메시지 프레이밍 전략에 따른 위암 생존자 자가관리 교육 효과

Abstract

Introduction Gastric cancer is the 5th-most common cancer and has the 4th-highest mortality rate worldwide (Ferlay et al., 2020). In South Korea, it is the most prevalent cancer, but its 5-year survival rate has been increasing and now exceeds 77% (National Cancer Information Center, 2021). More gastric cancer survivors are becoming aware of the importance of quality of life after treatment. However, gastric cancer survivors who undergo gastrectomy, the primary treatment for gastric cancer, still face many challenges. After gastrectomy, patients can experience several physical symptoms, such as postprandial fullness, epigastric pain, indigestion, diarrhea, and nausea after diet as a result of reduced stomach volume and reconstruction (Gharagozlian et al., 2020). These symptoms lead to malnutrition in 50–60% of patients (Muscaritoli et al., 2017; Son et al., 2017), most of whom lose more than 10% of their usual weight (Climent et al., 2017) and 6–16% of their muscle mass (Yamazaki et al., 2020). These losses are the most rapid in the 6 months following surgery. Due to limitations on what they can eat, how long it takes them to eat, and their physical symptoms after diet, many patients avoid having meals with other people, which can cause feelings of isolation and lead to psychosocial issues. Therefore, it is important to manage these physical and psychosocial issues to maintain their health and well-being after gastrectomy (McCorkle et al., 2011; National Cancer Institute, 2018). Several self-management intervention studies enhancing gastrectomy patients’self-management abilities have been published(Cuthbert et al., 2019; Simonsmeier et al., 2021). They suggested future directions for self-management intervention research, such as using various theories for study frameworks, having more consistent study interventions, and using various educational strategies. Therefore, this study examined how message-framing in self-management education programs affected gastric cancer survivors’self-management. Message-framing is the way in which health information is presented and can affect recipients’health behaviors(Rothman & Salovey, 1997). The two types of framing are gain-message framing, which emphasizes the expected gains of health behaviors, and loss-message framing, which emphasizes the expected losses caused by not doing health behaviors. This originated in Prospect Theory, which suggests that people respond differently to perceived gains or losses related to health behaviors (Kahneman & Tversky, 1979). According to Prospect Theory, when people are confronted with a choice in which risk is embedded, they tend to become risk-seeking. Numerous studies have found that gain-message framing increases the degree to which people engage in health-promoting behaviors where undesirable consequences are not embedded in the behaviors, such as exercise. In contrast, loss-message framing is more effective for risk-embedded health behaviors, such as cancer screening because patients can be diagnosed with cancer (risk) (Keller & Lehmann, 2008; Gallagher & Updegraff, 2012; Williamson et al., 2020). However, these findings may not be applied to cancer survivors because most of these studies were conducted on general populations, not those with health problems. Cancer survivors may have different risk perceptions of health behaviors due to their circumstances (Latimer et al., 2010; Gallagher & Updegraff, 2012; Updegraff & Rothman, 2013). Only one study evaluated the effect of message-framing strategies on physical activity among colorectal cancer survivors after 1 and 12 months (Hirschey et al., 2016). Two groups of participants were provided with a brochure about physical activity that had either a gain or loss framing. The loss groups engaged in more amount of exercise at either point in time, but it was not a statistically significantly different. However, a brochure using message-framing may not have been strong enough to cause cancer survivors’behavioral changes. Hence, a study examining the effects of message-framing in a stronger and more comprehensive intervention that also addresses nutrition and distress is needed. This study was conducted to test the effect of message-framing in a self-management education program for gastric cancer survivors on health outcomes and health behaviors about diet, physical activity, and distress. Methods Program development The education program consisted of three components: 1) face-to-face education using a booklet; 2) gain- or loss-framed text messages; and 3) self-monitoring of health behaviors. Booklet content was invented based on several guidelines for cancer survivors (Rock et al., 2012; Park et al., 2015; Muscaritoli et al., 2021; National Comprehensive Cancer Network, 2021, 2022) and was validated by four clinicians working in gastric cancer departments and three nurses with at least 5-years of experience caring for gastric cancer survivors. The text messages were approximately 100 words long and were sent once per week for 8 weeks. Their content was validated by a health communication expert. Participants monitored their health behaviors by responding to a survey online, the link for which was sent to them by text message every other week. The survey included 9 items that asked participants to reflect on their health behaviors of the preceding week. Program evaluation Study design This study had a two-group block randomization pre-post study design with a block size of 4. Participants were randomly assigned to groups using a random number generator. The analysis was double-blinded because participants were assigned to groups after having received the initial education and text messages were sent automatically. Participants The participants were 64 gastric cancer survivors who visited the outpatient clinic of a university hospital in Seoul, South Korea after undergoing gastrectomy. Convenient sampling was employed and the inclusion criteria were as follows: 1) have had early gastric cancer and 2) visited the hospital for their 3-month follow-up appointment. The exclusion criteria were as follows: 1) have difficulty in using cell phones or their text messaging function, 2) have difficulty understanding the education, 3) have a cancer history beyond the gastric cancer, 4) have been concurrently diagnosed with gastric and another form of cancer, and 5) not being aware of their cancer diagnosis. The sample size was calculated based on a previous study that investigated the effect of message-framing strategies on physical activity among colorectal cancer survivors (Hirschey et al., 2016) using G*power (Faul, 2007) (effect size = 0.53, power = 0.8, p-value = 0.05). Participants were randomly allocated to two groups of gain-message framing group and loss-message framing group and each group required 40 participants to produce statistically meaningful results. However, only 64 participants were recruited because of restrictions on data collection due to the COVID-19 pandemic. Measures 1. Health outcomes 1) Nutritional status The Patient-generated Subjective Global Assessment short form was used to measure nutritional status (Ottery, 1996). It is comprised of 6 items across 4 domains that measure respondents’height, current weight, weight a month ago, food intake, physical symptoms, and level of functioning. The range of possible scores is 0–36 with higher scores indicating poorer nutritional status. 2) Physical activity The International Physical Activity Questionnaire was used to measure participants’physical activity. It calculates the metabolic equivalents (MET) of walking and moderate- and vigorous-intensity activities engaged in in the preceding week (Craig et al., 2003). The original and Korean versions of the instrument have Spearman’s rho correlation coefficients of 0.30 and 0.27, respectively, indicating that they were valid. 3) Exercise intensity The Borg Category-Ratio-Scale was used to measure the average exercise intensity over the preceding week (Borg, 1982, 1998, 2004). It consists of a single question and the range of possible scores is 0–10 with higher scores indicating greater intensity. 4) Distress The Distress Thermometer developed by National Comprehensive Cancer Network was used to measure the overall distress (National Comprehensive Cancer Network, 2022). It consists of a single question and the range of possible scores is 0–10 with higher scores indicating higher distress levels. 2. Health behaviors The health behavior assessment tools were developed by the research team for participants to self-monitor their health behaviors throughout the intervention period. 1) Dietary habits Dietary habits were evaluated with 4 questions about participants’ eating habits over the preceding week: “Did you chew your food well?”, “Did you eat small amounts of food at one time?”, “Did you drink water after at least 30 minutes after you finished your food?”, and “Did you relax for about 30 minutes after having meals?” Responses could be made along a 5-point Likert scale with higher scores indicating better dietary habits. 2) Physical activity performance Physical activity performance was evaluated with 4 questions about participants’physical activities over the preceding week aside from physical activity measured by the International Physical Activity Questionnaire: “Did you do physical activities regularly?”, “Did you do muscle training?”, “Did you try to reduce time in sedentary?”, and “Did you increase your exercise intensity within the last two weeks?” Responses could be made along a 5-point Likert scale with higher scores indicating better performance of physical activity. 3) Distress management Distress management was evaluated with a single question that asked participants the number of minutes they had spent managing their distress over the preceding week. 3. Participant characteristics Participants’demographic characteristics included their gender, age, marital status, educational achievement, employment status, presence of a caregiver. Their clinical characteristics included their underlying diseases, the type of gastrectomy they received, and the days passed since gastrectomy. 4. Control variables 1) Perceived benefits The perceived benefits measurement tool was developed by Moon (1990) and revised and validated by Im and Jeon (2019). It consists of 7 items that can be responded to along a 5-point Likert scale. The Cronbach’s αs of the original version, revised version, and the version used in this study were 0.73, 0.94, and 0.90 respectively. 2) Perceived barriers The perceived barriers measurement tool was also developed by Moon (1990) and revised and validated by Im and Jeon (2019). It consists of 7 items that can be responded to along a 5-point Likert scale. The Cronbach’s αs of the original version, revised version, and the version used in this study were 0.67, 0.81, and 0.71 respectively. Data collection The pre-test was conducted at the outpatient clinic of a university hospital in Seoul, South Korea between May 2021 and November 2021. The researcher met those who were eligible to participate in a private conference room. Only those who agreed to participate completed the pre-test questionnaire. The researcher provided the participants with a copy of the of health outcomes and health behaviors survey sealed in an envelope for post-test. Participants then received education on that day and began receiving the text message intervention on the following day. They were asked to complete health behavior surveys after 2, 4, and 6 weeks. The post-test was conducted by phone 8 weeks after the pre-test. Patients’general characteristics were obtained from their electronic medical records. Data analysis SPSS 25 was used to analyze the data analysis and statistical significance was defined as a p-value of 0.05 or less. General characteristics, health outcome, and health behavior data are displayed in terms of mean, standard deviation, frequency, and percentage. The two groups were compared using independent t-tests, chi-square tests, and fisher’s exact tests. The group and time interaction effects on health outcomes and behaviors were analyzed using the general estimating equation while controlling for perceived benefits and barriers. Ethical considerations This study was approved by the Institutional Review Board of the Yonsei University Health Center and Department of Surgery. After the researcher explained about the study, their right to withdraw from the study at any time, and expected benefits and possible harms of participating in the study, patients signed consent forms. The confidentiality of their data was assured throughout the study. Results 1. General characteristics and homogeneity More than half of the participants were men (56.3%), and the average age was 59.3±8.2 years old. Majority were married (92.2%), high school graduates (76.6%), and had spouses as their main caregivers (92.2%). Most of them had underlying comorbidities (65.6%), were treated with total gastrectomy (96.9%) and were diagnosed through National Cancer Screening Program without explicit physical symptoms (56.3%). The homogeneity test of the two groups in terms of demographic and clinical characteristics, there were no significant differences in gender, age, marital status, education level, and primary caregivers, underlying disease, surgical type, postoperative period, and clinical characteristics of diagnostic path. Also, homogeneity at baseline was achieved for all dependent variables and control variables in two groups. 2. Health outcomes The nutritional status score decreased from 5.1±3.9 to 2.7±2.1 in gain group, from 6.7±3.7, to 2.2±2.5 in loss group, indicating that the nutritional status improved after 8 months of intervention. This change was statistically significant (β=-2.41, p<.001). However, there was no significant difference between the groups (β=0.59, p=.090). There was a significant interaction effects between time and group: the loss group at T1 had significantly better nutritional status compared to the gain group at T0(β=-2.09, p=.020). The physical activity in both groups were improved at T1 compared to T0 and the time effect was marginally significant (β=641.13, p=.06). Compared to the loss group, the degree of physical activity in the gain group increased, but it was not statistically significant (β=29.68, p=.697). The interaction effects between time and group was also not statistically significant (β=-257.30, p=.575). The exercise intensity score increased from 3.5±1.3 to 3.9±1.0 in gain group, and from 3.3±1.5 to 3.7±1.1 in the loss group, and the time effect was marginally significant (β=0.44 and p=.052). The group effect of the exercise intensity was not statistically significant (β=0.08, p=.539), and the interaction effects between time and group was also not statistically significant (β=-0.03, p=.930). The distress score decreased from 3.0±2.1 to 2.3±2.1 in the gain group and from 3.2±2.4 to 2.6±1.8 in the loss group, and the time effect was marginally significant (β=-0.75, p=.007). The distress score decreased in the gain group compared to the loss group, but was not statistically significant (β=-0.08, p=.670), therefore, there was no group effect. The score of the gain group at T1 decreased compared to the that of the loss group at T0, it was not statistically significant (β=-0.13, p=.821), so there was no group and time effects. 3. Health behaviors Dietary habits score increased over time, and the time effect at T1, T2, T3 was significant compared to T0 (β=0.29, p=.018; β=0.11, p=.008; β=0.13, p=.004). For the group effect, the dietary habits score in the gain group was higher the loss group, but it was not statistically significant (β=0.03, p=.731). The interaction effect between time and group, the loss group scored higher than the gain group score at all points, but only at T4, the interaction effect between time and group (β=0.36, p=.033). The physical activity performance score continued to increase in both groups, and it was statistically significant higher at all points in T1-T4 compared to T0 (β=0.57, p<.001; β=0.60, p<.001; β=0.38, p=.014; β=0.63, p<.001). The group comparison showed that the exercise performance score in the gain group increased more than in the loss group, but was not statistically significant (β=0.02, p=830). The interaction between time and group was not significant at all four points. The time effect of The distress management time was not statistically significant at all points in T1-T4. The distress management time of the gain group was greater than that of the loss group, but was not statistically significant (β=13.28, p=.643). The interaction effects between time and group was not significant at all points (β=27.38, p=.610; β=-44.30, p=.462; β=-8.64, p=.897; β=-0.94; p=.881). Discussion The purpose of this study was to examine the effect of gain and loss message-framing for patients with gastric cancer on health behaviors and outcomes. The loss message-framing group had better outcomes than the gain message-framing group for some outcomes. However, studies have shown that gain-message framing is more effective at encouraging people to engage in health behaviors with no risks, such as diet and physical activity, and loss-framed messages is more effective at motivating people to engage in health behaviors when they perceive those behaviors as risky (Gallagher & Updegraff, 2012). The inconsistency in these studies’ and this study’s results may be a product of the fact that those studies examined the general population (Latimer et al., 2010; Williams et al., 2019; Williamson et al., 2020), as previous studies examining the effects of message-framing that targeted patients showed similar findings (Hirschey et al., 2016; Li et al., 2017). It can be assumed that the patients and the general population might perceive health behaviors differently. Therefore, message recipients’risk perceptions of health behaviors’risks should be assessed in a future study because it could moderate the effect of message-framing. Dietary habits and nutritional status were the only variables that differed significantly by group over time. This result was likely a product of the fact that effects of message-framing interact with message recipients’personal interests, characteristics and experiences, or issue of involvement and so affects their health behaviors (Rothman et al., 1993; Rothman et al., 2006; Updegraff & Rothman, 2013). Often the biggest issue that gastrectomy patients find painful is the symptoms related to food intake, so they may find eating itself unpleasant (Davis & Ripley, 2017; Dumon & Dempsey, 2019; Gharagozlian et al., 2020). Therefore, participants might be more responsive to loss-framed message only in health information that patients are interested in, or highly involved in. Furthermore, participants may have had a higher issue-of-involvement regarding diet due to quick response times. When gastrectomy patients do not follow dietary recommendations, they experience physical symptoms relatively quickly, approximately 15–30 minutes after eating (Seoul National University Bundang Hospital, 2021), whereas not exercising or managing distress does not cause an immediate effect (Clark, 2016). Thus, participants may have felt that nutrition and dietary information was more closely related. All outcome variables improved after 8 weeks. The average nutritional status score for the pre-test was 6, which is classified as moderate malnutrition, but it decreased to 2 by the post-test. Both groups had physical activity levels of 1,800–2,000 and 2,000–2400 METs at the pre- and post-tests, respectively, although both exceeded the minimum of 600 METs required for classification as moderate physical activity. Both groups rated their distress as 3 out of 10 at pre-test, and 2 after 8 weeks. This study did not have a control group, so it is unclear whether the changes observed in this study were a result of the education program, participant maturation, or the natural result of the passage of time. However, most participants expressed high levels of satisfaction with and gratitude for the education program when asked for feedback after the post-test, so the self-management education effects may have exceeded those of the message-framing. Future studies should have control groups and evaluate the effectiveness of education programs as well as message-framing. This study had three limitations. First, this study did not have the targeted sample size due to delays in data collection as a result of the COVID-19 pandemic. The second limitation was that there were significantly fewer participants over 73 years old than those in other age groups because of their reluctance or inability to use text messaging and other barriers to participation. Of the 13 candidates who were excluded from the study, 12 were excluded because of difficulties using text messaging and completing the surveys. This pattern indicates that specific self-management interventions should be developed that minimize burdens on older patients. The third limitation was that the health behavior variables were not subject to reliability or validity testing, which could limit the internal validity of the study. Conclusion This study tested the effects of gain and loss message-framing on gastric cancer survivors’ health behaviors. The results showed that loss-framed messages were significantly more effective than gain-framed messages at promoting nutrition-related health behaviors, but otherwise there was no significant difference between the two. The effects of message-framing on patient education should be further studied. These studies should also examine the survivors of other types of cancer and patients with chronic diseases to better understand message-framing. Other variables should also be examined, such as perceptions of health behavior- and illness-related risks, educational needs, and other personal characteristics and evaluate how they moderate the relationship between message-framing and health outcomes. Text messages are limited in terms of the content that can be delivered, so other media should be tested, such as video clips. This study's results can be used to highlight the importance of educational strategies that can enhance the effectiveness of patient self-management.

위암은 국내 1위 발생 암으로, 주 치료법인 위 절제술 이후 위암 생존자는 다양한 영양문제, 체중 및 근력감소, 및 심리사회적 문제를 경험한다. 이에 이들의 자가관리의 중요성이 대두되나 기존 관련 연구들은 시공간적 제약이 크며, 효과적인 교육 전략에 대한 고려가 부족하다. 본 연구의 목적은 위 절제술 이후 3개월 시점의 위암 생존자를 대상으로 영양, 신체활동, 디스트레스에 있어서 메시지 프레이밍 전략을 활용한 자가관리 교육 프로그램을 개발하고, 이를 토대로 메시지 프레이밍 종류에 따른 자가관리 교육의 효과를 평가하는 것이다. 연구의 이론적 기틀로 건강신념모델과, 메시지 프레이밍 개념의 근간이 되는 전망이론을 활용하였고, 연구 추진은 크게 프로그램 개발과 프로그램 적용의 두 단계로 이루어졌다. 먼저 프로그램 개발 단계에서는 여러 위암 생존자 가이드라인을 참고하여 영양, 신체활동, 디스트레스를 교육 주제로 선정한 후, 면대면 소책자 교육, 이득/손실 메시지 프레이밍을 활용한 문자 메시지 중재, 건강행위 자가 모니터링의 세 가지 요소로 구성된 총 8주간의 교육 프로그램을 개발하였다. 프로그램 초안에 대해서는 전문가 타당도 검증을 거쳤으며 3명의 위암 생존자에게 타당성 조사를 시행한 후 최종 교육 프로그램을 도출하였다. 프로그램 적용 단계는 두 군 블록 무작위 배정 사전사후 설계 연구로 서울시 일 상급종합병원에서 위암 진단 후 위 절제술 시행하고 3개월이 경과한 암 생존자들 64명을 이득 메시지 프레이밍 군과 손실 메시지 프레이밍 군으로 무작위 배정하여 교육 프로그램 효과를 평가하였다. 종속변수 및 측정도구로는 영양상태(Patient-Generated Subjective Global Assessment), 신체활동(International Physical Activity Questionnaire), 운동강도(Borg Category Ratio-Scale), 디스트레스(National Comprehensive Cancer Network Distress Thermometer)의 건강결과변수와, 대상자들의 자가 모니터링을 위해 연구자가 개발한 식습관 실천, 운동 수행, 디스트레스 관리에 대한 건강행위변수를 포함하였다. 자료분석은 SPSS 25 프로그램을 이용하여 기술통계, Independent t-test, Chi-square test를 수행하였고, 시간과 군의 교호작용에 대한 교육 프로그램 효과 평가는 일반화추정방정식으로 분석하였으며, 이 때 건강신념모델을 토대로 지각된 유익성 및 지각된 장애성 사전 점수는 보정하였다. 구체적인 연구결과는 다음과 같다. 1. 대상자 특성을 살펴보면 남성이 36명(56.3%), 평균 연령은 59.3±8.2세였으며 대부분 기혼(92.2%), 고졸이상(76.6%), 배우자가 주 보호자(92.2%)였다. 기저질환이 있는 경우가 과반수 이상(65.6%)으로 대부분은 위아전절제술을 받았으며(96.9%), 신체증상없이 국가암진단으로 위암 진단을 받았다(56.3%). 사전조사 시 영양상태, 신체활동, 운동강도, 디스트레스, 식습관, 운동, 디스트레스 관리, 지각된 유익성 및 지각된 장애성은 동질성이 확보되었다. 2. 이득군의 사전점수에 비해 손실군의 사후점수가 유의하게 호전되는 것으로 나타나(β=-2.09, p=.020) 부가설 1-1인 “두 군의 영양상태는 시간에 따라 통계적으로 유의한 차이가 있을 것이다”는 지지 되었으나, 부가설 1-2인 “두 군의 신체활동 정도는 시간에 따라 통계적으로 유의한 차이가 있을 것이다.”, 부가설 1-3인 “두 군의 운동강도는 시간에 따라 통계적으로 유의한 차이가 있을 것이다.", 부가설 1-4인 “두 군의 디스트레스는 시간에 따라 통계적으로 유의한 차이가 있을 것이다.” 는 모두 기각되었다. 3. T4시점에서 이득군에 비해 손실군의 점수가 유의하게 높아(β=0.36, p=.033) 부가설 2-1 “두 군의 식습관 실천정도는 시간에 따라 시간에 따라 통계적으로 유의한 차이가 있을 것이다.” 는 부분적으로 지지되었다. 반면 부가설 2-2 “두 군의 운동 수행정도는 시간에 따라 통계적으로 유의한 차이가 있을 것이다.”, 부가설 2-3 “두 군의 디스트레스 관리정도는 시간에 따라 통계적으로 유의한 차이가 있을 것이다.”는 모두 기각되었다. 위 절제술 후 3개월 시점의 위암 생존자에게서 메시지 프레이밍을 활용하여 영양, 신체활동, 디스트레스에 대한 자가관리 교육 프로그램을 적용한 결과, 영양상태와 식습관 실천에 있어서 이득군에 비해 손실군에서 시간에 따른 효과가 있었음을 확인할 수 있었고, 신체활동 및 디스트레스 관련 변수들에서는 시간과 군의 유의한 교호작용이 확인되지는 않았다. 다만 두 군 모두 사후조사 시점에서 모든 측정 변수가 호전되는 양상을 확인할 수 있었다. 이와 같은 연구 결과는 위암 생존자의 관심도 또는 교육 요구도가 높은 주제에 있어서 이득 메시지 프레이밍보다 손실 메시지 프레이밍 전략이 효과적일 수 있음을 시사한다. 추후 관심도, 교육요구도뿐만 아니라 건강행위나 질병 자체에 대한 위험 인식 등이 메시지 프레이밍과 건강행위에 있어서 조절변수로써 작용하는지 확인하는 연구가 수행되어야 하겠다.