Attempt to engage, yet failure to obtain successful bowel cancer screening: more likely in Māori, Pacific peoples, Asians, men and high-deprivation areas

Colorectal cancer (CRC) is the third most commonly diagnosed cancer globally and is ranked as the world’s second leading cause of cancer death.¹ A global 80% increase in new CRC cases is expected over the next two decades.^2,3 In New Zealand, CRC is consistently the second highest cause of cancer death for both genders.⁴ New Zealand survival outcomes are poor when compared to other nations.^1,5 One reason for New Zealand’s high mortality rate is delayed diagnosis, with 24% of all new CRC cases being metastatic compared with 19% in Australia and 17% in the United Kingdom (UK).⁶

The New Zealand national Census in 2018 recorded that 70.2% identify as European, followed by Māori (16.5%), Asian (15.1%), Pacific peoples (8.1%) and other ethnicities (2.7%). Māori are the Indigenous people of New Zealand, while Pacific peoples refers to the people of 40 different Pacific ethnic groups living in New Zealand. Māori and Pacific peoples encounter many barriers to healthcare in New Zealand.^7,8 The Asian ethnic group is often excluded from inequality discussion and is perceived to outperform or be similar to Europeans.⁹ Specific Māori health equity and diversity initiatives have been implemented to address health inequality, though with questionable effectiveness.¹⁰ Compared with Europeans, Māori and Pacific peoples are diagnosed with CRC at a younger age, at a more advanced stage and experience poorer health outcomes.^5,10

In 2017, New Zealand’s National Bowel Screening Programme (NBSP) began rolling out a staged faecal immunochemical test (FIT) kit CRC screening programme. Eligible residents (aged 60 to 74) are sent a test kit, instructions on how to use it in 11 languages, a consent form and a ReplyPaid envelope for the return of the completed kit. The programme was designed on evidence from a pilot scheme that began in 2012 and subsequent modelling studies.¹¹ Globally, FIT kits increase screening participation, which leads to improved health outcomes with CRC.¹² However, many factors including ethnicity, gender, language barriers and socio-economic status reduce participation in such screening.¹³ In New Zealand, the proportions of Māori and Pacific peoples participating in CRC screening are lower than expected.^14,15 Targeted interventions to improve Māori and Pacific peoples participation show modest 2–5% increases¹⁶ or even reduced participation.¹⁷ A lack of research examining the pre-diagnostic period for CRC patients in New Zealand⁵ highlights a need for greater understanding around screening participation.

One factor that is seldom a focus of research is FIT kits that have been returned but are unable to be processed, which are known as “spoilt kits”. Spoilt kits are technical errors, including reasons such as consent forms not being signed or dated, barcodes not being attached to the sample, insufficient sample collected, kits being outside expiry dates and other similar errors. Individuals who return such kits are termed “participants” in NBSP, but because their sample is not processed, they remain unscreened for CRC. NBSP follow up all spoilt kits and send replacement kits. Māori and Pacific participants are contacted by phone by the National Coordination Centre. Thus, participants who initially return a spoilt kit subsequently return a kit that can be processed and are screened for CRC. However, some participants who return a spoilt kit never go on to receive successful screening. These participants, despite being sent one or more replacement kits, do not re-attempt the screening or they send multiple kits that are spoilt. These participants are a unique population in that they have initially engaged with screening and have thus attempted but failed to be screened (ABF). As there is a general lack of research on this unique population, the aim of this current study was to examine the ABF proportions compared to total participation in New Zealand. Additionally, we aimed to examine if gender, age, ethnicity or deprivation were associated factors for ABF. This may allow targeted interventions for those individuals most at risk of ABF and improve the NBSP.

Methods

Our study is a cross-sectional analysis of data held by NBSP. At retrieval, the dataset was identified by NBSP as “Definitive spoilt kit rate all”. NBSP describe the data as counting people, not FIT kits. The dataset consisted of data on all adults who were eligible (aged between 60 and 74 years) that were invited and attempted to participate by returning a FIT kit in the NBSP between January 2012 and May 2022.

After examining the data, we defined participants and ABF in NBSP for this study as follows. Participants were individuals who returned a FIT kit for screening, including kits that could be screened and those which were spoilt. ABF were participants who remained unscreened despite having returned a kit. For example, an individual who initially returned a spoilt kit and then subsequently returned a kit that allowed that individual to be successfully screened was not ABF. However, an individual who had returned two spoilt kits and at the time of data retrieval had not returned a third kit would be classified ABF.

Data retrieved were year and month, ethnicity, sex, age group, deprivation level, number of people who were recorded ABF and number of total participants. NBSP information on residents was obtained from the New Zealand National Health Index (NHI) and the National Enrolment Service (NES). Age groups were recorded by NBSP into three categories: 60 to 64, 65 to 69 and 70 to 74 years of age. For deprivation, NBSP records Otago University’s New Zealand Index of Deprivation (NZDep)¹⁸ as an indicator of socio-economic deprivation. The NZDep estimates the relative deprivation of an area using census data relating to income, home ownership, employment, qualifications, family structure, housing, access to transport and communications. Quintile one is the least deprived and quintile five is the most deprived.

As the NBSP dataset is only available as frequency tables, this prevented the use of logistic regression. Numbers of participants and ABF were collated in contingency tables for the following categories: gender, NBSP age category, ethnicity, deprivation. The total number of participants and the proportion ABF were compared by year graphically. Identified trends were then examined using Pearson correlation coefficient and linear regression. Pearson correlation coefficient was chosen after the Kolmogorov–Smirnov test confirmed data was a Normal distribution. ABF was compared by gender using Fisher’s exact test. Differences between ethnic groups, ages and between deprivation levels were compared using Chi-squared test of independence. Each ethnicity was compared with other ethnicities using Fisher’s exact test. This was also repeated for deprivation levels by ethnicity. Odds ratios were generated with Fisher’s exact test or Chi-squared test for each contingency table examined. All data were analysed using Prism version 4.0 (GraphPad Version 4.01, San Diego, CA, USA, www.graphpad.com).

Ethical approval was obtained from the Eastern Institute of Technology (EIT) Research Ethics and Approvals Committee (ref EA03110322), Hawke’s Bay. Locality approval was obtained from the Hawke’s Bay District Health Board Research Committee (ref 22/04/354). Comment from Health New Zealand – Te Whatu Ora, National Screening Unit for publication was obtained on 17 July 2023.

Results

Between January 2012 and May 2022, NBSP recorded a total of 569,665 participants. Of these, 7,176 (1.26%) were recorded as ABF. The ABF proportions decreased by year (r=-0.98, p=0.0178) from 2012 to 2016. However, ABF proportions increased by year (r=0.90, p=0.0148) from 2016 to 2021 (Figure 1).

View Figure 1, Table 1–2.

Males were more likely (p<0.0001) to ABF compared with females (odds ratio 1.11, 95% CI 1.06 to 1.17) (Table 1).

ABF proportions differed by ethnicity (χ2=715.7, p<0.0001). Asian, Māori and Pacific ethnicities were all more likely to be ABF compared with Europeans (Table 2). Māori (χ2=21.66, p<0.0001) and Pacific peoples (χ2=36.46, p<0.0001) were more likely to be ABF compared with Asian ethnicities. Pacific peoples were not (χ2=3.01, p=0.0829) more likely to be ABF when compared with Māori.

Deprivation level had the most significant associations with odds to be classified ABF (Figure 2). Māori deprivation level five (most deprived) had 4.12 (95% CI 3.64 to 4.67, p<0.0001) the odds for being recorded as ABF compared with European deprivation level one (least deprived). At every level of deprivation, European ethnicity had lower ABF proportions compared with Asian, Māori and Pacific ethnicities. For Europeans, a change from deprivation level one to level two was significant (odds ratio 1.176, 95% CI 1.08 to 1.28, p=0.0002).

Younger (ages 60 to 64) compared with older (ages 65 to 69) and oldest (ages 70 to 74) had higher ABF proportions (χ2=151.2, p<0.0001). This same trend was statistically significant for European ethnicity (χ2=107.7, p<0.0001), but did not reach significance for Asian (χ2=5.393, p=0.0674), Māori (χ2=5.850, p=0.0537) or Pacific peoples (χ2=0.08104, p=0.9603).

Discussion

The aim of this study was to examine ABF proportions compared to total participation in New Zealand CRC screening. Additionally, we aimed to examine if gender, age, ethnicity and deprivation were associated with ABF. We suggest that as CRC screening is being rolled out across New Zealand ABF proportions are increasing, which is a concerning trend. Māori and Pacific peoples have greater odds to be classified ABF, and to a lesser extent Asian ethnicities, when compared with Europeans. Even accounting for deprivation, which generated the largest significant differences in the odds to ABF, Māori, Pacific peoples and Asian ethnicities have greater odds to ABF compared with Europeans. Men and younger individuals are disproportionally ABF when compared with female and older individuals. Targeted interventions for participants most at risk of ABF in the NBSP are required in New Zealand to achieve health equity.

We identified that the proportion of participants classified as ABF initially decreased (p=0.0178) per year from 2012 to 2016. As NBSP only began rolling out a staged FIT kit screening scheme in 2017, this initial trend was on data from the pilot scheme that began in one New Zealand location only. We speculate that the pilot scheme was successful at identifying and correcting the ABF problem. However, we have identified a concerning new trend. While the NBSP scheme has steeply increased the number of total participants (a four-fold increase in participants between 2017 and 2021), the proportion of these participants who are classified ABF is also increasing by year (r=0.90, p=0.0148). We suggest that greater efforts in reducing ABF proportions are required in New Zealand.

Asian, Māori and Pacific ethnicity participants all (p<0.0001) had greater odds to ABF compared with Europeans, demonstrating a health inequity. When controlling for deprivation, Europeans experienced lower odds for ABF compared with Asian, Māori and Pacific ethnicities. Recently, deprivation was also noted as a large factor in spoilt kits in a UK location.¹⁹ Ethnicity patterns of deprivation have remained virtually unchanged for 30 years in New Zealand and are a key determinant of health inequity.²⁰ That ethnicity and deprivation have strong associations with ABF proportions in CRC screening is concerning. Māori and Pacific peoples are diagnosed with CRC at a younger age, at a more advanced stage and experience poorer CRC health outcomes.^5,10 There is a clear need to make NBSP screening work for Māori and Pacific peoples, yet the evidence here suggests it has clearly been suboptimal to date.

Younger (ages 60 to 64) compared with older (ages 65 to 69) and oldest (ages 70 to 74) ages had higher ABF proportions. Although the age inequity was not significant for Māori and Pacific peoples, the lower relative numbers of ABF may have limited this analysis. A recent meta-analysis²¹ of 69 studies on FIT kit participation rates reported younger participants were less likely to participate in CRC screening. We suggest younger participants are also more likely to ABF. Males are more likely (p<0.001) to ABF compared with females. Men have been found to be at higher risk of returning spoilt kits, generally, in the Hawke’s Bay Region of New Zealand.¹⁵ A recent meta-analysis¹² on the efficacy of mailed FIT kits highlighted the low uptake in rural men. Men exhibit a general reluctance to engage in health promotion activities such as screening.²² A New Zealand study aiming to identify knowledge and attitudes to CRC screening²³ highlighted several New Zealand male perceptions that are likely to influence screening. These include a perceived marginalisation of men’s health, health perceived as a feminine matter and a stereotypical male “macho image”. The same factors may explain the gender bias in ABF. Not only are men less likely to take part in CRC screening generally, our findings suggest men are also less likely to be successful if they attempt to participate.

We provided the results of this study to the NBSP for comment. NBSP made clear they regarded the ABF population as a failure of the NBSP and not of the individuals who are classified ABF. As of 2023, the NBSP state they have introduced new initiatives designed to make completing the test clearer to reduce spoilt kit rates. These include a new design rolled out for the test kit, national media and primary campaigns and offering the test instructions in different languages on the NBSP website. While these are encouraging new initiatives, we note that previously it was found that Māori, Pacific peoples and males were more likely to complete the FIT kit if they had a drop-off checking location such as community pharmacies, community laboratories or primary healthcare providers.²⁴ Considering Māori, Pacific peoples, Asian ethnicities, Europeans experiencing high deprivation, men and younger age groups all experience higher odds to ABF, we suggest drop-off checking locations for all participants should be considered by NBSP. A drop-off checking service may reduce spoilt kit rates because of consent forms not being signed or dated, or barcodes not being attached to the sample, which could be immediately corrected by the participant under guidance. Additionally, the opportunity to explain in person from a local provider why the kit is spoilt may facilitate future bowel cancer screening participation. It may be possible to introduce culturally appropriate checking services even in rural settings, for example a Māori-run checking service at a local community centre.

There are several limitations from our findings. Firstly, our findings are based on secondary data collected by NBSP, meaning we cannot control its quality nor the potential for bias. Also, NBSP themselves use NHI and NES data, which has its own limitations.^25,26 We are unable to control for the effect of COVID-19 on New Zealand bowel cancer screening and spoilt kits.²⁷ NBSP was electively suspended for 3 months because the COVID-19 pandemic hit postal services, and transit delay caused the rate of spoiled kits to exceed 10%. The database examined does not allow a detailed analysis of the factors associated with a kit return and thus does not allow more specific corrective measures to be proposed. For example, establishing how a rural location influences ABF was not possible. While we have been able to identify who is at risk of ABF, we do not know why. We sought to compare these results to those found in other countries engaging in similar programmes. However, while general spoilt kit rates are commonly reported, the ABF population has not been a focus of any previous research to our knowledge. Also, the definition of spoilt kits is not interchangeable. In the UK for example, technical failures are considered separately from spoilt kits.²⁸ Further research is required in New Zealand and globally in this area.

In conclusion, we have described a unique population that have attempted but failed to be screened in a national bowel cancer screening programme. As CRC screening is being rolled out across New Zealand the proportions of ABF are increasing, a concerning trend that needs addressing. Māori, Pacific peoples and Asian ethnicities, men and younger individuals are disproportionally experiencing ABF in New Zealand. We suggest drop-off location checking services for all participants are required across New Zealand.