Ninth Annual Maternal and Child Health Epidemiology Conference / December 10-12, 2003

Results of Neural Tube Defects Case Ascertainment Process Combining Vital Records Datasets and Birth Defects Surveillance Data, for Years 1996 - 2001 in Puerto Rico

HECTOR GARCIA MALDONADO: Good afternoon to all.  I don’t think that the microphone is a problem for me here, so with my voice only I can give you a presentation.  For the last seven years, we have been working in Puerto Rico with a birth defects surveillance system.  And we began to focus on NTD in the island, thanks to the effort of a woman that was a pediatrician looking after potential causes and situations regarding neural tube defects.  And this excerpt that I’m presenting right now is from an article written by Larry Edmonds and other guys from the CDC, *Margaret *Watkins; my supervisor, *Joe *Millionaire, about that surveillance useful for monitoring birth defects as well as trying to look for distribution and changes of.  And I know that each one of us knows about this, but in Puerto Rico, we will--like in the past, trying to look for prevalence of NTDs since for the last 40 years we just attend a lot of cases with this condition, but we didn’t have*through all the State Department of Health or any other effort, prevalence rate or statistics that could prove about this situation. 

Basically, our surveillance system is based on collaboration from private and public hospitals.  We don’t have legislative mandate in Puerto Rico about registering birth defects.  And since 1996, the CDC collaborated together with the Puerto Rico Department of Health to put on a CDC Fellow prior to me to design an implement the surveillance system, especially beginning with birth defects surveillance.  And it’s an interesting fact, the surveillance came after public policy about folic acid in Puerto Rico, thanks to the effort of Dr. *Mountas in our island, most of the live births are covered in the island, mostly, 99 percent.  And since that, we try to develop an active case ascertainment of NTDs in the island.  We began first with a passive system, so most hospitals reported NTD cases to the Department of Health.  But since 1996, we hired five abstractors, and they went back to 1995, trying to look at any missing case that we could not identify through the passive system that we had. 

As I mentioned before, in ’96, CDC collaborated with a CDC Fellow to design an implement the BDSS.  In our active case ascertainment, we consider live births, stillbirths, terminations, and therapeutic and spontaneous abortion as part of our case ascertainment for NTDs.  And the most important fact about this, that most of the cases that we have is from live births.  And sometimes we could find some terminations, but in Puerto Rico, physicians have problems trying to put on the records about this kind of category for us to ascertain in case of neural tube defects, and I think other states happen the same thing.  This is a graphic that we have developed.  We have observed a decline of NTDs in Puerto Rico.  We know that Puerto Rican women do not take folic acid vitamins daily, as all of us know.  And we know that this decline is due to some situation in Puerto Rico.  We believe that fortification is part of that explanation; but right now, we’re trying to go into a folic levels study to determine if this effect for this decline.  The purpose of the study was trying to look on the vital records and see if we really have an additional ascertainment of cases of neural tube defects. 

Since we have an active ascertainment in Puerto Rico through the BDSS, for us it was important trying to identify potential cases that maybe we just lost throughout the active ascertainment if that was the case.  And we went through the vital records data sets that includes live births, fetal deaths, and death certificates, and with the collaboration of the vital records office, we’re trying to match some of those cases in order for us to look at potential new cases and those that we already have in our database.  The only cases that we consider as a fact was NTD, and those data sets that we received was about 500 to 700 birth defects obtained from the vital records office.  And from that, we just only got 310 cases of NTD cases identified in vital records only.  So we decided to determine what kind of procedure we would just look on in order for us to match those cases that we find out in the vital records together with our database in the surveillance system.  And we just use personal identifiers such as child and parents name, social security number, date of birth, and for identifying potential NTD cases in the vital records and BDS data together in the match process, we just consider hospital name, date of birth, parent and child name, social security, congenital anomaly, and the mother’s town of residence or the county by year of birth.  We just check also for any kind of transposition and typographical errors. 

Very different from the American culture, in Puerto Rico--as well as in the rest of the Hispanic countries throughout Central Latin America--we use the two surnames, the mother’s and the father’s side.  For example, I am Hector Garcia Maldonado, and “Garcia” is my parents’ name.  “Maldonado” is my mother’s name.  So that kind of situation, we have to look on two of the vital records because we have “Maria Rodriguez,” but we could have 1000 “Maria Rodriguez.”  So we have to check them for each record to look on if that Maria Rodriguez is the same Maria Rodriguez from the vital records.  And we have to look onto the two surnames for that, as well as for the infant names. 

As far as the analysis in our methodology, we’re trying to screen for the NTD cases using both codes, ICD-9, ICD-10.  You know that ICD-10 is very cumbersome to use it because now, they are categorizing many of the birth defects so detailed.  And we’re just trying to look for those potential codes combinations, so we assured that we can really match those cases and put out the process.  We consider only babies with one year of age or less than that, including (inaudible) between both sources.  And we also will be considering the validity of the vital records in terms of ascertainment for sensitivity and specificity and positive predicted value, using the surveillance, the gold standard.  And those all match cases, we were going to the hospital to review again all the records in order for us to look on if there were NTDs as reported in the vital records or if there were not. 

As far as the results, we found out that from those 310 cases that we found out from NTDs throughout the years, 198 cases only were found by records, and 405 were considered from the surveillance system, NTDs from 1996 to 2001.  Then, as a result of the process, we found out that vital records contributed with an additional seven percent of the 93 percent that we have already collected from the BDSS, and that seven percent we find out 30 new cases of NTDs not collected by the surveillance system, but find out in the vital records.  This is more or less the distribution per year of that ascertainment process between both two sources.  One of the things that we find out throughout the match process and through the revision of the hospital records of the NTD cases that we find out in the vital records is that there is a lack of report of those cases by the health care staff and at the hospital, especially physicians and some nurses, sometimes because of emotional burden or something like that.  Also, in the logbooks, it’s very difficult to find sometimes the NTDs.  And I have a personal experience. 

I went to do some audit of the cases that we have (inaudible) in Puerto Rico, and in the worst part, I was in a hospital that we didn’t find a case, for example, for Down Syndrome in the logbook.  When I had an interview with a physician of the nursery, he told me that he found out about one potential Down Syndrome case.  And we went to the logbook again to find out if that was really reported, and it was not there.  So those kind of problems we find regularly in Puerto Rico throughout the hospitals.  From those 198 cases that we find as NTD cases, 88 were not matched.  One hundred ten were matched with the surveillance system, but 88 was not.  And that’s the distribution throughout the period of time measured this, and this potential crisis, what really was trying to give them to those doctors, and they were back to the hospital again, trying to review and find out what happened with those cases. 

This table, we’re just trying to show you a distribution of what happened with those cases.  As you can see, we find out--33 percent of the 88, we found that there were new cases that we didn’t collect it through the BDS, through the surveillance system.  And 10 of them were not an NTD case, and they are pending.  As of this moment, there are 48 cases that we are trying to go back again, especially because most of them are from the San Juan area, metropolitan area, and the abstractor that works in that area, we have to retrain him in order for her to look on those cases again and trying to see if we are missing or not in San Juan those 48 cases.  This graph, what it really shows displayed is that we have an average of 27 percent of our NTD cases from *BDS data were reported in vital records.  That means in the opposite way that our surveillance system as an active one is very reliable and is very effective in the ascertainment process.  Our sensitivity in terms of all NTD types combined is 27 percent from the vital records data sets, and that range from 58 percent for anencephaly and 13 percent for spina bifida. 

The most strange thing about all this is that we find out that vital record--we find out more anencephaly potential cases in vital records than spina bifida.  In the opposite manner, in the surveillance system, we find more spina bifida than anencephaly.  And that is the most experiential thing in all the states, too.  So again, because many physician for emotional thing don’t want to report anencephaly or terminations or whatever.  The PPB value for NTD combined was 79 percent, and it ranged from 88 percent for anencephaly and 75 for spina bifida from the vital records.  These are tables how we would just develop, remembering that our gold standard is surveillance system.  We find out throughout this study that it’s almost obvious that that higher ascertainment that we have from the surveillance system was really achieved thanks to the efforts that we have put on trying to go to the hospital and finding all those cases in Puerto Rico, especially for--giving some intervention services in the case of a spina bifida per se through all the CH and CM centers that we have throughout the island, seven of them.

And despite of the effectiveness of our NTD section, we have to recognize the situation that vital records have contributed with an additional seven percent of ascertainment.  So it’s potentially a good source for us to depend on trying to look on for those new cases.  And of that seven percent, most of them we have to go to make a second review so we’ll be very, very sure that we are getting the cases and we’re not missing it.  Again, our ascertainment from surveillance is useful.  It’s very, very, very useful for establishing the completeness of data, and that’s something that most of us that work with birth defects is looking on when we are doing ascertainment of this kind of defect.  And even when predictive positive value was relatively high for NTD and reported from the vital records, we have to go again to do some revision of those cases. 

Three things I think that means for public health in Puerto Rico.  That means first that even when the studies show that most of the ascertainment comes from surveillance, we need vital records to ascertain additional cases.  And this study is the first one in Puerto Rico that proved those kind of things.  And empirically, we have the suspicion before that vital records could help us a lot, even when it’s not a very highly potential source for ascertainment for neural tube defects.  But it really gave us an additional avoiding any missing cases that we could have for entities in Puerto Rico.  And the third thing is that for us, it’s very important--and we are doing it right now--education for health care providers in terms of recording in writing those NTD cases in the medical records.  There is something lazy about them to write a lot of papers, you know; but I think that if we can put a lot of effort trying to not force them, but trying to make acquaintance of the importance of reporting NTDs throughout the hospital records, it will be much easier for us to find new cases.