We kicked off the new year with a lively #JHSJC discussion on two articles from Volume 45, Issue 1 of The Journal of Hand Surgery: Development and Validation of a Prognostic, Risk-Adjusted Scoring System for Operative Upper-Extremity Infections and Establishing a National Registry for Hand Surgery.
Feel free to continue the conversation below in the comments! And please check out previous #JHSJC transcripts (please scroll to the bottom of the page).
J Hand Surg Am- ASSH @JHandSurg:
January’s #JHSJC starts now! Let’s get the discussion going!
Brent Graham @BgTalkinman:
Welcome to the January 2020 edition of #jhsjc! #jhsjc
Please remember to use the #jhsjc hashtag to follow the discussion.
Lets start with the article on risk stratification for re-op in upper ext #infection. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD:
Although this study did not really take into account social determinants of health #SDoH, it may still generalizable to most centers #jhsjc
Nina Suh @NinaSuh9:
I found the article’s risk factors were intuitive which was nice to have confirmation in a study #jhsjc
Brent Graham @BgTalkinman:
This study used a rigorous design, a split sample approach; half of the patients used to create a model validated in the other 1/2 #jhsjc
Nina Suh @NinaSuh9:
yes, agreed. the c-stat was also very good #jhsjc
Brent Graham @BgTalkinman
There were over 600 pts. Did the sample seem generalizable? Are these the pnts you see? #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @BgTalkinman:
This study didn’t take into account social determinants of health #SDoH, it may still generalizable to most centers #jhsjc
Charles Goldfarb @Congenitalhand:
Yes, generalizable w reasonable range. #jhsjc
Nina Suh @NinaSuh9:
Clinically I would still use judgement to know whether an infection has recurred but it provides stratification to tell patients #jhsjc
Yes, these would be patients I would see so I found the study useful and generalizable #jhsjc
Brent Graham @BgTalkinman to @NinaSuh9:
Given the predictors in the model, would that reflect your clinical judgement? #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD:
#SDoh Social determinants of health – socioeconomic status, education, neighborhood and physical environment, employment, social support networks, nutrition status as well as access to health care – #JHSJC pic.twitter.com/w8o5QwVBRq
Nina Suh @NinaSuh9 to @BgTalkinman:
yes, I believe the predictors coincide with my clinical judgement so I enjoyed this paper! #jhsjc
Brent Graham @BgTalkinman:
Pt factors, clinical findings, etiologic factors all part of the model prob improves validity #jhsjc
That is the value of the split-sample approach they used #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @BgTalkinman and @NinaSuh9:
high sensitivity model like this will have low specificity #jhsjc
Warren @whammert:
Good from a scientific standpoint, but hard to remember the points for each condition, so more helpful for research than everyday #jhsjc
Brent Graham @BgTalkinman:
How about the weighting of the factors. Did that seem to make sense? #jhsjc
Brent Graham @BgTalkinman to @RayRavenMD and @NinaSuh9:
That is an important observation. Does that matter? #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @BgTalkinman:
that means it’s good for predicting persistent infection. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @BgTalkinman and @NinaSuh9:
Experience and clinical judgment are needed to avoid overtreatment. #jhsjc
Brent Graham @BgTalkinman to @whammert:
I’m sure it will be an app for your phone in the very near future! #jhsjc
Warren @whammert to @BgTalkinman:
that would certainly make it easier – automated treatment -plug in the condition and it tells you what to do #jhsjc
Brent Graham @BgTalkinman to @RayRavenMD:
Right! Sensitivity prob matters less because where the decision is further surgery specificity is more important. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @whammert and @BgTalkinman:
#clinicaldecisionsupport systems #jhsjc
Brent Graham @BgTalkinman to @RayRavenMD and @NinaSuh9:
Like all decision aids, it is meant to be a guide. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD:
#SDoH can be responsible for up to 80 percent of a health outcome. Therefore, any scoring or #clinicaldecisionsupport system should include #SDoH (individually or collectively) as an independent risk factor(s) #JHSJC
Brent Graham @BgTalkinman to @RayRavenMD and @whammert:
Another positive feature is that it can help the less experienced make a good decision. #jhsjc
Nina Suh @NinaSuh9:
I agree to use it as a guide and it is helpful to have residents read papers like this to help them too for decision making. #jhsjc
Brent Graham @BgTalkinman to @NinaSuh9:
Agreed! #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD:
Well-designed #clinicaldecisionsupport systems have the ability to translate #evidencebased knowledge into actionable guidance for treatment. #JHSJC
Brent Graham @BgTalkinman to @NinaSuh9:
There are many conditions for which this kind of approach would have excellent applicability. #jhsjc
RyanC @RPCalfeeSTL:
Good points Nina and Ray #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD:
#clinicaldecisionsupport systems can increase diagnostic accuracy and efficiency, improve outcomes, and reduce cost of care. They will become an important tool in the push towards #ValueBasedCare #JHSJC
Brent Graham @BgTalkinman to @NinaSuh9:
I also like the fact that the output of the tool is a probability – much more realistic #jhsjc
Brent Graham @BgTalkinman to @whammert:
There is going to be an increasing proportion of care that looks like this in the future #jhsjc
Brent Graham @BgTalkinman:
It looks like there is general agreement on the merits of this study! #jhsjc
Nina Suh @NinaSuh9:
this type of study is not as common in hand sx literature so good to see and yes, other conditions would be suitable too #jhsjc
Warren @whammert to @BgTalkinman:
treat algorithms. Takes some of the art out of medicine #jhsjc
Nina Suh @NinaSuh9:
yes, probabilities are more realistic to handle clinically I think too #jhsjc
Dr. Amy M. Moore @AmyMMooreMD to @BgTalkinman:
Just landed in Columbus! Sorry to miss the conversation. #jhsjc
Avi Giladi @theaviram:
would we see different results with more data. surgeon/technique? timing? also, surprised at model strength with such low EPV #jhsjc
Christopher J. Dy @ChrisDyMD:
Hi all! Following along after finishing up an on call case. #jhsjc
Brent Graham @BgTalkinman to @whammert:
I agree but there will always be room for judgement. Hopefully those judgements can be better informed and less variable. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @whammert, @BgTalkinman, and @NinaSuh9:
Scoring system would be stronger if included antibiotics, blood tests results, and #SDoH #jhsjc
Brent Graham @BgTalkinman:
Dr. Moore (@AmyMMooreMD) is a senior author in this paper. Good work! Ask her questions! #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @AmyMMooreMD:
why did you decide not to include social determinants of health? #jhsjc
Brent Graham @BgTalkinman to @theaviram:
I don’t think so. The split sample approach is excellent but it may reflect a regional variation in indications… #jhsjc
Avi Giladi @theaviram to @BgTalkinman:
i agree that the included variables will remain important, but I believe care delivery has at least some impact #jhsjc
Brent Graham @BgTalkinman to @theaviram:
also definitions, other local factors could be important. Suspect it would be repeatable elsewhere with little variation #jhsjc
That could well be true. Hopefully somebody validates this model in a new setting and reports the results. #jhsjc
Brent Graham @BgTalkinman to @RayRavenMD and @AmyMMooreMD:
It’s a good question! #jhsjc
Nina Suh @NinaSuh9:
I agree @theaviram that regional variation is a fair concern.Some large IM models have been wrong when applied elsewhere #jhsjc
Brent Graham @BgTalkinman to @NinaSuh9 and @theaviram:
Start studying this and tell us what you find! #jhsjc
I would love to see this exact model validated somewhere else. Replication is important! #jhsjc
Nina Suh @NinaSuh9 to @BgTalkinman:
Okay!!! 🙂 #jhsjc
Brent Graham @BgTalkinman to @NinaSuh9:
We are committed to publishing good replication studies! #jhsjc
Avi Giladi @theaviram to @NinaSuh9:
as you said earlier, findings support clinical experience. likely they hold up. weight/impact and other pertinents may vary #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @AmyMMooreMD and @BgTalkinman:
It is a great study. And it tells us a lot about infections as well as the utility #clinicaldecisionsupport systems! #JHSJC
Brent Graham @BgTalkinman to @theaviram and @NinaSuh9:
Agreed! #jhsjc
Dr. Amy M. Moore @AmyMMooreMD to @RayRavenMD and @BgTalkinman:
Thanks so much. I give a ton of credit to Ketan Sharma- who as a resident wanted a better algorithm to treat hand infections. #jhsjc
Brent Graham @BgTalkinman:
Great discussion! Let’s move on to the other paper – establishing a hand surgery registry #jhsjc
We don’t usually discuss our review content but this is a very important topic and should make for interesting discussion. #jhsjc
Collecting data in a registry sounds great but there could be some issues beyond cost. Ideas? #jhsjc
Nina Suh @NinaSuh9:
I believe a hand registry for implants makes most sense as the outcomes are very defined #jhsjc
other hand surgery conditions may be harder to make a registry due to treatment variations #jhsjc
Avi Giladi @theaviram to @BgTalkinman:
maybe a registry helps solve the issues we were just discussing 🙂 #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @BgTalkinman:
I am glad we are discussing this. #jhsjc
Brent Graham @BgTalkinman to @NinaSuh9:
Agreed however implants are not commonly used in hand surgery. The proposal is to be much more comprehensive #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @theaviram and @BgTalkinman:
Exactly. #jhsjc
Warren @whammert:
Biggest challenge will be getting surgeons or others to enter data #jhsjc
Charles Goldfarb @Congenitalhand:
HAKIR is a good talking point as a baseline/ example for a national registry. Strengths are breadth but limited outcomes. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @BgTalkinman and @NinaSuh9:
registries aren’t just for implants. They are for all clinical data. #jhsjc
Brent Graham @BgTalkinman to @NinaSuh9:
Treatment variations for sure – but that could lead to important insights. #jhsjc
Warren @whammert to @Congenitalhand:
different health care system. Works there, but will be hard to replicate here. #jhsjc
Brent Graham @BgTalkinman to @RayRavenMD and @NinaSuh9:
That’s exactly what they are proposing. Like HAKIR in Sweden. #jhsjc
Nina Suh @NinaSuh9:
I thought implants would be easier for an initial registry bc the surgeon group would likely be smaller to convince #jhsjc
Charles Goldfarb @Congenitalhand to @whammert:
Not so easy even there. But, that model could work here too. Just very limited data in HAKIR. #jhsjc
Nina Suh @NinaSuh9:
also, who would house the data and who would get access to do studies from? #jhsjc
Warren @whammert to @BgTalkinman, @RayRavenMD, and @NinaSuh9:
they have a small number of facilities where all the data is grouped #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @whammert and @Congenitalhand:
#RCT is not really practical for studying many hand surgery conditions. #jhsjc
RyanC @RPCalfeeSTL:
If can collect national data big potential for advancing understanding but needs focus #jhsjc
Brent Graham @BgTalkinman to @whammert and @Congenitalhand:
I agree with @whammert. How will it account for the huge variation in health care settings in the US? #jhsjc
Brent Graham @BgTalkinman to @RPCalfeeSTL:
Isn’t that what large administrative databases already do? #jhsjc
Charles Goldfarb @Congenitalhand to @RayRavenMD and @whammert:
Agree- so hard to study everything. Focused conditions seems key. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @whammert, @BgTalkinman, @NinaSuh9, and @RPCalfeeSTL:
There is an increasing focus placed on the patients’ perspective #PatientReportedOutcome #JHSJC
Brent Graham @BgTalkinman to @RPCalfeeSTL:
I’m not sure I see what would be different. #jhsjc
Avi Giladi @theaviram to @RPCalfeeSTL:
A major challenge with this proposal, as with HAKIR, is that hand surgery technique varies widely. focus is critical. #jhsjc
Avi Giladi @theaviram to @BgTalkinman and @RPCalfeeSTL:
admin data devoid of quality outcomes. and based on coding. #jhsjc
Brent Graham @BgTalkinman to @theaviram and @RPCalfeeSTL:
I don’t see that as a huge obstacle because some of those variations are important to observe. #jhsjc
Charles Goldfarb @Congenitalhand to @theaviram and @RPCalfeeSTL:
Agree but variation is ok- may helpful inform patients & surgeons. Those with poorer outcomes adjust #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD:
A registry is an apparatus for a collection of observational data on a particular population. #jhsjc
Brent Graham @BgTalkinman to @theaviram and @RPCalfeeSTL:
What I see as more problematic are things like varying participation by hand surgeons #jhsjc
Warren @whammert to @BgTalkinman, @theaviram, and @RPCalfeeSTL:
selection bias #jhsjc
Avi Giladi @theaviram:
many participating here collect PROs as part of regular care. perhaps we can all coordinate. power in numbers. #jhsjc
Brent Graham @BgTalkinman to @theaviram and @RPCalfeeSTL:
Those are clear limitations. However, how will a hand surgery registry overcome that? #jhsjc
Warren @whammert to @theaviram:
to this point, there is not a consensus on what PRO to collect #jhsjc
Brent Graham @BgTalkinman to @theaviram:
That would be great but concern would be that entering all cases is unlikely, especially where there are poor outcomes. #jhsjc
Charles Goldfarb @Congenitalhand to @whammert and @theaviram:
Agree. So challenging unless we stay general (PROMIS). #jhsjc
Brent Graham @BgTalkinman to @RayRavenMD:
Agree completely. The problem is whether all data gets entered. It does in an admin db because that is linked to payment. #jhsjc
Avi Giladi @theaviram to @whammert:
agree. major challenge. PROMIS helps but has many (recently published) limitations #jhsjc
Nina Suh @NinaSuh9:
I think for overburdened MDs, getting buy-in will be difficult unless there is incentives or penalties unfortunately #jhsjc
Charles Goldfarb @Congenitalhand to @BgTalkinman and @RayRavenMD:
I am not sure how joint registries get around this concern of selective patient enrollment? #jhsjc
Brent Graham @BgTalkinman to @Congenitalhand, @whammert, and @theaviram:
I can see that happening bc PROMIS is going to be generally collected even in admin db 1 day. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD:
Registries must work through EHRs and interoperability is a challenge. Would have to integrate with multiple systems. More than 600 #EHR vendors exist #JHSJC
Interoperability must be based on open standards that enable providers to interface with any applicable registry without requiring customization or permission from EHR vendor #jhsjc
Brent Graham @BgTalkinman to @NinaSuh9:
That may prove to be the case – payment for “quality”. Getting ahead of the curve is smart but difficult. #jhsjc
Warren @whammert to @Congenitalhand and @theaviram:
I think this proposal is more than PRO – looking for procedures, costs, so hard for pt to enter #jhsjc
Brent Graham @BgTalkinman to @Congenitalhand:
I’m not sure but the implants themselves may be tracked by industry. Harder to track a CTR. #jhsjc
Avi Giladi @theaviram to @whammert and @Congenitalhand:
agree but PRO are major gap in admin data. we are building our own registry to try and overcome this. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @Congenitalhand and @BgTalkinman:
It’s not a study, the data can be anonymized. Why do patients have to enroll? #jhsjc
Brent Graham @BgTalkinman to @whammert, @Congenitalhand, and @theaviram:
In general the more detailed the registry, the less likely it is to be consistently filled in. #jhsjc
And the easier it is to complete, the less information it contains. #jhsjc
Charles Goldfarb @Congenitalhand to @BgTalkinman:
Agree. Also less exciting. Honestly, this is why focused conditions seem to make the most sense. #jhsjc
Megan Conti Mica, MD @megancontimica to @NinaSuh9:
I agree. It needs to be easy with a positive impact on practice #jhsjc
Warren @whammert to @RayRavenMD, @Congenitalhand, and @BgTalkinman:
not enroll, but someone has to enter data – pt can enter PROMIS/ PRO but not other things #jhsjc
Brent Graham @BgTalkinman to @RayRavenMD and @Congenitalhand:
They wouldn’t be enrolling. This would be a surgeon completed registry. #jhsjc
Brent Graham @BgTalkinman to @whammert, @RayRavenMD, and @Congenitalhand:
Exactly! And that is the prob. Bad results, busy surgeons, etc. might not consistently happen #jhsjc
Avi Giladi @theaviram:
dont “need” to enroll, but with privacy/data concerns mounting (ie Facebook) backlash is possible. opt out option may be needed. #jhsjc
Megan Conti Mica, MD @megancontimica to @RayRavenMD:
Do you think different patient populations will muddy any data conllection? #jhsjc
Charles Goldfarb @Congenitalhand to @BgTalkinman and @RayRavenMD:
Surgeon completed for surgical data. Subjective data has to allow emailed links for pts. #jhsjc
Brent Graham @BgTalkinman to @whammert, @RayRavenMD, and @Congenitalhand:
There will be costs and those will be borne by health care providers. #jhsjc
Megan Conti Mica, MD @megancontimica to @BgTalkinman:
What about surgeon factors? #jhsjc
Warren @whammert:
Surgeons will need to see a benefit to buy in and contribute date – lofty goal, but very difficult to execute #jhsjc
Brent Graham @BgTalkinman to @megancontimica and @RayRavenMD:
For sure! But the positive would be greater generalizability of the findings. #jhsjc
Brent Graham @BgTalkinman to @megancontimica:
Yup! That too. In the end, not sure it would be so different than admin db. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @megancontimica:
underserved populations always alter the data. But that is where most of the cost and bad outcomes are #jhsjc
Brent Graham @BgTalkinman to @whammert:
The carrot might be compulsion to participate as a demonstration of “quality”. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD:
Assessments of #quality differ substantially when using clinical registry versus administrative claims data #jhsjc
Megan Conti Mica, MD @megancontimica to @RayRavenMD:
Was just at a conference talking about how we always exclude at risk populations but those are truly the ones that need the research! #jhsjc
Brent Graham @BgTalkinman to @megancontimica:
I think the research questions will have to be very general. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD:
Administrative data contain valuable resource utilization data, medication data, and information on other non-hand diagnoses/procedures that may not be collected in disease or procedure-specific registries. #jhsjc
Warren @whammert to @RayRavenMD:
I think you indicate you are contributing to quality by participating as the carrot to get people to contribute. #jhsjc
Brent Graham @BgTalkinman to @RayRavenMD:
That may well be true. That could be a result of differing data quality. #jhsjc
Brent Graham @BgTalkinman to @whammert and @RayRavenMD:
Yes, that’s what I mean. Hard to see how that happens nationally but I think that would be the strategy. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD:
Comparing #AdministrativeData & #ClinicalData data is an effective #quality tool to identify discrepancies in each source of data #jhsjc
Brent Graham @BgTalkinman to @RayRavenMD:
Agree! I think that is the kind of work Robin Kamal is doing. #jhsjc
Warren @whammert to @BgTalkinman and @RayRavenMD:
that or make it a requirement for MOC/ continuous certification and the boards get access to data #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @megancontimica:
Yep. That’s where the savings come from. #BendTheCostCurve #jhsjc
Brent Graham @BgTalkinman to @whammert and @RayRavenMD:
That might also be a good strategy but I suspect the boards will not see that as their mandate. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @whammert and @BgTalkinman:
that is a fantastic idea! #MOC #jhsjc
Brent Graham @BgTalkinman to @Congenitalhand:
Agree that some kind of combined strategy is likely but still a major logistic undertaking. #jhsjc
Warren @whammert:
agree, but I think it will take something like that to get surgeons to participate #jhsjc
Charles Goldfarb @Congenitalhand to @BgTalkinman:
Great discussion. Will require committed surgeons/ institutions. #jhsjc
Brent Graham @BgTalkinman:
To come full circle: will it be better than PearlDiver? #jhsjc
Avi Giladi @theaviram to @BgTalkinman:
yes #jhsjc
Charles Goldfarb @Congenitalhand to @theaviram and @BgTalkinman:
Done right- yes. #jhsjc
Ray Raven MD, MBA, FAAOS @RayRavenMD to @Congenitalhand, @theaviram, and @BgTalkinman:
I agree. #jhsjc
Brent Graham @BgTalkinman to @theaviram:
Maybe! Roll up your sleeves! There is going to be a lot of work required from smart people! #jhsjc
Brent Graham @BgTalkinman to @Congenitalhand and @theaviram:
doing it “right” – that is the key point and a good one to end on! See you next month! #jhsjc
Nina Suh @NinaSuh9:
Great discussion! See everyone next month #jhsjc
