User login
Feasibility and acceptance of a telehealth intervention to promote symptom management during treatment for head and neck cancer
Treatment for head and neck cancer is most often a rigorous regimen of combination therapies, producing a multitude of distressing symptoms and side effects. While it is nearly impossible to circumvent the physical and psychosocial insults caused by such treatment, some interventions directed toward educating and supporting patients during active treatment have met with success.[1], [2], [3] and [4] Conversely, other efforts have demonstrated little impact[5] and [6] or have been poorly received,7 pointing to the need for effective, acceptable means to provide support during such difficult treatment.
Over the past 10 years, telemedicine technology has enabled innovative approaches for improving patient education, assessment, support, and communication during treatment for both acute and chronic diseases. A recent policy white paper8 described telemedicine technology as including “the electronic acquisition, processing, dissemination, storage, retrieval, and exchange of information for the purpose of promoting health, preventing disease, treating the sick, managing chronic illness, rehabilitating the disabled, and protecting public health and safety” (p. 2). This same paper suggests that national telemedicine initiatives are essential to health-care reform based upon their proven cost–effectiveness and clinical efficacy. However, cost savings and clinical effectiveness will be unrealized outcomes if the interventions are not feasible in practice or acceptable to the targeted population.
In the arena of cancer care, telephone-based systems have been used to report and monitor cancer symptoms with favorable compliance noted even when patients are expected to initiate calls on a regular basis.[9], [10], [11] and [12] Favorable acceptance ratings have also been reported by both patients and clinicians regarding computerized systems used to assess symptoms and quality of life (QOL) in cancer patients.[13], [14], [15], [16], [17], [18] and [19] In the United Kingdom, a handheld computer system was successfully used to monitor and support patients receiving chemotherapy for lung or colorectal cancer,20 and a study testing a dialogic model of cancer care expecting patients to respond to telehealth messaging on a daily basis over 6 months reported an 84% cooperation rate.21 In these studies, the majority of patients reported ease of use and acceptability of the technology. Survey research has found both urban and rural cancer patients to be receptive to medical and psychiatric services provided via telehealth.22
Published reports describing use of telehealth and computerized interventions during head and neck cancer treatment are less prevalent. Touch-screen computers were successfully used in the Netherlands to collect QOL and distress data from head and neck cancer patients.16 Videoconferencing has been used successfully to overcome geographical barriers to patient assessment[23], [24] and [25] and to provide speech–language pathology services to people living with head and neck cancers in remote areas. Reported use of telehealth management appears promising for providing timely access to care for those who are geographically isolated.26
A research group based in the Netherlands developed and tested a comprehensive electronic health information support system for use in head and neck cancer care.27 The system had four patient-related functions: facilitating communication between patients and health-care providers, providing information about the disease and its treatment, connecting patients with other patients similarly diagnosed, and monitoring patients after hospital discharge. The system was found to be well-accepted and appreciated by participating patients, and its use enabled early identification and direct intervention for patient problems.27 A clinical trial of the telehealth application showed improved QOL in five of 22 studied parameters for the treatment group.28 However, 20 of the 59 patients eligible for the intervention group refused participation; 11 (55%) of these stated computer-related concerns as their reason for nonparticipation.
Knowing that head and neck cancer patients experience a high burden of illness and often have significant communication, socioeconomic, and geographic barriers to care, our team developed a telehealth intervention using a simple telemessaging device to circumvent communication barriers and perceived technical challenges associated with computer-based systems to provide education and support to patients in their own home and on their own time schedule.29 Overall, we hypothesized that patients receiving the intervention would experience less symptom distress, improved QOL, increased self-efficacy, and greater satisfaction with symptom management than those in the control group. However, as a first step toward examining the efficacy and effectiveness of this intervention, this study examined both quantitative and qualitative indicators of its feasibility and acceptance among patients undergoing treatment for head and neck cancer.
Methods
Design
Subsequent to study approval by the University of Louisville's Human Subjects Protection Office, a randomized clinical trial comparing the telehealth intervention to standard care was conducted using a two-group parallel design. This study reports on the intervention's feasibility and acceptance in the treatment group of 44 patients.
Site
Participants were recruited from patients receiving care from the Multidisciplinary Head and Neck Cancer Team at the James Graham Brown Cancer Center (JGBCC) over a 2-year period (June 2006 through June 2008). The team consisted of head and neck surgeons, medical oncologists, radiation oncologists, nurses, a pathologist, a speech therapist, a registered dietician, a psychologist, and a social worker. This team developed a comprehensive assessment and treatment plan during each patient's initial visit to the clinic and coordinated patient care throughout the treatment process.
Sample
Patients eligible for study participation met the following inclusion criteria: (1) initial diagnosis of head or neck cancer including cancers of the oral cavity, salivary glands, paranasal sinuses and nasal cavity, pharynx, and larynx; (2) involvement in a treatment plan including one or more modalities (ie, surgery, chemotherapy, radiation, or any combination); (3) capacity to give independent informed consent; and (4) ability to speak, read, and comprehend English at the eighth-grade level or above. Patients were excluded from participation if they had no land telephone line, had a thought disorder, were incarcerated, or had compromised cognitive functioning.
All patients scheduled for assessment received an explanation of the research study via print materials prior to their first clinic visit. During their first scheduled clinic visit, all patients identified as eligible were approached by a member of the research study staff, who briefly explained the study and asked if they might be interested in study participation. Because of the stress and content of this first clinic visit, interested patients were contacted later by phone to schedule an additional visit to review the study and obtain informed consent.
During the informed consent meeting, the study procedures were explained in detail. If the patient agreed and signed a consent form, a randomization grid which considered the patient's particular treatment plan was used to assign the patient to either the control or the experimental group. Baseline data were also collected during this first visit.
Description of the Intervention
The technology selected for implementing the intervention was the Health Buddy® System, a commercially available, proprietary system produced and maintained by Robert Bosch Healthcare Palo Alto, CA. The Health Buddy, the appliance used for interaction between the participant and the health-care provider, is a user-friendly, easily visible, electrical device that attaches to the user's land phone line (see Figure 1). Questions and information are displayed on the liquid crystal display (LCD) screen of the 6 × 9–inch appliance. The individual responds to questions by pressing one of the four large buttons below the screen. The research team selected the technology provider based on the ability of the technology to perform in accordance with the research objectives.
Symptom control algorithms developed using participatory action research (surveys of current and past patients and clinicians) and evidence-based practice were programmed into the telehealth messaging system (see article by Head et al,29 which details the algorithm topic selection and development process). The algorithms addressed 29 different symptoms and side effects of treatment, consisting of approximately 100 questions accompanied by related educational and supportive responses. Patients were asked three to five questions daily related to the symptoms anticipated during their treatment scenario. Depending upon their response, they would receive specific information related to symptom self-management, including recommendations as to when to contact their clinicians. The algorithms were constructed with the goal of encouraging self-efficacy and independent action on the part of the participant. See Figure 2 for an example of the branching algorithms.
Participants randomly assigned to the treatment group immediately had the Health Buddy connected to a land telephone line in their home. Most (40%) chose to place it in their kitchen, while another 26% placed it in their bedrooms; most often, it was in a highly visible location, serving to remind the participant to respond. Research study staff delivered, installed, and demonstrated how to operate the equipment. Installation was simple and required only minutes. A tutorial programmed into the Health Buddy taught participants how to reply to questions appearing on the monitor using the four large keys below the possible answers or a rating scale which would appear depending on the type of question asked.
During the early hours of the morning, the device would automatically call a toll-free number. Responses to the previous day's questions were uploaded, and questions and related information for the next day were downloaded over the telephone line onto a secure server. Phone service was never disrupted by the device; if the phone was in use, the system would connect later to retrieve and download information. Once new content was transferred, a green light on the device would flash to alert the participant that new questions were available for response. Once the participant pressed any of the keys, the new algorithms would begin appearing on the monitor screen.
Participants were instructed to begin responding on the first day they received treatment or on the first day after returning home from surgery. They were asked to continue responding daily (unless hospitalized for treatment) throughout the treatment period and for approximately 2 weeks posttreatment as treatment-induced symptoms continue during that period of time. Study staff contacted participants when treatment was complete and scheduled a date to pick up the appliance and end daily responding. Daily patient responses required 5–10 minutes.
Participant responses could be viewed by study staff via Internet access 1 day after being answered. Responses were monitored daily by study nurses. Symptoms unrelieved over time or symptoms targeted as requiring immediate intervention (ie, serious consideration of suicide) would result in the study nurse contacting the patient directly by phone and/or contacting clinicians to assure immediate intervention. However, it is important to note that this direct intervention by study staff was infrequent as most symptoms were addressed independently by the participant as desired. If a participant had not reported a period of planned hospitalization and did not respond for 3 consecutive days, study staff would contact the patient by phone to ascertain the reason for noncompliance.
Measures
The following indicators were selected as measures of acceptance (accrual rate), feasibility (utilization, nurse-initiated contacts), and/or satisfaction (satisfaction ratings). Narrative responses and a poststudy survey provided additional data examining acceptance, feasibility, and satisfaction with the intervention. In addition, demographic and medical information as well as measures assessing primary study outcomes were collected from each participant. Table 1 lists all measures and the study time point when they were administered.
| MEASURES | PRETREATMENT | DURING TREATMENT | POSTTREATMENT | CUMULATIVE |
|---|---|---|---|---|
| Demographics | X (baseline) | X | ||
| Accrual rate | X | |||
| Utilization rate | X | |||
| FACT-H&N | X (baseline) | X (mid-tx) | X (2–4 weeks post-tx) | |
| MSAS | X (baseline) | X (mid-tx) | X (2–4 weeks post-tx) | |
| Satisfaction with technology | X | |||
| Nurse-initiated contacts | X | |||
| Exit interview | X (end of treatment) | |||
| Poststudy written survey | X (60–90 days post-tx) |
tx = treatment
Accrual rate
The number of individuals assessed for study eligibility, reasons for exclusion or noncompletion, and numbers included in the analysis were all recorded to examine acceptance of the intervention and identify issues with the intervention or technology affecting participation.
Utilization
Feasibility was operationalized as device utilization using the percentage of days on which a participant responded to the Health Buddy. This was calculated using the number of days the participant responded to the telehealth device divided by the number of days the participant had the device and was expected to respond. These data were maintained and provided by the telehealth provider (Robert Bosch Healthcare).
Nurse-initiated contacts with participants and/or clinicians
The number of occasions on which a nurse decided to intervene was used as an indicator of feasibility under the premise that the goal of the intervention was to support and encourage patient-driven efforts to seek care for persistent or troubling symptoms. If a patient reported a symptom, he or she was given management information and encouraged to discuss problems further with the clinician either by phone or during clinic visits. If a patient continued to report an unresolved symptom or if the symptom required immediate intervention (ie, suicide threat), the research nurse reviewing responses would contact the patient and/or clinician to ascertain why and/or assist with its resolution. These nurse-initiated contacts should be infrequent if the intervention is achieving the goal of developing patient self-efficacy.
Satisfaction ratings
Items assessing satisfaction with the technology were also administered to participants via the telehealth messaging device. Questions related to satisfaction with the initial setup of the telehealth appliance were asked at the beginning of the intervention. Ongoing satisfaction with the device, messaging content, and the health-care provider were assessed every 90 days. The specific questions asked are detailed in Table 4.
Narrative data
Upon completion of the intervention, participants in the treatment group completed an exit interview using open-ended questions regarding the utility of the intervention, relevance of the algorithms, value or burden of item repetition in the algorithms, symptoms or problems experienced that were not addressed by the intervention, and general comments.
Poststudy survey
A final survey was mailed to participants several months after completion of the study, asking for additional feedback about the impact of the intervention. Specifically, participants (both treatment and control groups) were asked about their overall satisfaction with the treatment and services at the cancer center, their satisfaction with information received about their treatment, the response(s) received when they attempted contact with the health-care team after hours, the amount of support received, their current smoking and alcohol usage, and several demographic questions not earlier assessed or available through record review (years of education, highest degree, income range). Those receiving the intervention were also asked about the impact of the Health Buddy on their care and actions taken in response to the algorithms.
Demographic and medical information
Demographic information was collected using the initial survey, and information about the participant's medical history, condition, treatments received, treatment timing, complications, comorbidities, and treatment response was collected via retrospective medical record review subsequent to completion of the clinical trial.
Outcome measures
While outcomes of the clinical trial are not the subject of this article, the results of QOL and symptom burden measures for the treatment group only are included here because of their relationship with device utilization. The two measures included the Functional Assessment of Cancer Therapy–Head and Neck Scale and the Memorial Symptom Assessment Scale and were administered at baseline (before beginning treatment), mid-treatment, and posttreatment.
• Functional Assessment of Cancer Therapy–Head and Neck Scale (FACT-H&N). The FACT-G (general) is a multidimensional QOL instrument designed for use with all cancer patients. The instrument has 28 items divided into four subscales: Functional Well-Being, Physical Well-Being, Social Well-Being, and Emotional Well-Being. This generic core questionnaire was found to meet or exceed requirements for use in oncology based upon ease of administration, brevity, reliability, validity, and responsiveness to clinical change.30 Added to the core questionnaire is the head and neck–specific subscale, consisting of 11 items specific to this cancer site. A Trial Outcome Index (TOI) is also scored and is the result of the physical, functional, and cancer-specific subscales. List et al31 found the FACT-H&N to be reliable and sensitive to differences in functioning for patients with head and neck cancers (Cronbach's alpha was 0.89 for total FACT-G and 0.63 for the head and neck subscale in this study of 151 patients). Additionally, head and neck cancer patients found the FACT-H&N relevant to their problems and easy to understand, and it was preferred over other validated head and neck cancer QOL questionnaires.32 The FACT-H&N was chosen for this study because it (1) is nonspecific related to a treatment modality or subsite among head and neck cancers, (2) allows comparison across cancer diagnoses while still probing issues specific to head and neck cancer, (3) is short and can be completed quickly, (4) includes the psychosocial domains of social/family and emotion subscales as well as physical and functional areas, and (5) is self-administered.
• Memorial Symptom Assessment Scale (MSAS). This multidimensional scale measures the prevalence, severity, and distress associated with the most common symptoms experienced by cancer patients. Physical and emotional subscale scores as well as a Global Distress Index (GDI, considered to be a measure of total symptom burden) can be generated from patient responses. The MSAS has demonstrated validity and reliability in both in- and outpatient cancer populations.[33], [34] and [35] Initial psychometric analysis by Portenoy et al34 used factor analysis to define two subscales: psychological symptoms and physical symptoms with Cronbach alpha coefficients of 0.88 and 0.83, respectively; convergent validity was also established. It was chosen for this study because of its proven ability to measure both the presence and the intensity of experienced symptoms.[33], [35], [36], [37] and [38]
Data Analysis
Quantitative data were documented and analyzed using the Statistical Package for the Social Sciences (SPSS, Inc., Chicago, IL), version 16. Descriptive statistics were calculated to describe the sample and assess study outcomes, including feasibility and acceptability of the intervention. To ascertain relationships between usage of the device and demographic and medical information, a series of correlational analyses using Spearman's rho were conducted. This nonparametric test was chosen over Pearson's r because of the small sample size, the lack of a normal distribution for several of the variables, and the ordinal nature of several of the variables. Multiple regression analyses were also planned, but lack of significant bivariate correlations precluded multivariate analysis.
Qualitative responses to open-ended questions were analyzed to identify themes and direct quotations illustrating those themes.
Descriptive analysis of the treatment group's responses to the outcome measures (QOL and symptom burden) was done to ascertain changes over the course of the intervention using the mean scores at baseline, during treatment, and posttreatment.
Results
Description of Participants
Participants randomly assigned to the intervention group (n = 45) were an average age of 59 years (±11.7), and most were covered by private (34%) or public (48%) insurance. On average, participants had completed 13.5 years (±3.0) of formal education. Thirty-nine (87%) of the participants were male and 91% were Caucasian.
With regard to medical information, participants were predominantly diagnosed with stage II cancers of the head and neck (36%). The most prevalent site was the larynx (12 patients), followed by the tongue and the base of the tongue (seven patients) and unknown primary (seven patients). The vast majority received chemotherapy (32, or 71%) and/or radiation (42, or 93%).
Additional details regarding demographic and medical characteristics of the sample are provided in Table 2.
| FREQUENCY | VALID PERCENT | |
|---|---|---|
| Gender (n = 44) | ||
| Male | 39 | 88.6 |
| Female | 5 | 11.3 |
| Race (n = 44) | ||
| Caucasian | 40 | 90.9 |
| African American | 4 | 9.0 |
| Tumor stage (n = 44) | ||
| I | 7 | 15.9 |
| II | 15 | 34.0 |
| III | 11 | 25.0 |
| IV | 4 | 9.0 |
| Unable to determine | 5 | 11.4 |
| Unknown | 2 | 4.5 |
| Site of cancer (n = 44) | ||
| Larynx | 12 | 27.2 |
| Tongue, base of tongue | 7 | 15.9 |
| Unknown primary | 7 | 15.9 |
| Tonsillar | 4 | 9.0 |
| Other H&N sites | 14 | 31.8 |
| Insurance status (n = 44) | ||
| No insurance | 8 | 18.2 |
| Medicaid | 1 | 2.3 |
| Medicare | 2 | 4.5 |
| Medicaid and Medicare | 1 | 2.3 |
| Medicare and supplement | 9 | 20.5 |
| Medicare and VA benefits | 2 | 4.5 |
| Veteran benefits only | 6 | 13.6 |
| Private insurance | 15 | 34.1 |
| Highest educational degree (n = 20)a | ||
| Less than high school | 3 | 15.0 |
| High school or GED | 9 | 45.0 |
| Associate's/bachelor's degree | 4 | 20.0 |
| Masters, PhD, or MD | 2 | 10.0 |
| Other | 2 | 10.0 |
| Income range (n = 18)a | ||
| $20,000 or less | 5 | 27.8 |
| $20,001–50,000 | 5 | 27.8 |
| $70,001–100,000 | 5 | 27.8 |
| Over $100,000 | 3 | 16.7 |
| Percent of poverty in zip code area (n = 44) | ||
| 2.8–5.1% | 11 | 25.0 |
| 5.9–8.6% | 11 | 25.0 |
| 9.0–11.9% | 10 | 22.7 |
| 12.3–45.9% | 12 | 27.2 |
Feasibility and Acceptability
Accrual rate
Of the 185 patients assessed for eligibility during the 2-year recruitment period, 105 were excluded. See Figure 3 for a detailed depiction of study accrual for both the treatment and control groups. Thirty-three (31%) were excluded because they did not have a land phone line, a requirement for transmitting the algorithms to the Health Buddy appliance. Most of these had cell phones only. No potential participants refused participation due to issues related to operation of the technology itself.
Device utilization
Participants used the telehealth device for an average of 70.7 days (±26.7), which constituted 86.3% (±15.0) of the total days available for use. Of note, the median percentage of use was 94.2% and the modal percentage was 100%, indicating that the vast majority of participants consistently used the telehealth device. The participant with the lowest usage rate used the device 46% of the days available.
By far, the most common reason for Health Buddy nonresponse was patient hospitalization. Two subjects traveled out of town frequently on weekends and would leave the Health Buddy at home. One subject had accidentally unhooked the Health Buddy, and a home visit was made to reconnect the device into the patient's phone line.
Nurse-initiated contacts with participants and/or clinicians
Of the 45 enrolled patients, 33 required additional contact with a research nurse (see Table 3). The most common reasons patients were contacted were nonresponse for 3 consecutive days (38.3%), repeated reporting of high levels of unrelieved pain (30%), and suicidal thoughts (10%). In all, 120 calls were placed: one call for every 25.9 response days. In every case, the problem was resolved.
| NUMBER OF PATIENTS | PROBLEM | OUTGOING CALLS | RESOLUTION |
|---|---|---|---|
| 15 | No response on Health Buddy for 3 consecutive days | 46 | Patient teaching |
| 17 | Pain-related issues | 36 | Advocacy/referral/patient teaching |
| 5 | Suicidal thoughts | 12 | Advocacy/referral |
| 7 | G-tube problems | 8 | Patient teaching |
| 5 | Sadness/depression | 6 | Advocacy/referral |
| 3 | Multiple symptoms | 3 | Advocacy/referral |
| 3 | Nausea/vomiting | 4 | Referral/patient teaching |
| 2 | Coughing/excessive secretions | 2 | Patient teaching |
| 2 | Constipation | 2 | Patient teaching |
| 1 | Stomatitis | 1 | Referral |
Satisfaction ratings
Responses to surveys programmed into the Health Buddy system are displayed in Table 4. Overall, respondents responded favorably, finding the installation to be easy, the content to be helpful, and the overall experience to be positive.
| PERCENT OF RESPONDENTS | |
|---|---|
| Installation satisfaction | |
| Installation problems? | |
| Yes | 2 |
| No | 98 |
| Any difficulty completing the first training questions? | |
| Yes | 7 |
| No | 94 |
| Length of installation? | |
| 2–5 minutes | 52 |
| 6–10 minutes | 41 |
| 11–15 minutes | 4 |
| 16–20 minutes | 2 |
| Content satisfaction | |
| Overall, I think the Health Buddy questions are | |
| Very easy | 44 |
| Somewhat easy | 16 |
| Neutral | 32 |
| Somewhat difficult | 4 |
| Difficult | 4 |
| Repeating questions reinforced knowledge and understanding | |
| Strongly agree | 56 |
| Somewhat agree | 28 |
| Neutral | 12 |
| Somewhat disagree | 4 |
| Strongly disagree | 0 |
| Understanding of my health condition | |
| Much better | 64 |
| Somewhat better | 20 |
| Neutral | 16 |
| Somewhat worse | 0 |
| Much worse | 0 |
| Managing my health condition | |
| Much better | 52 |
| Somewhat better | 44 |
| Neutral | 4 |
| Somewhat worse | 0 |
| Much worse | 0 |
| Recommend the device to others | |
| Very willing | 80 |
| Somewhat willing | 12 |
| Neutral | 4 |
| Somewhat unwilling | 0 |
| Very unwilling | 4 |
| Overall satisfaction | |
| Satisfaction with device | |
| Very satisfied | 45 |
| Satisfied | 35 |
| Somewhat satisfied | 15 |
| Not very satisfied | 5 |
| Satisfaction with the communication between you and your doctor or nurse | |
| More satisfied | 65 |
| No difference | 30 |
| Less satisfied | 5 |
| Ease of using the device | |
| Very easy | 85 |
| Easy | 15 |
| Not easy | 0 |
| Overall experience with the device | |
| Positive | 85 |
| Neutral | 15 |
| Negative | 0 |
| Continue to use the device | |
| Very likely | 40 |
| Likely | 40 |
| Somewhat likely | 15 |
| Not very likely | 0 |
Narrative comments
During the exit interview, participants were asked, “How was having the Health Buddy helpful to you?” Responses could be categorized into two major themes: (1) the Health Buddy provided needed information and (2) the Health Buddy improved my self-management during treatment.
Statements made related to the information provided included the following:
- • It gave me information on what could be expected from treatment
• It was a constant reminder of things to watch for
• It kept me abreast of my total condition at all times
• It gave good directions so I didn't have to ask at the cancer center
• It gave good suggestions on treatments (home remedies) such as gargles, care of feeding tube, exhaustion, and everyday symptoms
Statements made indicative that the Health Buddy improved self-management included the following:
- • I learned what I could do to make myself feel better
• It helped me manage my symptoms
• It taught me about symptom management and how to handle problems
• It let me know whether to contact a doctor or use self-care
• It gave me who to call for problems and some things to try
• It kept me aware of what I needed to do in order to make the period easier
• It reminded me to take my meds and exercise
Additionally, some participants noted the support they felt from having the Health Buddy interventions during treatment in saying the following:
- • It kind of helped my depression through acknowledging it and giving me something to do
• It made me feel I was not the only one who had experience with these things
• It comforted me because I knew what was going to happen
Poststudy survey
Twenty (45%) of the 44 patients who received the intervention responded to the mailed poststudy survey. When asked if they felt they received better care because they had the device, 13 of the 20 (65%) responded that they did. Eighteen (90%) of the treatment group responders stated they were very satisfied with their care (one stated “somewhat satisfied”) and 20 (100%) said they would recommend the cancer center for treatment. Nineteen (95%) stated they received adequate support during treatment.
Outcome Measures
Mean scores on the FACT-H&N and subscales and the MSAS and subscales taken pre-, during, and posttreatment are displayed in Table 5. As expected, average QOL scores declined during treatment, while symptom distress increased, with a return to near baseline scores posttreatment.
| SCALE/SUBSCALE | PRETREATMENT | DURING TREATMENT | POSTTREATMENT |
|---|---|---|---|
| Total FACT-H&N | 100.3 | 85.6 | 101.5 |
| FACT-G | 74.3 | 69.4 | 78.5 |
| Trial Outcome Index | 62.6 | 46.0 | 65.0 |
| Physical Well-Being | 21.2 | 17.6 | 21.1 |
| Functional Well-Being | 15.6 | 12.5 | 17.4 |
| Emotional Well-Being | 21.1 | 22.3 | 22.2 |
| Social Well-Being | 21.1 | 22.3 | 22.2 |
| Total MSAS | 0.7 | 1.1 | 0.8 |
| Global Distress Index | 1.1 | 1.8 | 1.3 |
| Physical | 0.7 | 1.5 | 1.1 |
| Psychological | 1.1 | 1.2 | 0.8 |
Correlations
The relationships between percentage usage per patient and the following variables were evaluated: age, income, years of education, tumor stage, and percent poverty in patient's zip code. Percent poverty in zip code area was intended to be a surrogate measure of the patient's socioeconomic status. Results are displayed in Table 6. No significant correlations were noted, although years of education and percentage poverty in zip code showed a trend toward significance.
| VARIABLE (VS % USAGE) | RELATIONSHIP | |
|---|---|---|
| SPEARMAN'S RHO RS | SIGNIFICANCE (ONE-TAILED) | |
| Percent poverty in zip code | 0.213 | 0.083 |
| Age | 0.146 | 0.173 |
| Years of education | −0.325 | 0.081 |
| Income | −0.292 | 0.120 |
| Tumor stage | 0.196 | 0.122 |
| Physical Well-Being (during treatment) | 0.310 | 0.048 |
| Emotional Well-Being (during treatment) | 0.315 | 0.042 |
Although a multivariate model was planned, the lack of significant bivariate correlations precluded the need for multivariate analysis.
When percent usage was correlated with FACT-H&N total and subscales taken at baseline, during active treatment, and posttreatment, significant positive correlations were found between the percentage used and the Physical Well-Being subscale score during treatment (Spearman's rho = 0.310, P = 0.048) and between percentage used and the Emotional Well-Being subscale during treatment (Spearman's rho = 0.315, P = 0.042).
There were no significant correlations between percentage usage and the scores on the MSAS.
Discussion
Both qualitative and quantitative measures indicate that using telehealth to support symptom management during aggressive cancer treatment is both feasible and well-accepted. Patient users were not intimidated by this particular technology as it was simple to set up and use and required no previous computer training to operate. The Health Buddy was viewed as providing important and useful information. Overall, users felt that it improved their ability to self-manage their disease and the side effects of treatment and provided a sense of support and security.
Unlike other studies which use telehealth devices to monitor patient symptoms, our goal was to increase patient self-management of the symptoms experienced during intensive medical treatment, therefore avoiding increased burden on the medical system. The fact that the research nurse overseeing the responses needed to intervene only once every 25.9 days speaks to the ability of the intervention to have a positive impact on utilization of medical services.
The lack of significant relationships between usage and descriptive variables such as age and years of education suggests that the intervention was equally acceptable to all subgroups. Factors such as age, previous computer literacy, educational obtainment, and socioeconomic status did not significantly differentiate our study population in terms of compliance as verified by usage percentages.
The significant relationships found between the percentage used and the subscale scores on Physical Well-Being and Emotional Well-Being during treatment may indicate that increased use of the telemessaging intervention during treatment resulted in better physical and emotional aspects of QOL.
The high rate of daily compliance with the intervention in spite of differentiating personal variables and the severity of the treatment regimen may have been due to one or a combination of the following factors:
- • the simplicity of the technology
• the visibility of the appliance (often placed in the kitchen or living area of the home) and its flashing green light as cues to the need to respond
• the usefulness of the information provided
• the use of simple messaging language presented in an encouraging, positive manner
• affirmations related to application of the symptom management protocols suggested
• curiosity related to the day's messaging and the motivational saying which always appeared at the end
• knowledge that someone was reviewing the responses, tracking and intervening when the participant did not respond for several days
Our study supports the benefits of telehealth interventions noted by providers in a study by Sandberg et al39: opportunities for more frequent contact, greater relaxation and information due to the ability to interact in one's own home, increased accessibility by those frequently underserved, and timely medical information and monitoring. Similar to the study done in the Netherlands,[27] and [28] this study noted technological problems as the primary disadvantage; but in our intervention, we had no problems with the technology or equipment.
Although computerized technology served as a barrier to previous telehealth research, the lack of a land-based phone line was a factor preventing participation in the current study. Indeed, many participants maintained only wireless communication devices, which were not compatible with the version of the Health Buddy that was employed in this study. However, improvements in the technology since completion of this study now allow for wireless access to the appliance or provision of an independent wireless messaging device for those without such access in their own homes.
Although the data generally support the feasibility and acceptability of the telehealth-based intervention, the results should be interpreted in the context of a few study limitations. In particular, the sample size was somewhat small, and data pertaining to the socioeconomic status of participants were not available for all participants. Second, the study did not include measures of the patient's direct interactions with health-care providers during the study or specific data related to their health-care utilization (eg, emergency room visits, preventable inpatient hospitalizations, emergency calls to clinicians). The collection of more exhaustive measures of health-care utilization was limited by resources but is planned for subsequent studies. Finally, concerns regarding subject burden limited assessment of the usability of the telehealth device.
Although compliance with utilization expectations and completion of study measures was excellent during the course of the intervention, response to the follow-up survey mailed several months later was less than 50%. This low response rate was most likely due to several factors: (1) this survey was sent at the conclusion of the entire study (by this time, patients were 0–21 months past their active participation); (2) it was a mailed survey with no additional contact or follow-up effort to increase response rate; (3) participants may have felt that they had already shared their opinions in the exit interview and may have felt overburdened by study measures at this point; and (4) participants may have died, moved, or been medically unable to respond. This lack of response did limit our ability to evaluate the longitudinal impact of the intervention.
Conclusions
This telehealth intervention proved to be an acceptable and feasible means to educate and support patients during aggressive treatment for head and neck cancer. Patient compliance with telehealth interventions during periods of extreme symptom burden and declining QOL is feasible if simple technology cues the patient to participate, offers positive support and relevant education, and is targeted or tailored to their specific condition.
1 C.D. Llewellyn, M. McGurk and J. Weinman, Are psycho-social and behavioural factors related to health related-quality of life in patients with head and neck cancer?: A systematic review, Oral Oncol 41 (5) (2005), pp. 440–454. Article | 
2 K.T. Vakharia, M.J. Ali and S.J. Wang, Quality-of-life impact of participation in a head and neck cancer support group, Otolaryngol Head Neck Surg 136 (3) (2007), pp. 405–410. Article |
3 P.J. Allison et al., Results of a feasibility study for a psycho-educational intervention in head and neck cancer, Psychooncology 13 (2004), pp. 482–485. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (22)
4 L.H. Karnell et al., Influence of social support on health-related quality of life outcomes in head and neck cancer, Head Neck 29 (2) (2007), pp. 143–146. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (19)
5 K.M. Petruson, E.M. Silander and E.B. Hammerlid, Effects of psychosocial intervention on quality of life in patients with head and neck cancer, Head Neck 25 (7) (2003), pp. 576–584. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (29)
6 J.R.J. deLeeuw et al., Negative and positive influences of social support on depression in patients with head and neck cancer: a prospective study, Psychooncology 9 (2000), pp. 20–28. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (46)
7 J. Ostroff et al., Interest in and barriers to participation in multiple family groups among head and neck cancer survivors and their primary family caregivers, Fam Process 43 (2) (2004), pp. 195–208. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (17)
8 R.L. Bashshur et al., National telemedicine initiatives: essential to healthcare reform, Telemed J E Health 15 (6) (2009), pp. 1–11.
9 K. Davis et al., An innovative symptom monitoring tool for people with advanced lung cancer: a pilot demonstration, J Support Oncol 5 (8) (2007), pp. 381–387. View Record in Scopus | Cited By in Scopus (8)
10 K.H. Mooney et al., Telephone-linked care for cancer symptom monitoring: A pilot study, Cancer Pract 10 (3) (2002), pp. 147–154. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (30)
11 R.H. Friedman et al., The virtual visit: using telecommunications technology to take care of patients, J Am Med Inform Assoc 4 (1997), pp. 413–425. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (64)
12 A. Weaver et al., Application of mobile phone technology for managing chemotherapy-associated side-effects, Ann Oncol 18 (11) (2007), pp. 1887–1892. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (10)
13 D. Berry et al., Computerized symptom and quality-of-life assessment for patients with cancer: Part 1: Development and pilot testing, Oncol Nurs Forum 31(5 (2004), pp. E75–E83. Full Text via CrossRef
14 K. Mullen, D. Berry and B. Zierler, Computerized symptom and quality-of-life assessment for patients with cancer: Part II: Acceptability and usability, Oncol Nurs Forum 31 (5) (2004), pp. E84–E89. Full Text via CrossRef
15 B. Fortner et al., The Cancer Care Monitor: psychometric content evaluation and pilot testing of a computer administered system for symptom screening and quality of life in adult cancer patients, J Pain Symptom Manage 26 (6) (2003), pp. 1077–1092. Article |
16 R. de Bree et al., Touch screen computer-assisted health-related quality of life and distress data collection in head and neck cancer patients, Clin Otolaryngol 33 (2) (2008), pp. 138–142. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (6)
17 H. Huang et al., Developing a computerized data collection and decision support system for cancer pain management, Comput Inform Nurs 21 (4) (2003), pp. 206–217. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (17)
18 D.J. Wilkie et al., Usability of a computerized pain report in the general public with pain and people with cancer pain, J Pain Symptom Manage 25 (3) (2003), pp. 213–224. Article |
19 K. Kroenke et al., Effect of telecare management on pain and depression in patients with cancer: a randomized trial, JAMA 304 (2) (2010), pp. 163–171. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (7)
20 N. Kearney et al., Utilizing handheld computers to monitor and support patients receiving chemotherapy: results of a UK-based feasibility study, Support Care Cancer 14 (7) (2006), pp. 742–752. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (17)
21 N.R. Chumbler et al., Remote patient–provider communication and quality of life: empirical test of a dialogic model of cancer care, J Telemed Telecare 13 (2007), pp. 20–25. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (5)
22 A.L. Grubaugh et al., Attitudes toward medical and mental health care delivered via telehealth applications among rural and urban primary care patients, J Nerv Ment Dis 196 (2) (2008), pp. 166–170. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (6)
23 J. Stalfors et al., Accuracy of tele-oncology compared with face-to-face consultation in head and neck cancer case conferences, J Telemed Telecare 7 (2001), pp. 338–343. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (5)
24 C. Dorrian et al., Head and neck cancer assessment by flexible endoscopy and telemedicine, J Telemed Telecare 15 (2009), pp. 118–121. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (3)
25 J. Stalfors et al., Haptic palpation of head and neck cancer patients—implications for education and telemedicine, Stud Health Technol Inform 81 (2001), pp. 471–474. View Record in Scopus | Cited By in Scopus (8)
26 C. Myers, Telehealth applications in head and neck oncology, J Speech Lang Pathol Audiol 29 (3) (2005), pp. 125–127.
27 J.L. van den Brink et al., Involving the patient: a prospective study on use, appreciation and effectiveness of an information system in head and neck cancer care, Int J Med Inform 74 (10) (2005), pp. 839–849. Article |
28 J.L. van den Brink et al., Impact on quality of life of a telemedicine system supporting head and neck cancer patients: a controlled trial during the postoperative period at home, J Am Med Inform Assoc 14 (2) (2007), pp. 198–205. Article |
29 B. Head et al., Development of a telehhealth intervention for head and neck cancer patients, Telemed J E Health 15 (1) (2009), pp. 100–108. View Record in Scopus | Cited By in Scopus (1)
30 D.F. Cella et al., The Functional Assessment of Cancer Therapy (FACT) scale: development and validation of the general measure, J Clin Oncol 11 (3) (1993), pp. 570–579. View Record in Scopus | Cited By in Scopus (1626)
31 M.A. List et al., The Performance Status scale for head and neck cancer patients and the Functional Assessment of Cancer Therapy-Head and Neck scale: A study of utility and validity, Cancer 77 (11) (1996), pp. 2294–2301. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (169)
32 H.M. Mehanna and R.P. Morton, Patients' views on the utility of quality of life questionnaires in head and neck cancer: a randomised trial, Clin Otolaryngol 31 (4) (2006), pp. 310–316. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (10)
33 V. Chang et al., The Memorial Symptom Assessment Scale short form, Cancer 89 (2000), pp. 1162–1171. Full Text via CrossRef
34 R.K. Portenoy et al., The Memorial Symptom Assessment Scale: an instrument for the evaluation of symptom prevalence, characteristics and distress, Eur J Cancer 30A (9) (1994), pp. 1326–1336. Abstract |
35 J.E. Tranmer et al., Measuring the symptom experience of seriously ill cancer and noncancer hospitalized patients near the end of life with the Memorial Symptom Assessment Scale, J Pain Symptom Manage 25 (5) (2003), pp. 420–429. Article |
36 V.T. Chang et al., Symptom and quality of life survey of medical oncology patients at a Veterans Affairs medical center: a role for symptom assessment, Cancer 88 (5) (2000), pp. 1175–1183. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (121)
37 J.F. Nelson et al., The symptom burden of chronic critical illness, Crit Care Med 32 (7) (2004), pp. 1527–1534. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (55)
38 L.B. Harrison et al., Detailed quality of life assessment in patients treated with primary radiotherapy for squamous cell cancer of the base of the tongue, Head Neck 19 (3) (1997), pp. 169–175. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (132)
39 J. Sandberg et al., A qualitative study of the experiences and satisfaction of direct telemedicine providers in diabetes case management, Telemed J E Health 15 (8) (2009), pp. 742–750. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (1)
Treatment for head and neck cancer is most often a rigorous regimen of combination therapies, producing a multitude of distressing symptoms and side effects. While it is nearly impossible to circumvent the physical and psychosocial insults caused by such treatment, some interventions directed toward educating and supporting patients during active treatment have met with success.[1], [2], [3] and [4] Conversely, other efforts have demonstrated little impact[5] and [6] or have been poorly received,7 pointing to the need for effective, acceptable means to provide support during such difficult treatment.
Over the past 10 years, telemedicine technology has enabled innovative approaches for improving patient education, assessment, support, and communication during treatment for both acute and chronic diseases. A recent policy white paper8 described telemedicine technology as including “the electronic acquisition, processing, dissemination, storage, retrieval, and exchange of information for the purpose of promoting health, preventing disease, treating the sick, managing chronic illness, rehabilitating the disabled, and protecting public health and safety” (p. 2). This same paper suggests that national telemedicine initiatives are essential to health-care reform based upon their proven cost–effectiveness and clinical efficacy. However, cost savings and clinical effectiveness will be unrealized outcomes if the interventions are not feasible in practice or acceptable to the targeted population.
In the arena of cancer care, telephone-based systems have been used to report and monitor cancer symptoms with favorable compliance noted even when patients are expected to initiate calls on a regular basis.[9], [10], [11] and [12] Favorable acceptance ratings have also been reported by both patients and clinicians regarding computerized systems used to assess symptoms and quality of life (QOL) in cancer patients.[13], [14], [15], [16], [17], [18] and [19] In the United Kingdom, a handheld computer system was successfully used to monitor and support patients receiving chemotherapy for lung or colorectal cancer,20 and a study testing a dialogic model of cancer care expecting patients to respond to telehealth messaging on a daily basis over 6 months reported an 84% cooperation rate.21 In these studies, the majority of patients reported ease of use and acceptability of the technology. Survey research has found both urban and rural cancer patients to be receptive to medical and psychiatric services provided via telehealth.22
Published reports describing use of telehealth and computerized interventions during head and neck cancer treatment are less prevalent. Touch-screen computers were successfully used in the Netherlands to collect QOL and distress data from head and neck cancer patients.16 Videoconferencing has been used successfully to overcome geographical barriers to patient assessment[23], [24] and [25] and to provide speech–language pathology services to people living with head and neck cancers in remote areas. Reported use of telehealth management appears promising for providing timely access to care for those who are geographically isolated.26
A research group based in the Netherlands developed and tested a comprehensive electronic health information support system for use in head and neck cancer care.27 The system had four patient-related functions: facilitating communication between patients and health-care providers, providing information about the disease and its treatment, connecting patients with other patients similarly diagnosed, and monitoring patients after hospital discharge. The system was found to be well-accepted and appreciated by participating patients, and its use enabled early identification and direct intervention for patient problems.27 A clinical trial of the telehealth application showed improved QOL in five of 22 studied parameters for the treatment group.28 However, 20 of the 59 patients eligible for the intervention group refused participation; 11 (55%) of these stated computer-related concerns as their reason for nonparticipation.
Knowing that head and neck cancer patients experience a high burden of illness and often have significant communication, socioeconomic, and geographic barriers to care, our team developed a telehealth intervention using a simple telemessaging device to circumvent communication barriers and perceived technical challenges associated with computer-based systems to provide education and support to patients in their own home and on their own time schedule.29 Overall, we hypothesized that patients receiving the intervention would experience less symptom distress, improved QOL, increased self-efficacy, and greater satisfaction with symptom management than those in the control group. However, as a first step toward examining the efficacy and effectiveness of this intervention, this study examined both quantitative and qualitative indicators of its feasibility and acceptance among patients undergoing treatment for head and neck cancer.
Methods
Design
Subsequent to study approval by the University of Louisville's Human Subjects Protection Office, a randomized clinical trial comparing the telehealth intervention to standard care was conducted using a two-group parallel design. This study reports on the intervention's feasibility and acceptance in the treatment group of 44 patients.
Site
Participants were recruited from patients receiving care from the Multidisciplinary Head and Neck Cancer Team at the James Graham Brown Cancer Center (JGBCC) over a 2-year period (June 2006 through June 2008). The team consisted of head and neck surgeons, medical oncologists, radiation oncologists, nurses, a pathologist, a speech therapist, a registered dietician, a psychologist, and a social worker. This team developed a comprehensive assessment and treatment plan during each patient's initial visit to the clinic and coordinated patient care throughout the treatment process.
Sample
Patients eligible for study participation met the following inclusion criteria: (1) initial diagnosis of head or neck cancer including cancers of the oral cavity, salivary glands, paranasal sinuses and nasal cavity, pharynx, and larynx; (2) involvement in a treatment plan including one or more modalities (ie, surgery, chemotherapy, radiation, or any combination); (3) capacity to give independent informed consent; and (4) ability to speak, read, and comprehend English at the eighth-grade level or above. Patients were excluded from participation if they had no land telephone line, had a thought disorder, were incarcerated, or had compromised cognitive functioning.
All patients scheduled for assessment received an explanation of the research study via print materials prior to their first clinic visit. During their first scheduled clinic visit, all patients identified as eligible were approached by a member of the research study staff, who briefly explained the study and asked if they might be interested in study participation. Because of the stress and content of this first clinic visit, interested patients were contacted later by phone to schedule an additional visit to review the study and obtain informed consent.
During the informed consent meeting, the study procedures were explained in detail. If the patient agreed and signed a consent form, a randomization grid which considered the patient's particular treatment plan was used to assign the patient to either the control or the experimental group. Baseline data were also collected during this first visit.
Description of the Intervention
The technology selected for implementing the intervention was the Health Buddy® System, a commercially available, proprietary system produced and maintained by Robert Bosch Healthcare Palo Alto, CA. The Health Buddy, the appliance used for interaction between the participant and the health-care provider, is a user-friendly, easily visible, electrical device that attaches to the user's land phone line (see Figure 1). Questions and information are displayed on the liquid crystal display (LCD) screen of the 6 × 9–inch appliance. The individual responds to questions by pressing one of the four large buttons below the screen. The research team selected the technology provider based on the ability of the technology to perform in accordance with the research objectives.
Symptom control algorithms developed using participatory action research (surveys of current and past patients and clinicians) and evidence-based practice were programmed into the telehealth messaging system (see article by Head et al,29 which details the algorithm topic selection and development process). The algorithms addressed 29 different symptoms and side effects of treatment, consisting of approximately 100 questions accompanied by related educational and supportive responses. Patients were asked three to five questions daily related to the symptoms anticipated during their treatment scenario. Depending upon their response, they would receive specific information related to symptom self-management, including recommendations as to when to contact their clinicians. The algorithms were constructed with the goal of encouraging self-efficacy and independent action on the part of the participant. See Figure 2 for an example of the branching algorithms.
Participants randomly assigned to the treatment group immediately had the Health Buddy connected to a land telephone line in their home. Most (40%) chose to place it in their kitchen, while another 26% placed it in their bedrooms; most often, it was in a highly visible location, serving to remind the participant to respond. Research study staff delivered, installed, and demonstrated how to operate the equipment. Installation was simple and required only minutes. A tutorial programmed into the Health Buddy taught participants how to reply to questions appearing on the monitor using the four large keys below the possible answers or a rating scale which would appear depending on the type of question asked.
During the early hours of the morning, the device would automatically call a toll-free number. Responses to the previous day's questions were uploaded, and questions and related information for the next day were downloaded over the telephone line onto a secure server. Phone service was never disrupted by the device; if the phone was in use, the system would connect later to retrieve and download information. Once new content was transferred, a green light on the device would flash to alert the participant that new questions were available for response. Once the participant pressed any of the keys, the new algorithms would begin appearing on the monitor screen.
Participants were instructed to begin responding on the first day they received treatment or on the first day after returning home from surgery. They were asked to continue responding daily (unless hospitalized for treatment) throughout the treatment period and for approximately 2 weeks posttreatment as treatment-induced symptoms continue during that period of time. Study staff contacted participants when treatment was complete and scheduled a date to pick up the appliance and end daily responding. Daily patient responses required 5–10 minutes.
Participant responses could be viewed by study staff via Internet access 1 day after being answered. Responses were monitored daily by study nurses. Symptoms unrelieved over time or symptoms targeted as requiring immediate intervention (ie, serious consideration of suicide) would result in the study nurse contacting the patient directly by phone and/or contacting clinicians to assure immediate intervention. However, it is important to note that this direct intervention by study staff was infrequent as most symptoms were addressed independently by the participant as desired. If a participant had not reported a period of planned hospitalization and did not respond for 3 consecutive days, study staff would contact the patient by phone to ascertain the reason for noncompliance.
Measures
The following indicators were selected as measures of acceptance (accrual rate), feasibility (utilization, nurse-initiated contacts), and/or satisfaction (satisfaction ratings). Narrative responses and a poststudy survey provided additional data examining acceptance, feasibility, and satisfaction with the intervention. In addition, demographic and medical information as well as measures assessing primary study outcomes were collected from each participant. Table 1 lists all measures and the study time point when they were administered.
| MEASURES | PRETREATMENT | DURING TREATMENT | POSTTREATMENT | CUMULATIVE |
|---|---|---|---|---|
| Demographics | X (baseline) | X | ||
| Accrual rate | X | |||
| Utilization rate | X | |||
| FACT-H&N | X (baseline) | X (mid-tx) | X (2–4 weeks post-tx) | |
| MSAS | X (baseline) | X (mid-tx) | X (2–4 weeks post-tx) | |
| Satisfaction with technology | X | |||
| Nurse-initiated contacts | X | |||
| Exit interview | X (end of treatment) | |||
| Poststudy written survey | X (60–90 days post-tx) |
tx = treatment
Accrual rate
The number of individuals assessed for study eligibility, reasons for exclusion or noncompletion, and numbers included in the analysis were all recorded to examine acceptance of the intervention and identify issues with the intervention or technology affecting participation.
Utilization
Feasibility was operationalized as device utilization using the percentage of days on which a participant responded to the Health Buddy. This was calculated using the number of days the participant responded to the telehealth device divided by the number of days the participant had the device and was expected to respond. These data were maintained and provided by the telehealth provider (Robert Bosch Healthcare).
Nurse-initiated contacts with participants and/or clinicians
The number of occasions on which a nurse decided to intervene was used as an indicator of feasibility under the premise that the goal of the intervention was to support and encourage patient-driven efforts to seek care for persistent or troubling symptoms. If a patient reported a symptom, he or she was given management information and encouraged to discuss problems further with the clinician either by phone or during clinic visits. If a patient continued to report an unresolved symptom or if the symptom required immediate intervention (ie, suicide threat), the research nurse reviewing responses would contact the patient and/or clinician to ascertain why and/or assist with its resolution. These nurse-initiated contacts should be infrequent if the intervention is achieving the goal of developing patient self-efficacy.
Satisfaction ratings
Items assessing satisfaction with the technology were also administered to participants via the telehealth messaging device. Questions related to satisfaction with the initial setup of the telehealth appliance were asked at the beginning of the intervention. Ongoing satisfaction with the device, messaging content, and the health-care provider were assessed every 90 days. The specific questions asked are detailed in Table 4.
Narrative data
Upon completion of the intervention, participants in the treatment group completed an exit interview using open-ended questions regarding the utility of the intervention, relevance of the algorithms, value or burden of item repetition in the algorithms, symptoms or problems experienced that were not addressed by the intervention, and general comments.
Poststudy survey
A final survey was mailed to participants several months after completion of the study, asking for additional feedback about the impact of the intervention. Specifically, participants (both treatment and control groups) were asked about their overall satisfaction with the treatment and services at the cancer center, their satisfaction with information received about their treatment, the response(s) received when they attempted contact with the health-care team after hours, the amount of support received, their current smoking and alcohol usage, and several demographic questions not earlier assessed or available through record review (years of education, highest degree, income range). Those receiving the intervention were also asked about the impact of the Health Buddy on their care and actions taken in response to the algorithms.
Demographic and medical information
Demographic information was collected using the initial survey, and information about the participant's medical history, condition, treatments received, treatment timing, complications, comorbidities, and treatment response was collected via retrospective medical record review subsequent to completion of the clinical trial.
Outcome measures
While outcomes of the clinical trial are not the subject of this article, the results of QOL and symptom burden measures for the treatment group only are included here because of their relationship with device utilization. The two measures included the Functional Assessment of Cancer Therapy–Head and Neck Scale and the Memorial Symptom Assessment Scale and were administered at baseline (before beginning treatment), mid-treatment, and posttreatment.
• Functional Assessment of Cancer Therapy–Head and Neck Scale (FACT-H&N). The FACT-G (general) is a multidimensional QOL instrument designed for use with all cancer patients. The instrument has 28 items divided into four subscales: Functional Well-Being, Physical Well-Being, Social Well-Being, and Emotional Well-Being. This generic core questionnaire was found to meet or exceed requirements for use in oncology based upon ease of administration, brevity, reliability, validity, and responsiveness to clinical change.30 Added to the core questionnaire is the head and neck–specific subscale, consisting of 11 items specific to this cancer site. A Trial Outcome Index (TOI) is also scored and is the result of the physical, functional, and cancer-specific subscales. List et al31 found the FACT-H&N to be reliable and sensitive to differences in functioning for patients with head and neck cancers (Cronbach's alpha was 0.89 for total FACT-G and 0.63 for the head and neck subscale in this study of 151 patients). Additionally, head and neck cancer patients found the FACT-H&N relevant to their problems and easy to understand, and it was preferred over other validated head and neck cancer QOL questionnaires.32 The FACT-H&N was chosen for this study because it (1) is nonspecific related to a treatment modality or subsite among head and neck cancers, (2) allows comparison across cancer diagnoses while still probing issues specific to head and neck cancer, (3) is short and can be completed quickly, (4) includes the psychosocial domains of social/family and emotion subscales as well as physical and functional areas, and (5) is self-administered.
• Memorial Symptom Assessment Scale (MSAS). This multidimensional scale measures the prevalence, severity, and distress associated with the most common symptoms experienced by cancer patients. Physical and emotional subscale scores as well as a Global Distress Index (GDI, considered to be a measure of total symptom burden) can be generated from patient responses. The MSAS has demonstrated validity and reliability in both in- and outpatient cancer populations.[33], [34] and [35] Initial psychometric analysis by Portenoy et al34 used factor analysis to define two subscales: psychological symptoms and physical symptoms with Cronbach alpha coefficients of 0.88 and 0.83, respectively; convergent validity was also established. It was chosen for this study because of its proven ability to measure both the presence and the intensity of experienced symptoms.[33], [35], [36], [37] and [38]
Data Analysis
Quantitative data were documented and analyzed using the Statistical Package for the Social Sciences (SPSS, Inc., Chicago, IL), version 16. Descriptive statistics were calculated to describe the sample and assess study outcomes, including feasibility and acceptability of the intervention. To ascertain relationships between usage of the device and demographic and medical information, a series of correlational analyses using Spearman's rho were conducted. This nonparametric test was chosen over Pearson's r because of the small sample size, the lack of a normal distribution for several of the variables, and the ordinal nature of several of the variables. Multiple regression analyses were also planned, but lack of significant bivariate correlations precluded multivariate analysis.
Qualitative responses to open-ended questions were analyzed to identify themes and direct quotations illustrating those themes.
Descriptive analysis of the treatment group's responses to the outcome measures (QOL and symptom burden) was done to ascertain changes over the course of the intervention using the mean scores at baseline, during treatment, and posttreatment.
Results
Description of Participants
Participants randomly assigned to the intervention group (n = 45) were an average age of 59 years (±11.7), and most were covered by private (34%) or public (48%) insurance. On average, participants had completed 13.5 years (±3.0) of formal education. Thirty-nine (87%) of the participants were male and 91% were Caucasian.
With regard to medical information, participants were predominantly diagnosed with stage II cancers of the head and neck (36%). The most prevalent site was the larynx (12 patients), followed by the tongue and the base of the tongue (seven patients) and unknown primary (seven patients). The vast majority received chemotherapy (32, or 71%) and/or radiation (42, or 93%).
Additional details regarding demographic and medical characteristics of the sample are provided in Table 2.
| FREQUENCY | VALID PERCENT | |
|---|---|---|
| Gender (n = 44) | ||
| Male | 39 | 88.6 |
| Female | 5 | 11.3 |
| Race (n = 44) | ||
| Caucasian | 40 | 90.9 |
| African American | 4 | 9.0 |
| Tumor stage (n = 44) | ||
| I | 7 | 15.9 |
| II | 15 | 34.0 |
| III | 11 | 25.0 |
| IV | 4 | 9.0 |
| Unable to determine | 5 | 11.4 |
| Unknown | 2 | 4.5 |
| Site of cancer (n = 44) | ||
| Larynx | 12 | 27.2 |
| Tongue, base of tongue | 7 | 15.9 |
| Unknown primary | 7 | 15.9 |
| Tonsillar | 4 | 9.0 |
| Other H&N sites | 14 | 31.8 |
| Insurance status (n = 44) | ||
| No insurance | 8 | 18.2 |
| Medicaid | 1 | 2.3 |
| Medicare | 2 | 4.5 |
| Medicaid and Medicare | 1 | 2.3 |
| Medicare and supplement | 9 | 20.5 |
| Medicare and VA benefits | 2 | 4.5 |
| Veteran benefits only | 6 | 13.6 |
| Private insurance | 15 | 34.1 |
| Highest educational degree (n = 20)a | ||
| Less than high school | 3 | 15.0 |
| High school or GED | 9 | 45.0 |
| Associate's/bachelor's degree | 4 | 20.0 |
| Masters, PhD, or MD | 2 | 10.0 |
| Other | 2 | 10.0 |
| Income range (n = 18)a | ||
| $20,000 or less | 5 | 27.8 |
| $20,001–50,000 | 5 | 27.8 |
| $70,001–100,000 | 5 | 27.8 |
| Over $100,000 | 3 | 16.7 |
| Percent of poverty in zip code area (n = 44) | ||
| 2.8–5.1% | 11 | 25.0 |
| 5.9–8.6% | 11 | 25.0 |
| 9.0–11.9% | 10 | 22.7 |
| 12.3–45.9% | 12 | 27.2 |
Feasibility and Acceptability
Accrual rate
Of the 185 patients assessed for eligibility during the 2-year recruitment period, 105 were excluded. See Figure 3 for a detailed depiction of study accrual for both the treatment and control groups. Thirty-three (31%) were excluded because they did not have a land phone line, a requirement for transmitting the algorithms to the Health Buddy appliance. Most of these had cell phones only. No potential participants refused participation due to issues related to operation of the technology itself.
Device utilization
Participants used the telehealth device for an average of 70.7 days (±26.7), which constituted 86.3% (±15.0) of the total days available for use. Of note, the median percentage of use was 94.2% and the modal percentage was 100%, indicating that the vast majority of participants consistently used the telehealth device. The participant with the lowest usage rate used the device 46% of the days available.
By far, the most common reason for Health Buddy nonresponse was patient hospitalization. Two subjects traveled out of town frequently on weekends and would leave the Health Buddy at home. One subject had accidentally unhooked the Health Buddy, and a home visit was made to reconnect the device into the patient's phone line.
Nurse-initiated contacts with participants and/or clinicians
Of the 45 enrolled patients, 33 required additional contact with a research nurse (see Table 3). The most common reasons patients were contacted were nonresponse for 3 consecutive days (38.3%), repeated reporting of high levels of unrelieved pain (30%), and suicidal thoughts (10%). In all, 120 calls were placed: one call for every 25.9 response days. In every case, the problem was resolved.
| NUMBER OF PATIENTS | PROBLEM | OUTGOING CALLS | RESOLUTION |
|---|---|---|---|
| 15 | No response on Health Buddy for 3 consecutive days | 46 | Patient teaching |
| 17 | Pain-related issues | 36 | Advocacy/referral/patient teaching |
| 5 | Suicidal thoughts | 12 | Advocacy/referral |
| 7 | G-tube problems | 8 | Patient teaching |
| 5 | Sadness/depression | 6 | Advocacy/referral |
| 3 | Multiple symptoms | 3 | Advocacy/referral |
| 3 | Nausea/vomiting | 4 | Referral/patient teaching |
| 2 | Coughing/excessive secretions | 2 | Patient teaching |
| 2 | Constipation | 2 | Patient teaching |
| 1 | Stomatitis | 1 | Referral |
Satisfaction ratings
Responses to surveys programmed into the Health Buddy system are displayed in Table 4. Overall, respondents responded favorably, finding the installation to be easy, the content to be helpful, and the overall experience to be positive.
| PERCENT OF RESPONDENTS | |
|---|---|
| Installation satisfaction | |
| Installation problems? | |
| Yes | 2 |
| No | 98 |
| Any difficulty completing the first training questions? | |
| Yes | 7 |
| No | 94 |
| Length of installation? | |
| 2–5 minutes | 52 |
| 6–10 minutes | 41 |
| 11–15 minutes | 4 |
| 16–20 minutes | 2 |
| Content satisfaction | |
| Overall, I think the Health Buddy questions are | |
| Very easy | 44 |
| Somewhat easy | 16 |
| Neutral | 32 |
| Somewhat difficult | 4 |
| Difficult | 4 |
| Repeating questions reinforced knowledge and understanding | |
| Strongly agree | 56 |
| Somewhat agree | 28 |
| Neutral | 12 |
| Somewhat disagree | 4 |
| Strongly disagree | 0 |
| Understanding of my health condition | |
| Much better | 64 |
| Somewhat better | 20 |
| Neutral | 16 |
| Somewhat worse | 0 |
| Much worse | 0 |
| Managing my health condition | |
| Much better | 52 |
| Somewhat better | 44 |
| Neutral | 4 |
| Somewhat worse | 0 |
| Much worse | 0 |
| Recommend the device to others | |
| Very willing | 80 |
| Somewhat willing | 12 |
| Neutral | 4 |
| Somewhat unwilling | 0 |
| Very unwilling | 4 |
| Overall satisfaction | |
| Satisfaction with device | |
| Very satisfied | 45 |
| Satisfied | 35 |
| Somewhat satisfied | 15 |
| Not very satisfied | 5 |
| Satisfaction with the communication between you and your doctor or nurse | |
| More satisfied | 65 |
| No difference | 30 |
| Less satisfied | 5 |
| Ease of using the device | |
| Very easy | 85 |
| Easy | 15 |
| Not easy | 0 |
| Overall experience with the device | |
| Positive | 85 |
| Neutral | 15 |
| Negative | 0 |
| Continue to use the device | |
| Very likely | 40 |
| Likely | 40 |
| Somewhat likely | 15 |
| Not very likely | 0 |
Narrative comments
During the exit interview, participants were asked, “How was having the Health Buddy helpful to you?” Responses could be categorized into two major themes: (1) the Health Buddy provided needed information and (2) the Health Buddy improved my self-management during treatment.
Statements made related to the information provided included the following:
- • It gave me information on what could be expected from treatment
• It was a constant reminder of things to watch for
• It kept me abreast of my total condition at all times
• It gave good directions so I didn't have to ask at the cancer center
• It gave good suggestions on treatments (home remedies) such as gargles, care of feeding tube, exhaustion, and everyday symptoms
Statements made indicative that the Health Buddy improved self-management included the following:
- • I learned what I could do to make myself feel better
• It helped me manage my symptoms
• It taught me about symptom management and how to handle problems
• It let me know whether to contact a doctor or use self-care
• It gave me who to call for problems and some things to try
• It kept me aware of what I needed to do in order to make the period easier
• It reminded me to take my meds and exercise
Additionally, some participants noted the support they felt from having the Health Buddy interventions during treatment in saying the following:
- • It kind of helped my depression through acknowledging it and giving me something to do
• It made me feel I was not the only one who had experience with these things
• It comforted me because I knew what was going to happen
Poststudy survey
Twenty (45%) of the 44 patients who received the intervention responded to the mailed poststudy survey. When asked if they felt they received better care because they had the device, 13 of the 20 (65%) responded that they did. Eighteen (90%) of the treatment group responders stated they were very satisfied with their care (one stated “somewhat satisfied”) and 20 (100%) said they would recommend the cancer center for treatment. Nineteen (95%) stated they received adequate support during treatment.
Outcome Measures
Mean scores on the FACT-H&N and subscales and the MSAS and subscales taken pre-, during, and posttreatment are displayed in Table 5. As expected, average QOL scores declined during treatment, while symptom distress increased, with a return to near baseline scores posttreatment.
| SCALE/SUBSCALE | PRETREATMENT | DURING TREATMENT | POSTTREATMENT |
|---|---|---|---|
| Total FACT-H&N | 100.3 | 85.6 | 101.5 |
| FACT-G | 74.3 | 69.4 | 78.5 |
| Trial Outcome Index | 62.6 | 46.0 | 65.0 |
| Physical Well-Being | 21.2 | 17.6 | 21.1 |
| Functional Well-Being | 15.6 | 12.5 | 17.4 |
| Emotional Well-Being | 21.1 | 22.3 | 22.2 |
| Social Well-Being | 21.1 | 22.3 | 22.2 |
| Total MSAS | 0.7 | 1.1 | 0.8 |
| Global Distress Index | 1.1 | 1.8 | 1.3 |
| Physical | 0.7 | 1.5 | 1.1 |
| Psychological | 1.1 | 1.2 | 0.8 |
Correlations
The relationships between percentage usage per patient and the following variables were evaluated: age, income, years of education, tumor stage, and percent poverty in patient's zip code. Percent poverty in zip code area was intended to be a surrogate measure of the patient's socioeconomic status. Results are displayed in Table 6. No significant correlations were noted, although years of education and percentage poverty in zip code showed a trend toward significance.
| VARIABLE (VS % USAGE) | RELATIONSHIP | |
|---|---|---|
| SPEARMAN'S RHO RS | SIGNIFICANCE (ONE-TAILED) | |
| Percent poverty in zip code | 0.213 | 0.083 |
| Age | 0.146 | 0.173 |
| Years of education | −0.325 | 0.081 |
| Income | −0.292 | 0.120 |
| Tumor stage | 0.196 | 0.122 |
| Physical Well-Being (during treatment) | 0.310 | 0.048 |
| Emotional Well-Being (during treatment) | 0.315 | 0.042 |
Although a multivariate model was planned, the lack of significant bivariate correlations precluded the need for multivariate analysis.
When percent usage was correlated with FACT-H&N total and subscales taken at baseline, during active treatment, and posttreatment, significant positive correlations were found between the percentage used and the Physical Well-Being subscale score during treatment (Spearman's rho = 0.310, P = 0.048) and between percentage used and the Emotional Well-Being subscale during treatment (Spearman's rho = 0.315, P = 0.042).
There were no significant correlations between percentage usage and the scores on the MSAS.
Discussion
Both qualitative and quantitative measures indicate that using telehealth to support symptom management during aggressive cancer treatment is both feasible and well-accepted. Patient users were not intimidated by this particular technology as it was simple to set up and use and required no previous computer training to operate. The Health Buddy was viewed as providing important and useful information. Overall, users felt that it improved their ability to self-manage their disease and the side effects of treatment and provided a sense of support and security.
Unlike other studies which use telehealth devices to monitor patient symptoms, our goal was to increase patient self-management of the symptoms experienced during intensive medical treatment, therefore avoiding increased burden on the medical system. The fact that the research nurse overseeing the responses needed to intervene only once every 25.9 days speaks to the ability of the intervention to have a positive impact on utilization of medical services.
The lack of significant relationships between usage and descriptive variables such as age and years of education suggests that the intervention was equally acceptable to all subgroups. Factors such as age, previous computer literacy, educational obtainment, and socioeconomic status did not significantly differentiate our study population in terms of compliance as verified by usage percentages.
The significant relationships found between the percentage used and the subscale scores on Physical Well-Being and Emotional Well-Being during treatment may indicate that increased use of the telemessaging intervention during treatment resulted in better physical and emotional aspects of QOL.
The high rate of daily compliance with the intervention in spite of differentiating personal variables and the severity of the treatment regimen may have been due to one or a combination of the following factors:
- • the simplicity of the technology
• the visibility of the appliance (often placed in the kitchen or living area of the home) and its flashing green light as cues to the need to respond
• the usefulness of the information provided
• the use of simple messaging language presented in an encouraging, positive manner
• affirmations related to application of the symptom management protocols suggested
• curiosity related to the day's messaging and the motivational saying which always appeared at the end
• knowledge that someone was reviewing the responses, tracking and intervening when the participant did not respond for several days
Our study supports the benefits of telehealth interventions noted by providers in a study by Sandberg et al39: opportunities for more frequent contact, greater relaxation and information due to the ability to interact in one's own home, increased accessibility by those frequently underserved, and timely medical information and monitoring. Similar to the study done in the Netherlands,[27] and [28] this study noted technological problems as the primary disadvantage; but in our intervention, we had no problems with the technology or equipment.
Although computerized technology served as a barrier to previous telehealth research, the lack of a land-based phone line was a factor preventing participation in the current study. Indeed, many participants maintained only wireless communication devices, which were not compatible with the version of the Health Buddy that was employed in this study. However, improvements in the technology since completion of this study now allow for wireless access to the appliance or provision of an independent wireless messaging device for those without such access in their own homes.
Although the data generally support the feasibility and acceptability of the telehealth-based intervention, the results should be interpreted in the context of a few study limitations. In particular, the sample size was somewhat small, and data pertaining to the socioeconomic status of participants were not available for all participants. Second, the study did not include measures of the patient's direct interactions with health-care providers during the study or specific data related to their health-care utilization (eg, emergency room visits, preventable inpatient hospitalizations, emergency calls to clinicians). The collection of more exhaustive measures of health-care utilization was limited by resources but is planned for subsequent studies. Finally, concerns regarding subject burden limited assessment of the usability of the telehealth device.
Although compliance with utilization expectations and completion of study measures was excellent during the course of the intervention, response to the follow-up survey mailed several months later was less than 50%. This low response rate was most likely due to several factors: (1) this survey was sent at the conclusion of the entire study (by this time, patients were 0–21 months past their active participation); (2) it was a mailed survey with no additional contact or follow-up effort to increase response rate; (3) participants may have felt that they had already shared their opinions in the exit interview and may have felt overburdened by study measures at this point; and (4) participants may have died, moved, or been medically unable to respond. This lack of response did limit our ability to evaluate the longitudinal impact of the intervention.
Conclusions
This telehealth intervention proved to be an acceptable and feasible means to educate and support patients during aggressive treatment for head and neck cancer. Patient compliance with telehealth interventions during periods of extreme symptom burden and declining QOL is feasible if simple technology cues the patient to participate, offers positive support and relevant education, and is targeted or tailored to their specific condition.
Treatment for head and neck cancer is most often a rigorous regimen of combination therapies, producing a multitude of distressing symptoms and side effects. While it is nearly impossible to circumvent the physical and psychosocial insults caused by such treatment, some interventions directed toward educating and supporting patients during active treatment have met with success.[1], [2], [3] and [4] Conversely, other efforts have demonstrated little impact[5] and [6] or have been poorly received,7 pointing to the need for effective, acceptable means to provide support during such difficult treatment.
Over the past 10 years, telemedicine technology has enabled innovative approaches for improving patient education, assessment, support, and communication during treatment for both acute and chronic diseases. A recent policy white paper8 described telemedicine technology as including “the electronic acquisition, processing, dissemination, storage, retrieval, and exchange of information for the purpose of promoting health, preventing disease, treating the sick, managing chronic illness, rehabilitating the disabled, and protecting public health and safety” (p. 2). This same paper suggests that national telemedicine initiatives are essential to health-care reform based upon their proven cost–effectiveness and clinical efficacy. However, cost savings and clinical effectiveness will be unrealized outcomes if the interventions are not feasible in practice or acceptable to the targeted population.
In the arena of cancer care, telephone-based systems have been used to report and monitor cancer symptoms with favorable compliance noted even when patients are expected to initiate calls on a regular basis.[9], [10], [11] and [12] Favorable acceptance ratings have also been reported by both patients and clinicians regarding computerized systems used to assess symptoms and quality of life (QOL) in cancer patients.[13], [14], [15], [16], [17], [18] and [19] In the United Kingdom, a handheld computer system was successfully used to monitor and support patients receiving chemotherapy for lung or colorectal cancer,20 and a study testing a dialogic model of cancer care expecting patients to respond to telehealth messaging on a daily basis over 6 months reported an 84% cooperation rate.21 In these studies, the majority of patients reported ease of use and acceptability of the technology. Survey research has found both urban and rural cancer patients to be receptive to medical and psychiatric services provided via telehealth.22
Published reports describing use of telehealth and computerized interventions during head and neck cancer treatment are less prevalent. Touch-screen computers were successfully used in the Netherlands to collect QOL and distress data from head and neck cancer patients.16 Videoconferencing has been used successfully to overcome geographical barriers to patient assessment[23], [24] and [25] and to provide speech–language pathology services to people living with head and neck cancers in remote areas. Reported use of telehealth management appears promising for providing timely access to care for those who are geographically isolated.26
A research group based in the Netherlands developed and tested a comprehensive electronic health information support system for use in head and neck cancer care.27 The system had four patient-related functions: facilitating communication between patients and health-care providers, providing information about the disease and its treatment, connecting patients with other patients similarly diagnosed, and monitoring patients after hospital discharge. The system was found to be well-accepted and appreciated by participating patients, and its use enabled early identification and direct intervention for patient problems.27 A clinical trial of the telehealth application showed improved QOL in five of 22 studied parameters for the treatment group.28 However, 20 of the 59 patients eligible for the intervention group refused participation; 11 (55%) of these stated computer-related concerns as their reason for nonparticipation.
Knowing that head and neck cancer patients experience a high burden of illness and often have significant communication, socioeconomic, and geographic barriers to care, our team developed a telehealth intervention using a simple telemessaging device to circumvent communication barriers and perceived technical challenges associated with computer-based systems to provide education and support to patients in their own home and on their own time schedule.29 Overall, we hypothesized that patients receiving the intervention would experience less symptom distress, improved QOL, increased self-efficacy, and greater satisfaction with symptom management than those in the control group. However, as a first step toward examining the efficacy and effectiveness of this intervention, this study examined both quantitative and qualitative indicators of its feasibility and acceptance among patients undergoing treatment for head and neck cancer.
Methods
Design
Subsequent to study approval by the University of Louisville's Human Subjects Protection Office, a randomized clinical trial comparing the telehealth intervention to standard care was conducted using a two-group parallel design. This study reports on the intervention's feasibility and acceptance in the treatment group of 44 patients.
Site
Participants were recruited from patients receiving care from the Multidisciplinary Head and Neck Cancer Team at the James Graham Brown Cancer Center (JGBCC) over a 2-year period (June 2006 through June 2008). The team consisted of head and neck surgeons, medical oncologists, radiation oncologists, nurses, a pathologist, a speech therapist, a registered dietician, a psychologist, and a social worker. This team developed a comprehensive assessment and treatment plan during each patient's initial visit to the clinic and coordinated patient care throughout the treatment process.
Sample
Patients eligible for study participation met the following inclusion criteria: (1) initial diagnosis of head or neck cancer including cancers of the oral cavity, salivary glands, paranasal sinuses and nasal cavity, pharynx, and larynx; (2) involvement in a treatment plan including one or more modalities (ie, surgery, chemotherapy, radiation, or any combination); (3) capacity to give independent informed consent; and (4) ability to speak, read, and comprehend English at the eighth-grade level or above. Patients were excluded from participation if they had no land telephone line, had a thought disorder, were incarcerated, or had compromised cognitive functioning.
All patients scheduled for assessment received an explanation of the research study via print materials prior to their first clinic visit. During their first scheduled clinic visit, all patients identified as eligible were approached by a member of the research study staff, who briefly explained the study and asked if they might be interested in study participation. Because of the stress and content of this first clinic visit, interested patients were contacted later by phone to schedule an additional visit to review the study and obtain informed consent.
During the informed consent meeting, the study procedures were explained in detail. If the patient agreed and signed a consent form, a randomization grid which considered the patient's particular treatment plan was used to assign the patient to either the control or the experimental group. Baseline data were also collected during this first visit.
Description of the Intervention
The technology selected for implementing the intervention was the Health Buddy® System, a commercially available, proprietary system produced and maintained by Robert Bosch Healthcare Palo Alto, CA. The Health Buddy, the appliance used for interaction between the participant and the health-care provider, is a user-friendly, easily visible, electrical device that attaches to the user's land phone line (see Figure 1). Questions and information are displayed on the liquid crystal display (LCD) screen of the 6 × 9–inch appliance. The individual responds to questions by pressing one of the four large buttons below the screen. The research team selected the technology provider based on the ability of the technology to perform in accordance with the research objectives.
Symptom control algorithms developed using participatory action research (surveys of current and past patients and clinicians) and evidence-based practice were programmed into the telehealth messaging system (see article by Head et al,29 which details the algorithm topic selection and development process). The algorithms addressed 29 different symptoms and side effects of treatment, consisting of approximately 100 questions accompanied by related educational and supportive responses. Patients were asked three to five questions daily related to the symptoms anticipated during their treatment scenario. Depending upon their response, they would receive specific information related to symptom self-management, including recommendations as to when to contact their clinicians. The algorithms were constructed with the goal of encouraging self-efficacy and independent action on the part of the participant. See Figure 2 for an example of the branching algorithms.
Participants randomly assigned to the treatment group immediately had the Health Buddy connected to a land telephone line in their home. Most (40%) chose to place it in their kitchen, while another 26% placed it in their bedrooms; most often, it was in a highly visible location, serving to remind the participant to respond. Research study staff delivered, installed, and demonstrated how to operate the equipment. Installation was simple and required only minutes. A tutorial programmed into the Health Buddy taught participants how to reply to questions appearing on the monitor using the four large keys below the possible answers or a rating scale which would appear depending on the type of question asked.
During the early hours of the morning, the device would automatically call a toll-free number. Responses to the previous day's questions were uploaded, and questions and related information for the next day were downloaded over the telephone line onto a secure server. Phone service was never disrupted by the device; if the phone was in use, the system would connect later to retrieve and download information. Once new content was transferred, a green light on the device would flash to alert the participant that new questions were available for response. Once the participant pressed any of the keys, the new algorithms would begin appearing on the monitor screen.
Participants were instructed to begin responding on the first day they received treatment or on the first day after returning home from surgery. They were asked to continue responding daily (unless hospitalized for treatment) throughout the treatment period and for approximately 2 weeks posttreatment as treatment-induced symptoms continue during that period of time. Study staff contacted participants when treatment was complete and scheduled a date to pick up the appliance and end daily responding. Daily patient responses required 5–10 minutes.
Participant responses could be viewed by study staff via Internet access 1 day after being answered. Responses were monitored daily by study nurses. Symptoms unrelieved over time or symptoms targeted as requiring immediate intervention (ie, serious consideration of suicide) would result in the study nurse contacting the patient directly by phone and/or contacting clinicians to assure immediate intervention. However, it is important to note that this direct intervention by study staff was infrequent as most symptoms were addressed independently by the participant as desired. If a participant had not reported a period of planned hospitalization and did not respond for 3 consecutive days, study staff would contact the patient by phone to ascertain the reason for noncompliance.
Measures
The following indicators were selected as measures of acceptance (accrual rate), feasibility (utilization, nurse-initiated contacts), and/or satisfaction (satisfaction ratings). Narrative responses and a poststudy survey provided additional data examining acceptance, feasibility, and satisfaction with the intervention. In addition, demographic and medical information as well as measures assessing primary study outcomes were collected from each participant. Table 1 lists all measures and the study time point when they were administered.
| MEASURES | PRETREATMENT | DURING TREATMENT | POSTTREATMENT | CUMULATIVE |
|---|---|---|---|---|
| Demographics | X (baseline) | X | ||
| Accrual rate | X | |||
| Utilization rate | X | |||
| FACT-H&N | X (baseline) | X (mid-tx) | X (2–4 weeks post-tx) | |
| MSAS | X (baseline) | X (mid-tx) | X (2–4 weeks post-tx) | |
| Satisfaction with technology | X | |||
| Nurse-initiated contacts | X | |||
| Exit interview | X (end of treatment) | |||
| Poststudy written survey | X (60–90 days post-tx) |
tx = treatment
Accrual rate
The number of individuals assessed for study eligibility, reasons for exclusion or noncompletion, and numbers included in the analysis were all recorded to examine acceptance of the intervention and identify issues with the intervention or technology affecting participation.
Utilization
Feasibility was operationalized as device utilization using the percentage of days on which a participant responded to the Health Buddy. This was calculated using the number of days the participant responded to the telehealth device divided by the number of days the participant had the device and was expected to respond. These data were maintained and provided by the telehealth provider (Robert Bosch Healthcare).
Nurse-initiated contacts with participants and/or clinicians
The number of occasions on which a nurse decided to intervene was used as an indicator of feasibility under the premise that the goal of the intervention was to support and encourage patient-driven efforts to seek care for persistent or troubling symptoms. If a patient reported a symptom, he or she was given management information and encouraged to discuss problems further with the clinician either by phone or during clinic visits. If a patient continued to report an unresolved symptom or if the symptom required immediate intervention (ie, suicide threat), the research nurse reviewing responses would contact the patient and/or clinician to ascertain why and/or assist with its resolution. These nurse-initiated contacts should be infrequent if the intervention is achieving the goal of developing patient self-efficacy.
Satisfaction ratings
Items assessing satisfaction with the technology were also administered to participants via the telehealth messaging device. Questions related to satisfaction with the initial setup of the telehealth appliance were asked at the beginning of the intervention. Ongoing satisfaction with the device, messaging content, and the health-care provider were assessed every 90 days. The specific questions asked are detailed in Table 4.
Narrative data
Upon completion of the intervention, participants in the treatment group completed an exit interview using open-ended questions regarding the utility of the intervention, relevance of the algorithms, value or burden of item repetition in the algorithms, symptoms or problems experienced that were not addressed by the intervention, and general comments.
Poststudy survey
A final survey was mailed to participants several months after completion of the study, asking for additional feedback about the impact of the intervention. Specifically, participants (both treatment and control groups) were asked about their overall satisfaction with the treatment and services at the cancer center, their satisfaction with information received about their treatment, the response(s) received when they attempted contact with the health-care team after hours, the amount of support received, their current smoking and alcohol usage, and several demographic questions not earlier assessed or available through record review (years of education, highest degree, income range). Those receiving the intervention were also asked about the impact of the Health Buddy on their care and actions taken in response to the algorithms.
Demographic and medical information
Demographic information was collected using the initial survey, and information about the participant's medical history, condition, treatments received, treatment timing, complications, comorbidities, and treatment response was collected via retrospective medical record review subsequent to completion of the clinical trial.
Outcome measures
While outcomes of the clinical trial are not the subject of this article, the results of QOL and symptom burden measures for the treatment group only are included here because of their relationship with device utilization. The two measures included the Functional Assessment of Cancer Therapy–Head and Neck Scale and the Memorial Symptom Assessment Scale and were administered at baseline (before beginning treatment), mid-treatment, and posttreatment.
• Functional Assessment of Cancer Therapy–Head and Neck Scale (FACT-H&N). The FACT-G (general) is a multidimensional QOL instrument designed for use with all cancer patients. The instrument has 28 items divided into four subscales: Functional Well-Being, Physical Well-Being, Social Well-Being, and Emotional Well-Being. This generic core questionnaire was found to meet or exceed requirements for use in oncology based upon ease of administration, brevity, reliability, validity, and responsiveness to clinical change.30 Added to the core questionnaire is the head and neck–specific subscale, consisting of 11 items specific to this cancer site. A Trial Outcome Index (TOI) is also scored and is the result of the physical, functional, and cancer-specific subscales. List et al31 found the FACT-H&N to be reliable and sensitive to differences in functioning for patients with head and neck cancers (Cronbach's alpha was 0.89 for total FACT-G and 0.63 for the head and neck subscale in this study of 151 patients). Additionally, head and neck cancer patients found the FACT-H&N relevant to their problems and easy to understand, and it was preferred over other validated head and neck cancer QOL questionnaires.32 The FACT-H&N was chosen for this study because it (1) is nonspecific related to a treatment modality or subsite among head and neck cancers, (2) allows comparison across cancer diagnoses while still probing issues specific to head and neck cancer, (3) is short and can be completed quickly, (4) includes the psychosocial domains of social/family and emotion subscales as well as physical and functional areas, and (5) is self-administered.
• Memorial Symptom Assessment Scale (MSAS). This multidimensional scale measures the prevalence, severity, and distress associated with the most common symptoms experienced by cancer patients. Physical and emotional subscale scores as well as a Global Distress Index (GDI, considered to be a measure of total symptom burden) can be generated from patient responses. The MSAS has demonstrated validity and reliability in both in- and outpatient cancer populations.[33], [34] and [35] Initial psychometric analysis by Portenoy et al34 used factor analysis to define two subscales: psychological symptoms and physical symptoms with Cronbach alpha coefficients of 0.88 and 0.83, respectively; convergent validity was also established. It was chosen for this study because of its proven ability to measure both the presence and the intensity of experienced symptoms.[33], [35], [36], [37] and [38]
Data Analysis
Quantitative data were documented and analyzed using the Statistical Package for the Social Sciences (SPSS, Inc., Chicago, IL), version 16. Descriptive statistics were calculated to describe the sample and assess study outcomes, including feasibility and acceptability of the intervention. To ascertain relationships between usage of the device and demographic and medical information, a series of correlational analyses using Spearman's rho were conducted. This nonparametric test was chosen over Pearson's r because of the small sample size, the lack of a normal distribution for several of the variables, and the ordinal nature of several of the variables. Multiple regression analyses were also planned, but lack of significant bivariate correlations precluded multivariate analysis.
Qualitative responses to open-ended questions were analyzed to identify themes and direct quotations illustrating those themes.
Descriptive analysis of the treatment group's responses to the outcome measures (QOL and symptom burden) was done to ascertain changes over the course of the intervention using the mean scores at baseline, during treatment, and posttreatment.
Results
Description of Participants
Participants randomly assigned to the intervention group (n = 45) were an average age of 59 years (±11.7), and most were covered by private (34%) or public (48%) insurance. On average, participants had completed 13.5 years (±3.0) of formal education. Thirty-nine (87%) of the participants were male and 91% were Caucasian.
With regard to medical information, participants were predominantly diagnosed with stage II cancers of the head and neck (36%). The most prevalent site was the larynx (12 patients), followed by the tongue and the base of the tongue (seven patients) and unknown primary (seven patients). The vast majority received chemotherapy (32, or 71%) and/or radiation (42, or 93%).
Additional details regarding demographic and medical characteristics of the sample are provided in Table 2.
| FREQUENCY | VALID PERCENT | |
|---|---|---|
| Gender (n = 44) | ||
| Male | 39 | 88.6 |
| Female | 5 | 11.3 |
| Race (n = 44) | ||
| Caucasian | 40 | 90.9 |
| African American | 4 | 9.0 |
| Tumor stage (n = 44) | ||
| I | 7 | 15.9 |
| II | 15 | 34.0 |
| III | 11 | 25.0 |
| IV | 4 | 9.0 |
| Unable to determine | 5 | 11.4 |
| Unknown | 2 | 4.5 |
| Site of cancer (n = 44) | ||
| Larynx | 12 | 27.2 |
| Tongue, base of tongue | 7 | 15.9 |
| Unknown primary | 7 | 15.9 |
| Tonsillar | 4 | 9.0 |
| Other H&N sites | 14 | 31.8 |
| Insurance status (n = 44) | ||
| No insurance | 8 | 18.2 |
| Medicaid | 1 | 2.3 |
| Medicare | 2 | 4.5 |
| Medicaid and Medicare | 1 | 2.3 |
| Medicare and supplement | 9 | 20.5 |
| Medicare and VA benefits | 2 | 4.5 |
| Veteran benefits only | 6 | 13.6 |
| Private insurance | 15 | 34.1 |
| Highest educational degree (n = 20)a | ||
| Less than high school | 3 | 15.0 |
| High school or GED | 9 | 45.0 |
| Associate's/bachelor's degree | 4 | 20.0 |
| Masters, PhD, or MD | 2 | 10.0 |
| Other | 2 | 10.0 |
| Income range (n = 18)a | ||
| $20,000 or less | 5 | 27.8 |
| $20,001–50,000 | 5 | 27.8 |
| $70,001–100,000 | 5 | 27.8 |
| Over $100,000 | 3 | 16.7 |
| Percent of poverty in zip code area (n = 44) | ||
| 2.8–5.1% | 11 | 25.0 |
| 5.9–8.6% | 11 | 25.0 |
| 9.0–11.9% | 10 | 22.7 |
| 12.3–45.9% | 12 | 27.2 |
Feasibility and Acceptability
Accrual rate
Of the 185 patients assessed for eligibility during the 2-year recruitment period, 105 were excluded. See Figure 3 for a detailed depiction of study accrual for both the treatment and control groups. Thirty-three (31%) were excluded because they did not have a land phone line, a requirement for transmitting the algorithms to the Health Buddy appliance. Most of these had cell phones only. No potential participants refused participation due to issues related to operation of the technology itself.
Device utilization
Participants used the telehealth device for an average of 70.7 days (±26.7), which constituted 86.3% (±15.0) of the total days available for use. Of note, the median percentage of use was 94.2% and the modal percentage was 100%, indicating that the vast majority of participants consistently used the telehealth device. The participant with the lowest usage rate used the device 46% of the days available.
By far, the most common reason for Health Buddy nonresponse was patient hospitalization. Two subjects traveled out of town frequently on weekends and would leave the Health Buddy at home. One subject had accidentally unhooked the Health Buddy, and a home visit was made to reconnect the device into the patient's phone line.
Nurse-initiated contacts with participants and/or clinicians
Of the 45 enrolled patients, 33 required additional contact with a research nurse (see Table 3). The most common reasons patients were contacted were nonresponse for 3 consecutive days (38.3%), repeated reporting of high levels of unrelieved pain (30%), and suicidal thoughts (10%). In all, 120 calls were placed: one call for every 25.9 response days. In every case, the problem was resolved.
| NUMBER OF PATIENTS | PROBLEM | OUTGOING CALLS | RESOLUTION |
|---|---|---|---|
| 15 | No response on Health Buddy for 3 consecutive days | 46 | Patient teaching |
| 17 | Pain-related issues | 36 | Advocacy/referral/patient teaching |
| 5 | Suicidal thoughts | 12 | Advocacy/referral |
| 7 | G-tube problems | 8 | Patient teaching |
| 5 | Sadness/depression | 6 | Advocacy/referral |
| 3 | Multiple symptoms | 3 | Advocacy/referral |
| 3 | Nausea/vomiting | 4 | Referral/patient teaching |
| 2 | Coughing/excessive secretions | 2 | Patient teaching |
| 2 | Constipation | 2 | Patient teaching |
| 1 | Stomatitis | 1 | Referral |
Satisfaction ratings
Responses to surveys programmed into the Health Buddy system are displayed in Table 4. Overall, respondents responded favorably, finding the installation to be easy, the content to be helpful, and the overall experience to be positive.
| PERCENT OF RESPONDENTS | |
|---|---|
| Installation satisfaction | |
| Installation problems? | |
| Yes | 2 |
| No | 98 |
| Any difficulty completing the first training questions? | |
| Yes | 7 |
| No | 94 |
| Length of installation? | |
| 2–5 minutes | 52 |
| 6–10 minutes | 41 |
| 11–15 minutes | 4 |
| 16–20 minutes | 2 |
| Content satisfaction | |
| Overall, I think the Health Buddy questions are | |
| Very easy | 44 |
| Somewhat easy | 16 |
| Neutral | 32 |
| Somewhat difficult | 4 |
| Difficult | 4 |
| Repeating questions reinforced knowledge and understanding | |
| Strongly agree | 56 |
| Somewhat agree | 28 |
| Neutral | 12 |
| Somewhat disagree | 4 |
| Strongly disagree | 0 |
| Understanding of my health condition | |
| Much better | 64 |
| Somewhat better | 20 |
| Neutral | 16 |
| Somewhat worse | 0 |
| Much worse | 0 |
| Managing my health condition | |
| Much better | 52 |
| Somewhat better | 44 |
| Neutral | 4 |
| Somewhat worse | 0 |
| Much worse | 0 |
| Recommend the device to others | |
| Very willing | 80 |
| Somewhat willing | 12 |
| Neutral | 4 |
| Somewhat unwilling | 0 |
| Very unwilling | 4 |
| Overall satisfaction | |
| Satisfaction with device | |
| Very satisfied | 45 |
| Satisfied | 35 |
| Somewhat satisfied | 15 |
| Not very satisfied | 5 |
| Satisfaction with the communication between you and your doctor or nurse | |
| More satisfied | 65 |
| No difference | 30 |
| Less satisfied | 5 |
| Ease of using the device | |
| Very easy | 85 |
| Easy | 15 |
| Not easy | 0 |
| Overall experience with the device | |
| Positive | 85 |
| Neutral | 15 |
| Negative | 0 |
| Continue to use the device | |
| Very likely | 40 |
| Likely | 40 |
| Somewhat likely | 15 |
| Not very likely | 0 |
Narrative comments
During the exit interview, participants were asked, “How was having the Health Buddy helpful to you?” Responses could be categorized into two major themes: (1) the Health Buddy provided needed information and (2) the Health Buddy improved my self-management during treatment.
Statements made related to the information provided included the following:
- • It gave me information on what could be expected from treatment
• It was a constant reminder of things to watch for
• It kept me abreast of my total condition at all times
• It gave good directions so I didn't have to ask at the cancer center
• It gave good suggestions on treatments (home remedies) such as gargles, care of feeding tube, exhaustion, and everyday symptoms
Statements made indicative that the Health Buddy improved self-management included the following:
- • I learned what I could do to make myself feel better
• It helped me manage my symptoms
• It taught me about symptom management and how to handle problems
• It let me know whether to contact a doctor or use self-care
• It gave me who to call for problems and some things to try
• It kept me aware of what I needed to do in order to make the period easier
• It reminded me to take my meds and exercise
Additionally, some participants noted the support they felt from having the Health Buddy interventions during treatment in saying the following:
- • It kind of helped my depression through acknowledging it and giving me something to do
• It made me feel I was not the only one who had experience with these things
• It comforted me because I knew what was going to happen
Poststudy survey
Twenty (45%) of the 44 patients who received the intervention responded to the mailed poststudy survey. When asked if they felt they received better care because they had the device, 13 of the 20 (65%) responded that they did. Eighteen (90%) of the treatment group responders stated they were very satisfied with their care (one stated “somewhat satisfied”) and 20 (100%) said they would recommend the cancer center for treatment. Nineteen (95%) stated they received adequate support during treatment.
Outcome Measures
Mean scores on the FACT-H&N and subscales and the MSAS and subscales taken pre-, during, and posttreatment are displayed in Table 5. As expected, average QOL scores declined during treatment, while symptom distress increased, with a return to near baseline scores posttreatment.
| SCALE/SUBSCALE | PRETREATMENT | DURING TREATMENT | POSTTREATMENT |
|---|---|---|---|
| Total FACT-H&N | 100.3 | 85.6 | 101.5 |
| FACT-G | 74.3 | 69.4 | 78.5 |
| Trial Outcome Index | 62.6 | 46.0 | 65.0 |
| Physical Well-Being | 21.2 | 17.6 | 21.1 |
| Functional Well-Being | 15.6 | 12.5 | 17.4 |
| Emotional Well-Being | 21.1 | 22.3 | 22.2 |
| Social Well-Being | 21.1 | 22.3 | 22.2 |
| Total MSAS | 0.7 | 1.1 | 0.8 |
| Global Distress Index | 1.1 | 1.8 | 1.3 |
| Physical | 0.7 | 1.5 | 1.1 |
| Psychological | 1.1 | 1.2 | 0.8 |
Correlations
The relationships between percentage usage per patient and the following variables were evaluated: age, income, years of education, tumor stage, and percent poverty in patient's zip code. Percent poverty in zip code area was intended to be a surrogate measure of the patient's socioeconomic status. Results are displayed in Table 6. No significant correlations were noted, although years of education and percentage poverty in zip code showed a trend toward significance.
| VARIABLE (VS % USAGE) | RELATIONSHIP | |
|---|---|---|
| SPEARMAN'S RHO RS | SIGNIFICANCE (ONE-TAILED) | |
| Percent poverty in zip code | 0.213 | 0.083 |
| Age | 0.146 | 0.173 |
| Years of education | −0.325 | 0.081 |
| Income | −0.292 | 0.120 |
| Tumor stage | 0.196 | 0.122 |
| Physical Well-Being (during treatment) | 0.310 | 0.048 |
| Emotional Well-Being (during treatment) | 0.315 | 0.042 |
Although a multivariate model was planned, the lack of significant bivariate correlations precluded the need for multivariate analysis.
When percent usage was correlated with FACT-H&N total and subscales taken at baseline, during active treatment, and posttreatment, significant positive correlations were found between the percentage used and the Physical Well-Being subscale score during treatment (Spearman's rho = 0.310, P = 0.048) and between percentage used and the Emotional Well-Being subscale during treatment (Spearman's rho = 0.315, P = 0.042).
There were no significant correlations between percentage usage and the scores on the MSAS.
Discussion
Both qualitative and quantitative measures indicate that using telehealth to support symptom management during aggressive cancer treatment is both feasible and well-accepted. Patient users were not intimidated by this particular technology as it was simple to set up and use and required no previous computer training to operate. The Health Buddy was viewed as providing important and useful information. Overall, users felt that it improved their ability to self-manage their disease and the side effects of treatment and provided a sense of support and security.
Unlike other studies which use telehealth devices to monitor patient symptoms, our goal was to increase patient self-management of the symptoms experienced during intensive medical treatment, therefore avoiding increased burden on the medical system. The fact that the research nurse overseeing the responses needed to intervene only once every 25.9 days speaks to the ability of the intervention to have a positive impact on utilization of medical services.
The lack of significant relationships between usage and descriptive variables such as age and years of education suggests that the intervention was equally acceptable to all subgroups. Factors such as age, previous computer literacy, educational obtainment, and socioeconomic status did not significantly differentiate our study population in terms of compliance as verified by usage percentages.
The significant relationships found between the percentage used and the subscale scores on Physical Well-Being and Emotional Well-Being during treatment may indicate that increased use of the telemessaging intervention during treatment resulted in better physical and emotional aspects of QOL.
The high rate of daily compliance with the intervention in spite of differentiating personal variables and the severity of the treatment regimen may have been due to one or a combination of the following factors:
- • the simplicity of the technology
• the visibility of the appliance (often placed in the kitchen or living area of the home) and its flashing green light as cues to the need to respond
• the usefulness of the information provided
• the use of simple messaging language presented in an encouraging, positive manner
• affirmations related to application of the symptom management protocols suggested
• curiosity related to the day's messaging and the motivational saying which always appeared at the end
• knowledge that someone was reviewing the responses, tracking and intervening when the participant did not respond for several days
Our study supports the benefits of telehealth interventions noted by providers in a study by Sandberg et al39: opportunities for more frequent contact, greater relaxation and information due to the ability to interact in one's own home, increased accessibility by those frequently underserved, and timely medical information and monitoring. Similar to the study done in the Netherlands,[27] and [28] this study noted technological problems as the primary disadvantage; but in our intervention, we had no problems with the technology or equipment.
Although computerized technology served as a barrier to previous telehealth research, the lack of a land-based phone line was a factor preventing participation in the current study. Indeed, many participants maintained only wireless communication devices, which were not compatible with the version of the Health Buddy that was employed in this study. However, improvements in the technology since completion of this study now allow for wireless access to the appliance or provision of an independent wireless messaging device for those without such access in their own homes.
Although the data generally support the feasibility and acceptability of the telehealth-based intervention, the results should be interpreted in the context of a few study limitations. In particular, the sample size was somewhat small, and data pertaining to the socioeconomic status of participants were not available for all participants. Second, the study did not include measures of the patient's direct interactions with health-care providers during the study or specific data related to their health-care utilization (eg, emergency room visits, preventable inpatient hospitalizations, emergency calls to clinicians). The collection of more exhaustive measures of health-care utilization was limited by resources but is planned for subsequent studies. Finally, concerns regarding subject burden limited assessment of the usability of the telehealth device.
Although compliance with utilization expectations and completion of study measures was excellent during the course of the intervention, response to the follow-up survey mailed several months later was less than 50%. This low response rate was most likely due to several factors: (1) this survey was sent at the conclusion of the entire study (by this time, patients were 0–21 months past their active participation); (2) it was a mailed survey with no additional contact or follow-up effort to increase response rate; (3) participants may have felt that they had already shared their opinions in the exit interview and may have felt overburdened by study measures at this point; and (4) participants may have died, moved, or been medically unable to respond. This lack of response did limit our ability to evaluate the longitudinal impact of the intervention.
Conclusions
This telehealth intervention proved to be an acceptable and feasible means to educate and support patients during aggressive treatment for head and neck cancer. Patient compliance with telehealth interventions during periods of extreme symptom burden and declining QOL is feasible if simple technology cues the patient to participate, offers positive support and relevant education, and is targeted or tailored to their specific condition.
1 C.D. Llewellyn, M. McGurk and J. Weinman, Are psycho-social and behavioural factors related to health related-quality of life in patients with head and neck cancer?: A systematic review, Oral Oncol 41 (5) (2005), pp. 440–454. Article |
2 K.T. Vakharia, M.J. Ali and S.J. Wang, Quality-of-life impact of participation in a head and neck cancer support group, Otolaryngol Head Neck Surg 136 (3) (2007), pp. 405–410. Article |
3 P.J. Allison et al., Results of a feasibility study for a psycho-educational intervention in head and neck cancer, Psychooncology 13 (2004), pp. 482–485. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (22)
4 L.H. Karnell et al., Influence of social support on health-related quality of life outcomes in head and neck cancer, Head Neck 29 (2) (2007), pp. 143–146. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (19)
5 K.M. Petruson, E.M. Silander and E.B. Hammerlid, Effects of psychosocial intervention on quality of life in patients with head and neck cancer, Head Neck 25 (7) (2003), pp. 576–584. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (29)
6 J.R.J. deLeeuw et al., Negative and positive influences of social support on depression in patients with head and neck cancer: a prospective study, Psychooncology 9 (2000), pp. 20–28. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (46)
7 J. Ostroff et al., Interest in and barriers to participation in multiple family groups among head and neck cancer survivors and their primary family caregivers, Fam Process 43 (2) (2004), pp. 195–208. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (17)
8 R.L. Bashshur et al., National telemedicine initiatives: essential to healthcare reform, Telemed J E Health 15 (6) (2009), pp. 1–11.
9 K. Davis et al., An innovative symptom monitoring tool for people with advanced lung cancer: a pilot demonstration, J Support Oncol 5 (8) (2007), pp. 381–387. View Record in Scopus | Cited By in Scopus (8)
10 K.H. Mooney et al., Telephone-linked care for cancer symptom monitoring: A pilot study, Cancer Pract 10 (3) (2002), pp. 147–154. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (30)
11 R.H. Friedman et al., The virtual visit: using telecommunications technology to take care of patients, J Am Med Inform Assoc 4 (1997), pp. 413–425. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (64)
12 A. Weaver et al., Application of mobile phone technology for managing chemotherapy-associated side-effects, Ann Oncol 18 (11) (2007), pp. 1887–1892. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (10)
13 D. Berry et al., Computerized symptom and quality-of-life assessment for patients with cancer: Part 1: Development and pilot testing, Oncol Nurs Forum 31(5 (2004), pp. E75–E83. Full Text via CrossRef
14 K. Mullen, D. Berry and B. Zierler, Computerized symptom and quality-of-life assessment for patients with cancer: Part II: Acceptability and usability, Oncol Nurs Forum 31 (5) (2004), pp. E84–E89. Full Text via CrossRef
15 B. Fortner et al., The Cancer Care Monitor: psychometric content evaluation and pilot testing of a computer administered system for symptom screening and quality of life in adult cancer patients, J Pain Symptom Manage 26 (6) (2003), pp. 1077–1092. Article |
16 R. de Bree et al., Touch screen computer-assisted health-related quality of life and distress data collection in head and neck cancer patients, Clin Otolaryngol 33 (2) (2008), pp. 138–142. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (6)
17 H. Huang et al., Developing a computerized data collection and decision support system for cancer pain management, Comput Inform Nurs 21 (4) (2003), pp. 206–217. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (17)
18 D.J. Wilkie et al., Usability of a computerized pain report in the general public with pain and people with cancer pain, J Pain Symptom Manage 25 (3) (2003), pp. 213–224. Article |
19 K. Kroenke et al., Effect of telecare management on pain and depression in patients with cancer: a randomized trial, JAMA 304 (2) (2010), pp. 163–171. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (7)
20 N. Kearney et al., Utilizing handheld computers to monitor and support patients receiving chemotherapy: results of a UK-based feasibility study, Support Care Cancer 14 (7) (2006), pp. 742–752. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (17)
21 N.R. Chumbler et al., Remote patient–provider communication and quality of life: empirical test of a dialogic model of cancer care, J Telemed Telecare 13 (2007), pp. 20–25. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (5)
22 A.L. Grubaugh et al., Attitudes toward medical and mental health care delivered via telehealth applications among rural and urban primary care patients, J Nerv Ment Dis 196 (2) (2008), pp. 166–170. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (6)
23 J. Stalfors et al., Accuracy of tele-oncology compared with face-to-face consultation in head and neck cancer case conferences, J Telemed Telecare 7 (2001), pp. 338–343. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (5)
24 C. Dorrian et al., Head and neck cancer assessment by flexible endoscopy and telemedicine, J Telemed Telecare 15 (2009), pp. 118–121. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (3)
25 J. Stalfors et al., Haptic palpation of head and neck cancer patients—implications for education and telemedicine, Stud Health Technol Inform 81 (2001), pp. 471–474. View Record in Scopus | Cited By in Scopus (8)
26 C. Myers, Telehealth applications in head and neck oncology, J Speech Lang Pathol Audiol 29 (3) (2005), pp. 125–127.
27 J.L. van den Brink et al., Involving the patient: a prospective study on use, appreciation and effectiveness of an information system in head and neck cancer care, Int J Med Inform 74 (10) (2005), pp. 839–849. Article |
28 J.L. van den Brink et al., Impact on quality of life of a telemedicine system supporting head and neck cancer patients: a controlled trial during the postoperative period at home, J Am Med Inform Assoc 14 (2) (2007), pp. 198–205. Article |
29 B. Head et al., Development of a telehhealth intervention for head and neck cancer patients, Telemed J E Health 15 (1) (2009), pp. 100–108. View Record in Scopus | Cited By in Scopus (1)
30 D.F. Cella et al., The Functional Assessment of Cancer Therapy (FACT) scale: development and validation of the general measure, J Clin Oncol 11 (3) (1993), pp. 570–579. View Record in Scopus | Cited By in Scopus (1626)
31 M.A. List et al., The Performance Status scale for head and neck cancer patients and the Functional Assessment of Cancer Therapy-Head and Neck scale: A study of utility and validity, Cancer 77 (11) (1996), pp. 2294–2301. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (169)
32 H.M. Mehanna and R.P. Morton, Patients' views on the utility of quality of life questionnaires in head and neck cancer: a randomised trial, Clin Otolaryngol 31 (4) (2006), pp. 310–316. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (10)
33 V. Chang et al., The Memorial Symptom Assessment Scale short form, Cancer 89 (2000), pp. 1162–1171. Full Text via CrossRef
34 R.K. Portenoy et al., The Memorial Symptom Assessment Scale: an instrument for the evaluation of symptom prevalence, characteristics and distress, Eur J Cancer 30A (9) (1994), pp. 1326–1336. Abstract |
35 J.E. Tranmer et al., Measuring the symptom experience of seriously ill cancer and noncancer hospitalized patients near the end of life with the Memorial Symptom Assessment Scale, J Pain Symptom Manage 25 (5) (2003), pp. 420–429. Article |
36 V.T. Chang et al., Symptom and quality of life survey of medical oncology patients at a Veterans Affairs medical center: a role for symptom assessment, Cancer 88 (5) (2000), pp. 1175–1183. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (121)
37 J.F. Nelson et al., The symptom burden of chronic critical illness, Crit Care Med 32 (7) (2004), pp. 1527–1534. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (55)
38 L.B. Harrison et al., Detailed quality of life assessment in patients treated with primary radiotherapy for squamous cell cancer of the base of the tongue, Head Neck 19 (3) (1997), pp. 169–175. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (132)
39 J. Sandberg et al., A qualitative study of the experiences and satisfaction of direct telemedicine providers in diabetes case management, Telemed J E Health 15 (8) (2009), pp. 742–750. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (1)
1 C.D. Llewellyn, M. McGurk and J. Weinman, Are psycho-social and behavioural factors related to health related-quality of life in patients with head and neck cancer?: A systematic review, Oral Oncol 41 (5) (2005), pp. 440–454. Article |
2 K.T. Vakharia, M.J. Ali and S.J. Wang, Quality-of-life impact of participation in a head and neck cancer support group, Otolaryngol Head Neck Surg 136 (3) (2007), pp. 405–410. Article |
3 P.J. Allison et al., Results of a feasibility study for a psycho-educational intervention in head and neck cancer, Psychooncology 13 (2004), pp. 482–485. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (22)
4 L.H. Karnell et al., Influence of social support on health-related quality of life outcomes in head and neck cancer, Head Neck 29 (2) (2007), pp. 143–146. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (19)
5 K.M. Petruson, E.M. Silander and E.B. Hammerlid, Effects of psychosocial intervention on quality of life in patients with head and neck cancer, Head Neck 25 (7) (2003), pp. 576–584. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (29)
6 J.R.J. deLeeuw et al., Negative and positive influences of social support on depression in patients with head and neck cancer: a prospective study, Psychooncology 9 (2000), pp. 20–28. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (46)
7 J. Ostroff et al., Interest in and barriers to participation in multiple family groups among head and neck cancer survivors and their primary family caregivers, Fam Process 43 (2) (2004), pp. 195–208. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (17)
8 R.L. Bashshur et al., National telemedicine initiatives: essential to healthcare reform, Telemed J E Health 15 (6) (2009), pp. 1–11.
9 K. Davis et al., An innovative symptom monitoring tool for people with advanced lung cancer: a pilot demonstration, J Support Oncol 5 (8) (2007), pp. 381–387. View Record in Scopus | Cited By in Scopus (8)
10 K.H. Mooney et al., Telephone-linked care for cancer symptom monitoring: A pilot study, Cancer Pract 10 (3) (2002), pp. 147–154. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (30)
11 R.H. Friedman et al., The virtual visit: using telecommunications technology to take care of patients, J Am Med Inform Assoc 4 (1997), pp. 413–425. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (64)
12 A. Weaver et al., Application of mobile phone technology for managing chemotherapy-associated side-effects, Ann Oncol 18 (11) (2007), pp. 1887–1892. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (10)
13 D. Berry et al., Computerized symptom and quality-of-life assessment for patients with cancer: Part 1: Development and pilot testing, Oncol Nurs Forum 31(5 (2004), pp. E75–E83. Full Text via CrossRef
14 K. Mullen, D. Berry and B. Zierler, Computerized symptom and quality-of-life assessment for patients with cancer: Part II: Acceptability and usability, Oncol Nurs Forum 31 (5) (2004), pp. E84–E89. Full Text via CrossRef
15 B. Fortner et al., The Cancer Care Monitor: psychometric content evaluation and pilot testing of a computer administered system for symptom screening and quality of life in adult cancer patients, J Pain Symptom Manage 26 (6) (2003), pp. 1077–1092. Article |
16 R. de Bree et al., Touch screen computer-assisted health-related quality of life and distress data collection in head and neck cancer patients, Clin Otolaryngol 33 (2) (2008), pp. 138–142. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (6)
17 H. Huang et al., Developing a computerized data collection and decision support system for cancer pain management, Comput Inform Nurs 21 (4) (2003), pp. 206–217. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (17)
18 D.J. Wilkie et al., Usability of a computerized pain report in the general public with pain and people with cancer pain, J Pain Symptom Manage 25 (3) (2003), pp. 213–224. Article |
19 K. Kroenke et al., Effect of telecare management on pain and depression in patients with cancer: a randomized trial, JAMA 304 (2) (2010), pp. 163–171. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (7)
20 N. Kearney et al., Utilizing handheld computers to monitor and support patients receiving chemotherapy: results of a UK-based feasibility study, Support Care Cancer 14 (7) (2006), pp. 742–752. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (17)
21 N.R. Chumbler et al., Remote patient–provider communication and quality of life: empirical test of a dialogic model of cancer care, J Telemed Telecare 13 (2007), pp. 20–25. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (5)
22 A.L. Grubaugh et al., Attitudes toward medical and mental health care delivered via telehealth applications among rural and urban primary care patients, J Nerv Ment Dis 196 (2) (2008), pp. 166–170. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (6)
23 J. Stalfors et al., Accuracy of tele-oncology compared with face-to-face consultation in head and neck cancer case conferences, J Telemed Telecare 7 (2001), pp. 338–343. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (5)
24 C. Dorrian et al., Head and neck cancer assessment by flexible endoscopy and telemedicine, J Telemed Telecare 15 (2009), pp. 118–121. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (3)
25 J. Stalfors et al., Haptic palpation of head and neck cancer patients—implications for education and telemedicine, Stud Health Technol Inform 81 (2001), pp. 471–474. View Record in Scopus | Cited By in Scopus (8)
26 C. Myers, Telehealth applications in head and neck oncology, J Speech Lang Pathol Audiol 29 (3) (2005), pp. 125–127.
27 J.L. van den Brink et al., Involving the patient: a prospective study on use, appreciation and effectiveness of an information system in head and neck cancer care, Int J Med Inform 74 (10) (2005), pp. 839–849. Article |
28 J.L. van den Brink et al., Impact on quality of life of a telemedicine system supporting head and neck cancer patients: a controlled trial during the postoperative period at home, J Am Med Inform Assoc 14 (2) (2007), pp. 198–205. Article |
29 B. Head et al., Development of a telehhealth intervention for head and neck cancer patients, Telemed J E Health 15 (1) (2009), pp. 100–108. View Record in Scopus | Cited By in Scopus (1)
30 D.F. Cella et al., The Functional Assessment of Cancer Therapy (FACT) scale: development and validation of the general measure, J Clin Oncol 11 (3) (1993), pp. 570–579. View Record in Scopus | Cited By in Scopus (1626)
31 M.A. List et al., The Performance Status scale for head and neck cancer patients and the Functional Assessment of Cancer Therapy-Head and Neck scale: A study of utility and validity, Cancer 77 (11) (1996), pp. 2294–2301. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (169)
32 H.M. Mehanna and R.P. Morton, Patients' views on the utility of quality of life questionnaires in head and neck cancer: a randomised trial, Clin Otolaryngol 31 (4) (2006), pp. 310–316. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (10)
33 V. Chang et al., The Memorial Symptom Assessment Scale short form, Cancer 89 (2000), pp. 1162–1171. Full Text via CrossRef
34 R.K. Portenoy et al., The Memorial Symptom Assessment Scale: an instrument for the evaluation of symptom prevalence, characteristics and distress, Eur J Cancer 30A (9) (1994), pp. 1326–1336. Abstract |
35 J.E. Tranmer et al., Measuring the symptom experience of seriously ill cancer and noncancer hospitalized patients near the end of life with the Memorial Symptom Assessment Scale, J Pain Symptom Manage 25 (5) (2003), pp. 420–429. Article |
36 V.T. Chang et al., Symptom and quality of life survey of medical oncology patients at a Veterans Affairs medical center: a role for symptom assessment, Cancer 88 (5) (2000), pp. 1175–1183. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (121)
37 J.F. Nelson et al., The symptom burden of chronic critical illness, Crit Care Med 32 (7) (2004), pp. 1527–1534. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (55)
38 L.B. Harrison et al., Detailed quality of life assessment in patients treated with primary radiotherapy for squamous cell cancer of the base of the tongue, Head Neck 19 (3) (1997), pp. 169–175. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (132)
39 J. Sandberg et al., A qualitative study of the experiences and satisfaction of direct telemedicine providers in diabetes case management, Telemed J E Health 15 (8) (2009), pp. 742–750. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (1)