Cancer is an important cause of morbidity and mortality worldwide where 18.1 million new cases of cancer and 9.6 million deaths from cancer were estimated in 2018. The cumulative risk of incidence indicates that 1 in 8 men and 1 in 10 women will develop the disease in a lifetime [1]. Long-term cancer survivors represent a sizeable portion of the population. Plant-based foods may enhance the prevention of cancer-related outcomes in these patients [2]. Plant-based diets are a diverse family of dietary patterns defined as infrequent consumption of animal foods along with frequent intake of plant-based foods in the usual diet. A high fruit and vegetable diet reduces the risk of cancer in the mouth, esophagus, lung, stomach, colon and rectum, evidence of probable risk reduction in cancer of the larynx, pancreas, breast, and bladder [3]. By the World Cancer Research Fund (WCRF) cancer prevention recommendations, 30–50% of all cancer cases are preventable by following a healthy diet and lifestyle [4]. Although these facts are widely known, only a fraction of the population follows the WCRF recommendations. Urgent action is needed to promote healthy plant-based foods in dietary guidelines to effectively reduce the risk of cancer.
Most people are now using smartphones and these devices, coupled with embedded sensors and modern communication technologies, make them an attractive technology for enabling monitoring of an individual’s health [5]. To monitor the alcohol consumption of college students, a study was conducted using smartphones and wearables which concluded by providing insights into mobile technology [6]. Mobile health (mHealth) applications (apps) have gained popularity in intervention for health behavior change [7]. Research into developing apps aimed at modifying key lifestyle behaviors associated with chronic diseases and other health issues have yielded positive findings [8]. For example, system monitoring during chemotherapy via an app has been found to lengthen survival − 75% of patients using the app were still alive, compared to 49% of non-users [9]. Reviews conducted by [10] and [11] provide comprehensive evidence that app-based mobile interventions are effective and highly promising for changing nutrition behaviors and nutrition-related health outcomes. However, there is a lack of evidence base and medical professional involvement in the development of current health apps. For example, [12] examined 185 health apps that focused on breast cancer awareness and found that only 13% were developed with professional medical input.
Despite the number of mHealth apps to recommend diets for weight control and general health management, there is still none that have attempted to recommend plant-based diets for cancer patients. To design such an app, there is a need to understand public perception about plant-based diet and nutrition apps. To do so, 73 links were web scraped to understand public perception of what they want and think about the different aspects of a plant-based diet. Web scraping approach allows an immense amount of information found on the internet to be compiled and analyzed [13], and to help aid cancer patients [14].
In this paper, we have examined many online forums from Malaysia and Singapore to address the following five research questions:
-
RQ1. Do cancer patients use plant-based diet and is it effective?
-
RQ2. What are the types, volume, frequency, and methods of cooking (where applicable) of fruits and vegetables used for prevention and condition management of cancer?
-
RQ3. How does eating a plant help with a particular organ? How does it prevent or/and help during and after the different stages of cancer?
-
RQ4. What are the existing myths/fake news about cancer?
-
RQ5. What are the requirements of a nutrition app for cancer patients?
To facilitate answering the last question, three sub-questions emerged:
-
RQ5.1. What do cancer patients need in a nutrition app?
-
RQ5.2. What are the missing features in the current cancer apps they might be using?
-
RQ5.3. How do cancer patients/caretakers search for nutrition information/recipes online?
The rest of the paper is organized as follows, with the next section detailing related work. The following sections outline the methodology used and findings of public’s perception, attitude and requirements related to plant-based diets, followed by discussions. This paper contributes to providing greater insights and understanding on how a nutrition app can be designed to help improve cancer patients’ diet and outcomes.
1.1 The role of plant-based diet in cancer prevention
Plant-based diet has been shown to protect against the 15 leading causes of death in the world, including many cancers, and may offer benefits as nutrition interventions to improve the management and treatment of these conditions [15]. Although the role of diet and lifestyle factors in health and disease is gaining more attention and emphasis, the benefits are still underestimated and undervalued. Common cited reasons for not eating fruits and vegetables are a lack of knowledge (e.g., not knowing how to cook them) and dislike for its texture, smell, and taste [16]. However, the study found that people are willing to try them when they are taught about their health benefits or how to prepare/eat them. In addition, the kind of plant-based diet and amount of intake depend on various factors such as whether a person is undergoing chemotherapy, the stage of cancer and his/her age, gender, sex, and other psychosocial factors. It is also important to ascertain if the type of nutrition diet is being recommended as prevention or to manage a particular stage of cancer. Unfortunately, there is still a lack of mHealth app that can provide detailed dietary guidance or recommendations of fruits and vegetables as prevention and condition management for cancer.
Finding of a recent study indicates that greater adherence to plant-based diet index (PDI) is inversely associated with risk of breast cancer [17]. A plant-based diet is also valuable in the primary and secondary prevention of colorectal cancer where epidemiological studies show a 46% − 88% reduced risk of colorectal cancer for those following a plant-based diet [18]. A study involving a good representation of an ethnically diverse population (including both men and women of Asian, American Indian, Black, and Caucasian ethnicities of different ages, smoker/non-smokers, consumer, and non-consumers of alcohol) reveals that a lower consumption of vegetables, fruits, fiber, and whole grains is associated with higher pancreatic cancer risk [19].
The National Cancer Society Malaysia, for example, provides useful nutrition tips for people living with cancer to be taken during cancer treatment and recovery [20]. However, the types of food, volume, and frequency change during and after treatment, and thus a personalized diet is essential. This is to ensure that users will be able to adapt to what is recommended according to their body’s changing nutritional needs. Through a cognitive approach by understanding the patients’ needs, values, and psychosocial factors involved in nutritional behavior and food-related decisions alongside other variables (sex, age, race), researchers have found that it is possible to achieve important clinical targets, to develop a personalized approach and to support concrete actions towards healthier diets thus preventing recurrences, monitoring chronic conditions, and supporting a good quality of life [21].
1.2 The impact of mobile health applications for cancer
The appeal of smartphones for assistance in health promotion concurs with the trend that more people are seeking health information via mobile devices [22]. Researchers like Wang et al. [23] have advocated for smartphone interventions for long-term health management of chronic diseases. In this context, apps provide the opportunity to bring behavioral interventions into real-life situations where people make decisions about their health [24]. A study done by Viitala et al. [25] showed that patients’ sense of security and freedom increased after using mHealth apps targeted for cancer. Research has shown that compared to those without mHealth apps, individuals with mHealth apps have significantly higher odds of using their smart devices to track progress on a health-related goal to make a health-related decision and in health-related discussions with care providers. Middelweerd et al. [27] highlights that smartphone users value health behavior apps that require low effort, are pleasant to use, are developed by credential experts, enable self-monitoring, provide advice on how to change (dietary) behavior, include positively framed alerts/reminders, provide accurate tracking functions, incorporate adequate privacy settings, and clearly show what the app will do. These factors need to be taken into consideration to improve the engagement and retention of the user [28].
The use of mobile health apps to provide help with nutrition has yielded positive findings [10, 11]. There are nutrition apps catered for cancer, such as OncoFood to help patients track their daily dietary habits [29]. However, there are limitations to the app such as patients having to take too much time to input data and many of them wish for recipe suggestions as well as to be able to make changes to existing and past data on food and prepared meals. Another study by Keaver et al. [30] reviews the quality, nutrition content and behavior techniques of 1149 apps aimed at those with cancer, but after two rounds of screening, only 12 apps were identified. There was a lack of strategies for implementation and a lack of indication on whether the information available is catered for specific cancer types or at specific stages of cancer or treatment. Out of the 12 apps, 6 apps were also providing non-evidence-based information. This study concludes that there is little nutrition information that is currently available on publicly available apps for cancer. However, only 3% of those apps have had their content developed or evaluated by health providers leaving behind the question of whether those apps are reliable or not. The challenge is in developing apps that are appropriate for health tracking, monitoring, and interventions using evidence-based strategies. In addition, there is a lack of understanding of how wearable or smartphone sensors can be used for personalized diet management and interventions [31]. Research done by Cai et al. [32] further emphasize on the need of having patients, nurses, and healthcare professionals to collaborate in the design of a mHealth app. Uncertainty is a common factor in general healthcare and knowing how to navigate it is quite useful in any health technology [33]. In Malaysia, a cancer dietary app was developed to provide healthy eating guide (advice from healthy eating to eating problems, weight loss prevention and increasing protein intake) which are uniquely tailored to the local food choices, preferences, and ingredients [34].
1.3 The use of web scraping and text analysis to identify requirements
Web scraping is a method used to extract a huge amount of information automatically from websites. It is also known as screen scraping, web data scraping or web harvesting [35]. Web scraping allows the immense amount of information found on the internet to be compiled and analyzed to make sense of what is happening in a short amount of time [13]. Currently, web scraping has mostly been used to research about food prices [36] and to extract recipes [37]. Previously, to understand the perception towards diet and food, surveys, records, 24-hour recall and questionnaires were most used [38]. The use of wearable cameras was previously implemented to understand food consumption life cycle [16]. A So far, web scraping has not been used as a digital ethnography method to find, analyze, and understand human perception on food, diet, and cancer.
A lot of people, including cancer patients, have been using social media to express how they feel, what they are going through and share with each other their own experience. For example, a breast cancer patient, Lisa Bonchek Adams, used twitter and tweeted over 176,000 times to talk about her own cancer experience [39]. Shaiket et al. [40] highlights that data analysis has been done previously to conduct an online diagnosis of diabetes with twitter data, to find out about the average happiness of cancer patients by using patient tweets, to conduct a sentiment analysis on breast cancer screening as well as many others.
To design an app for cancer patients, it is important to know what the cancer patients want and need in an app. It is common to use focus groups to find out more about a particular cancer [41, 42]. However, the use of text analysis to identify requirements for the design and development of cancer apps is still rare and little understood. While a plant-based diet is effective to improve outcomes for cancer patients, there is a lack of mHealth apps that provide reliable plant-based dietary information and recommendations for cancer patients. As a result, the public tends to just take in whatever information they have, which may not be true. As such, we conducted a text analysis to understand the perception that the public has on the above to identify requirements for a nutrition app to help improve cancer patients’ diet and outcomes.
Sentiment analysis, also known as opinion mining, is the process of automating information such as opinions, attitudes, emotions, and feelings. Sentiment analysis is usually applied to reviews and social media. It calculates the aggregate sentiment polarity and classifies the sentiment as positive, neutral, or negative [43] In sentiment analysis, results are represented in score for each term as follows: positive score (si+), neutral score (si0) and negative score (si−). Each score is used to determine how that sentence is perceived [44]. Microsoft Azure Machine Learning embodies cloud services and can be used to calculate the contribution score of the user based on the metrics and has about 100 techniques including regression, classification, text analysis and recommendation [45]. This study applied these methods to identify requirements for a plant-based nutrition app, particularly suitable for cancer patients.