[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure 1.  Assessment for Eligibility and Follow-up
Assessment for Eligibility and Follow-up

Flow diagram of patient flow through the study.

Figure 2.  Mean Edmonton Symptom Assessment System (ESAS) Scores and Mean Patient Health Questionnaire-9 (PHQ-9) Scores
Mean Edmonton Symptom Assessment System (ESAS) Scores and Mean Patient Health Questionnaire-9 (PHQ-9) Scores

A, Mean ESAS scores are measured on a numerical scale of 0 (absent) to 10 (worse). B, Mean PHQ-9 scores are measured on a numerical scale of 0 (not at all) to 3 (nearly every day) and scored from 0 to 27 where a score of 5 is a cutpoint for mild depression, a score of 10 is a cutpoint for moderate depression, a score of 15 is a cutpoint for moderately severe depression, and a score of 20 is a cutpoint for severe depression. P < .001 for difference in change in mean symptom scores between groups from baseline to 12-month follow-up.

Table 1.  Baseline Characteristics of Study Participants in Clinics A, B, C, D, and Ea
Baseline Characteristics of Study Participants in Clinics A, B, C, D, and Ea
Table 2.  Baseline Characteristics of Study Participants in Clinics F, G, H, and Ia
Baseline Characteristics of Study Participants in Clinics F, G, H, and Ia
Table 3.  Data on Health Care Use
Data on Health Care Use
1.
Basch  E, Deal  AM, Dueck  AC.  Patient-reported symptom monitoring during chemotherapy—reply.  JAMA. 2017;318(19):1935-1936. doi:10.1001/jama.2017.14903PubMedGoogle ScholarCrossref
2.
Basch  E, Deal  AM, Dueck  AC,  et al.  Overall survival results of a trial assessing patient-reported outcomes for symptom monitoring during routine cancer treatment.  JAMA. 2017;318(2):197-198. doi:10.1001/jama.2017.7156PubMedGoogle ScholarCrossref
3.
Basch  E, Deal  AM, Kris  MG,  et al.  Symptom monitoring with patient-reported outcomes during routine cancer treatment: a randomized controlled trial.  J Clin Oncol. 2016;34(6):557-565. doi:10.1200/JCO.2015.63.0830PubMedGoogle ScholarCrossref
4.
Reilly  CM, Bruner  DW, Mitchell  SA,  et al.  A literature synthesis of symptom prevalence and severity in persons receiving active cancer treatment.  Support Care Cancer. 2013;21(6):1525-1550. doi:10.1007/s00520-012-1688-0PubMedGoogle ScholarCrossref
5.
van den Beuken-van Everdingen  MH, Hochstenbach  LM, Joosten  EA, Tjan-Heijnen  VC, Janssen  DJ.  Update on prevalence of pain in patients with cancer: systematic review and meta-analysis.  J Pain Symptom Manage. 2016;51(6):1070-1090.e9. doi:10.1016/j.jpainsymman.2015.12.340PubMedGoogle ScholarCrossref
6.
Greco  MT, Roberto  A, Corli  O,  et al.  Quality of cancer pain management: an update of a systematic review of undertreatment of patients with cancer.  J Clin Oncol. 2014;32(36):4149-4154. doi:10.1200/JCO.2014.56.0383PubMedGoogle ScholarCrossref
7.
Kamal  AH, Nipp  RD, Bull  J, Stinson  CS, Abernethy  AP.  Symptom burden and performance status among community-dwelling patients with serious illness.  J Palliat Med. 2015;18(6):542-544. doi:10.1089/jpm.2014.0381PubMedGoogle ScholarCrossref
8.
Deandrea  S, Montanari  M, Moja  L, Apolone  G.  Prevalence of undertreatment in cancer pain: a review of published literature.  Ann Oncol. 2008;19(12):1985-1991. doi:10.1093/annonc/mdn419PubMedGoogle ScholarCrossref
9.
Patel  MI, Williams  DC, Wohlforth  C, Fisher  G, Wakelee  HA, Blayney  DW.  Are patients with thoracic malignancies at risk for uncontrolled symptoms?  J Oncol Pract. 2015;11(1):e98-e102. doi:10.1200/JOP.2014.001502PubMedGoogle ScholarCrossref
10.
Nipp  RD, El-Jawahri  A, Moran  SM,  et al.  The relationship between physical and psychological symptoms and health care utilization in hospitalized patients with advanced cancer.  Cancer. 2017;123(23):4720-4727. doi:10.1002/cncr.30912PubMedGoogle ScholarCrossref
11.
Zaza  C, Baine  N.  Cancer pain and psychosocial factors: a critical review of the literature.  J Pain Symptom Manage. 2002;24(5):526-542. doi:10.1016/S0885-3924(02)00497-9PubMedGoogle ScholarCrossref
12.
Barkley  R, Soobader  MJ, Wang  J, Blau  S, Page  RD.  Reducing cancer costs through symptom management and triage pathways.  J Oncol Pract. 2019;15(2):e91-e97. doi:10.1200/JOP.18.00082PubMedGoogle ScholarCrossref
13.
Patel  MI, Periyakoil  VS, Blayney  DW,  et al.  Redesigning cancer care delivery: views from patients and caregivers.  J Oncol Pract. 2017;13(4):e291-e302. doi:10.1200/JOP.2016.017327PubMedGoogle ScholarCrossref
14.
Patel  MI, Moore  D, Bhattacharya  J, Milstein  A, Coker  TR.  Perspectives of health care payer organizations on cancer care delivery redesign: a national study.  J Oncol Pract. 2019;15(1):e46-e55. doi:10.1200/JOP.18.00331PubMedGoogle ScholarCrossref
15.
Patel  MI, Periyakoil  VS, Moore  D, Nevedal  A, Coker  TR.  Delivering end-of-life cancer care: perspectives of providers.  Am J Hosp Palliat Care. 2018;35(3):497-504. doi:10.1177/1049909117719879PubMedGoogle ScholarCrossref
16.
Patel  MI, Moore  D, Milstein  A.  Redesigning advanced cancer care delivery: three ways to create higher value cancer care.  J Oncol Pract. 2015;11(4):280-284. doi:10.1200/JOP.2014.001065PubMedGoogle ScholarCrossref
17.
Patel  MI, Ramirez  D, Agajanian  R, Agajanian  H, Bhattacharya  J, Bundorf  KM.  Lay health worker-led cancer symptom screening intervention and the effect on patient-reported satisfaction, health status, health care use, and total costs: results from a tri-part collaboration  [published online September 24, 2019].  J Oncol Pract. doi:10.1200/JOP.19.00152PubMedGoogle Scholar
18.
Rocque  GB, Pisu  M, Jackson  BE,  et al; Patient Care Connect Group.  Resource use and Medicare costs during lay navigation for geriatric patients with cancer.  JAMA Oncol. 2017;3(6):817-825. doi:10.1001/jamaoncol.2016.6307PubMedGoogle ScholarCrossref
19.
Ogrinc  G, Davies  L, Goodman  D, Batalden  P, Davidoff  F, Stevens  D.  SQUIRE 2.0 (Standards for Quality Improvement Reporting Excellence): revised publication guidelines from a detailed consensus process.  BMJ Qual Saf. 2016;25(12):986-992. doi:10.1136/bmjqs-2015-004411PubMedGoogle ScholarCrossref
20.
Bruera  E, Kuehn  N, Miller  MJ, Selmser  P, Macmillan  K.  The Edmonton Symptom Assessment System (ESAS): a simple method for the assessment of palliative care patients.  J Palliat Care. 1991;7(2):6-9. doi:10.1177/082585979100700202PubMedGoogle ScholarCrossref
21.
Kroenke  K, Spitzer  RL, Williams  JB.  The PHQ-9: validity of a brief depression severity measure.  J Gen Intern Med. 2001;16(9):606-613. doi:10.1046/j.1525-1497.2001.016009606.xPubMedGoogle ScholarCrossref
22.
Centers for Medicare & Medicaid Services. Risk adjustment. https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Risk-Adjustors.html. Accessed February 5, 2020.
23.
Kahan  BC.  Accounting for centre-effects in multicentre trials with a binary outcome—when, why, and how?  BMC Med Res Methodol. 2014;14:20. doi:10.1186/1471-2288-14-20PubMedGoogle ScholarCrossref
24.
STATACorp LLC. New in Stata 16. https://www.stata.com/new-in-stata/. Accessed August 15, 2019.
25.
Kroenke  K, Theobald  D, Norton  K,  et al.  The Indiana Cancer Pain and Depression (INCPAD) trial design of a telecare management intervention for cancer-related symptoms and baseline characteristics of study participants.  Gen Hosp Psychiatry. 2009;31(3):240-253. doi:10.1016/j.genhosppsych.2009.01.007PubMedGoogle ScholarCrossref
26.
Yount  SE, Rothrock  N, Bass  M,  et al.  A randomized trial of weekly symptom telemonitoring in advanced lung cancer.  J Pain Symptom Manage. 2014;47(6):973-989. doi:10.1016/j.jpainsymman.2013.07.013PubMedGoogle ScholarCrossref
27.
French  EB, McCauley  J, Aragon  M,  et al.  End-of-life medical spending in last twelve months of life is lower than previously reported.  Health Aff (Millwood). 2017;36(7):1211-1217. doi:10.1377/hlthaff.2017.0174PubMedGoogle ScholarCrossref
28.
Garcia  SF, Wortman  K, Cella  D,  et al.  Implementing electronic health record–integrated screening of patient-reported symptoms and supportive care needs in a comprehensive cancer center.  Cancer. 2019;125(22):4059-4068. doi:10.1002/cncr.32172PubMedGoogle ScholarCrossref
29.
Strasser  F, Blum  D, von Moos  R,  et al; Swiss Group for Clinical Cancer Research (SAKK).  The effect of real-time electronic monitoring of patient-reported symptoms and clinical syndromes in outpatient workflow of medical oncologists: E-MOSAIC, a multicenter cluster-randomized phase III study (SAKK 95/06).  Ann Oncol. 2016;27(2):324-332. doi:10.1093/annonc/mdv576PubMedGoogle ScholarCrossref
30.
Holch  P, Warrington  L, Bamforth  LCA,  et al.  Development of an integrated electronic platform for patient self-report and management of adverse events during cancer treatment.  Ann Oncol. 2017;28(9):2305-2311. doi:10.1093/annonc/mdx317PubMedGoogle ScholarCrossref
31.
Levine  DM, Lipsitz  SR, Linder  JA.  Trends in seniors’ use of digital health technology in the United States, 2011-2014.  JAMA. 2016;316(5):538-540. doi:10.1001/jama.2016.9124PubMedGoogle ScholarCrossref
32.
Rocque  GB, Williams  CP, Jones  MI,  et al.  Healthcare utilization, Medicare spending, and sources of patient distress identified during implementation of a lay navigation program for older patients with breast cancer.  Breast Cancer Res Treat. 2018;167(1):215-223. doi:10.1007/s10549-017-4498-8PubMedGoogle ScholarCrossref
33.
Patel  MI, Sundaram  V, Desai  M,  et al.  Effect of a lay health worker intervention on goals-of-care documentation and on health care use, costs, and satisfaction among patients with cancer: a randomized clinical trial.  JAMA Oncol. 2018;4(10):1359-1366. doi:10.1001/jamaoncol.2018.2446PubMedGoogle ScholarCrossref
34.
Patel  MI, Moore  D, Coker  TR.  End-of-life cancer care redesign: patient and caregiver experiences in a lay health worker-led intervention.  Am J Hosp Palliat Care. 2019;36(12):1081-1088. doi:10.1177/1049909119847967PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    Health Policy
    March 16, 2020

    Association of a Lay Health Worker Intervention With Symptom Burden, Survival, Health Care Use, and Total Costs Among Medicare Enrollees With Cancer

    Author Affiliations
    • 1Division of Oncology, Stanford University School of Medicine, Stanford, California
    • 2Medical Services, Veterans Affairs Palo Alto Health Care System, Palo Alto, California
    • 3Center for Primary Care and Outcomes Research/Health Research and Policy, Stanford University School of Medicine, Stanford, California
    • 4CareMore Health, Cerritos, California
    • 5The Oncology Institute of Hope and Innovation, Downy, California
    • 6Seattle Children’s Research Institute, Seattle, Washington
    • 7Department of Pediatrics, University of Washington School of Medicine, Seattle
    JAMA Netw Open. 2020;3(3):e201023. doi:10.1001/jamanetworkopen.2020.1023
    Key Points español 中文 (chinese)

    Question  Is a lay health worker–led screening and referral intervention associated with patient symptom burden, acute care use, and total costs of care?

    Findings  In this multisite quality improvement study of 425 Medicare Advantage enrollees with a diagnosis of cancer who were provided usual cancer care augmented by a lay health worker trained to proactively screen patient symptoms, discuss symptoms with a physician assistant, and refer patients with uncontrolled symptoms to palliative care and behavioral medicine compared with 407 control patients diagnosed and treated in the year prior, patients in the intervention group were associated with significant reductions in symptom burden over time, and patients in the control group were associated with worsening symptoms over time. Patients who received the intervention were associated with fewer inpatient and emergency department visits and lower median total costs, and there were no differences in survival.

    Meaning  The findings suggest that a lay health worker–led symptom screening and proactive referral intervention may improve value-based cancer care.

    Abstract

    Importance  Undertreated patient symptoms require approaches that improve symptom burden.

    Objective  To determine the association of a lay health worker–led symptom screening and referral intervention with symptom burden, survival, health care use, and total costs among Medicare Advantage enrollees with a new diagnosis of solid or hematologic malignant neoplasms.

    Design, Setting, and Participants  This quality improvement study conducted at 9 community oncology practices from November 1, 2016, to October 31, 2018, compared newly diagnosed Medicare Advantage enrollees with solid or hematologic malignant neoplasms with patients diagnosed and treated 1 year prior. Analysis was conducted from August 1, 2019, to January 11, 2020.

    Interventions  Usual care augmented by a lay health worker trained to screen symptoms and refer patients to palliative care and behavioral medicine.

    Main Outcomes and Measures  The primary outcome was change in symptoms using the Edmonton Symptom Assessment Scale and the 9-item Patient Health Questionnaire at baseline and 6 and 12 months after enrollment. Secondary outcomes were between-group comparison of survival, 12-month health care use, and costs.

    Results  Among 425 patients in the intervention group and 407 patients in the control group, the mean (SD) age was 78.8 (8.3) years, 345 (41.5%) were female, and 407 (48.9%) were non-Hispanic white. Patients in the intervention group experienced a lower symptom burden as measured by the Edmonton Symptom Assessment Scale score over time compared with patients in the control group (mean [SD] difference, −1.9 [14.2]; 95% CI, −3.77 to −0.19; P = .01 for the intervention group and 2.32 [17.7]; 95% CI, 0.47 to 4.19; P = .02 for the control group). Similar findings were noted in 9-item Patient Health Questionnaire depression scores (mean [SD] difference, −0.63 [3.99]; 95% CI, −1.23 to −0.028; P = .04 for the intervention group and 1.67 [5.49]; 95% CI, 0.95 to 2.37; P = .01 for the control group). Patients in the intervention group compared with patients in the control group had fewer mean (SD) inpatient visits (0.54 [0.77]; 95% CI, 0.47-0.61 vs 0.72 [1.12]; 95% CI, 0.61-0.83; P = .04) and emergency department visits (0.43 [0.76]; 95% CI, 0.36-0.50 vs 0.57 [1.00]; 95% CI, 0.48-0.67; P = .002) per 1000 patients per year and lower total costs (median, $17 869 [interquartile range, $6865-$32 540] vs median, $18 473 [interquartile range, $6415-$37 910]; P = .02). A total of 180 patients in the intervention group and 189 patients in the control group died within 12 months. Among those who died, patients in the intervention group had greater hospice use (125 of 180 [69.4%] vs 79 of 189 [41.8%]; odds ratio, 3.16; 95% CI, 2.13-4.69; P < .001), fewer mean (SD) emergency department and hospital visits (emergency department: 0.10 [0.30]; 95% CI, 0.06-0.14 vs 0.30 [0.46]; 95% CI, 0.24-0.38; P = .001; hospital: 0.27 [0.44]; 95% CI, 0.21-0.34 vs 0.43 [0.82]; 95% CI, 0.32-0.55; P = .02), and lower costs (median, $3602 [interquartile range, $1076-$9436] vs median, $12 726 [interquartile range, $5259-$22 170]; P = .002), but there was no significant difference in inpatient deaths (18 of 180 [10.0%] vs 30 of 189 [15.9%]; P = .14).

    Conclusions and Relevance  This study suggests that a lay health worker–led intervention may be one way to improve burdensome and costly care.

    Introduction

    Despite efforts to improve symptoms among adults with cancer,1-4 suboptimal management of symptoms continues to affect many patients.5-8 Undertreated symptoms are associated with acute care use,9 inpatient deaths,10 psychological and physical distress,11 and financial burden to patients and the health care system.12 Barriers to appropriate management of symptoms include lack of infrastructure,13,14 clinician time,15 and reimbursement.15,16

    A lay health worker (LHW)–led (ie, personnel who are not clinically trained) symptom screening intervention for adults with advanced cancer was previously developed.17 The goal was to proactively identify symptoms to prevent psychological and physical distress and decrease acute care use. The LHW was trained to assess symptoms and relay them to a supervising physician assistant (PA), who conducted any necessary interventions. Compared with patients who received a diagnosis of cancer the year prior, patients in the intervention group had better experiences of care, a better self-reported overall health status, a better self-reported mental and emotional health status, fewer emergency department (ED) visits and hospitalizations, lower health care costs, and no differences in hospice use compared with patients who were not in the intervention group.

    In this study, we expanded the LHW-led intervention for all newly diagnosed patients regardless of cancer stage and implemented it across 9 community-based clinics. We also trained the LHW to refer patients with uncontrolled symptoms to palliative care and behavioral medicine, with the intended goal to improve symptom burden early and prevent acute care use. Although 1 prior study assessed the navigation of distress screening by LHWs,18 to our knowledge, there have been no studies that examine an LHW-led symptom screening and supportive services referral intervention across multiple community practices.

    The primary objective of this study was to evaluate the association of this LHW-led symptom screening intervention with patient-reported symptom burden. The secondary objectives were to evaluate association of this LHW-led symptom screening intervention with survival, health care use, total health care costs, and end-of-life care among patients who died within 12 months of diagnosis.

    Methods
    Study Design and Population

    We conducted a multisite quality improvement study using a pre-post design with CareMore Health, a Medicare Advantage health plan, and 9 of 10 oncology clinics within the Oncology Institute of Hope and Innovation in Los Angeles, California. We excluded 1 clinic because it was involved in the previously reported intervention.17 We enrolled all CareMore Medicare Advantage beneficiaries with newly diagnosed solid tumors or hematologic malignant neoplasms in these 9 clinics and excluded patients who did not receive medical oncology care. Patients were screened by a new patient coordinator, enrolled from November 1, 2016, to October 31, 2017, and followed up for 12 months after enrollment or until death, whichever was first. The comparison group (control group) was composed of CareMore Medicare Advantage beneficiaries with newly diagnosed solid tumors or hematologic malignant neoplasms who received medical oncology care at these same 9 clinics in the year prior to the intervention, from November 1, 2015, to October 31, 2016. The control group received usual cancer care without standardized symptom screening or management. We identified the control group through both CareMore claims data for all patients in these 9 clinics with 2016 International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Clinical Modification codes C00-D49, indicating receipt of cancer services during the dates of interest, and verified by the Oncology Institute of Hope and Innovation cancer registry, which contains data on patients’ names, payers, clinics, cancer stages, cancer diagnoses, symptom screening dates and scores, and diagnosis dates. The Stanford University Institutional Review Board reviewed and approved this study as quality improvement. Because the study was implemented as a process change for all patients in the clinic, the institutional review board waived formal patient consent. This study followed the Standards for Quality Improvement Reporting Excellence (SQUIRE) reporting guidelines.19

    Expanded LWH-Led Symptom Screening Intervention

    The 12-month telephonic intervention consisted of 1 LHW who assessed symptoms using the validated Edmonton Symptom Assessment Scale (ESAS; score range, 0-10, where 0 indicates no symptoms and 10 indicates the worst possible symptoms)20 and screened for depression using the validated 9-item Patient Health Questionnaire (PHQ-9; score range, 0-27, where a score of 5 was a cutpoint for mild depression, a score of 10 was a cutpoint for moderate depression, a score of 15 was a cutpoint for moderately severe depression, and a score of 20 was a cutpoint for severe depression) (eAppendix 1 in the Supplement).21 All patients in both groups completed the ESAS and PHQ-9 at the first oncology visit after their cancer diagnosis and again at 6 and 12 months after diagnosis. The LHW contacted all eligible patients across the 9 clinic sites 1 week prior to their first oncology appointment. Patients were risk stratified into a high-risk category if they received a diagnosis of metastatic cancer, were undergoing active chemotherapy, or reported moderate or greater symptom burden (ESAS score ≥4) or moderate or greater depression risk (PHQ-9 score ≥10). All other patients were stratified into the low-risk category. The supervising PA risk stratified patients weekly and reassigned them based on changes in clinical characteristics. The LHW called patients in the high-risk category weekly and called those in the low-risk category monthly. The LHW documented all scores in the electronic health record and forwarded an encounter note to the PA and oncology clinician on the same day for review. To improve symptom relief for patients with uncontrolled symptoms identified by the LHW, the LHW proactively referred patients with a change in symptom scores of greater than 2 points in subsequent assessments to palliative care and to behavioral medicine if their PHQ-9 score reflected moderate depression. Telephone calls were conducted Monday through Friday between 8:30 am and 5:00 pm. The LHW also provided patients with basic education about palliative care services. The LHW had a Bachelor of Arts degree, had no prior clinical experience, worked 40 hours per week, and completed a 10-hour training led by the PA. The PA dedicated 25% full-time equivalent for this program, met in-person with the LHW once weekly to discuss patients, and reviewed all LHW documentation daily.

    CareMore Health supported this intervention by providing financial support for the LHW, 25% full-time equivalent for the PA, and a $150 per member per month incentive to cover any additional costs, such as clinical time, incurred to the oncology clinicians for conducting the intervention. CareMore Health preauthorized all LHW referrals to palliative care and behavioral medicine.

    Outcomes

    Prior to conducting the study, investigators, payer executives, and clinicians agreed to measure outcomes to determine the intervention’s success. The decision was not to measure intermediary variables (eg, referrals to palliative care) but to measure associations wtih patient-reported symptoms, healthcare use (ie, ED visits and hospitalizations), and total health care spending. A preplanned subset analysis evaluated the association with end-of-life care and costs (ie, ED visits, hospitalizations and total costs in the last month of life, hospice use, and acute care facility deaths) for the patients who died within 12 months of diagnosis. Process metrics included intervention deviations, percentage of patients contacted 1 week before their first oncology appointment, and mean number of LHW-patient contacts.

    Survival

    We collected data on vital status and death dates from CareMore Health registrars. These data were validated by the Oncology Institute of Hope and Innovation cancer registry.

    Health Care Use and Health Care Costs

    We collected data for the following cancer care services and total costs from CareMore Health claims data for all patients: palliative care use; behavioral medicine use; dates of all ED visits, hospital admissions, and hospital discharges; and the use of hospice services. We measured all health care use within the 12-month enrollment period (which also represented a 12-month postdiagnosis period) for the intervention group and the 12-month postdiagnosis period for the control group. We collected all claims data from CareMore Health to measure total costs, including inpatient and outpatient costs for the intervention and control groups. Using these same data sources for patients who died during the 12-month enrollment or postdiagnosis period, we measured ED visits, hospitalizations, and total costs in the last 30 days of life as well as acute care facility deaths.

    Demographic and Clinical Characteristics

    We collected data on age, sex, self-reported race/ethnicity, cancer diagnosis, cancer stage, and clinic site from the electronic health record. We collected data on the Centers for Medicare & Medicaid Services Hierarchical Condition Category risk adjustment factor,22 which is based on a patient’s underlying health conditions, sex, age, Medicaid enrollment, and disabled status using CareMore Health claims data.

    Statistical Analysis

    We conducted the statistical analysis from August 1, 2019, to January 11, 2020. We used descriptive statistics to compare differences between the groups for clinical and demographic characteristics. We compared between-group differences in symptom scores from baseline to 6 and 12 months using mixed-effect generalized linear regression models for repeated measures. We compared survival using Kaplan-Meier methods and risk of death using Cox proportional hazards regression models. We compared health care use (ED and hospitalizations) per patient using Poisson regression models with an offset term for length of follow-up and normalized ED visits and hospital admission rates per 1000 members per year. We compared health care costs using generalized linear regression models with a gamma link-log function to account for skewed data with an offset term for length of follow-up. All models were adjusted for age, sex, stage of disease, cancer diagnosis, and risk adjustment factor. All models included random effects to account for clustering of the individual within the clinic site.23 All P values were from 2-sided tests, and the results were deemed statistically significant at P = .05. We used Stata, version 15.1 (StataCorp)24 to conduct all analyses. Regression model results are presented as between-group differences with 95% CIs. We accounted for a 10-month administrative delay in claims.

    Results

    A total of 407 patients were identified in the control group and 425 patients participated in the intervention (Figure 1). There were no patients who dropped out of the study or who were lost to follow-up (Figure 1). There were no differences in age, race/ethnicity, cancer diagnosis, cancer stage, or risk adjustment factor between the groups across clinics; however, clinic A had a higher proportion of male patients in the control group compared with the intervention group (50 of 75 [66.7%] vs 42 of 80 [52.5%]; Table 1 and Table 2).

    Process Metrics

    There were no deviations to the intervention processes during the study. The LHW contacted all patients 1 week prior to their first oncology visit. The mean (SD) number of LHW-patient contacts was 32.8 (11.2) (eAppendix 2 in the Supplement).

    Symptom Scores

    Figure 2 depicts mean symptom scores by group at baseline and 6 and 12 months after diagnosis. The patients in the intervention group experienced reductions in symptoms over time (mean [SD] change in total ESAS score, −1.9 [14.2]; 95% CI, −3.77 to −0.19; P = .01), whereas the patients in the control group experienced worsening symptoms over time (mean [SD] change in total ESAS score, 2.32 [17.7]; 95% CI, 0.47-4.19; P = .02). The mean [SD] between-group differences were statistically significant (–4.75 [1.01]; 95% CI, –6.72 to –2.78; P = .02). With the exception of drowsiness and appetite (which did not change significantly among patients in either group), the patients in the intervention group experienced significant reductions in all symptoms compared with the patients in the control group over time (eFigure 1 in the Supplement). The patients in the intervention group also experienced significant reductions in the severity of depression (change in mean [SD] total PHQ-9 scores, −0.63 [3.99]; 95% CI, −1.23 to −0.03; P = .04); however, the patients in the control group experienced an increase in mean PHQ-9 scores over time (change in mean [SD] total PHQ-9 scores, 1.67 [5.49]; 95% CI, 0.95-2.37; P = .01) (Figure 2). These mean (SD) between-group differences were statistically significant (–2.07 [0.37]; 95% CI, –2.80 to –1.34; P = .001).

    Survival

    At 12-month follow-up, there were no differences in median survival across both groups, with 180 of 425 patients (42.4%) having died among the intervention group and 189 of 407 patients (46.4%) having died among the control group (eFigure 2 in the Supplement). In Cox proportional hazards regression models, there were no differences in the risk of death after adjusting for age, stage of cancer, sex, cancer diagnosis, and risk adjustment factor across groups (hazard ratio for intervention vs control, 0.92; 95% CI, 0.75-1.13; P = .41).

    Health Care Use and Total Health Care Costs

    Table 3 reports differences in health care use and costs between groups. The patients in the intervention group experienced greater use of palliative care (207 of 425 [48.7%] vs 101 of 407 [24.8%]; odds ratio, 2.83; 95% CI, 1.05-7.93; P = .04) and behavioral medicine (126 of 425 [29.6%] vs 46 of 407 [11.3%]; odds ratio 3.41; 95% CI, 2.76-4.21; P < .001) compared with the patients in the control group. The patients in the intervention group experienced fewer ED visits (mean [SD], 0.43 [0.76]; 95% CI, 0.36-0.50 vs 0.57 [1.00]; 95% CI, 0.48-0.67; P = .002) and hospitalizations (mean [SD], 0.54 [0.77]; 95% CI, 0.47-0.61 vs 0.72 [1.12]; 95% CI, 0.61-0.83; P = .04) per 1000 members per year compared with the patients in the control group. The patients in the intervention group also experienced lower median total costs ($17 869 [interquartile range, $6865-$32 540] vs $18 473 [interquartile range, $6415-$37 910]; P = .02) compared with the patients in the control group.

    Health Care Use and Total Costs of End-of-Life Care

    Among 189 patients in the control group and 180 patients in the intervention group who died within 12 months after diagnosis, there were significant differences in health care use and total costs at the end of life (Table 3). The patients in the intervention group experienced fewer ED visits (mean [SD], 0.10 [0.30]; 95% CI, 0.06-0.14 vs 0.30 [0.46]; 95% CI, 0.24-0.38; P = .001) and hospitalizations (0.27 [0.44]; 95% CI, 0.21-0.34 vs 0.43 [0.82]; 95% CI, 0.32-0.55; P = .02) compared with the patients in the control group. A greater proportion of patients in the intervention group than patients in the control group used hospice care (125 of 180 [69.4%] vs 79 of 189 [41.8%]; odds ratio, 3.16; 95% CI, 2.13-4.69; P < .001). The proportion of acute care facility deaths was not significantly different between intervention and control patients (18 of 180 [10.0%] vs 30 of 189 [15.9%]; odds ratio, 0.58; 95% CI, 0.28-1.20; P = .14). In the last 30 days of life, patients in the intervention group experienced lower median total costs compared with patients in the control group ($3602 [interquartile range, $1076-$9436] vs $12 726 [interquartile range, $5259-$22 170]; P = .002). There were also significant differences in the distribution of end-of-life total costs between the 2 groups (eFigure 3 in the Supplement).

    Discussion

    This multisite LHW-led proactive symptom screening intervention for patients with newly diagnosed cancer was associated with lower symptom burden, acute care use, and total costs. The intervention was also associated with better quality of end-of-life care, less acute care use, and lower total end-of-life costs. Our findings of lower symptom burden associated with proactive symptom screening are similar to previous work17 and consistent with the literature.1,25,26 In this intervention, we expanded the LHW role to refer patients to palliative care and behavioral medicine in response to worsening symptoms and found improved symptoms, lower acute care use, and lower total costs within 1 year of diagnosis and at the end of life compared with a control group that received no standardized symptom screening interventions.

    Undertreated symptoms are associated with prolonged hospitalizations and acute care readmissions.10 At the end of life, we found less health care cost variation among patients in the intervention group. Our findings may reflect associated lower acute care use and hospitalization length among patients in the intervention group due to lower symptom burden but could also reflect uniformity of services delivered, given that variations in end-of-life costs are often due to differing combinations of services provided in the acute care setting.27 More investigations are required to understand these findings.

    Our study differs from technology-based approaches such as electronic portals or mobile phone–based symptom-reporting applications.28-30 These approaches can create barriers for patients, such as older adults, who may not prefer or be as comfortable with the use of technology-based health approaches.31 Furthermore, we chose to use a personnel-based approach based on patient and caregiver feedback.13 Unlike other personnel-led interventions that use health care professionals,2,25,26 we trained an LHW and expanded the intervention to all patients regardless of their cancer stage. Compared with studies using a lay navigation program,32 our intervention focused on symptom screening and referral and assessed their associations with health care use and cost and, perhaps more important, symptom burden over time. Although effective in advance care planning,33 LHW-led symptom screening and proactive referral to supportive care throughout the care continuum is unique. It is possible that our approach may have unmeasured benefits that technology-based approaches may lack. These unmeasured aspects, such as social support and counseling,34 in the LHW-patient relationship could be important in understanding the greater use of palliative care, behavioral medicine, and hospice care that we report. Future studies should evaluate these aspects.

    Another unique aspect of our study was our collaboration with a payer that financially supported the implementation and community oncology practices that delivered the intervention. This stakeholder-engaged initiative13-15 allowed us to design benefits that could meet the needs of the patients, clinicians, and payers involved. One such example was the preauthorization for all LHW referrals to palliative care and behavioral medicine. By collaborating with the payer to bypass impediments to care, such as insurance authorization, patients in the intervention group could access timely and appropriate care for their reported symptoms. Another example is the $150 per member per month incentive from the payer to the oncology practice that enabled financial reimbursement for additional time incurred by the practice to conduct the intervention. Although early identification of symptoms was not associated with unscheduled clinic visits, given that the PA conducted all interventions via telephone, oncologists spent additional time daily reviewing LHW and supportive care services documentation and approving interventions conducted by the PA.

    Limitations

    This study has some limitations. The patients in the intervention group were compared with a retrospective cohort of patients; thus, the causality of the intervention and our findings cannot be inferred. Although the demographic characteristics between the control and intervention groups were similar, there were differences, for example, by sex in 1 clinic. Although we are unaware of changes in practice patterns or secular trends during this time frame, it is possible that there could be differences in unmeasured variables independent of this intervention that could bias our findings. Second, we were unable to measure the PA interventions made in response to symptoms, which may be important in understanding our outcomes. Another limitation was the practice setting. We conducted the intervention among 9 community-based clinics in Los Angeles, California, that were associated with 1 cancer institute and limited the intervention to elderly CareMore Health Medicare Advantage beneficiaries. Thus, generalizability to other populations and organizations may be limited. Furthermore, 1 LHW and 1 PA conducted this intervention. It is possible that they may have been aware of the innovation being tested and optimized the intervention. The feasibility and effect of using multiple LHWs and supervising personnel needs to be further investigated. In addition, although we collected information on palliative care and behavioral medicine use, we did not measure LHW referrals, and therefore we were unable to assess the effect of the intervention on these intermediary variables. In addition, we did not adjust for multiple comparisons, which could have affected the P values we reported.

    Conclusions

    This LHW-led symptom screening and referral intervention was associated with lower symptom burden, acute care use, and total costs, without survival differences between the intervention group and the control group. At the end of life, the intervention was associated with lower acute care use and total costs and greater hospice use. This intervention may be one approach to improve the value of cancer care delivery in community practices.

    Back to top
    Article Information

    Accepted for Publication: January 26, 2020.

    Published: March 16, 2020. doi:10.1001/jamanetworkopen.2020.1023

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Patel MI et al. JAMA Network Open.

    Corresponding Author: Manali I. Patel, MD, MPH, MS, Division of Oncology, Stanford University School of Medicine, 875 Blake Wilbur Dr, Stanford, CA 94305 (manalip@stanford.edu).

    Author Contributions: Drs Patel and R. Agajanian had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: All authors.

    Acquisition, analysis, or interpretation of data: Patel, R. Agajanian.

    Drafting of the manuscript: Patel.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Patel.

    Obtained funding: Patel, Ramirez.

    Administrative, technical, or material support: Patel, Ramirez, R. Agajanian, H. Agajanian.

    Supervision: All authors.

    Conflict of Interest Disclosures: Dr Patel reported receiving grants from the National Institutes of Health during the conduct of the study. No other disclosures were reported.

    Funding/Support: Research reported in this publication was supported, in part, by the National Institute on Minority Health and Health Disparities of the National Institutes of Health under award K23MD013474.

    Role of the Funder/Sponsor: The National Institute on Minority Health and Health Disparities of the National Institutes of Health under award K23MD013474 supported the interpretation of the data and preparation and review of the manuscript.

    Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

    Meeting Presentation: This study was presented at the Annual Meeting of the American Society of Clinical Oncology; June 1, 2019; Chicago, Illinois; and the Quality Symposium Meeting of the American Society of Clinical Oncology; September 7, 2019; San Diego, California.

    References
    1.
    Basch  E, Deal  AM, Dueck  AC.  Patient-reported symptom monitoring during chemotherapy—reply.  JAMA. 2017;318(19):1935-1936. doi:10.1001/jama.2017.14903PubMedGoogle ScholarCrossref
    2.
    Basch  E, Deal  AM, Dueck  AC,  et al.  Overall survival results of a trial assessing patient-reported outcomes for symptom monitoring during routine cancer treatment.  JAMA. 2017;318(2):197-198. doi:10.1001/jama.2017.7156PubMedGoogle ScholarCrossref
    3.
    Basch  E, Deal  AM, Kris  MG,  et al.  Symptom monitoring with patient-reported outcomes during routine cancer treatment: a randomized controlled trial.  J Clin Oncol. 2016;34(6):557-565. doi:10.1200/JCO.2015.63.0830PubMedGoogle ScholarCrossref
    4.
    Reilly  CM, Bruner  DW, Mitchell  SA,  et al.  A literature synthesis of symptom prevalence and severity in persons receiving active cancer treatment.  Support Care Cancer. 2013;21(6):1525-1550. doi:10.1007/s00520-012-1688-0PubMedGoogle ScholarCrossref
    5.
    van den Beuken-van Everdingen  MH, Hochstenbach  LM, Joosten  EA, Tjan-Heijnen  VC, Janssen  DJ.  Update on prevalence of pain in patients with cancer: systematic review and meta-analysis.  J Pain Symptom Manage. 2016;51(6):1070-1090.e9. doi:10.1016/j.jpainsymman.2015.12.340PubMedGoogle ScholarCrossref
    6.
    Greco  MT, Roberto  A, Corli  O,  et al.  Quality of cancer pain management: an update of a systematic review of undertreatment of patients with cancer.  J Clin Oncol. 2014;32(36):4149-4154. doi:10.1200/JCO.2014.56.0383PubMedGoogle ScholarCrossref
    7.
    Kamal  AH, Nipp  RD, Bull  J, Stinson  CS, Abernethy  AP.  Symptom burden and performance status among community-dwelling patients with serious illness.  J Palliat Med. 2015;18(6):542-544. doi:10.1089/jpm.2014.0381PubMedGoogle ScholarCrossref
    8.
    Deandrea  S, Montanari  M, Moja  L, Apolone  G.  Prevalence of undertreatment in cancer pain: a review of published literature.  Ann Oncol. 2008;19(12):1985-1991. doi:10.1093/annonc/mdn419PubMedGoogle ScholarCrossref
    9.
    Patel  MI, Williams  DC, Wohlforth  C, Fisher  G, Wakelee  HA, Blayney  DW.  Are patients with thoracic malignancies at risk for uncontrolled symptoms?  J Oncol Pract. 2015;11(1):e98-e102. doi:10.1200/JOP.2014.001502PubMedGoogle ScholarCrossref
    10.
    Nipp  RD, El-Jawahri  A, Moran  SM,  et al.  The relationship between physical and psychological symptoms and health care utilization in hospitalized patients with advanced cancer.  Cancer. 2017;123(23):4720-4727. doi:10.1002/cncr.30912PubMedGoogle ScholarCrossref
    11.
    Zaza  C, Baine  N.  Cancer pain and psychosocial factors: a critical review of the literature.  J Pain Symptom Manage. 2002;24(5):526-542. doi:10.1016/S0885-3924(02)00497-9PubMedGoogle ScholarCrossref
    12.
    Barkley  R, Soobader  MJ, Wang  J, Blau  S, Page  RD.  Reducing cancer costs through symptom management and triage pathways.  J Oncol Pract. 2019;15(2):e91-e97. doi:10.1200/JOP.18.00082PubMedGoogle ScholarCrossref
    13.
    Patel  MI, Periyakoil  VS, Blayney  DW,  et al.  Redesigning cancer care delivery: views from patients and caregivers.  J Oncol Pract. 2017;13(4):e291-e302. doi:10.1200/JOP.2016.017327PubMedGoogle ScholarCrossref
    14.
    Patel  MI, Moore  D, Bhattacharya  J, Milstein  A, Coker  TR.  Perspectives of health care payer organizations on cancer care delivery redesign: a national study.  J Oncol Pract. 2019;15(1):e46-e55. doi:10.1200/JOP.18.00331PubMedGoogle ScholarCrossref
    15.
    Patel  MI, Periyakoil  VS, Moore  D, Nevedal  A, Coker  TR.  Delivering end-of-life cancer care: perspectives of providers.  Am J Hosp Palliat Care. 2018;35(3):497-504. doi:10.1177/1049909117719879PubMedGoogle ScholarCrossref
    16.
    Patel  MI, Moore  D, Milstein  A.  Redesigning advanced cancer care delivery: three ways to create higher value cancer care.  J Oncol Pract. 2015;11(4):280-284. doi:10.1200/JOP.2014.001065PubMedGoogle ScholarCrossref
    17.
    Patel  MI, Ramirez  D, Agajanian  R, Agajanian  H, Bhattacharya  J, Bundorf  KM.  Lay health worker-led cancer symptom screening intervention and the effect on patient-reported satisfaction, health status, health care use, and total costs: results from a tri-part collaboration  [published online September 24, 2019].  J Oncol Pract. doi:10.1200/JOP.19.00152PubMedGoogle Scholar
    18.
    Rocque  GB, Pisu  M, Jackson  BE,  et al; Patient Care Connect Group.  Resource use and Medicare costs during lay navigation for geriatric patients with cancer.  JAMA Oncol. 2017;3(6):817-825. doi:10.1001/jamaoncol.2016.6307PubMedGoogle ScholarCrossref
    19.
    Ogrinc  G, Davies  L, Goodman  D, Batalden  P, Davidoff  F, Stevens  D.  SQUIRE 2.0 (Standards for Quality Improvement Reporting Excellence): revised publication guidelines from a detailed consensus process.  BMJ Qual Saf. 2016;25(12):986-992. doi:10.1136/bmjqs-2015-004411PubMedGoogle ScholarCrossref
    20.
    Bruera  E, Kuehn  N, Miller  MJ, Selmser  P, Macmillan  K.  The Edmonton Symptom Assessment System (ESAS): a simple method for the assessment of palliative care patients.  J Palliat Care. 1991;7(2):6-9. doi:10.1177/082585979100700202PubMedGoogle ScholarCrossref
    21.
    Kroenke  K, Spitzer  RL, Williams  JB.  The PHQ-9: validity of a brief depression severity measure.  J Gen Intern Med. 2001;16(9):606-613. doi:10.1046/j.1525-1497.2001.016009606.xPubMedGoogle ScholarCrossref
    22.
    Centers for Medicare & Medicaid Services. Risk adjustment. https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Risk-Adjustors.html. Accessed February 5, 2020.
    23.
    Kahan  BC.  Accounting for centre-effects in multicentre trials with a binary outcome—when, why, and how?  BMC Med Res Methodol. 2014;14:20. doi:10.1186/1471-2288-14-20PubMedGoogle ScholarCrossref
    24.
    STATACorp LLC. New in Stata 16. https://www.stata.com/new-in-stata/. Accessed August 15, 2019.
    25.
    Kroenke  K, Theobald  D, Norton  K,  et al.  The Indiana Cancer Pain and Depression (INCPAD) trial design of a telecare management intervention for cancer-related symptoms and baseline characteristics of study participants.  Gen Hosp Psychiatry. 2009;31(3):240-253. doi:10.1016/j.genhosppsych.2009.01.007PubMedGoogle ScholarCrossref
    26.
    Yount  SE, Rothrock  N, Bass  M,  et al.  A randomized trial of weekly symptom telemonitoring in advanced lung cancer.  J Pain Symptom Manage. 2014;47(6):973-989. doi:10.1016/j.jpainsymman.2013.07.013PubMedGoogle ScholarCrossref
    27.
    French  EB, McCauley  J, Aragon  M,  et al.  End-of-life medical spending in last twelve months of life is lower than previously reported.  Health Aff (Millwood). 2017;36(7):1211-1217. doi:10.1377/hlthaff.2017.0174PubMedGoogle ScholarCrossref
    28.
    Garcia  SF, Wortman  K, Cella  D,  et al.  Implementing electronic health record–integrated screening of patient-reported symptoms and supportive care needs in a comprehensive cancer center.  Cancer. 2019;125(22):4059-4068. doi:10.1002/cncr.32172PubMedGoogle ScholarCrossref
    29.
    Strasser  F, Blum  D, von Moos  R,  et al; Swiss Group for Clinical Cancer Research (SAKK).  The effect of real-time electronic monitoring of patient-reported symptoms and clinical syndromes in outpatient workflow of medical oncologists: E-MOSAIC, a multicenter cluster-randomized phase III study (SAKK 95/06).  Ann Oncol. 2016;27(2):324-332. doi:10.1093/annonc/mdv576PubMedGoogle ScholarCrossref
    30.
    Holch  P, Warrington  L, Bamforth  LCA,  et al.  Development of an integrated electronic platform for patient self-report and management of adverse events during cancer treatment.  Ann Oncol. 2017;28(9):2305-2311. doi:10.1093/annonc/mdx317PubMedGoogle ScholarCrossref
    31.
    Levine  DM, Lipsitz  SR, Linder  JA.  Trends in seniors’ use of digital health technology in the United States, 2011-2014.  JAMA. 2016;316(5):538-540. doi:10.1001/jama.2016.9124PubMedGoogle ScholarCrossref
    32.
    Rocque  GB, Williams  CP, Jones  MI,  et al.  Healthcare utilization, Medicare spending, and sources of patient distress identified during implementation of a lay navigation program for older patients with breast cancer.  Breast Cancer Res Treat. 2018;167(1):215-223. doi:10.1007/s10549-017-4498-8PubMedGoogle ScholarCrossref
    33.
    Patel  MI, Sundaram  V, Desai  M,  et al.  Effect of a lay health worker intervention on goals-of-care documentation and on health care use, costs, and satisfaction among patients with cancer: a randomized clinical trial.  JAMA Oncol. 2018;4(10):1359-1366. doi:10.1001/jamaoncol.2018.2446PubMedGoogle ScholarCrossref
    34.
    Patel  MI, Moore  D, Coker  TR.  End-of-life cancer care redesign: patient and caregiver experiences in a lay health worker-led intervention.  Am J Hosp Palliat Care. 2019;36(12):1081-1088. doi:10.1177/1049909119847967PubMedGoogle ScholarCrossref
    ×