Targeting and Tailoring Health Care Messages Suzanne Bakken

Targeting and Tailoring Health Care Messages Suzanne Bakken, RN, DNSc School of Nursing and Department of Medical Informatics Columbia University New York, NY 10032

Purpose • Review principles of targeting and tailoring health care messages • Illustrate principles and steps with examples – CAP-IT – MI Heart – Heart. Care • Describe medical informatics vs. contentspecialist roles and challenges

Definitions • Targeting - intended to reach a specific subgroup of the population typically based upon demographic characteristics – Culturally-sensitive video on cervical screening resulted in in Pap smear in Latinas (Yancey & Walden, 1994) • Personalized - use of name to draw attention to a generic message (Kreuter et al, 2000)

Definitions • Tailoring - intended to reach a specific individual based on unique characteristics related to the outcome of interest and obtained through an assessment or profile (Kreuter et al, 2000) – Within ethnicity-diversity when exposed to a culturally-similar (CSV) and culturally-dissimilar video for HIV/AIDS intervention – Only CSV effective in African American youths who scored themselves as “know a lot about AIDS” (Stevenson et al. , 1995)

individualized Tailored communication Personalized communication Targeted communication generic Content of Communication Interpersonal communication Generic communication Not assessment based Level of Assessment (Kreuter et al, 2000) Based on Assessment of individuals

Where is the State of the Art in Diabetes Care?

Rationale for Tailoring • Based upon elaboration likelihood model (Petty et al, 1994) • Superfluous information is eliminated • Information is perceived as more personally relevant • Persons may pay more attention to information they perceive as personally relevant • Personally relevant + attended to is more likely to lead to thoughtful consideration of actors related to behavior change • Personally relevant + attended to + thoughtful consideration will be more useful to help enact desired behavior change (Kreuter et al, 2000)

RCTs of Tailored vs. Non-tailored Health Messages • • Dietary Change Smoking Cessation Physical activity Mammography Weight control Cholesterol screening Nutritional label reading

Tailored vs. Non-tailored Health Messages Are More Likely To: • • • Catch attention Be saved Be discussed with others Be perceived by readers as interesting Be perceived by readers as personally relevant • Be perceived by readers as personally written for them

Steps in the Tailoring Process • • • Analyze the health problem Develop a program framework Develop a tailored assessment Design feedback Write tailored messages Create tailoring algorithms Automate tailoring process Implement the program Evaluate the program

Examples • CAP-IT • MI Heart • Heart. Care

Example • Client Adherence Profiling and Intervention Tailoring RCT (Holzemer & Bakken, NR 04846) • To compare CAP-IT vs. standard care on: – HAART adherence rates – CD 4 count, viral load, viral resistance – Quality of life – Health services utilization

Steps in the Tailoring Process 3 Analyze the health problem 3 Develop a program framework • Develop a tailored assessment • Design feedback • Write tailored messages • Create tailoring algorithms • Automate tailoring process • Implement the program • Evaluate the program

Ickovics & Meisler Framework (1997)

Steps in the Tailoring Process • Analyze the health problem • Develop a program framework 3 Develop a tailored assessment • Design feedback • Write tailored messages • Create tailoring algorithms • Automate tailoring process • Implement the program • Evaluate the program

CAP-IT Intervention RCT • Client Adherence Profiling (CAP) – Client characteristics – Knowledge about medication regimen – Knowledge about side effects and symptom management of side effects – Self-care behaviors – Treatment complexity – Engagement with provider – Support systems

Steps in the Tailoring Process • Analyze the health problem • Develop a program framework • Develop a tailored assessment 3 Design feedback 3 Write tailored messages 3 Create tailoring algorithms • Automate tailoring process • Implement the program • Evaluate the program

CAP-IT Intervention RCT • Intervention Tailoring (IT) – At least one intervention from each category based upon nursing diagnosis In CAP-IT, this is done by nursing judgment, not by algorithm

Potential Gains from an Informatics Approach • Automatic scoring of profile from tailored assessment • Standardized selection of interventions based upon score using decision support rules • Measurement of and ability to measure “dose” of intervention • Data re-use (e. g. , from CIS) • Resource re-use (content re-used from message library) • 24 x 7 availability

Potential Downsides from an Informatics Approach • • Access, digital divide issues Computer literacy Literacy and health literacy Where is provider in the loop?

Information-Technology-Based Patient Education for Decreasing Prehospital Delay of Patients Presenting with Acute Myocardial Infarct: The MI-HEART Project James J. Cimino, MD, Principal Investigator Rita Kukafka, Dr. PH Yves A. Lussier, MD Vimla L. Patel, Ph. D Department of Medical Informatics, Columbia University Supported by National Library of Medicine and the National Heart, Lung and Blood Institute N 01 -LM 3534

Research Design and Methods: Overview • Randomized, controlled trial • 300 patients will be recruited and randomized into one of three groups • Follow-up at 1, 3, and 6 -months

Research Design and Methods: Outcomes • Likelihood of action (seeking help in response to symptoms) as reported by patients • Changes in attitudes and beliefs associated with patient delay

MI-HEART Tailoring Variables Medical Record Demographics Health history • Diabetes • Angina Family History Computer-Related • font size • display Online Questionnaire Variables associated with increased delay • Somatic and emotional awareness • Expectation of symptoms • Perceived threat • Self-efficacy • Response efficacy

Heart. Care Project • Post-operative CABG care • Web TV interface • Web page with access to relevant resources – Targeted information: based upon phase of the recovery process – Tailored information: based upon nurse interview during hospitalization (Brennan et al)

Tailoring Recovery Resources to Patients • Establishing the tailoring model – Patient Profiles – Access (TM) database • Delivering WWW resources ‘on-thefly’, across the recovery period – Active server pages sorting nurseidentified or developed WWW pages

Discussion • What are potential areas of focus for computer-based educational and behavioral interventions? • Which of the current barriers to education could potentially be addressed through computer-based approaches?

Discussion • What are factors that have been used or could be used to target educational and behavioral change interventions? • What are factors that have been used or could be used to tailor educational and behavioral change interventions? • Which of these factors are routinely collected and stored in electronic form? • How can such approaches extend the provider and yet engage the client?

Areas for Medical Informatics Contribution • Security and confidentiality of the data • Knowledge representation for re-use – Standardized assessments – Clinical terminology – Logic for decision support – Sharable message libraries • User interface