Disclosure Electra Paskett, Ph.D., MPH I have the following relationships with commercial interests: Merck Research grant, in-kind donation Pfizer Stock ownership Meridian Bioscience Stock ownership A commercial interest is any entity producing, marketing, reselling, or distributing health care goods or services consumed by, or used on, patients.
Patient Navigation in Clinical Care: Efficacy and Implementation Electra D. Paskett, Ph.D. Ohio State University The Ohio State University Comprehensive Cancer Center Arthur G. James Cancer Hospital and Richard J. Solove Research Institute
Objectives Define Patient Navigation Present evidence of Patient Navigation efficacy in clinical care Describe how to implement Patient Navigation in health care settings
What is Patient Navigation (PN)?
Historical Perspective PN concept started in 1990 by Harold Freeman at Harlem Hospital NCI funded Patient Navigation Research Program sites in 2004 Patient Navigation Act signed in 2005 CMS PN sites started in 2006 6 sites HRSA PN sites in 2008 6 sites
Initial Contact Conclude Navigation Abnormal Finding Resolution General Framework of Patient Navigation OUTREACH PATIENT NAVIGATION Abnormal finding/diagnosis to resolution Eliminate critical delivery gap for populations experiencing disparities Abnormal results/ Diagnosis REHABILITATION Diagnosis Treatment Survivorship
Patient Navigators Address Individual Patient Barriers Translate medical next steps and what to expect into lay language Promote understanding of healthcare system pathways Increase access to clinical trials Coaching and cultural, emotional and psychosocial support Assistance with physical needs and other barriers to care Facilitate referrals to community resources and social services
Underlying Premise of PN PN will get patients in to the health care system ever or faster Abnormalities will be resolved OR cancers will be treated earlier Resulting in better outcomes: Morbidity and mortality (McLaughlin et al, JCO, 2012) Quality of care/satisfaction Costs
What PN is Navigators are assigned to patients Identify and address ANY barrier to receiving the care recommended by the HCP Follow-up and re-assess; repeat as appropriate Work closely with clinical team, as well Navigators can be: Lay or professional (ie, nurses, social workers) Embedded in the clinic or telephonic
PN is NOT: Get Personalized Pampering at Stanford Cancer Center For example, navigators roam the hallways, offering beverages, newspapers and magazines, and providing a friendly ear. They may also drive people in the tram to appointments, find housing and restaurants for people from out of town, play games with patients while they are receiving infusion therapy, or find a wig or scarf for a chemotherapy patient. Not a Beverage Navigator
What is the present evidence of the efficacy of PN in Clinical Care?
Does Navigation Make a Difference? Before PNRP Data from randomized clinical trials are sparse Observational data support navigation Navigation has been studied along some points of the cancer continuum and not others Certain cancer sites have not received attention Much of the work has been qualitative Paskett, Harrop, Wells CA J Clin, 2011
Continuum of Cancer Navigation in PNRP Prevention Abnormal Screen Diagnosis (T1) Cancer Diagnosis Treatment Start (T2) Throughout Treatment Survivorship Palliative Care PNRP
Literature Review of PN After PNRP Using PN s 24 studies conducted from 2008-2012 (including 5 from individual PNRP sites) Settings included screening as well as diagnostic follow-up and start of treatment Designs used: RCT, cohort, quasi-experimental Majority of studies showed positive benefit of PN
The Patient Navigation Research Program: Overview
PNRP Grantee Map
Healthcare System Settings (n=95) Other 37% Neighborhood Health Center 30% Private Hospital Ambulatory Care 5% Public Hospital Ambulatory Care 28% 58% Public Clinics
Methodology of PNRP Over 12,000 participants recruited Common training protocol for navigators Uniform definitions for eligibility and outcomes Abnormal screening tests or symptoms or Diagnosed cancer Breast Cervical Colorectal Prostate
Patient Navigation and Timeliness of Diagnostic Evaluation
Continuum of Cancer Navigation Prevention Abnormal Screen Diagnosis (T1) Cancer Diagnosis Treatment Start (T2) Throughout Treatment Survivorship Palliative Care
Methods Time to diagnostic resolution Date of Abnormal screen Date of Diagnostic resolution
Examples of Diagnostic Resolution Abnormal Mammogram BIRADS0 Additional films to result in BIRADS1 or 2 Abnormal Mammogram BIRADS4 Biopsy needle, core, excision Referral for abnormal Breast Exam Clinical opinion of specialist of normal findings Evaluation deemed clinically appropriate (eg. normal imaging, biopsy)
Methods Different study designs across PNRP sites Prospective Meta-analysis Outcomes: Time to resolution Hazards Ratios Adjusted for covariates
Sample Size by Screening Abnormality Abnormal Screen Intervention Arm N Control Arm N Total N (%) Breast 3,075 3,643 6,718 (64) Cervical 1,455 1,226 2,681 (26) Colorectal 192 244 436 (4) Prostate 306 311 617 (6) Total 5028 5424 10,452 (100)
Demographic Covariates Variable Value Outcome 1 Diagnostic Evaluation (N=10,521) Intervention Control N (%) N (%) Race /Ethnicity White 1,224 24% 1,370 25% African American 1,487 29% 1,843 34% Hispanic 2,142 42% 1,964 36% Other 207 4% 185 3% Insurance Private 1,202 24% 1,599 29% Public 1,969 39% 2,290 42% Uninsured 1,837 36% 1,548 28% Gender Female 4,665 92% 5,006 92% Marital Married 1,772 35% 1,588 29% Age (yrs) Mean ± SD 43.6 ± 14.8 47.2 ± 14.9 Cancer Type Breast 3,083 61% 3,643 67% Cervical 1,455 29% 1,226 22% Colorectal 219 4% 278 5% Prostate 306 6% 311 6%
Meta-analysis of Impact of Patient Navigation on Diagnostic Resolution after Cancer Screening Abnormality from 0 90 Days Cancer Type N ahr (95% CI) A Breast 1722 B Breast 339 D Breast E Breast 634 472 F Breast 444 G Breast 995 H Breast 1911 A Cervix 1267 B Cervix 533 E Cervix F Cervix D Colorectal 235 595 234 G Colorectal 172 C Prostate 482 D Prostate 129 Overall (I-squared = 86.4%, p = 0.000) * 1.05 (0.96, 1.15) 2.25 (1.84, 2.76) 0.85 (0.67, 1.07) 0.93 (0.51, 1.69) 1.19 (0.90, 1.57) 0.90 (0.77, 1.05) 0.85 (0.69, 1.04) 1.39 (1.11, 1.74) 2.16 (1.63, 2.86) 1.05 (0.83, 1.33) 1.03 (0.75, 1.42) 1.05 (0.66, 1.68) 0.63 (0.38, 1.05) 0.93 (0.71, 1.22) 1.71 (1.11, 2.64) 1.14 (0.96, 1.35) *I squared addresses the heterogeneity of the model, and is not the overall effect of the intervention.3.5 1 2 4 8 Favors Control Favors Navigation JNCI, 2014
Meta-analysis of Impact of Patient Navigation on Diagnostic Resolution after Cancer Screening Abnormality from 91-365 Days Cancer Type N ahr (95% CI) A Breast 1722 B Breast 339 D Breast 634 E Breast 472 F Breast 444 G Breast 995 H Breast 1911 A Cervix 1267 B Cervix 533 E Cervix 235 F Cervix 595 D Colorectal G Colorectal 234 172 C Prostate 482 D Prostate 129 Overall (Isquared = 84.5%, p = 0.000) * 1.05 (0.96, 1.15) 2.25 (1.84, 2.76) 2.44 (1.72, 3.46) 1.36 (0.67, 2.77) 1.19 (0.90, 1.57) 2.08 (1.08, 4.00) 0.70 (0.43, 1.15) 1.39 (1.11, 1.74) 2.16 (1.63, 2.86) 1.05 (0.83, 1.33) 1.23 (0.73, 2.07) 2.17 (1.14, 4.13) 2.41 (0.89, 6.53) 1.41 (0.96, 2.08) 1.71 (1.11, 2.64) 1.51 (1.23, 1.84) *I squared addresses the heterogeneity of the model, and is not the overall effect of the intervention.3.5 1 2 4 8 Favors Control Favors Navigation JNCI, 2014
Effect of Patient Navigation on Time from Definitive Diagnosis to Initiation of Treatment (T2)
Continuum of Cancer Navigation Prevention Abnormal Screen Diagnosis (T1) Cancer Diagnosis Treatment Start (T2) Throughout Treatment Survivorship Palliative Care
Methods Analysis performed on the individual patient level. Kaplan-Meier survival estimates. Cox proportional hazard analysis performed for entire 365 day follow-up period.
T2 Study Population Variable Value Outcome 2 Cancer Treatment (N=2,105) Intervention Control N (%) N (%) Race /Ethnicity White 285 28% 376 35% African American 385 37% 425 40% Hispanic 338 33% 213 20% Other 16 2% 39 4% Insurance Private 342 33% 461 43% Public 448 43% 492 46% Uninsured 236 23% 119 11% Gender Female 874 85% 920 86% Marital Married 383 37% 397 37% Age (yrs) Mean ± SD 51.7 ± 15.0 53.8 ± 15.3 Cancer Type Breast 605 59% 683 64% Cervical 245 24% 207 19% Colorectal 52 5% 58 5% Prostate 130 13% 125 12%
Results: Time to Start of Treatment Impact of Navigation during Diagnostic and Treatment Phases Days 0 90 Adjusted HR (95% CI) Days 91-365 Adjusted HR (95% CI) Diagnostic phase Treatment phase 1.1 (0.96 1.3) 1.5 (1.2 1.8) 0.85 (0.7 1.01) 1.4 (1.1 1.9) JNCI, 2014
Who benefits from PN?
Does Everyone Need PN? Is there a subset of patients who benefit? Can we determine who benefits a priori? Reduce resources for the same benefit = Hot Spotting Percent Resolved 100 80 60 40 20 0 Cervical Abnormality Site A 0 50 100 150 200 250 300 350 400 Days Control Navigated
Navigation Improved Care the Most at Clinical Sites with Greater Delays 100 90 80 70 60 50 40 Control Navigated 30 20 10 0 A B C D F G H I J K L M N 0 P Q
How to Decide Who Needs PN? Who is at risk of delay/loss to follow-up? Who has barriers? Who is helped most by navigation?
Analyses in PNRP Who is at risk of delay/loss to follow-up? More co-morbidities and barriers Incomes <$10,000 Unemployed Less education Renters vs home owners Non-married 2+ dependents
Example: Effects of Barriers on TTR Among 1995 breast and 1194 cervical ppts Range: 0-12 barriers 2/3 of breast and ½ of cervical ppts had at least 1 barrier TTR for any barrier: Breast: ahr=0.744, p<.001 Cervical: ahr=0.792, p<.001 Katz et al., WHI, 2014
Predictor Who Reports a Barrier to Care? (OSU data) Impact of events avoidance score (5-pt increase) OR (95% CI) for having a barrier p-value 1.13 (1.01 1.26) 0.0358 Any comorbidity vs. none 1.55 (0.98 2.46) 0.0635 SF-36 social function = 100 vs. <100 0.55 (0.35 0.86) 0.0093 Multivariable model for barriers (n=380)
Who Does PN Help the Most?
Navigation Eliminated Disparities by Income Household Income Adjusted HR for Control Arm (95% C.I.) Adjusted HR for Navigation Arm (95% C.I.) < $10,000 0.81 (0.64, 1.02) 0.96 (0.76, 1.21) $10,000 $19,999 0.90 (0.71, 1.16) 1.06 (0.83, 1.34) $20,000 - $49,999 0.87 (0.68, 1.10) 1.09 (0.87, 1.36) $50,000 + Ref. 0.95 (0.75, 1.19)
Navigation Eliminated Disparities by Employment Employment Adjusted HR for Control Arm (95% C.I.) Adjusted HR for Navigation Arm (95% C.I.) Full time Ref. 1.15 (1.00, 1.34) Part time 1.00 (0.82, 1.23) 1.32 (1.11, 1.57) Unemployed 0.85 (0.74, 0.98) 1.12 (0.98, 1.29)
Impact of Navigation for patients with other Comorbidities Time to Diagnostic Resolution for Participants with 2+ Major Comorbidities
Types of Barriers to Care Delay Time to Diagnostic Resolution 100 90 80 70 %Resolved 60 50 40 30 20 No Barriers Other Barriers Social Barriers 10 0 0 50 100 150 200 250 300 350 400 Primeau et al, in press, JGIM Number of days to diagnostic resolution
Summary Patient navigation results in: Reduction in delays in care Reduction in # of women who never complete needed care Most effective where the gaps and needs are the greatest
Implementation of PN in health care settings
Policy Standards American College of Surgeons: By 2015 PN required for accreditation PN Assistance Act : Re-introduced, hopefully, soon Money for PN in CMS programs Affordable Care Act (ACA): Provisions for PN Not clearly specified how will be paid for
Clinical Care Implications Practices can identify highest risk patients for delays, those who would benefit: Practices with delays in care Low income, unemployed Housing insecurity Comorbidities Barriers to Care
Which Tasks and Networks Improve Care? Tasks/Network Patient Provider Staff Support EMR Navigate with specific patient (tell, inquire, support, coach) Facilitate for specific patient (coordinate care, seek advice, find supports) Maintain system for all patients (find potential pts, build int/ext networks) Document/Review (record info, actions, results) Other (provide clinic back-up, do non-nav tasks)
Which Tasks and Networks Improve Care? Tasks/Network Patient Provider Staff Support EMR Navigate with specific patient (tell, inquire, support, coach) Yes Facilitate for specific patient (coordinate care, seek advice, find supports) Yes Yes Yes Maintain system for all patients (find potential pts, build int/ext networks) Document/Review (record info, actions, results) Other (provide clinic back-up, do non-nav tasks) Yes
Which Tasks and Networks Detract from Care? Tasks/Network Patient Provider Staff Support EMR Navigate with specific patient (tell, inquire, support, coach) Facilitate for specific patient ((coordinate care, seek advice, find supports) Maintain system for all patients (find potential pts, build int/ext networks) Document/Review (record info, actions, results) Other (provide clinic back-up, do non-nav tasks) X X
Conclusions Overwhelming evidence that PN can: Reduce delays to receiving follow-up care for cancer abnormalities Reduce delays to starting cancer treatment Reduce those lost to follow-up care Hot spotting of at risk patients Policy implications for PN in current laws
Acknowledgements Patient Navigation Research Project (PNRP) Funded by the NCI Center to Reduce Cancer Health Disparities (RFA- CA-05-019) in partnership with the American Cancer Society and the AVON Foundation Special Thanks to: The PNRP Investigators, Staff, NOVA Research Company, Community Partners, and Participants
Thank you! The Ohio State University Comprehensive Cancer Center Arthur G. James Cancer Hospital and Richard J. Solove Research Institute