Russ Altman: What really happens when you mix medications?
Russ Altman: Vad händer när man kombinerar läkemedel?
Russ Altman uses machine learning to better understand adverse effects of medication. Full bio
Double-click the English transcript below to play the video.
and get some tests.
och det tas några prover.
that you have high cholesterol
att du har högt kolesterol
from medication to treat it.
skulle få dig att må bättre.
that this is going to work.
att det här kommer fungera.
a lot of studies, submitted it to the FDA.
utförde en massa studier,
granskade och godkände.
skeptically, they approved it.
hur läkemedlet fungerar,
of what the side effects are.
of a conversation with your physician
because you've been blue,
att du har varit nedstämd,
in life quite as much as you usually do.
på samma sätt som tidigare.
I think you have some depression.
jag tror att du lider av depression.
you another pill."
ännu en tablett till dig."
about two medications.
of people have taken it,
miljontals människor har tagit den,
the FDA looked at it -- all good.
FDA har kontrollerat - bra.
these two together?
de här två tillsammans?
"post-marketing surveillance,"
"post marketing surveillance",
if bad things are happening
om det händer dåliga saker
who has several diagnoses
med flera diagnoser
and really, in my opinion,
och enligt mig
to understand these interactions
vi kan förstå dessa interaktioner
of different sources of data
av data från olika håll
when drugs can be used together safely
är säkra att användas samtidigt
because that's his name.
för det är så han heter.
to understand how drugs work
hur läkemedel fungerar,
and how they work separately,
och hur de fungerar enskilt,
en fantastisk databas.
an amazing database.
download it right now --
ladda ner det nu -
of adverse event reports
companies, pharmacists.
företag, farmaceuter.
that the patient has,
som patienten har,
or side effects, that they experience.
that are occurring in America today,
som sker i USA idag,
of thousands of drugs.
and we know it's involved with diabetes.
att det har koppling till diabetes.
glucose response.
glukosresponsen."
look at the side effects of a drug
hitta biverkningarna av en medicin
is likely to change glucose or not."
ändrar glukosnivåer eller inte gör det."
that were known to change glucose
som är kända glukospåverkare
that don't change glucose,
som inte påverkar glukos
in their side effects?
"Vad är skillnaden i biverkningarna?
In urination habits?"
to give him a really good predictor.
gav honom en bra förutsägare.
with 93 percent accuracy
med 93 % sannolikhet
han behöver få självförtroende.
you have to build his confidence.
knows all the drugs that change glucose,
vilka mediciner som påverkar glukoshalten,
but not really that interesting,
men inte så intressant,
I thought you might say that."
so I did one other experiment.
så jag gjorde ett annat experiment.
who were on two drugs,
som tog två mediciner
glucose-changing signals,
på glukosförändring,
did not change glucose,
inte ändrar glukoshalten,
när de användes samtidigt."
Good idea. Show me the list."
not very exciting.
inte så intressant.
was, on the list there were two drugs:
var ett par mediciner:
a cholesterol medication.
en kolesterolmedicin.
of Americans on those two drugs."
amerikaner som tar båda medicinerna."
at the time, 15 million on pravastatin,
paroxetin och 15 miljoner pravastatin
en miljon använde bägge.
with their glucose
that he did in the FDA database
som han gjorde i FDA-databasen
with the mumbo jumbo,
med allt svammel,
evidence that we have."
som vi har."
electronic medical record.
elektroniska medicinska arkiv.
som kan användas i forskning,
that's OK for research,
on these two drugs
som tar båda medicinerna
and thousands of people
that take paroxetine and pravastatin.
som tar paroxetin och pravastatin.
and had a glucose measurement,
och var med om en glukosmätning,
another glucose measurement,
och var med om en annan glukosmätning,
something like two months.
ungefär två månader.
we found 10 patients.
had a bump in their glucose
we call this P and P --
vi kallar det P och P -
the second one comes up,
sedan kommer den andra,
20 milligrams per deciliter.
20 milligram per deciliter.
if you're not diabetic,
om du inte är diabetiker,
about a potential diagnosis of diabetes.
en diabetesdiagnos.
don't have a paper,
and -- give me a break --
at Harvard and Vanderbilt,
vid Harvard och Vanderbilt,
Vanderbilt i Nashville,
Vanderbilt in Nashville,
medical records similar to ours.
elektroniska medicinska arkiv.
similar patients
the glucose measurements
glukosmätningarna
in one week found 40 such patients,
40 patienter på en vecka,
samma trend.
from three diverse medical centers
från tre olika sjukhus
getting these two drugs
med dessa två mediciner
somewhat significantly.
we had left out diabetics,
have messed up glucose.
at the glucose of diabetics,
hos diabetiker,
per deciliter, not just 20.
på 60 mg/dl, inte bara 20.
"We've got to publish this."
"Vi måste publicera det här."
ett enda riktigt experiment,
was in review, went to the lab.
gick Nick till labbet.
who knew about lab stuff.
but I don't do pipettes.
men har inga pipetter.
ger mediciner till möss.
one P, paroxetine.
ett P, paroxetin.
of mice both of them.
fick båda medicinerna.
20 to 60 milligrams per deciliter
från 20 till 60 mg/dl
based on the informatics evidence alone,
på bevis i form av data,
if you give these to mice, it goes up.
om de tar dem.
could have ended there.
historien kunna sluta.
thinking about all of this,
of it, but somebody said,
men någon sa,
som tar de här två medicinerna
who are taking these two drugs
of hyperglycemia.
one new medication or two,
en ny medicin eller två
or the one drug you're taking,
their search logs with us,
dela sin sökhistorik med oss,
these kinds of searches.
sådana här sökningar.
denied our request.
who works at Microsoft Research
från Microsoft Research
the Bing searches."
companies in the world,
största företagen i världen
to make him feel better.
you might not understand.
för att göra googlesökningar,
to do searches at Google,
for research purposes only."
i forskningssyfte."
my friend at Microsoft.
min vän på Microsoft.
som en vanlig person skulle skriva
that a regular person might type in
"urinerar ofta", "kissar ofta" -
"urinating a lot," "peeing a lot" --
som man kanske skriver.
of the things you might type in.
som vi kallade "diabetesord".
that we called the "diabetes words."
that about .5 to one percent
till en procent
involve one of those words.
innehåller någon av de fraserna.
eller "Seroxat" - det är samma sak -
or "Paxil" -- those are synonyms --
of diabetes-type words,
that there's that "paroxetine" word.
ett ord för paroxetin.
to about three percent from the baseline.
så är andelen tre procent.
are present in the query,
finns med i frågan,
som vi undersökte
that we were interested in,
or hyperglycemia-type words.
their side effects indirectly
om sina biverkningar indirekt
to the attention of the FDA.
i sociala medier
surveillance programs
for doing this, and others,
och andra,
som enskilt eller tillsammans,
either individually or together,
Why tell this story?
of big data and medium-sized data
"big data" och medelstora data
a new ecosystem
and to optimize their use.
och optimera användningen av dem.
at Columbia now.
vid Columbia nu.
for hundreds of pairs of drugs.
samma sak med hundratals mediciner.
very important interactions,
is a way that really works
är ett fungerande sätt
interaktioner mellan mediciner.
of drugs at a time.
två mediciner åt gången.
on three, five, seven, nine drugs.
tre, fem, sju, nio mediciner.
to their nine-way interaction?
den niofaldiga interaktionen?
A och B, A och C, A och D,
A and B, A and C, A and D,
E, F, G, alla tillsammans,
D, E, F, G all together,
eller försvagar varandra,
more effective or less effective
that are unexpected?
for us to use data
interaktioner mellan mediciner.
the interaction of drugs.
that we were able to generate
hur mycket vi kunde få fram
frivilligt rapporterade sina biverkningar
volunteered their adverse reactions
through themselves, through their doctors,
genom sina läkare,
vid Stanford, Harvard, Vanderbilt,
at Stanford, Harvard, Vanderbilt,
och säkerhet - med rätta.
and security -- they should be.
som stänger ut den typen av data,
that closes that data off,
and it was a little bit of a sad story.
och det var ganska sorgligt.
the two drugs very carefully together,
med försiktighet,
när man skriver ut mediciner.
when you're prescribing.
två eller tre mediciner
two drugs or three drugs
nya effekter av mediciner
för biverkningar,
of causing a side effect,
för depression, för diabetes -
for depression, for diabetes --
TED Talk on a different day,
en annan dag -
of drugs in combination
av medicinkombinationer
of our patients even better?
våra patienter ännu bättre?
ABOUT THE SPEAKER
Russ Altman - Big data techno-optimist and internistRuss Altman uses machine learning to better understand adverse effects of medication.
Why you should listen
Professor of bioengineering, genetics, medicine and computer science at Stanford University, Russ Altman's primary research interests are in the application of computing and informatics technologies to problems relevant to medicine. He is particularly interested in methods for understanding drug actions at molecular, cellular, organism and population levels, including how genetic variation impacts drug response.
Altman received the U.S. Presidential Early Career Award for Scientists and Engineers, a National Science Foundation CAREER Award and Stanford Medical School's graduate teaching award. He has chaired the Science Board advising the FDA Commissioner and currently serves on the NIH Director’s Advisory Committee. He is a clinically active internist, the founder of the PharmGKB knowledge base, and advisor to pharmacogenomics companies.
Russ Altman | Speaker | TED.com