Predictors of cancer rehabilitation medicine referral and utilization among cancer survivors

doi:10.21203/rs.3.rs-2129841/v1

Download PDF

Research Article

Predictors of cancer rehabilitation medicine referral and utilization among cancer survivors

https://doi.org/10.21203/rs.3.rs-2129841/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 16 Mar, 2023

Read the published version in Supportive Care in Cancer →

You are reading this latest preprint version

Purpose

Cancer survivors experience high rates of physical inactivity that often go unaddressed. The My Wellness Check program (MWC) is an EHR-integrated screening and referral system that includes surveillance of physical activity and triage to cancer rehabilitation medicine services. This study examined assessment of physical activity and subsequent referrals to cancer rehabilitation medicine.

Methods

A secondary analysis was performed for survivors who completed the MWC between April 2021 and January 2022. Univariable and multivariable logistic regression modeled determinants of qualification for a physical activity referral and provider completion of referral to cancer rehabilitation medicine. Referral was based on the Moving through Cancer questionnaire. Adjusted odds ratios (aOR) and corresponding 95% confidence intervals (95% CI) were calculated.

Results

There were 1,174 survivors who completed the physical activity survey, of which 46% (n = 540) qualified for a physical activity referral. After controlling for group differences, qualification for a referral included moderate-severe physical dysfunction (aOR: 1.815; 95% CI: 1.177, 2.798), and Hispanic or Latino ethnicity (aOR: 0.744; CI: 0.574, 0.963). Only 31% (n = 168) received a completed physician referral to cancer rehabilitation medicine following screening. No patient-level factors were significantly associated with receiving a completed physician referral. Following referral, 8% (n = 13) utilized cancer rehabilitation medicine services.

Conclusions

Patient-level and clinical factors may predict qualification for physical activity referrals; however, they don’t appear to predict referral completion to cancer rehabilitation medicine services. Future research should focus on potential provider- and organization-level factors that interact and influence referral to and utilization of cancer rehabilitation medicine services.

patient-reported outcomes

cancer rehabilitation

oncology

supportive care

symptom monitoring

physical activity

Adults with a history of cancer are at risk of short-term, long-term, and late-onset physical and psychosocial impairments [1] that can limit functioning in life roles [2] and reduce quality of life [3]. Engagement in physical activity and rehabilitation services have the potential to improve individuals' functioning from cancer diagnosis through survivorship [4, 5]. Historically, assessment of these physical needs and use of rehabilitation services to improve physical activity remains limited in ambulatory oncology settings [6]. Barriers to the timely receipt of cancer rehabilitation services and exercise prescription exist at the level of the patient, provider, system, payer, and policy. Missed opportunities to prevent and manage cancer-related disability are often due to lags in early identification of functional decline, limited discussion surrounding physical activity, or lack of timely referral to cancer rehabilitation [1, 7–9]. At the patient-level, cancer survivors may have difficulty articulating or identifying pre-clinical disablement or physical inactivity [8]. Likewise, healthcare providers share reduced confidence in exercise prescription and/or may not prioritize education about physical activity. Furthermore, electronic health record (EHR) systems often lack routine screening and surveillance of symptoms from diagnosis through survivorship [10], as well as algorithms to improve referral efficiency [11, 12]. A culmination of barriers at each level leads to under-reported symptoms, deteriorating function, and greater burden on health care systems such as higher rates of hospitalization [13].

The My Wellness Check program was created in part to overcome the multi-level barriers to access and delivery of quality cancer survivorship care in an urban academic medical center. My Wellness Check integrates patient-reported outcomes (PROs) in EHR systems to routinely assess physical and emotional symptoms, and supportive care and practical needs of cancer survivors from diagnosis through survivorship [14, 15]. Delivered in English and Spanish languages, it is considered the first ambulatory oncology program to capture EHR-integrated PROs such as PROMIS® computerized adaptive tests (CATs) [16] linked with supportive care referrals for Spanish-speaking Hispanics/Latinos [15].

One aspect of this comprehensive EHR-integrated screening and referral system includes surveillance of physical activity and subsequent referral to cancer rehabilitation medicine services. Cancer rehabilitation medicine is defined as evaluation or treatment of functional decline and physical impairments by a licensed physiatrist with expertise in oncology. Engagement in physical activity is a known risk factor for cancer development, cancer-related disability, and other adverse cancer-related outcomes [5]. Appropriate exercise or rehabilitation interventions to improve physical activity must consider the survivors’ functional level to be safe and effective [17]. The My Wellness Check program 1) identifies physical inactivity-related risk based on a two-item questionnaire [18, 19], 2) alerts the oncology care team of physical inactivity when opening a patient chart via a best practice alert (BPA), and 3) guides the clinician to select the appropriate intervention (exercise education or referral to cancer rehabilitation medicine) (Fig. 1). After receiving the BPA, clinicians are presented with a decision tree to facilitate selection of an appropriate physical activity intervention. Intervention selection is based on the survivors’ current functional status using the Eastern Cooperative Oncology Group Performance Status Scale (ECOG) [20]. If the cancer survivor is ambulatory and capable of all self-care tasks (ECOG Scores of 0–2), the provider is encouraged to provide educational materials to promote physical activity and exercise. If the cancer survivor displays difficulty with ambulation and needs assistance withs self-care tasks (ECOG Score ≥ 3), the provider is prompted to refer the cancer survivor to cancer rehabilitation medicine for exercise medical clearance and support. To date, evaluation of this screening-referral pathway has not been examined to ensure appropriate access and delivery of cancer rehabilitation medicine services.

The purpose of this study was three-fold. First, we described and analyzed the demographic, clinical characteristics, and PROs of ambulatory oncology patients who do and do not qualify for referral to cancer rehabilitation medicine services in the My Wellness Check program. Furthermore, we examined predictive factors for cancer rehabilitation medicine referral beyond physical activity and ECOG performance status. Second, to promote an understanding of which cancer survivors receive a completed cancer rehabilitation medicine referral, we examined patient-level factors associated with a completed physician referral. Lastly, we evaluated the prevalence of cancer rehabilitation medicine utilization following referral. This analysis provides important insights into access and delivery of cancer rehabilitation medicine services, including recommendations to strengthen referral pathways and system-level processes.

Program Description

My Wellness Check is applicable for all patients with an International Classification of Diseases, Tenth Revision cancer diagnosis and a scheduled second ambulatory oncology appointment. Patients were contacted through a preferred method of communication (I.e., text, e-mail, or the patient portal) 72 hours before their appointment and asked to complete the My Wellness Check assessments. Patients completed responses to multiple self-repot questionnaires. Patient-Reported Outcomes Measurement Information System (PROMIS®) computer adaptive tests (CATs) captured symptom severity of depression, anxiety, fatigue, pain interference, and physical function in their preferred language [16]. PROMIS® CATs were treated as dichotomous variable based on the patient reporting moderate-severe symptom severity or none-mild symptom severity. Thresholds for moderate-severe severity include T-scores at a ≥ 70 for pain interference and fatigue, ≥ 65 for anxiety, ≥ 60 for depression, and ≤ 30 for physical function [21]. Quality of life was captured using the Functional Assessment of Cancer Therapy – General 7 [22]. The Moving through Cancer questionnaire two-item questionnaire [18, 19] captured physical activity levels through the following yes-no formatted questions: During the past week have you performed three or more days of physical activity where your heart beats faster and your breathing is harder than normal for 30 minutes or more?; and 2) During the past week have you performed two or more days of physical activity to increase muscles strength, such as lighting weights? If a survivor responded “no” to both questions, a BPA was sent to the medical team via electronic health record system, denoting the survivor is not meeting recommended physical activity requirements [5, 18]. After the patient triggered the physical activity BPA, the clinician could respond by either providing a referral to cancer rehabilitation medicine services or temporarily dismissing the alert (Fig. 1).

Study Design

This is a secondary data analysis of cancer survivors who participated in the My Wellness Check program as part of their ambulatory oncology care between April 15, 2021 thru January 31, 2022. Details of the survey design, recruitment procedures, and implementation of the program were published in earlier work [15, 23, 24]. Ethical approval of this study was provided by the University of Miami Institutional Review Board (Protocol: 20211220; PI: Penedo).

Participants and Procedures

The study sample was composed of individuals who met the following criteria: 1) Partial or full completion of at least one My Wellness Check program PRO assessment battery; and 2) Complete data responses on the 2-item physical activity questionnaire between April 2021 and January 2022. Complete data on the two-item physical activity survey was required as the combination of responses triggered the specific physical activity in-basket alert for potential cancer rehabilitation medicine intervention. For participants who had a record of more than one survey submission during the course of their cancer care, the first survey submission was selected for analysis.

Potential Determinants of Rehabilitation Medicine Referral

A combination of sociodemographic, clinical characteristics, and PROs were extracted as explanatory variables. Self-reported sociodemographic information extracted from EHR includes age, sex, race, ethnicity, preferred language, marital status the NCI Comorbidity Index [21, 22], and insurance coverage Clinical characteristics of survivors included initial cancer diagnosis, time since diagnosis calculated in years, and history of cancer-related treatments. Initial cancer diagnosis was categorized in two ways: 1) solid tumor or hematologic; and 2) cancer type (breast, colorectal, genitourinary, gynecological, head and neck, hematologic, lung, melanoma, prostate, stomach and esophagus, thyroid, other). Patient reported outcomes of survivors accounted for cancer-related pain, fatigue, physical function, depression, anxiety, and quality of life.

Outcome Variables

Each study objective had a different outcome of interest (Fig. 1). The first outcome of interest was qualification for a physical activity referral. For the second objective, the outcome of interest was cancer rehabilitation medicine referral completion. Lastly, we sought to see the number of individuals that utilized cancer rehabilitation services following physician referral. Utilization of cancer rehabilitation medicine services was a dichotomous variable based on whether the patient received a cancer rehabilitation medicine consult, evaluation, or treatment at Sylvester Comprehensive Cancer Center following physician referral through My Wellness Check program.

Data Analysis

Data for the My Wellness Check program was extracted from the electronic data warehouse at the University of Miami Miller School of Medicine. Service utilization data was based on national provider identifiers and medical claims associated with the two cancer physiatrists affiliated with the health care system. For each objective, demographic and clinical characteristics were analyzed by t-test and chi-squared test, or non-parametric alternative as appropriate. Logistic regression models were developed using forward stepwise selection techniques based on the Akaike Information Criterion to examine determinants of both outcomes of interest with prespecified level of significance for univariate models (p ≤ .30) and multivariate (p ≤ .05). Furthermore, the multivariable model was then tested to evaluate regression estimates while controlling for significant groups differences based on demographic, clinical characteristics, or PROs. The odds ratio (OR) with 95% confidence intervals (95% CI) were computed as well. All statistical analyses were performed using SAS version 9.4 statistical software (SAS Institute, Cary, NC).

Between April 15, 2021 and January 31, 2022, there were 1,461 cancer survivors who completed at least one MWC survey in either Spanish or English. Due to incomplete data on the 2-item physical activity questionnaire, 287 survivors were excluded. The final sample included 1,174 individuals (Table 1). More than half of the sample was composed of individuals identifying as male (59%, n = 695) or who were greater than or equal to 65 years old (57%, n = 668). The sample was predominantly White (89%, n = 1,044) and 35% of the sample was Hispanic or Latino (n = 413). Patients primarily had a solid tumor as an initial cancer diagnosis (79%, n = 929) inclusive but not limited to lung (17%, n = 207), prostate (15%, n = 178), head and neck (12%, n = 143), and colorectal (9%, n = 99). Seventeen percent (n = 203) of the sample had an initial diagnosis of hematologic origin. Most patients received chemotherapy (58%, n = 679) as a part of their cancer-related treatment. Patients received healthcare coverage through commercial or non-governmental insurance plans (62%, n = 728), public or government-assisted programs (36%, n = 425, or self-pay (2%, n = 21). Overall, the sample had an average NCI Comorbidity Index Score of 1.7 (Standard Deviation (SD) = 2.0).

Table 1

Patient characteristics of individuals who completed the physical activity questionnaire in My Wellness Check program
Variable	Total Sample (n = 1174)	No Physical Activity Alert (n = 634, 54%)	Trigger Physical Activity Alert (n = 540, 46%)	p-value
Sex, Male, n (%)	695 (59)	397 (63)	298 (45)	0.0017
Age (years), n (%) < 65 years ≥ 65 years Mean (SD) Median (Minimum, Maximum)	506 (43) 668 (57) 65.8 (12.1) 67 (20–96)	272 (43) 362 (57) 65.6 (11.9) 67 (20–93)	234 (43) 306 (57) 66.0 (12.3) 67 (21–96)	0.8818
Race, n (%) White Black or African American Asian or Other	1044 (89) 112 (10) 18 (1)	567 (89) 60 (10) 7 (1)	477 (88) 52 (10) 11 (2)	0.4267
Ethnicity, n (%) Hispanic or Latino Non-Hispanic or Latino Missing	413 (35) 714 (61) 47 (4)	240 (38) 368 (48) 26 (4)	173 (32) 346 (64) 21 (4)	0.0330
Preferred Language, n (%) English Spanish	222 (19) 952 (81)	489 (77) 145 (23)	463 (86) 77 (14)	0.0002
Marital Status, n (%) Single Married or Partnered Divorced, Widowed, or Separated Other/Unknown	173 (15) 792 (68) 197 (17) 12 (1)	89 (14) 425 (67) 116 (18) 4 (1)	84 (16) 367 (68) 81 (15) 8 (1)	0.2172
Insurance, n (%) Commercial or Non-Governmental Public or Governmental Self-Pay	728 (62) 425 (36) 21 (2)	394 (62) 227 (36) 13 (2)	334 (62) 198 (37) 8 (1)	0.8157
Initial Cancer Diagnosis, n (%) Solid Tumor Hematologic Tumor Unknown or Missing	922 (79) 203 (17) 49 (4)	492 (78) 115 (18) 27 (4)	430 (80) 88 (16) 22 (4)	0.6933
Cancer Therapies Received: Radiation, n (%) Chemotherapy, n (%) Surgery, n (%)	449 (38) 679 (58) 554 (47)	240 (38) 351 (55) 293 (46)	209 (39) 328 (60) 261 (48)	0.7655 0.4686 0.0629
PROMIS® Anxiety, Moderate-Severe, n(%)	82 (7)	46 (4)	36 (3)	0.6932
PROMIS® Depression, Moderate-Severe, n(%)	109 (9)	47 (4)	62 (5)	0.0167
PROMIS® Fatigue, Moderate-Severe, n(%)	93 (8)	31 (5)	62 (12)	< .0001
PROMIS® Pain, Moderate-Severe, n(%)	50 (4)	15(1)	35 (3)	0.0005
PROMIS® Physical Function, Moderate-Severe, n(%)	143 (12)	53 (8)	90 (17)	< .0001
Quality of Life	19.2 (5.7)	19.8 (5.6)	18.4 (5.8)	< .0001
Time Since Initial Diagnosis (years)	3.0 (3.13)	3.04 (3.10)	2.84 (3.14)	0.3003
NCI Comorbidity Index	1.7 (2.0)	1.59 (1.97)	1.82 (2.11)	0.0534
Note: Patient-reported outcomes of moderate-severe threshold are based T-scores at ≥ 70 for pain interference and fatigue, ≥ 65 for anxiety, ≥ 60 for depression, and ≤ 30 for physical function. Not all respondents had complete data on PRO data. PROMIS® measures, Time since Initial Diagnosis, and Quality of Life are based on the following sample sizes (n): PROMIS® Anxiety: 1133; PROMIS® Depression: 1133; PROMIS® Fatigue: 1150; PROMIS® Pain: 1139; PROMIS® Physical Function: 1149; Time Since Initial Diagnosis (years): 1157; Quality of Life: 1122 Abbreviations: PROMIS®: Patient-Reported Outcomes Measurement Information System; NCI: National Cancer Institute; SD: standard deviation

Physical Activity Alerts

Based on responses to the physical activity questionnaire, 540 (46%) triggered a physical activity alert (Table 1). There were significant differences between the group of individuals reported adequate physical activity and those did not. Patients who triggered an alert had a higher proportion of individuals reporting moderate-severe depressive symptoms (p = 0.0167), moderate-severe pain (p = 0.005), moderate-severe fatigue (p < 0.001), moderate-severe physical function (p < 0.001), and worse reported health-related quality of life (p < 0.001). The group who did not trigger an alert had a significantly higher proportion of males (p < 0.0017) and higher proportion of individuals who preferred Spanish as a primary language (p = 0.0002).

Univariate analysis of baseline characteristics associated with qualifying for a physical activity referral are found in Table 2. Sex, depressive symptoms, fatigue, pain interference, physical function, and quality of life were significantly associated with qualifying for a physical activity referral. After adjusting for group differences, the model revealed two factors that were independently associated with qualifying for a physical activity alert: physical function and ethnicity (Table 2). Individuals who reported moderate-severe physical function were 75% more likely to trigger a physical activity alert (Adjusted odds ratio [aOR]: 1.75; 95% CI: 1.137, 2.693). Individuals who identified as Hispanic or Latino were approximately 78% less likely to report physical inactivity (aOR = 0.720, 95% CI: 0.556, 0.932) compared to individuals identifying as Non-Hispanic or Latino.

Table 2

Risk factors associated with a physical activity alert among My Wellness Check participants
Unadjusted Univariable Risk Factor	OR (95% CI)	p-value
Sex: Female vs. Male	1.360 (1.077,1.719)	0.0099
Age: ≥65 years vs. <65 years	0.983 (0.779, 1.239)	0.8818
Race:
Black or African American vs. White	1.030 (0.697, 1.523)	0.8814
Asian or Other vs. White	1.868 (0.718, 4.856)	0.1999
Ethnicity: Hispanic or Latino vs. Non-Hispanic or Latino	0.767 (0.600, 0.979)	0.0331
Marital Status:
Married or Partner vs. Single	0.915 (0.658, 1.272)	0.5699
Divorced, Widowed, or Separated vs Single	0.740 (0.490, 1.117)	0.1514
Insurance: Public or Governmental Insurance vs. Private or Non-Governmental Insurance	1.029 (0.809, 1.308)	0.8157
Treatments Received:
Chemotherapy vs. No Chemotherapy	1.247 (0.988, 1.575)	0.0631
Surgery vs. No Surgery	1.089 (0.865, 1.370)	0.4686
Radiation vs. No Radiation	1.037 (0.819, 1.313)	0.7654
PROMIS® Anxiety: Moderate-Severe vs. None-Mild	1.095 (0.697, 1.721)	0.6943
PROMIS® Depression: Moderate-Severe vs. None-Mild	1.620 (1.088, 2.411)	0.0175
PROMIS® Fatigue: Moderate-Severe vs. None-Mild	2.523 (1.613, 3.947)	< 0.001
PROMIS® Pain: Moderate-Severe vs. None-Mild	2.859 (1.544, 5.294)	0.0008
PROMIS® Physical Function: Moderate-Severe vs. None-Mild	2.192 (1.528, 3.146)	< 0.001
FACT-G7 Quality of Life	0.957 (0.938, 0.977)	< .0001
Time Since Diagnosis (years)	0.975 (0.939, 1.012)	0.1791
NCI Comorbidity Index	1.057 (0.999,1.118)	0.0540
Stepwise logistic regression adjusting for group differences	OR (95% CI)	p-value
PROMIS® Fatigue: Moderate-Severe vs. None-Mild	1.619 (0.937, 2.795)	0.0840
PROMIS® Physical Function: Moderate-Severe vs. None-Mild	1.750 (1.137, 2.693)	0.0109
FACT-G7 Quality of Life	0.981 (0.956, 1.007)	0.1515
Ethnicity: Hispanic or Latino vs. Non-Hispanic or Latino	0.720 (0.556, 0.932)	0.0126
PROMIS® Depression: Moderate-Severe vs. None-Mild	1.000 (0.621, 1.610)	0.9989
PROMIS® Pain: Moderate-Severe vs. None-Mild	1.674 (0.810, 3.461)	0.1645
Sex: Female vs. Male	1.232 (0.958, 1.585)	0.1041
Note: Unadjusted univariable logistic regression used all observations except for the following variables due to missing data PROMIS® Anxiety: 1133; PROMIS® Depression: 1133; PROMIS® Fatigue: 1150; PROMIS® Pain: 1139; PROMIS® Physical Function: 1149; Time Since Initial Diagnosis (years): 1157; Quality of Life: 1122. Adjusted multivariable models excluded a total of 192 individuals due to missing data on explanatory variables. Abbreviations: PROMIS®: Patient-Reported Outcomes Measurement Information System; CI: Confidence Interval; NCI: National Cancer Institute; OR: Odds Ratio

Cancer Rehabilitation Referral Completion

Of the 540 individuals that prompted a physical activity alert, only 168 (31%) were provided with a completed physician referral for a cancer rehabilitation medicine consultation. Interestingly, there were no significant differences regarding demographic or clinical characteristics, or PROs between those who received a referral and those who had the physical activity alert cancelled (Table 3).

Table 3

Patient characteristics of individuals who received a completed referral to cancer rehabilitation following physical activity alert in My Wellness Check program
Variable	Total Sample (n = 540)	Referral Cancelled (n = 372)	Referral Completed (n = 168)	p-value
Sex, Male, n (%)	298 (55)	205 (55)	93 (55)	0.9569
Age (years), n (%) < 65 years ≥ 65 years Mean (SD) Median (Minimum, Maximum)	234 (43) 306 (57) 66.0 (12.3) 67 (21–96)	151 (41) 221 (59) 66.1 (11.9) 68 (21–90)	83 (49) 85 (51) 65.8 (13.1) 65 (26–96)	0.0557
Race, n (%) White Black or African American Asian or Other	477 (88) 52 (10) 11 (2)	325 (87) 38 (10) 9 (3)	152 (91) 14 (8) 2 (1)	0.5591
Ethnicity, n (%) Hispanic or Latino Non-Hispanic or Latino Missing	176 (33) 346 (64) 21 (3)	114 (31) 243 (65) 15 (4)	59 (35) 103 (61) 6 (4)	0.3150
Preferred Language, n (%) English Spanish	463 (86) 77 (14)	318 (85) 54 (15)	145 (86) 23 (14)	0.7995
Marital Status, n (%) Single Married or Partnered Divorced, Widowed, or Separated Other/Unknown	84 (16) 367 (68) 81 (15) 8 (1)	57 (15) 252 (68) 59 (16) 4 (1)	27 (16) 115 (69) 22 (13) 4 (2)	0.5782
Insurance, n (%) Commercial or Non-Governmental Public or Governmental Self-Pay	334 (62) 198 (37 8 (1)	222 (60) 143 (38) 7 (2)	112 (67) 55 (32) 1 (1)	.1770
Initial Cancer Diagnosis, n (%) Solid Tumor Hematologic Tumor Unknown or Missing	430 (81) 88 (17) 15 (2)	295 (79) 62 (17) 15 (4)	135 (80) 26 (16) 7 (4)	0.9273
Cancer Therapies Received: Radiation, n (%) Chemotherapy, n (%) Surgery, n (%)	209 (39) 328 (61) 261 (48)	152 (41) 228 (61) 175 (47)	57 (33) 100 (59) 86 (51)	0.1258 0.6972 0.3719
PROMIS® Anxiety, Moderate-Severe, n (%)	36 (7)	27 (7)	9 (5)	0.4123
PROMIS® Depression, Moderate-Severe, n (%)	62 (12)	47 (12)	15 (9)	0.2111
PROMIS® Fatigue, Moderate-Severe, n (%)	62 (12)	42 (11)	20 (12)	0.8357
PROMIS® Pain, Moderate-Severe, n (%)	35 (7)	27 (7)	8 (5)	0.2754
PROMIS® Physical Function, Moderate-Severe, n (%)	90 (17)	69 (18)	21 (13)	0.0808
Quality of Life	18.4 (5.8)	18.2 (5.8)	19.0 (5.8)	0.5745
Time Since Initial Diagnosis (years)	2.9 (3.14)	2.9 (3.2)	2.8 (3.1)	0.1552
NCI Comorbidity Index	1.8 (2.1)	1.8 (2.12)	1.8 (2.08)	0.5858
Note: : Patient-reported outcomes of moderate-severe threshold are based T-scores at ≥ 70 for pain interference and fatigue, ≥ 65 for anxiety, ≥ 60 for depression, and ≤ 30 for physical function. Not all respondents had complete data on PRO data. PROMIS® measures, Time since Initial Diagnosis, and Quality of Life are based on the following sample sizes (n): PROMIS® Anxiety: 523; PROMIS® Depression: 521; PROMIS® Fatigue: 531; PROMIS® Pain: 522; PROMIS® Physical Function: 528; Time Since Initial Diagnosis (years): 533; Quality of Life: 517 Abbreviations: PROMIS®: Patient-Reported Outcomes Measurement Information System; NCI: National Cancer Institute; SD: Standard Deviation

In univariable analyses, there were no significant predictors associated with receiving a completed referral to cancer rehabilitation medicine services (Table 4). When conducting the forward stepwise multiple logistic regression with parameters of p < 0.3 for entry and p < 0.05 for factors to remain in the model, there were no factors that significantly predicted completion of referral versus cancellation (Table 4). As an exploratory analysis, we enlarged the stay parameter to p < 0.1. In this model, moderate-severe pain interference (OR:0.446, 95% CI: 0.179, 1.110), older age (OR: 0.705, 95% CI: 0.478, 1.039), and history of radiation therapy (0.737, 95% CI: 0.494, 1.100) provided good model fit (χ² = 0.0089, df = 1, p = 0.9250); however, these factors only trended toward significance (Table 4).

Table 4

Risk factors associated with a completed referral to cancer rehabilitation medicine following physical activity alert from the My Wellness Check
Unadjusted Univariable Risk Factor	OR (95% CI)	p-value
Sex: Female vs. Male	0.990 (0.686, 1.428)	0.9569
Age: ≥65 years vs. <65 years	0.7 (0.485, 1.009)	0.0562
Race:
Black or African American vs. White	0.562 (0.129, 2.438)	0.5034
Asian or Other vs. White	0.804 (0.424, 1.524)	0.4413
Ethnicity: Hispanic or Latino vs. Non-Hispanic or Latino	1.221 (0.827, 1.803)	0.3153
Marital Status:
Married or Partner vs. Single	0.787 (0.403, 1.539)	0.4842
Divorced, Widowed, or Separated vs Single	0.963 (0.403, 1.539)	0.8857
Insurance: Public or Governmental Insurance vs. Private or Non-Governmental Insurance	0.762 (0.519, 1.121)	0.1673
Treatments Received:
Chemotherapy vs. No Chemotherapy	0.929 (0.640, 1.347)	0.6972
Surgery vs. No Surgery	1.181 (0.820, 1.700)	0.3721
Radiation vs. No Radiation	0.743 (0.508, 1.087)	0.1263
PROMIS® Anxiety: Moderate-Severe vs. None-Mild	1.382 (0.635, 3.008)	0.4142
PROMIS® Depression: Moderate-Severe vs. None-Mild	0.678 (0.368, 1.251)	0.2135
PROMIS® Fatigue: Moderate-Severe vs. None-Mild	1.062 (0.602, 1.871)	0.8385
PROMIS® Pain: Moderate-Severe vs. None-Mild	0.639 (0.284, 1.438)	0.2792
PROMIS® Physical Function: Moderate-Severe vs. None-Mild	0.627 (0.371, 1.063)	0.0829
FACT-G7 Quality of Life	1.024 (0.991, 1.057)	0.1553
Time Since Diagnosis (years)	0.995 (0.938, 1.055)	0.8699
NCI Comorbidity Index	0.976 (0.892, 1.065)	0.5852
Stepwise logistic regression adjusting for group differences	OR (95% CI)	p-value
PROMIS® Pain: Moderate-Severe vs. None-Mild	0.446 (0.179, 1.110)	0.0828
Age: ≥65 years vs. <65 years old	0.705 (0.478, 1.039)	0.0772
Radiation vs. No Radiation	0.737 (0.494, 1.100)	0.1354
Note: Adjusted multivariable model excluded a total of 59 individuals due to missing data on explanatory variables. Abbreviations: PROMIS®: Patient-Reported Outcomes Measurement Information System; CI: Confidence Interval; NCI: National Cancer Institute; OR: Odds Ratio

Utilization of Cancer Rehabilitation Medicine Following Referral. Of the 168 individuals received a physician referral, 13 (8%) utilized cancer rehabilitation medicine services. An additional six individuals had been scheduled for an evaluation appointments, however, five cancelled the appointment and did not reschedule and one individual did not show to the appointment.

There has been a call to action to test models of care that enhance access to survivorship services, such as cancer rehabilitation medicine [25, 26]. The MWC program was responsive to this call by creating an EHR-integrated screening and referral system to facilitate patient-provider conversations about physical activity, as well as generate cancer rehabilitation medicine referrals. Our findings revealed that nearly half of cancer survivors did not achieve the recommended guidelines for physical activity and were significantly likely to report moderate-to-severe fatigue and physical dysfunction. Furthermore, only 31% of individuals received a completed referral from a provider to cancer rehabilitation medicine, yet there were no significant determinants that predicted between those who did and did not receive a completed referral. These findings present an opportunity to refine existing pathways and overcome barriers to rehabilitation access and utilization.

Individuals reporting moderate-to-severe fatigue or physical function were significantly associated with increased odds of qualifying for a physical activity referral. Fatigue and physical function represent two modifiable risk factors that are amenable to cancer rehabilitation medicine interventions [27–29]. Fatigue and physical function tend to have a bidirectional as well as multiplicative influence on physical activity. Individuals with this combination of multiple moderate-severe symptoms will likely experience significant limitations in physical activity compared to those with fewer or mild symptoms, and vice versa [30]. Based on the current MWC system, these symptoms trigger multiple intervention options. However, complex medical schedules and management of multiple interventions can result in ineffective management, poor adherence to interventions, and limited patient satisfaction. Given and colleagues (2007) recommend that the most efficient and effective way to manage symptoms clusters is by offering a multi-purpose interventions that yields the greatest benefit [30]. Cancer rehabilitation medicine services present an opportunity to efficiently address these co-occurring symptoms. Future pathway refinement and research may consider testing cancer rehabilitation medicine as the unified solution for individuals who trigger alerts for physical function, fatigue, and physical activity simultaneously, as well explore this pathway’s influence on access and delivery of services.

Our findings also identified that individuals identifying as Hispanic or Latino were approximately 25% less likely to qualify for a physical activity referral even after adjusting for key covariates. Features of Hispanic or Latino ethnicity and healthcare culture may influence responsivity to the physical activity survey questions [31]. Individuals with intersectional identities may experience barriers to cancer care that adversely impact screening.[32] Additional stakeholder engagement may reveal multi-level factors that underlie these findings.

Less than one-third of individuals who qualified for cancer rehabilitation medicine services received a completed referral. However, there were no patient-level factors that significantly predicted the differences in referral. Access to services stems beyond patient-level factors alone. Provider- or organization-level barriers may be the root of differences between those who did and did not receive a referral. As EHR systems are used to optimize patient care, providers may experience “alert fatigue” associated with desensitization to notifications, complex workloads, and challenges distinguishing between informative and uninformative alerts. Likewise, referring providers may continue to be unaware of the benefits of physical activity, cancer rehabilitation medicine, and their role in survivorship care [7, 8, 14, 33]. Use of focus groups to engage key stakeholders associated with the referral process, such as physicians and advanced practitioners, may provide greater insight into the decision-making process. Application of qualitative methods may also reveal organizational processes that influence referral [34, 35] including the way in which alerts are delivered, strategies to improve collaboration between cancer rehabilitation medicine and standard cancer care delivery, and barriers to appointment scheduling.

Lastly, our findings revealed that only 8% of individuals who received a referral utilized cancer rehabilitation medicine services. These results point to the need for expanded efforts to better understand utilization and delivery of cancer rehabilitation medicine services. Recent evidence suggests that only 2%[36] − 10%[2, 6] utilize cancer rehabilitation services, inclusive of physical therapy, occupational therapy, speech language pathology, etc. Although our findings report similar levels of utilization for individuals seeking care of physical therapy or occupational therapy, data were specific to cancer rehabilitation delivered by a physiatrist only. Future research should evaluate how the PROs captured in the MWC program influence utilization of physical therapy and occupational therapy. Access to these services may have higher rates of utilization following MWC program alerts given widespread accessibility within community-based outpatient settings, additional provider availability, patient familiarity, and treatments for mild impairments or activity limitations. Initial referral to non-pharmacological intervention methods of physical therapy or occupational therapy may also be prioritized before a provider considers use of cancer physiatry as a specialty service. Examination of survivors’ decision-making processes to access cancer rehabilitation medicine services is warranted.

The results from this study made evident the value of modifiable risk factors, captured by PROs, to justify need for cancer rehabilitation services. It also highlighted the potential role of provider- and organization-level factors that may influence referral and utilization. However, study limitations may influence our findings. This study analyzed referral and utilization data at a single cancer center that may not detect delivery of services at other settings. Analyses did not include education or stage of disease as these variables had greater than 50% missing data in the original sample. Education might provide insight into socioeconomic factors that may be associated with referral and/or utilization. Future studies may also consider incorporating stage of disease into the analysis as need for and interest in rehabilitation increases with the presence of metastasis or advanced cancer stage [36, 37]. Discrepancies in referral completion may also be due to most patients having a ECOG scores ≤ 2. We were unable to evaluate the influence of ECOG scores on referral completion and utilization statistics. Future work should consider the interaction of physical activity and ECOG performance status on uptake of rehabilitation interventions. Univariable and multivariable modeling incorporated fewer observations than total sample size for each objective due to missing data across demographic, clinical characteristics, and patient-reported outcomes. For all analyses, missing data was less than 1–4% on any given variable and did not follow a pattern. Lastly, utilization rates do not capture patients who may have received services outside of the medical institution.

To the best of our knowledge, this is the first study to examine access to and utilization of cancer rehabilitation medicine services using EHR-integrated PRO assessments. There is preliminary evidence that physical activity screenings can assist in triaging patients to cancer rehabilitation medicine services. Findings revealed that provider- and system-level barriers may influence discussion about physical activity as well as referral to and utilization of cancer rehabilitation medicine services. Future work should examine provide decision-making processes, response to multiple symptom alerts, and system-level changes that more efficiently triage patients to supportive care services.

Funding: This study was funded in part by grant P30 CA240139 (Nimer, S: PI; Penedo, F: Cancer Control Co-Leader).

Conflicts of Interest/Competing Interests: Dr. Frank J Penedo is a paid consultant for Blue Note Therapeutics. All other authors declares no conflicts of interests.

Availability of Data: Deidentified data from this study are not available in a public archive. Deidentified data from this study will be made available (as allowable according to institutional IRB standards) by e-mailing the corresponding author.

Code Availability: Availability of the coding is available upon request to the corresponding author.

Ethics Approval: This study was approved by the University of Miami Institutional Review Board (IRB#20211220) and certifies that the study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments.

Consent to Participate:

This study was part of a performance improvement project within standard of care; therefore and was deemed exempt from institutional review board review, and informed consent was waived.

Consent for Publication: Not applicable

Author Contributions: Study concepts and design were developed by all authors. Data acquisition and assembly was conducted by RB, AN, and FP. Quality control of data was completed by RB and AN. Data analysis was completed by RB, AN, and TKS. Data interpretation was completed by all authors. Manuscript writing and accountability of work was completed by all authors.

Alfano CM, Kent EE, Padgett LS, Grimes M, de Moor JS. Making Cancer Rehabilitation Services Work for Cancer Patients: Recommendations for Research and Practice to Improve Employment Outcomes. PM&R. 2017;9(9, Supplement 2):S398-S406. doi: https://doi.org/10.1016/j.pmrj.2017.06.019.
Brick R, Bender C, Skidmore E. Impact of cancer and cancer-related treatments on participation restrictions. British Journal of Occupational Therapy. 2020:0308022620923858. doi: 10.1177/0308022620923858.
Newman RM, Alfano CM, Radomski MV, Pergolotti M, Wolf TJ, Sleight AG, et al. Catalyzing Research to Optimize Cancer Survivors' Participation in Work and Life Roles. OTJR (Thorofare N J). 2019:1539449219844749. doi: 10.1177/1539449219844749.
Alfano CM, Leach CR, Smith TG, Miller KD, Alcaraz KI, Cannady RS, et al. Equitably improving outcomes for cancer survivors and supporting caregivers: A blueprint for care delivery, research, education, and policy. CA Cancer J Clin. 2019;69(1):35–49. doi: 10.3322/caac.21548.
Ligibel JA, Bohlke K, May AM, Clinton SK, Demark-Wahnefried W, Gilchrist SC, et al. Exercise, Diet, and Weight Management During Cancer Treatment: ASCO Guideline. Journal of Clinical Oncology. 2022:JCO.22.00687. doi: 10.1200/JCO.22.00687.
Pergolotti M, Deal AM, Lavery J, Reeve BB, Muss HB. The prevalence of potentially modifiable functional deficits and the subsequent use of occupational and physical therapy by older adults with cancer. J Geriatr Oncol. 2015;6(3):194–201. doi: 10.1016/j.jgo.2015.01.004.
Cheville AL, Mustian K, Winters-Stone K, Zucker DS, Gamble GL, Alfano CMJPM, et al. Cancer rehabilitation: an overview of current need, delivery models, and levels of care. 2017;28(1):1–17.
Brick R, Lyons KD, Bender C, Eilers R, Ferguson R, Pergolotti M, et al. Factors influencing utilization of cancer rehabilitation services among older breast cancer survivors in the USA: a qualitative study. Supportive Care in Cancer. 2021. doi: 10.1007/s00520-021-06678-9.
Alfano CM, Cheville AL, Mustian K. Developing High-Quality Cancer Rehabilitation Programs: A Timely Need. Am Soc Clin Oncol Educ Book. 2016;35:241–9. doi: 10.14694/EDBK_1561641200/EDBK_156164.
Alfano Catherine M, Oeffinger K, Sanft T, Tortorella B. Engaging TEAM Medicine in Patient Care: Redefining Cancer Survivorship From Diagnosis. American Society of Clinical Oncology Educational Book. 2022(42):921–31. doi: 10.1200/EDBK_349391.
Alfano CM, Pergolotti M. Next-Generation Cancer Rehabilitation: A Giant Step Forward for Patient Care. Rehabil Nurs. 2018;43(4):186–94. doi: 10.1097/rnj.0000000000000174.
Stout NL, Silver JK, Raj VS, Rowland J, Gerber L, Cheville A, et al. Toward a National Initiative in Cancer Rehabilitation: Recommendations From a Subject Matter Expert Group. Arch Phys Med Rehab. 2016;97(11):2006–15. doi: 10.1016/j.apmr.2016.05.002.
Panattoni L, Fedorenko C, Greenwood-Hickman MA, Kreizenbeck K, Walker JR, Martins R, et al. Characterizing Potentially Preventable Cancer- and Chronic Disease–Related Emergency Department Use in the Year After Treatment Initiation: A Regional Study. Journal of Oncology Practice. 2018;14(3):e176-e85. doi: 10.1200/JOP.2017.028191.
Pergolotti M, Alfano CM, Cernich AN, Yabroff KR, Manning PR, de Moor JS, et al. A health services research agenda to fully integrate cancer rehabilitation into oncology care. 2019;125(22):3908–16.
Penedo FJ, Medina HN, Moreno PI, Sookdeo V, Natori A, Boland C, et al. Implementation and Feasibility of an Electronic Health Record–Integrated Patient-Reported Outcomes Symptom and Needs Monitoring Pilot in Ambulatory Oncology. JCO Oncology Practice. 2022:OP.21.00706. doi: 10.1200/OP.21.00706.
Cella D, Choi S, Garcia S, Cook KF, Rosenbloom S, Lai J-S, et al. Setting standards for severity of common symptoms in oncology using the PROMIS item banks and expert judgment. Quality of Life Research. 2014;23(10):2651–61. doi: 10.1007/s11136-014-0732-6.
Covington KR, Marshall T, Campbell G, Williams GR, Fu JB, Kendig TD, et al. Development of the Exercise in Cancer Evaluation and Decision Support (EXCEEDS) algorithm. Supportive Care in Cancer. 2021;29(11):6469–80. doi: 10.1007/s00520-021-06208-7.
Schmitz KH, Stout NL, Maitin-Shepard M, Campbell A, Schwartz AL, Grimmett C, et al. Moving through cancer: Setting the agenda to make exercise standard in oncology practice. Cancer-Am Cancer Soc. 2021;127(3):476–84. doi: https://doi.org/10.1002/cncr.33245.
Schmitz KH, Campbell AM, Stuiver MM, Pinto BM, Schwartz AL, Morris GS, et al. Exercise is medicine in oncology: Engaging clinicians to help patients move through cancer. CA: A Cancer Journal for Clinicians. 2019;69(6):468–84. doi: https://doi.org/10.3322/caac.21579.
Oken MM, Creech RH, Tormey DC, Horton J, Davis TE, McFadden ET, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. American Journal of Clinical Oncology. 1982;5(6).
Garcia SF, Wortman K, Cella D, Wagner LI, Bass M, Kircher S, et al. Implementing electronic health record–integrated screening of patient-reported symptoms and supportive care needs in a comprehensive cancer center. Cancer-Am Cancer Soc. 2019;125(22):4059–68. doi: https://doi.org/10.1002/cncr.32172.
Cella DF, Tulsky DS, Gray G, Sarafian B, Linn E, Bonomi A, et al. The Functional Assessment of Cancer Therapy scale: development and validation of the general measure. J Clin Oncol. 1993;11(3):570–9.
Natori A, Sookdeo VD, Koru-Sengul T, Schlumbrecht M, Calfa C, MacIntyre J, et al. Predictors of adherence to patient reported outcomes and psychosocial needs questionnaire in a culturally diverse ambulatory oncology setting: The My Wellness Check Program. Journal of Clinical Oncology. 2021;39(28_suppl):173-. doi: 10.1200/JCO.2020.39.28_suppl.173.
Penedo FJ, Oswald LB, Kronenfeld JP, Garcia SF, Cella D, Yanez B. The increasing value of eHealth in the delivery of patient-centred cancer care. The Lancet Oncology. 2020;21(5):e240-e51. doi: https://doi.org/10.1016/S1470-2045(20)30021-8.
Hewitt M, Greenfield S, Stovall E. From cancer patient to cancer survivor: Lost in translation. Washington D.C.: National Academies Press; 2006.
Alfano CM, Zucker DS, Pergolotti M, Ness KK, Jones LW, Price ND, et al. A Precision Medicine Approach to Improve Cancer Rehabilitation’s Impact and Integration with Cancer Care and Optimize Patient Wellness. Current Physical Medicine and Rehabilitation Reports. 2017;5(1):64–73. doi: 10.1007/s40141-017-0145-x.
Hunter EG, Gibson RW, Arbesman M, D’Amico M. Systematic Review of Occupational Therapy and Adult Cancer Rehabilitation: Part 1. Impact of Physical Activity and Symptom Management Interventions. American Journal of Occupational Therapy. 2017;71(2):7102100030p1-p11. doi: 10.5014/ajot.2017.023564%J American Journal of Occupational Therapy.
Harrington SE, Stout NL, Hile E, Fisher MI, Eden M, Marchese V, et al. Cancer Rehabilitation Publications (2008–2018) With a Focus on Physical Function: A Scoping Review. Physical Therapy. 2020;100(3):363–415. doi: 10.1093/ptj/pzz184.
Sleight A, Gerber LH, Marshall TF, Livinski A, Alfano CM, Harrington S, et al. Systematic Review of Functional Outcomes in Cancer Rehabilitation. Arch Phys Med Rehab. 2022. doi: https://doi.org/10.1016/j.apmr.2022.01.142.
Given BA, Given CW, Sikorskii A, Hadar N. Symptom Clusters and Physical Function for Patients Receiving Chemotherapy. Seminars in Oncology Nursing. 2007;23(2):121–6. doi: https://doi.org/10.1016/j.soncn.2007.01.005.
Yanez B, McGinty HL, Buitrago D, Ramirez AG, Penedo FJ. Cancer outcomes in Hispanics/Latinos in the United States: An integrative review and conceptual model of determinants of health. Educational Publishing Foundation; 2016. p. 114–29.
Kelly-Brown J, Palmer Kelly E, Obeng-Gyasi S, Chen JC, Pawlik TM. Intersectionality in cancer care: A systematic review of current research and future directions. Psycho-Oncology. 2022;31(5):705–16. doi: https://doi.org/10.1002/pon.5890.
Cheville AL, Kornblith AB, Basford JR. An examination of the causes for the underutilization of rehabilitation services among people with advanced cancer. Am J Phys Med Rehabil. 2011;90(5 Suppl 1):S27-37. doi: 10.1097/PHM.0b013e31820be3be.
Urquhart R, Jackson L, Sargeant J, Porter GA, Grunfeld E. Health System-Level Factors Influence the Implementation of Complex Innovations in Cancer Care. Healthc Policy. 2015;11(2):102–18.
Taplin SH, Anhang Price R, Edwards HM, Foster MK, Breslau ES, Chollette V, et al. Introduction: Understanding and Influencing Multilevel Factors Across the Cancer Care Continuum. JNCI Monographs. 2012;2012(44):2–10. doi: 10.1093/jncimonographs/lgs008.
Cheville AL, Rhudy L, Basford JR, Griffin JM, Flores AM. How Receptive Are Patients With Late Stage Cancer to Rehabilitation Services and What Are the Sources of Their Resistance? Arch Phys Med Rehab. 2017;98(2):203–10. doi: https://doi.org/10.1016/j.apmr.2016.08.459.
Kim YM, Kim D-Y, Chun MH, Jeon J-Y, Yun GJ, Lee MS. Cancer Rehabilitation: Experience, Symptoms, and Needs. jkms. 2011;26(5):619–24. doi: 10.3346/jkms.2011.26.5.619.

Competing interest reported. Dr. Frank J Penedo is a paid consultant for Blue Note Therapeutics. All other authors declares no conflicts of interests.

Download PDF

Journal Publication

published 16 Mar, 2023

Read the published version in Supportive Care in Cancer →

Editorial decision: Major revision
25 Jan, 2023
Reviews received at journal
13 Dec, 2022
Reviewers agreed at journal
21 Nov, 2022
Reviewers invited by journal
27 Oct, 2022
Editor assigned by journal
26 Oct, 2022
Submission checks completed at journal
13 Oct, 2022
First submitted to journal
03 Oct, 2022

You are reading this latest preprint version

Predictors of cancer rehabilitation medicine referral and utilization among cancer survivors

Status:

Journal Publication

Version 1

Abstract

Purpose

Methods

Results

Conclusions

Figures

Background

Methods

Program Description

Study Design

Participants and Procedures

Potential Determinants of Rehabilitation Medicine Referral

Outcome Variables

Data Analysis

Results

Physical Activity Alerts

Cancer Rehabilitation Referral Completion

Discussion

Conclusion

Declarations

References

Additional Declarations

Status:

Journal Publication

Version 1