What is the best way for providers to ask patients about antiretroviral adherence?

Transcription

1 Wht is the best wy for providers to sk ptients bout ntiretrovirl dherence? Wynne Cllon, BA, Somnth Sh, MD, MPH, 2 P. Todd Korthuis, MD, MPH, 2 Ir B. Wilson, MD, 3 Richrd D. Moore, MD, MHS, Jonthon Cohn, MD, 4 Mry Ctherine Bech, MD, MPH Johns Hopkins University, Bltimore, Mrylnd; 2 Oregon Helth & Science University, Portlnd, OR; 3 Brown University, Providence, RI; 4 Wyne Stte University, Detroit, MI MOTIVATION MIXED METHOD ANALYSIS: QUANTITATIVE & QUALITATIVE DISCUSSION Previous literture reports 9-53% nondherence rte to ARVs Physicin s estimtes of ptient dherence re highly PROJECT inccurte AIMS Exmine how often ptients request ressurnce from their physicins Exmine the types nd nture of these in clinicl settings requests METHODS No previous studies directly ssessing the optiml wy to sk bout ARV non-dherence Understnding optiml ptient-provider communiction is necessry first step to llow for effective SDM STUDY AIMS. Assess how providers sk bout ARV dherence in clinicl setting 2. Determine the types of questions tht elicit higher rtes of ccurte ptient disclosure of non-dherence STUDY SAMPLE & SETTINGS Prent study: Enhncing Communiction nd HIV Outcomes (ECHO) Studies - Audio-recorded encounters of 434 providers nd 45 ptients - Sites: Bltimore, NYC, Portlnd, Detroit Our smple: Ptients reporting non-dherence in 3-dy recll conducted by reserch ssistnt - 34 providers nd 58 ptients. Isolted ptient-provider dilogue bout ntiretrovirl medictions 2. Identified ll provider questions nd noted ptient response 3. Converstion nlysis: identified provider question ctegories inductively 4. Coded ll provider questions into different question types 5. Clculted reltive risk of ccurte disclosure by question type 6. Multi-vrite nlysis djusting for question order nd whether leding or not PATIENTS (n = 58) - 45% femle - 67% blck, 2% white - 66% HS degree - 90% > yer with provider; 43% > 5 yers with provider ptient discloses non-dherence fter clinicin sks n=32 6 questions Smple Chrcteristics Dilogue Chrcteristics clinicin lredy knew bout nondherence n=8 58 visits with nondherent ptient ptient does NOT disclose nondherence fter clinicin sks n=6 4 questions Open-ended How you been doin on tking your medictions? Clrifying Are you still tking the Combivir, Vired, nd Sustiv? So re you on the sme medicines still? RESULTS PROVIDERS (n = 34) - 50% femle - 7% white, 8% Asin, 9% blck - 74% MDs, 8% NPs, 9% PAs unprompted ptient disclosure of non-dherence n=8 clinicin does not sk bout ARVs n=4 Question Types Disclosure Rtes by Question Type Question Type Multivrite Anlysis Disclosed Brod (n = 2) 4 (33%) Clrifying (n = 23) 3 (3%) Positively-frmed (n = 7) 5 (29%) Negtively-frmed (n = 23) 20 (87%) Undjusted Adjusted for type of drug Reltive Risk 95% CI Reltive Risk 95% CI Negtivelyfrmed No Yes 3.77*** *** Order ** Leding No Yes.69* *p<0.0, **p<0.05, ***p<0.00 Positively-frmed Do you tke your HIV medicines everydy? Negtively-frmed Hve you missed ny doses in the lst four weeks? Do you miss sometimes? Missed vs. Tking - Missed is specific: provider s intent is cler; less mbiguity - Missed is normlizing: gives ptient implicit permission to disclose Brod questions - Non-specific wht is the provider sking? - Provider s intent not cler Clrifying - Asked with gol of mediction reconcilition - Aim is not to ssess dherence LIMITATIONS Smple limited to ptient who dmitted non-dherence in post-encounter interview CONCLUSIONS A substntil minority of providers did not elicit disclosure of non-dherence Brod, clrifying, nd positively-frmed questions re not sufficient to elicit ccurte disclosure Asking directly bout missed doses increses likelihood of ccurte disclosure Honest discussion of ptient s dherence sttus is n importnt first step to llow for subsequent SDM

2 WHY ARE YOU GIVING ME THIS? PHYSICIAN-PATIENT RELATIONS IN COLONOSCOPY SCREENING AMONG AFRICAN-AMERICANS, BALTIMORE, MD Zchry Obinn Enumh, MA, Rebkh Atnfou, MPH, Robert Blum, MD, PhD, MPH Johns Hopkins Urbn Helth Institute (UHI) ABSTRACT Colorectl cncer (CRC) is the third leding cuse of new cncer cses nd second leding cuse of cncer mortlity. Moreover, Africn-Americn women nd men re both more likely to die from colorectl cncer thn their white counterprts. In qulittive study, nine focus group discussions were conducted with 27 prticipnts t distinct centers cross Bltimore. As whole, prticipnts expressed tht the following ll contribute to promotion of CRC screening : ) fmily, friend, nd socil support, 2) the impct of hving fmily member with CRC, 3) symptomology, 4) positive physicin-ptient reltionships, nd 5) dvertisements (rdio, TV, flyers). Prticipnts lso noted mny brriers nd obstcles to completing CRC screening: ) chllenges with colonoscopy preprtion, 2) colonoscopic procedurl nd test-specific concerns, 3) insurnce nd cost concerns, 4) generlized fer, nd 5) poor reltions with physicins nd lck of physicin recommendtion. An understnding of key promoters nd brriers cn led to improved shred decision mking between ptients nd providers. Deeper converstions on bowel preprtion nd test specific concerns my improve screening rtes. Interventions trgeted t incresing eduction, knowledge of CRC, nd open converstions between ptients nd providers my be vible options to improve on the current CRC trends nd ssist in decresing the rcil disprity in CRC nd CRCS. METHODS PROMOTERS BARRIERS Grounded Theoreticl Approch Nvivo for Mc Open Coding Comments coded to vrious themes (fer, insurnce concerns, support systems, etc.) Source: QSR Interntionl Keywords: colorectl cncer, urbn popultion, screening, physicin-ptient reltions BACKGROUND Colorectl Cncer (CRC) is the third leding cuse of new cncer cses in USA CRC is the second leding cuse of deth in USA (50,30 in 204) Africn-Americns more likely to die from CRC thn other ethnicities Estimted 35,000+ cses of CRC dignosed in 204 Over 30% of Africn-Americns hve never been screened for CRC RESULTS 27 totl prticipnts Averge ge = 58 yers old Over 95% of prticipnts were 45+ yers old Approximtely 6% of prticipnts hd obtined CRCS Mix of insured/uninsured ptients CONCLUSIONS OBJECTIVES Develop deeper sensibility of colorectl cncer nd colorectl cncer screening burden mong Bltimore residents Determine mjor promoters nd brriers to obtining nd/or completing colonoscopy mong Est Bltimore residents Physicl nd personl brriers persist tht impct ptients bilities to both obtin nd complete colonoscopy mong Africn-Americns in Est Bltimore Interventions bsed on known promoters nd brriers cn led to improved physicinptient understndings nd reltionships in n effort to promote incresed screening rtes. Current work is underwy to explore how questions of fer, notions of msculinity, nd culturl ttitudes ffect CRCS, s well s the role tht medi cmpigns nd motivtionl interviewing my ply in improving screening rtes for CRC. Develop deeper sensibility on how knowledge of promoters nd brriers to CRCS cn led to more informtive nd beneficil physicin-ptient reltionships Acknowledgements: This project would not hve been possible without the support of the Johns Hopkins Urbn Helth Institute (UHI), Mrylnd Deprtment of Helth nd Mentl Hygiene, the Mrylnd Cigrette Restitution Fund, nd the Johns Hopkins Sidney Kimmel Cncer Center.

3 HIV Ptient-Provider Discussion of End-of-Life Helthcre Preferences Amy Knowlton, Mry Mitchell, Jenne Keruly, 2 Cynd Rushton, 3 Nncy Hutton, 2 Jennifer Wolff, Tom Smith 2 Johns Hopkins Bloomberg School of Public Helth, 2 Johns Hopkins School of Medicine, 3 Johns Hopkins School of Nursing (Sponsored by R0 NR04050 & P30 AI09489) Contct: knowlt@jhu.edu Objective Advnced cre plnning (ACP) is importnt for ensuring preferred cre t end of life (EOL) nd is now legislted in Mrylnd. Yet ACP is seldom ddressed by medicl providers, prticulrly with vulnerble Africn Americn ptients. Smple chrcteristics nd undjusted nd djusted odds of discussing end of life cre preferences with medicl providers mong HIV ptients with drug use disorders in the AFFIRM study (203-8). (Multivrite logistic regression) Undjusted Adjusted N (%) or Odds Rtios Odds Rtios Vribles Men (SD) [95% CI] [95% CI] Fmily history of disgreement in helthcre 2.26** 2.52*** 84 (30.4) of n incpcitted fmily [.34,3.8] [.45,4.39] Among smple of vulnerble, comorbid JHU outptient clinic ptients, only minority (4%) reported tlking to their doctor bout helthcre Evidence is needed to member (yes) preferences if they were inform provider trining in ACP communiction. Fmily problem solving (gree) 9 (69.5).57 [0.92,2.68] 2.08* [.6,3.7] incpcitted. Fmily fctors of history We exmined HIV ptient of disgreement in nd fmily fctors ssocited Physicl functionl helthcre of n with EOL cre discussion with limittions incpcitted fmily doctors. member, nd fmily Methods A lot 42 (5.2).86 [0.90,3.85] 2.6* [.00,4.65] problem solving, were ssocited with AFFIRM Study prticipnts (n=276) were Moore HIV clinic ptients who currently or formerly used hrd drugs. 97% were Africn Americn nd 30% reported detectble virl lod. EOL cre preference discussion with providers ws defined s (yes/no): Hve you ever tlked with your doctor bout medicl tretments you would wnt if you weren t ble to mke decisions for yourself? Fmily history of EOL helthcre disgreements ws defined s (yes/no): Hs your Somewht 3 (47.5) [0.82,2.37] [0.93,2.92] None 03 (37.3) Reference Reference Age (Men [SD]) 53 (6.2) * [0.93,.00] [0.92,.00] Sex (femle) 20 (43.5).2.06 [0.76,.92] [0.65,.74] Eduction (> high school) 56 (20.3) [0.46,.53] [0.44,.6] Discussed end-of-life helthcre preferences with your doctor 3 (40.9) Conclusions discussion of their own personl preferences in EOL helthcre. Greter reported impirment in physicl functioning ws lso significnt. Age ws negtively ssocited with EOL cre discussion with medicl providers. Anlysis controlled for sex nd eduction, which were not sttitsiclly significnt. fmily ever hd mjor disgreements bout medicl tretment for fmily member who ws criticlly ill nd not ble to tell doctors wht tretment they wnted? Multivrite logistics regression ws used. Even in the context of stte legisltion (Medicl Orders for Life Sustining Tretment; MOLST) promoting dvnced cre plnning, only minority of vulnerble, comorbid ptients in our study reported EOL cre discussion with their helthcre provider. Exploring fmily history of conflict over EOL helthcre of fmily member, nd ddressing ACP s mtter of fmily problem solving, mong this predominntly Africn Americn ptient popultion, my fcilitte providers discussion of ACP with them. The ge effect my be explined by lower tretment dherence, nd thus helth outcomes, observed mong younger vulnerble HIV ptients.

4 The Johns Hopkins After Cre Clinic: A Bridge to Cre Fernndo Men-Crrsco, MSW, BSN, RN, Sophi Pemberton, MSN, BSN, CRNP, Arjun Chnmugm, MD, MBA2, Roslyn Stewrt, MD, MS, MBA Johns Hopkins University, Bltimore, Mrylnd, USA. Deprtment of Medicine 2. Deprtment of Emergency Medicine The Centers for Medicre & Medicid Services estimtes ntionl redmission rte to be ~8% More thn $7 billion nnully for voidble redmissions. 20% of ptients experience preventble dverse events.2 Post-dischrge cre is unvilble to most highrisk ptients. Innovtion Description Retrospective cohort nlysis of Electronic Medicl Record nd billing dt. Ptients were strtified bsed on helthcre utiliztion, for scheduled ACC ppointment between April 20 nd November 30, 205. Ptients were clssified into 2 groups: Cohort - ptients who completed n ACC visit. Cohort 2 - ptients who were scheduled but did not complete their ACC visit. A totl of 706 referrls were included: 80.9% cme from the ED. 8.6% from HUs. Approvl to conduct this study ws given by the Institutionl Review Bord (IRB). Vribles ACC Complete (N=752) (44%) ACC NOT Complete (N=954) (55.9%) All Ptients (N= 706) PRE ED Visits.0 (.0).3 (.8).2 (.5) PRE HU Visits 0.3 (0.5) 0.3 (0.5) 0.3 (0.5) PRE Redmitted 29 (3.9%) 80 (8.4%) 09 (6.4%) POST ED Visits 0.2 (0.7) 0.5 (.9) 0.4 (.5) POST HU Visits POST Redmitted 0.0 (0.2) 0. (0.3) 3 (.7%) 9 (2.0%) Sttisticl Anlysis Chnges in count for ED nd HU visits were evluted using Poisson regression nlyses djusted for ge, gender, rce, nd referrl source. Cost nlysis ssumed tht the verge cost per ED visit t JHH is $,589, nd tht the verge cost per HU visit t JHH is $5, (0.3) ACC NOT DID: Difference ACC Complete Complete in Rtios ACC Rtio of Post Rtio of Post vs Complete Utiliztion vs Pre vs. NOT Pre Type (95% CI) (95% CI) (95% CI) Visits Sved Totl Cost p vlue Svings ED Visits 0.8 (0.5 to 0.22) 0.39 (0.35 to 0.43) 0.48 (0.38 to 57.9 ED 0.58) $250,935 <0.00 HU Visits 0.7 (0.2 to 0.24) 0.32 (0.25 to 0.4) 0.52 (0.34 to 33.9 HU 0.8) $54, (.9%) Totl Svings: $765,65 One FY Qurter Totl Svings: $327, Bonus Findings Ptient Demogrphics Sex: 54% mle, 46% femle. Rce/Ethnicity: 65% Blck, 6.4% White, nd 8.6% Other. Medin ge: 46.8 youngest 8, oldest % insured, with the lrgest pyer source being Medicid (Figure 2). Estblished >500 (~ /3) people with primry cre within JHH. 97% of men/women who hve given feedbck fter their visit report being very hppy with the services provided by the ACC. Tble 2 Visit & Cost Svings Attributed to ACC (Difference in Differences) Tble PRE vs POST ACC Difference in Utiliztion Study Design Discussion Socil Work 5% Phlebotomy 3% Nursing 42% Phrmcy 30% Commercil Insurnce 8% BCBS 5% Self Py 26% Medicre 8% Medicid 8% Medicid MCO 25% Figure 2 Pyer Distribution The Johns Hopkins Hospitl implemented new comprehensive cre coordintion model: The After Cre Clinic (ACC). Clinic gol is short-term intensive interventions prior to (re)enggement with medicl home Hypothesis: The ACC will significntly reduce emergency deprtment (ED) nd hospitl unit (HU) utiliztion for t-risk ptients. Follow-up cre for ptients dischrged from the ED or HUs. Trget ptient popultion is people t risk for: Poorly coordinted cre Being lost to follow-up Procedurl complictions Redmission People with no primry cre physicin (PCP) People unble to quickly see specilists Stffed by interprofessionl tem tht collbortively focuses on (Figure ): Ptient-fmily disese eduction Mediction reconcilition Self-cre skills Coordintion of primry nd home cre services ACC tem members: finncil counselors, medicl ssistnts, nurses (RNs), socil workers, Phrm-Ds, ED nd internl medicine MDs, medicl residents, nurse prctitioners (NP), NP students, nd medicl students. Methods Figure Interprofessionl Collbortion* Bckground *Medicl Providers see 00% of the Ptients Outcomes Primry Ptients who struggle with mnging their helth void further ED & HU visits (Tble ). The ACC model sved the Johns Hopkins Systems >$765,000 by verting ED nd HU visits (Tble 2). Limittions Secondry Pyer distribution includes privte, public, nd sfety net sources (Figure 2). The clinic hs independently hd positive return on investment of ~6%. Currently operting t cpcity with vilble stff nd clinic spce. Engging ptients prior to ACC visit hs been chllenging. Out of,706 ptients referred, 752 completed their ACC visit (44% show rte) Inbility to confirm ppointments for ptient who lck working telephone number. Trnsitionl medicine is importnt nd positively effects ptient outcomes. Smooth nd effective trnsitions of cre trnsltes into sustinble pthwys into primry cre for underserved ptients. The ACC model cn be replicted cross helth systems nd diverse ptient popultions. This model helps students: Negotite complex clinicl & socil situtions, demonstrte professionl responsibilities, dhere to ethicl principles, nd demonstrte sensitivity to diverse ptient popultion. Future Steps Refine dt collection efforts to better understnd lnguge nd culturl profile of ptient popultion. Develop nd implement ACC expnsion: Increse clinic spce nd stff Increse pre-acc ptient enggement Assign trnsition-guide RNs t the time of referrl in ED nd HU. Identify t risk ptients prior to HU dischrge by collborting with inptient cre tems. Recruit bilingul/biculturl stff. Cpitlize on phrmcists expertise by providing home phrmcy services for ACC ptients who hve complex medicl profiles. References. U.S. Deprtment of Helth nd Humn Services. (7 My 204). New HHS Dt Shows Mjor Strides Mde in Ptient Sfety, Leding to Improved Cre nd Svings. 2. Jencks Sf, et l.,. N Engl J Med 2009;360(4):48-28.

5 Correltes of End of Life Helthcre Decision Mking with Friends nd Fmily Mry M. Mitchell, PhD, Amy R. Knowlton, ScD, Cynd H. Rushton, PhD 2, Thoms J. Smith, MD 3, Jenne C. Keruly, MS 4, Nncy Hutton, MD 5, Jennifer Wolff, PhD 6 Johns Hopkins Bloomberg School of Public Helth, Deprtment of Helth, Behvior nd Society, Bltimore, MD.2 Deprtment of Acute nd Chronic Cre, School of Nursing, Johns Hopkins University, Bltimore, MD, 3 The Johns Hopkins School of Medicine, Deprtment of Oncology, Bltimore, MD, 4 School of Medicine, Division of Infectious Diseses, Johns Hopkins University, Bltimore, MD, 5 School of Medicine, Peditrics, Johns Hopkins University, Bltimore, MD, 6 Johns Hopkins Bloomberg School of Public Helth, Deprtment of Helth Policy nd Mngement, Bltimore, MD Bckground Smple Description of Logistic Regression Anlyses Study Numerous reserchers hve concluded tht end of life cre improves both pin nd symptom mngement, s well s mentl helth. (Hrding et l., 2005; Hung, 203). From our work with disdvntged, current or former drug-using persons living with HIV/AIDS (PLHIV) nd their cregivers, we hve found tht PLHIV often rely on their friends nd fmily for informl cregiving support, including discussion of medicl tretments nd helth dvice. Africn Americns, compred with whites, discussed end of life tretment less with their medicl providers, while femles communicted more thn mles (Wenger et l., 200). There hs been very little reserch ttention to plliitive cre discussions with friends nd fmily members, one or more of whom re likely to be medicl decision mkers when the PLHIV is incpcitted. Our objective for the present study ws to exmine biopsychosocil correltes of PLHIV hving end of life helthcre discussions with their fmily nd friends. Procedure Dt were from the AFFIRM study ( ), which exmined socil environmentl fctors ssocited with helth outcomes nd well-being mong disdvntged PLHIV nd their informl cregivers. Prticipnts recruited from HIV/AIDS helth cre hospitl-bsed clinic. Selection criteri included being dult, HIV seropositive, being either current or former hrd drug user (e.g. cocine, heroin), nd being willing to invite one s min supportive tie(s) to prticipte in the study. The study ws pproved by the Johns Hopkins University IRB nd both PLHIV nd their informl cregivers completed consent forms prior to study prticiption. SD = Stndrd Devition b HIV Tble I. Chrcteristics of HIV cregivers nd former or current injection drug using cre recipients nd their ssocitions with Bckground Vribles Mesures SD = Stndrd Devition b HIV Tble I. Chrcteristics of HIV cregivers nd former or current injection drug using cre recipients nd their ssocitions with Outcome Mesure: Hve you ever tlked with your fmily or friends bout wht medicl tretments you would wnt if you were not ble to mke decisions for yourself? Predictor Mesures: During the pst 30 dys, how much did pin interfere with your norml work or ctivities, including both work outside the home nd housework? Life stisfction scle? Dt Anlysis SD = Stndrd Devition b HIV Tble I. Chrcteristics of HIV cregivers nd former or current injection drug using cre recipients nd their ssocitions with Frequencies nd mens clculted in SPSS v23. Clculted Cronbch s Alph for the Life Stisfction Survey. Rn undjusted logistic regression nlyses. N(%) Sex (Femle) 20 (43.5) Rce/Ethnicity (Africn Americn/Blck) 268 (97.) Eduction (more thn high school) 56 (20.3) Current Drug Use (heroin, cocine, or 02 (37.0) stimulnts) Men(SD) Age 53 (6.2) Independent vribles significnt t p <.0 were retined nd tested in the finl model. SD = Stndrd Devition b HIV Tble I. Chrcteristics of HIV cregivers nd former or current injection drug using cre recipients nd their ssocitions with Vrible OR (95% CI) AOR (95% CI) Pin interference.29 (.05,.58)*.33 (.06,.66)* Stisfction with.0 (.00,.02).02 (.00,.03)** life scle Sex (Femle) 2.8 (.72, 4.59)*** 2.63 (.59, 4.36)*** Age.97 (.93,.00).96 [.92,.00] OR = Odds Rtio, AOR = Adjusted Odds Rtio, CI = Confidence Intervl p <.0, * p <.05, ** p <.0, *** p <.00. Frequencies of Predictors SD = Stndrd Devition. b Cronbch s Alph =.92. c Rnge for ech item: 0-0. d Rnge for item is 0 = not t ll to 4 = extremely. Independent Vribles Men (SD) Life Stisfction Summed Scle b, c, d 84.2 (9.) Question stem, How stisfied re you with Meeting bsic needs 7.9 (2.3) Mentl helth 7.4 (2.4) Physicl helth 7.2 (2.3) Achievements in life 7.3 (2.4) Reltionships with friends 7.5 (2.4) Feeling sfe outside your home 7. (2.6) Being ble to do things 8.0 (2.2) Living qulity life 8.5 (2.) Spiritulity/religion 8.5 (2.0) Life s whole 8.2 (2.) Chronic pin e.4 (.2) SD = Stndrd Devition. b Cronbch s Alph =.92. c Rnge for ech item: 0-0. d Rnge for scle: 0-0. e Rnge for item is 0 = not t ll to 4 = extremely. Acknowledgements: The study ws supported by grnts R0 DA0943 nd R0NR04050, nd P30 AI09489 from the Ntionl Institutes of Helth. Slightly more thn hlf (57%) hd discussed medicl tretments they would wnt if they were not ble to mke decisions for themselves. Ech one-point increse on the chronic pin item ws ssocited with 33% increse in the odds of hving hd discussion on end of life helthcre issues with friends or fmily. Ech one-point increse on the life stisfction scle ws ssocited with 2% increse in the odds of hving hd discussion with friends or fmily. Femles were.63 times more likely thn mles to hve hd n end of life discussion with friends or fmily. Ech yer increse in ge ws ssocited with 4% reduction in the odds of hving hd n end of life helthcre discussion. Conclusions & Implictions SD = Stndrd I. Chrcteristics of HIV cregivers nd former or current injection drug using cre recipients nd their ssocitions with Nerly hlf of the PLHIV in our study hd not hd end-of-life tretment discussions with their friends or fmily members, one or more of whom would likely ssume the role of primry medicl decision mker if the individul were incpcitted. It is likely tht PLHIV with more chronic pin hd poorer helth, which my hve prompted hving medicl converstions. It is possible tht femles nd individuls who hve greter life stisfction hve closer reltionships with friends nd fmily, which my hve prompted these converstions. These fcilittors nd hindrnces my be considertions in trgeting PLHIV for interventions to encourge discussions with their friends nd fmily tht focus on their wishes for medicl tretments if they were rendered incpcitted. Hrding, R., Krus, D., Esterbrook, P., Rveis, V. H., Higginson, I. J., & Mrconi, K. (2005). Does pllitive cre improve outcomes for ptients with HIV/AIDS? A systemtic review of the evidence. Sexully trnsmitted infections, 8(), 5-4. Hung, Y. T. (203). Chllenges nd responses in providing pllitive cre for people living with HIV/AIDS. Interntionl journl of pllitive nursing, 9(5). Wenger, N. S., Knouse, D. E., Collins, R. L., Liu, H., Schuster, M. A., Gifford, A. L.,... & Shpiro, M. F. (200). End-of-life discussions nd preferences mong persons with HIV. JAMA, 285(22),

6 Developing stted-preference instrument to ssess the brriers nd fcilittors to the self-mngement of type 2 dibetes Allison H Okes, BA ; Vincent Grmo, MHS ; Jodi B Segl, MD 2 ; John FP Bridges, PhD,2 Johns Hopkins Bloomberg School of Public Helth ; 2 Johns Hopkins School of Medicine Bckground Successful self-mngement of dibetes is crucil to controlling disese progression nd preventing the onset of serious dverse events []. Previous reserch hs identified severl brriers nd fcilittors to effective self-mngement, however the existing literture hs not quntified these ttributes vi the ptient perspective [2,3]. Objective This study utilized n dpted 5-stge frmework for instrument development [4] to crete stted-preference instrument to ssess brriers nd fcilittors to selfmngement (Figure ). Through community engged pproch, the instrument ws designed to prioritize nd weight ech of the brriers nd fcilittors to the selfmngement of type-2 dibetes from the ptient perspective. Methods Overview The Dibetes Action Bord (DAB) ws creted to engge community members nd clinicl experts, while encourging dilogue between the two groups. A trgeted literture serch nd three community bsed focus groups were conducted to identify preliminry list of brriers nd fcilittors (Tble 2) [5]. This list of objects ws further refined during in-person semi-structured cognitive interviews of ptients with type 2 dibetes living in Bltimore. Prticipnts (n=25) completed combintion of rnking nd rting exercises to elicit their stted-preferences (Tble ). Dt Anlysis A self-explicted method (SEM), reltively simple sttedpreference pproch, ws used to combine ttribute importnce (rnking) with ttribute desirbility (rting).. Evidence Synthesis 2. Expert Enggement A trgeted literture serch ws conducted to identify existing brriers nd fcilittors to the selfmngement of type 2 dibetes nd to develop priors The Dibetes Action Bord (DAB) ws creted to engge clinicl experts nd encourge communiction with community members Tble Smple chrcteristics Chrcteristic Subjects (N = 25) Age Yers (men, rnge) 57 (3, 89) Gender Mle 9 White 2 Rce Blck 20 Hispnic 2 Time since dignosis Yers (men, rnge) 0 (0, 37) Dibetes Dignosed nd medicted 8 Tble 2 Rnking nd rting score by fctor Preliminry Objects Prior Rnk Rte Regulr ccess to helthy food +/ My personl understnding of dibetes +/ Communiction with my hcre provider +/ My cpcity to mnge my dibetes +/ My current helth insurnce Other helth conditions tht I hve Access to convenient plces to exercise +/ My fmily commitments My fith nd religious prctices My lnguge nd culture Resources in my locl community Access to n ctive support group Stress bout time commitments (--) Strong brrier (-) Mild (+/-) Neutrl (+) Mild (++) Strong fcilittor Figure 2 Brriers nd fcilittors vi stndrdized SEM Helthy food Personl understnding Communiction Cpcity to mnge Current helth insurnce Other helth conditions Plces to exercise Fmily commitments Fith nd religion My lnguge nd culture Resources in locl community Active support group Time commitments -4 Potentil -2 Brriers 0 Potentil Fcilittors Community Enggement 4. Pretest Interviews 5. Pilot Testing Community bsed focus groups (n=3) were conducted to further identify nd refine preliminry list of 3 brriers nd fcilittors Ptients with type 2 dibetes (n=25) completed rnking nd rting exercises using the preliminry fctors vi semistructured cognitive interview The refined list of items nd ech of their descriptive questions should be incorported into forml pilot study Conclusions Discussion Prticipnts rnked helthy food, communiction, nd personl understnding s hving the most positive impct on self-mngement. Lest positively rnked items included my lnguge nd culture, resources in locl community nd time commitments. Prticipnts rted the mjority of fctors s hving strong positive effect on their bility to self-mnge (Tble 2). There ws high correltion between rnking nd rting scores (rho=.93). Rnkings nd rtings were combined vi SEM, yielding eight sttisticlly significnt fctors (p<.00). The lck of vrince in the rting exercise led to difficulty in detecting brriers, however this ws ccomplished through stndrdized SEM scores (Figure 2). Bsed on feedbck, the list of objects ws further refined nd descriptive questions were dded for ech. Understnding brriers nd fcilittors from the ptient perspective cn help providers in crfting tretment plns with the specific im of improving self-mngement. The revised ttributes should be used in forml pilot nd future stted-preference studies. Although the SEM is well suited for eliciting preferences in the fce of mny fctors, the SEM does not cpture trdeoffs nd ws limited in its bility to detect brriers. Future studies should consider best-worst scling s elicittion method. References The uthors thnk the members of the DAB. Work ws supported through PCORI Methods Progrm Awrd ME nd through the JH-FDA CERSI (U0FD ). Reserch ws pproved through JHSPH IRB (600). [] Hs, L [2] vn den Arend, IJ [3] Glsgow, RE [4] Jnssen, EM [5] Purnell, TS. 206.

7 Ptient Enggement Trining: Helping Providers Improve Ptient Stisfction nd Outcomes Nicole Schechter, Trcy Novk 2, Susn Donovn 2, 3, Michelle Hwkins 2, Lur Torres 2, Teres Eyer 2, Crlessi Hussein 3, & Stephen T. Wegener Deprtment of Physicl Medicine nd Rehbilittion, Johns Hopkins University School of Medicine, Bltimore, MD 2 Johns Hopkins HelthCre LLC, Glen Burnie, MD 3 Beneficiry Advisory Council, Johns Hopkins Medicine Allince for Ptients LLC, Glen Burnie, MD BACKGROUND: Given the incresing burden of chronic disese nd the evolving focus on vlue-bsed helth cre, providers re now responsible for mngement of the ptient over time nd spce while ptients nd fmilies re required to tke more ctive nd centrl role in their helth cre. Unfortuntely, mny providers, ptients nd fmilies re unprepred for these demnds. Effective provider communiction skills re key in engging ptients nd fmilies in these efforts. The gols of this project re to develop, implement, nd ssess the effectiveness of stff nd provider trining progrm to improve ptient enggement nd stisfction in the Johns Hopkins Helth System. PROGRAM DESCRIPTION: Ptient Enggement Trining (PET) incorportes the principles nd skill set from motivtionl interviewing (MI) nd communiction science. Guiding principles re: Ptient Enggement Trining involves three phses. Plnning phse - work with the orgniztion s ledership to identify needs, gols, logisticl brriers, chmpions for mintennce phse, nd develop trining pln to mtch. Trining phse - provide one -hour grnd rounds session for ll stff, 4 to 7 hour bsic trining with skills demonstrtion nd prctice for clinicins, nd -hour trining for identified ptient enggement chmpions. Skill building nd mintennce phse - coch chmpions in the development nd implementtion of skill building pln with monthly nd/or qurterly ctivities, disseminte Tip of the Month vi emil, nd monitor the use of PET skills in trinees. Plnning Evidence-bsed for mny ptient popultions Time efficient Verstile Lerned by prcticing Tught by clinicins Trining Skill Building & Mintennce Ptient Enggement Trining (PET) is bsed on principles nd skills from Motivtionl Interviewing, Stges of Chnge, nd Communiction Science. Primry tennts of PET ssume tht ptients wnt to be helthy, providers wnt to provide ptient-centered cre, nd the focus is on the ptient-provider reltionship nd communiction. Through PET, trinees lern to estblish new ttitude or spirit to guide ptient interction nd set of communiction skills tht to foster rpport, shred decision mking nd promote motivtion for chnge. Lerning techniques include: interctive didctics, videos for skill modeling, pired prctice, trid prctice, smll group prctice, nd specific cse exmples with role ply. Stges of Chnge Motivtionl Interviewing Principles & Skills ACE: Acceptnce, Compssion Evoction OARS: Open Ended Questions, Affirmtions, Reflections, Summries Communiction Styles Informing Asking Listening Directing Guiding Following EVALUATION MODEL Lerner Outcomes Stisfction with trining Self-efficcy for using ptient enggement skills & principles Knowledge & ttitudes Behviorl mesure of effective utiliztion of skills Ptient Outcomes Helth cre & provider stisfction scores Rtes of cncelled or missed ppointments Hospitl dmissions for preventble complictions LEARNER OUTCOMES N= 350 providers trined t 5 different JHM entities 43 Ptient Enggement Chmpions trined Men Stisfction scores for initil trining = 4.7/5 00% trinees engged in mintennce ctivities led by chmpions CONCLUSIONS PET s gol is to improve provider communiction skills tht will fcilitte shred decision mking. Lerners re highly stisfied with the trining nd preliminry dt indictes improvements in self-efficcy, knowledge, ttitudes nd skills. The PET progrm ppers fesible within existing clinicl nd orgniztionl structures. Long term follow up is needed to determine if initil effects re sustined nd if improvements in clinicin skills trnslte into better outcomes for ptients.

8 RESEARCH POSTER PRESENTATION DESIGN 205 CHALLENGES OF COMMUNICATION ON PATIENTS PARTNERING: PERCEPTION OF PATIENTS IN AN ONCOLOGY SERVICE IN BRASILIA, BRAZIL Lriss Sores & Lriss Polejck, PhD University Hospitl of Brsíli, University of Brsíli - Brzil Bckground Communiction of the dignosis of cncer presents chllenging tsk for helth workers. It brings series of.2 feelings nd behviors, such s fer, nxiety nd isoltion, which cn be enhnced by the socil representtion of the disese s deth sentence. 3 Communiction plys centrl role on qulity of cre nd ptients enrollment on tretment is key spect for good outcomes. But how much informtion ptient wnts to know? How do they wnt to receive this informtion? How cn we gurntee ptients utonomy nd prepre the tem to know nd to respect their choices? Wht is fmily role on this process? Objectives This study imed to )Know nd nlyze the vision of the person treting cncer bout communiction; 2)Identify the min ptient s difficulties in the communiction process with helth providers;3) Identify if ptients wnt to be informed bout oncologic disese nd tretment options checking lso if they wnted tht fmily to be notified bout the disese nd the tretment nd; 4)Propose ctions tht improve helth communiction qulity in ttention oncology. Mteril nd Methods Explortory nd descriptive study of qulittive pproch, previously uthorized by the Committee Ethics in Helth Sciences School of University of Brsili, under No. 365,966 / 3. The firstphse involved literture review, which supported the preprtion of semi-structured interviews pplied in the second phse. The study enrolled ten ptients on tretment t University Hospitl from from August to December 203. The tools used were: A) Timeline:It is technique in which orients the prticipnt to drw line tht is to strt n event importnt. It is widely used in psychotherpy nd it imed to help to recover prticipnt s nrrtive bout their own experience with. B) semi-structured interview sociodemogrphic dt - sex, ge, mritl sttus, eduction, employment sttus nd perceptions bout tretment. Ptients s records were lso seen to check bout their clinicl condition, type of cncer, stging nd time since dignosis. The interviews were recorded nd trnscribed nd nlysed ccordingly Brdin s content Anlysis. The study included ten prticipnts (n = 0) of which six were women. The verge ge of respondents ws 49 yers nd rnged 3-67 yers. Schooling rnged incomplete elementry school to university full, with prevlence of complete high school (n = 6). Regrding clinicl stging, there ws prepondernce of people with dvnced disese (n = 7), chrcterized by tumors in stges III nd IV of different types of cncer.most prticipnts (n = 8) highlighted tht they wnt to receive informtion bout the disese. However, they do not feel ble to prticipte in the decision-mking process on the tretment (n = 9). Dt nlysis lso reveled six themtic ctegories: the importnce of knowing bout the disese; ccess to helth services; interest in knowing bout the disese nd prticipting in tretment; Fmily prticiption in tretment; brriers to communiction nd; Helth worker role s fcilittor of qulity helth communiction. Conclusions The reveltion of truth is expected by most study prticipnts, even involving bd news, indicting interest to be included in the communiction process bout the disese nd tretment Still remins the chllenge for the interdisciplinry tem to provide conditions building bonds, vluing the knowledge of the subject nd promoting utonomy., it is suggested tht studies further integrte interdisciplinry tem providing ssistnce to persons with cncer nd extend the study of the importnce of subjectivity on communiction process. It is urgent to prepre better helth providers to lern bout communiction nd to lern how to del with emotions nd personl experiences relted to illness. References For references plese check complete rticle: /4395 E-mil: Lrisspolejck@unb.br

9 Erly Evidence Humn-Centered Decision Aids Help People Mke More Approprite Cre Decisions INTRODUCTION Chronic low bck, hip, nd knee pin nd low-risk prostte cncer re often treted with surgery before other options. Terms like hernited disc led people to ctstrophize their pin. Men with low-risk prostte cncer often opt for surgery out of fer. An RCT showed multimedi progrm creted with I liked lerning the ntomy nd knowing tht most of the things we hve herd of being problems nd cusing pin re not relly cuses or much of concern. humn-centered pproch reduced nxiety.* Humn-centered multimedi decision ids cn lso ddress emotions, understnding, nd promote clm delibertion. METHODS Multimedi decision ids developed using humn-centered pproch were prescribed to people who needed to mke tretment decision to help them understnd their condition, options, the pros nd cons, trdeoffs, nd their vlues nd preferences. After viewing progrm, people could opt to tke survey. I feel encourged I might be ble to use exercise to help my sitution nd void surgery RESULTS 7,635 ptient surveys cross 5 EmmiDecide progrms found: Of the 8,87 ptients who took the survey fter viewing decision ids 97 % now understnd there s more thn one wy to tret their condition for chronic low bck, hip, or knee pin or low-risk prostte cncer 37-44% re now interested in less ggressive tretment like surgery: 3,779 with low bck pin (37%) 96 % now hve better understnding of the pros nd cons of their tretment options,797 with hip pin (43%) 92 % now hve better sense of which tretment(s) mke the most sense for them 2,492 with knee pin (40%) 749 with low-risk prostte cncer (44%)

10 Understnding Ptients Rel Concerns to Enble True Shred Decision Mking Ptients with ulcertive colitis re more concerned bout complictions of their disese thn side effects of medictions. How deep n understnding of UC & tretments do you wnt? 460 US nd Austrlin ptients with UC Thompson, K.D., Connor, S.J., Wlls, D., Gollins, J., Stewrt, S.K., Bewtr, M. Bumbltt, G.L., Holubr, S.D., Greenup, A., Sechi, A., Girgis, A., Rubin, 5% D.T., & Siegel, C.A. Inflmm Bowel Dis. 206 Apr;22(4): Strong desire 46% to be informed INTRODUCTION Clinicins sometimes misjudge ptient fers nd concerns. In the cse of ulcertive colitis (UC), physicins ssume ptients fer mediction side effects 2% the most nd focus on tht converstion. We sought to better understnd wht is most concerning to ptients nd how they would like to be informed bout tretment options. This ffects shred decision mking (SDM) discussions, where Wnt Bsics Wnt to be knowledgeble Wnt in-depth understnding surgicl options re often not discussed until they re the lst resort. This cn mke surgery more dngerous, limit surgicl options, nd does not chieve SDM. When thinking bout your ulcertive METHODS A web-bsed survey ws sent to UC ptients throughout the United Sttes (US) colitis, wht re you most concerned bout? nd Austrli (AUS). Ptients showed their strength of greement or disgreement using moment-to-moment ffect trce. Stndrd quntittive nlysis ws used for the survey results, nd cluster nlysis ws performed on the ffect trce responses. 37% 29% 4% 4% Number of ptients with UC (370 US, 90 AUS) tht responded to the survey 460 Men ge 49 (8-8) Incresed risk of colon cncer Possible need for n ostomy Unble to do the things I most wnt to Possible side effects At wht point do you believe surgery is resonble tretment option for UC? 4% Other 2% None of the bove 78% 87 % wnted to shre in tretment decisions with their doctors. 7% 4% 2% When you first lern tht you hve ulcertive colitis After trying medictions tht were not effective Only s lst resort Never

11 Intensivist-reported fcilittors nd brriers to discussing post-dischrge outcomes with ICU surrogtes: A qulittive study Alison E. Turnbull, DVM, MPH, PhD,,2,3, Wesley E. Dvis, BA,2, Dle M. Needhm, FCPA, MD, PhD,2,4, Dougls B. White, MD, MAS 5,6,7, Michelle N. Ekin, PhD,2 Outcomes fter Criticl Illness nd Surgery Group (OACIS), Johns Hopkins University, Bltimore, MD; 2 Division of Pulmonry nd Criticl Cre Medicine, Deprtment of Medicine; 3 Deprtment of Epidemiology, Johns Hopkins Bloomberg School of Public Helth, Bltimore, MD; 4 Deprtment of Physicl Medicine nd Rehbilittion, Johns Hopkins University School of Medicine, Bltimore, MD; 5 Center for Bioethics nd Helth Lw, Deprtment of Criticl Cre Medicine, University of Pittsburgh, Pittsburgh, PA; 6 Clinicl Reserch, Investigtion, nd Systems Modeling of Acute Illness (CRISMA) Center, Deprtment of Criticl Cre Medicine, University of Pittsburgh, Pittsburgh, PA; 7 Progrm on Ethics nd Decision Mking in Criticl Illness, Deprtment of Criticl Cre Medicine, University of Pittsburgh, Pittsburgh, PA Introduction Intensive cre unit (ICU) ptients expected postdischrge outcomes re rrely discussed in fmily meetings -4 despite this informtion being centrlly importnt to ptients nd their fmilies. 5 Intensivists reported tension between their professionl responsibility to discuss likely post-dischrge outcomes versus uncertinty bout their bility to predict those outcomes. They cited 3 min brriers: ) unrelisticlly optimistic expecttions for recovery mong ICU Objectives To chrcterize intensivist-identified brriers nd fcilittors to discussing post-dischrge outcomes with surrogtes (usully fmilies) of ICU ptients. surrogtes, 2) hving little or no contct with survivors fter ICU dischrge, nd 3) miniml confidence pplying existing outcomes reserch to individul ptients. Experience tlking to ICU surrogtes, seeing ICU survivors in the out-ptient setting, nd trusted reserch were identified s importnt fcilittors to discussing Mterils nd Methods likely post-dischrge outcomes with surrogtes. Qulittive study conducted vi one-on-one, semistructured telephone interviews with 23 intensivists from Tble 2: Themes nd illustrtive quotes 20 hospitls with ccredittion council for grdute medicl eduction (ACGME) criticl cre medicine progrms in 6 sttes. A limited ppliction of grounded theory methods were used to code trnscribed interviews nd crete txonomy of themes nd illustrtive quotes. Tble : Demogrphics of Interviewed Intensivists (n = 23) Chrcteristics N (%) Mle, n (%) 20 (87%) Specilty Medicine 9 (83%) Surgicl 3 (3%) Anesthesiology 2 (9%) Yers since medicl school grdution, medin (rnge) Willingness to discuss withdrwing life support score, medin (rnge) Current prctice region b 22 (6-46) 25 (0-40) Northest (48%) South 7 (30%) Midwest 4 (7%) Professionl responsibility Seeing ICU survivors s out-ptients Experience tlking to fmilies Outcomes reserch Unrelistic expecttions of surrogtes No contct with ICU survivors Lck of experience tlking to fmilies Outcomes reserch Fcilittors to discussing post-dischrge outcomes I hve responsibility s physicin to t lest mke conjecture to help give them some guidnce, prticulrly when they re trying to mke tough decisions. Intensivist 2 I hve enough experience with dischrging people nd seeing them follow up [in clinic] tht I cn mke tht judgement. Intensivist 9 [I m comfortble discussing post-dischrge outcomes] becuse I ve been prcticing in the intensive cre unit for 25 yers. I ve tlked with lot of ptients. Intensivist There is vrious disese dt published in terms of how people do t 3, 6, 9, 2 months down the rod bsed upon wht they re like when the leve the unit. Intensivist Brriers to discussing post-dischrge outcomes [One problem] is the misperception on the prt of lot of people tht somehow medicine hs got insights, knowledge, technologies nd powers tht only God hs. There is n unrelistic expecttion on the prt of the popultion. Intensivist 2 As criticl cre physicins we re seeing them for short window of cute illness nd then BOOM - we re shipping them off to the hospitl or we re shipping them off to rehb I only see them cutely in the ICU so I don t see the downstrem effects. Intensivist 4 I wsn t trined to think tht wy [bout post-dischrge outcomes] during my residency or fellowship trining. I ve hd to lern bout this in my own prctice. Intensivist 20 I don t think tht the stte of the science permits us to dequtely predict wht the future will hold from long term outcomes perspective for ptients. Intensivist 7 West (4%) Score derived from prticiption in previous tril on reported willingness to bring up the option of withdrwing life support. Possible scores rnge from 0 40 with higher scores indicted the prticipnt reported they would bring up the option under more hypotheticl scenrios. b Region is defined s the U.S. census region References. White DB, et l. Prognostiction during physicin-fmily discussions bout limiting life support in intensive cre units. Crit Cre Med 2007;35: Dvidson JE. Neglect of qulity-of-life considertions in ICU fmily meetings for long-sty intensive cre unit ptients. Crit Cre Med 202;40: White DB, et l. The lnguge of prognostiction in intensive cre units. Med Decis Mk 200;30: Scheunemnn LP, et l. How Clinicins Discuss Criticlly Ill Ptients Preferences nd Vlues With Surrogtes: An Empiricl Anlysis. Crit Cre Med 205; 43(4): Anderson WG, et l. A multicenter study of key stkeholders perspectives on communicting with surrogtes bout prognosis in intensive cre units. Ann Am Thorc Soc 205;2: Finncil Support This reserch ws supported by Ptient nd Fmily Enggement Erly-creer Investigtor wrd from the Gordon nd Betty Moore Foundtion.

12 Prepring Medicl ICU Proxies to Prtner in Shred Decision-Mking: A Pilot Study This reserch ws generously funded by the Johns Hopkins Division of Pulmonry nd Criticl Cre Medicine. Alison E. Turnbull, DVM, MPH, PhD,2,3, Anhit Rbiee, MD,,2 Mohmed Hshem, MD,,2 Mohmmd Sher, MD, An To, BS, Croline Chessre, BA, Rchel Coffin, RN, BSN, 2 Dle M. Needhm, FCPA, MD, PhD,2,4 Outcomes After Criticl Illness nd Surgery Group, Johns Hopkins University, Bltimore, MD, 2 Division of Pulmonry & Criticl Cre Medicine, School of Medicine, Johns Hopkins University, Bltimore, MD 3 Deprtment of Epidemiology, Bloomberg School of Public Helth, Johns Hopkins University, Bltimore, MD 4 Deprtment of Physicl Medicine nd Rehbilittion, School of Medicine, Johns Hopkins University, Bltimore, MD Introduction Most medicl intensive cre unit (MICU) ptients don t hve decisionl cpcity, leving proxies (usully fmily members) to prticipte in tretment decisions. -2 These proxies report Figure : Flowchrt Unique ptients screened 343 Tble 2: Interviewed proxies (N = 22) Age*, medin (IQR) 5 (39-6) Femle*, n (%) 83 (68%) Yers of eduction, medin (IQR) 4 (2-6) Tble 3: Proxy feedbck (N = 22) Sttement n, (%) Agree or Strongly Agree high levels of conflict with helthcre providers, 3-4 nd often experience depression, nxiety nd post-trumtic stress symptoms, 5-9 sometimes lsting for yers. Objectives To determine whether short booklet prepring MICU proxies to prtner in shred decision mking is cceptble, importnt, nd upsetting. Secondry objectives included ssessing concordnce between preferred use of life support nd ctul code sttus orders, Eligible ptients 238 Ptients with enrolled proxy Totl proxies interviewed 22 Not eligible 05 <24 hours in the MICU 56 No proxies visited ptient 42 No proxies spoke English 7 No proxy enrolled 27 Did not meet proxy in MICU 0 Proxy declined prticiption 9 No clinicin permission 8 2 enrolled proxies for single ptient Hve you ever supported loved one in n ICU before? Yes (%) 75 (6%) ICU dy of interview, medin (IQR) 3 (2-4) Self-identified rce, n (%) Blck or Africn Americn 55 (45%) White 55 (45%) Other 8 (7%) Withheld 4 (3%) Reltion to Ptient Wife/Girlfriend 3 (25%) Husbnd/Boyfriend 5 (2%) Dughter 3 (25%) Son 9 (7%) Other 36 (29%) *Missing for 2 proxies who did not wnt to shre identifying informtion The booklet is pproprite for dult friends nd fmilymembers of ICU ptients. It is importnt for fmilies of ICU ptients to know the informtion in the booklet. Prts of the booklet re upsetting. 20 (98%) 9 (98%) 54 (44%) nd exploring proxy understnding of their loved one s code sttus. Mterils nd Methods A booklet ws developed with input from members of the Johns Hopkins Hospitl Ptient nd Fmily Advisory Council, MICU socil work tem, Spiritul Cre tem, Ptient Reltions, nd the Legl Deprtment. The booklet ws written t 5th grde reding level. Proxies were eligible for enrollment 24 hours fter their loved one s MICU dmission nd remined eligible for 7 dys. We ttempted to enroll proxies dily for 2 hour period, 7 dys week. Englishspeking, dult fmily members who were present in the MICU were pproched with permission from the ttending physicin. After orl consent, the booklet ws red loud nd given to ech enrolled proxy. The Institutionl Review Bord of Johns Hopkins University pproved this study. Proxies were sked whether they considered the informtion in the booklet cceptble, importnt, nd upsetting. They were lso sked bout their loved one s preferred use of life support using stndrdized survey. Tble : Ptients chrcteristics by proxy enrollment sttus proxy enrolled No proxy enrolled* N = N = 27 P-vlue** Age, medin (IQR) 58 (48-69) 58 (45-68) 0.83 Femle, n (%) 56 (50%) 67 (53%) 0.82 Rce, n (%) Blck or Africn Americn 48 (43%) 7 (56%) White 49 (44%) 38 (30%) 0.07 Other 4 (3%) 8 (4%) Medin income of zip code, n (%)*** <$40,000 3 (28%) 47 (37%) $40,000 - $70,000 4 (37%) 46 (36%) $70,00 - $00, (20%) 25 (20%) 0.0 >$00,000 6 (4%) 7 (6%) Unknown or missing (%) 2 (2%) ICU sty in dys, medin (IQR) 5. ( ) 2.6 (.7-4.3) <0.00 Died in ICU, n (%) 20 (8%) 8 (4%) 0.53 ** Wilcoxon rnk sum test for continuous vribles, Chi-squred Test for ctegoricl *** US Census Bureu ; $4,89 medin household income for Bltimore City; $74,94 medin household income for Mrylnd Conclusions A booklet explining the role of proxies is cceptble nd importnt even mong fmily members who find it upsetting. Mny proxies in the medicl ICU do not know or understnd their loved one s code sttus. References. Prendergst TJ, et l. A ntionl survey of end-of-life cre for criticlly ill ptients. Am J Respir Crit Cre Med. 998;58(4): Torke AM, et l. Scope nd outcomes of surrogte decision mking mong hospitlized older dults. JAMA Intern Med. 204;74(3): Abbott KH, et l. Fmilies looking bck: one yer fter discussion of withdrwl or withholding of life-sustining support. Crit Cre Med 200;29: Schuster RA, Hong SY, Arnold RM, White DB. Investigting conflict in ICUs-is the clinicins perspective enough? Crit Cre Med 204;42: Jones C, et l. Post-trumtic stress disorder-relted symptoms in reltives of ptients following intensive cre. Intensive Cre Med 2004;30: Azouly E, et l. Risk of post-trumtic stress symptoms in fmily members of intensive cre unit ptients. Am J Respir Crit Cre Med 2005;7: Anderson WG, et l. Posttrumtic stress nd complicted grief in fmily members of ptients in the intensive cre unit. J Gen Intern Med 2008;23: McAdm JL et l. Psychologicl symptoms of fmily members of high-risk intensive cre unit ptients. Am J Crit Cre 202;2: Cmeron JI, et l. One-Yer Outcomes in Cregivers of Criticlly Ill Ptients. N Engl J Med. 206;374(9): Tble 4: Concordnce* between preferred use of life support nd code sttus In your opinion, how does [nme] wnt doctors nd nurses in the ICU to tret him/her? Which of the following sttements sounds most like wht [nme] would sy? Use life-support mchines to keep me live no mtter wht. If my hert stops, do CPR. Use life-support mchines to keep me live no mtter wht. But if my hert stops don t do CPR. Try to help me get better, but don t use life support mchines nd if my hert stops don t do CPR. Focus on keeping me s comfortble s possible, even if tht mens I die sooner. Code sttus t time of interview Full code DNR only DNR / DNI Totl (N = 03) (N = 9) (N = 0) (N = 22) (39%) 6 8 (7%) (8%) (9%) I don t know wht he/she would sy (25%) Refused (2%) Shded boxes defined s discordnt. Responses in 35 of 22 interviews (29%, 95% CI 2% - 38%) were discordnt. Tble 5: Proxy understnding* of code sttus t time of interview Which of the following best describes how doctors nd nurses in the ICU re treting [nme] right now? (choose )** If life-support mchine is needed they ll use it. If his/her hert stops they ll do CPR. If life-support mchine is needed they'll use it. But if his/her hert stops they won't do CPR. They re trying to help him/her get better, but they will not use life support mchines nd they will not do CPR if his/her hert stops. They re focusing on keeping him/her s comfortble s possible, even if tht mens he/she dies sooner. Code sttus t time of interview Full code DNR only DNR / DNI Totl (N = 03) (N = 9) (N = 0) (N = 22) (50%) (7%) 3 5 (4%) 3 5 (4%) I don t know (26%) Question not sked 0 0 (9%) *Proxies in 42 of 22 interviews (34%, 95% CI 26% - 44%; shded boxes) either did not know or understnd their loved one's code sttus. **The wording of response choices depended on the response to the previous question: "Is [nme] on life-support right now?"

13 Why re Some House Stff Uncomfortble Discussing MOLST Forms with Sndr E. Zeh, MD, MS ; Mrgret M. Hyes MD 2 ; Cynthi S. Rnd, PhD 3 ; Alison E. Turnbull, DVM, MPH, PhD 3,4,5 Deprtment of Medicine, Johns Hopkins University School of Medicine, BlNmore, MD; 2 Division of Pulmonry nd CriNcl Cre Medicine, Beth Isrel Deconess Medicl Center/ Hrvrd Medicl School, Boston, MA; 3 Division of Pulmonry nd CriNcl Cre Medicine, Johns Hopkins University School of Medicine, BlNmore, MD; 4 Outcomes Rer CriNcl Illness nd Surgery (OASIS) Group, Johns Hopkins University School of Medicine, BlNmore, MD; 5 Deprtment of Epidemiology, Johns Hopkins Bloomberg School of Public Helth, BlNmore, MD IntroducNon The Mrylnd Medicl Orders for Life Figure : Comfort discussing choices on MOLST from verses score on Tble 2: Themes iden@fied in interviews with uncomfortble house stff Sustining Tretment (MOLST) form test of knowledge the legl nd regultory issues pertining to Theme Quote cretes enduring orders for the use of life MOLST forms (n = 358) sustining therpies Rer hospitl I cre more bout wht the p0ent s gols dischrge. Stte lw requires tht ll dults dmiyed to n ssisted living fcility, nursing home, hospice, dilysis MOLST s pperwork disconnected from discussions with p@ents. of cre re nd I don t relly cre ll tht much bout the MOLST. The MOLST is sheet of pper tht the stte of Mrylnd require we fill out. center or home helth gency hve completed MOLST form. I think tht s n Dending convers0on nd I would feel I ws overstepping my In Mrylnd, house stff re leglly permiyed to complete MOLST forms. House stff perceived themselves s posi0on if I were to bring tht up. Some house stff re uncomfortble discussing the opnons on MOLST form unuthorized or inpproprite to conduct discussions bout orders on the MOLST form. There should be n expert tem tht knows how to hve rel convers0on bout those things tht s relly not the with pnents despite understnding the most efficient use of specilized people s legl nd regultory issues pertining to 0me. MOLST forms. To the p0ent it comes out of nowhere. ObjecNve To understnd why some house stff who understnd the legl nd regultory issues The red dshed line indictes the men nd medin score of 6. Widths of boxes re proportionl to the number of house stff in ech group. Discussing tretment preferences t dischrge ws perceived s being illogicl or surprising to the p@ent. It s like, let me sk you bout this relly difficult ques0on right now when you re leving the hospitl, in beder condi0on thn you cme in, nd now I m going to pertining to MOLST re uncomfortble tlk to you bout end of life. discussing the choices on MOLST form with Tble : Demogrphics of house stff with test score 6 by self- pnents. reported comfort discussing op@ons on MOLST forms I lern lot when we do it with the Mterils nd Methods We performed semi-structured, one-on- How comfortble do you feel with discussing the choices on MOLST form with p@ents?! Post grdute trining yer, N (%) Comfortble or very comfortble (N = 85) Neutrl, uncomfortble, or very uncomfortble (N=43) Interviewed (N=8) House stff vlued wtching plli@ve cre specilists model comple@ng MOLST forms with p@ents. plli0ve cre tem, so you consult plli0ve cre, you see wht they do nd then you re like: well, this is much beder wy to do this. one telephone interviews with 8 house stff t Johns Hopkins Hospitl. Intern Yer Second Yer Third Yer+ 47 (25%) 55 (30%) 83 (45%) 0 (23%) (26%) 22 (5%) 5 (28%) 7 (39%) 6 (33%) Conclusions Interviewed house stff demonstrted bove verge knowledge of the legl nd regultory issues pertining to MOLST forms relnve to their peers on n Post grdute trining progrm, N (%) Internl Medicine Obstetrics nd Gynecology Other SpecilNes 2 74 (40%) 6 (9%) 95 (5%) How frequently do you complete MOLST form with pnents? N (%) 2 (28%) 7 (6%) 24 (56%) 9 (50%) 4 (22%) 5 (28%) Severl fctors contributed to house stff discomfort including viewing MOLST forms s pperwork disconnected from pnent preferences, lck of uthority to complete MOLST forms, nd the Nming of MOLST form complenon. insntunon-wide survey. (Figure, Tble ) Trnscripts of interviews were coded using grounded theory to crete txonomy of themes nd illustrnve quotes. (Tble 2) Never Less thn once month >once month but <once week >once week but <dily At lest once ech dy 2 (6%) 47 (25%) 3 (7%) 77 (42%) 8 (0%) 5 (2%) 5 (35%) 8 (9%) (26%) 4 (9%) 0 (0%) 4 (22%) 3 (7%) 7 (39%) 4 (22%) Internl Medicine nd Obstetrics nd Gynecology residents were purposefully oversmpled within our interviewees s they were more likely to complete MOLST forms with ptients thn their collegues. 2 Other specilties included Generl Surgery, Surgery sub-specilties, Neurology, Psychitry, Emergency Medicine, nd Physicl Medicine & Rehbilittion. House stff my benefit from communicnon skills trining nd ressurnce tht they re uthorized to discuss the opnons on MOLST forms with pnents. Finncil Support This reserch ws generously supported by the Johns Hopkins Hospitl Deprtment of Medicl Affirs.