This isa question ¡that I get very often: What’s the differencebetween ¡patient engagement and patient activation? Arethey the same? Arethey different? 1
I think I’ll addressthat first.Butwe’ll also talk abouthowpatient activation is linked with the use of ¡information;whatis the evidence that activation is predictiveof ¡ outcome, such ¡as behaviors,utilization,costs;and ¡what do we knowabouthow to increase thisin patients. I’ll spend justa minute talkingabouthowdelivery systems are usingthis kind ¡of measurement in innovativewaysand ¡then,finally, end ¡u with ¡what the implicationsof all of thisis for supportingtheuseof information in healthcare decisionsamong consumers. 2
Whatis patientengagement? Asyou ¡all know,this term isused ¡pretty loosely.People all say the same thing,but they mean ¡differentthings.A common ¡way it is used ¡is to describeinterventionsto increase involvementor participation;or theresulting participation ¡or involvement;or both;or somethingelse. When ¡we started ¡the work on ¡ patient activation we spent a lot of time on definitionsto be very clear whatit was we were tryingto measure. 3
The definition ¡thatwe came u with ¡after workingwith ¡a national expertconsensus panel, with ¡patients, and ¡goingto the literatureisthat an ¡activated ¡consumer or patient is onewho hasthe motivation,knowledge, skill,and confidence to takeon the roleof managingtheir health ¡or their healthcare. That is, an ¡individual who understandswhat their roleis and feels competent and ableto do it. Thiswas one of the firstthingsthat we learned ¡after workingon ¡measurementin ¡this area is in ¡almost any population ¡ group thatwe havelooked at you see a full range.You see peoplewho are very passive abouttheir health ¡and ¡people who arevery proactive.And ¡that doesn’tmatter if they areeighty-‑fiveyear oldsyou’relookingator if it is Medicaid;you’ll see that full distribution.Whatyou ¡do seeis that the mean ¡will movea bit. For some groupsthe mean ¡will be lower and ¡for some groupsthemean ¡will be higher.Butthe pointis that you ¡don’tknow. Just because someone is maybedisadvantaged ¡in ¡certain ¡ways, that doesn’t mean ¡that they aren’t proactiveabouttheir health. And,in ¡fact, demographics -‑ age, education, income, gender -‑ account for aboutfiveor six percentof the variation ¡ in patient ¡activation scores. But ¡it ¡is there.it ¡is just ¡not ¡very powerful.Welooked at ¡the same question per health literacy and it isabout twenty-‑fivepercentof the variation is accounted ¡for by thosesame variables. 4
Here’s the measure. A as you can see there arejust declarativestatements about-‑-‑ and ¡ they’re ¡very general-‑-‑ that peoplemight make abouttheir health and they respond with degrees of agreement or disagreement. 5
We were able tosee that there appears tobe levels or stages that people gothroughontheir waytobecoming effective ¡self-‑managers. Andveryearlyonwe didin-‑depth ¡interviews with ¡ people along this dimension ¡whounderstand ¡how dotheyunderstand ¡their role, how dothey cope and respond. Andit was quite informative. We sawthat people whomeasuredlowonthis scale theyhadmuchmore experience withfailure;theywere muchmore likelytosaythings like, “It doesn’t reallymatter what I do. I can’t have a positive impact onmyhealth.” Theywere discouraged. Theywere overwhelmed ¡with ¡the taskof managingtheir health. Theyhad ¡very poor or low competence and ¡manyhad ¡poor problem-‑solving skills. Andthenthe other surprising thing was manyof them didn’t reallyunderstand ¡what their role was.Theythought it was tobe passive inthe medical ¡encounter. All ¡of that has lot of implications, whichI’lltalk about at the end-‑-‑ about what does that mean ¡for engaging people with ¡information. We ended ¡up using a Rasch analysistocreate the measure. There are twocharacteristics I want to just mentionhere. One is that it is interval-‑level ¡measurement. Soit is more like ruler with equal distance between ¡the marks on ¡the ruler and ¡that means that the measure is more precise and ¡consistent than ¡most social-‑science-‑basedmeasures. The mathof it tells us that we’re actuallytapping intoone underlying idea. The math ¡doesn’t tellus what that idea is, but what we’re measuringis a person’s self-‑concept as a managerof theirownhealth. Andit may not be conscious on ¡the individual’s part, but that’s kind ¡of what they’re telling us in ¡answering these questions. At this point the measure’s been ¡translated ¡intotwenty-‑three different languages and ¡we have been ¡able toevaluate the psychometricproperties of about half of those translations. Andsowhat we see isthat concept itselfis robust andseems toworkacross culture andlanguage. 6
Atthis pointthereare over 200 publicstudiesthatquantify patientactivation ¡and ¡what we see, generally,is that prior activated individualsaremorelikely to engagein positive behaviorsand ¡havebetter health ¡outcomes. 7
Just to give you an example of how it translatesinto behaviors,wehavelooked at the percentage of peopleat each level of activation engagein behaviors.Thisisthe behavioral domain ¡of managinghypertension:“Do you ¡takeyour medication ¡as recommended,know whatyour blood pressureshouldbe?”astaking moreownership; monitoring, thatis, beingmore proactive;and ¡keepinga diary,even ¡moreso. Looking at lots of behavioral ¡domainswesee thiskind of stair -‑step approach wherethehigher activated aremore likely to engagein the behavior and theless activated, less likely to. But we also saw that as the behaviorsbecome moredifficult,requiresustained ¡action, less people in all the levels actually down then. 8
Here’s howpeople behavein ¡the medical encounter.Howdo you ¡get a new prescription? Do you ¡read ¡aboutsideeffects? Do you ¡bringalist of questionsto your office visit? When ¡you ¡don’tunderstand,areyou ¡persistentin ¡askinguntil you ¡do? And ¡ do you ¡look at the doctors’qualificationswhen ¡choosinga new doctor? 9
Whatwe saw was that the higher activated ¡did ¡most of these behaviors,butthe others were moreof a stair-‑step approach.So after lookingatlots of these behavior mapsand lookingat who doeswhat, what we came away with was the insightthat a lot of the behaviorsthatwe’reasking peopleto do areonly doneby this highestlevel of activation. So when ¡we focuson ¡complex and ¡difficultbehaviors,firstand ¡only,maybe we’re discouragingthosewho areless activated. And ¡when ¡we give peopletoo much ¡ information or suggesttoo many changes arewe discouragingtheleastactivated and, essentially, are we setting them up for failure.Whatwewant to do,really,is set people up for success. So what we took away fromthiswas let’s start with behaviorsthatare morefeasible for peopleto takeon, and ¡that mightmean ¡breakingthem down ¡into small steps, and in doingso we can increase the opportunity forpeopleto experience success. And whatwe have observed isthat when people do experiencesuccess, even if there’rejustsmall steps, their motivation goes up.And then they’remoreready to take on that nextchallenge. So thiswas all sortof empirical,lookingat the data and thinkingaboutwhatdoes this mean? How should we proceed? I am goingto jump ahead hereto some of the research on howthis relateto outcomes. 10
This isa study that was published ¡afew yearswherewe wereworkingat the large delivery system,where they werecollecting patientactivation scoresfrompatientsthat went into the electronic medical record.Wehad ¡the opportunity to seehow doesthe patient activation scorerelate to all ¡of the quality metrics and the electronicmedical ¡ record? Thisiscontrollingfor age,income, gender and chronicillness. It is only telling us whatwas statistically significant.It isnot tellingus the magnitudeof the relationship.Buteverythingisin the expected direction.So peoplewho are more activated aremore likely to get screenings,they’re less likely to beobese, less likely to be smokers,less likely to havecostly utilization and to havetheir clinical ¡indicators within ¡rangefor most of these measures. Later we had ¡the opportunity to follow peopleover time in the same delivery systemand here we can see a little bit more aboutthe magnitudeof the relationship. 11
We’relookingatthat 2010 PAMscorepredictingoutcomesattwo years later It is comparingpeople ¡who are ¡at level one ¡to the ¡other three ¡levels. And,for example,if you look at the PHQ-‑9-‑-‑ thatis, arepeople in normal rangein the PHQ-‑9,they were ¡ abouttwo times more likely to bein ¡the normal rangetwo yearsago that had ¡a PAM scorethat was a level four.So that’s how you ¡would interpretthis.So you still see that kind of stair -‑step even though this is multivariateanalysiswe’recontrollingfor all the thingsI mentioned before. 12
Recommend
More recommend
