Friday, March 7, 2014

MANA Study Part 2: Neonatal Mortality Rates

You are reading Part 2 of my MANA study series. Click to view:

MANA Study Part 1: Intrapartum Mortality Rates

MANA Study Part 2: Neonatal Mortality Rates (currently reading)

MANA Study Part 3: Total Mortality Rates

MANA Study Part 4: Vaginal Birth After Cesarean

****

Neonatal refers to the newborn period of life, specifically the first month. Neonatal mortality rate refers to the death of a baby any time after live birth and through the first 27 days. If a baby dies prior to labor or prior to birth, the baby will not be included in the neonatal mortality rate. Only babies born alive but die some time in those first full 27 days are included in the neonatal mortality rate.

Neonatal mortality rates can be broken up into early neonatal (death in the first week) and late neonatal (death after the first week but before the 28th day). I will be looking at TOTAL neonatal mortality rates, which is simply early neonatal mortality plus late neonatal mortality.

The overall neonatal mortality rate for the United States is 4/1000 (4 deaths for every 1000 live births -- which is the same neonatal mortality rate as Canada). This rate is includes all gestational ages (including extreme prematurity), all risks, all complications, lethal anomalies, etc. For obvious reasons, this is not the correct comparison to the MANA study neonatal mortality rate.

Just as I did for the comparison of intrapartum mortality rates, I am going to compare the neonatal mortality rates from 3 different data sets:

1. The MANA study - planned home births with a midwife in the USA, attended mostly by CPMs, LMs and LDMs (Certified Professional Midwives, Licensed Midwives and Licensed Direct-Entry Midwives)

2. The Birth Center study - planned births in select CABC-accredited birth centers with a midwife in the USA, attended mostly by CNMs (Certified Nurse Midwives)

3. USA hospitals - births in hospitals in the USA


First, let's look at comparing MANA to the Birth Center study.

Here are the neonatal mortality rates, both rates are excluding anomalies:

MANA study = 0.77/1000

Birth Center study = 0.40/1000

That is an increase risk for babies born at home of just shy of two times. Small yes, but per 10,000 it is an additional 3-4 babies that die at home births during the neonatal period.

****

Now, let's compare to hospitals using the CDC Wonder Database. Since it is impossible to get an exact comparison group for hospital births, I will use different hospital groups for the comparison: one that is as close of a match as I can get to the MANA study, and one that is broader. I will define each of them below.

In the first comparison, I will try to match the MANA study as close as possible. So the search results will be based off of the following criteria:

Years 2004-2009 (same as the MANA study)
White Women (92.3% of women in the MANA study were white)
Singles and Twins (small percent were twins in the MANA study, the rest were singletons)
37 Weeks and Above (97.5% of women in the MANA study were 37 weeks and above)
Birth weight of 2500 grams or more (99.2% of babies in the MANA study were 2500 grams & up)
Excluding most lethal anomalies*

*Excluded are congenital malformations, deformations and chromosomal abnormalities from the hospital group. This may not include all lethal anomalies, as some may be labeled as something else, like cardiac disease, for example.

My search includes everything else... all ages (maternal), all education levels, all marital statuses, etc.

2004-2006:

2007-2009:


Neonatal mortality rates for hospitals based on this search criteria are:
Medical Doctor = 0.33/1000
Certified Nurse Midwife = 0.25/1000

Please note: The hospital CNM group is mostly low risk women; the hospital MD group is low risk and high risk. MD refers not only to OB/GYNs but high risk pregnancy specialists, such as perinatologists and even other medical doctors who may deliver babies, even on rare occasions (such as ER doctors).

Comparing this similar hospital group to the MANA study and we have the following for neonatal mortality rates:

MANA study = 0.77/1000

Hospital CNMs = 0.25/1000

Hospital MDs = 0.33/1000

MANA study rate is 3 times greater than the hospital CNM rate and 2 times greater than the hospital MD group. Per 10,000 it is an additional 5-6 babies that die at home births during the neonatal period, compared to the hospital CNM group; compared to the hospital MD group, per 10,000 it is an additional 4-5 babies that die at home births during the neonatal period.

****

CDC neonatal mortality data is very accurate, despite what MANA and Citizens for Midwifery have claimed (I will get to that in another post). However, birth certificate data isn't perfect. We are potentially sacrificing the high accuracy the more search criteria we use. So that is why I will do another comparison with hardly any criteria selected.

The rates below are based on the following:

Years 2004-2009
Birth weight of 2500 grams or more
Excluding most lethal anomalies*

*Excluded are congenital malformations, deformations and chromosomal abnormalities from the hospital group. This may not include all lethal anomalies, as some may be labeled as something else, like cardiac disease, for example.

My search includes everything else... all races, all ages (maternal), all gestational ages, all singles and multiples, all education levels, all marital statuses, etc.

2004-2006:

2007-2009:


Neonatal mortality rates for hospitals based on this search criteria are:
Medical Doctor = 0.42/1000
Certified Nurse Midwife = 0.28/1000

Please note: As I stated above, the hospital CNM group is mostly low risk women; the hospital MD group is low risk and high risk. MD refers not only to OB/GYNs but high risk pregnancy specialists, such as perinatologists and even other medical doctors who may deliver babies, even on rare occasions (such as ER doctors).

Comparing this similar hospital group to the MANA study and we have the following for neonatal mortality rates:

MANA study = 0.77/1000

Hospital CNMs = 0.28/1000

Hospital MDs = 0.42/1000

MANA study rate is just shy of 3 times greater than the hospital CNM rate and just shy of 2 times greater than the hospital MD group. Per 10,000 it is an additional 4-5 babies that die at home births during the neonatal period, compared to the hospital CNM group; compared to the hospital MD group, per 10,000 it is an additional 3-4 babies that die at home births during the neonatal period.

Again we are seeing an increase to babies born at home compared to hospitals, even when we are looking at a much higher risk hospital group for comparison.

****

At least 5 babies from the MANA study that died during the neonatal period were due to hypoxia or ischemia. This basically means those babies were deprived of oxygen at some point. I cannot help but wonder how many of these 5 cases (and potentially more) were due to lack of proper monitoring during labor, especially during second stage (pushing).

For a look at the intrapartum mortality rates, click here.

Up next, I will put it all together and we can look at the total mortality rates.


References:

1. MANA study: http://onlinelibrary.wiley.com/doi/10.1111/jmwh.12172/full

2. Birth Center study: http://onlinelibrary.wiley.com/doi/10.1111/jmwh.12003/full

3. USA Hospitals: http://wonder.cdc.gov/

****


MANA Study Part 1: Intrapartum Mortality Rates

MANA Study Part 2: Neonatal Mortality Rates

MANA Study Part 3: Total Mortality Rates

MANA Study Part 4: Vaginal Birth After Cesarean


6 comments:

  1. Hi. I like your blog. I am not a scientist but I highly doubt that the differences found between these groups for this particular issue is statistically significant.

    ReplyDelete
    Replies
    1. Hi. Thank you!

      Can you explain why you don't think the neonatal mortality rates are statistically significant? You think a 275% increase is due to chance?

      Also, what are your thoughts on the intrapartum and total mortality rates?

      Delete
  2. I have a question regarding the following: "*Excluded are congenital malformations, deformations and chromosomal abnormalities from the hospital group. This may not include all lethal anomalies, as some may be labeled as something else, like cardiac disease, for example." Were these excluded in the MANA study also? I only ask because I wonder if it's a possibility that some of the deaths among the homebirths in the MANA study were attributed to any of these. I see this as a possibility because there are women who homebirth who choose to forego the screens and tests that would detect these anomalies, which would then include the baby in the neonatal mortality rates among homebirth but not have been a result of the decision to have a home birth. If they are included in the MANA study, and not in the hospital numbers used to compare to, how can we be sure there is an actual increase in neonatal mortality in homebirth? Especially since there are studies that show little to no significant difference between outcomes in home and hospital birth, such as this one here: http://www.bmj.com/content/330/7505/1416.

    ReplyDelete
    Replies
    1. There were some babies in the home birth group that did have congenital anomalies. They are excluded from the numbers above for home births and for hospital births.

      I also did a comparison without excluding lethal anomalies for both:

      http://whatifsandfears.blogspot.com/2014/02/mana-study-comparing-mortality-rates.html

      Delete
    2. Also, the BMJ study you included is very flawed. It actually shows the opposite, that home birth increases the risk for babies.

      First, the authors compared intervention rates between home and hospital for the same year, 2000. But when they compared mortality rates between home and hospital, they used the year 2000 for the home birth rates BUT for hospital rates, they used several papers dating back to 1969. Why didn't they compare hospital rates for the same year? Because if they would have, you find the rate to be double to triple the risk. The authors claimed that they did not have the hospital data available for the year 2000 when they published their study, which is why they used the dated papers. You can read a critique here which shows the numbers:

      http://www.sciencebasedmedicine.org/a-critique-of-the-leading-study-of-american-homebirth/

      It is also important to note that crucial data is missing from the BMJ study. In Autumn of 1999, NARM contacted all registered CPMs regarding this study. Not everyone could be contacted, not everyone planned to re-certify, not everyone agreed to participate. So they were left with 409 midwives. All 409 midwives AGREED to take part in this study. Yet, 18 midwives ended up not having their births used in the study - those 18 midwives were not able to re-certify because participation was mandatory in order to do so.... however, this does NOT mean those midwives are not still practicing. Where is the data on those 18 midwives and their births? Why did they AGREE initially to take part? What if they each had an infant or mother die under their care and that's why they did not complete the process? This essentially means midwives who had bad outcomes could bow out after the fact and those outcomes would then not be included in the study.

      These are two serious flaws. This is a study I once was totally behind.... this is what I had my husband (and others) read when we were planning for our home birth, this was my "proof" of home birth safety. I was completely unaware of these flaws at that time. I even wrote a response on the article in the British Medical Journal regarding the missing data on those 18 midwives but I have yet to hear a response.

      Delete
  3. Thank you for the clarification.

    ReplyDelete