According to the Reuters Institute Digital News Report 2019 [30], The Star Online occupied 29% of the local weekly online news media use, which ranked the highest among all English newspaper websites. Sin Chew Online occupied 12% of it, which ranked the highest score among all local Chinese newspaper websites.

The unit of analysis of this study was the news article. News articles were retrieved using a keyword search of the homepages of individual newspaper websites. The keywords such as “cancer” OR “cancer prevention” OR “cancer treatment” OR “breast cancer” OR “lung cancer” OR “prostate cancer” OR “colon cancer,” etc, appearing in the headlines, subheadlines, first 3 paragraphs and lead paragraphs were used to find articles about MI on cancer prevention. We excluded news articles that covered obituaries of public figures who died from cancer and did not focus on cancer treatment or suggestions for cancer prevention. A total of 841 news articles were identified from 2017 to 2019. The process of selecting and excluding news articles for analysis is shown in Figure 1.

Interrelated coding book and coding sheet were predesigned for the data collection purpose. There are 5 variables in the coding instruments that were conceptualized based on previous studies. There are several options under each variable; the number of options that should be chosen depends on the measurement and operationalization of each variable. All the variables were identified as nominal measurements. The operational definitions of each coding variable are shown in Multimedia Appendix 1.

Two coders (the first author and a graduate student) randomly coded 10% of the identified news articles for intercoder reliability check. The graduate student received extensive training before coding the articles. We employed a Cohen kappa reliability test [37]. Based on the Altman strength of agreement, Cohen kappa value ranging from .61 to .80 indicates good agreement while values falling from .81 to 1.0 show very good agreement [38]. Our results demonstrated that some coding items obtained good agreement, such as tactical MI (k=.73, 95% CI .59-.87) and medical cancer risks (k=.74, 95% CI .59-.88), and the rest of items gained very good agreement between 2 coders, like cancer type (k=.89, 95% CI .78-.98) and presence of MI (k=.94, 95% CI .85-1.0).

The obtained data were analyzed via SPSS (version 26.0, IBM Corp). First, we applied descriptive analysis and chi-square test to answer RQ1 and RQ2, including the frequencies, percentages, and differences of each news component. We then employed binary logistic regression (forward likelihood ratio method) to determine the associations between different news components and the presence of MI in online cancer news (RQ3). The level of significance in this study was set as P<.05.

As the media has been instrumental in creating public awareness about cancer and chronic disease prevention, understanding the media’s role in mobilizing the public regarding cancer prevention is of utmost importance. This study provides the first comprehensive and comparative assessment of MI in cancer news on the selected Malaysian English and Chinese newspapers with online versions. Of importance, our results show that there were nearly 70% of the online cancer news articles in The Star Online and Sin Chew Online presented at least one type of MI. It indicates that the selected Malaysian newspaper websites have an awareness of mobilizing and calling health behavior change for its readers. However, our findings are contrary to previous studies that examined MI in cancer coverage in Western countries [22,24-26]. Several possible explanations could support our findings. First, since the past studies that examined MI in cancer news coverage were conducted at least 10 years ago, the patterns of cancer news coverage have already changed, especially for online news media [22,24-26]. Second, Hoffman [20] introduced another type of MI called interactive MI, which is only applicable for online news content. This could be one of the reasons why online cancer news contained more MI as compared to print media. Third, the increasing rates of cancer incident and mortality around the world [1], the seriousness of the disease, health information needs among the people, and the media attention of cancer issues that differ from the past have become another dependable rationale to support our results. Fourth, culture and social background would give impacts to health coverage [39]. Thus, our results report that the characteristics of cancer news coverage in the selected Malaysian newspaper websites were quite different from the Western countries. Of note, we also found that cancer news articles in Sin Chew Online were more likely to present MI than those in The Star Online. This may due to the increased cancer incident rate reported among the Chinese population [4,5], and Sin Chew Online realized that it is crucial to promote and mobilize cancer prevention to its targeted readers.

Our results illustrated that there was an unbalanced monthly distribution in terms of the number of cancer news articles that provided MI. In certain months, there were more cancer news articles found providing MI, such as in February, October, and November. This may due to the fact that World Cancer Day falls in February, Children’s Cancer Awareness Month falls in September, and Breast Cancer Awareness Month falls in October. These occasions gained more media attention because of shared journalistic values; journalists usually cover health topics according to the allocation of relevant campaigns or events [40]. In the same vein, Varga and colleagues [40] reported that the media coverage on certain cancers mushroomed during their awareness months, such as breast cancer and prostate cancer, and especially on social media platforms. However, in certain months, there were fewer cancer news articles reporting MI, such as in June and July. These months are the busier months throughout the year as there are many local festivals and school holidays happen in these months. The Hari Raya Aidilfitri, Gawai Dayak, King’s birthday, and Dragon Boat Festival happen in these 2 months. Negative news such as reporting cancer issues and related death tolls would likely give way to positive news and announcements such as big sales and discounts, tourism, and Open House. Hence, the news distribution of MI on cancer prevention was found less in June and July. Speaking of the types of MI, we found that the majority of cancer news provided tactical MI, which contained how-to health information and detailed suggestions about cancer prevention. Such news provides a clear picture to the readers of how to prevent cancers in their daily life.

We found several differences between The Star Online and Sin Chew Online regarding news components providing MI in online cancer news. First, the results indicated that there was a significant difference in terms of cancer types portrayed in the 2 studied online news websites. More than half of cancer news that provided MI focused more than one specific type of cancer. It mainly covered health campaigns or events related to cancer prevention, followed by highly detectable cancer and highly preventable cancer. This phenomenon is consistent with the previous study, which applied the same categorization of cancer types [31]. By comparison, cancer news from Sin Chew Online focused more on general cancer topics while The Star Online is more likely to cover highly preventable cancer issues.

For news source, the results showed that the majority of cancer coverage that provided MI interviewed professionals from medical institutions, such as public and private hospitals, universities or laboratories, followed by NGOs. This corroborates with previous studies that examined sourcing practices in various health coverage [41,42]. Interestingly, we found that cancer news articles that provided MI in The Star Online were more likely to cite medical journals from overseas countries, while those in Sin Chew Online were more likely to interview government agencies and NGOs from the local Chinese society, such as ancestral connective organizations and religious organizations serve for Chinese communities (ie, Chinese temples and Chinese-dominant churches). This shows that cancer coverage from different languages has varying source preferences subject to their journalistic values and social/cultural norms [39].

Pertaining to news focus, most cancer news providing MI focused on primary cancer prevention, such as local educational events or social marketing campaigns related to cancer issues, followed by secondary cancer prevention, like mobilizations on cancer screening for early detection and other medical interventions. Since the characteristics of MI include triggers and call the readiness among individuals, it is reasonable that the majority of cancer coverage that provided MI focused on prevention, rather than other topics like cancer treatment or cancer research [20]. We found that cancer news providing MI in The Star Online focused slightly more on topics related to secondary cancer prevention and medical treatment as compared to the coverage reported in Sin Chew Online, which was more likely to concentrate on primary cancer prevention and social support topics. This phenomenon reflects Malaysian media ecology when it comes to health news coverage.

As for cancer risk factors, our results showed different findings compared to the previous study where this description was adopted [35]. More than half of the examined news coverage mentioned demographic risks, like cancer-related with a specific gender or age. Most of these demographic risks pointed to female cancers, such as cervical cancer prevention, HPV vaccination, and breast cancer self-checking. Almost half of cancer news articles mentioned medical risks and lifestyle risks when providing MI. It corroborates the dominant characteristics of MI that promotes behavioral change on health issues [20], which advocates changes and adoptions of health behavior, as well as an understanding of medical knowledge and preventive strategies among the readers. However, it is worth noting that there was rarely cancer coverage that provided MI on environmental risks, such as natural resources pollutions, radiations, and occupational risks. The reason could be due to lack of knowledge on environmental issues and their association with cancer morbidity and mortality [43]. By comparing the two newspaper websites, we found that cancer news from Sin Chew Online was more likely to mention lifestyle risks when providing MI. It may subject to the cancer incidence in local Chinese communities; reducing risk lifestyle is one of the dominant approaches to avoid cancer occurrence [5,6].

The logistic regression analysis unveiled the in-depth findings of this study. This method is inspired by previous studies that used the same approach to analyze other media messages [34,44]. We examined the associations between 3 critical news components (ie, news sources, news focus, and cancer risk factors) and the presence of MI in online cancer news separately. For the news sources, our results illustrated that the cancer news articles were most likely to present MI when NGOs were interviewed as the source, followed by medical institutions and medical journals. This result agrees with a previous study finding that elite sources are the dominant news sources in health news coverage [32]. This study suggests that health journalists should be more active in engaging with NGOs and medical institutions. On this count, NGOs could bring up-to-date resources for events and campaigns regarding cancer prevention, and medical institutions can provide credible and applicable health information for cancer prevention.

A finding of interest was that MI was more likely to be presented in online cancer news when articles focused on primary cancer prevention, such as health education, consultation, health events or campaigns, followed by secondary cancer prevention, like promotions of early detection, screening, or diagnosis. On the other hand, MI was less likely to be presented when articles covered topics related to medical research. Thus, we recommend that health journalists should devote more attention on how to cover topics related to cancer education, local events, and campaigns, which would mobilize and persuade the newsreaders to take proactive actions to prevent cancer instead of merely providing medical research-related news retrieved from collaborative news agencies.

Regarding cancer risk factors, we found that even though every type of cancer risk was positively associated with the presence of MI, the probability level was different. The articles that reported lifestyle risks such as obesity, alcohol consumption, or smoking causing cancer were positively associated with the presence of MI in online cancer news. This is consistent with previous findings which showed that since various cancers occurring today are related to unhealthy lifestyles, cancer news coverage switched their major attention from highlighting environmental causes to discussing lifestyle causes [35]. Our results also showed that the mentions of demographic risks, like risks related to gender and age, were positively associated with the presence of MI. However, we found out that the majority of news articles that mentioned demographic risks pointed to females, which suggests that future events or campaigns could focus on mobilizing females to prevent breast and cervical cancer, as well as educating this target group to raise their awareness on cancers that most prevalent among females. There were limited news articles on demographic risks provided MI that targeted males even though prostate cancer and testicular cancer are two serious health issues for males [45]. We suggest that health journalists need to notice that cancer coverage should not skew to cancers related to females only. It is also essential to raise awareness and mobilize males to prevent cancers that are associated with them. The lack of coverage on cancers related to males calls for the need for research on effective strategies to reduce health information disparities among the male population.

This study contains some limitations. First, we only applied a keyword searching strategy for searching online news articles, and we might have missed some relevant news articles due to human errors. Future researchers should improve news sample selecting strategy by applying technology-based approaches to minimize human errors. Second, we only analyzed MI in online cancer news from the selected Malaysian newspaper websites. We did not consider the cancer news published in Malay and Tamil newspaper websites, which are two other major languages spoken in Malaysia. The representativeness of this study, hence, is limited. More studies should give attention to cancer news published in Malay and Tamil; it would draw a more comprehensive picture in terms of covering cancer issues and providing MI in Malaysian media. Third, we determined the original definition of tactical MI is ambiguous, which points to “the explicit and implicit instructions for certain behavior” [19,20]. The understanding and operationalization of this concept highly depend on different researchers and newsreaders. Thus, communication scholars need to argue and re-theorize this definition in future studies. Fourth, it is worth noting that there is a lot of variation in the products and content of the agencies used as sources, such as their stature, professional staff, mission, and other characteristics that would explain or influence the variance in data collected. In this study, the researchers only categorized the sources in generic forms. Future studies could apply textual analysis, examining the latent content of news coverage or case study to explore these variations in detail. Fifth, this study did not compare the differences or correlations of representation of MI on the internet, print, and broadcast media. Hence, future research could examine whether if different types of MI are presented differently on different mediums. Last, our results only could represent the way of presenting MI by the online news media, but the effects of MI toward individuals are unknown. Future studies should examine the effects of MI and behavioral intention or outcomes among newsreaders.

This study is one of the first scholarly attempts to throw light on analyzing and comparing MI of cancer news on selected English and Chinese newspaper websites. The findings and discussions corroborate new practical, theoretical, and methodological knowledge in health communication and journalism that would help public health researchers, medical professionals, health journalists, health policymakers, and NGOs in their researches and decision makings. The comprehensive picture of this study shows the characteristics and in-depth relationships between MI and different news components. For future studies, researchers could focus on online cancer news in other languages and also compare the differences between those published in native/national language and English. Also, experimental designs should be developed to examine the effects of MI on cancer prevention among different individuals. Furthermore, our methodological use could be applied in future studies that examine other variables in health news coverage, such as frame building and supplemental information.