Randomized Controlled Trial of a Nurse-led Brief Behavioral Intervention for Dyspnea in Patients with Advanced Lung Cancer

Abstract

Purpose:

In patients with lung cancer, dyspnea is one of the most prevalent and disabling symptoms, for which effective treatments are lacking. We examined the efficacy of a nurse-led brief behavioral intervention to improve dyspnea in patients with advanced lung cancer.

Methods:

Patients with advanced lung cancer reporting at least moderate breathlessness (N=247) enrolled in a randomized trial of a nurse-led two-session intervention (focused on breathing techniques, postural positions, and fan therapy) versus usual care. At baseline and weeks 8 (primary endpoint), 16, and 24, participants completed measures of dyspnea (Modified Medical Research Council Dyspnea Scale [mMRCDS]; Cancer Dyspnoea Scale [CDS]), quality of life (Functional Assessment of Cancer Therapy-Lung [FACT-L]), psychological symptoms (Hospital Anxiety & Depression Scale), and activity level (Godin-Shephard Leisure Time Physical Activity Questionnaire). To examine intervention effects, we conducted ANCOVA and longitudinal mixed effects models.

Results:

The sample (Age_mean=66.15 years; 55.9% female) primarily included patients with advanced non-small cell lung cancer (85.4%). Compared to usual care, the intervention improved the primary outcome of patient-reported dyspnea on the mMRCDS (B=−0.33, 95% CI=−0.61, −0.05) but not the CDS total score at 8 weeks. Intervention patients also reported less dyspnea on the CDS “sense of discomfort” subscale (B=−0.59, 95% CI=−1.16, −0.01) and better functional wellbeing per the FACT-L (B=1.39, 95% CI=0.18, 2.59) versus the control group. Study groups did not differ in overall quality of life, psychological symptoms, or activity level at 8 weeks or longitudinally over 24 weeks.

Conclusion:

For patients with advanced lung cancer, a scalable behavioral intervention alleviated the intractable symptom of dyspnea. Further research is needed on ways to enhance intervention effects over the long-term and across additional outcomes.

ClinicalTrials.gov Identifier:

NCT03089125

Introduction

Among patients with advanced lung cancer, over 70% report clinically significant dyspnea,^1–3 which is the subjective experience of distress or discomfort related to the sensation of feeling breathless.⁴ Occurring episodically, upon exertion, or continuously even at rest,^5–7 the symptom often persists or worsens during treatment and especially at the end of life.^8–11 Dyspnea is associated not only with worse physical symptoms, psychological distress, interference with daily activities, and poor quality of life (QOL), but also increased healthcare utilization.^12–23

Despite the morbidity associated with dyspnea, effective treatments for patients with cancer are limited.^24,25 Pharmacologic interventions generally demonstrate poor outcomes. Opioids may offer short-term relief, but long-term efficacy studies show minimal to no benefit, and these medications carry serious risks, with at least 25% of patients discontinuing due to side effects.^24,26–28 Similarly, dyspnea does not improve with corticosteroids or anxiolytics.^{24,27,29–33} Non-pharmacologic interventions, such as supplemental oxygen delivered via high-flow nasal canula, behavioral skills, and fan therapy, have shown some promise, with minimal to no risks.^34–36 Thus, both the American Society of Clinical Oncology and European Society of Medical Oncology have released clinical guidelines for dyspnea management, prioritizing non-pharmacologic interventions as first-line therapy, followed by pharmacologic treatments.^24,37

Even with the accumulating evidence, rigorous randomized controlled trials (RCT) of non-pharmacologic interventions targeting dyspnea in patients with cancer are needed.²⁵ Moreover, patients with advanced lung cancer experience high symptom burden, making trial participation and engagement with supportive care interventions challenging. In response to these barriers, our team conducted a single-group pilot feasibility trial of a two-session behavioral intervention for dyspnea management in this population. To optimize access, oncology nurses delivered the intervention in tandem with other medical visits. Participants reported not only high acceptability, but also significantly improved dyspnea on the primary outcome measure (i.e., Modified Medical Research Council Dyspnea Scale, [mMRCDS]).³⁸

Building on this successful pilot study, we conducted a randomized trial of the nurse-led brief behavioral intervention for dyspnea versus usual care in patients with advanced lung cancer. The primary aim was to test the efficacy of the intervention for improving patient-reported dyspnea, with secondary outcomes targeting QOL, psychological symptoms, and activity level.

Methods

Study Design and Setting

Between 5/2017 and 10/2022, we conducted an RCT to evaluate the efficacy of a nurse-led, two-session behavioral intervention for managing dyspnea in patients with advanced lung cancer reporting at least moderate dyspnea. Participants received their oncology care at the Massachusetts General Hospital (MGH), Mass General/North Shore Cancer Center-Danvers, or Dana-Farber Cancer Institute (DFCI). The Dana-Farber/Harvard Cancer Center IRB approved the study protocol.

Eligibility Criteria

Participant eligibility criteria included: (a) a diagnosis of advanced lung cancer (i.e., non-small cell, small cell) or mesothelioma not being treated with curative intent, (b) self-reported shortness of breath (mMRCDS score≥ 2), (c) age≥18 years, (d) Eastern Cooperative Oncology Group (ECOG) performance status=0–2 per clinician documentation,³⁹ (e) receipt of oncology care at a participating site, and (f) ability to complete questionnaires in English. We excluded patients who had acute cognitive or psychiatric conditions that could interfere with study participation per the judgment of the treating clinicians.

Procedures

Research staff identified eligible patients by reviewing the electronic health record (EHR) and clinic schedules, informing the attending oncology clinicians of each patient’s potential eligibility. If the clinicians did not respond with reservations, research staff approached the patient to describe the study and administer the mMRCDS. Patients who screened eligible and agreed to participate completed informed consent documents and baseline self-report measures either in clinic or via REDCap, a secure web-based platform.⁴⁰ An independent Office of Data Quality then randomly assigned participants using a computer-generated number sequence and 1:1 permuted block randomization (block size of 4) to receive the dyspnea intervention versus usual care, stratified by cancer type (non-small cell lung cancer versus small cell lung cancer/mesothelioma) and study site.

Participants completed the follow-up self-report measures at weeks 8 (primary endpoint), 16, and 24 with a +/− three-week window around each timepoint. Participants also received a wrist actigraph device, which they wore for three days both at baseline and at 8 weeks. Research assistants, blinded to group assignment, delivered the follow-up questionnaires in clinic, by mail, over the telephone, or via REDCap.⁴⁰ We selected the assessment timepoints based on our prior supportive care trials with this population.^38,41,42

Study Groups

Brief Behavioral Intervention for Dyspnea (Supplemental Table 1).

Patients assigned to the intervention participated in two nurse-led sessions within 8 weeks of enrollment. The sessions were 30–45 minutes each, taking place two to three weeks apart. Trained nurses delivered the intervention in clinic or via telephone or video. The first session focused on teaching patients evidence-based approaches for dyspnea management including: 1) psychoeducation about the relationship between dyspnea and the physiological stress response, 2) behavioral techniques for managing acute episodes of breathlessness (i.e., pursed-lip breathing, postural positions, and fan therapy), and 3) skills for reducing physiologic stress and strengthening lung capacity (e.g., slow diaphragmatic breathing). Participants received handouts describing these techniques, a handheld fan, and an audio-recording of a guided breathing exercise to facilitate home-based practice. The second session addressed patient adherence to the intervention, including review of participants’ use of the breathing exercises and techniques for acute dyspnea, problem-solving any barriers to practice, and reinforcement of the skills.

Usual Care Control Group.

Patients in both groups received usual oncology care, including any medical treatments for dyspnea prescribed by their oncology clinicians. At the time of enrollment, research staff informed patients in the control group that they could meet with a study nurse for one session after the final 24-week assessment to review the intervention skills and receive the materials.

Intervention Fidelity

The nurse interventionists underwent a comprehensive training with the PI (JG), a licensed clinical psychologist, and listened to at least three audio-recordings of exemplar sessions before meeting with participants. Throughout the study, research staff assessed fidelity to the intervention in two ways. First, after each session, the nurses documented the topics they discussed with the participants using a study-specific standardized survey. Second, we audio-recorded intervention visits and sampled at least 20% of each session for review to ensure the study nurses delivered the intervention per protocol using a pre-defined checklist of assigned topics. Finally, the PI led supervision with the study nurses every two to four weeks to review and address any challenges with intervention delivery.

Study Measures

The primary outcome was participant-reported dyspnea severity at 8 weeks on the mMRCDS.⁴³ Given increased recognition in the field regarding the multidimensional nature of dyspnea,⁴ we also administered the Cancer Dyspnoea Scale (CDS) as a co-primary outcome measure.⁴⁴ Secondary outcomes included QOL, psychological symptoms, and physical activity level at 8 weeks, as well as longitudinal analysis of all the outcomes across the four study time points.

Co-Primary Outcome Measures

Dyspnea:

The mMRCDS is a widely-used, single-item scale to evaluate dyspnea-related functional impairment with scores ranging from 0 (‘only get breathless with strenuous exercise’) to 4 (‘too breathless to leave the house’).⁴³ Investigators have reported a minimal clinically important difference of at least 0.5–1.0 points on the mMRCDS in response to treatment.^45–47 The CDS is a validated 12-item measure that includes three subscales assessing additional dimensions of dyspnea-related effort, anxiety, and physical discomfort. Higher scores on the CDS total scale (range: 0–48) and subscales reflect worse dyspnea.⁴⁴ The co-primary outcomes were the mMRCDS and CDS total scores, and the CDS subscales were exploratory outcomes.

Secondary Outcome Measures

Quality of Life:

We used the 36-item Functional Assessment of Cancer Therapy-Lung (FACT-L) scale to assess QOL. The FACT-L assesses physical, social, emotional, and functional well-being and lung cancer specific symptoms over the past week. Higher scores on the total scale (range: 0–136) and subscales indicate better QOL.⁴⁸

Psychological Symptoms:

Participants completed the Hospital Anxiety and Depression Scale (HADS), which consists of two 7-item subscales that assess anxiety and depression symptoms over the previous week. Subscale scores range from 0 (no distress) to 21 (maximum distress).⁴⁹

Activity Level:

Participants reported their activity level using the Godin-Shephard Leisure Time Physical Activity Questionnaire (GSLTPAQ), a 4-item measure assessing the frequency and extent of physical activity over the course of one week. Higher scores on the GSLTPAQ represent greater physical activity.^50,51 Also, as an objective measure of activity level, participants wore a wrist actigraph device during a consecutive three-day period at baseline and at week 8, which we used to calculate the percentage of time spent immobile while awake.⁵²

Sample Characteristics Measures

Sociodemographic and Clinical Characteristics.

Patients reported their demographic information and tobacco use on a sociodemographic questionnaire. Research staff also reviewed the EHR to record the date of cancer diagnosis, cancer type, ECOG performance status, medical treatments, and comorbidities (to calculate the Charlson Comorbidity Index [CCI]).⁵³

Statistical Analyses

All analyses were performed using R version 4.2.3. Before fitting regression models, we generated descriptive statistics and data visualizations for all longitudinal outcomes by week and study group. In accordance with the intention-to-treat principle, each participant was included in their randomly assigned group across all analyses.

For the two co-primary dyspnea outcomes (mMRCDS and CDS total scores at 8 weeks), we fit analysis of covariance (ANCOVA) models with terms for baseline dyspnea severity scores, cancer type, study site, and randomized group. For these models, we reported the intervention group coefficient estimates with 95% confidence intervals (CI) along with the Bonferroni-adjusted p-values to control overall type I error at 5%. Primary ANCOVA models were fit using a complete-case analysis. Sensitivity analyses using multiple imputation for nonresponse (excluding truncation due to death) were conducted for the two co-primary dyspnea outcomes; imputation models included all variables used in the ANCOVA models plus ECOG performance status and CCI. We also fit a partially nested mixed effect model including all terms used in the primary ANCOVA model plus a random nurse effect to account for potential clustering by nurse in the intervention group.⁵⁴ In a post hoc analysis, we examined the proportions of participants in each group reporting at least a minimal clinically important improvement, no change, or worsening on the mMRCDS using a proportional odds model with terms for cancer type, study site, and randomized group. We calculated Cohen’s d intervention effect sizes, with 0.2, 0.5, and 0.8 representing small, medium, and large effects, respectively.⁵⁵

We fit similar ANCOVA models for all secondary and exploratory subscale outcomes, reporting coefficient estimates, 95% CIs, and unadjusted p-values. To compare dyspnea severity change through week 24, we fit linear mixed effects models with random participant intercepts with fixed effects for cancer type, study site, randomization group, continuous study week (0, 8, 16, and 24), and a group-by-week interaction. The group-by-week estimate (with 95% CI) and unadjusted p-value were reported; this term represents the difference in slopes between the study groups. Given descriptive evidence for non-linear trajectories for the mMRCDS endpoint, we also fit post hoc mixed effects models with categorical study week, using the Akaike information criterion (AIC) to compare the continuous and categorical time mixed effects models.

For the power analysis, we had estimated that with 200 study completers the probability would be 80% that the study could detect a treatment difference in the 8-week primary and secondary patient-reported outcomes at a two-sided 0.05 significance level if the true difference between groups was 0.40 times the standard deviation. Given attrition in this morbid population, we increased the sample size to 250 participants.

Results

Sample Characteristics

Study staff approached 1,961 patients, of whom 58.5% reported less than moderate dyspnea, 6.5% asked to be reapproached in the future, and 34.9% met the dyspnea threshold to participate. Three hundred fifty-three patients (51.5%) signed the consent form, and 250 were registered and randomized, though three participants were excluded post registration, resulting in a final sample of 247 (Figure 1). As shown in Table 1, participants’ mean age was 66.15 years, with the majority of the sample identifying as female (55.9%), White (90.3%), and married/partnered (70.7%). Study groups were generally balanced with respect to clinical characteristics and use of medical interventions that could impact dyspnea (Supplemental Table 2).

Figure 1.

CONSORT Diagram

* Note: Early in the trial, 86 patients were deemed ineligible after signing consent form prior to IRB amendment allowing staff to screen for dyspnea before consenting procedures.

Table 1.

Baseline Sociodemographic and Clinical Characteristics

Variable	Usual Care N=121	Intervention N=126	Entire Sample N=247
Age, Mean (SD)	65.44 (10.08)	66.83 (11.47)	66.15 (10.81)
Sex, N (%)
Female	70 (57.9)	68 (54.0)	138 (55.9)
Male	51 (42.1)	58 (46.0)	109 (44.1)
Race, N (%)
American Indian or Alaskan Native	0 (0.0)	0 (0.0)	0 (0.0)
Asian	2 (1.7)	6 (4.8)	8 (3.2)
African American or Black	2 (1.7)	5 (4.0)	7 (2.8)
Native Hawaiian or Pacific Islander	0 (0.0)	0 (0.0)	0 (0.0)
White	113 (93.4)	110 (87.3)	223 (90.3)
Multiple Races	2 (1.7)	4 (3.2)	6 (2.4)
Other	2 (1.7)	1 (0.8)	3 (1.2)
Hispanic/Latino, N (%)
Yes	1 (0.8)	3 (2.5)	4 (1.7)
No	118 (99.2)	117 (97.5)	235 (98.3)
Missing	2	6	8
Religion, N (%)
Catholic	55 (45.5)	66 (53.2)	121 (49.4)
Other Christian (e.g., Protestant)	32 (26.4)	27 (21.8)	59 (24.1)
Jewish	11 (9.1)	7 (5.6)	18 (7.3)
Muslim	0 (0.0)	3 (2.4)	3 (1.2)
Other	4 (3.3)	4 (3.2)	8 (3.3)
Atheist	0 (0.0)	2 (1.6)	2 (0.8)
None	19 (15.7)	15 (12.1)	34 (13.9)
Missing	0	2	2
Relationship Status, N (%)
Married/partner	93 (77.5)	81 (64.3)	174 (70.7)
Single	2 (1.7)	7 (5.6)	9 (3.7)
Divorced/separated	12 (10.0)	21 (16.7)	33 (13.4)
Widowed/loss of partner	13 (10.8)	17 (13.5)	30 (12.2)
Missing	1	0	1
Education, N (%)
High school graduate or less	44 (36.4)	30 (23.8)	74 (30.0)
Associate degree/technical school	20 (16.5)	39 (31.0)	59 (23.9)
College graduate	25 (20.7)	28 (22.2)	53 (21.5)
Master’s or doctoral level degree	32 (26.4)	29 (23.0)	61 (24.7)
Annual Income, N (%)
Less than $25,000	12 (10.0)	16 (12.8)	28 (11.4)
$25,000-$49,999	19 (15.8)	19 (15.2)	38 (15.5)
$50,000-$99,999	31 (25.8)	32 (25.6)	63 (25.7)
$100,000-$149,999	18 (15.0)	19 (15.2)	37 (15.1)
$150,000 or more	24 (20.0)	20 (16.0)	44 (18.0)
Prefer not to answer	16 (13.3)	19 (15.2)	35 (14.3)
Missing	1	1	2
Cancer Type, N (%)
Non-small cell lung cancer	102 (84.3)	109 (86.5)	211 (85.4)
Small cell lung cancer	12 (9.9)	11 (8.7)	23 (9.3)
Mesothelioma	7 (5.8)	6 (4.8)	13 (5.3)
Cancer Treatment, N (%)
Radiation	20 (16.5)	19 (15.1)	39 (15.8)
Intravenous chemotherapy	64 (52.9)	67 (53.2)	131 (53.0)
Oral/targeted chemotherapy	35 (28.9)	33 (26.2)	68 (27.5)
Immunotherapy alone/or combination	45 (37.2)	52 (41.3)	97 (39.3)
Smoking Status, N (%)
Current smoker	7 (5.8)	5 (4.0)	12 (4.9)
Prior smoker	84 (69.4)	80 (63.5)	164 (66.4)
Never smoker	30 (24.8)	41 (32.5)	71 (28.7)
ECOG Performance Status, N (%)
0	17 (14.0)	15 (11.9)	32 (13.0)
1	73 (60.3)	88 (69.8)	161 (65.2)
2	31 (25.6)	23 (18.3)	54 (21.9)
Time Since Diagnosis of Metastatic Disease, Median Months (IQR)	13.15 (4.13 – 31.95)	11.64 (4.34 – 28.98)	12.95 (4.15 – 31.40)
Charlson Comorbidity Index, M (SD)	1.50 (1.58)	1.37 (1.62)	1.44 (1.60)
Dyspnea Severity, M (SD)
mMRCDS	2.60 (0.75)	2.59 (0.72)	2.60 (0.73)
CDS	12.61 (7.51)	11.13 (6.51)	11.85 (7.05)
Quality of Life, M (SD)
FACT-L	87.23 (16.36)	90.19 (18.59)	88.74 (17.56)
Psychological Symptoms, M (SD)
HADS-Anxiety	6.68 (3.76)	5.37 (3.80)	6.01 (3.83)
HADS-Depression	6.40 (3.16)	5.73 (3.72)	6.06 (3.47)
Activity Level, M (SD)
GSLTPAQ	11.19 (18.20)	12.29 (16.91)	11.75 (17.53)
Actigraphy (percent time immobile)	50.48 (13.62)	47.61 (11.19)	49.00 (12.48)

Note: Lung cancer was not included in the calculation of the Charlson Comorbidity Index. ECOG=Eastern Cooperative Oncology Group; mMRCDS=Modified Medical Research Council Dyspnea Scale; CDS=Cancer Dyspnoea Scale; FACT-L=Functional Assessment of Cancer Therapy-Lung Scale; HADS=Hospital Anxiety & Depression Scale; GSLTPAQ=Godin-Shephard Leisure Time Physical Activity Questionnaire.

Intervention Delivery

Fourteen nurses led the dyspnea intervention sessions (8 MGH, 6 DFCI), with the number of participants per nurse ranging from 1 to 31 (median 3.5). Of the 126 participants assigned to the intervention, 84.1% (n=106) completed the first session and 73.8% (n=93) completed the second session. Those who did not finish the intervention either withdrew (11.9%, n=15); were too ill, enrolled in hospice, or died during the study (7.1%, n=9); declined the session (4.0%, n=5); or were lost to follow-up or transferred care (3.2%, n=4). The median days from baseline to the first session was 21 (IQR=14–29) and from the first to second session was 21 (IQR=14–27). With respect to intervention fidelity, study nurses completed 100% of post-session surveys, and staff review of session audio-recordings showed that the study nurses addressed 96.6% and 95.0% of topics per protocol for the first and second sessions, respectively. Per the nurse survey for session two documenting use of skills, 94.6% of intervention participants practiced at least one skill taught in the first session, including pursed-lip (87.1%) and diaphragmatic (78.5%) breathing, followed by the postural positions (65.2%), guided breathing audio-recording (58.2%), and hand-held fan (56.5%). No participants reported adverse events related to the intervention.

Co-Primary Outcomes

Results from ANCOVA models (Table 2) demonstrated that patients assigned to the intervention reported significantly lower dyspnea versus the usual care group on the primary mMRCDS outcome at 8 weeks (B=−0.33, 95% CI=−0.61, −0.05, Bonferroni-corrected p=.038, between-group Cohen’s d=0.32). Of note, both study groups reported improved mMRCDS scores from baseline to week 8 (within-group Cohen’s d=1.47 for the intervention vs 0.94 for usual care). As shown in Figure 2, the percentage of patients who improved on the mMRCDS from baseline to week 8 by at least one point (i.e., the minimal clinically important difference) was higher in the intervention group compared to usual care group (72.0% vs 59.2%). A proportional odds model, controlling for baseline factors, showed that the intervention group was more likely to experience at least a one-point improvement on the mMRCDS versus no change versus worsening compared to participants in the control group (OR=2.02, 95% CI=1.09, 3.74, p=.025). With respect to the multidimensional measure of dyspnea, the intervention led to significantly lower scores on the CDS Discomfort subscale (B=−0.59, 95% CI=−1.16, −0.01, p=.045) at 8 weeks, but not the co-primary outcome of the CDS total score or other exploratory CDS subscale outcomes (Table 2). The categorical time mixed effects model (AIC=1936.40), which fit the longitudinal data better than the continuous time model (AIC= 2001.51), confirmed the 8-week mMRCDS dyspnea results (B=−0.35, 95% CI=−0.65, −0.06). The missing data sensitivity analysis using multiple imputation for 8-week mMRCDS also led to similar results (B=−0.35, 95% CI=−0.59, −0.11), as did the sensitivity analysis accounting for potential clustering by study nurse (B=−0.33, 95% CI=−0.63, −0.04).

Table 2.

ANCOVA Models of Study Group Differences in 8-Week Outcome Measures

Outcome Measure	Adjusted Mean (95% CI)
	Usual Care	Intervention	Difference (95% CI)	p-value	N
Dyspnea Severity
mMRCDS*	1.88 (1.69, 2.07)	1.55 (1.35, 1.75)	−0.33 (−0.61, −0.05)	0.038	190
CDS Total*	10.01 (8.87, 11.15)	9.47 (8.28, 10.65)	−0.55 (−2.20, 1.10)	>0.99	191
CDS Effort	5.33 (4.72, 5.94)	5.33 (4.70, 5.97)	0.00 (−0.88, 0.88)	>0.99	191
CDS Anxiety	1.84 (1.39, 2.29)	1.71 (1.24, 2.17)	−0.14 (−0.78, 0.51)	0.680	191
CDS Discomfort	2.92 (2.52, 3.32)	2.33 (1.92, 2.75)	−0.59 (−1.16, −0.01)	0.045	191
Quality of Life
FACT-L Total	91.40 (88.91, 93.89)	93.00 (90.43, 95.57)	1.60 (−1.98, 5.18)	0.380	188
Physical Wellbeing	18.93 (18.17, 19.70)	18.82 (18.02, 19.61)	−0.12 (−1.22, 0.99)	0.837	191
Social Wellbeing	22.35 (21.65, 23.04)	22.39 (21.68, 23.11)	0.05(−0.95, 1.05)	0.926	190
Emotional Wellbeing	16.30 (15.65, 16.94)	16.98 (16.31, 17.65)	0.68 (−0.25, 1.62)	0.152	189
Functional Wellbeing	15.44 (14.61, 16.28)	16.83 (15.97, 17.70)	1.39 (0.18, 2.59)	0.024	189
Lung Cancer Subscale	18.13 (17.42, 18.85)	18.13 (17.39, 18.87)	−0.01 (−1.04, 1.03)	0.991	188
Psychological Symptoms
HADS Anxiety	5.85 (5.27, 6.42)	5.84 (5.25, 6.43)	−0.01 (−0.83, 0.82)	0.990	191
HADS Depression	6.09 (5.50, 6.67)	5.59 (4.99, 6.20)	−0.49 (−1.34, 0.35)	0.251	191
Activity Level
GSLTPAQ	12.73 (8.43, 17.03)	11.24 (6.50, 15.98)	−1.49 (−7.90, 4.92)	0.647	164
Actigraphy (percent time immobile)	45.80 (43.06, 48.53)	47.42 (44.46, 50.38)	1.63 (−2.45, 5.71)	0.431	102

^*Bonferroni adjusted p-values calculated as min{2 x unadjusted p-value, 1.00}v.

Note: All models control for cancer type (i.e., non-small cell lung cancer versus small cell lung cancer/mesothelioma), study site, and baseline scores of outcome variable. mMRCDS=Modified Medical Research Council Dyspnea Scale; CDS=Cancer Dyspnoea Scale; FACT-L=Functional Assessment of Cancer Therapy-Lung Scale; HADS=Hospital Anxiety & Depression Scale; GSLTPAQ=Godin-Shephard Leisure Time Physical Activity Questionnaire.

Figure 2.

Clinically Meaningful Change in Patient-Reported mMRCDS Scores from Baseline to Week 8

Note: “Improved” = a decrease of at least one point on the mMRCDS; “No Change” = same mMRCDS score at both time points; and “Worsened” = an increase of at least one point on the mMRCDS.

Secondary Outcomes

Intervention patients also reported better functional wellbeing at week 8 per the FACT-L (B=1.39, 95% CI=0.18, 2.59, p=.024) versus the usual care group. However, study groups did not differ significantly at the 8-week endpoint with respect to the FACT-L total score or other subscales, psychological symptoms, or physical activity level per self-report or actigraphy (Table 2). Finally, the longitudinal analyses showed that, while both study groups reported significantly improved dyspnea on the mMRCDS and CDS total scores from baseline to 24 weeks, the trajectories of these linear slopes did not differ between the groups for any of the outcome measures (Table 3).

Table 3.

Linear Mixed Effects Models of Study Group Differences in Patient-reported Outcomes from Baseline to 24 Weeks

Outcome Measure	Estimated change over 24 weeks (95% CI)
	Usual Care	Intervention	Difference (95% CI)	p-value	N1	N2
Dyspnea Severity
mMRCDS	−0.89 (−1.16, −0.61)	−0.95 (−1.24, −0.66)	−0.06 (−0.38, 0.26)	0.710	247	196
CDS Total	−2.63 (−4.13, −1.13)	−2.26 (−3.84, −0.67)	0.38 (−1.40, 2.16)	0.677	247	196
CDS Effort	−1.43 (−2.25, −0.60)	−1.24 (−2.11, −0.37)	0.19 (−0.79, 1.17)	0.707	247	196
CDS Anxiety	−1.11 (−1.72, −0.50)	−0.49 (−1.13, 0.15)	0.62 (−0.10, 1.34)	0.092	247	196
CDS Discomfort	−0.17 (−0.72, 0.38)	−0.60 (−1.17, −0.02)	−0.42 (−1.07, 0.23)	0.204	247	196
Quality of Life
FACT-L Total	4.86 (1.57, 8.14)	2.91 (−0.60, 6.41)	−1.95 (−5.87, 1.97)	0.330	247	194
Physical Wellbeing	0.64 (−0.44, 1.73)	0.66 (−0.49, 1.81)	0.01 (−1.28, 1.30)	0.984	247	196
Social Wellbeing	−0.26 (−1.12, 0.60)	0.26 (−0.66, 1.17)	0.52 (−0.51, 1.54)	0.324	247	196
Emotional Wellbeing	1.11 (0.23, 1.98)	0.34 (−0.59, 1.26)	−0.77 (−1.81, 0.27)	0.147	247	195
Functional Wellbeing	1.21 (0.13, 2.29)	0.53 (−0.61, 1.67)	−0.68 (−1.96, 0.61)	0.303	247	195
Lung Cancer Subscale	2.32 (1.35, 3.28)	1.31 (0.30, 2.33)	−1.00 (−2.14, 0.14)	0.086	247	194
Psychological Symptoms
HADS Anxiety	−1.00 (−1.73, −0.28)	−0.28 (−1.04, 0.49)	0.73 (−0.13, 1.59)	0.098	247	196
HADS Depression	−0.18 (−0.89, 0.53)	−0.10 (−0.85, 0.66)	0.08 (−0.76, 0.93)	0.845	247	196
Activity Level
GSLTPAQ	0.54 (−5.30, 6.38)	1.28 (−5.09, 7.65)	0.74 (−6.33, 7.77)	0.837	240	186

Note: All models control for cancer type (i.e., non-small cell lung cancer versus small cell lung cancer/mesothelioma) and study site. mMRCDS=Modified Medical Research Council Dyspnea Scale; CDS=Cancer Dyspnoea Scale; FACT-L=Functional Assessment of Cancer Therapy-Lung Scale; HADS=Hospital Anxiety & Depression Scale; GSLTPAQ=Godin-Shephard Leisure Time Physical Activity Questionnaire. N1 includes baseline measure and N2 excludes baseline measure.

Discussion

The results of this RCT demonstrate that a nurse-led brief behavioral intervention improved patient-reported dyspnea per the mMRCDS at the 8-week primary outcome timepoint in patients with advanced lung cancer compared to usual care. Exploratory analyses also suggest that the intervention may improve the sense of dyspnea-related discomfort and functional wellbeing in this population, though these latter findings require confirmation. The intervention did not impact overall QOL, psychological symptoms, or physical activity level. Finally, longitudinal analyses showed that both the intervention and usual care groups maintained improvements in dyspnea from baseline to 24 weeks, but these trajectories did not differ between the groups across the subsequent follow-up time points.

Patient-reported dyspnea improved in both study groups, each with corresponding within-group large effect sizes. As shown in Figure 2, the majority of intervention and control participants experienced a clinically meaningful reduction in dyspnea, though a greater proportion of the intervention group reported this improvement compared to the usual care group. These findings were surprising, given prior research showing that dyspnea tends to worsen in patients with advanced lung cancer over time.^8–10 Several factors could account for the natural course of improved dyspnea, including advances in lung cancer therapies; regression to the mean in a sample that was screened for at least moderate breathlessness as an eligibility criterion; and the expectation of usual care participants to receive the intervention at the end of the study. Although participants in both groups received similar types and lines of cancer treatment prior to and during the study, we are unable to rule out the potential confound of response to cancer therapy.

The intervention primarily focused on learning skills for taking slow deep breaths and coping with moments of acute dyspnea by using pursed-lip breathing, postural relief, and a handheld fan. Therefore, the intervention understandably improved dyspnea-related functional disability (as per the mMRCDS and FACT-L Functional Wellbeing subscale) and discomfort (per the CDS-Discomfort subscale, which assesses ease of inhaling, exhaling, and breathing slowly). The lack of intervention effects on the other domains of dyspnea (e.g., anxiety) and secondary outcomes are likely due to the targeted focus of the intervention, low dose, or both.

While more comprehensive exercise and rehabilitation programs may also improve dyspnea in patients with advanced lung cancer, the scalability and acceptability of such intensive programs for a highly morbid population remains a salient question.^56,57 A key goal of our brief behavioral intervention was to enhance access for patients to learn evidence-based skills for managing dyspnea within the flow of oncology care and from trusted nurses, which was confirmed in the high degree of fidelity to the intervention protocol. Although the intervention patients did not utilize every skill, most reported that they practiced at least the pursed-lip and diaphragmatic breathing techniques. Similarly, a small trial of a single session, 20-minute mindful breathing exercise rapidly reduced dyspnea in patients with lung disease,⁵⁸ providing corroborating evidence that such brief interventions can be beneficial. Nonetheless, further research is needed on ways to boost intervention effects over time while retaining the “low-touch.”

Although this study has many strengths in its rigorous RCT design, several limitations warrant consideration. The sample was relatively homogeneous with respect to the clinical setting and participant characteristics, limiting generalizability of the results to community cancer centers and underrepresented groups. Also, the rate of attrition was higher in the intervention group compared to the usual care group, which may have impacted the ability to detect intervention effects on outcomes over time. Although we designed the intervention to limit burden on patients, it primarily focused on dyspnea-related functional disability rather than fully addressing the multidimensional nature of the symptom. Finally, the lack of a systematic assessment of potential adverse events may have resulted in under-documentation of participants’ experiences.

Despite considerable advances in cancer therapy, dyspnea remains one of the most distressing and debilitating symptoms that patients with advanced lung cancer experience, with limited treatments to provide relief. This trial demonstrates the efficacy of a scalable, brief behavioral intervention for improving dyspnea, which oncology nurses can easily provide in their practice.

Context Summary.

Key Objective:

What is the effect of a nurse-led brief behavioral intervention for dyspnea in patients with advanced lung cancer?

Knowledge Generated:

Compared to usual care, a two-session behavioral intervention led by oncology nurses, which focused on breathing techniques, postural positions, and use of a hand-held fan, led to a reduction of dyspnea in patients with advanced lung cancer at the primary eight-week outcome timepoint. The intervention did not impact overall quality of life, psychological distress, or activity level.

Relevance:

This minimally intensive intervention for dyspnea holds promise for broader scalability in oncology nursing practice. Additional research should investigate methods to enhance effects across additional outcomes over a longer time period.