Interventions for cancer screening among Chinese Americans: A systematic review and meta-analysis

Background Cancer is the leading cause of death among Chinese Americans (CAs). Although death rates of cancers can be significantly reduced by screening cancers at an early stage, cancer screening (CS) rates are low among CAs. Interventions on CS may increase the uptake rates of CS and help to decrease the death rates of cancers in CAs. Objectives This study aims to summarize the intervention methods on CS among CAs and compare effects of various intervention methods on the outcomes of CS, including knowledge levels of CS, intentions to complete CS, and actual completions of CS. Methods A systematic review and meta-analysis design was used. Keyword searching was conducted on PubMed, Google Scholar, PsycINFO, and CINAHL. Inclusion and exclusion criteria were applied. The PEDro scale was used to evaluate the quality of the studies. Data was analyzed using Review Manager Version 5.4 software. Random effect model and subgroup analyses were conducted. Results The search yielded 13 eligible studies. All of the reviewed interventions were culturally tailored. Systematic review results were categorized by intervention delivery objects, intervention led, intervention contact, intervention types, and intervention focus according to group consensus. Meta-analysis results showed that the interventions on CS had a positive effect on all outcomes, including a 1.58 (95% CI, 1.17–2.14; P = 0.003), 1.78 (95% CI, 1.27–2.48; P = 0.0007), and 1.72 (95% CI, 1.22–2.42; P = 0.002) effect on knowledge of CS, intentions to complete CS, and completions of CS, respectively, compared to the control group. The subgroup analysis suggested that physician-led, individual-based, face-to-face client-focused interventions with multiple components increased CS among CAs, with the OR ranging from 1.60 (95% CI, 1.08–2.39; P = 0.02) to 3.11 (95%CI, 1.02–9.49; P = 0.05). Discussion Interventions on CS significantly increased CAs’ knowledge of CS, intentions to complete CS, and completions of CS. Physician-led, individual-based, face-to-face client-focused interventions with multiple components should be utilized for CAs.

for Chinese Americans are generally lower than those for non-Hispanic whites and are even lower among those with limited English proficiency. During 2013 and 2014, rates for cervical cancer screening with the pap test among Chinese Americans and non-Hispanic whites in the United States were 65.8% vs. 82.8%, for breast cancer screening with mammograms were 65.6% vs. 68.9%, and for colorectal cancer screening with endoscopy/FOBT were 53.6% vs. 60.5% [15], respectively. Among older Chinese Americans, prior research also found that participation in early detection cancer screening was less likely, compared to other Americans [16,17].

Cancer screening interventions
To increase the uptake rate of cancer screening, interventions which aimed to increase community demand, community access, and provider delivery have been conducted. Several studies have been conducted to evaluate the effects of these interventions on the uptake rates of breast, cervical, and colorectal cancer screening [18,19]. Researchers found that both client-focused interventions (e.g., client reminders [18,19], outreach, education, navigation, and small media including videos or tailored or untailored printed materials, such as letters, brochures, pamphlets, flyers, or newsletters distributed by healthcare systems or community groups [18]) and provider-focused interventions (e.g., clinician reminders [19], face-to-face education of clinicians [19], and provider assessment and feedback Involving evaluation of provider performance in delivering or offering screening to clients and presenting providers with information about their performance in providing screening services [18]) seem to be effective in increasing the uptake rates of screening for cancers [18]. Also, researchers found that combinations of interventions were associated with greater increases compared to single components; and repeated interventions were associated with increased annual FBT completion [18].

Outcomes of Knowledge of cancer screening, intention to screening cancers and completion of cancer screening
Participants' uptake rates of cancer screenings were significantly related to their knowledge about screenings. Previous studies have revealed that knowledge promotes women's participation in different kinds of cancer screenings [20][21][22][23]. A study conducted with participants aged 50-75 years old in South Carolina showed that higher level of knowledge was associated with a greater likelihood of having ever been screened for colorectal cancer (odds ratio [OR]: 1.05; 95% CI: 1.02-1.41; p < 0.001) [24]. Similarly, in the study conducted by Chen et al. [25] and the study conducted by Guo, Zhang, and Wu [26], results revealed that knowledge level influenced willingness towards and behaviors related to cervical cancer screening and breast cancer prevention intentions in Chinese women, respectively.
In addition, participants' intentions to screening cancers were essential for them to complete cancer screening. Researchers found that participants who formed implementation intentions (e.g., the intentions motivate the individual to act, also the individual has developed strategies and plans that promote behavioral enactment [27]) were much more likely to complete screening, compared to the participants who didn't form implementation intentions (92% vs. 69%) [28]. Evidence also suggests that implementation intentions attenuated the relationship between previous delay behavior and subsequent attendance for cervical cancer screening [28].

Research question, purpose, and significance of the study
Intervention projects on cancer screening can increase the uptake rate of cancer screening among high-risk populations. Despite findings from previous intervention studies which provided information to increase cancer screening rates among the US population, these studies suggested a need for more studies to assess one-on-one education, group education interventions, etc. [18]. Furthermore, several systematic reviews and meta-analyses have been done to examine the effects of cancer screening interventions on the uptake rates of screening among the US population [18,19]. However, to our best knowledge, no systematic review and meta-analyses have been done to examine the effects of cancer screening interventions on the uptake rates of cancer screening in Chinese Americans to date. With the supposition that the uptake of cancer screening could be impacted by culture, researchers have highlighted the importance of culture on behavior and indicated a need to assess culturally sensitive, theorybased interventions to encourage screening and reduce cancer-related health disparities [29]. From this point, a systematic review and meta-analysis to examine the effects of cancer screening interventions (e.g., culturally fitted interventions) on the uptake rate of cancer screenings among Chinese Americans is necessary.
The research questions aimed to be answered in this study were two-fold: (1) What intervention methods have been used for increasing cancer screening rates among Chinese Americans in the past ten years? and (2) Which intervention methods are effective and how effective are they? The purpose of this systematic review and meta-analysis is to investigate and summarize the intervention methods focusing on cancer screening among Chinese Americans and compare the effects of intervention methods on the outcomes of cancer screening, including the knowledge levels of cancer screening, intentions to complete cancer screening, and completions of cancer screening. This study will provide a comprehensive picture of the intervention programs which have been done on cancer screenings for Chinese Americans over the past ten years. It will also suggest an optimal way to increase cancer screening rates among Chinese Americans.

Materials & methods
We conducted a systematic review and meta-analysis to explore the study aim, according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines.

Data sources and searches
In this study, databases including PubMed, Google Scholar, PsycINFO, and CINAHL were searched. Keywords for searching were 1) Chinese Americans and 2) cancer screening related keywords, including cancer screening, mammogram, colonoscopy, FOBT, sigmoidoscopy, prostate-specific antigen, PSA, Pap, HPV, Cancer prevent � , lung cancer screening, low dose CT, and low dose computed tomography. For example, we used the Boolean search strategy: Chinese American � AND (cancer screening OR mammogram OR colonoscopy OR FOBT OR sigmoidoscopy OR prostate-specific antigen OR PSA OR Pap OR HPV OR Cancer prevent � OR lung cancer screening OR low dose CT OR low dose computed tomography) in the PubMed database to search the eligible literature. Equivalent words with similar meanings were also searched.

Study selection and eligibility criteria
To exhaust the research articles that addressed interventions on cancer screening among Chinese Americans in the past ten years, we checked the titles of the articles first, then screened the abstract and text of the articles, and inspected references from the eligible articles for further inclusion. The inclusion criteria for selecting eligible articles were: 1) peer-reviewed articles, 2) intervention studies focusing on cancer screening among Chinese Americans, or minority populations including Chinese Americans, 3) published in the English language in the last ten years through June 20, 2021, and 4) with full text available. Studies were excluded if they were 1) study protocols or other informal articles (e.g., letter to editors, commentaries, etc.) without data supported; or 2) not meeting inclusion criteria.

Data synthesis and study quality
The first and second author of this study did the initial searching in the databases separately. After identifying the eligible articles respectively, the authors had an in-depth discussion about which articles should be included and excluded. Disagreements were solved by consulting another researcher in the field. Information on the purposes, samples, study designs, methods, and results of the studies were exacted to a table of evidence to facilitate data analysis.
The methodological quality of each study was assessed using the PEDro scale [30]. The scale comprises a list of 11 criteria. Each criterion is valued by either a 0 ("No) or 1 ("Yes"), with only 10 of them used (item 2 to 11) to calculate the total score, yielding a maximum score of 10 points for each assessed study. The item 1 is used to evaluate studies' external validity, which is not included when accessing studies' PEDro score. Higher scores indicated superior methodological quality. Studies with a score lower than 4 are considered 'poor' quality, 4 to 5 are considered 'fair', 6 to 8 are considered 'good' and 9 to 10 are considered 'excellent' [30]. The agreement between the two reviewers was evaluated with the intraclass correlation coefficient (ICC).

Data analysis
We used the Review Manager Version 5.4 software to conduct the meta-analysis. Random effect model was applied in the analysis. A range of exploratory post-hoc subgroup analyses were conducted to examine the effects of the intervention delivery objects, intervention led, intervention contact, intervention types, and intervention focus on participants' completion of cancer screening. Intervention effect sizes for participants' completion of cancer screening were calculated using Hedge's g statistic and were weighed by the sample size of the studies. The Hedge's g-values were then averaged to calculate the overall effect size and converted to a z value. The Tau 2 and I 2 statistics were utilized to evaluate the included studies' heterogeneity and reveal the variance among the studies. The I 2 statistics values were categorized into no (0%-25%), low (25%-50%), moderate (50%-75%), and high (75%-100%) heterogeneity [31]. When necessary, raw data (e.g., mean with standard deviation) in the studies were converted to the other type of data (e.g., percentage). We assessed risk of publication bias within studies according to the PRISMA recommendation using a tool based on Agency for Healthcare Research and Quality's guidance. Moreover, forest plots were prepared to visualize the effect size and the odds ratio with 95% CI. Publication bias was examined visually using funnel plots. An asymmetrical funnel plot represents a potential publication bias. The first author did the data analysis and the second author reviewed and verified the results.

Descriptions of interventions based on the systematic review
Cultural characteristics of the interventions. For the interventions which were conducted in the 13 studies, all of them were culturally tailored interventions which were delivered  Cancer prevention seminar Each seminar lasted 2 hours, including 1 hour for the seminar (PowerPoint presentation) and a total of 1 hour for a brief questionnaire administered before and after the seminars. The topics in the cancer prevention seminar included the following: 1) "What is cancer and who gets cancer?" (how cancer is the leading cause of death in Asian Americans); 2) "cancer myths and facts" (cancer is not contagious, and some cancers are curable); 3) "common cancers among Chinese men and women" (lung, colorectal, liver, stomach, and prostate cancers among men; lung, colorectal, breast, cervical, and thyroid cancers among women); 4) "risk factors for common cancers; 5) "early warning signs of common cancers"; 6) "American Cancer Society cancer screening guidelines" (for breast, cervical, colorectal, and prostate cancers); and 7) "general cancer risk reduction" (what individuals can do, including weight control, regular exercise, no smoking, drink alcohol in moderation, and using sun protection in the Mandarin and/or Cantonese spoken language, or Chinese written language. Cultural characteristics such as Chinese beliefs (e.g., fatalistic views of cancer, yin-yang balance in the body, attitudes toward Western examinations, embarrassment towards diseases), social and family support, and language barriers were considered when designing the group-based interventions [42,43]. Culturally adapted materials for the individual-based interventions were provided in both Chinese and English. Several Chinese culture elements were reflected in the videos for the individual-based interventions [41,42]. For example, they used an all-Chinese cast, a soap opera set within the lives of a Chinese family, Chinese dialog featuring appropriate idioms, Chinese foods and decorations at the settings, and Chinese background music [42]. Delivery characteristics of the interventions. We organized the characteristics of the intervention delivery methods and outcomes into logical categories according to group consensus [45]. The primary comparator was usual care for the randomized control trails and the pre-intervention for the pre-post intervention studies. For trials with multiple arms, we assessed the outcomes of the culturally tailored interventions compared to usual care. The characteristics about the delivery methods and outcomes related to the interventions are summarized in Table 3.
Intervention led. Three studies were physician-led intervention studies [34,35,37] and ten were community worker or educator-led studies [32,33,36,[38][39][40][41][42][43][44]. Two of the three physician-led intervention studies included two components, which were the physician-targeted components and patient-targeted components [34,37]. The last physician-led intervention study had only one physician-targeted component, which aimed to indirectly increase the uptake rate of cancer screening among their patients [35]. In the three physician-led studies, physicians received trainings or seminars, or information materials related to screenings [34,35,37], and their patients received mailers [34] or small group sessions [37].
Intervention types. In the studies, four types of interventions were identified, including patient education, clinician education, screening kit outreach, and patient navigator (a barriers-focused intervention). Among the 13 studies, nine studies used single component interventions, including seven studies which only used the patient education method [32,33,38,39,41,42,44], and two studies used the patient navigator method [40] and clinician education method [35], respectively; the other four studies [34,36,37,43] used multiple-component interventions which included two or three components of the four intervention types.
Due to a vague and inconsistent definition of beliefs and attitudes in the available studies which tested effect of interventions on participants' attitudes and beliefs toward cancer screening, which could bring possible bias to the results, this study did not conduct further metaanalysis exploring effects of interventions on participants' attitudes and beliefs toward cancer screening. Only effects of interventions on participants' knowledge of cancer screening, intentions to complete cancer screening, and completion of cancer screening were analyzed.
Effect on participants' knowledge of cancer screening. Of the nine studies which tested the effects of interventions on participants' knowledge of cancer screening [32,33,[36][37][38][41][42][43][44], two studies were not included in the meta-analysis due to missing data on the total points which were used to measure knowledge level [36] and a vague measurement of knowledge in the report [43]. Results showed that compared to the control group, the group that received interventions on cancer screening had a significantly increased knowledge on cancer screening at post-intervention. The pooled summary effect of the interventions included was about one and a half times higher in comparison to the control (OR, 1.58; 95% CI, 1.17-2.14; P = 0.003). However, a moderate level of heterogeneity was noticed across the study results (Tau 2 = 0.1, ChI 2 = 15.39, df = 6, p = 0.02, I 2 = 61%) (Fig 2).
Effect on participants' intention to complete cancer screening. Eight studies tested effects of interventions on participants' intentions to complete cancer screening [32,33,35,37,38,[42][43][44]. Two studies were not included in the data analysis, because one study included participants' completions of cancer screening data and intentions to complete cancer screening data together [33], the other study had missing data of the total points which were used to measure participants' intentions to complete cancer screening [43]. Results showed that compared to the control group, the interventions on cancer screening significantly increased participants' intentions to complete cancer screening. The pooled summary effect of the interventions included was about 1.78 times higher in comparison to the control (OR, 1.78; 95% CI, 1.27-2.48; P = 0.0007). Also, a moderate level of heterogeneity was noticed across these study results (Tau 2 = 0.10, ChI 2 = 13.38, df = 5, p = 0.02, I 2 = 63%) (Fig 3).
Effect on participants' completion of cancer screening. Of the ten studies testing effects of interventions on participants' completions of cancer screening [33][34][35][36][37][38][39][40][41]44], one study was not included in the data analysis, because it included participants' completions of cancer screening data and intentions to complete cancer screening data together [33]. Results showed that compared to the control group, the interventions on cancer screening significantly increased participants' completions of cancer screening. The pooled summary effect of the interventions included was about 1.72 times higher in comparison to the control group (OR, 1.72; 95% CI, 1.22-2.42; P = 0.002). Nevertheless, these results should be interpreted with caution due to the presence of a high level of heterogeneity (Tau 2 = 0.18, ChI 2 = 32.25, df = 8, p<0.0001, I 2 = 75%) (Fig 4).
Comparison of the effects of physician-VS. community worker or educator-led interventions on participants' completion of cancer screening. Of the nine included studies which tested effects of interventions on participants' completions of cancer screening [34-37, 38-41, 44], three studies were physician-led intervention studies [34,35,37] and six were community worker or educator-led studies [36,[38][39][40][41]44]. Results showed that compared to the control group, both the physician-led and the community worker or educator-led interventions on cancer screening significantly increased participants' completions of cancer screening. The pooled summary effects of the physician-led and the community worker or educator-led interventions were about 2.83 times, and 1.44 times higher in comparison to the control, respectively (OR, 2.83; 95%CI, 1.18-6.79; P = 0.02 and OR, 1.44; 95% CI, 1.13-1.83; P = 0.003). Although the total heterogeneity was high across the studies (Tau 2 = 0.18, ChI 2 = 32.25, df = 8, p<0.0001, I 2 = 75%), the heterogeneity across the studies significantly decreased in the subgroup analysis on community worker or educator-led studies (Tau 2 = 0.02, ChI 2 = 6.64, df = 5, p = 0.25, I 2 = 25%) (Fig 6).

Publication bias
For each main outcome of interest, respective funnel plots were generated for evaluation of publication bias. The distribution of data points provided limited evidence for small study publication bias (Fig 10A-10C).

Discussion
This study examined the effects of interventions on cancer screening among Chinese Americans. Outcomes investigated in this study included participants' knowledge of cancer screening, intentions to complete cancer screening, and completions of cancer screening. Results showed that the interventions on cancer screening have a positive effect on all outcomes, including a 1.58, 1.78, 1.72 effect on knowledge of cancer screening, intentions to complete cancer screening, and completions of cancer screening, respectively, compared to the control group. In addition, subgroup analysis suggested individual-based, physician-led, and face-toface interventions might be a good way to increase cancer screening among Chinese Americans, with the OR ranging from 1.82 to 2.83. To date, to our best knowledge, this is the first study examined the effects of evidence-based interventions on cancer screening among Chinese Americans. Findings from this study could potentially be used for developing sensitive intervention programs to increase cancer screening rates among Chinese Americans. Utilizing appropriate intervention methods to expand Chinese Americans' knowledge on cancer screening is effective in increasing cancer screening rates among this population. Previous studies have showed that lacking knowledge toward cancer screening is a barrier for Chinese Americans to obtaining the tests [46][47][48]. Chinese Americans' lack of knowledge about the screening tests, availability of facilities that perform the tests as well as the extent of cost coverage act as barriers to screening. High-risk Chinese Americans' lack of knowledge about cancer screening might be caused by the language barrier. Health care providers can help them by providing information about cancer screening and expanded insurance coverage. In addition, educational materials written in Chinese that provide resources on cancer screening to improve their knowledge would be necessary in improving their screening rates.
Furthermore, interventions which are both linguistically and culturally adapted to Chinese Americans seem effective in increasing cancer screening rates among this population. Seminars or counseling conducted in Mandarin or Cantonese, information materials written in fifth grade reading level Chinese [30], and visual media featuring Chinese cultural characteristics could help with the screening.
In addition, findings from the meta-analyses showed a physician-led, individual-based, direct in-person face-to-face client-focused intervention with multiple components could be the optimal way to enhance cancer screening uptake among Chinese Americans. Compared to the community worker or educator-led, group-based, clinician-focused, indirect remote or self-learning interventions with a single component, the aforementioned-methods have multiple strengths which could bring benefits to high-risk Chinese Americans. First, individualbased interventions could be more personally targeted. Sensitive and individually targeted information could be provided and discussed to overcome the language barrier which may exist in the group-based interventions. Second, compared to community workers or educators, physicians are widely trusted among Chinese Americans [49]. The level of trust-in-physicians among U.S. Chinese older adults was 42.0 out of 55 [49] on the Trust in Physician Scale [50]. By building a close rapport between physicians and high-risk Chinese Americans, interventions on cancer screening could be more effective. A physician-led intervention which aims to increase both the physicians' and their patients' knowledge levels of cancer screening could be a cost-effective way to increase cancer screening rates. Through the interventions on physicians, a larger portion of high-risk patients could be identified and reached. Patients who are eligible for cancer screenings could also be further recommended to receive screening by their physicians. Third, a direct in-person face-to-face intervention could provide opportunities for high-risk Chinese Americans to interact with the interveners. Any questions raised from the intervention could be answered immediately. Also, a direct in-person face-to-face intervention could help to facilitate relationships between interveners and high-risk Chinese American participants, which is beneficial to build rapport and further increase the uptake rate of cancer screening. Fourth, single component interventions are often insufficient to lead to sustainable change. On the contrary, multiple-component interventions not only affect the desired outcomes but also multiple associated outcomes [51]. It is necessary to have multiple components to address multi-level influences simultaneously, since multiple strategies are generally more effective than a single strategy for increasing cancer screening [52]. Lastly, compared to the clinician-focused intervention, the client-focused intervention was found to be more effective in increasing the completion rates of cancer screening among Chinese Americans. Multiple client-focused interventions, such as client reminders, one-on-one education, and group education should be implemented to help high-risk Chinese Americans to raise their awareness about screening cancers, increase their knowledge level about screening cancers, and overcome the barriers to screening cancers. However, strategies need to be efficient in developing multi-component interventions since such types of intervention could be labor-and costintensive.
Lastly, a trend of the intervention methods was noticed shifting from the video [32,41,42] or phone-based or -assisted [36,40] interventions to in-person face to face educational seminars [33-35, 38, 44] from 2011 to 2018. Possible reasons for the trend may be related to the difficulties in assessing the effects of video or phone-based/assisted interventions, since participants' utilization levels of the materials in the interventions are hard to evaluate; also, given that in person face to face interventions are more effective and easier to conduct than video or phone-based interventions, further technology development would be essential to utilize video or phone-based/assisted interventions.

Limitations
This study has some limitations. First, as in all systematic reviews and meta-analyses, publication and other reporting biases may have affected our findings. Second, we found substantial heterogeneity among study effects, which diminishes the precision of our estimates for intervention effect sizes. We suspect this heterogeneity was due to the varied intervention methods, but I 2 was only partially reduced by adjusting this factor. However, given the intervention categories in which all point estimates and virtually almost all limits of 95% CIs included clinically important, we are confident about the interventions' benefit. Third, studies included in this meta-analysis were either good or fair quality studies; none of them were high quality studies given none of them met the three blinding criteria in the scale (blinding of the participants/ therapists/assessors). As they were intervention studies which aimed to increase participants' knowledge levels, participant and therapist blinding was not feasible, but blinding of the assessors was feasible given the design of the studies. Thus, in terms of the purpose of the studies, the insufficient strength of evidence reported in this review should not be interpreted as evidence that the interventions are not effective but, rather, as encouragement for additional research before effectiveness can be established.

Conclusions
This systematic review and meta-analysis showed a statistically significant increase on participants' knowledge of cancer screening, intentions to complete cancer screening, and completions of cancer screening with the implementation of cancer screening interventions. An individual-based, physician-led, and direct in-person face-to-face intervention method was suggested to be utilized in the cancer screening interventions. Future research programs and clinical practice which aim to increase the uptake rates of cancer screening among high-risk Chinese Americans should utilize language sensitive and individually targeted materials to increase this population's knowledge levels of cancer screening; provide guidelines and aid service for physicians to initiate discussions around cancer screening; and offer in-person faceto-face opportunities for high-risk Chinese Americans to share their thoughts toward cancer screening, thus increasing this population's intention to complete cancer screening and eventually increase the screening completion rate among this population. In addition, investigation about the trend of cancer screening uptake rates among the general Chinese American population is also necessary, which could help researchers and health care providers to better understand the status of cancer screening uptake among Chinese Americans.