|
   |
| ORIGINAL ARTICLE |
|
|
|
| Ahead of print
publication |
| |
Global pattern and trend of cervical cancer incidence from 1993 to 2012: Joinpoint regression and age-period-cohort analysis
Yuvaraj Krishnamoorthy, Sathish Rajaa, Dinesh K Giriyappa
Department of Preventive and Social Medicine, Jawaharlal Institute of Postgraduate Medical Education and Research, Puducherry, India
| Date of Submission | 11-Dec-2019 |
| Date of Decision | 13-Dec-2019 |
| Date of Acceptance | 25-Mar-2020 |
| Date of Web Publication | 14-May-2021 |
Correspondence Address:
Yuvaraj Krishnamoorthy,
Department of Preventive and Social Medicine, Jawaharlal Institute of Postgraduate Medical Education and Research, Puducherry
India
 Source of Support: None, Conflict of Interest: None DOI: 10.4103/ijc.IJC_1043_19
Background: Cervical cancer ranks fourth in global cancer incidence and mortality among women. A comparison of the global trends in cervical cancer would help us to identify high focus regions and serves an opportunity to evaluate the impact of the screening programs. Hence, the current study was done to assess the global trend in the incidence of cervical cancer from 1993 to 2012 among individuals aged between 30 and 79 years.
Methods: This secondary data analysis was conducted using the World Health Organization (WHO) Cancer Incidence data of five continents plus database (America, Asia, Europe, and Oceania) on the incidence of cervical cancer. Joinpoint regression was performed to determine the average annual percent change (AAPC) in cervical cancer incidence. We performed an age-period-cohort analysis to obtain age, period, and cohort-specific deviations and rate ratio (RR).
Results: Out of the four regions studied, all the regions showed a declining trend in cervical cancer incidence. The maximum decline was found in Oceania (AAPC = −3.3%) followed by America (AAPC = −2.0%). There was a consistent rise in cervical cancer incidence across the age groups in all the four continents with the maximum burden among the elderly. All the regions showed a steady decline in the rate of cervical cancer through the periods 1998–2002 to 2007–2012. There was also a steady decline in cervical cancer incidence across the cohorts from 1923–1927 to 1978–1982 in all the regions except America.
Conclusion: To summarize, cervical cancer incidence showed a declining trend globally, with the maximum decline in the Oceania region from 1993 to 2012
Keywords: Cohort studies, regression analysis, uterine cervical neoplasms
Key Message Incidence of cervical cancer has declined globally over the past two decades, with maximal decline in Oceania and American region.
How to cite this URL:
Krishnamoorthy Y, Rajaa S, Giriyappa DK. Global pattern and trend of cervical cancer incidence from 1993 to 2012: Joinpoint regression and age-period-cohort analysis. Indian J Cancer [Epub ahead of print] [cited 2021 Jun 13]. Available from: https://www.indianjcancer.com/preprintarticle.asp?id=316111 |
| » Introduction | |  |
Cancer is the major cause of death among women in both high-income and middle-income countries.[1] Cervical cancer ranks fourth in both global cancer incidence and mortality among the global women population.[2] It accounts for nearly 569,847 new cases in 2018. This burden varies considerably worldwide, with around 85% of the worldwide burden contributed by the low-middle-income countries.[3] Even across the nations, evidence suggests that women belonging to the poorest income quintile, those with lesser education levels, living in rural areas, and those facing adverse gender issues are less likely to benefit from timely prevention and detection.[4]
Several risk factors such as persistent infection with human papillomavirus (HPV), numerous sexual partners, smoking, poor genital hygiene, use of oral contraceptives, early age of sexual intercourse, and multiparity are linked with this high burden.[5],[6],[7],[8],[9] In addition to the burden and mortality, these cancers also cause substantial costs through direct medical, non-medical, and indirect costs.[9] Tackling these risk factors, through primary prevention activities, could avert about one-third to half of the total cancer burden and mortality.[10]
In recent years, global cervical cancer incidence and mortality have decreased considerably, especially among the high-income countries following the introduction of various screening strategies.[11],[12],[13],[14] The coverage of these screening programs is not globally uniform, accounting for a steady rise in cancer mortality in many low-middle-income countries.[11],[13] Thus, to achieve the elimination of cervical cancer within the 21st century according to Universal Health Coverage (UHC), it is necessary to combine intensive vaccination, screening, and treatment strategies.[15]
Worldwide comparisons of cervical cancer trends would enable us to identify higher focus regions and provides an opportunity to evaluate the effects of screening program initiated by them. A newer approach to determine the changing trends in incidence and mortality is by joinpoint regression. The time points at which significant change occurs will be better identified and evaluated using this approach.[16] Another model routinely used to depict the difference in rates across various age groups, periods, and cohorts are annual percent change (APC) analysis.[17],[18] Age effect depicts the changing risks associated across multiple age groups, while the period effect depicts the change in rates associated with all the age groups simultaneously. The cohort effect is related to a change in frequency in successive age groups across consecutive periods.[19] Hence, the study aimed at assessing the trends and deviations in the incidence of cervical cancer from 1993 to 2012 among individuals aged between 30 and 79 years using the World Health Organization (WHO) Cancer Incidence data of five continents.
| » Materials and Methods | | |
Study design and data sources
Secondary data analysis was performed using joinpoint regression and age-period-cohort model to determine the burden and trend in the incidence of cervical cancer. Incidence data for the years 1993 to 2012 were obtained from the WHO Cancer Incidence in five continents plus database (CI5 plus).[20] CI5 Plus provides access to cancer data on annual incidence rates for 41 countries from 108 cancer registries up to 2012. However, national data are available for only 21 countries. For the rest of the nations, regional data from individual records were aggregated as a proxy indicator for national incidence.
The five continents were Asia, America, Africa, Europe, and Oceania. However, we cannot include Africa in the analysis as the data were available for only one register with many zero values, which are difficult to analyze. Data were available from 1993 on Oceania while other continents had data available from 1998.
Data on new cervical cancer cases were obtained concerning site (International Classification of Disease-10 code C53), year of diagnosis, and 5-year age groups from all the available cohort-specific cancer registries (national or regional).
Statistical analysis
Joinpoint regression
First, age-standardized incidence rates per 100,000 population were calculated by direct standardization. The world standard population proposed by Segi (1960) and modified by Doll et al. (1966) was used for standardization.[21],[22]
The Joinpoint Regression Program (version 4.7.0) is a software developed by the US National Cancer Institute (NCI) to perform trend analysis for incidence and mortality rates. The annual percent change (APC) using joinpoint regression was used to examine the age-standardized rates (ASRs) during the study period.[23] The best-fitting point called “joinpoint,” where a statistically significant change occurs was identified by joinpoint analysis. Also, the trends between the joinpoints can be assessed using this analysis. The grid search method was used for fitting the segmented line regression and, hence, to determine the best fit for each model.[24] Significance was tested using the Monte-Carlo permutation method with 4,499 randomly selected datasets. P < 0.05 were considered as statistically significant.[25]
The joinpoint regression model of the natural log-transformed rates, with a maximum number of three joinpoints, was used to calculate APC, which was used to determine whether the incidence rate of cervical cancer in the model differs from the null hypothesis. 95% confidence interval (CI) for each APC was calculated to determine the statistical significance of APC in each segment. The magnitude and direction of recent trends were determined using average APC (AAPC) with 95% CI.
Age-period-cohort analysis
Oceania had ten 5-year age interval (30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–79 years) and four 5-year calendar periods (1993–1997, 1998–2002, 2003–2007 and 2008–2012) for calculation of incidence rate; whereas, only three 5-year calendar periods from 1998 were available for Asia, America, and Europe. Age and period classifications were used to identify 13 (10 + 4 −1 = 13) consecutive 5-year birth cohorts for Oceania data (i.e., 1918–1922, 1923–1927, 1928–1932, 1933–1937, 1938–1942, 1943–1947, 1948–1952, 1953–1957, 1958–1962, 1963–1967, 1968–1972, 1973–1977, 1978–82) and 12 (10 + 3 −1 = 12) consecutive 5-year birth cohorts for Asia, America, and Europe except 1918–1922. The secular trend for the incidence rate of cervical cancer was assessed for the period 1993–2012, and the age-specific incidence rates for each birth cohort were calculated.
A web tool (https://analysistools.nci.nih.gov/apc/) developed by Division of Cancer Epidemiology and Genetics under the National Cancer Institute (NCI), United States was used for age-period-cohort analysis.[26] There had been a longstanding problem in conducting analysis using the age-period-cohort model because of perfect collinearity that existed between the variables age, period, cohort. Identifiability issue arises as a result of intrinsic mathematical relation (age + cohort = period). This results in a methodological challenge as standard techniques of regression cannot be applied.[27] Hence, weighted least squares regression has been employed to partition the effects of age, period, and cohort thereby applying it to health outcomes. It overcomes the issue of identifiability by using an algorithm that optimizes the estimation for all the three components separately.
Annual percentage change (local drifts) of the age-specific rates with its test of significance was calculated. Other parameters calculated were age, period, and cohort deviations. These were done to show the local changes in the trend independent of the direction and magnitude of the overall trend. The cross-sectional age curve (i.e., expected age-specific rates in reference period adjusted for cohort effects) was also calculated and depicted. The result of the period and cohort parameters were assessed using the period rate ratio (RR) and cohort RR with 95%CI.[26]
Wald tests which follow Chi-square distribution was used for hypothesis testing and were performed to check the statistical significance of trend and deviations in the incidence rates of cervical cancer according to age, period, and cohort factors. P < 0.05 was considered statistically significant.
Ethical approval
Since the data were freely available in the public domain and the data were properly anonymized, ethical approval was not required.
| » Results | | |
Age-standardized rate (ASR) of cervical cancer incidence
Asia had the highest age-standardized rate (ASR) (19.28 per 100,000 people) of cervical cancer incidence, followed by Europe (17.53 per 100,000 people) in 2012. Oceania had the least ASR of cervical cancer incidence with 11.33 cases per 100,000 people in 2012. We also examined the trend of ASR for all these four continents using joinpoint regression during the period.
Joinpoint regression
America
In America, ASR per 100,000 population declined from 16.35 in 1998 to 12.44 in 2012 [Figure 1]a. The annual declining trend in cervical cancer incidence during the study period was significant (AAPC = −2.0%; 95% CI: −2.4% to −1.6%). The number of joinpoints (i.e., statistically significant change in the trend) was two (2004, 2009). APC in segment-1 (1998–2004) was −2.9% (95% CI: −3.3% to 2.4%), segment-2 (2004–2009) was −0.6% (−1.5% to 0.3%) and segment-3 (2009–2012) was −2.6% (−4.1% to −1.2%). | Figure 1: Age-standardized incidence trend of cervical cancer by geographical regions among females aged between 30 and 79 years (a) America (b) Asia (c) Europe (d) Oceania
Click here to view |
Asia
ASR in Asia showed a significant decline at an annual rate of −1.8% (95% CI: −2.4% to −1.2%). It declined from 24.46 in 1998 to 19.28 in 2012 per 100,000 population. [Figure 1]b. The number of joinpoint found was one (2006). APC in segment-1 (1998–2006) was −2.5% (95% CI: −3.2% to −1.8%) and segment-2 (2006–2012) was −0.9% (95% CI: −2.0% to 0.3%).
Europe
ASR in Europe declined from 20.32 in 1998 to 17.53 in 2012 per 100,000 population [Figure 1]c. There was a significant annual decline at a rate of −1.1% (95% CI: −1.9% to −0.3%). The number of joinpoints was two (2003, 2009). APC in segment-1 (1998–2003) was −1.9% (95% CI: −3.2% to −0.6%), segment-2 (2003–2009) was 0.4% (−0.9% to 1.7%) and segment-3 (2009–2012) was −2.5% (−5.5% to 0.5%).
Oceania
Oceania had a significant decline in the ASR from 19.82 in 1993 to 11.33 in 2012 per 100,000 population [Figure 1]d. APC during the study period was −3.3% (95% CI: −4.0% to −2.6%). The number of joinpoint was one (2002). APC in segment-1 (1998–2002) was −6.8% (95% CI: −7.8% to −5.7%) and segment-2 (2002–2012) was −0.1% (95% CI: −1.1% to 1.0%).
Age-period-cohort analysis
America
The cross-sectional age curve showed a consistent increase in earlier age groups from 30–34 years to 45–49 years and then showed a steady declining trend from 50–54 years to 75–79 years [Figure 2]e. Period RR showed a significant steady decline (P < 0.001) every 5 years from 1998–2002 to 2008–12 [Figure 2]f. Cohort RR showed a significant steady decline (P < 0.001) across the cohorts with a slight upward surge in the later groups between 1968–1972 and 1978–1982 [Figure 2]g. | Figure 2: Trend of Cervical Cancer incidence in America during the period 1998–2012 an age period cohort analysis; RR: Rate Ratio
Click here to view |
Asia
[Figure 3]a shows an increasing trend in the younger and middle age groups between 30 and 54 years, followed by the declining pattern of age-specific incidence rates over time among the elderly (P < 0.001). The age deviation found was a declining trend in the extremes of the age groups and an increasing trend in the middle-age groups (P < 0.001) [Figure 3]b. The analysis of period deviations [Figure 3]c showed that there were an alternate declining and increasing trend across three periods (P < 0.001). The study of cohort deviations [Figure 3]d showed an alternate increasing and decreasing trend across the cohorts (P < 0.001). | Figure 3: Trend of Cervical Cancer incidence in Asia during the period 1998–2012 an age period cohort analysis; RR: Rate Ratio
Click here to view |
The cross-sectional age curve showed a consistent increase in the age groups from 30–34 years to 55–59 years, followed by a declining trend among elderly age groups [Figure 3]e. Period RR showed a consistent declining pattern (P < 0.001) with every 5 years from 1998–2002 to 2008–2012 [Figure 3]f. Cohort RR showed a steady declining pattern (P < 0.001) across all the cohorts [Figure 3]g.
Europe
In Europe, there was an increasing trend of age-specific incidence rates over time in the earlier age groups from 30–34 years to 35–39 years following which declining trend was found [Figure 4]a with maximum decline among the elderly. The incidence was differing significantly across the age groups (P < 0.001) [Figure 4]b. The analysis of period deviations [Figure 4]c showed that there was no significant change across all three periods (P = 0.79). [Figure 4]d showed an alternate increasing and decreasing trend across all the cohorts (P < 0.001). | Figure 4: Trend of Cervical Cancer incidence in Europe during the period 1998–2012 an age period cohort analysis; RR: Rate Ratio
Click here to view |
The cross-sectional age curve showed a consistent increase in the age groups from 30–34 years to 40–44 years, followed by a declining trend from 45–49 years to 75–79 years [Figure 4]e. Period RR showed a decline (P < 0.001) with every 5 years from 1998–2002 to 2008–2012 periods [Figure 4]f. Cohort RR showed a declining trend (P < 0.001) with each of the cohorts [Figure 4]g.
Oceania
In Oceania, there was an increasing trend of age-specific incidence rates over time in the earlier age groups from 30–34 years to 50–54 years following which declining trend was found with the maximum decline among the elderly between 60 and 64 years [Figure 5]a. A significant age deviation was observed (P < 0.001) as shown in [Figure 5]b. [Figure 5]c shows a declining trend from 1993–1997 to 2003–2007 periods, followed by an increase in 2008–2012 (P < 0.001). [Figure 5]d shows an alternate increasing and decreasing trend across the cohorts (P < 0.001). | Figure 5: Trend of Cervical Cancer incidence in Oceania during the period 1993–2012 an age period cohort analysis; RR: Rate Ratio
Click here to view |
The cross-sectional age curve showed an alternate increasing and decreasing trend across all the age groups [Figure 5]e. The period RR showed in [Figure 5]f showed a declining trend with every five years from 1993–1997 to 2008–2012 periods and the cohort RR [Figure 5]g showed a declining trend (P < 0.001) across all the cohorts.
Discussion
This study demonstrates the global epidemiological profile of cervical cancer incidence. We found a substantial geographical variation in the rate, which was higher among developing regions when compared to developed nations. Asia had the highest ASR of cervical cancer incidence (19.28 per 100,000 populations in 2012) while Oceania had the lowest ASR with 11.33 per 100,000 persons in 2012. WHO has recommended that ASR of cervical cancer incidence should be reduced to less than 4 per 100,000 women-years to be considered as no longer a significant public health problem. However, the current study shows that the ASR is still way above the threshold and remains an important public health issue across all the regions. All countries must reduce the incidence well below this threshold to reach elimination status. However, all the areas showed a declining trend with a maximum decline in Oceania (AAPC = −3.3%) followed by America (AAPC = -2.0%). A similar pattern was found in a study conducted by Vaccarella et al. (2013), where a declining trend was seen in all the regions during their study period (1998–2002).[28] A decline in the Oceania region observed in the current study is more compared to the previous studies.[28],[29]
We further explored the cancer incidence trend by accessing the age, cohort, and period-specific trends. Based on this, we found that the decline in Oceania can be partly attributable to the success and rapid uptake of the National Cervical cancer Screening Programme during the early 2000s which reached the coverage rate of more than 70% during the study period.[30],[31] The screening program would typically have a period effect, and the period effect was significant in Oceania. However, many of the island countries in Oceania do not have an effective screening program. Though the period deviation is significant in Asia, the decline in this region is mainly attributed to the socioeconomic improvement with higher empowerment of women. Some of the registries in China have reported a dramatic increase in cervical cancer incidence following a steep decline, primarily due to the altered sexual behavior in the younger population. Period deviations were not significant in America and Europe, which means a factor other than screening would attribute to a decreasing trend in America and Europe. One possible factor might be the HPV prevalence in those regions. Also, this decline rate is expected to increase with the introduction of the National HPV vaccination program for girls aged 12 years in 2007.[32]
There was a consistent rise in cervical cancer incidence across the middle-aged women in all the four continents and a declining trend in the elderly age groups. This is a worrying trend as younger adults (30–44 years) had an increase in incidence during the study period. The present study also found that there was a consistent decline in cervical cancer incidence across the cohorts from 1923–1927 to 1978–1982 in continents such as Europe, Asia, and Oceania, while there was a slight upward surge in the older age groups found in America. Changes in the pattern of risk factors manifest themselves as variations in risk ratio across the successive birth cohorts. Hence, such a trend might be due to a decline in the risk factors related to cervical cancer like early age of first sexual contact, higher parity, and increased uptake of smoking among women across the regions.[33]
Limitations of our study were limited availability or unavailability of data for many countries hindering the understanding of the real global burden of cervical cancer incidence. We did not analyze based on global changes in government policy, priorities, or the overall environment in which health and social spending operate. Also, this is secondary data analysis—may not have been a true reflection of all the cases in the community, as there might be an issue of completeness or coverage of cancer registry. Future studies should explore specific epidemiological risk factors responsible and understand the complexity of the global burden of cervical cancer.
Despite these limitations, the present study results have several programmatic implications. This observed decline in cervical cancer incidence across all the regions is a pleasing finding as it is profoundly beneficial to the healthcare system and the economic status of the country. However, if the trend should be further improved to achieve the WHO goal for cervical cancer elimination, it will also reduce the demand for services that will decline dramatically in the future, and the service delivery systems can be put to use in other major public health problems. Still, there is an extensive gap existing in the coverage of screening programs especially in the low middle and lower-income regions.[31] This necessitates the need for decentralized testing and treating policy. The countries should implement population-based screening with appropriate quality.
To stay on the path for cervical cancer elimination, commitment to reaching the following targets by 2030 is necessary: 90% of girls fully vaccinated with the HPV vaccine by the age of 15 years; 70% are screened for cervical cancer at the age of 35 and 45 years, and 90% of those identified with the disease should receive appropriate care and treatment.[15] Several strategies like HPV vaccination, screening, and treatment of precancerous lesions, early detection, prompt treatment of invasive cancers, and palliative care services have been proven cost-effective across varied settings. Together they help the health system in addressing cervical cancer across the care continuum of control, elimination and eradication. Thus, it is essential to address these services by strengthening health systems, thereby achieving WHO’s goal of Universal Health Coverage (“UHC”).[15]
| » Conclusion | | |
To summarize, we found out that the incidence of cervical cancer shows a declining trend throughout the world, with a maximum decline in Oceania and America over the past two decades. However, the alarming finding is the rising trend among the middle age groups population especially among the younger adults between 30 and 44 years across all the continents. The more comprehensive packages of interventions ensuring early detection and adequate management of cervical cancer should be devised.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
| » References | | |
| 1. | Torre LA, Islami F, Siegel RL, Ward EM, Jemal A. Global cancer in women: Burden and trends. Cancer Epidemiol Biomarkers Prev 2017 26:444-57.  |
| 2. | |
| 3. | Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. GLOBOCAN 2008, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 10 [Internet]. Lyon, France: International Agency for Research on Cancer; 2010. Available from: http://globocan.iarc.fr. |
| 4. | Ginsburg O, Bray F, Coleman MP, Vanderpuye V, Eniu A, Kotha SR, et al. The global burden of women’s cancers: A grand challenge in global health. Lancet 2017;389:847-60. |
| 5. | Sadhiya JS, Sudhakar SS, Shoba K, Srinivasan K. Epidemiological trend of head and neck cancer-a retrospective study. Int J Cancer Res Prev 2016;9. |
| 6. | Plummer M, Peto J, Franceschi S. Time since first sexual intercourse and the risk of cervical cancer. Int J Cancer 2012;130:2638-44. |
| 7. | McKay J, Tenet V, Franceschi S, Chabrier A, Gheit T, Gaborieau V, et al. Immuno-related polymorphisms and cervical cancer risk: The IARC multicentric case-control study. PloS One 2017;12:e0177775. |
| 8. | International Collaboration of Epidemiological Studies of Cervical Cancer. Carcinoma of the cervix and tobacco smoking: Collaborative reanalysis of individual data on 13,541 women with carcinoma of the cervix and 23,017 women without carcinoma of the cervix from 23 epidemiological studies. Int J Cancer 2006;118:1481-95. |
| 9. | Das BC, Gopalkrishna V, Hedau S, Katiyar S. Cancer of the uterine cervix and human papillomavirus infection. Curr Sci 2000;78:12. |
| 10. | |
| 11. | Bray F, Loos AH, McCarron P, Weiderpass E, Arbyn M, Møller H, et al. Trends in cervical squamous cell carcinoma incidence in 13 European countries: changing risk and the effects of screening. Cancer Epidemiol Biomarkers Prev 2005;14:677-86. |
| 12. | Reimers LL, Anderson WF, Rosenberg PS, Henson DE, Castle PE. Etiologic heterogeneity for cervical carcinoma by histopathologic type, using comparative age-period-cohort models. Cancer Epidemiol Biomarkers Prev 2009;18:792-800. |
| 13. | Arbyn M, Raifu AO, Weiderpass E, Bray F, Anttila A. Trends of cervical cancer mortality in the member states of the European Union. Eur J Cancer 2009;45:2640-8. |
| 14. | Peto J, Gilham C, Fletcher O, Matthews FE. The cervical cancer epidemic that screening has prevented in the UK. Lancet 2004;364:249-56. |
| 15. | |
| 16. | Yu B, Huang L, Tiwari RC, Feuer EJ, Johnson KA. Modelling population-based cancer survival trends by using joinpoint models for grouped survival data. J R Stat Soc Ser A Stat Soc 2009;172:405-25. |
| 17. | Robertson C, Gandini S, Boyle P Age–period–cohort models: A comparative study of available methodologies. J Clin Epidemiol 1999;52:569-83. |
| 18. | Holford TR. Understanding the effects of age, period, and cohort on incidence and mortality rates. Ann Rev Public Health 1991;12:425-57. |
| 19. | Tobin K, Gilthorpe MS, Rooney J, Heverin M, Vajda A, Staines A, et al. Age-period-cohort analysis of trends in amyotrophic lateral sclerosis incidence. J Neurol 2016;263:1919-26. |
| 20. | Ferlay J, Bray F, Steliarova-Foucher E, Forman D. Cancer incidence in five continents, CI5plus: IARC CancerBase No. 9 [Internet]. Available from: http://ci5.iarc.fr. [Last accessed on 2019 Dec 01]. |
| 21. | Segi, M. Cancer Mortality for Selected Sites in 24 Countries (1950–57). Sendai, Japan: Department of Public Health, Tohoku University School of Medicine. |
| 22. | Doll R, Payne P, Waterhouse J, editors. Cancer Incidence in Five Continents: A Technical Report. Berlin: Springer-Verlag (for UICC); 1966. |
| 23. | Clegg LX, Hankey BF, Tiwari R, Feuer EJ, Edwards BK. Estimating average annual per cent change in trend analysis. Stat Med 2009;28:3670-82. |
| 24. | Lerman PM. Fitting segmented regression models by grid search. Appl Stat 1980;29:77-84. |
| 25. | Kim HJ, Fay MP, Feuer EJ, Midthune DN. Permutation tests for joinpoint regression with applications to cancer rates. Stat Med 2000;19:335-51. |
| 26. | Rosenberg PS, Check DP, Anderson WF. A web tool for age-period-cohort analysis of cancer incidence and mortality rates. Cancer Epidemiol Biomarkers Prev 2014;23:2296-302. |
| 27. | McNally RJ, Alexander FE, Staines A, Cartwright RA. A comparison of three methods of analysis for age-period-cohort models with application to incidence data on non-Hodgkin’s lymphoma. Int J Epidemiol 2017;26:32-46. |
| 28. | Vaccarella S, Lortet-Tieulent J, Plummer M, Franceschi S, Bray F. Worldwide trends in cervical cancer incidence: Impact of screening against changes in disease risk factors. Eur J Cancer 2013;49:3262-73. |
| 29. | Garland SM, Bhatla N, Ngan HY. Cervical cancer burden and prevention strategies: Asia Oceania perspective. Cancer Epidemiol Biomarkers Prev 2012;21:1414-22. |
| 30. | Australian Institute of Health and Welfare. Cervical Screening in Australia 2004–2005. Cancer series No. 38. Canberra: Australian Institute of Health and Welfare; 2007. |
| 31. | |
| 32. | Garland SM, Brotherton JM, Skinner SR, Pitts M, Saville M, Mola G, et al. Human papillomavirus and cervical cancer in Australasia and Oceania: Risk-factors, epidemiology, and prevention. Vaccine 2008;26:M80-8. |
| 33. | Momenimovahed Z, Salehiniya H. Incidence, mortality, and risk factors of cervical cancer in the world. Biomed Res Ther 2017;4:1795-811. |
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]
|
