + 2 = GEST.

Data quality - Examining the 2014 abortion results, it was immediately obvious that the distribution of weeks seemed off. A spreadsheet was made. The spreadsheet compares the distribution of the 2014 POSTFERT field with the distributions observed for the 2012-2013 GESTATIO field.

So far, it seems:

* 2012 and 2013 data are similar
* 2014 data are significantly different
* 2014 data have a bias toward higher gestational age
* For example, average weeks for 2014 (8.4) vs for 2013 (7.3)
* For example, bar chart is shifted to the right for 2014 data
* It seems likely that gestational age is still being entered by many
* For weeks at abortion, the new data seem less accurate than previously

It's possible the above analysis is mixed up. If anything seems amiss, please let us know.

Beyond possibly make the data maybe less reliable, a drawback of the change is that it makes discussing the data more confusing. If confusing to data analysts and computer programmers, pity the poor legislative aide or senator trying to understand the information.

Method #1 to use 2014 abortion data within VitalPro pregnancy module - One possible solution is to simply continue reporting GEST. During the data importing step, PFA is converted to GEST by adding 2. Given that the legislation does not seem to require reporting PFA, and because GEST is the standard measure, and in the interests of simplicity, using GEST seems a reasonable course. DSHS staff have agreed to accept this method for now.

Sample output - Text format
Sample output - Spreadsheet format

Other possibilities - If Method #1 proves insufficient at some point, other methods are possible. The separate abortion module (Method #2) might be helpful. Another possibility would be to import both fields (PFA and GEST) into the pregnancy module, as separate fields (Method #3). This would be a significant amount of work. Also, this method is not as simple for the user as Method #1.

Perhaps there are other innovative solutions to make best use of this important but somewhat compromised field. Given that the data are messed up (it seems), there is not going to be any really good solution. At a minimum, we need to be aware that this field is as accurate as before.