Category Archives: General

Crowdsourcing Medical Research

Medical resource is hard to do. It’s tough to come up with an idea no one has explored, design the study, accumulate subjects (for clinical research), analyze it, and then write a good paper. You’re probably familiar with my lamentations over all of the small,retrospective studies that seem to dominate the medical literature.

Crowdsourcing takes advantage of the “human computing power” of ordinary people all around us. Some of you may remember the protein folding computer game that was distributed for free about 10 years ago, allowing everyone to try their hand at designing protein configurations. Turns out, masses of regular people are better than computers for doing this kind of stuff. And the results were impressive. Or look at the power of Amazon’s Mechanical Turk, a crowdsourcing platform for a variety of tasks. Or KickStarter for funding projects that would have a tough time getting money on their own.

Researchers at Penn used crowdsourcing in a study to map the locations of all automated external defibrillators in Philadelphia. They called it the MyHeartMap challenge. The crowd quickly identified and catalogued over 1400 of them. They are now using crowdsourcing to perform literature sources to collect and analyze health-related studies using free medical literature search sites.

The crowd does best on studies involving problem solving, data processing, monitoring and surveying. The downside is that there is built-in variability when using the crowd, which can make it difficult to replicate and confirm validity. But the sheer numbers that can be accumulated are far larger than what can be expected using traditional research methods.

Bottom line: Crowdsourced research has significant potential to accelerate and improve the quality of medical research (and in other fields as well). Anyone engaged in research needs to look at their own projects to determine if any of them can be improved using crowdsourcing. Look at what crowds are good at (solving problems, observing and reporting), and use them to bolster and improve our knowledge base.

Reference: Crowdsourcing – Harnessing the masses to advance health and medicine, a systematic review. J General Int Med DOI: 10.1007/s11606-013-2536-8, 2013.

Print Friendly, PDF & Email

Motorcycle Helmet Type Makes A Difference

It’s well established that motorcycles helmets make a difference during a crash. EAST has developed an evidence based review that provides recommendations regarding helmet use. Click here to see them. Unfortunately, many states don’t have helmet laws, and several have been persuaded to repeal perfectly good ones.

Although there is strong literature support for helmet use, the specific type of helmet (full-face vs all others) has not been prospectively studied. Helmets that are not full-face (FF) do not cover areas that protect lower parts of the brain and the upper brainstem.

A group in Baltimore did a one year prospective study on the effect of helmet type on craniofacial injury. They treated 176 motorcycle crash victims during that time, and were able to identify the helmet type in 151. A total of 84 wore FF helmets, and 67 wore other types. Here are the interesting findings:

  • Facial fractures occurred in 16% of patients. Only 7% of FF helmeted patients had these fractures vs 27% wearing other helmets.
  • Skull fractures were found in 6%, with only 1% in FF helmet wearers vs 11% in other helmet types
  • Mortality decreased from 7.5% to 4.8% (36% reduction) in the FF helmeted patients (not statistically significant)
  • Cervical spine fractures decreased 20% from 11.9% to 8.5% in the helmeted group (also not significant)

Bottom line: Choosing a motorcycle helmet carefully is important. Remember, motorcyclists are far less protected than automobile drivers. Improving protection has definitively been shown to decrease injuries of nearly all types. Although it certainly is patient choice as to what they wear and if they ride, it can make a big difference to them (and society) if they ignore these recommendations.

Related posts:

Reference: Choice of motorcycle helmet makes a difference: a prospective observational study. J Trauma 75(1):88-91, 2013.

Print Friendly, PDF & Email

Identifying Bowel and Mesenteric Injury by CT

CT scan is an invaluable tool for evaluating blunt abdominal trauma. Although it is very good at detecting solid organ injury, it is not so great with intestinal and mesenteric injuries. Older studies have suggested that CT can detect mesenteric injuries if done right, but a newly published study has shown good accuracy with a few imaging tweaks.

A Taiwanese study looked at a series of prospectively studied victims of blunt abdominal trauma. Patients with abdominal pain or a positive FAST were entrolled (total 106). IV contrast was given, and scans during the arterial, portal, and equilibrium contrast phases were performed using a multidetector scanner. Images were read in a blinded fashion.

A total of 13 of 23 patients who underwent laparotomy were found to have a bowel or mesenteric injury. Five had bowel injury, 4 had mesenteric hemorrhage, and 4 had both. Mesenteric contrast extravasation was seen in 7 patients, and this correlated with mesenteric bleeding at laparotomy.

The authors found that the following signs on CT scan indicated injury:

  • Full or partial thickness change in bowel wall appearance
  • Increased mesenteric density
  • Free fluid without solid organ injury

Bottom line: This study shows that CT scan can detect bowel and mesenteric injury reliably if you scan the patient 3 times! This seems like over-radiation and overkill. A more intelligent way to approach this would be to perform a normal trauma abdominal scan. If a suspicious area of mesenteric or bowel thickening is seen, then a limited rescan through the affected area only for equilibrium phase images may be warranted. If actual contrast extrvasation is seen, no further scanning is needed. A quick trip to the OR is in order. And higher risk patients (e.g. seat belt sign) should have a lower threshold for diagnosis!

Reference: Contrast-enhanced multiphasic computed tomography for identifying life-threatening mesenteric hemorrhage and transmural bowel injuries. J Trauma 71(3):543-548, 2011.

Print Friendly, PDF & Email

June Trauma MedEd Newsletter Is Out!

The June newsletter is now available! Click the image below or the link at the bottom to download. This month’s topic is orthopedics. 

In this issue you’ll find articles on:

  • Pelvic trauma radiographs demystified
  • Bleeding and pelvic fractures
  • Predicting pelvic fracture bleeding
  • Posterior hip dislocation
  • Captain Morgan hip reduction technique
  • The fading 8-hour open fracture rule
  • Smoking is bad for fractures, really!

Subscribers had the newsletter emailed to them over the weekend. If you want to subscribe (and download back issues), click here.

image

Download the newsletter here!

Print Friendly, PDF & Email

Conservative Management Of Pancreatic Injury

There has been a slow shift toward nonoperative management of many injuries that used to demand a quick trip to the operating room. Liver and spleen injury is one of the best examples, with extremely good success rates (95%). Kidneys fall into this category, too.

The pancreas is another solid organ. Perhaps we can do the same thing? A number of pediatric surgeons have been attempting to manage children with pancreatic injury. Low grade injuries (principally contusions) have been managed expectantly for some time. Could higher grade injuries (duct injury) be managed this way as well? How about using repeat imaging, percutaneous drainage, stenting via ERCP, and TPN to avoid the OR in hemodynamically stable kids?

A recent paper looks at this practice critically. Nine years of registry data at two Level I pediatric centers was reviewed to identify all high grade (III-IV) pancreatic injuries. They isolated 39 children with this injury (which is quite a few!). They were separated into two groups based on initial management plan, operative (15) or nonoperative (24). Here are the results of interest (all statistically significant):

  • Average ISS was higher in the nonop group (23 vs 15)
  • Hospital length of stay was longer in the nonop group (28 vs 15 days)
  • TPN was required for a longer period in the nonop group (22 vs 8 days)
  • There were more complications in the nonop group (17 vs 4 children), with 13 developing a pseudocyst (none in the op group)

Bottom line: Nonoperative management of high grade pancreatic injury in kids is just not ready for prime time. It may seem that avoiding a big abdominal operation would be a good thing. Distal pancreatectomy usually keeps children in the hospital for 5-7 days, and then they are done unless they have other serious injuries. Nonoperative management results in a lengthy stay in the hospital, multiple imaging studies (radiation), getting stuck with big drainage needles, and TPN with its attendant infection risks. The old fashioned way, going to the OR, is still the best!

Related post:

Reference: Non-operative management of high-grade pancreatic trauma: is it worth it? J Ped Surg 48:1060-1064, 2013.

Print Friendly, PDF & Email