0:02

This is Thinking About OPGYN

0:04

, with your hosts Antonia Roberts

0:06

and Howard Harrell .

0:17

Antonia .

0:18

Howard .

0:19

What are we thinking about o today's episode ?

0:22

Well , we're going to talk about the safety of

0:24

rural versus urban hospitals and

0:27

then the rates of vaginal cuff dehiscence

0:29

with different routes of hysterectomy . And

0:31

then we're ambitious . We've got a lot of different things to

0:33

talk about . Then we're also going to talk about new

0:35

evidence about salpingectomy to prevent

0:37

ovarian cancer . And then there was also

0:40

a nice analogy between lab

0:42

tests and research studies in a journal

0:44

, so we'll try to talk about that too .

0:46

Ambition .

0:47

Yes , but first what's the thing we do for no reason , Howard

0:49

.

0:49

Okay . Well , how about screening asymptomatic

0:51

women for bacterial vaginosis in

0:54

an effort to prevent preterm labor ? Or , for

0:56

that matter , even screening women

0:58

who come to triage We've had some contractions

1:00

perhaps with a vaginitis probe or a

1:02

wet mount to look for bacterial vaginosis

1:04

as a routine workup for their preterm

1:06

contractions .

1:07

Antonia , okay , sure , I have certainly seen

1:09

that latter scenario much more often

1:11

and I'm guessing it might be more a common place

1:14

than just screening every

1:16

pregnant woman without symptoms . But I could

1:18

be wrong . But we do know that bacterial

1:21

vaginosis is a risk factor

1:23

for preterm birth . It increases that

1:25

risk anywhere from two-fold

1:27

up to seven-fold depending on the

1:29

gestational age at diagnosis . Essentially

1:32

, the earlier you diagnose it , the higher risk

1:34

for preterm birth , and we've

1:37

known this for a long time . But the question

1:39

all along has been whether

1:41

or not screening for it and treating for

1:43

it does anything to reduce

1:45

that risk . So , in other words , is

1:47

this like a modifiable risk factor

1:50

or is it more like a

1:52

sign of some common , underlying

1:54

common denominator , like a predisposition

1:57

towards both preterm labor

1:59

and towards BV , maybe

2:01

like an inflammatory state or immune dysfunction

2:04

kind of thing where you could treat the BV but

2:06

it wouldn't affect that underlying common

2:08

denominator ?

2:10

Yeah , another example might be pre-dentition

2:12

. We also know that that's a significant

2:14

risk factor for preterm birth , but it turns

2:16

out that just treating poor dentition

2:18

during pregnancy doesn't do anything to change

2:21

the patient's risk for preterm birth , even

2:23

though we tried that for a while . And much of the same

2:25

story seems to be true for bacterial vaginosis

2:27

. Poor dentition and BV may

2:29

both be signs , as you said , of some underlying

2:32

immune issue , some immune deficiency

2:34

or poor immune response that's also

2:36

associated with preterm birth , which

2:39

has an inflammatory or perhaps infectious

2:41

etiology as well , rather than

2:43

a direct causation .

2:44

Yeah , so it's the same old correlation

2:46

doesn't equal causation debate

2:49

. This is something that goes back and forth

2:51

in the OB literature , though there's been at

2:53

least three meta-analyses looking

2:56

at whether or not screening and treating

2:58

for BV in pregnancy has any

3:00

impact on preterm labor , and

3:02

depending on which studies you choose to include

3:05

or not , then you see different results

3:07

and different conclusions . One review

3:09

by Lamont looked at five

3:12

studies and they reported a

3:14

benefit of screening and treating

3:16

with clindamycin . But then the Cochran

3:19

database looked at 21 studies

3:21

and based on the entirety

3:24

of those , they did not recommend screening

3:26

. And this was just in Cochran's most

3:29

recent review of this topic . And even more

3:31

recently , the US Preventative Services Task

3:33

Force looked at 48

3:35

studies , and that included a

3:37

large one that came after

3:39

those other reviews I just mentioned

3:41

, and they also concluded , based

3:44

on all those 48 studies , that there's no

3:46

benefit for screening or detecting

3:48

or treating asymptomatic BV

3:50

in the general pregnant population . And

3:53

then looking specifically at high risk women

3:55

, like women who have previously had a preterm

3:57

delivery , even for them the data

3:59

was inconclusive .

4:01

Right . So , despite all what you just

4:03

said , it's still a pretty common practice

4:05

that folks will screen or screen

4:07

at least patients who come to triage . I think that OB

4:09

residents are taught fairly routinely

4:11

to think of bacterial vaginosis as a cause

4:14

of preterm contractions or preterm labor

4:16

, rather than just an associated risk factor

4:18

. And in the spirit of having a thorough differential diagnosis

4:20

or maybe just wanting to treat something

4:23

, to find something wrong with the patient

4:25

and get the pain out of triage , they're expected

4:27

to look for it in patient who present to labor

4:29

and delivery , complaining of some contractions

4:31

even if they're not found to be in preterm labor

4:34

. In other words , they just have preterm contractions . And

4:36

yeah , some might even be taught to screen for

4:38

BV in a new OB visit since

4:40

, as you said , the earlier it's detected , the bigger

4:42

the risk for preterm labor . So

4:44

find it early and treat it . But the best evidence

4:47

says that treating it does not help .

4:49

Yeah , this is an example of something

4:51

where there's mixed data or maybe

4:53

contradictory results or guidelines

4:55

, again , depending on which studies you

4:58

look at . So some people just decide

5:00

well , if it's inconclusive , I'll err on the

5:02

side of doing more because it might

5:04

help someone , whereas on the flip

5:06

side , other people will take it as clear

5:09

evidence of no benefit , like , if it's not

5:11

conclusive , then it's not beneficial .

5:13

Yeah , there's a bias either to

5:15

do , I think , too much , which we call the commission

5:17

bias , or a bias to do too little

5:20

, which is an emission bias , and people

5:22

fall out on one side of the other of this . And

5:24

this is especially true , this presence

5:27

of a commission bias , I think , with a

5:29

lot of the things surrounding preterm

5:31

labor that we've discussed many times before , including

5:33

things like the tocolytics we talked about , progesterone

5:36

or bedrest activity restriction , pelvic

5:38

rest , lots of other things where the data

5:41

simply says it doesn't work . But if there's

5:43

ever been a single report or speculation

5:45

or plausible physiologic mechanism

5:47

by which it might work , that people keep doing

5:50

it , thinking well , it's not harming and maybe

5:52

it'll help someone , even though it doesn't work in large

5:54

studies .

5:55

Yeah , lots of examples of this , this same

5:58

bias tendency . I think progesterone

6:00

for recurrent miscarriage , like

6:02

early in the first trimester , is another example

6:04

.

6:05

Yeah , definitely . Well , there's a new

6:07

trial in July 17 , 2023

6:10

, from a randomized controlled trial in France called

6:12

the ALTOP trial AUTOP

6:14

they have French clever titles there

6:16

. I think this ends up being now the largest

6:19

trial ever to look at this problem . So this was

6:21

done at 19 French Perinatal Centers over

6:23

a two and a half year period and low risk

6:25

women before 20 weeks gestation were enrolled

6:27

. Low risk means that they essentially didn't

6:29

have any prior preterm births or late

6:32

miscarriages something 1918 week

6:34

miscarriages . They were randomized in a one-to-one

6:36

manner to either have a BV screen

6:38

and then treatment based upon the results , or

6:40

to just usual care . They used a PCR

6:42

test and overall they found that the

6:44

screen and treat approach was not effective

6:47

at reducing the rates of preterm birth in that population

6:49

. Now , as usual in a breakout

6:51

group here the small subgroup they suggested

6:53

there might still be a benefit among nulliparous

6:55

patients and unfortunately this just

6:57

kicks the ball down the field a little bit more

6:59

when we draw again inappropriate conclusions

7:02

from underpowered subset groups . But they

7:04

didn't handle this well . They didn't claim to have

7:06

found something , they just said that might

7:08

be an area where another study could be

7:10

done .

7:11

Yeah , that's fine to pull out

7:13

tiny subgroups . It's appropriate

7:15

actually , I think , to look at data if you do

7:18

a study and then determine just where

7:20

you think the new research directions might

7:22

be . But no one should take this

7:25

negative study about screening

7:27

and treating BV in pregnancy and decide that

7:29

they're just going to do it in nulliparous

7:31

patients rather than low risk

7:34

multips , especially when we have

7:36

so many other negative studies already

7:38

. The assumption should be that this comment

7:40

they make on the nulliparous patients

7:42

is perhaps an idea

7:45

or even an unverified conclusion

7:47

in an underpowered subgroup

7:49

, not that it's any kind of hard

7:51

or even moderately

7:54

reasonable scientific evidence that

7:56

would direct treatment .

7:58

I do think we confuse a lot me

8:00

being the medical community confuse

8:02

a lot what is preliminary hypothesis

8:04

generating types of studies or

8:07

types of analyses and what is

8:09

tends to be hypothesis confirming

8:11

types of studies . So almost all

8:14

retrospective or cross-sectional

8:17

type data , survey data

8:19

, risk factor association type data

8:21

. All that is hypothesis generating

8:23

and almost never should it inform

8:25

actual clinical practice . And

8:27

the same thing is true for the subsets

8:30

and different kind of analysis of stratified

8:32

data in studies like this . It is hypothesis

8:35

generating . It's not conclusion forming

8:37

, if you will . But yeah , we see this mistake all

8:39

the time when people take an underpowered

8:41

subgroup or data from a subgroup

8:44

that wasn't intended to find some result

8:46

and run with it and act like it is

8:48

the truth all of a sudden . But the conclusion

8:50

here is that we should not be routinely screening

8:53

and treating asymptomatic women in an effort

8:55

to reduce bacterial vaginosis . The data

8:57

just doesn't support it in either group and until

8:59

new studies do that's what it says

9:01

it's most likely that , as we said

9:03

before , the bacterial vaginosis is not the direct

9:05

cause of preterm labor , but that it represents

9:08

something different about that particular woman's

9:10

again I don't know immune system or

9:12

susceptibility to inflammatory processes

9:14

, something like that . So that bacterial

9:16

vaginosis and poor dentition for that matter

9:18

, and probably other similar things

9:21

and preterm labor are both symptoms

9:23

of an underlying common pathology

9:25

.

9:26

All right then , but what about the women who

9:28

present with symptoms

9:30

of preterm labor ? If they come

9:32

into the hospital contracting

9:35

, does that mean that they could be symptomatic

9:37

of possibly having BV

9:40

? Because we just mentioned the residents doing

9:42

this test a lot on these women .

9:44

I think that's the argument sometimes , but

9:46

no , I would say that those women who have preterm

9:49

contractions that that is not a symptom

9:51

of bacterial vaginosis , the symptoms of

9:53

bacterial vaginosis . When we say symptomatic

9:55

, we don't mean uterine contractions . Symptoms

9:57

are a foul-smelling vaginal discharge

10:00

. So no , there's no evidence that we should

10:02

be routinely screening women for BV when

10:04

they present with preterm contractions , unless they

10:06

also , by the way , complain of foul-smelling

10:09

or irritating discharge .

10:10

All right , then let's move on . I

10:12

know you've been doing some work in

10:14

Tennessee trying to advocate

10:16

for shoring up resources and access

10:19

to rural hospitals that do obstetrics

10:21

and also trying to help reverse

10:23

the trend of the number of maternity

10:26

deserts that exist . We've talked

10:28

about maternity deserts before . Well , you

10:30

probably saw there's a new article in

10:32

the JAMA Health Forum that looks

10:35

at the risk of severe maternal morbidity

10:37

in patients who gave birth at rural

10:39

versus urban US hospitals

10:42

.

10:42

Yeah , this is another study derived

10:44

from a database we've talked about before

10:46

involving about 11 million births

10:49

, and they found that the risk of severe

10:51

maternal morbidity was elevated for

10:53

women who gave birth at rural hospitals

10:55

that had fewer than 460

10:57

births per year . Now , of course there's a range

11:00

it's not a black and white cut off at exactly

11:02

460 , but the highest risk was at

11:04

hospitals with fewer than 110

11:06

births per year , and the maternal risks incrementally

11:08

decreased as the hospital had

11:10

increasing volume . Up to 240

11:13

births was a bit safer than less than 110

11:15

, and up to 460 births was still

11:18

safer yet , but the safest was a hospital

11:20

with more than 460 births a year

11:22

, at least the way they stratified the data in

11:24

a rural setting . There was a pretty wide range

11:26

of outcomes , though Some small hospitals provided

11:29

a high standard of excellent outcomes , while some

11:31

large hospitals appeared to have more poor outcomes

11:33

. So there are obviously a lot of variables , but

11:35

on the whole , on average , the lower the volume

11:37

at rural hospitals , the higher the

11:39

significant morbidity and mortality rates

11:42

.

11:42

The fascinating thing about what

11:44

they found was that this

11:46

was not the trend in urban hospitals

11:49

, and they had to compare this in

11:51

several ways . So in

11:53

one of their charts they included all

11:55

urban hospitals within this data set of

11:58

11 million births , and for that

12:00

they had to use different definitions

12:02

for low , medium , high volume

12:04

deliveries than what we just listed

12:06

for rural , because urban hospitals

12:08

they range on a much larger scale . You won't

12:11

find a rural hospital that delivers

12:13

like 10,000 babies per year

12:15

, but there are urban hospitals that do so

12:17

. For urban they considered high volume

12:19

to be more than 2,000

12:22

births a year , and remember rural was 460

12:24

. For urban , low volume

12:26

was anything up to 500

12:29

deliveries , and then they had medium category

12:31

500 to 1,000 , and then a medium

12:34

high category 1,000

12:36

to 2,000 . So in this

12:38

comparison the very highest volume

12:40

for a rural hospital would have been

12:43

equivalent to the very lowest volume

12:45

urban hospital category . So

12:47

in a way that's useful because it captures all

12:49

the urban hospitals , but

12:52

in another way it makes it harder

12:54

to draw conclusions . So they

12:56

did also find some urban hospitals

12:58

that had very few deliveries , including

13:00

some that had less than 110 per year , and

13:03

so then they made a separate comparison all

13:05

the rural hospitals compared to just those

13:08

urban ones that had very

13:10

relatively low volume , so that they

13:12

actually matched up the same delivery numbers

13:14

per category . So there was a less than 110

13:16

category , up to 240 , up

13:18

to 460 , and so on .

13:20

Well , that probably represents a very small

13:22

minority of urban hospitals , of course . So again

13:24

, maybe difficult to draw firm conclusions

13:27

from the smaller comparison . Generally

13:29

, there's fundamental differences in how

13:31

hospitals are resourced and how

13:33

they operate , based upon their location

13:36

. Rural and urban hospitals have different resources

13:38

, different staffs , different things . So even

13:40

if the number categories were the same , we

13:42

might still be comparing apples to oranges .

13:44

Yeah , that's true and I think they do discuss

13:46

that . But I think it's nice that they compared

13:49

it in both ways so you can at least get

13:51

a glimpse of what the differences are when we

13:53

actually arc matching up the delivery

13:55

volumes . Maybe

13:58

it was an attempt to head off those

14:00

criticisms about how you wouldn't be able

14:02

to compare them to rural hospital

14:04

460 deliveries to an urban hospital

14:06

that has 10,000 deliveries . But you

14:08

are right , the total number of patients

14:11

in that smaller urban comparison

14:13

was less than the

14:15

patients at the rural hospitals by at least

14:17

100,000 women . So

14:20

maybe not enough patients to even

14:22

draw proper conclusion , like you said . But

14:24

taking all those caveats into

14:26

consideration , this smaller comparison

14:29

with the smaller urban hospitals found that

14:31

only that lowest volume category

14:34

of less than 110

14:36

deliveries per year , only that category

14:38

had more severe maternal

14:40

morbidity than those other groups . And even then

14:43

it wasn't as big of a difference

14:47

as that linear trend that you described

14:49

with the rural ones , where it was the worst

14:51

with the least volume and linearly

14:53

decreased with increasing volume . So

14:55

I think they're at least trying to suggest

14:58

here that rural hospitals

15:00

are specifically the ones that

15:02

suffer from having a low delivery volume

15:04

, not just any hospital anywhere

15:06

that happens to have a low volume . They're

15:09

trying to show that it's just the rural ones .

15:11

Yeah , if anything we learned from that , it's maybe

15:13

that once your hospital's doing less

15:15

than 10 deliveries a month , that may be

15:17

a critical place where it's hard to maintain

15:19

a quality program , whether it's in a rural

15:22

setting or an urban setting . And it might

15:24

be tempting from all that to

15:26

conclude that urban hospitals

15:29

must be safer than rural hospitals

15:31

, but the overall rate of severe

15:33

maternal morbidities they're still very

15:35

comparable between the two . In fact , for the worst

15:37

groups so in other words the very lowest

15:39

volume hospitals , the rate of severe maternal

15:41

morbidity was the same at about 0.7%

15:44

, whether they were urban or rural . It's just that

15:46

further increase of delivery volumes

15:48

were associated with different trends in

15:50

maternal outcomes depending on rural or urban settings

15:53

. So while more than 460 deliveries

15:55

was the safest in the rural setting , with a

15:57

severe maternal morbidity rate of 0.47%

16:00

, the mid-lower volume centers were

16:02

actually even safer in the urban centers

16:04

at a rate of 0.4% , which

16:06

is about the same . That actually represents

16:08

a more drastic difference than with the rural hospitals

16:11

and in their discussion the authors attribute this

16:13

to a better ability for those mid

16:15

to low volume small urban hospitals

16:17

to refer patients out as needed , and

16:20

in this small comparison they didn't separate

16:22

out high versus low risk patients , like

16:24

they did in the main comparison . So it's very likely

16:26

that improved rates at the mid to low

16:28

urban hospitals represents

16:30

having less high risk patients . It may

16:32

be that the larger hospital all

16:35

the high risk patients go there , and the smaller mid

16:37

to low volume hospital is where low risk patients

16:39

aggregate . And the authors didn't actually discuss

16:41

at all how to make specifically the ultra

16:44

low volume urban hospitals any safer , even

16:46

though they had the same severe maternal

16:48

morbidity rates as the smallest rural hospitals

16:50

. Again , it just may be that under 110

16:52

, it's hard to do anything in either setting .

16:54

Yeah , and it is good that at

16:57

least with the comparison where they had

16:59

all of the urban hospitals

17:01

, even the 10,000 plus ones

17:03

, that they separated out high versus low

17:05

risk patients , because the high

17:08

risk patients will tend to get sent

17:10

out to those higher volume facilities and that would

17:12

definitely skew the morbidity rates if

17:14

we weren't paying attention to it . So when

17:16

we look at those patient

17:18

risk comparisons we

17:20

find in this study that specifically

17:22

the low risk rural patients

17:24

had the biggest disparities in

17:27

maternal morbidity outcomes based

17:29

on the delivery volume of where they had their

17:31

babies . So again , regardless

17:34

of the maternal risk status at urban hospitals

17:36

, there really were no differences

17:39

in those maternal morbidity

17:41

rates in that main comparison where

17:43

it was like 500 deliveries , 1000 , 2000

17:46

plus . So again we

17:48

have to take into account the caveats of

17:50

having the different delivery numbers being

17:52

different categories .

17:54

They discussed some of the possible reasons and solutions

17:56

for this disparity in maternal outcomes

17:58

, and obviously shutting down those

18:00

rural hospitals or taking resources away

18:03

from them would be even worse . So many

18:05

women would go from a suboptimally

18:07

resourced rural hospital

18:09

to a maternity desert with no hospital

18:11

or no medical care at all , at least within a safe

18:14

and reasonable distance from home . So the

18:16

authors call for more resources for rural

18:18

hospitals and mention a couple of programs like

18:20

the CMS , birthing friendly hospital and the

18:22

rural emergency hospital designations that are meant

18:24

to help improve those specific services

18:26

in rural hospitals , and there's a few other

18:28

suggestions I could easily spend at least the next

18:30

hour or more talking about .

18:32

We should get into some of that of what you've been working

18:34

on maybe not for a whole hour right now , but I

18:36

don't know maybe 30 minutes to talk about

18:39

that . But this finding of the low-risk

18:41

patients having worse outcomes does raise

18:43

questions about what's going on with them in the rural settings

18:45

. Maybe the most attention and resources

18:47

in those remote hospitals get pulled away

18:50

to the high-risk patients because they

18:52

didn't have that much of a disparity

18:54

in outcomes . And then in those small

18:56

hospitals with limited staffing , maybe

18:58

that detracts from routine care , detracts

19:01

from properly identifying when does a low-risk

19:03

patient actually become high-risk and then

19:05

she just goes on to be mislabeled as low-risk . Or

19:08

maybe they're just not that well prepared for emergencies

19:10

like hemorrhage or sepsis or the

19:12

things that would lead to severe maternal morbidity . And

19:14

then , regardless of what the process

19:17

is for this disparity , you have to

19:19

also wonder is it purely from a lack of resources

19:21

? Or is it also from a lack of volume

19:23

that leads to a lack of clinical experience

19:26

and judgment ? Because those very lowest volume

19:28

hospitals are also the least likely to have any

19:30

kind of training program , especially residency

19:32

training program . And if it's not enough volume

19:34

for a resident to achieve competence over

19:37

four years , then how do you ensure that it's

19:39

enough volume for an attending physician

19:41

to maintain competence year in and year

19:43

out ?

19:44

Well , of course those aren't one-to-one associations

19:46

, but I mean a residency program with

19:48

24 residents a year might

19:50

have 80 faculty and a

19:52

small hospital might have four or

19:54

five folks . So I don't know that correlates

19:57

to folks in rural areas having

19:59

less care or less care , but

20:01

it does probably get into if you're

20:03

a hospitalist or something and you're doing hundreds of deliveries

20:06

a year , versus 90 deliveries

20:08

a year for the average of a GYN . Maybe that

20:10

makes a difference .

20:11

Yeah . Or if someone is doing like

20:13

less than one delivery a week , for example

20:16

, that's yeah , right , yeah .

20:17

Which easily might be the case in these very

20:19

low hospitals . But I think for the most part stuff

20:22

scales . It's just that at a certain point it doesn't

20:24

scale Like . You need a certain number of nurses

20:26

per delivering patients per year and

20:28

you might work at a busier hospital , but instead of

20:31

being one of four or five doctors you're one

20:33

of 50 doctors and instead of being one

20:35

of 20 nurses you're one of 120

20:37

nurses . So at a certain point there is

20:39

a correct amount , a right amount , and it's hard

20:42

to meet that at the fringes . I think the two

20:44

busy hospitals may overwork folks

20:47

and their shortcuts taken and the

20:49

two small hospitals may not provide

20:51

enough experience . There is a hospital this

20:53

week that was cited by the state of California

20:55

in Los Angeles . It's one of the busiest hospitals

20:58

in Los Angeles . They were cited for a

21:00

patient who died after a C-section and they specifically

21:02

talked about basically being busy and taking

21:04

shortcuts . So they weren't doing quality improvement projects

21:07

, they didn't have a hemorrhage protocol , they

21:09

weren't staffed sufficiently to look for signs

21:11

of hemorrhage in the postoperative patient . So you can be too

21:14

busy and not have that right fit number , and

21:16

you can be just too slow and not

21:18

have the right fit number too or not have

21:20

the right clinical experience . But I do think residency

21:23

programs and educational programs

21:25

add a layer of protection

21:27

, if you will , in the sense of there's more eyes on

21:29

the problem , more discussion , et cetera .

21:31

Yeah , definitely there's several

21:33

additional layers of attention

21:36

to patients when you have residents

21:38

and four different year levels of residents

21:40

on your team and all of them are

21:42

learning and questioning things and watching

21:45

you closely and asking

21:47

what guideline was that from again . So

21:49

that's a little bit of built in , I

21:51

think , accountability . Now , obviously

21:54

the majority of hospitals are not

21:56

academic , so that's a rare luxury

21:58

, if you will . But certainly outside of the

22:00

academic setting , the physicians have

22:03

to put all of that rigor on themselves

22:05

voluntarily , because it's not built in for

22:07

them . There's no one that's following them

22:09

around and asking questions and mimicking

22:12

them and writing things up for a weekly

22:14

M&M conference or a weekly quiz

22:16

or anything like that . So I imagine , even

22:19

especially at a small , maybe poorly

22:21

resourced community hospital , even the most

22:23

conscientious doctor that's every day

22:25

thinking about am I doing things right

22:27

? What are the newest guidelines ? That doctor

22:30

may still have difficulties getting the time or the

22:32

support to go to conference and do their

22:34

maintenance training for whatever they're getting rusty

22:36

on , because their hospital is just so short

22:38

staffed and they can't go .

22:40

Yeah , I think it cuts both ways . I definitely think

22:42

all that's true , and physicians

22:44

in general are resistant to new

22:46

programs like the hemorrhage or

22:48

other hypertension bundles and things

22:50

like that that we're trying to put into hospitals . Recently

22:53

in Tennessee , we worked on

22:55

optimal cord clamping in the birthing

22:57

hospitals in Tennessee and yeah there

22:59

was a lot of resistance in small centers

23:01

, in some places in different sizes

23:04

, but in general , at the larger facilities

23:06

there was more acceptance and uptake , probably

23:08

just from peer pressure . If you're one person

23:10

in four who doesn't want to do optimal cord

23:13

clamping and you're a big personality

23:15

, then you probably get away with it , but if

23:17

you're the one outlier in 30 , there's

23:19

more peer pressure . So , and again , when

23:21

you add in trainees , whether it's medical

23:23

students or residents or whatever , those doctors

23:25

or students , student doctors make

23:28

a difference , those doctors in those

23:30

settings that don't have any academic program

23:32

at all , or students or whatever , then they , as you

23:34

said , they don't have folks watching them , they don't have

23:36

folks asking questions . There's less scrutiny , there's

23:38

less layers of accountability . There are

23:41

no perhaps set down rounds and checkout

23:43

rounds or M&M conferences

23:45

, things like that . So the burden is entirely

23:47

on you to keep up to practice and

23:50

up to date with evidence-based care

23:52

and stay current and change

23:54

without peer support . To change I mean

23:56

also changing in an isolated situation

23:59

without a lot of peers around you . Changing at

24:01

the same time or indicating this is the right thing

24:03

to do can be scary for people , so they stick with

24:05

what they know . And if you do have

24:07

something go wrong and a case is reviewed by

24:10

someone internally , well , it may be

24:12

reviewed by a friend of theirs

24:14

it's a very small department or

24:16

somebody who just doesn't have any more

24:18

expertise than they do . So there's all kinds

24:20

of variation in how hospitals and departments handle

24:23

continuing education and clinical quality

24:25

, and those things tend to be more formalized at

24:27

larger centers . On the flip side , as

24:30

I said , if hospitals get too big , like this LA

24:32

hospital , there's probably a loss of

24:34

individualized care and certainly continuity

24:36

of care , and , yes , more learners are

24:38

involved in the mix , and things , though

24:40

, because of that , tend to become more like a

24:42

depersonalized assembly line and

24:45

they can get understaffed , which is probably

24:47

the case with this LA hospital . That's not

24:49

just a rule problem , especially when they have volume

24:51

surges . So there may be a sweet spot

24:53

beyond which more deliveries at

24:55

least per physician , per nurse

24:58

, per staffing unit or bed

25:00

size or whatever is no longer make

25:02

, no longer makes things better . But in this

25:05

data , only when they did not

25:07

adjust for patient clinical characteristics

25:09

severe maternal morbidity was consistently

25:11

lower in the medium delivery groups , 500

25:13

to 1000 . In other groups it was maybe

25:15

lower for low risk patients in one category

25:17

, then only lower for high risk patients . In another

25:20

, but again only in unadjusted data , and

25:22

these patient characteristics , adjusted

25:24

for , included maternal age , race , education

25:26

status , insurance status and a co-morbidity

25:28

score . They didn't include a breakdown of how

25:30

these characteristics related , though , to

25:32

severe maternal morbidity outcomes

25:35

.

25:35

Yeah , but they did show separate

25:38

breakdowns of these patient characteristics

25:40

that they made their adjustments with in each

25:43

of the urban versus rural

25:45

states that they looked at and

25:47

in each separate hospital

25:49

volume category within each of those

25:51

states . But it sounds like

25:54

those characteristics at any given

25:56

hospital are something that can be tracked but

25:58

they can't control .

26:00

Yeah , they did point out that the Medicaid mix

26:02

, of course , is much greater in the

26:04

rule access areas , along with your

26:07

inheriting in that environment patients with

26:09

poor social determinants , and of

26:11

course the implication of that

26:13

is that payment enhancements for lower volume hospitals

26:15

that primarily treat Medicaid patients

26:18

could help address some of these resource constraints

26:20

and the availability of clinicians and

26:22

just the financial viability

26:24

that small hospitals face and staffing

26:26

and things like that .

26:28

And are those the sort of recommendations

26:30

you've been promoting for Tennessee ?

26:32

Yeah , I mean , I think the lesson here is that there

26:34

is a right size and that patients

26:36

need to be in the right center . So

26:39

there are many patients who the rural

26:41

hospital just serves as a transfer

26:43

point to get them , or their unborn

26:46

fetus that's at high risk for needing

26:48

a NICU or something , to the right place where the right

26:50

care can be delivered . But even for low risk

26:52

patients there's a right size , there's a right

26:54

number of deliveries per clinician

26:57

, per nurse , per whatever , and

26:59

you know that . I think that's what this gets at . And on the edges

27:01

of not being right sized you start to see

27:04

things missed . So we need to find

27:06

something workable and that isn't just

27:08

pushing rural patients into larger urban

27:10

hospitals that probably aren't suited to

27:12

handle that extra volume anyway . You might just

27:14

end up increasing the maternal morbidity

27:16

and depersonalized care and things like

27:18

that in those urban hospitals , in addition

27:21

to significantly disrupting the lives and communities

27:23

of those rural patients who are getting transferred to

27:25

that urban setting . So that just is a lose

27:27

lose situation . This data definitely supports

27:30

that . Even now , a rural hospital that

27:32

can deliver , say , between 460,000

27:35

babies a year can deliver world class care

27:37

that's comparable to that of any urban hospital

27:40

, if it's appropriately resourced and , of

27:42

course , if it appropriately transfers patients

27:44

that need other resources out of that hospital

27:46

to a tertiary center , and that's

27:48

true for both low risk and high risk

27:50

patient populations . The challenge , though , is

27:52

to improve outcomes for patients who currently

27:54

end up in the very lowest volume

27:56

hospitals , and if those are still

27:59

the closest facilities , within an hour

28:01

or two of where a patient might live , then the

28:03

answer still may not be to shut that hospital

28:05

down or make all the pregnant women

28:07

in that area travel to the city

28:09

for their care or their antenatal , probably

28:12

and delivery care . The answer is

28:14

to help that hospital get up to the level

28:16

of a higher volume that's needed

28:19

to maintain basic quality care

28:21

for basic obstetric safety .

28:23

So what are some specific things

28:25

that you're advocating for to make this problem

28:27

better ?

28:28

Well . As the article points out , medicaid

28:30

needs essentially to pay a premium for

28:32

women who deliver in a rural access hospital

28:35

. States could also step in with dollars

28:37

or tort reform measures

28:39

that help with liability concerns . These

28:41

small hospitals have to carry liability insurance

28:44

like everybody else does , and the financial aspects

28:46

surrounding that insurance can just be devastating

28:48

to the budgets . A reason to close the unit eventually

28:51

, and things like loan repayment that favors physicians

28:53

. Moving to more rural areas

28:56

after training rather than an urban area

28:58

, could help attract physicians and retain physicians

29:00

into areas that need them where there

29:02

are maternity deserts , and the same thing could be

29:04

done for nursing staff as well . Nurses

29:06

in many rural areas travel

29:08

to the nearest big city to work because of slightly

29:11

better payment or whatever . Unfortunately

29:13

, all of the payment structures

29:15

incentivize people to live in urban

29:18

areas . But state legislators do have some

29:20

control over these issues , and the federal government

29:22

does through the CMS program

29:24

, which right now unfortunately favors

29:26

higher reimbursement to urban areas and

29:28

lower reimbursement to rural areas . And

29:30

there's some sense of that , because it

29:33

costs more to run a hospital or

29:35

have a healthcare resource in an urban

29:37

area due to cost of living . But it's so skewed

29:40

that right now , for the most part , hospitals

29:42

are giving up on rural areas because

29:44

the payment is so much better . If you're going to build a new

29:46

service or maintain a service , the zip

29:49

code it's in determines how much money you make , and

29:51

that shouldn't be .

29:52

Yeah , I can see how that would be a

29:54

big challenge . I just went

29:56

through the process myself of looking

29:59

for where to settle down after completing the

30:01

military service and being able to pick

30:03

for myself where I want to live , and I saw

30:05

lots of jawbats for some

30:07

pretty rural locations that were advertising

30:10

massive salaries . It's not just about

30:12

the money , though , so they have to offer enough

30:14

money to overcome all these other

30:16

things that it's hard to compete

30:18

against a city for , like well , being

30:20

close to family and friends and good schools

30:22

and things you can do in a city , like

30:24

nice parks , gyms , whatever , shopping

30:27

kind of stuff that you might not find

30:29

all of those things in a rural area . So

30:31

, yeah , I could they get

30:33

less money , but then they have to offer more money just

30:35

to get someone to work there . Yeah , I can

30:37

see how that would be tricky , and obviously all

30:39

of that is probably outside the scope of

30:41

the city . It isn't .

30:43

It isn't . It isn't when you realize that

30:45

in rural Tennessee , for example , in

30:47

most of the counties that have hospitals

30:50

in Tennessee , the healthcare system

30:52

and hospital system is the largest employer

30:54

. So keeping folks there

30:57

who are in the healthcare services

30:59

nurses , et cetera , pharmacists

31:01

, physicians , whoever that have these

31:03

good , high paying jobs associated

31:05

with it , or is what builds a strong community

31:07

and the good schools and the tax spaces that

31:09

you're talking about . So we're , unfortunately , we're on

31:12

the other end of a decades

31:14

long redistribution of money

31:16

in the healthcare dollars at least , from

31:18

urban to rural areas , and so when a

31:20

rural hospital closes , it

31:22

devastates the local industries . Major

31:25

manufacturing plants and industries don't want to look

31:27

for communities to settle in that don't have

31:29

hospitals . So it's a vicious cycle . You've

31:31

got to have high quality healthcare

31:33

and a hospital in the community to attract

31:35

the jobs that increase the tax

31:37

basis , that build the parks and build

31:39

the schools et cetera . So CMS

31:41

does have control over this and they've contributed this

31:44

problem over the last 30 years .

31:46

Yeah , so they need to take notice and start

31:48

turning the tides . I know one of the other

31:50

short term fixes that happens right now

31:52

is where doctors will take a locum's

31:55

period , maybe even just a weekend

31:57

at a time or a couple of weeks at a time , to

31:59

fly out to somewhere and maybe live

32:01

in a rental or hotel and

32:03

just take call and then fly back to wherever

32:06

they're from . And obviously that has downsides

32:08

to it , but right , now ?

32:10

well , it's financially unsustainable

32:12

, though . I mean , there's so much more

32:14

expensive to hire that sort of physician

32:17

and they're also not gonna come in and be dedicated

32:19

to types of quality improvement projects

32:21

and things that stabilize the hospital

32:23

. They're just in and out and they charge a lot

32:25

for that service .

32:27

Yeah , it's like a bandaid , so yeah

32:29

, so that's not what we should be trying

32:31

to push for more of . But anyway

32:33

, if anyone , especially in Tennessee

32:35

, is looking to get involved with that sort of

32:37

advocacy , certainly they can email us

32:39

and we could help point them in the right direction

32:42

. But I'm sure many other states

32:44

in the US are working on similar projects

32:46

and I think if anyone even outside

32:48

of the US wants to tell us how they've

32:50

seen the rural obstetric care being

32:52

handled , that would be really interesting to hear about . So

32:55

let us know .

32:56

There are maternity deserts in urban areas

32:58

as well . This problem isn't just

33:00

actually with rural areas , and

33:02

the same sorts of problems is creating

33:04

some of those urban air maternity

33:07

deserts . So as rural volumes increasingly

33:09

shift into the city and the city , hospitals

33:12

become overwhelmed with new patients

33:14

and then city hospitals can also

33:16

are just financially they're consolidating

33:19

. The effect has been to create some very , very

33:21

high volume hospitals in

33:23

a city that are very , very , very busy . In

33:25

this paper , the highest volume hospitals had more than 10,000

33:27

deliveries a year and none of them had

33:30

severe maternal morbidity rates as

33:32

low as some of the hospitals in that 500,000

33:34

per year range . In a larger city , with

33:37

traffic and hospital consolidation

33:39

forcing everybody to one or two places

33:41

and then poor transportation issues

33:43

, a woman who lives in the city

33:46

or on the perimeter of the city might still find herself

33:48

an hour away from a hospital

33:50

where she can get obstetric care

33:52

in the event of an obstetric emergency

33:54

. And there are hospitals within the

33:56

big city that don't have OBGYNs

33:59

on staff because they've shifted it all to one place

34:01

. So she may go to the emergency department with

34:03

postpartum cardiomyopathy and not have the

34:05

ability to get a consultation from an OBGYN

34:07

because there are no OBGYNs there anymore

34:09

, and then , when

34:11

she gets to that big hospital that does 10,000

34:14

deliveries , she may find inattentive

34:16

care due to a lack of space and

34:18

just getting out fires and getting to the next

34:20

emergency , because those hospitals have

34:22

limited resources too . So that could create

34:24

a situation just as bad , if not worse

34:26

, than what many women in rural areas

34:29

face with downsizing , enclosures

34:31

and a lack of access . We should be promoting

34:33

access everywhere .

34:35

Yeah , that example you just brought up

34:37

with postpartum cardiomyopathy harkens

34:39

back to what we talked about a little while ago

34:41

with racial disparities , with

34:43

out-of-hospital maternal mortality

34:46

rates and also with severe morbidities

34:48

. And for that condition , especially

34:51

the cardiomyopathy , delayed diagnosis

34:53

worsens the outcome significantly

34:55

and it's going to be delayed . If it's being

34:57

assessed by someone who's not an obstetrician

35:00

, it's not on their mind to

35:02

look for that , and many of

35:04

the women in the inner city probably

35:06

are African American women who

35:09

are basically living in these urban obstetric

35:11

deserts , as you mentioned . So that might

35:13

also be a big reason for their disparities

35:16

in mortality and morbidity .

35:17

Yeah , one of the areas we've identified to help

35:19

with that postpartum cardiomyopathy

35:22

problem in particular to address is not

35:24

training OBGYNs , is training non-OBGYNs

35:27

who see these patients to recognize that

35:29

that woman needs care . And so

35:31

we know , that's true that these women are going

35:33

into urban maternity deserts

35:36

and not getting diagnosed .

35:37

Agreed , let's move on to a different one .

35:39

All right . Well , in the June 20th edition

35:42

of the Journal of the American Medical Association

35:44

, there was a couple of interesting editorials . One

35:46

is about opportunistic salpingectomy for

35:49

ovarian cancer prevention by Rebecca Stone

35:51

and her colleagues at Hopkins . The practice

35:53

of opportunistic salpingectomy

35:55

to prevent ovarian cancer really started , originally

35:58

based upon some observational data and

36:00

theory , but last year , in 2022

36:02

, there was a trial published by Canadian

36:05

researchers . That was really the first piece

36:07

of prospective evidence that we've

36:09

had that shows that this is a successful

36:11

strategy . So they followed 25,889

36:14

women who underwent opportunistic salp

36:16

injectomy and compared them to over

36:18

32,000 women who had either had

36:20

a hysterectomy alone or a tubal ligation

36:23

alone so a partial salp injectomy between

36:25

2008 and 2017 . And

36:28

at the time of follow-up they reported no

36:30

high-grade serous cancers and five or

36:32

fewer epithelial cancers in the

36:34

women who had undergone salp injectomy

36:36

. So a quick refresher serous is the most

36:38

common subset of epithelial , which

36:40

is the most common type overall among

36:43

ovarian cancers . Now , based on the

36:45

rates of cancers in the control group and

36:47

what's already previously been known about the incidence

36:50

of ovarian cancer , they would have expected

36:52

to see 5.2 compared

36:54

to zero and 8.6 compared

36:57

to under five cancers in those

36:59

two groups . So this is the first real-world

37:01

evidence that this approach is working

37:03

.

37:04

Yeah , I think we've been waiting for some

37:06

hard evidence on this and

37:08

I think it's invigorated the idea that salp

37:11

injectomy complete removal of

37:13

the fallopian tube should be done whenever possible

37:15

when a woman desires sterilization

37:17

and she's already getting a surgery where

37:20

that can be easily done . So if

37:22

just a hysterectomy , for example

37:24

leaving the fallopian tubes already

37:26

, is going to reduce the ovarian cancer risk

37:29

by about 30% , that would translate

37:31

to the average woman going from

37:33

like a 1.4% lifetime risk

37:35

to slightly less than 1% risk

37:37

after getting that procedure . The

37:39

suspected explanation

37:41

is there's less blood flow to the ovaries

37:43

in fallopian tubes or something . But

37:46

that's how that's . What's been observed is

37:48

about 30% risk reduction . But

37:50

this study now is

37:52

the first to really clearly show that , yes

37:54

, complete fallopian tube removal is even

37:57

better than just removing the uterus

37:59

or just removing part of the tubes . But

38:01

the reductions , the fact that

38:03

they still saw from eight expected

38:06

cancers down to five or less of

38:08

the epithelial ovarian , might

38:10

seem like a fractional or even maybe

38:13

even a disappointing improvement . But

38:15

we have to remember that even if you did a complete

38:18

oophorectomy , you wouldn't fully eliminate

38:20

the risk of ovarian cancer . In

38:23

that case someone still gets it . You might call it primary

38:25

peritoneal cancer instead , but it's essentially

38:27

the same entity . So either

38:30

way , we're always going to have to accept some residual

38:32

cancer risk of ovarian

38:35

cancer , and we already know that

38:37

doing an opportunistic oophorectomy

38:39

is net harmful in an average

38:41

risk asymptomatic younger

38:44

woman . So we're not talking about that

38:46

at all . That's not for debate . We're

38:48

talking about the more benign practice of removing

38:51

the entire fallopian tube in someone who

38:53

is otherwise going to get them tied off or

38:55

have a hysterectomy . She was already not

38:58

going to use them anymore to

39:00

get pregnant . So in real numbers , what

39:02

this found was that out of the 25,000-ish

39:06

women who got complete

39:08

self-injectomy in this study , maybe

39:10

about nine or 10 of them were

39:12

spared from developing ovarian

39:14

cancer because of that complete self-injectomy

39:17

over just a few years . Probably over

39:19

time more and more of them would

39:21

have been spared .

39:22

Yeah , that was just over . I think the average follow-up

39:24

was just a smidge under five years . So obviously

39:26

those numbers will scale as time goes

39:29

on . But the editorial , though

39:31

, raises an interesting idea . So while

39:33

OBGYNs have been looking to do

39:35

opportunistic salpingectomies at the time

39:38

of hysterectomy or for sterilization

39:40

for some time , women undergo lots

39:42

of other surgeries in the course of a year

39:44

for lots of other reasons , and opportunities

39:47

are being missed . So , for example , hundreds

39:49

of thousands of women in the US each year

39:51

undergo surgeries like cholecystectomies

39:53

or appendectomies or hernia repairs

39:56

or other intra-abdominal surgeries where

39:58

fallopian tubes could be removed if

40:00

the patient didn't desire future fertility

40:02

. And this broader reach , this

40:05

idea of a more expansive

40:07

program of prophylactic self-injectomy

40:10

, could save over time a couple of thousand

40:12

lives a year and significant money

40:14

, since more of these surgeries are being

40:16

done each year than hysterectomies anyway

40:18

. And the projection is that if we had full

40:20

uptake of this procedure at the time of hysterectomy

40:23

, just for that we could save 2,000 deaths

40:25

a year and half a billion dollars . So

40:27

we could do better than that if we added in

40:29

opportunistic self-injectomy at the time

40:31

of other general surgical procedures .

40:34

Yeah , especially when recently we've

40:36

heard about worldwide shortages and common

40:38

chemotherapy drugs . Saving

40:41

any ovarian cancer cases , especially

40:43

with something as easy as just fallopian

40:45

tube removal , would be great

40:47

. And it's an interesting idea to put

40:49

this on a non-gynecologist

40:52

because and in this study they

40:55

also pointed it out that there's no surgical

40:57

cancer prevention program right

40:59

now that encompasses multiple specialties

41:02

. So if someone

41:04

getting an appendectomy was

41:07

going to also be offered an opportunistic

41:10

, we would need the general

41:12

surgeons either to do that themselves

41:15

, so they would have to learn how to do it , and I'm

41:17

sure they would learn it very easily .

41:19

Not hard .

41:20

Or it's not hard , it's very easy , it's

41:22

like intern level , or we

41:24

would have to have a system where they call in

41:27

intraoperatively a gynecologist

41:29

who would come in and just do that portion of the surgery

41:31

. But in either case , with the primary

41:34

procedure being that general surgery

41:36

case , the patient would need to be appropriately

41:38

identified and counseled and educated

41:41

beforehand and offered

41:43

this procedure and had the risks

41:45

and benefits explained . And right now

41:48

, in many cases , especially depending

41:50

on insurance status , the

41:52

patient would need to sign these federal sterilization

41:55

papers that have a 30-day wait time

41:58

, and so that's not going to be doable for

42:00

any unscheduled procedures , like many

42:02

appendectomies or cholecystectomies are

42:04

, unfortunately .

42:05

Yeah , yeah . And they point out that having opportunistic

42:07

self-injectomy labeled as a

42:09

form of sterilization in our CPT codes

42:12

is a problem due to this legally

42:14

mandated waiting period for sterilization

42:16

. If it were coded as something else , like cancer

42:18

prevention , and it was clear that the surgery was being

42:20

done for that reason , then

42:22

that would dramatically improve those circumstances

42:25

and those opportunities . But I do think it's an interesting

42:27

idea and an opportunity for multi-specialty

42:30

collaboration . If the surgeries were done

42:32

at the time of other general surgeries , the cost

42:34

for a patient would be very small and likely

42:37

very cost effective , given the number of lives

42:39

and dollars saved over time by

42:41

preventing ovarian cancer .

42:43

Yeah , they wouldn't have to go through two separate surgeries

42:45

for two minor or

42:47

minimally invasive procedures . That could be done just

42:50

with the single OR day single

42:52

anesthesia . I'd be curious

42:54

to hear what general surgeons think

42:57

about this proposed collaboration

42:59

from this study because I'm sure

43:01

if the tables were turned let's say

43:04

it was shown that some kind of general

43:06

surgical procedure that was done

43:08

opportunistically , like an appendectomy

43:10

for example , let's say it was shown to have

43:12

a huge net benefit , then I'm sure

43:15

that many gynecologists would also

43:17

make a point of becoming proficient at that

43:19

as an add-on to whatever GYN

43:21

procedure they're doing , like let's do a hysterectomy

43:24

and then this opportunistic whatever . And

43:26

in fact they used to do appendectomies

43:29

commonly and I still see it

43:31

show up on different

43:33

privilege forms as an option for

43:35

me to tick off . Do I want to which ? I don't tick

43:37

them off , I haven't been trained in it .

43:39

But yeah , obgyns used to do over a million

43:41

appendectomies a year . We actually used to do more

43:43

than surgeons , because they would be done at every

43:45

hysterectomy and every C-section

43:47

, practically .

43:48

Yeah , yeah , because from my understanding

43:51

it was believed before that an

43:53

opportunistic appendectomy for

43:55

someone that doesn't have appendicitis or anything would

43:57

save them from future appendicitis . But

44:00

we don't do it now , we don't do it anymore because

44:02

there's extra surgical risk and complications

44:05

that you're exposing them to and

44:07

I guess the risk-benefit ratio

44:09

there is just not worth it . So anyway

44:11

, you said there was another editorial

44:13

in that same journal edition that you also

44:16

liked . Is this the analogy one ?

44:17

Yeah , so yeah , there's an editorial that discusses

44:20

the analogy between clinical trials

44:22

and diagnostic tests , and this is something I

44:24

talk a lot about . This was written to stimulate

44:26

conversations about what a negative controlled

44:29

trial means , and the editorial finds

44:31

it useful to liken clinical trials to

44:34

diagnostic tests . So clinicians use

44:36

diagnostic tests every day and are , hopefully

44:38

, intimately familiar with their strengths

44:40

, limitations , quirks , etc . Whereas they don't usually

44:42

have the same working knowledge or intimate understanding

44:45

of clinical trials . Now I love

44:47

to talk about this . In fact , the math and statistics

44:49

involved in the interpretation of clinical trials is

44:51

exactly the same math and statistics needed

44:53

for the interpretation of clinical laboratory test interpretation

44:56

. A p-value in a clinical trial

44:58

represents the tails of the standard

45:00

distribution of expected results and similarly the

45:02

high and low ends of normal results

45:04

of any quantitative clinical test . So

45:07

not just positive or negative , but you know the

45:09

range of numbers , the stuff that turns up red

45:11

. They merely demarcate the median plus

45:13

or minus two standard deviations in the same way

45:15

that the p-value does . So

45:17

the results beyond those parameters have the same statistical

45:20

meaning as a p-value , and the editorial

45:22

highlights this point and also notes that

45:24

it's not a new concept . That was published in the

45:26

journal over 30 years ago . This is a large

45:28

emphasis of what I write about in

45:30

my book Clinical Reasoning , and essentially

45:32

the editors , without saying it , are encouraging

45:34

a Bayesian approach to interpreting both

45:37

clinical trials and lab results . A

45:39

normal lab result or a negative

45:41

clinical trial don't necessarily mean that

45:43

a patient doesn't have a disease or that

45:45

the intervention being studied doesn't have merit . But

45:48

in the same way the opposites are also true

45:50

An abnormal lab result or a positive

45:52

clinical trial doesn't necessarily mean that

45:54

a patient does have a disease or that

45:56

a clinical hypothesis is true . So in

45:58

both cases , pretest probability matters

46:00

. Now they encourage a likelihood or ratio

46:03

approach to thinking about how the results

46:05

of a test or the results of a clinical

46:07

trial alter what a Bayesian would call

46:09

the prior probability of your hypothesis

46:11

. But it's all the same stuff . It uses slightly different

46:13

terms , to put it in the language commonly used

46:16

by frequentist statisticians , and

46:18

talks about likelihood ratios rather than Bayes'

46:20

factors for the Bayesian and other equivalent

46:22

substitutions . But the spirit of it's excellent and

46:24

their reason for it is maybe different than

46:26

my reason for approaching it , but it is

46:29

high time that folks realize that

46:31

these trials are not

46:33

all or nothing , that p-value doesn't demarcate

46:35

truth from falsity and also clinical lab

46:38

tests . Normal and abnormal tests

46:40

aren't the arbiter of whether a patient

46:42

has a disease or not .

46:44

Yeah , I think the whole purpose of this article

46:46

is to translate something about

46:48

clinical studies into

46:51

terms that maybe many

46:53

practicing clinicians are more familiar

46:56

with , where they order lab tests all day

46:58

, read lab tests all day , but they're not every

47:01

day designing trials or

47:03

reviewing trials or that

47:05

familiar with the pitfalls of trials

47:07

, so to speak . So just to translate

47:09

all of this if you test someone

47:12

for a disease who probably doesn't have

47:14

it so like for me , who

47:16

has zero joint pain if you tested me for

47:18

arthritis there's some lab tests

47:20

that could indicate that and the tests came

47:22

back positive that result

47:25

would probably mean something different

47:27

than if you did that same test on

47:29

someone who very obviously

47:31

already has it and you're

47:33

almost just confirming , like someone with severe joint

47:36

pain , otherwise unexplained , and

47:38

family history of arthritis

47:40

. Test them and the test is positive , that

47:42

confirms it . But again , if you test me

47:44

, then you would think maybe something

47:47

else , something got messed up with that

47:49

test . It's not a perfect test . And then

47:51

the inverse is true . So if you test

47:53

someone who almost certainly has the disease

47:55

and it's negative , then you'd say , well

47:57

, I still think they have it very likely

47:59

. So that might be a false negative , because

48:01

there are false negatives with this test , whereas

48:04

that same negative result in me , who

48:06

already had the low chance . That's probably

48:08

a reliable , true negative , because

48:10

we know from our everyday

48:13

practice that clinical tests are not perfect

48:15

. They have to be used appropriately and interpreted

48:18

with caution . The exact same things are true

48:20

for clinical trials , because they also

48:22

are not perfect and have to be used

48:24

responsibly , done responsibly

48:26

and then interpreted again with caution

48:29

. And you go into great detail of all

48:31

of that in your book .

48:32

Well , it's good to see more and more articles like this

48:34

being published . I do think we've moved away

48:36

. Years ago , when I was a med student carried a Fagan

48:39

nomogram in my pocket , we talked about likelihood

48:41

ratios . I don't see that that often anymore

48:43

. We've moved away in the electronic era

48:45

from thinking about these tests

48:48

in a likelihood ratio , pre-test , post-test

48:50

probability manner , and so this

48:52

article is a reminder that's how we interpret tests

48:54

and extending that also to clinical

48:56

trials . If you don't know what a Fagan nomogram is

48:59

or a likelihood ratio , please read the book

49:01

. But yeah , we're quietly seeing a revolution , I

49:03

think , in Bayesian inference or Bayesian

49:05

updating entering into medical thinking , and unfortunately

49:07

we haven't seen it translate you much into clinical practice

49:10

, but we're getting there . But okay , well

49:12

, one thing I hear a lot about is that we've made

49:14

a ton of progress by teaching endoscopic

49:16

surgeons to take bigger and bigger bites or

49:18

use different suture-assisted devices or things like

49:21

that , and that we've been making all this progress

49:23

in reducing what once was a

49:25

very horribly high rate of vaginal

49:27

cuff dehiscence with endoscopic hysterectomy

49:30

. But we've reduced it down now to an acceptable

49:32

level because we've taught better techniques about

49:34

all these things .

49:36

Yeah , there's definitely truth to

49:38

that . We have made a lot of progress from

49:40

those horrible rates . Historically in

49:42

the early days of laparoscopic hysterectomy

49:44

the rates were around 3% in the literature

49:47

. So everyone in 30 or

49:49

so patients would require a return

49:51

to the OR to fix that cuff dehiscence

49:53

. So to address that , endoscopic

49:56

surgeons over time have been taught

49:58

to take larger bites than they

50:00

think they need to and just realize

50:02

that the effect of the magnification on their

50:05

laparoscope is causing them to misjudge

50:07

the size of the bite they're taking

50:09

. It's easy to think something

50:11

is a centimeter , but really it's only about five

50:13

millimeters because you're so zoomed in it

50:16

looks like a big distance but it's really a tiny

50:18

distance . But you've always maintained

50:20

that despite many of these improvements

50:23

, there's a limit in how low

50:25

you can reduce complications of laparoscopic

50:27

hysterectomy like the cuff dehiscence , because

50:30

of how the colpotomy is done . So if

50:32

you're doing everything laparoscopically

50:34

, you have to use a thermal device

50:36

to make the colpotomy

50:39

, which is the vaginal incision that ends

50:41

up getting sutured back together and

50:43

that causes a lot more damage to

50:45

the tissue compared to if you

50:47

would use a scalpel during vaginal or

50:49

abdominal hysterectomy . So

50:52

even the most perfectly sized suture

50:54

closure is going to have less healthy tissue to

50:56

work with and more risk of dehiscence

50:59

right .

50:59

Yeah Well , there's a new paper in the March

51:01

2023 edition of the Journal of Minimally

51:03

Invasive Gynecology which looks at trends

51:06

and risk factors for vaginal cuff dehiscence

51:08

by mode of hysterectomy over time

51:10

, which should tell us a story of how improvements

51:13

in techniques and emphasis on this

51:15

problem helped to make the situation better .

51:17

All right . What did they find ?

51:18

Well , they looked at 4,059

51:20

hysterectomies of all routes over

51:23

an 11-year period at one single

51:25

facility . So these included abdominal

51:27

, vaginal , total laparoscopic-assisted vaginal

51:29

hysterectomies and robotic-assisted hysterectomies

51:31

. They , of course , excluded super cervical

51:33

hysterectomies from their data , since there's no cuff

51:35

to close . Now , overall , the rate of

51:37

vaginal cuff dehiscence in the study

51:39

was highest among robotic-assisted hysterectomies

51:42

, at 0.66%

51:44

, so way better than 3% , followed

51:46

by 0.32% risk

51:48

with total laparoscopic hysterectomies and

51:50

0.27% risk with

51:52

total abdominal hysterectomies . And there were

51:54

no cuff dehiscences cuff in the laparoscopic-assisted

51:57

vaginal or in the total vaginal hysterectomies

51:59

. Interestingly , they emphasized that this demonstrated

52:01

a much lower rate of cuff to hystinces than was previously

52:04

reported in the literature , which definitely does

52:06

, though this is still a relatively

52:08

rare complication . So they admit

52:10

that it may be underpowered , but it still looks

52:12

like there's been an improvement in the rates of cuff

52:14

to hystinces over time with these improved techniques

52:17

.

52:17

Okay . Well , let's go back to the part where

52:19

they said there were zero dehiscences among

52:22

the vaginal hysterectomy group and

52:24

also none in the laparoscopic-assisted

52:26

vaginal hysterectomy group . That probably

52:29

shows that the key factor

52:31

there is how the colpotomies are made . Although

52:34

they don't verify that every single colpotomy

52:36

for these cases was done by a cold

52:38

knife , some people even will use a

52:40

Bovie at TVH , but it's reasonable

52:43

to assume that most of them probably were

52:45

done by scalpel and not by energy

52:47

.

52:48

Yeah , I do think that the authors bury the lead here

52:50

in focusing on the reduction in

52:52

the rates of the endoscopic hysterectomies

52:54

, because the story is there were none , as

52:57

you said , with the vaginal hysterectomies or the laparoscopic-assisted

52:59

vaginal hysterectomies . I definitely think that

53:02

if you make the colpotomy with a cold knife

53:04

, as you do at vaginal hysterectomy , that

53:06

you'll have better tissue healing and a lower rate

53:08

of dehiscence than if you make it with an energy device

53:10

. This has been well enough established . I think

53:12

now for a while that we can consider

53:15

it just a fact . In over a thousand vaginal

53:17

hysterectomies that I've performed or assisted

53:19

, I've never seen a cuff dehiscence . But

53:21

I think vaginal surgeons , for the same reason , should avoid

53:23

, as you said , using the Bovie to make

53:25

their colpotomies . Now there are other differences

53:28

than what tool you use for the colpotomy

53:30

. In both vaginal hysterectomy and laparoscopic-assisted

53:33

vaginal hysterectomy , the vaginal cuff is closed by hand , directly

53:36

from below , and this may still result

53:38

in better bites than a closure made using

53:41

a camera and laparoscopic tools

53:43

coming from above . Or they also had

53:45

a men abdominal hysterectomies and the same thing you're

53:47

fighting for exposure deep in the pelvis

53:49

. You think you got the edge of it . It's hard to see

53:51

. Well , it's easy to see vaginally . But I don't

53:53

believe that I take necessarily bigger

53:56

bites in vaginal hysterectomy than endoscopic

53:58

surgeons do today , at least after

54:01

being retrained . In fact , I think they take bigger

54:03

bites than I do . So I really do think this

54:05

comes down to the use of an energy device

54:07

, which is just a limiting step in

54:09

this procedure . It really is quite interesting

54:12

why the technique of laparoscopic hysterectomy doesn't

54:14

just encourage doing the colpotomies

54:16

and the closure is vaginally , in other words

54:18

, a laparoscopic-assisted vaginal hysterectomy . It

54:20

would be much cheaper , safer for the patient

54:22

, faster in many cases . But I think

54:24

the industry has promoted the total laparoscopic approach

54:27

, particularly when using the robot , especially

54:29

with the colpotomizers , the suture cyst devices

54:31

, all these things that you need to buy

54:33

if you're going to do a total endoscopic approach

54:36

, that you don't need if you're doing a laparoscopic-assisted

54:38

vaginal hysterectomy .

54:39

Yeah , I agree with that . It seems like there's so

54:41

many products that cater and encourage

54:43

people doing hysterectomies to

54:45

do them fully laparoscopically and

54:47

not partially laparoscopically , and

54:49

I haven't seen , for example , a product

54:52

that's catered around doing

54:54

a colpotomy from below and

54:56

then maintaining an airtight seal and then

54:58

coming above and manipulating

55:00

the uterus . I do think I

55:03

have done some V-notes . I think

55:05

V-notes is starting to change that , although

55:07

you could also make it work with just

55:09

using an insulation ring in the vagina

55:11

, and that would probably still be more airtight

55:13

than improvising with a sterile glove

55:15

to keep the air from leaking out . But there aren't

55:18

any custom fit products

55:20

that are meant to perfectly make

55:22

that the most easiest way to do

55:24

the hysterectomy .

55:25

Well , if you're doing it as a laparoscopic-assisted , you

55:28

just do the top and you go down and finish

55:30

at the bottom and close . So you don't

55:32

necessarily need to go back up , and

55:34

so it's just a different technique . And another fundamental

55:36

limitation of the endoscopic

55:39

approach , particularly when you do the lower portion

55:41

is the increased risk of injury

55:43

to the ureters , which are pulled

55:45

in closer to the uterine artery rather

55:47

than pulled away from them when you descend the uterus

55:49

with a vaginal hysterectomy . So I think that's

55:51

the other fundamental thing that endoscopic

55:54

hysterectomy cannot overcome , right .

55:56

This study also showed that

55:58

the trend increased towards

56:00

minimally invasive hysterectomies

56:02

. Over their study period , it went

56:05

from 2010 to 2021

56:07

. There was a 14 percentage

56:09

point increase in hysterectomies

56:12

done by one of those minimally invasive approaches

56:14

, but overall there was a 35

56:16

percent point increase in

56:18

utilizing the robot , which

56:20

shows that the robot is eating up the other , less

56:23

expensive , minimally invasive approaches

56:25

.

56:26

Yeah , and that's what typically happens . It's

56:28

true that the total number of minimally invasive

56:30

hysterectomies are increasing as you see the robot

56:32

introduced over time . But that

56:35

was going to happen anyway had there been no

56:37

robot , as more gynecologists were graduating

56:39

from residency programs trying to do laparoscopic

56:41

hysterectomies or , in Howard fantasy

56:44

land , vaginal hysterectomies . But there's really

56:46

no evidence that the robot in and of itself

56:48

has been the driving force of a

56:50

decrease in open abdominal

56:52

hysterectomies in the United States , but instead

56:55

it seems to just be cannibalizing both

56:57

vaginal and laparoscopic hysterectomies

56:59

, replacing a minimal hysterectomy with

57:01

another minimal approach . So this again just shows

57:04

the influence of industry on our profession .

57:06

It's interesting that the authors don't even

57:08

mention how vaginal hysterectomy

57:10

is really superior in

57:12

regards to the outcome that they were focusing

57:14

on . It's almost just taken for granted

57:17

that we still have to do most

57:19

of these hysterectomies endoscopically , and

57:22

we should be proud of ourselves that the robot

57:24

only has a two third of a percent risk

57:26

that the woman will have a cuff breakdown and

57:28

require a second surgery , even when there's already

57:31

another route that yielded

57:33

zero cuff breakdowns at least zero

57:35

in this study . They just

57:37

overlapped it .

57:38

Well , yeah , that's the trend for literature

57:40

in general about hysterectomy . The literature

57:42

overall is consistent over

57:44

time that vaginal hysterectomy is the best approach for

57:46

the patient and the route associated with the best

57:49

outcomes and the lowest risk of re-operation

57:51

, the lowest cost , etc . Yet the

57:53

trend is away from it because of many of the reasons

57:55

that are certainly beyond the scope of this discussion

57:57

, but industry influence has a lot to do

57:59

with it .

58:00

Well , speaking of that , we will

58:02

have to talk about v-notes at some point

58:04

. It seems like , at least for some people , it's

58:06

like an attempt , with the help of industry

58:08

, to reintroduce vaginal surgery to

58:11

gynecologists who have some trepidation

58:13

of operating without a laparoscope

58:15

or for some reason are averse to

58:18

doing vaginal hysterectomy . So

58:20

we'll have a more extensive discussion

58:22

on it later , I think .

58:24

I think V-notes trumps robot , but

58:26

at some point you got to take the training

58:28

wheels off .

58:29

Yeah , that's fair .

58:31

Okay , well at least V-notes has the ability

58:33

to overcome those two fundamental differences

58:35

in terms of how you make the colpotomy

58:37

and the descending the uterus with

58:40

the time of ligation of the uterine artery

58:42

. So there aren't large studies

58:44

that give fair comparisons , but in

58:46

the range of less bad , it

58:48

has to be the less bad

58:50

one .

58:51

Yeah , I think later we'll see more head-to-head

58:53

to the other types of minimally

58:55

invasive . It's still a little bit too early

58:58

to say quantitatively

59:00

and all of that stuff , but we'll come back to

59:02

it later . But for today we'll wrap up

59:04

. So the Thinking About OBGYN website

59:07

will have links to a lot of the things we talked about and

59:09

then we'll be back in a couple weeks .

59:15

Thanks for listening . Find us online at

59:17

Thinking About OBGYNcom

59:20

. Be sure to subscribe . Look for

59:22

new episodes every two weeks .