Wednesday, September 25, 2013

How safe is home birth in Delaware? The only statistical analysis ever done

(That I can find, at least ...)

This is the text of a report that will be released tomorrow night at the Division of Public Health meeting on non-nurse midwives tomorrow night in Clayton.  Follow the link for details.

This will be a public meeting that begins the process of fixing the very bad mistake that the General Assembly and DPH made last year in passing HB 194 and essentially criminalizing the entire profession of midwifery and the practice of home birth.

What most astounded me when I began talking with the bureaucrats and legislators who supported and passed this law is that they did so in response to a handful of "bad outcomes" to home births in the last year or so.  Let's not mince words, "bad outcomes" means perinatal or prenatal death, or something equally horrible for the mother.  But what I discovered that there was the typical rush to "do something" to be seen addressing the problem, and nobody stopped to determine whether these outcomes were representative of a disturbing trend or were simply a statistical clump.  I kept asking the question, and nobody knew.

So I started researching the topic and I discovered another problem:  once HB 194 became law, for midwives or the OB/GYNs who often consult with them (whether they are officially licensed or not) to talk about their clinical experiences was to risk investigation, arrest, and prosecution.  You see, in the wake of the Earl Bradley scandal, in 2010 the General Assembly passed nine different bills aiming at protecting patients, including one that imposes an ironclad "duty to report" requirement on physicians that intersects with HB 194 to make it impossible to actually have a conversation about clinical practices in home birth. 

But I'm not a physician nor any form of licensed medical practitioner, so guess what?  Under statute I have no duty to report.  I can talk to anybody, especially about events that occurred prior to 2010, and they can talk to me.

So if you are interested in knowing what nobody else in Delaware will tell you about the extent and safety of home birth, go below the fold and meet me in my academic guise ...

Home Birth in Delaware:  2000-2010
A Very Preliminary Statistical Analysis

Steven H. Newton
Professor of History and Political Science
Delaware State University

Abstract:  While the Delaware Department of Public Health collects statistics on the location of birth (hospital, birthing center, or home), and on perinatal and fetal mortality, there are significant obstacles to comparing of the safety of home birth to hospital birth. 
Nonetheless, statistics recorded by DPH for home births, supplemented by national statistics reported by the Center for Disease Control and Prevention and the private records of several midwives provided to this researcher make it possible to build a reasonably accurate statistical picture of both the risks and outcomes of home birth in Delaware between 2000-2010.
The results indicate non-nurse midwives dismiss about 10% of their patients from care well prior to labor as having pregnancies that are too high risk for home birth, and that about 15% of the labors attended by non-nurse midwives end up in transportation to a hospital or birthing center.  Of the transported patients during this period there are no consistent records available showing bad outcomes either to mother or child; most ended in a successful C-section.  About 85% of the clients of non-nurse midwives experience safe home birth with minimal or no complications for mother or child.
This research also suggests that non-nurse midwives have been consulting regularly with other medical professionals, even in the absence of collaborative agreements as required by State law.  This appears to occur in 75% of all cases; the medical professionals most likely to be consulted were (in descending order of frequency) OB/GYNs, Neonatologists, Nurse Midwives, and Primary Care Physicians.
During the decade examined, there were 812 reported home births in Delaware.  This number does not include those patients transferred to another facility, as the outcome would then be reported under the statistics from that facility.  However, it is possible to project that there were 978 incidents of labor that began as home births during the period.
In 2010 Delaware’s fetal mortality rate for all births was 5.7 per 1,000 [there were 62 recorded deaths].  Based on this rate, if the fetal mortality rate for home births was the same as for all births, we should expect to see about five fetal deaths in home births during the period.  This research finds only two, but is admittedly incomplete.  Nonetheless there is no evidence currently available to suggest that over the decade from 2000-2010 home birth in Delaware was statistically any more dangerous than hospital birth.

Home Birth in Delaware:  2000-2010
A Statistical Analysis

The Sources

Five major sources were used in the compilation of this research:

1.     Birth and infant mortality records compiled by the Delaware Division of Public Health for the period 2000-2010 [accessed at ].

2.     The January 2012 CDCP Report on Home Birth in the US from 1990-2009 [accessed at ].

3.     The study published in the Canadian Medical Association Journal in 2009 comparing birth outcomes for mothers and children between home and hospital births from 2000-2004 [accessed at ].

4.     Access to the confidential records of several non-nurse midwives operating in Delaware during the period under study.

5.     Interviews with non-nurse midwives and OB/GYNs operating in Delaware during the period under study.

Limitations of the sources

1.     The DPH records list 812 home births during the period under study.  However, this statistic includes (but does not distinguish between) accidental and planned home births.  Moreover, the statistics tell us nothing about who (if anyone) attended these births.  Nor do they reveal how many women may have initially been clients of non-nurse midwives and have been dismissed as high-risk or non-compliant before labor; how many women gave birth in a hospital after beginning labor at home and undergoing transport; or how many women left the State to give birth either in a private setting or a religiously affiliated birth center in Pennsylvania or Maryland.  Moreover, these statistics are completely silent on adverse outcomes in home births or hospital births occurring after transport from a home birth setting.  In other words, the 812 home births reported over the decade are likely to undercount the number of women who attempted to have a home birth.

2.     The CDCP study provides some national background and useful trends for comparison to Delaware statistics.  It is, however, a difficult study to use for direct, detailed comparison, as conditions, policies, and laws differ so broadly across the nation as to make such comparisons tricky.  To cite a single example:  According to the CDCP, midwives nationally attend 62% of home births.  But the CDCP breaks down that 62% into 19% Nurse Midwives and 43% non-nurse midwives.  This is clearly not an applicable statistic for Delaware, where Nurse Midwives do not attend home births.  This data must be used with care.

3.     The CMAJ study is well-crafted and returns very useful numbers regarding the risk/benefit of home births as compared to hospital births.  The problem here is that the CMAJ study muddies the waters (at least from a Delaware perspective) by placing what we, in the US, would consider both Nurse and non-nurse midwives into the same category, and not differentiating between them in terms of outcome.  This does not, however, reduce the usefulness of this study (and the CDCP study above) in examining the profiles of the women who choose home birth and the risk level of their pregnancies.

4.     The confidential records provided by several midwives for this study are both invaluable for understanding the realities of home birth in Delaware and a problem in terms of their reliability.  These records were provided to this researcher in a format that removed all names and identifying data of both patients and other medical professionals involved in these births.  They were formatted consistent with NARM (National Association of Registered Midwives) standards, and designed so that they could be incorporated into public health or other medical records if such became legal.  However, these records were provided anonymously to this researcher, and could not be directly validated by any medical professional without triggering the statutory “duty to report.”  As such, analysis done and conclusions drawn from these records depend on (a) their internal accuracy and reliability; and (b) this researcher’s accuracy and honesty in abstracting those statistics.   With respect to these limitations, two comments should be made:

a.     Internal accuracy and reliability:  The records as provided showed evidence of being real records with details entered on an Excel Spreadsheet.  Formatting inconsistencies, differing abbreviations, and spelling errors were consistent with the transfer of raw records into the spreadsheet via “cut and paste” rather than being generated in a single sitting.  Moreover, when questioned at random about specific records (see below) the attending midwives were able quite quickly to recall case specifics and provide additional anecdotal information that was consistent with the case notes provided.  Given that these records involved dozens of births over a multiple year span, this would be very difficult to fake in a convincing manner.  It should also be noted that this researcher has an extensive background in medical-administrative records over a twenty-year career in the US Army, and that these records showed no glaring anomalies to indicate they had been faked or doctored.  Conversely, however, the inability of the researcher to produce the raw data for public examination must cause the information presented here to be discounted somewhat in terms of credibility.

b.     The researcher:  This researcher has nearly three decades of experience in dealing with quantitative research, two decades of experience in dealing with medical-administrative records, and a publishing history that includes nine books and over thirty-five articles in refereed journals.  This researcher also conducted the first statewide DHSS study (1999) for the Interagency Committee on Adoption on the availability of mental health providers to offer services for adopted children and their families; and served as the Co-Chair of the State of Delaware Social Studies Curriculum Frameworks Commission (1992-1995).

5.     The interviews were conducted with five different midwives, six women who have experienced home births attended by non-nurse midwives, and four OB/GYNs.  All of these interviews were conducted in strict confidentiality, with no identifying information about the individuals recorded.  The intent of the interviews was to challenge and/or validate the statistical material from the midwife records and the provisional conclusions of this study.  These interviews have been extraordinarily useful in clarifying certain questions regarding home birth, but their anonymity and the inability of other researchers to verify their content independently requires them to be viewed in the same light as the confidential records discussed above.

The Statistics
Reported home births in Delaware have declined slightly since 2000, principally due to the 2002 requirement that all non-nurse midwives hold a collaborative agreement with a physician to practice legally.  Since that requirement was enacted, only one non-nurse midwife in Delaware has been able to get such an agreement signed, and her agreement specifically limits her practice to Amish and Mennonite families.
DPH reports the following numbers and percentages of home births during the period:                 
2000:  93 (0.8% of total live births)
2001:  63 (0.6%)
2002:  76 (0.7%)
2003:  79 (0.7%)
2004:  70 (0.6%)
2005:  65 (0.6%)
2006:  80 (0.7%)
2007:  61 (0.5%)
2008:  61 (0.5%)
2009:  86 (0.8%)
2010:  78 (0.7%)
Total:  812 (0.67%)

Of particular interest in these statistics are two facts.  Against a national home birth percentage during the period of slightly less than 1%, Delaware maintained (even with policy restrictions) a home birth percentage of near 0.7%, and was listed by the CDCP as one of the states in which home births held steady or increased over the past twenty years.
The second fact, however, is an indicator that the legislation requiring collaborative agreements between physicians and non-nurse midwives did depress what would likely have been a rapidly growing number of home births in Delaware, rather than a static figure.  In 2000, in Kent County, a staggering 3.8% of all births recorded for the year were home births; this number stands at four times the national average, and also stands out as an anomaly in Delaware records for the first decade of this century.
But this figure of 812 home births is clearly misleading.  It excludes or conflates all of the following:
1.     Women who began the process of having home birth under a non-nurse midwife’s care, and then were dismissed by that midwife either for non-compliance with instructions or high risk;
2.     Women who had accidental as opposed to planned home births;
3.     Women whose complications during labor and delivery resulted in transport to a medical facility that later recorded the birth;
4.     Women who had planned home births either unassisted or assisted only by lay individuals and not midwives.

Can we make educated guesses about the numbers of women in each category?
This is where the confidential records and interviews help us understand the profile of home birth in Delaware.  The records examined included the complete records of two non-nurse midwives across a two-year period, and encompass forty-five different cases.  Interviews with other midwives generally confirmed the same statistical outlines.  In addition, one other midwife provided partial records from another year that also confirmed the suggested picture.
It appears that approximately 10% of the pregnant women who initially became clients of non-nurse midwives left their care well prior to labor.  About 1% were dismissed by the midwives themselves for failure to comply with instructions.  The remainder were dismissed by the non-nurse midwives to the care of hospitals or birth centers because they identified factors associated with the pregnancy that indicated too much risk for safe home birth.
About 15% of patients attended by midwives in a home birth setting began labor but were later transported to a hospital or birthing center.  The breakdown would indicate that about 2% were sent to another setting very early in labor when the midwife noticed a serious problem.  Another 11% were sent to other settings based on “exhaustion” or “failure to descend,” and about 2% were transported due to serious complications that developed during labor.  We will discuss the outcomes of these cases later.
About 85% of the patients who enter labor in a home setting under the care of a non-nurse midwife deliver without major complications for mother or child.
National statistics reported by CDCP argue that 5% of home births were accidental, and that individuals other than midwives or other medical professionals attend 33% of home births.
Interviews with midwives, medical professionals, and home birth parents suggests that, on an annual basis, about 15 Delaware women from Amish or Mennonite communities choose to travel to religiously affiliated birthing centers in Maryland or Pennsylvania, who would otherwise have chosen to have a home birth if the services of a midwife were available to them legally.
This admittedly incomplete data suggests that instead of 812 home births in Delaware between 2000-2010, probably about 1,025 women entered the care of a non-nurse midwife with the intent of having a home birth or planned to have a home birth unassisted by a midwife.  Once those who left the midwife’s care for any reason prior to labor are removed, about 978 entered labor in a home birth setting, and of those 606 entered labor under the care of a non-nurse midwife, 49 entered labor at home by accident, and 323 were either unassisted (by plan) or assisted by lay people.
Of the 606 cases in which the patient entered labor under the care of a non-nurse midwife, statistics and interviews would suggest that 91 were eventually transferred to hospitals or birthing centers, while 515 experienced a successful home birth with minimal or no complications.
What were the causes and outcomes for the women who had to be transported out of a home birth setting during labor?  Overwhelmingly, midwife records list either “failure to descend” and/or “exhaustion” as the reason for transport [90%], and a much smaller number of cases [10%] where serious complications developed during labor.  This would mean that about 7-9 women annually were transported for “failure to descend,” and perhaps 2 women every 3 years for serious complications.  Again, these numbers refer ONLY to women actually under the care of a non-nurse midwife at the beginning of labor; of the 323 women we suspect underwent home birth unassisted by a midwife, there were certainly many other incidents of transport, but these are currently impossible to quantify.
What about the outcomes?  The midwife records listed outcomes for all cases, including transports.  The majority of transports (80%) ended up in successful C-section births; the record is unclear regarding complications for the mothers.  About 10% of midwife transports ended up in successful VBAC births, with almost all the rest resulting in simple vaginal deliveries.
Between the records and the interviews this research has thus far only documented two cases between 2000-2010 in which the transport of women in labor from a non-nurse midwife’s care resulted in fetal death.  There is no direct indication that these two deaths either occurred because of, or in spite of, anything the midwife did or did not do, could or could not do in a home birth setting.  We simply do not know at this point.  We have no reliable statistics on bad outcomes for the women themselves in these situations.
One completely unexpected result of this inquiry was the high number of consultations taking place between non-nurse midwives and other medical professionals during pregnancy and even labor.  Slightly over 80% of the case records examined showed evidence of consults with primarily OB/GYNs and also Neonatologists, Nurse Midwives, and Primary Care Physicians.  Many of these consults occurred on multiple occasions, and several are documented as having occurred during labor.  Interviews with other midwives, home birth parents, and several OB/GYNs (speaking with assurances of complete confidentiality) confirmed that these consultations do occur on a regular basis, regardless of law and public policy.


1.     Of an estimated 978 women in Delaware between 2000-2010 who entered labor in a home birth setting, 606 of them did so under the care of a non-nurse midwife.  Of those, 515 had a successful home birth experience, and 82 others experienced a successful birth via C-section, VBAC or SVD after transport.  About 9 women encountered serious complications during or after transport, and there were at least 2 fetal deaths.  Within the limitations of the information we currently have available, nothing suggests that on a statistical basis home birth under the care of a non-nurse midwife results in a higher level of bad outcomes than births in a hospital or birthing center.  A number of factors regarding the women who choose home birth covered in the CDCP and CMAJ studies may explain this.
2.     Non-nurse midwives in Delaware during the period appear neither to have been operating either irresponsibly or in a vacuum.  Women were regularly dismissed from care because their pregnancies included indicators of high risk.  Midwives routinely consulted other medical professionals.  There is no evidence to suggest that midwives as a group tended to wait too long after complications appeared in delivery before requesting transport.
3.     That there were 323 home births not attended by any qualified medical professional, and that these births would arguably have resulted in at least the same percentage of transports (arguably more), suggests the possibility for confusion on the part of medical professionals and public health authorities with respect to the statistical sources of complications and bad outcomes.  It is effectively impossible, from a hospital or birth center perspective, to distinguish between transports of the patients of midwives and transports of patients having unassisted home births.  This is due to the fact that the patients uniformly report their labor as “unassisted” even when a midwife is involved, due to current statute and policy.
4.     Many OB/GYNs in Delaware were willing during the period to continue consulting with midwives despite the fact that they knew that these non-nurse midwives were acting outside statute and policy.  On the other hand, none of the OB/GYNs who took the career risk of doing so were willing to sign formal collaborative agreements with the midwives with whom they communicated.
5.     This researchers suspects but cannot prove or even assert with statistical confidence (yet), that the greatest health risk in home birth in Delaware is not labor attended by non-nurse midwives, but unassisted labor.  Most if not all non-nurse midwives have ceased operating since the passage of HB 194 and the invoking of stiffer penalties for violating the statutes, and several reported anecdotally that the number of women determined to experience home birth—with or without assistance—has remained steady.  What this means is that without a revision to current statutes to make midwife care attainable for these women, Delaware can probably expect to see a spike in bad outcomes from home births over the next 4-5 years.  If this assessment is correct, such a spike is deplorable because it is completely preventable.


NCSDad said...

Nice work Steve.
I was born in an apartment without physician.
How many times and ways does this happen in every legislature every year. It's corrosive.

Doug said...

I'll be there tonight with the Wolf. This time it looks like we sit on the same side of the table so to speak. Good write up, I would suggest we look to the Bill's primary sponsor for a clue. Might want to look at her donors too.

Quite a bill for the party of "choice in Women's Health Care" ain't it?

tom said...

I realize that the numbers you are working with may be too small to be statistically reliable. But can you show any data on how the rates of "bad outcomes" compare to those of in-hospital births?

It is my understanding that while Delaware's infant mortality rate has been improving steadily over the past decade, it is still among the worst in the nation.

With 99+ percent of our births occurring in hospitals, it is difficult to believe that home births could be responsible, nor is there wiggle room to say that, if hospitals see a higher percentage of bad outcomes, it is because they deal with all of the difficult cases while most of the routine births happen elsewhere.