Colorado Care Tablet: The design of an interoperable Personal Health Application to help older adults with multimorbidity manage their medications

Our reference: YJBIN 1661 P-authorquery-v7

AUTHOR QUERY FORM

Journal: YJBIN

Article Number: 1661

Please e-mail or fax your responses and any corrections to:

E-mail: [email protected]

Fax: +31 2048 52799

Dear Author,

Any queries or remarks that have arisen during the processing of your manuscript are listed below and highlighted by flags in the proof. Pleasecheck your proof carefully and mark all corrections at the appropriate place in the proof (e.g., by using on-screen annotation in the PDF file) orcompile them in a separate list.

For correction or revision of any artwork, please consult http://www.elsevier.com/artworkinstructions.

Articles in Special Issues: Please ensure that the words ‘this issue’ are added (in the list and text) to any references to other articles in thisSpecial Issue.

Uncited references: References that occur in the reference list but not in the text – please position each reference in the text or delete itfrom the list.

Missing references: References listed below were noted in the text but are missing from the reference list – please make the list completeor remove the references from the text.

Location inarticle

Query / remarkPlease insert your reply or correction at the corresponding line in the proof

Q1 Please provide captions for Figs. 2 and 3 and ‘‘Table 1.

Q2 Please check the page range is okay as typeset for Refs. [3, 9, 13].

Q3 Please check and approve the Fig. 1 caption is okay as typeset.

Electronic file usageSometimes we are unable to process the electronic file of your article and/or artwork. If this is the case, we have proceeded by:

h Scanning (parts of) your article h Rekeying (parts of) your article h Scanning the artwork

Thank you for your assistance.

UNCORRECTE

DPROOF

1

2 Colorado Care Tablet: The design of an interoperable Personal Health Application to3 help older adults with multimorbidity manage their medications

4 Katie A. Siek a,*, Stephen E. Ross b, Danish U. Khan a, Leah M. Haverhals b,1, Steven R. Cali b,2, Jane Meyers a,3

5 aWellness Innovation and Interaction Laboratory, University of Colorado at Boulder, Department of Computer Science, 430 UCB, Boulder, CO 80309-0430, USA6 bUniversity of Colorado School of Medicine, Mail Stop B180, 12631 E. 17th Avenue, Room 8504, P.O. Box 6511, Aurora, CO 80045, USA

7

9a r t i c l e i n f o

10 Article history:11 Available online xxxx

12 Keywords:13 Personal Health Records14 Older adults15 Medication management16 User-centered approach17 Participatory design18 User interface design19

2 0a b s t r a c t

21Medication errors are common and cause serious health issues during care transitions, particularly for22older adults with multiple chronic conditions. In this paper, we discuss the design and evaluation of23the Colorado Care Tablet, a Personal Health Application (PHA) that helps older adults and their lay care-24givers manage their medication regimes during care transitions. We created a PHA that older adults with25limited computing experience could easily use by designing an application based on their real world arti-26facts and workflows.27! 2010 Published by Elsevier Inc.

28

29

30 1. Project goals and design requirements

31 Care for older adults with multiple chronic conditions is often32 fragmented and prone to care transitions. Older adults may receive33 care from a set of independent doctors and specialists as an outpa-34 tient and then transfer to a new set of inpatient doctors for acute35 care, followed by convalescent care at a skilled nursing facility until36 outpatient care can be resumed. Many times the older adult and37 caregiver are responsible for sharing informationbetweencare facil-38 ities [1], but do not feel they can effectively communicate or execute39 the care plan [2]. Poorly coordinated care often leads to medication40 errorswith adverse health consequences [3]. Since the care facilities41 do not have the time or resources to coordinate care [1], we took a42 grass-roots approach to empower older adults to manage their43 health during care transitions. Based on the ‘‘Care Transitions Inter-44 vention” [2,4], a proven program that improves transitional care for

45older adults using a paper Personal Health Record (PHR), we devel-46oped an electronic Personal Health Application (PHA),4 the Colorado47Care Tablet (CCT), that assists older adults in managing and sharing48medication regimes across care transitions.49The CCT design was informed by previous research about med-50ication management [2,5–9] and our multidisciplinary team expe-51riences to create a functional, electronic prototype that addresses52the needs of older adults and lay caregivers.5 In this paper, we pres-53ent an overview of our user centered iterative design cycle with an54emphasis on how artifacts and workflows from our user needs55assessment informed the design of the CCT. More specifically, our56primary design requirements were to:

571. Help participants create and maintain a personal medication58list when care is received from multiple, independent practices.592. Provide easily accessible, authoritative medication information.603. Help participants effectively communicate with health profes-61sionals about medication regimes and conditions.

62In addition, we explored how each design requirement could be63implemented.

642. Prototype description

65We optimized the CCT interface for a Lenovo ThinkPad X60 tab-66let PC because it provided portability along with a finger-touch67sensitive screen. We used a Socket Mobile Bluetooth Cordless Hand

1532-0464/$ - see front matter ! 2010 Published by Elsevier Inc.doi:10.1016/j.jbi.2010.05.007

* Corresponding author. Fax: +1 (303) 492 2944.E-mail addresses: [email protected] (K.A. Siek), [email protected] (S.E.

Ross), [email protected] (D.U. Khan), [email protected] (L.M.Haverhals), [email protected] (S.R. Cali), [email protected] (J.Meyers).

1 Present address: Veterans Affairs Medical Center, Denver, Colorado, REAP(Research Enhancement Award Program) to Improve Care Coordination, 1055Clermont St., 151 Research Denver, CO 80220, USA.

2 Present address: A-234 Life Sciences Building, East Lansing, MI 48824, USA.3 Present address: Center for Life Long Learning and Design, Department of

Computer Science, Campus Box 430, University of Colorado at Boulder, Boulder, CO80309-0430, USA.

4 Personal Health Application (PHA); Personal Health Record (PHR), Colorado CareTablet (CCT).

5 For simplicity, we refer to older adults and lay caregivers as participants in thismanuscript.

Journal of Biomedical Informatics xxx (2010) xxx–xxx

Contents lists available at ScienceDirect

Journal of Biomedical Informatics

journal homepage: www.elsevier .com/locate /y jb in

YJBIN 1661 No. of Pages 6, Model 5G

18 May 2010ARTICLE IN PRESS

Please cite this article in press as: Siek KA et al. Colorado Care Tablet: The design of an interoperable Personal Health Application to help older adults withmultimorbidity manage their medications. J Biomed Inform (2010), doi:10.1016/j.jbi.2010.05.007

UNCORRECTE

DPROOF

68 Scanner Series 7 to scan medication barcodes. The functional pro-69 totype was a server-side application developed in PHP, JavaScript,70 and HTML – thus CCT can be accessed from any machine, not just a71 tablet PC, with Internet access. The CCT architecture is shown in72 Fig. 2.73 TheCCTprototype provides participantswith an easy, linear nav-74 igation structure, to complete four main functions: medication list75 creation andmanagement,medication information retrieval, doctor76 visit preparation, and information onwhen to seek assistance. Fig. 377 provides a snapshot of the main functions. In this section, we will78 briefly discuss the navigation and functionality design.

79 2.1. Navigation

80 Older adults can most effectively navigate websites that have81 simplified selections per screen and flattened navigation structures82 [10]. Thus, we created a simple main application screen that had83 four icons representing the tasks most needed by participants dur-84 ing care transitions (Fig. 3) based on prior work [2,4] and our needs85 assessment [11]. Once an icon was selected, a linear, wizard-like86 navigation structure provided participants with clearly defined87 steps on how to complete each function. For example, CCT partic-88 ipants were walked through a step-by-step process on how to add89 medications (Fig. 1). Consistent with design guidelines for older90 adults [10], we ensured there were multiple ways for participants91 to complete the same task. For example, participants could sche-92 dule medications by pressing the medication schedule icon on93 the dock area or they could select the medication from their list94 and then select schedule.

952.2. Medication list creation and management

96The first step in medication management is creating and main-97taining an up-to-date medication list. While medication reconcili-

Fig. 1. Linear Navigation Structure Example: Adding a medication: The participant would select medications from the pharmacy fulfillment page and then navigate to themedication list confirmation page. The participant can add more medications or continue using CCT.

Fig. 2.

Q3

Q1

2 K.A. Siek et al. / Journal of Biomedical Informatics xxx (2010) xxx–xxx

YJBIN 1661 No. of Pages 6, Model 5G

18 May 2010ARTICLE IN PRESS

Please cite this article in press as: Siek KA et al. Colorado Care Tablet: The design of an interoperable Personal Health Application to help older adults withmultimorbidity manage their medications. J Biomed Inform (2010), doi:10.1016/j.jbi.2010.05.007

UNCORRECTE

DPROOF

98 ation applications have been developed for PHRs tethered to a sin-99 gle practice, this task can be particularly challenging when care is

100 received from multiple independent practices. The CCT walked101 participants step-by-step through the process to build a list. Partic-102 ipants could add medications by: selecting medications recently103 picked up from pharmacies; scanning the medication barcode; or104 entering the prescription medication name where a codified repre-105 sentation was captured. Medications were identified by image,106 generic name, and trade name. Medication lists were prominently107 featured on the main application screen.

108 2.3. Medication information retrieval

109 Once participants completed the medication list creation, they110 were directed to the CCT wizard main page where they could re-111 trieve information about their medications. Once they selected the112 medication from their list and selected ‘‘Drug Facts”, they were pre-113 sented with the medication picture and icons for the six commonly114 asked questions about medications. Participants could easily access115 information about ‘‘Why am I taking this medication?” to ‘‘What if I116 miss a dose?” (Fig. 3). Informationwas derived andmanually parsed117 from the National Library of Medicine’s MedlinePlus [12].

118 2.4. Doctor visit preparation

119 We designed a visit preparation wizard to empower patients120 with the information necessary for effective interactions with

121health professionals during short clinic visits. The wizard walked122participants through a series of steps to formulate questions and123send information to healthcare professionals prior to visits. The124wizard first suggested some common questions that could be per-125sonalized by each patient (e.g., ‘‘Why am I taking Flomax? Is there126a cheaper medication available?”). Then, the participant confirmed127their current medication list. Finally, the participant was shown a128memo that could be printed or emailed to their healthcare provider129that listed all of their questions, medication list, and any medica-130tion list discrepancies that the CCT automatically identified when131comparing the participant’s confirmed medication list and the132doctor’s.

1332.5. When to seek assistance

134A critical element of self-care in care transitions is being vigilant135for ‘‘red flag” signs that might lead to readmission without inter-136vention. We had to balance red flag input with limited free-text en-137try because research has shown that older adults are better using138touch screen interfaces than keyboards or mice [13,14]. In addition,139we wanted to design for the future where CCT could be interacted140with on any surface with an interface. We designed the CCT to pro-141vide users with stub statements (e.g., My fever goes over ____ de-142grees) to prompt patients to consider red flags in response to these143constraints. The stub statements were filled in with values recom-144mended by their doctors. If a participant checked a red flag, indi-

Fig. 3.

K.A. Siek et al. / Journal of Biomedical Informatics xxx (2010) xxx–xxx 3

YJBIN 1661 No. of Pages 6, Model 5G

18 May 2010ARTICLE IN PRESS

Please cite this article in press as: Siek KA et al. Colorado Care Tablet: The design of an interoperable Personal Health Application to help older adults withmultimorbidity manage their medications. J Biomed Inform (2010), doi:10.1016/j.jbi.2010.05.007

UNCORRECTE

DPROOF

145 cating the red flag had occurred, the CCT notified the participant146 that they should contact their doctor immediately.

147 3. Evaluation results

148 The CCT was developed over a two-year period that included a149 6-month needs assessment and an 18-month design, development,150 and evaluation period. The needs assessment findings informed the151 CCT interface design. We iteratively developed the CCT interface152 from low-fidelity, paper-based prototypes to a functional proto-153 type application. Although outside the scope of this paper, more154 information about our iterative design process is available in Table155 1 and [15].

156 3.1. User needs assessment

157 We conducted the user needs assessment with four focus158 groups and fourteen in situ interviews to understand the issues159 faced by participants when managing medications during care160 transitions. We documented how participants managed their med-161 ication regimes and health information through interview notes162 and digital photographs. These photographs – representing the real-163 ity of participants’ lives – informed the design of the CCT prototypes.164 Confirming previous research [7,9], we found that participants ar-165 ranged medications around their house based on their daily rou-166 tines and derived medication information from a variety of167 sources. Participants employed a variety of tools for keeping track168 of medications – including pillboxes, labeled envelopes, and Ziploc169 bags.

170 3.2. CCT iterative user studies

171 Iterative, participatory design of the CCT included six iterative172 user studies. We used a combination of Rapid Iterative Testing173 and Evaluation [16] and Instant Data Analysis [17] to quickly iden-174 tify and address usability issues. Each user study had 4–8 partici-175 pants for a total of 31 participants: 22 older adults (mean age:176 76.4 years old) and 9 caregivers (mean age: 52.7 years old). In177 the first two studies, we used paper-based, low-fidelity prototypes

178for rapid assessment and redevelopment. We iteratively designed179high-fidelity, tablet PC-based prototypes for the remaining four180studies.181The evaluations confirmed that participants wanted to see med-182ication images with their lists. Another consistent theme was the183need to simplify the user interface for this target population, even184if this meant limiting functionality. For instance, users found it con-185fusing to view their own personal medication list and the medica-186tion lists from various doctors. Comparing their list and a doctor’s187list side-by-side was more difficult than helpful – participants were188concerned that they were inadvertently corrupting their doctor’s189medication list. Ultimately, users preferred to view only fulfilled190medications (rather than also including prescribed medications)191whencreating their lists. Furthermore, they felt theCCT shouldauto-192matically identifymedication list discrepancies betweenparticipant193and doctor lists.194Their preferences for an image-rich, linear navigation structure195with fewer utilities and more prompts was borne out in limited196comparative testing of the CCT and the Google Health interface197in early 2009. Our participants successfully completed common198medication management tasks with the CCT, but could not effi-199ciently navigate the Google Health hierarchical and hyperlinked200navigation structure to create a medication list or find information201about medications.202Several challenges remained in user testing. While the CCT was203well accepted by those in their 60s and 70s, the oldest adults in our204study (80+ years old) declined to use any electronic medication205management system, regardless of how simplified its interface206might be. Also, while participants liked the concept of entering207medications by scanning medication barcodes and thought the208scanner was ergonomically suitable, in practice we found that209scanning was too inconsistent to be incorporated into a production210system.

2114. Discussion and implications

212We developed a PHA prototype for medication management213that was well accepted in user testing by participants. Acceptance214was due in part to our analysis and accommodation of the work-

Table 1

Study Studydate

N (# OlderAdults/#Caregivers)

Goals Borrowed from participant artifacts orworkflows

Outcomes

1 10/2007

4 (3/1) Design and evaluate paper-based prototypes informed byparticipants’ everyday reality

Health information management areas ofparticipants’ homes informed low-fidelitykitchen counter interface prototype

Participants wanted interfaces that borrowedexamples from their everyday life along withappropriate textual information

2 12/2007

4 (4/0) Improve the prototype designand defined medicationmanagement interfaces

Date and time labeled envelopes informed low-fidelity interface reminder clock prototype

Participants thought medication list was for doctor’suse; Participants wanted easier input mechanisms

3 2/2008

4 (4/0) Refine the medication listmanagement interface andevaluate alternate forms ofinput mechanisms

Participants’ handwritten lists on linednotebook paper informed medication listbackground. Medication bottle informednavigation link to home screen

Participants understood that they owned themedication list with the help of everyday examplesfrom their lives. Participants found alternativemechanisms of inputting medications wasconvenient

4 3/2008

4 (3/1) Verify medication listmanagement interface anddefine visit preparationinterface

A participant’s pocket calendar informed visitpreparation interface (Fig. 3)

Participants wanted to use visit preparationinterface for their doctor appointments

5 7/2008

8 (5/3) Verify visit preparationinterface and enhancemedication informationinterface

Post-it note (Fig. 1) informed the design ofenhanced medication list background toeliminate need to line up text with a linedbackground

Participants could easily obtain authoritativemedication information. Participants wanted tomonitor symptoms of their disease condition

6 1/2009

7 (3/4) Design and integrate diseasesymptom monitoringinterface. Evaluate the entiresystem

Red-flags informed the navigation link todisease symptom monitoring interface (Fig. 3)

Participants effectively used CCT to performcommon medication management tasks

4 K.A. Siek et al. / Journal of Biomedical Informatics xxx (2010) xxx–xxx

YJBIN 1661 No. of Pages 6, Model 5G

18 May 2010ARTICLE IN PRESS

Please cite this article in press as: Siek KA et al. Colorado Care Tablet: The design of an interoperable Personal Health Application to help older adults withmultimorbidity manage their medications. J Biomed Inform (2010), doi:10.1016/j.jbi.2010.05.007

UNCORRECTE

DPROOF

215 flows and artifacts used by older adults in their everyday lives. A216 simplified linear navigation structure also proved helpful. Ulti-217 mately, participants were able to manage medication lists in sim-218 ulated care transitions with data that will soon become available219 with greater interoperability (e.g., fulfillment data and Electronic220 Medical Record medication lists).221 These results confirm the notion that a ‘‘one size fits all” elec-222 tronic PHR is not likely to be adopted by older adults with multi-223 morbidity – a group that stands to benefit greatly from medication224 management assistance. Thus, this confirms the proposition that a225 common platform, consisting of a data layer and common func-226 tions, serving multiple tailored applications can produce more227 compelling PHAs.228 The CCT prototype could be made a reality in the near future.229 Currently only a minority of ambulatory doctors keep electronic230 medication lists, and virtually no hospital provides patients with231 an electronic discharge summary. More prescribing data will be232 available as electronic prescribing is adopted. Fulfillment data are233 also available from Surescripts, although this requires patients to234 explicitly give consent. Informatics communities must explore235 more automated mechanisms for user authorization and authenti-236 cation to make this data more readily available.

237 5. Overview of implications

238 We designed CCT to demonstrate how a linear interface naviga-239 tion structure could provide participants with the ability to com-240 plete medication management tasks that are critical in care241 transitions. It also shows how artifacts and workflows from the tar-242 get population’s everyday life can aid in the design of a usable,243 intuitive interface.

244 Conflict of interest statement

245 The authors of this document have no conflicts of interest –246 including any financial, personal, or other relationships with other247 people or organizations within three years of beginning the sub-248 mitted work that could inappropriately influence, or be perceived249 to influence, our work.

250 Acknowledgments

251 Contributors: Eric Coleman, MD, advised the authors on key is-252 sues in transitional care. We would like to thank our participants253 and advisory board for their valuable feedback.

254Funders: This research is funded by the Robert Wood Johnson255Foundation Project HealthDesign Grant RWJ59880 (PI Stephen E.256Ross).

257References

258[1] Coleman EA. Falling through the cracks: challenges and opportunities for259improving transitional care for persons with continuous complex care needs. J260Am Geriatr Soc 2003;51:549–55.261[2] Coleman EA, Smith JD, Frank JC, Min SJ, Parry C, Kramer AM. Preparing patients262and caregivers to participate in care delivered across settings: the Care263Transitions Intervention. J Am Geriatr Soc 2004;52:1817–25.264[3] Forster A, Murff H, Peterson J, Gandhi T, Bates D. The incidence and severity of265adverse events affecting patients after discharge from the hospital. Ann Intern266Med 2003;138(3):161–7.267[4] Coleman EA. Transition survival skills (2009) [cited 1 March 2010], Care268Transitions Program. Available from: http://www.caretransitions.org/269transitionskills.asp2009.270[5] Hayes TL, Hunt JM, Adami A., Kaye JA. An electronic pillbox for continuous271monitoring of medication adherence. In: 28th annual international conference272of the IEEE in engineering in medicine and biology society EMBS ’06. 2006. p.2736400–3.274[6] Kaushik P, Intille SS, Larson K. Observations from a case study on user adaptive275reminders for medication adherence. In: Proceedings of the second276international conference on pervasive computing technologies for healthcare.2772008. p. 250–3.278[7] Palen L, Aalokke S. Of pill boxes and piano benches: ‘‘home-made” methods for279managing medication. In: Proceedings of the 2006 20th anniversary280conference on computer supported cooperative work. 2006. p. 79–88.281[8] Wan D. Magic Medicine Cabinet: A Situated Portal for Consumer Healthcare,282Lecture Notes in Computer Science: Handheld and Ubiquitous Computing:283HUC ’99. 1999. p. 352–355.284[9] Moen A, Brennan PF. Health@Home: the work of health information285management in the household (HIMH): implications for consumer health286informatics (CHI) innovations. J Am Med Inform Assoc 2005;12(6):648–56.287[10] Demiris G, Finkelstein SM, Speedie SM. Considerations for the design of a Web-288based clinical monitoring and educational system for elderly patients. J Am289Med Inform Assoc 2001;8(5):468–72.290[11] Ross S, Darr C, Haverhals L, Siek K. Project health design: assisting older adults291with transitions of care – design phase proposal (2007) [cite 7 april 2010],2922007 project health design: assisting older adults with transitions of care –293design phase proposal. Available from: http://www.projecthealthdesign.org/294media/file/x%20-%20ColoPHD_WK3_DesignProposal.doc2007.295[12] MedlinePlus (2010). [cited 1 March 2010], US National Library of Medicine.296Available from: http://www.nlm.nih.gov/medlineplus/2010.297[13] Rau P-L, Hsu J-W. Interaction devices and web design for novice older users.298Educ Gerontol 2005;31(1):19–40.299[14] Charness N, Bosman EA, Elliott RG. Senior-friendly input devices: is the pen300mightier than the mouse? In: Proceedings of 103rd annual convention of the301American psychological association meeting. 1995.302[15] Khan DU, Siek KA, Meyers J, Haverhals L, Cali S, Ross SE. Don’t they do that?303Interdisciplinary iterative design of a personal health application for older304adults, TR# CU-CS 1064-10. 2010.305[16] Wixon D. Evaluating usability methods: why the current literature fails the306practitioner, Interactions 2003;10(4): 28–34 [ACM].307[17] Kjeldskov J, Skov MB, Stage J. Instant data analysis: conducting usability308evaluations in a day. In: Proceedings of the third Nordic conference on human-309computer interaction. 2004. p. 233–240.

310

Q2

K.A. Siek et al. / Journal of Biomedical Informatics xxx (2010) xxx–xxx 5

YJBIN 1661 No. of Pages 6, Model 5G

18 May 2010ARTICLE IN PRESS

Please cite this article in press as: Siek KA et al. Colorado Care Tablet: The design of an interoperable Personal Health Application to help older adults withmultimorbidity manage their medications. J Biomed Inform (2010), doi:10.1016/j.jbi.2010.05.007