Kevin Powell
1. Intravenous fluids – you have dedicated a lot of thought and research towards this topic, including prior presentations at PHM and an article (Healthy Skepticism and Due Process. J Hosp Med. 2018 Sep;13(9):654-655). If you got to wave a magic wand and change one thing about IV fluid use across North America, what would it be?
Although I raised the issue of maintenance IVF composition at a PHM conference in 2011, most of my work on the subject since 2014 has been focused on the bad science and bad policy making. If I could change one thing about maintenance fluids, it would be to have PHM focused on more important things and to stop promoting a practice change that has poor scientific reliability, inadequate safety data, and a nearly unmeasurable impact on patient care.
My work with IVF has largely been a vehicle for analyzing how errors are made in scientific validity in pediatric hospital medicine. It is worrisome to me that people believe that one can improve the safety of rare adverse events based on small RCTs. The risk of a serious bad outcome from maintenance IVF in a nonsurgical, previously healthy pediatric patient is about 1:150,000. Surgical patients are at 10-fold higher risk. RCTs with less than 100 patients per arm cannot provide reliable insight into those rare events. That is the second law of thermodynamics. A meta-analysis of 10 such studies totaling 1000 patients is still not useful. Even the large Minnesota study of 10,000 patients, which is the best study we have, demonstrated no statistically significant benefit and no clinical benefit. In that study, there were only 4 cases of moderate hyponatremia (128 or 129 mEq/L) during the whole year, and no serious adverse events. So that data cannot be extrapolated to predict events of profound hyponatremia. But over 20 articles have been published on this subject.
Worse, a guideline was created when, after 15 years of seeing these publications, 55% of pediatric hospitalists were using one method and 45% had rejected the erroneous science and chose to use the compromise, traditional approach. A committee of 17 people was formed with only 2 pediatric hospitalists involved. The committee failed to ensure the minority viewpoint was adequately represented. Without ever presenting their work at a PHM conference, indeed without most hospitalists even knowing about this committee’s existence, a guideline on maintenance IVF was perpetrated. Due to the rare, random and idiosyncratic nature of these adverse events, the guideline threatens anyone who preferred the traditional approach with huge malpractice lawsuits. Medicine recently has sought to eliminate variability in care. This is the correct goal when evidence is clear that one option is better. When evidence is unclear, we need to practice inclusivity, not tyranny.
I could have raised a larger stink about all this than I did, but I thought the specific issue of maintenance fluid was of too low a value to merit attention by the field of PHM. The bigger issues to me were that 1) PHM has been very weak at assessing and rejecting bad science and 2) the leadership for the AAP’s process for writing guidelines lacked rigor, inclusivity, and reliability.
Those issues became the focus of my research for six years while the time spent on clinical care has dwindled to nothing. As I evaluate the performance of the CDC and the White House Coronavirus Task Force over the past 5 months, I see the very same flaws in assessing the quality of science and in the polity of producing guidelines. The credibility of scientists has been severely compromised. Those flaws have cost tens of thousands of lives. I wish I had been more successful in my efforts, but at least I have been focused on the key issues.
2. You participated in the March for Science in St. Louis and wrote the listserv about how little attention it got. In addition to some notable phrases from your post, such as “In Peer Review We Trust,” you mentioned that Dr. Syndey Gellis used to provide an abstract service called Pediatric Notes (an annotated bibliography of notable research with commentary on its likely validity). Do you think it’s feasible to resume something like that on the SOHM library or the AAPHOSPMED Listserv as a group effort? If so, how do you envision it?
I believe the Listserv could do this and in a limited way it has served this purpose. PHM is too complex now for a solo person like Dr. Gellis to serve this function. I had hoped an online community could serve that function but over the years that hope has diminished.
People on AAPHospMed are reluctant to challenge another person’s ideas. Confrontations should always be civil and tactful, but if one values truth, there are times when confrontation is necessary. Tact is the art of making a point without making an enemy. Sometimes, though, you just have to make the point. MIT is an interesting community in that respect. Engineering problems are more likely than medical problems to have clean, verifiable, answers. The difficulty is finding them. Differences of opinion about the correct answer and the path for finding it are focused on solving the problem and less likely to be perceived as ad hominem. So stark disagreements can occur and are even valued for providing clarification. Many personality conflicts between researchers can be set aside in the quest for an absolute truth. The MIT approach can be perceived as rude in other communities.
Pediatrics is at the other extreme from MIT engineers. Pediatricians highly value being “nice.” I’d prefer the adjective polite, but the word nice is what I hear from other specialties describing pediatricians. Participants and lurkers on the AAPHospMed Listserv strongly value the collegiality of the community. They see the vitriol spewing in the vast majority of forums that have been created by the Internet and Facebook. They do not want to see that contaminate this community. So rigor and confrontation have not been prominent characteristics.
PHM has shown a willingness to challenge dogma when the dogma is seen as something created by others. The PHM community has been far less willing or able to confront wrong ideas if they are coming from within. Skepticism and an absolute dedication to truth are impeded by a desire to be nice.
A second factor appears to be fear that bosses will retaliate. The fear of criticizing someone else, with the risk that it might harm careers, is real. Dysfunctional academic hierarchies can be destructive to people and their careers. Based on what I have witnessed happen to others and on my personal experience, I don’t mentor people to ignore that fear. With time, I decided that since I was white, male, single, and able to work outside academia when necessary, if I couldn’t stand up to those miscreants and pay the price, who could? It may not be wise for others to follow that lead.
Could a small group of people with enough gravitas serve the role of peer reviewing on the Listserv? Maybe, but not in the current culture. I would focus more on making the SOHMLibrary a repository of local practice guidelines. By having those easily accessible, you’ve accomplished the important elements of peer review.
3. In your article “Reducing Overtreatment without backsliding” (J Hosp Med. 2017 Nov;12(11):937-938), you mention that “it is rare and difficult to eliminate, without substitution, ineffective treatments once they are in widespread use. This is the challenge facing the Choosing WiselyTM approach. Established habits of overtesting, overdiagnosis and overtreatment are refractory to correction, other than by replacing retirees with a new generation of physicians.” Aside from waiting for a new generation of physicians, what other options are there for physicians who want to change culture/bad habits at an institution/practice?
I have a great deal of expertise in engineering, in medicine, and in ethics. Your question is about managing human behavior. There are many people in PHM who have expertise in that area. Ask them. I’m not one of them.
I’ve tried to compensate for this with an unwavering quest for the truth, a belief in empowering the voices of all stakeholders, and a dedication to compassionately caring for children. The MBAs say that isn’t enough and I acknowledge they are right. But those are the only tools I’ve mastered.
4. I believe you gifted Dr. Christopher Landrigan a T-shirt saying “Enemy of the American People’ for his opinion piece in The Washington Post. You mentioned in a listserv post that you try to keep up with how mainstream news media portray issues in medical ethics. What trends have you noticed and how can we as hospitalists contribute or affect change (if need be)?
That was not me. I have no knowledge of that event or what was implied.
I have worked with philosophers and theologians who can synthesize systematic theology far better than I can. But I can speak their language while I live on the front lines of medicine. So my role is typically 1) using examples of real-life situations to point out weaknesses in their theories and 2) telling them when their abstract ideas, while logically true, don’t correspond to reality (see Euclid’s 5th postulate.) Rationing in an era of Covid-19 has manifested both issues.
The trend I have seen in mainstream news is an abdication of the journalist’s responsibility to find the truth. They present the two most diametrically opposed viewpoints they can find and expect the uninformed public to ferret out the truth somewhere on the spectrum. Modern journalism also has adopted surveys rather than investigation. CNN once presented a poll showing that most Americans though O.J. Simpson was guilty of murder. That poll was taken while the police were still chasing the white Ford Bronco, so nobody answering the poll had any actual knowledge of the facts. In an era of clickbait, this insanity has blossomed.
My observation has been that most scientists who go to church leave their science on the doorstep when they enter and pick it up again when they leave. The most effective way to contribute to medical ethics is to take the science inside and include it in conversations in the pews. It changes the public and it changes you. (Non-Christians will need to extrapolate my observation to apply it to their practices and beliefs.)
5. During a previous listserv post, you mentioned a French study where oral amoxicillin was used (in your words) "starting after 48 hr of IV amoxicillin (yes, AMOXicillin, not AMPicillin) for GBS disease and all 200 newborns did well. Additionally, the study measured serum levels and proved they were therapeutic – contrary to previous myths that neonates don’t reliably absorb oral drugs and that high doses of amoxicillin aren’t fully absorbed." Given that this study is 13 years old and IV therapy is still the standard, why do you think oral transition is not happening? Do you think that this is something, much like osteomyelitis, that might change in the future, especially given that the risk of a central line infection is far greater than the theoretical (and empirically near zero) risk of undertreating a meningitis? How might you design such a study?
Yes, I expect it will change in the future. That single study was so large and so well designed that doing another RCT is unnecessary. An observational study to track rare failures will be needed to refine the results.
The long delay in adopting this practice indicates that many/most physicians do not practice EBM. They practice M&M medicine – with their choices determined by the cases of morbidity and mortality they have seen. So the ID specialist, who has seen the worst cases, is more easily persuaded to act invasively. The hospitalist sees the routine complications and suffering from aggressive treatment. Only when the respect for the hospitalist is on par with that of the “expert” will things change.
This came up recently in a discussion with the orthopedic surgeon who replaced my hip. I asked for how long I should take antibiotic prophylaxis when I see the dentist. She said she prefers for the rest of my life, because she has seen joint infections occur years after the procedure. I pointed out that as a hospitalist, I get to routinely deal with allergic reactions, SJS, C. diff, and drug resistance. We compromised on two years.
6. Many of your opinion pieces, presentations, and articles center around staying apprised/abreast of literature, but also understanding the validity of studies. Many hospitalists (and their learners) struggle with this. Regarding methodologies, are there any tools or habits you have found helpful to you? How do you curate your literary diet/regimen from the high-volume output of the medical field? How do you screen abstracts and decide which papers to read fully?
Learn basic statistics. Early in a career join a journal club and learn from others how to critique research. Be ruthless.
Choose a few (3-6) key journals. Skim the table of contents and choose only articles that have the potential to be practice changing. Read the abstract and skim the methods. Be skeptical. Look for any reason to stop reading and trash the article. The literature is vast and you do not want to waste time and contaminate your memory reading bad research.
When you find a really good article, go beyond it and spend some time googling the subject. Do a mini-review and in 2-4 hours become an expert on that small area. That forms so many links in memory that the subject will stay with you for years and only needs updating when significant changes occur. Do one mini-review a month for 5 years and you will have covered most of PHM. The mini-reviews using PubMed are when you will find any important research in journals outside your standard routine.
Be wary of research that only promotes the positive findings. Demand safety data. Demand power calculations. Develop good habits of keeping up with reading. Having residents around can be a powerful motivator.
And remember that these are the methods that worked best in Jurassic and Cretaceous periods. You will need to evolve them. I reviewed the online c.v.’s of a few PHM researchers I admired. Some of them had 50-100 articles listed. I rarely had read more than 5 of them. Ioaniddis suggests that 50% of published research is wrong, and this limited sample suggests that another 40+% which is true is not worth reading. The current academic model for publication is seriously corrupt.
7. Jack Percelay often ends his list-serve commentary with the phrase “that’s just my 2 pennies.” What are your two cents?
Ultimately, I think the success of PHM will depend on three things.
1) Recruiting the best and brightest people possible to the field. I did not attend MIT as an undergraduate. My impression as a grad student is that the Institute graduates some the best in the world – but that is because they start the some of the best. I’m much less impressed with the incremental gains. So I think recruiting talent is important. Many things about PHM are less attractive than when I chose it. Many places, with no data to support it, opted for round the clock shifts in house. That is a lot of nights and weekends. The number of hours per FTE is much higher than Peds EM, who also work nights and weekends. There is now an extra 2 years of fellowship. What have we gained?
2) Establishing PHM’s niche of expertise and authority. The NICU and PICU have their niche. What is ours? I think newborn medicine should belong to PHM rather than the NICU. This is why I think the vision of the ABP, which dissed newborn care, is misguided. We are also the ones who can care for children in community hospitals without all the subspecialty support.
3) Having a high impact. Pediatric oncology in its first 15 years cured ALL. Infectious disease experts hope to create a vaccine for the pandemic in less than a year. In comparison, replacing prednisone with dexamethasone for patients in the ER with asthma and fiddling with IVF composition are not the best foci for the field of PHM.
Although I raised the issue of maintenance IVF composition at a PHM conference in 2011, most of my work on the subject since 2014 has been focused on the bad science and bad policy making. If I could change one thing about maintenance fluids, it would be to have PHM focused on more important things and to stop promoting a practice change that has poor scientific reliability, inadequate safety data, and a nearly unmeasurable impact on patient care.
My work with IVF has largely been a vehicle for analyzing how errors are made in scientific validity in pediatric hospital medicine. It is worrisome to me that people believe that one can improve the safety of rare adverse events based on small RCTs. The risk of a serious bad outcome from maintenance IVF in a nonsurgical, previously healthy pediatric patient is about 1:150,000. Surgical patients are at 10-fold higher risk. RCTs with less than 100 patients per arm cannot provide reliable insight into those rare events. That is the second law of thermodynamics. A meta-analysis of 10 such studies totaling 1000 patients is still not useful. Even the large Minnesota study of 10,000 patients, which is the best study we have, demonstrated no statistically significant benefit and no clinical benefit. In that study, there were only 4 cases of moderate hyponatremia (128 or 129 mEq/L) during the whole year, and no serious adverse events. So that data cannot be extrapolated to predict events of profound hyponatremia. But over 20 articles have been published on this subject.
Worse, a guideline was created when, after 15 years of seeing these publications, 55% of pediatric hospitalists were using one method and 45% had rejected the erroneous science and chose to use the compromise, traditional approach. A committee of 17 people was formed with only 2 pediatric hospitalists involved. The committee failed to ensure the minority viewpoint was adequately represented. Without ever presenting their work at a PHM conference, indeed without most hospitalists even knowing about this committee’s existence, a guideline on maintenance IVF was perpetrated. Due to the rare, random and idiosyncratic nature of these adverse events, the guideline threatens anyone who preferred the traditional approach with huge malpractice lawsuits. Medicine recently has sought to eliminate variability in care. This is the correct goal when evidence is clear that one option is better. When evidence is unclear, we need to practice inclusivity, not tyranny.
I could have raised a larger stink about all this than I did, but I thought the specific issue of maintenance fluid was of too low a value to merit attention by the field of PHM. The bigger issues to me were that 1) PHM has been very weak at assessing and rejecting bad science and 2) the leadership for the AAP’s process for writing guidelines lacked rigor, inclusivity, and reliability.
Those issues became the focus of my research for six years while the time spent on clinical care has dwindled to nothing. As I evaluate the performance of the CDC and the White House Coronavirus Task Force over the past 5 months, I see the very same flaws in assessing the quality of science and in the polity of producing guidelines. The credibility of scientists has been severely compromised. Those flaws have cost tens of thousands of lives. I wish I had been more successful in my efforts, but at least I have been focused on the key issues.
2. You participated in the March for Science in St. Louis and wrote the listserv about how little attention it got. In addition to some notable phrases from your post, such as “In Peer Review We Trust,” you mentioned that Dr. Syndey Gellis used to provide an abstract service called Pediatric Notes (an annotated bibliography of notable research with commentary on its likely validity). Do you think it’s feasible to resume something like that on the SOHM library or the AAPHOSPMED Listserv as a group effort? If so, how do you envision it?
I believe the Listserv could do this and in a limited way it has served this purpose. PHM is too complex now for a solo person like Dr. Gellis to serve this function. I had hoped an online community could serve that function but over the years that hope has diminished.
People on AAPHospMed are reluctant to challenge another person’s ideas. Confrontations should always be civil and tactful, but if one values truth, there are times when confrontation is necessary. Tact is the art of making a point without making an enemy. Sometimes, though, you just have to make the point. MIT is an interesting community in that respect. Engineering problems are more likely than medical problems to have clean, verifiable, answers. The difficulty is finding them. Differences of opinion about the correct answer and the path for finding it are focused on solving the problem and less likely to be perceived as ad hominem. So stark disagreements can occur and are even valued for providing clarification. Many personality conflicts between researchers can be set aside in the quest for an absolute truth. The MIT approach can be perceived as rude in other communities.
Pediatrics is at the other extreme from MIT engineers. Pediatricians highly value being “nice.” I’d prefer the adjective polite, but the word nice is what I hear from other specialties describing pediatricians. Participants and lurkers on the AAPHospMed Listserv strongly value the collegiality of the community. They see the vitriol spewing in the vast majority of forums that have been created by the Internet and Facebook. They do not want to see that contaminate this community. So rigor and confrontation have not been prominent characteristics.
PHM has shown a willingness to challenge dogma when the dogma is seen as something created by others. The PHM community has been far less willing or able to confront wrong ideas if they are coming from within. Skepticism and an absolute dedication to truth are impeded by a desire to be nice.
A second factor appears to be fear that bosses will retaliate. The fear of criticizing someone else, with the risk that it might harm careers, is real. Dysfunctional academic hierarchies can be destructive to people and their careers. Based on what I have witnessed happen to others and on my personal experience, I don’t mentor people to ignore that fear. With time, I decided that since I was white, male, single, and able to work outside academia when necessary, if I couldn’t stand up to those miscreants and pay the price, who could? It may not be wise for others to follow that lead.
Could a small group of people with enough gravitas serve the role of peer reviewing on the Listserv? Maybe, but not in the current culture. I would focus more on making the SOHMLibrary a repository of local practice guidelines. By having those easily accessible, you’ve accomplished the important elements of peer review.
3. In your article “Reducing Overtreatment without backsliding” (J Hosp Med. 2017 Nov;12(11):937-938), you mention that “it is rare and difficult to eliminate, without substitution, ineffective treatments once they are in widespread use. This is the challenge facing the Choosing WiselyTM approach. Established habits of overtesting, overdiagnosis and overtreatment are refractory to correction, other than by replacing retirees with a new generation of physicians.” Aside from waiting for a new generation of physicians, what other options are there for physicians who want to change culture/bad habits at an institution/practice?
I have a great deal of expertise in engineering, in medicine, and in ethics. Your question is about managing human behavior. There are many people in PHM who have expertise in that area. Ask them. I’m not one of them.
I’ve tried to compensate for this with an unwavering quest for the truth, a belief in empowering the voices of all stakeholders, and a dedication to compassionately caring for children. The MBAs say that isn’t enough and I acknowledge they are right. But those are the only tools I’ve mastered.
4. I believe you gifted Dr. Christopher Landrigan a T-shirt saying “Enemy of the American People’ for his opinion piece in The Washington Post. You mentioned in a listserv post that you try to keep up with how mainstream news media portray issues in medical ethics. What trends have you noticed and how can we as hospitalists contribute or affect change (if need be)?
That was not me. I have no knowledge of that event or what was implied.
I have worked with philosophers and theologians who can synthesize systematic theology far better than I can. But I can speak their language while I live on the front lines of medicine. So my role is typically 1) using examples of real-life situations to point out weaknesses in their theories and 2) telling them when their abstract ideas, while logically true, don’t correspond to reality (see Euclid’s 5th postulate.) Rationing in an era of Covid-19 has manifested both issues.
The trend I have seen in mainstream news is an abdication of the journalist’s responsibility to find the truth. They present the two most diametrically opposed viewpoints they can find and expect the uninformed public to ferret out the truth somewhere on the spectrum. Modern journalism also has adopted surveys rather than investigation. CNN once presented a poll showing that most Americans though O.J. Simpson was guilty of murder. That poll was taken while the police were still chasing the white Ford Bronco, so nobody answering the poll had any actual knowledge of the facts. In an era of clickbait, this insanity has blossomed.
My observation has been that most scientists who go to church leave their science on the doorstep when they enter and pick it up again when they leave. The most effective way to contribute to medical ethics is to take the science inside and include it in conversations in the pews. It changes the public and it changes you. (Non-Christians will need to extrapolate my observation to apply it to their practices and beliefs.)
5. During a previous listserv post, you mentioned a French study where oral amoxicillin was used (in your words) "starting after 48 hr of IV amoxicillin (yes, AMOXicillin, not AMPicillin) for GBS disease and all 200 newborns did well. Additionally, the study measured serum levels and proved they were therapeutic – contrary to previous myths that neonates don’t reliably absorb oral drugs and that high doses of amoxicillin aren’t fully absorbed." Given that this study is 13 years old and IV therapy is still the standard, why do you think oral transition is not happening? Do you think that this is something, much like osteomyelitis, that might change in the future, especially given that the risk of a central line infection is far greater than the theoretical (and empirically near zero) risk of undertreating a meningitis? How might you design such a study?
Yes, I expect it will change in the future. That single study was so large and so well designed that doing another RCT is unnecessary. An observational study to track rare failures will be needed to refine the results.
The long delay in adopting this practice indicates that many/most physicians do not practice EBM. They practice M&M medicine – with their choices determined by the cases of morbidity and mortality they have seen. So the ID specialist, who has seen the worst cases, is more easily persuaded to act invasively. The hospitalist sees the routine complications and suffering from aggressive treatment. Only when the respect for the hospitalist is on par with that of the “expert” will things change.
This came up recently in a discussion with the orthopedic surgeon who replaced my hip. I asked for how long I should take antibiotic prophylaxis when I see the dentist. She said she prefers for the rest of my life, because she has seen joint infections occur years after the procedure. I pointed out that as a hospitalist, I get to routinely deal with allergic reactions, SJS, C. diff, and drug resistance. We compromised on two years.
6. Many of your opinion pieces, presentations, and articles center around staying apprised/abreast of literature, but also understanding the validity of studies. Many hospitalists (and their learners) struggle with this. Regarding methodologies, are there any tools or habits you have found helpful to you? How do you curate your literary diet/regimen from the high-volume output of the medical field? How do you screen abstracts and decide which papers to read fully?
Learn basic statistics. Early in a career join a journal club and learn from others how to critique research. Be ruthless.
Choose a few (3-6) key journals. Skim the table of contents and choose only articles that have the potential to be practice changing. Read the abstract and skim the methods. Be skeptical. Look for any reason to stop reading and trash the article. The literature is vast and you do not want to waste time and contaminate your memory reading bad research.
When you find a really good article, go beyond it and spend some time googling the subject. Do a mini-review and in 2-4 hours become an expert on that small area. That forms so many links in memory that the subject will stay with you for years and only needs updating when significant changes occur. Do one mini-review a month for 5 years and you will have covered most of PHM. The mini-reviews using PubMed are when you will find any important research in journals outside your standard routine.
Be wary of research that only promotes the positive findings. Demand safety data. Demand power calculations. Develop good habits of keeping up with reading. Having residents around can be a powerful motivator.
And remember that these are the methods that worked best in Jurassic and Cretaceous periods. You will need to evolve them. I reviewed the online c.v.’s of a few PHM researchers I admired. Some of them had 50-100 articles listed. I rarely had read more than 5 of them. Ioaniddis suggests that 50% of published research is wrong, and this limited sample suggests that another 40+% which is true is not worth reading. The current academic model for publication is seriously corrupt.
7. Jack Percelay often ends his list-serve commentary with the phrase “that’s just my 2 pennies.” What are your two cents?
Ultimately, I think the success of PHM will depend on three things.
1) Recruiting the best and brightest people possible to the field. I did not attend MIT as an undergraduate. My impression as a grad student is that the Institute graduates some the best in the world – but that is because they start the some of the best. I’m much less impressed with the incremental gains. So I think recruiting talent is important. Many things about PHM are less attractive than when I chose it. Many places, with no data to support it, opted for round the clock shifts in house. That is a lot of nights and weekends. The number of hours per FTE is much higher than Peds EM, who also work nights and weekends. There is now an extra 2 years of fellowship. What have we gained?
2) Establishing PHM’s niche of expertise and authority. The NICU and PICU have their niche. What is ours? I think newborn medicine should belong to PHM rather than the NICU. This is why I think the vision of the ABP, which dissed newborn care, is misguided. We are also the ones who can care for children in community hospitals without all the subspecialty support.
3) Having a high impact. Pediatric oncology in its first 15 years cured ALL. Infectious disease experts hope to create a vaccine for the pandemic in less than a year. In comparison, replacing prednisone with dexamethasone for patients in the ER with asthma and fiddling with IVF composition are not the best foci for the field of PHM.