Machine-Generated Mortality Estimates and Nudges to Promote Advance Care Planning Discussion Among Cancer Patients

A Stepped-Wedge Cluster Randomized Trial Using Machine-Generated Mortality Estimates and Behavioral Nudges to Promote Advance Care Planning Discussion Among Cancer Patients

This study will use a stepped-wedge cluster randomized trial to evaluate the effect of a health system initiative using machine learning algorithms and behavioral nudges to prompt oncologists to have serious illness conversations with patients at high-risk of short-term mortality.

Study Overview

• Status: Completed
• Conditions: Oncology
• Intervention / Treatment: Behavioral: Nudge

Detailed Description

Patients with cancer often undergo costly therapy and acute care utilization that is discordant with their wishes, particularly at the end of life. Early serious illness conversations (SIC) improve goal-concordant care, and accurate prognostication is critical to inform the timing and content of these discussions. This study will use a stepped-wedge, cluster randomized trial to evaluate the effect of a health system initiative using machine learning algorithms and behavioral nudges to prompt oncologists to have serious illness conversations with patients at high-risk of short-term mortality. Oncology practices will be randomly assigned in sequential four-week blocks to receive the intervention.

• Study Type: Interventional
• Enrollment (Actual): 78
• Phase: Not Applicable

Contacts and Locations

Study Locations

• United States
  • Pennsylvania
    • Philadelphia, Pennsylvania, United States, 19103
      • Penn Medicine

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 18 years and older (Adult, Older Adult)
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

Description

Inclusion Criteria:

• Care for adults with cancer at the following clinics at Perelman Center for Advanced Medicine

  • Breast Oncology
  • Gastrointestinal Oncology
  • Genitourinary Oncology
  • Lymphoma
  • Melanoma and Central Nervous System Oncology
  • Myeloma
  • Thoracic / Head and Neck Oncology
• Care for adults with cancer at the Pennsylvania Hospital Oncology clinic

Exclusion Criteria:

• Providers who care for only patients with benign hematologic disorders
• Providers who see only genetic consults
• Providers who see less than 12 high-risk patients in either the pre- or post- intervention periods
• Visits for patients with lung cancer who are enrolled in an ongoing palliative care clinical trial that may lead to more SICs
• Patient visits that are for oncology genetics consults (such patients may still be included if they see their primary oncologist during the trial)
• Providers who have not undergone serious illness conversation program training (SIC)

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Health Services Research
• Allocation: Randomized
• Interventional Model: Parallel Assignment
• Masking: Double

Number of Arms

2

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
No Intervention: Control; Clinicians will receive current standard communications regarding serious illness performance.
Experimental: Mortality Estimates and Nudges; Clinicians will receive a weekly email with upcoming patients that have high mortality estimates to consider for a serious illness conversation. Clinicians will have the opportunity to review the list and pre-commit (using an opt-out design) to patients appropriate for a conversation. They will receive a nudge on the day of the patient visit through a text message reminding them of their pre-commitment to conduct a serious illness conversation	Behavioral: Nudge; Oncology practices will be randomly assigned to receive an intervention, in which individual clinicians will receive a weekly audit email detailing how many serious illness conversations (SIC) they have had compared to the recommended level, and a link to a list of their patients scheduled in clinic next week at high risk of short-term mortality as identified by a mortality prediction algorithm. Clinicians will have the chance to review the opt-out list and pre-commit to a serious illness conversation with appropriate patients. Clinicians will receive nudge on the day of the patient visit via text message reminding them of their pre-commitment to conduct a serious illness conversation.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Change in the proportion of patients with a documented serious illness conversation (SIC)	The change in the proportion of patients that have an outpatient oncology visit with documentation of a serious illness conversation (SIC)	16 weeks

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Change in the proportion of patients with a documented SIC among those identified as high-risk by the algorithm	The change in the proportion of patients who have an outpatient oncology visit and are identified as high-risk by the machine learning algorithm with documentation of a SIC	16 weeks
Change in the proportion of patients with a documented advanced care planning	The change in the proportion of patients with documentation of advanced care planning.	16 weeks
Change in the proportion of patients with a documented serious illness conversation (SIC) including follow-up	The change in the proportion of patients that have an outpatient oncology visit with documentation of a serious illness conversation (SIC) including follow-up	40 weeks
Change in the proportion of patients with a documented SIC among those identified as high-risk by the algorithm including follow-up	The change in the proportion of patients who have an outpatient oncology visit and are identified as high-risk by the machine learning algorithm with documentation of a SIC including follow-up	40 weeks
Change in the proportion of patients with a documented advanced care planning including follow-up	The change in the proportion of patients with documentation of advanced care planning including follow-up	40 weeks

Other Outcome Measures

Outcome Measure	Measure Description	Time Frame
Oncology Evaluation Center admissions	The number of Oncology Evaluation Center admissions	40 weeks
Healthcare utilization and receipt of chemotherapy in the last 30 days of life	Healthcare utilization in the last 30 days of life in Penn Medicine facilities including acute care utilization as above and receipt of chemotherapy	40 weeks
Number of Emergency department admissions	The number of emergency department admissions	40 weeks
Inpatient admissions	The number of inpatient hospital admissions	40 weeks
Intensive care unit admissions	The number of intensive care unit admissions	40 weeks

Collaborators and Investigators

• Sponsor: University of Pennsylvania
• Investigators: Principal Investigator: Mitesh S Patel, MD, University of Pennsylvania

Publications and helpful links

General Publications

• Parikh RB, Manz CR, Nelson MN, Ferrell W, Belardo Z, Temel JS, Patel MS, Shea JA. Oncologist Perceptions of Algorithm-Based Nudges to Prompt Early Serious Illness Communication: A Qualitative Study. J Palliat Med. 2022 Nov;25(11):1702-1707. doi: 10.1089/jpm.2022.0095. Epub 2022 Aug 18.
• Li EH, Ferrell W, Klaiman T, Kumar P, O'Connor N, Schuchter LM, Chen J, Patel MS, Manz CR, Parikh RB. Impact of Behavioral Nudges on the Quality of Serious Illness Conversations Among Patients With Cancer: Secondary Analysis of a Randomized Controlled Trial. JCO Oncol Pract. 2022 Apr;18(4):e495-e503. doi: 10.1200/OP.21.00024. Epub 2021 Nov 12.
• Manz CR, Parikh RB, Small DS, Evans CN, Chivers C, Regli SH, Hanson CW, Bekelman JE, Rareshide CAL, O'Connor N, Schuchter LM, Shulman LN, Patel MS. Effect of Integrating Machine Learning Mortality Estimates With Behavioral Nudges to Clinicians on Serious Illness Conversations Among Patients With Cancer: A Stepped-Wedge Cluster Randomized Clinical Trial. JAMA Oncol. 2020 Dec 1;6(12):e204759. doi: 10.1001/jamaoncol.2020.4759. Epub 2020 Dec 10. Erratum In: JAMA Oncol. 2022 Feb 24;:
• Manz CR, Parikh RB, Evans CN, Chivers C, Regli SH, Bekelman JE, Small D, Rareshide CAL, O'Connor N, Schuchter LM, Shulman LN, Patel MS. Integrating machine-generated mortality estimates and behavioral nudges to promote serious illness conversations for cancer patients: Design and methods for a stepped-wedge cluster randomized controlled trial. Contemp Clin Trials. 2020 Mar;90:105951. doi: 10.1016/j.cct.2020.105951. Epub 2020 Jan 23.

Study record dates

Study Major Dates

• Study Start (Actual): July 15, 2019
• Primary Completion (Actual): November 1, 2019
• Study Completion (Actual): April 19, 2020

Study Registration Dates

• First Submitted: June 10, 2019
• First Submitted That Met QC Criteria: June 10, 2019
• First Posted (Actual): June 13, 2019

Study Record Updates

• Last Update Posted (Actual): April 24, 2020
• Last Update Submitted That Met QC Criteria: April 23, 2020
• Last Verified: April 1, 2020

More Information

Terms related to this study

• Keywords: advance care planning; opt-out; nudge; serious illness conversations; mortality estimates; pre-commitment
• Additional Relevant MeSH Terms: Neoplasms
• Other Study ID Numbers: 833178

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No