Interventions to improve quality of life (QOL) and/or mood in patients with head and neck cancer (HNC): a review of the evidence

Background Patients with head and neck cancer (HNC) experience significant physical and psychological distress, which have a negative impact on their quality of life (QOL). Few strategies have been studied to help improve QOL in this patient population. Results In this article, we review the existing literature for intervention studies that focus on improving QOL and/or mood in HNC patients. Our review yielded 14 studies that met criteria. Types of interventions included educational, psychosocial, physical and psychological symptom management, mindfulness, pharmacologic, exercise, and telemedicine. Although the majority of the studies had small sample sizes or other methodological limitations, many showed preliminary feasibility and acceptability with some positive impacts on QOL and/or mood. Conclusions Larger studies are warranted with more robust randomized designs to determine efficacy of interventions to improve QOL and/or mood in patients with HNC. Additionally, future studies must also consider strategies for implementation and dissemination of these interventions into the health care system to improve the physical and psychological burden of HNC as a population.

Additionally, the symptom burden during treatment is significant as patients report a myriad of symptoms including severe pain, distress, dehydration, malnutrition, nausea, constipation, and sleep disturbances [2,3].
Throughout this intense treatment, psychological distress is particularly high in patients with HNC, with levels of depression, anxiety, and post-traumatic stress disorder (PTSD) among the highest reported of any cancer population [4,5]. Between 15 and 50% of HNC patients are affected by major depressive disorder at some point during their treatment [5]. Depression levels change over the treatment trajectory, with one study showing the highest levels at the completion of radiotherapy [6]. Suicide rates are four times higher than the general population [7]. Patients who experience higher levels of depression during treatment can face worse outcomes than those who experience lower levels of depression during treatment [4].
Patients with HNC and untreated psychological needs have a worse QOL, longer hospital stays, worse postoperative performance status, increased narcotic use, decreased treatment compliance, and increased complications with treatment [8]. Additionally, numerous studies demonstrate a link between depression and worse survival [9,10].
Not only is depression a major problem for patients with HNC, but those undergoing treatment also experience significant levels of anxiety and PTSD [4,11]. Gil et al noted that patients often experience the highest levels of anxiety at the time of diagnosis 12 . Additionally, in the face of a new diagnosis, 12% of patients and 29% of caregivers met criteria for PTSD [11]. HNC patients often deal with cosmetic issues due to radiation and surgical interventions. This lasting disfigurement increases the risk of PTSD [12,13]. In a study by Moschopoulou et al, the prevalence of PTSD at the 6-year mark was found to be 13%, which is significantly higher than the general population [12]. Thus, improving how we address psychological symptoms is paramount to patients' overall QOL and mood as we strive to optimize the care of our patients with HNC.
Among cancer survivors, evidence-based interventions have proven effective in reducing psychological distress associated with a cancer diagnosis or treatment and improving QOL [14]. However, few such interventions have been designed and implemented to ameliorate the severe psychological and functional burden of HNC. Thus, the primary objective of this paper is to review these interventions and evaluate their impact on psychological symptoms and QOL, with a particular focus on study methodology, outcome measures, and efficacy. We conclude with a summary of the methodological challenges in these trials and propose future research directions to improve the QOL and mood of patients with HNC.

Search and selection criteria
We conducted a literature search using the online PubMed database to identify studies in English through June 2018 of interventions during or after HNC treatment designed to improve psychological symptoms and/ or QOL in patients as either a primary or secondary endpoint. Authors JS and JB performed this search independently and then compared results, deciding together whether to include each study. Key words and phrases used alone and/or in combination for the search included HNC, QOL, intervention, depression, anxiety, PTSD, and/or randomized-controlled trial (RCT). Reference lists from citations were also reviewed for relevant publications. We included studies with a prospective intervention to improve QOL and or psychological symptoms during or after HNC treatment. Studies could be RCTs, non-randomized, controlled trials, or pilot studies. We excluded studies that were not prospective interventions (1 study), those that were ongoing or in development without outcomes reported (3 studies), or those that did not collect patient reported outcomes (1 study). Using the designated search strategy, 14 studies were identified. The methodological and statistical quality of the studies was assessed using a 7-item checklist with the following criteria: (A) sample characteristics, (B) Sample size, (C) Data collection, (D) Response Rates, (E) Outcome measurement, (F) Comparison groups, (G) Statistical Analyses adapted from previous published, standardized checklists [15]. Descriptive statistics were calculated using Microsoft Excel. Authors JS and JB scored all 14 studies independently and then discussed discrepancies until a consensus was formed.

Nurse-led interventions
Four studies summarized in Table 1 evaluated interventions that were delivered by nurses [16][17][18][19]. Katz et al implemented an educational intervention in which 19 patients were randomized to usual care versus receipt of an educational booklet outlining information on diagnosis, treatment, and coping strategies for patients undergoing surgery for oral cavity cancer, as well as two educational sessions with a nurse [16]. As compared with control patients, patients in the experimental group demonstrated significantly lower anxiety and higher knowledge at the three-month follow-up. There were no differences between the groups for QOL or depression. In a larger study by van der Meulen et al [19], 205 patients were randomized to receive either usual care or the psychological intervention NUCAI (nurse counseling and intervention). NUCAI consisted of six bimonthly sessions on problem-focused counseling to address physical, psychological, and social consequences of HNC and treatment. The patients receiving the intervention reported lower levels of depression 1 year after HNC treatment compared to control patients, as well as decreased physical symptoms including pain and swallowing troubles. The remaining two studies involved patients who screened positive for psychological dysfunction. Duffy and colleagues randomized patients who screened positive for tobacco use, alcohol use, and/or depression to usual care versus a nurse-administered cognitive behavioral therapy (CBT) intervention with medications when necessary [17]. There were no differences between intervention and control groups in depressive symptoms or alcohol use, both of which decreased during the course of the study across groups. However, patients in the intervention group were more likely to quit smoking than in the usual care group. Finally, Semple and colleagues conducted a non-randomized, controlled study of patients who screened positive for psychological dysfunction of an individualized, psychosocial intervention program delivered at home versus usual care for patients who did not want to participate (control group) [18]. Patients in the intervention group received between 2 and 6 home sessions lasting approximately 90 min each, which targeted pertinent social, psychological, and physical problems faced by patients after treatment. Patients who received the intervention had improvements in anxiety and depression, as well as improved components of QOL. Control patients had no significant improvements in psychological function or QOL over the study time period.
These data suggest that a variety of nurse-led psychosocial or educational interventions improve some aspects of mood and QOL for patients with HNC, but not consistently. The interventions themselves were markedly different, as were the outcome measures used, which makes it challenging to draw unifying conclusions or do cross intervention comparisons. However, it is important to note that patients found the intervention content acceptable, and the nurse-led delivery format was both feasible and acceptable. Patients may be more willing to participate in an intervention that is delivered by someone already involved in their care, such as a nurse, who is without the stigma of being a mental health provider (i.e. a psychologist or psychiatrist). Involving someone already on the care team also would aid in dissemination into the health system. Additionally, the problem focused intervention and the NUCAI intervention both showed significant improvements in patients' depression, which were their primary outcomes, underscoring that interventions targeting mood in HNC can be effective compared to usual care. There are important methodological limitations to these studies, however, that have implications on how we interpret the outcomes reported. The only study that reported race was majority white. Therefore, outcomes in a more heterogeneous population are unknown. Two of the studies [16,18] had small sample sizes with important baseline differences between the control and intervention groups that could account for differences noted in the outcomes. For example, more patients in the control group in Katz et al had more extensive surgeries than those in the experimental group, which could account for the poorer outcomes in the control group because of the greater morbidity of the surgery. In the Semple study [18], the experimental group reported higher levels of anxiety and depression than the control group, and the control group was self-selected for those who did not want to participate in the home intervention. These differences could account for the more pronounced improvements seen in the experimental group, making the effect of the intervention less clear.
Additionally, each intervention was complex and no study analyzed which aspects of the intervention were most helpful. Finally, the interventions were all conducted at different times during patients' cancer treatment courseone prior to any treatment, two posttreatment, and one involving patients at any timewhich can impact how the data are interpreted since psychological and physical symptoms worsen during treatment. Understanding the expected trajectory of mood and QOL throughout treatment is critical to interpreting outcomes. Finally, two of these interventions only enrolled patients who screened positive for psychosocial dysfunction. Thus, application of these interventions to a broader population of patients with HNC is unclear. Interestingly, the NUCAI intervention was a more inclusive population of all patients undergoing treatment for HNC, which showed decreases in depression across the entire study population indicating that benefits can be seen regardless of baseline mood. Overall, it appears that nurse-led interventions targeting physical and psychological functioning are beneficial, though outcomes must be interpreted with caution.

Psychologist-led interventions
Six studies in Table 2 employed interventions targeting psychological distress which were delivered by psychologists [20][21][22][23]28]. Hammerlid et al enrolled two small, single arm studies of patientsone group in a longterm group psychological therapy and the second in a week long, short-term psychoeducational program with their caregivers [20]. Patients in both intervention groups received supportive therapy in a group format and reported improvement in emotional, psychosocial, and depressive symptoms. The second study by Allison et al involved a feasibility study with the Nucare coping strategies program, which was a program to teach individuals how to cope with their cancer diagnosis and treatments by way of problem solving, relaxation training, and individualized goal setting either in a group or one-on-one setting with a therapist or at home without a therapist [21]. Following the intervention, patients reported higher QOL and lower depressive symptoms than at baseline. The third study by Kilbourn et al was also a single arm feasibility study that assessed the intervention Easing and Alleviating Symptoms during Treatment (EASE) [28]. All participants received telephone counseling sessions focused on coping and stress management throughout radiation treatment. The intervention was feasible and acceptable, however, baseline to postintervention QOL scores decreased slightly and pain scores did not change. The fourth study by Pollard et al was also a single arm feasibility study that assessed a mindfulness intervention for patients receiving radiation treatment [29]. This intervention was also feasible and acceptable, but baseline to post-intervention QOL scores decreased across the study timepoints. However, participants with higher post-intervention mindfulness did have higher QOL scores. The remaining two studies targeted patients with elevated distress. Kangas and colleagues conducted a pilot RCT of an early CBT intervention among 35 patients with elevated levels of PTSD, depression, or anxiety during radiation treatment for HNC [22]. Patients were randomized to receive seven weekly sessions with a clinical psychologist of either a multi-modal cognitive behavioral therapy (CBT) intervention or non-directive supportive counseling (SC). Both interventions equally reduced PTSD and anxiety, however, more patients in the CBT arm no longer met criteria for PTSD, anxiety, and/or depression at the 12-month time point compared to patients receiving SC. The final study was an RCT of a stepped care intervention targeting psychological distress in patients with HNC who screened positive for distress [23]. Patients were randomized to usual care or to the stepped care intervention. Stepped care consisted of watchful waiting, guided self-help, problem-solving therapy, and/or psychotherapy with or without medication, where care would be escalated to the next step if distress did not improve. Patients who received the intervention had greater improvements in symptoms of anxiety and depression compared to patients receiving usual care.
Four of these six studies demonstrate that psychologist-led interventions targeting psychological symptoms in patients with HNC can improve depression and anxiety symptoms, PTSD, and/or QOL. Together, these studies suggest that patients are likely to benefit from programs led by mental health professionals and highlight that they will participate in and value these types of interventions. For example, in Kangas et al, enrollment rate to the study was high (> 80% of those eligible enrolled), and > 70% of participants attended all counseling sessions during their HNC treatment [22]. The results from two of these studies were also more methodologically robust because they used a randomized, controlled design [22,23], which is more able to assess preliminary efficacy. The positive outcomes across four of the studies also suggest that HNC patients with or without baseline distress can benefit from such interventions since two of the trials included patients with HNC.
These studies also have important methodological limitations. Four had small sample sizes [20,21,28] and were single arm studies with no control group, and while Kangas et al was a randomized trial [22], both arms received an intervention with a psychologist, and thus there was no true control. Therefore, the improvements in outcomes noted in these studies could be attributed to a natural improvement over time in QOL and/or psychological symptoms, which is known to occur post-HNC treatment. On the opposite spectrum, the two studies [28,29] that did not show patient improvement only assessed timepoints at baseline and immediately post-intervention, when participants likely still had significant treatment side effects and thus an improvement of QOL was unlikely to be found. No study reported race or ethnicity. Additionally, there were significant amounts of missing data across the studies, which may have influenced the results. Most of the interventions were also time and likely cost intensive, which potentially limits how well they can be disseminated. Finally, although psychologists have appropriate expertise to deliver these interventions, barriers exist for patients to receive psychological care, which can thereby restrict adoption of the interventions despite known benefits.

Pharmacologic intervention
Only one study targeted depression in HNC treatment using a pharmacologic intervention, summarized in Table 3 [24]. Lydiatt et al conducted a randomized, double blind study with prophylactic escitalopram, a selective serotonin reuptake inhibitor approved to treat depression, versus placebo in non-depressed patients undergoing treatment for HNC [24]. The head and neck surgeon involved in the patient's care gave the overview of the study and a delayed consent process was used. All patients were followed for the study in the surgery clinic. The rate of depression in patients receiving escitalopram was significantly lower than the rate in those receiving placebo, and the QOL of patients receiving escitalopram was significantly higher. This was the first trial in a cancer population to show success in the prevention of depression, and the effect was most pronounced in those receiving radiation therapy where depression was highest. Although there were improved outcomes in both depression and QOL, the concerns of the study revolved around the cost of drug therapy, as well as the possible side effects, though the vast majority of participants reported adverse effects 25% of the time or less. Additionally, the population was ethnically and racially homogenous; thus, its application to a more heterogeneous population remains an important question. Further exploration of which HNC patients may benefit most from this intervention will be important.

Lifestyle intervention
One study, also summarized in Table 3, evaluated QOL and mood as a secondary outcome through a lifestyle intervention targeting physical activity and health education [25]. Capozzi et al conducted an RCT of an exercise intervention in patients with newly diagnosed HNC. The intervention included a physician referral and clinic support, health education sessions, and an individual as well as group exercise program with an exercise physiologist and certified personal trainer. Patients were randomized to the 12-week exercise and lifestyle intervention or a wait-list control group, which received the same intervention delayed by 12 weeks. There were no differences between groups in body composition, QOL, or mood at 12 or 24 weeks. It should be noted that patients randomized to the wait-list control (i.e. delayed intervention group) showed greater adherence to the exercise and lifestyle intervention than patients who received the intervention at study onset. Although this study did not show a difference in the primary or secondary outcomes, every participant eventually received the intervention. Thus, there was no true control group for long-term outcomes of 36-48 weeks, which is the time when a positive impact of the intervention may have become clearer. Thus, exploration of this intervention on longterm outcomes as the primary outcome is warranted.

System-based interventions
Two other studies, in Table 4, evaluated interventions targeting QOL or mood through a health-system based intervention with a telemedicine system [26] or a tailored information package [27]. Van den Brink, et al led a non-randomized, controlled trial evaluating a telemedicine system integrated into patients' care compared with a group of patients in a different hospital without access to the telemedicine system [26]. The intervention included a laptop for communication with the health care team, health information, a forum with patients, and home symptom monitoring. Patients in the intervention group had improvements in QOL at 6 weeks post-intervention compared to the control group, though these differences were no longer significant at 12 weeks. The second study by D'Souza et al was also a non-randomized, controlled trial integrating the delivery of a comprehensive information package into patient care compared to a control group at a different hospital who did not receive the information [27]. Patients who received the intervention reported less anxiety than the control group, while there was no difference in depression.
Although these studies are rather different in the interventions tested, each targets QOL and/or mood using novel strategies. There were mild intervention effects across the studies with possible improvements in QOL and anxiety, but both had notable methodological shortcomings. Both were non-randomized, controlled studies and thus baseline differences may have accounted for differences in outcomes. Additionally, the D'Souza study [27] measured baseline data after the intervention had started, which is problematic because the baseline data could have been influenced by the intervention. Finally, the timing of the outcome measures across the study may contribute, in part, to some of the negative results observed. For example, van den Brink [26] only measured outcomes through week 12, which would have been at the end of adjuvant treatment for any patients receiving radiation, a time point when QOL and mood are nadiring. This could explain why the initial differences in QOL between intervention and control groups were no longer evident. Later timepoints when patients have recovered from treatment fully are helpful to assess the long-term effects of the intervention. Given these methodological issues, the negative results of these trials should be interpreted cautiously.

Discussion
All 14 studies reviewed here used distinct, novel interventions to improve QOL and mood in patients with HNC. The studies included different HNC patient populations, cancer stages, follow-up times, outcome measures, and were of variable quality. However, despite these differences, the studies illustrate that these interventions are well-received by patients and can be beneficial to them. The majority of the interventions led to a positive impact on patient-reported QOL and/or mood. Although determining which intervention was the most successful is difficult, important themes emerged from the interventions as a whole. First, nurse-led and psychologist-led interventions were both successful in improving patient outcomes, but nurse-led interventions may be easier to integrate into patients' care with less stigma and higher availability, in particular when the intervention visits are coordinated with other treatment visits for their cancer or at the patients' homes. Second, education was a critical component of many of the interventions, underscoring the importance of patients' understanding of the cancer and treatment, and how improved understanding can positively impact QOL and/or mood. Third, most intervention components involved several sessions, emphasizing that ongoing reinforcement of the topics for patients is necessary to effect change. Fourth, multicomponent interventions were important because they were able to address multiple aspects of patient care simultaneously in order to impact QOL and/or mood. Finally, many of these intervention trials included all patients with HNC and showed positive effects across the entire population studied. Therefore, these types of interventions should not only be considered for those who are already distressed, but rather for the HNC population as a whole given the significant burden of treatment.

Challenges and future directions
The studies reviewed also highlight significant limitations of the research methodology employed that must be understood in order to improve future studies. First, many of these studies were either single arm studies, or had small sample sizes as seen in the quality assessment, limiting conclusions that can be drawn about efficacy. Necessary next steps will involve conducting larger RCTs, which can appropriately assess efficacy. Second, there was inconsistency across these studies in the outcome measures used to assess QOL and/or mood, as well as in the timing of outcome assessments. Future studies should be standardized in the outcome measures used as well as the timing of the measurement in order to be able to draw consistent conclusions. Ideally, patients should be enrolled at the same point in time in relation to their treatment (either prior to treatment start or at a specified point post-treatment) and then assessments should be measured at baseline, post treatment if applicable, and at 3, 6, and 12 months due to the impact of treatment and length of recovery. Including a mixed population is particularly problematic in HNC because patients in active treatment have worse QOL and mood than those before or after treatment. Thus, a mixed population of patients may introduce major differences in outcome data that could be misinterpreted as an intervention effect. Third, only 4 of the 14 studies measured long-term outcomes at 1 year or later, which is an important timepoint in patients' treatment trajectories in regard to long-term QOL and mood. Many studies also had significant drop off in response rates of the outcomes on later assessments, which could have biased results. Future studies should enroll patients at the same point in time in their treatment trajectory in order to minimize baseline differences, include long-term outcome measures, and maintain high response rates to the extent possible such that the data is representative of the population.
The majority of interventions were led by a healthcare professional (nurse, psychologist, clinician, exercise physiologist), which can be labor-or cost-intensive. However, recent technological advances have resulted in the development of novel digital health interventions that have the potential to positively benefit cancer survivors [30]. Although few such interventions have been evaluated in patients with HNC, one study found that oral cancer survivors rated a web-based symptom management program favorably and were interested in using online tools to improve QOL [31]. Thus, a combined program that integrates online resources with in-person support may be helpful in promoting positive psychological outcomes at reduced costs.
Only two of the interventions were focused on the healthcare delivery system in which patients were receiving care, and no intervention was focused on the dissemination and implementation of the intervention into the healthcare system. In order to improve health outcomes in patients with HNC on a population level, determining how to implement these interventions health system-wide will be critical to establishing feasibility.
Finally, this review focused primarily on interventions for patients with HNC. However, caregivers of these patients also report significant distress and poor QOL during the patient's treatment, which can also impact outcomes [15]. There are even fewer interventions targeting caregivers or caregiver and patient dyads in HNC. Thus, broadening future interventions to include a caregiver component will also be important to improve overall outcomes.
In conclusion, with the small number of studies meeting criteria for our review, it is clear that further research in interventions focusing on improving mood and QOL in patients with HNC is needed. Although many of these studies had positive outcomes, more than half had sample sizes of 60 of fewer patients, or were pilot in nature. In order to determine what interventions are most beneficial to our patients, we must move beyond pilot studies to well-designed, rigorous, RCTs that are adequately powered to detect meaningful improvements in short and long-term QOL and mood. We must also include caregivers in these studies where applicable. Once efficacy is clear, we must then determine how to best implement and disseminate the interventions health system-wide in order to reduce the significant physical and psychological burden of HNC for our patients and their caregivers.