Cybergrant Winter 2015
The Stanford Cyber Grant funds were used to collaborate with campus-wide colleagues on the following projects
I.
Role of Networked Technology in Engaging High-Risk Patients in Complex Care Programs
More
PI:
Donna Zulman, MD, MS. Other contributors: Colin W. O'Brien (Stanford medical student), Jessica Y. Breland, PhD; Cindie Slightam, MPH; Andrea Nevedal, PhD
Results to Date
- Qualitative interviews with 20 clinical representatives and leaders from 12 complex care programs revealed the ways in which programs use technology to improve patient engagement across three broad domains:
- Communication and actions to improve health (e.g., videoconferences to communicate new clinical information from patient’s home)
- Building relationships and trust (e.g., telemonitoring as a mechanism to help patients understand that the complex care team “has their back”)
- Insight and goal-setting (e.g., use of electronic assessments to elicit patient priorities and goals and develop appropriate care plans)
- Communication and actions to improve health (e.g., videoconferences to communicate new clinical information from patient’s home)
- Technology use among high-risk patients is limited barriers at the patient level and at the provider/system level:
- Patient barriers: lack of technological skill, lack of interest, competing demands
- Provider/system barriers: problems with technology operation, privacy concerns
These results have been presented at the Society for General Internal Medicine and published in the following article: (link to article).
Next Steps
- Building on these pilot findings, our team is developing proposals for two projects that integrate technology into clinical care for complex patients:
- Proposal for CTSA Collaborative Innovation Award (Mark Cullen, PI; Donna Zulman, Site-PI): Optimizing capture of social and behavioral determinants of health in EHRs
- Proposal for AHRQ: Partnered Care Plans: Integrating Patient Priorities into the Electronic Health Record
The long-term goal of this work is to transform care for patients with multiple chronic conditions by developing a scalable partnered care plan approach that integrates MCC individuals’ priorities and preferences into clinical decision-making. To meet this goal, we propose to evaluate 1) assessment tools for eliciting patients’ priorities before a clinic visit, and 2) implementation strategies for incorporating patient priorities into electronic health record care plans. We hypothesize that when effectively designed and implemented, partnered care plans will increase patient engagement, positively influence patient and providers’ experiences with clinical encounters, and result in clinical outcomes that are aligned with patients’ preferences.
II.
Oncology - Online Test result sharing with patients before confirming physician review/consultation
More
PI:
Cati Brown-Johnson, PhD. With RAs Sacha McBain & Paolo RigoParticipants: 18 Stanford oncologists participated in a short survey; 10 participated in focus groups or interviews lasting 15 minutes to 1.5 hours.
Results to Date
Oncologists discussed themes of information release ), patient-physician communication and trust, results), and the impact of the new test results release policy on workload and workflow6 of 10 participants preferred that patients have easy access to electronic records, but none preferred that this information be actively “pushed” to patients through emails outside of the medical information portal. . Policy suggestions included adjusting release triggers, revising the push method, and providing space within the online test for oncologists to make notes.
This research produced final projects for McBain and Rigo within the Stanford Health4All Certificate Program. The team shared interim and final results in presentation form with the Bedside Medicine research group, led by Abraham Verghese.
Final drafting of the paper “We chase a lot of shadows: Oncologist Perspectives on Electronically-Shared Test Results” is underway.
Next Steps
Publish paper.
III.
a - Surveying physicians to better understand mindsets and impact - including the role of networked technologies in helping or hindering the patient-physician relationship (Doctoral Thesis Primary Data Collection for Kari Leibowitz, Psychology).
More
PIs/Team
Participants: 385 Physicians from Stanford/the Palo Alto area and across the country.
PIs: Kari Leibowitz, BA, and Alia Crum, PhD (Stanford Mind Body Lab, Psychology)
Results to Date
Two upcoming conference presentations: the Society for Personality and Social Psychology conference and the International Positive Psychology Association conference; Presentations at Stanford: the Health for Healers January meeting.
Next Steps
Using insights from this project to design upcoming interventions for physicians, including our mindset intervention that will roll out with Primary Care 2.0 in several Stanford Primary Care clinics.
Request for additional funding is dependent on including information about relating to the Electronic Medical Record (EMR) as part of our mindset intervention.
III.
b - Empathy Evaluation in Medical Trainees and Practitioners
More
PIs/Team
Poonam Hosamani MD, Kari Leibowitz (Psychology Doctoral Student), Alia Crum, PhD (Psychology)
Results to Date
We are currently in the data collection phase of the project. We have preliminary results from our pilot survey distribution.
Next Steps
Continue with survey distribution, particularly with residents.
IV.
The influence of doctor-patient relationships on health outcomes: Measuring reactions to antihistamines by varying warmth in person
and technology use
More
PIs/Team
PIs: Dr. Alia Crum and Lauren Howe, Department of Psychology, Stanford University
Participants:
Study 1: Participants were 163 healthy adults (ages 18-30, 101 women, 59 men, 1 transgender, 2 not reported) recruited through a paid subject pool at Stanford University. We excluded participants who had taken antihistamines within three days of the study appointment as this diminishes allergic responses.
Study 2: So far 43 overweight and obese adults have been recruited from the community to participate in a follow up study.
Results to Date
Study 1: The study is part of a manuscript that has been submitted for publication and is currently on invited revision from the journal Health Psychology.
Study 2: Data collection for the study is underway and we expect to complete it by the end of March.
Next Steps
Study 1: In this study, we found that physician warmth played a powerful role in shaping positive and negative expectations about treatment. Focusing attention on a computer during a patient health background interview and failing to engage/maintain eye contact with participants undermined the sense of warmth participants felt interacting with physicians.
Study 2: We are extending this project by investigating the power of physicians exposing humanizing flaws for building rapport with patients. Overweight/obese individuals experience lower quality healthcare, facing negative evaluation from physicians and avoiding necessary healthcare as a result. We expected that exposing a physician’s own health shortcomings would benefit this stigmatized group by reducing concerns that physicians would judge overweight/obese individuals harshly. We are testing whether doctors who have unhealthy habits exposed or who openly disclose their own health habits are more desirable to people who frequently fear judgment from physicians in healthcare contexts (e.g., people who are overweight). We think that these flaws will decrease expectations of judgment and help to build rapport with patients, positively impacting health outcomes.
V.
How does technology in patient-provider communication affect care for diabetes?
More
PIs/Team
Investigators: Sukyung Chung, PhD, Palo Alto Medical Foundation Research Institute
Latha Palaniappan, MD, MS, Stanford University School of Medicine
Results to Date
Among patients with diabetes,
- The use of secure messaging increased dramatically in recent years. The increase coincides with financial incentives such as the removal of user fee (in 2011), and explicit payment of providers for responding to patient messages ($5 per message thread in 2012).
- Patients who make frequent visits to PCP are also more likely to use secure messaging.
- Messaging replaced some phone calls, but not office visits.
- Additional messages, given visit frequency, helps patients manage diabetes.
- Odds Ratio with 95% CI from Multivariate Analysis
- The effect size is larger than that of face-to-face visits.
Next Steps
Manuscript is under review by authors with plan to publish it in a peer reviewed journal within the next few months.
VI.
Wellness Interventions for Housestaff - Do Animals Beat Machines?
More
PIs/Team
Participants/Researchers - Mala Mandyam MD; Peter Bayley MD; Lisa Shieh (PI)
Collaborators - Junaid Zaman MD, Alexander Tzabazis MD
Results to Date
We've completed a pilot phase where we have been focusing on technique of data gathering. 20 residents/medical students participated in the pilot portion. In this pilot, the average subjective stress level went from 48 to 30 (on a scale of 0 to 100, with 100 being the most stressed) after subjects interacted with a dog. The stress level went from 44.5 to 45 after subjects used their smartphone. Because of the variability in data collection (tested out during the pilot), we do not think this data is accurate enough to use in final analyses. Data from the FirstBeat devices is still being analyzed.
Next Steps
- We plan to recruit 50 residents/students and have them participate in each arm
- We will conduct the remainder of the study from Jan to April, and do analyses in June. We'll do interim analyses for quality assurance
- We will recruit additional residents to serve as researchers to collect data on days when I am not available
VII.
Three Good Things to Improve Resilience for our patients
More
PIs/Team
Principal Investigator: Lisa Shieh, MD, PhD
Participants: Leon Castaneda, Steven Crane, Valeria Sindra and Cara Lai.
Results to Date
Currently, we have solidified our survey instruments and are continually refining our processes as we have now started enrolling and following up on participants. We have developed a database in Redcap and are currently testing its functionality before moving it to the production mode. We have recruited a couple key individuals for the project that will provide continuity and are currently in the process of recruiting more individuals to help with patient recruitment and follow up. We have compiled a demographic and clinical information that we desire for STRIDE to pull from Epic. We researched the availability and cost of a statistician via Spectrum, which unfortunately, we found very cost-prohibitive and are currently seeking alternatives.
Next Steps
- continue recruiting individuals to help with data collection
- continue recruiting and following up with patients to collect data
- we will continue to assess exclusion and dropout rates to ensure that we have enough participant data to achieve statistical power
- once the study panel is completed, we will pull demographic and clinical data from Epic via STRIDE
- we are currently writing our introduction/background and methods section of the paper
- we are seeking a statistician to be involved in the project
VIII.
Does wearable technology improve patient mobility and wellbeing during acute hospitalizations?
More
PIs/Team
Participants: Samantha Wang (Resident); David Svec, MD (Co-Investigator); Lisa Shieh, MD (PI)
Results to Date
The study anticipates preliminary results to be available by late-August 2017. Hypotheses and study objectives were considered based on previous studies that have used FitBit® in surgical and geriatric population. Please see attached Proposal for further details on background literature.
Next Steps
The project’s goal is to recruit 100 total participants in a single center randomized study of using FitBit® in acutely hospitalized medicine patients. Next steps in study infrastructure development and recruitment include:
- finalization of study survey to be distributed to all study participants at the end of the study
- purchase of study devices (15), each to be used 3-5 times
- education of staff members (charge nurses, floor nurses, NAs) of the study during the period of enrollment
- enrollment of patients
- establishment of a database to collect FitBit® variables across three domains: ambulation data / sleep data /heart rate data. This will be compared to outcomes between the two groups using: nursing/PT documents, tele monitoring data, nursing localization data, survey results, and hospitalization duration (length of stay) and discharge site
- collection of data
- distribution of survey to participants
IX.
Increasing Physician-Nurse Communication at the Bedside, using Infrared Locator Technology
More
PIs/Team
(All MDs) Rebecca Tisdale, Adam Sang, James Barnes, Jeff Chi, Lisa Shieh (PI)
Results to Date
- Currently obtaining baseline data of resident MD and nurse overlap at the patient bedside for November –January. We have had 100% enrollment of residents for both months.
- Outcomes from hospitalist pilot submitted to Society of Hospital Medicine Annual Meeting’s "Research, Innovations, and Clinical Vignettes (RIV)" competition.
Next Steps
- We plan to do our root cause analysis of why MD and RN are not at bedside together in January. Intervention is planned for February-March with the nurses on B3. C3 will remain our control unit.
- Post intervention data collection is planned for April-June.
X.
Mining Sepsis Data to Evaluate the effects of universal electronic sepsis screening and potential for alternative screening strategies
More
PIs/Team
Principal Investigator: Dr. Lisa Shieh
Participants: Dr. Kai Swenson, Dr. Norman Downing, Dr. Gomathi Krishnan
Results to Date
- Creation of two initial resident surveys and trial among small group. Survey 1 consists of resident knowledge, attitudes, and actions surrounding current sepsis screening at Stanford. Survey 2 consists of knowledge basis surrounding Sepsis-3 guidelines and use of qSOFA.
- Completion of initial STRIDE data collection (all hospitalized patients with floor-level care from September 2015 - July 2016, accumulating data on collection time and components of first positive sepsis screen (SIRS, qSOFA, or both), age, additional demographic information, and in-patient mortality and time of death.
Next Steps
- Distribute 1st resident survey, analysis of results
- Distribute 2nd resident survey, analysis of results
- Additional STRIDE data collection: Comorbidities, hospital diagnoses (sepsis, infection type), date of admission, admitting provider/service, critical care utilization (presence and timing) including vasopressor use and mechanical intubation
- Analysis of STRIDE data comparing qSOFA to SIRS screening algorithms, with particular emphasis on subgroups where SIRS-based approach may lack accuracy (oncology patients, transplant patients, surgery patients, and heart failure patients).
XI.
Mining Networked Data to explore how physician work factors influence patient outcomes and the physician-patient relationship.
More
PIs/Team
PI: Grant James Barnes
Participants:
- James Barnes – Former resident in internal medicine at Stanford and Health Services Research and Development Fellow at PCOR and Palo Alto VA.
- Nathaniel Myall – Former resident in internal medicine at Stanford and current Stanford Hematology and Oncology Clinical Fellow.
- Raymond Deng – Current Stanford Medical Student
- Lisa Shieh – Clinical Professor and Medical Director for Quality, Department of Medicine at Stanford
- David Chan – Assistant Professor of Medicine, Core Faculty member at the Center for Health Policy and PCOR
- Gomathi Krishnan – Clinical and Translational Research Informatics Specialist for the Stanford Center for Clinical Informatics
Results to Date
Preliminary Regression Results examining the effect of the day of admission show an increase in length of stay of about 0.2 days for patients admitted on day 4 of the call day cycle vs. day 1. These results are preliminary and many specifics of the model. Patient satisfaction data has been requested and we are in the process of acquiring it. Difficulties Encountered So far we have not encountered major difficulties. We did encounter some interesting findings in the data which has required further expansion of our dataset for additional analysis. This has required modifications to the IRB and further interaction with Gomathi Krishnan at STRIDE in order to acquire this data.
Next Steps
We plan to finalize our analysis plans, incorporate additional outcomes as. We have been approved for modifications to the IRB to examine baseline physiologic and lab data as well as incorporate patient satisfaction scores into the analysis. Lab and physiologic data will help us control for patient illness -related factors that do not rely on post-discharge administrative data-derived severity scores which may themselves be independently affected by workload factors. We are currently in the process of sorting through this data and coding it correctly for a series of regression models.