Abstract
Objectives
We compared clinical performance of three strategies of primary human papillomavirus (HPV) testing, primary cytology and co-testing for cervical cancer screening.
Design
A population-based prospective cohort study of clinical performance of screening strategy.
Setting
Patients recruited from community in Changzhi County, Shanxi Province, China.
Patient
3209 women aged 30–64 years without gynaecological issues.
Primary and secondary outcome measures
The performance of different screening strategies for detecting cervical intraepithelial neoplasia grade 2 or more severe (CIN2+).
Results
A total of 53 CIN2+ and 31 CIN3+ cases are detected. For CIN2+, sensitivity of primary HPV (95.9%) and co-testing (98.0%) are not statistically different, but significantly higher than primary cytology (48.0%). Specificity (86.8%), colposcopy referral rate (7.8%) and number of colposcopies required to detect one case (9.8) for primary HPV are better than co-testing (79.8%, 11.9%, 14.3%, respectively). For CIN3+, primary HPV, co-testing have 100% of sensitivity and specificity, which is significantly higher than primary cytology (56.7% and 90.2%). Number of colposcopies required to detect one case for primary HPV (15.9) is better than co-testing (23.8).
Conclusions
Compared with co-testing, HPV primary screening had comparable sensitivity and higher specificity for CIN2+ detection, and both of them showed better performance than cytology primary screening in cervical cancer screening.
Keywords: epidemiology, public health, infection control
Strengths and limitations of this study.
From 2018 to 2020, the low loss to follow-up rate (11.9%) increase the stability of our results.
A rigorous diagnostic procedure performed by cytologists and pathologists (a diagnostic panel of three senior doctors) ensure the accuracy of the end point.
The single-centre study limits the extrapolation of our results.
Three strategies based on one population might have cross-contamination, and rigorous randomised controlled trial is preferred.
Although results of cytology and pathology have a strict quality control process, human papillomavirus testing is based on only one company’s kit which may be biased.
Introduction
In May 2018, WHO made a global call for action towards the elimination of cervical cancer, which was to be achieved through human papillomavirus (HPV) vaccination, and effective cervical cancer screening, treatment and management.1 Disparities in economic development and uneven distribution of health resources threaten the progress of cervical cancer elimination in China.2 In China, the HPV vaccine was approved for marketing in 2016, a decade later than the earliest country to market it. A survey showed that the HPV vaccination rate for girls aged 9–14 years in China was <1% in 2020.3 The proposal of the National Committee of the Chinese People’s Political Consultative Conference in 2020 suggested that HPV vaccine should be included in the national immunisation programme (NIP). As of now, cities such as Erdos and Xiamen have started free HPV vaccination for school-aged girls. Then, 112 countries have included HPV vaccine in NIP as of 2021, but China has not yet done so; therefore, the HPV vaccination rate of Chinese school-aged girls is much lower than that of high-income countries with NIP.4
In 2004, The Guidelines for Cervical Cancer Screening and Early Diagnosis and Treatment in China was developed to guide the practice of cervical cancer prevention and treatment strategies for the cervical cancer population. Then in 2005, the ‘National Cervical Cancer Early Diagnosis and Treatment Demonstration Bases’ were established in Shenzhen and Xiang Yuan County in urban and rural areas, respectively. From 2006 to 2008, 43 regions in 31 provinces nationwide were supported to carry out cervical cancer prevention and treatment.5 At the initial stage of the cervical cancer screening programme, the detection methods used were highly subjective and poorly sensitive, so that the problem of misdiagnosis, missed diagnosis or repeated screening was serious, with an average annual screening of 3.8 million people and low population coverage. From 2009 to 2018, the total number of cervical cancer screenings exceeded 85 million.
In addition, in 2015, a cross-sectional survey was conducted in 298 districts/counties across 31 provincial-level administration divisions (PLADs), which were selected randomly from over 2400 districts/counties to be representative of the whole population in the Mainland China as well as the population in every PLAD and specific regions, including urban and rural areas.6 Results showed that provincial-level data further showed that screening rates varied widely across the 31 PLADs. The screening rates in five provinces, including Beijing and Shanghai, exceeded 35%, whereas three provinces, Tibet, Anhui and Hebei, had screening rates of <15%.
To achieve the WHO’s goal that 70% of women be screened, a sustainable cervical cancer screening strategy should be adapted to improve screening coverage in the real world.1 Cytology had been the primary method for screening in the past decades. However, growing evidence have shown that inclusion of HPV testing alone or combined with cytology (co-testing) for screening, compared with cytology alone, is associated with a subsequent reduction in precancerous lesions. Cytology is affordable and is easily adapted to different settings, but its accuracy depends largely on the cytologist.7 8 On the other hand, HPV DNA testing is often preferred for its high accuracy and reproducibility. However, strict environmental and equipment requirements have limited its usage in low-income and middle-income countries (LMICs).9–11 Furthermore, the majority of HPV infection will be cleaned within 6–24 months in women even without any intervention.12 13 Therefore, screening with HPV testing alone could result in overdiagnosis, because HPV-positive individuals without the risk of cervical cancer may seek treatment.14 15 A combination of HPV DNA and cytology (named ‘co-testing’) could greatly improve clinical accuracy, as it combines the high specificity of cytology testing and high sensitivity of HPV testing, but co-testing is always along with higher costs.
According to the clinical performance of the test and the resource availability of the area, different sustainable screening strategies should be tailored in China. Therefore, in this prospective screening cohort, we compare clinical performance of stand-alone and co-testing strategies with cytology and HPV tests in one population.
Methods
Participants recruitment
The Changzhi County, Shanxi Province is an area with a high prevalence of HPV infection16–18 and with low health and economic resource. The 10 townships were divided into 3 clusters based on geographical location, and then 1 township was randomly selected from each cluster. A total of 3386 women without gynaecological issues were involved at baseline. The inclusion criteria were as follows: (1) the patients were 30–64 years old with intact uterus; (2) there was no history of cervical cancer treatment or cervical surgery; (3) after screening, sexual life was prohibited within 48 hours, and no vaginal medication, vaginal contraceptives or vaginal lotion were found within 48 hours; (4) women without suspected clinical pregnancy symptoms; pregnant women could participate in the study at 8 weeks after the end of pregnancy; (5) informed consent was signed, cytology or biopsy samples were collected for testing. Exclusion criteria: those who do not meet any of the above inclusion criteria. Finally, 177 women were excluded and 3209 women were tested by both cytology and HPV DNA at baseline (figure 1). All women with positive baseline result were followed up for 3 years, and in the last year, all participants were also tested by both HPV and cytology.
Figure 1.
Participant flow of enrolment in the screening cohort. ASC-US, atypical squamous cells of undetermined significance; CIN, cervical intraepithelial neoplasia; HPV, human papillomavirus; NILM, negative for intraepithelial lesion or malignancy.
Sample size calculation
According to our previous research (data are not shown), the cumulative incidence of cervical intraepithelial neoplasia grade 2 or more severe (≥CIN2) at 3 years in a population with normal baseline cytology ranged from approximately 1.13% to 2.1%. We hypothesise that the number of cases with ≥CIN2 on pathology at the end of follow-up should be no less than 20. Therefore, the cumulative incidence of ≥CIN2 at 3 years combined with the above-mentioned cumulative incidence of ≥CIN2, it is estimated that the population requiring NILM is 20/1.13%=1770 cases. The total sample size required was estimated based on the proportion of NILM in the total cervical cancer screening population. According to our previous research (data are not shown), the percentage of cervical NILM in Chinese women is about 83% of the total population, and considering a 30% missed follow-up rate, about 2772 women need to be recruited for cervical cancer screening. Considering other factors, to ensure a sufficient population and number of cases required for combined screening in the total screening population, at least 3000 women were planned to be included in this study.
Three screening strategies
All 3209 women were tested by both cytology and HPV DNA at baseline and follow-up for 3 years by cytology. In the last year, all participants were also test by both HPV and cytology. According to the Chinese Guidelines for Cervical Cancer Screening,19 three sets of data were extracted from one population in our study to evaluate the clinical performance of three screening strategy.
In strategy 1 (primary HPV with triage by 16/18 genotyping and cytology), participants were primarily screened with HPV testing. Those tested positive for HPV 16/18 were referred to colposcopy, and the 12 other high-risk (HR)-HPV(+) were triaged with cytology. Among them, those diagnosed with ≥ASC-US were referred to colposcopy.
In strategy 2 (primary cytology with triage by HPV 16/18 genotyping), participants were primarily screened with cytology alone. Among them, those diagnosed with>ASC-US were referred to colposcopy. ASC-US were triaged with reflex HPV testing, and HPV 16/18+ were referred to colposcopy.
In strategy 3 (co-testing with genotyping and cytology triage: HPV 16/18 and ASC-US HPV+ threshold) women were screened by both cytology and HPV testing, and those with either HPV 16/18 +, or >ASC-US, or ASC-US and HPV+ were referred to colposcopy. More details are shown in figure 2. Notably, three strategies were conducted in one same population-based cohort, but data were analysed by different strategies.
Figure 2.
Screening strategies in this population-based cohort. Strategy 1: primary HPV with triage by 16/18 genotyping and cytology: ASC-US threshold. Strategy 2: cytology with triage by HPV 16/18 genotyping. Strategy 3: co-testing with genotyping and cytology triage: HPV 16/18 and ASC-US HPV+ threshold. ASC-US, atypical squamous cells of undetermined significance; HPV, human papillomavirus; HR, high-risk; NILM, negative for intraepithelial lesion or malignancy.
Sample collection
All patients first underwent gynaecological examinations. Cervical exfoliated cells were collected using a cytology brush (Hologic, Bedford, Massachusetts, USA) and stored in the tubes with preservation solution for the Thinprep cytology test (Hologic). A 2.5 mL PreservCyt Liquid was separated from each cytological specimen and placed in a special specimen tube for HPV DNA test (Tegen, Shanghai). The remaining PreservCyt Liquid was subjected to cytological examination.
Laboratory tests
All of the following tests and diagnostic procedures were strictly double-blinded. Cytology slides were read by two pathologists results and reported according to the Bethesda 2014 classification. The cytological results were as follows: negative for intraepithelial lesion or malignancy (NILM), atypical cells of undetermined significance (ASC-US), low-grade squamous intraepithelial lesion (LSIL), atypical squamous cells-cannot exclude high-grade squamous intraepithelial lesion, high-grade squamous intraepithelial lesion (HSIL), atypical glandular cells and cervical cancer cells. Diagnoses were reported if the diagnoses by two cytologists were consistent. Otherwise, a third cytologist was consulted.
A commercial assay was used for HPV DNA testing. And HPV testing method was Biochip Method, the manufacturer was Beijing Bohui Innovative Optoelectronic Technology and the China Food and Drug Administration (CFDA) approval number was obtained (registration certificate n0: 20163401108). The results were categorised as following: HPV−, HR-HPV (result positive for 1 or more of 14 HR types (HPV 16, HPV 18, HPV 31, HPV 33, HPV 35, HPV 39, HPV 45, HPV 51, HPV 52, HPV 56, HPV 58, HPV 59, HPV 66 and HPV 68)), HPV 16/18+ (result positive for either genotype 16 or 18) and HPV non-16/18+ (result negative for genotype 16/18 and positive for 1 or more of 12 other HR types).
Final screening assessment
All women with positive HPV 18 and/or HPV 16 testing or abnormal cytology (ASC-US or worse) were referred to colposcopy. If the colposcopy is fully exposed and the lesion site exists, biopsy should be carried out in the abnormal part, and the specific location of the specimen should be clearly marked; if the colposcopy exposure is insufficient, cervical curettage should be performed. Two pathologists independently made diagnosis according to the 2014 WHO Classification of Tumours of the Female Genital Tract. If the diagnoses were concordant, they were reported as the pathological diagnosis. Otherwise, two pathologists will read all cytology slides, and for all positive results and 10% of negative results, a third pathologist will read the slides and then take the results that are in agreement between the three doctors, and for the results that are not in agreement a decision will be made by all three. All technicians, cytologist and pathologist involved in HPV testing and cytology slides reading were double-blind in the whole process of our study. The histological diagnoses of cervical lesions were divided into normal, LSIL/CIN1 (including the condylomatous variant), HSIL/CIN2, HSIL/CIN3 (including adenocarcinoma in situ) and carcinoma (squamous cell carcinoma or adenocarcinoma). HPV testing, cytology and pathological examination were performed with blinding to results of each test. Women with both negative HPV and cytology results were not referred to colposcopy and regarded as normal.
Follow-up and end point
The end point is histologically confirmed ≥CIN2. At baseline, those with positive HPV 18 and/or HPV 16 testing or abnormal cytology (ASC-US or worse) were immediately referred to colposcopy, and pathological results below CIN2 would be followed up with cytology every year. Those with cytology (NILM) and non-HPV 16/18 (+) would be followed up with cytology every year; those with normal cytology (NILM) and HPV(−) would be followed up with cytology every 3 years. More details are shown in figure 1. At the end of follow-up, all women were screened with cytology and HPV DNA testing.
Statistical analysis
In this study, pathological diagnosis results were viewed as ‘Golden Criteria’ to compare three screening strategies. In addition, for those with negative screening results at baseline and third year, they would not be clinically recommended to conduct colposcopy due to the low prevalence rate of cervical lesions,20 therefore they were viewed as normal pathological results in final statistical analysis. The clinical performance of the three strategies includes accuracy and cost of screening technology, the accuracy of detecting high-grade CIN is first calculated using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), colposcopy referral rate and number needed to refer (NNR) to colposcopy to find one case of CIN2+ or CIN3+. And 95% CIs for proportions were calculated according to the efficient-score method (corrected for continuity) described by Robert Newcombe, based on the procedure outlined by Wilson.21 22 Second, the cost of each strategy was also calculated in the real world, and the standard price for each screening strategy from the National Rural-Area Two-Cancer Screening Project was used: 35 RMB for a cytology examination, 112 RMB for a HPV DNA test and 220 RMB for a histopathological biopsy. Three screening strategies were from the same one cohort, therefore participants would referred to colposcopy if any condition in any screening strategy was satisfied. In cases where a participant was diagnosed with CIN2+ by one screening strategy at baseline, while the other strategy determined that follow-up was needed, and the participant would be viewed as end point, therefore it is unclear whether other strategy could detect CIN2+ in the following years. Sensitivity analysis was conducted to improve the stability of the results. SAS statistical software (V.9.4) was used for the statistical analyses.
Patient and public involvement
Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research. However, patient groups were involved in approving the patient information leaflet for sample collection.
Results
In 2017, a total of 3209 women were screened and 687 women with positive baseline results were needed to follow-up for 3 years. From 2018 to 2020, 18.9% (130/687) of participants were lost to follow-up. In 2020, all participants were recalled and compliance rate was 88.1% (2932/3328). In this screening cohort, a total of 157 cases of CIN1, 53 cases of CIN2 or worse (CIN2+) and 31 cases of CIN3 or worse (CIN3+) were detected, and population characteristics are shown in online supplemental table 1. In our cohort, 77, 68 and 130 participants were lost to follow-up in strategy 1, 2 and 3, respectively. Strategy 1 has detected 47 cases of CIN2+ and 29 cases of CIN3+, 24 cases of CIN2+ and 17 cases of CIN3+ in strategy 2, 49 cases of CIN2+ and 29 cases of CIN3+ in strategy 3. Four cases were indicated as needing follow-up but diagnosed by other strategies, four, two and three cases in strategy 1, 2 and 3. More details are shown in online supplemental table 2.
bmjopen-2022-063622supp001.pdf (71.3KB, pdf)
Distribution of HPV and cytological diagnosis among general population at baseline
Among all screened participants, 13.9% (446/3209) of women had abnormal cytology, 14.4% (463/3209) were HR-HPV positive and 4.7% (152/3209) were HPV 16/18 positive. The abnormal rate of cytology was highest in people aged 60–64 years (38.9%), followed by 50–59 years (21.8%), 40–49 years (9.7%) and 30–39 years (6.2%). The prevalence of HR-HPV peaked at age 60–64 years (17.9%), followed by 50–59 years (16.7%), 30–39 years (14.5%) and 40–49 years (12.6%). The positive rate of HPV16/18 was 6.2% in the 60–64 years group, 5.0% in 30–39 years group, 4.9% in 50–59 years group and 4.4% in 40–49 years group (table 1).
Table 1.
Distribution of HPV and cytological diagnosis at different age groups at baseline
| Age group (years) | Cytological diagnosis | HPV infection | HR-HPV infection | Total (%) | |||
| NILM (%) | ≥ASC-US (%) | Non-HR-HPV (%) | HR-HPV (%) | 16/18 (%) | 12OT (%) | ||
| 30–39 | 636 (93.8) | 42 (6.2) | 580 (85.5) | 98 (14.5) | 34 (5.0) | 64 (9.4) | 678 (100.0) |
| 40–49 | 1308 (90.3) | 140 (9.7) | 1266 (87.4) | 182 (12.6) | 63 (4.4) | 119 (8.2) | 1448 (100.0) |
| 50–59 | 720 (78.2) | 201 (21.8) | 767 (83.3) | 154 (16.7) | 45 (4.9) | 109 (11.8) | 921 (100.0) |
| 60–64 | 99 (61.1) | 63 (38.9) | 133 (82.1) | 29 (17.9) | 10 (6.2) | 19 (11.7) | 162 (100.0) |
| Total | 2763 (86.1) | 446 (13.9) | 2746 (85.6) | 463 (14.4) | 152 (4.7) | 311 (9.7) | 3209 (100.0) |
12OT: HPV 16, HPV 18, HPV 31, HPV 33, HPV 35, HPV 39, HPV 45, HPV 51, HPV 52, HPV 56, HPV 58, HPV 59, HPV 66 and HPV 68.
ASC-US, atypical squamous cells of undetermined significance; HPV, human papillomavirus; HR, high-risk; NILM, negative for intraepithelial lesion or malignancy; OT, other HR-HPV genetype.
Distribution of HR-HPV and cytological diagnosis among different grades of cervical lesions at baseline
A total of 17.4% (558/3209) of the participants were referred to colposcopy. Among them, 5.2% (166/3209) were pathologically abnormal, including 128 cases with CIN1, and 38 with CIN2+. HR-HPV infection prevalence was 38.0% (212/558): 59.4% (126/212) with normal pathology and 40.6% (86/212) with abnormal pathology. Among those with CIN2+, only 5.3% (2/38) was not infected with HR-HPV and the most prevalent subtype was HPV 16 (72.2%), followed by HPV 18 (16.7%), HPV 58 (16.7%), HPV 52 (13.9%) and HPV 33 (11.1%). The total abnormal cytology rate was 75.4% (421/558): 70.1% (295/421) with normal pathology and 29.9% (126/421) with abnormal pathology. Among those with CIN2+, 55.3% (21/38) were diagnosed with ASC-US+ and 44.7% (17/38) with NILM. There was no correlation between the severity of cytological diagnosis and the grade of cervical lesions (table 2).
Table 2.
Results of HR-HPV and cytological diagnosis among different grades of cervical lesions at baseline
| Pathological diagnosis | Total (%) | |||
| Normal (%) | CIN1 (%) | CIN2+ (%) | ||
| HPV infection* | ||||
| Non-HR-HPV | 266 (76.9) | 78 (22.5) | 2 (0.6) | 346 (100.0) |
| HR-HPV | 126 (59.4) | 50 (23.6) | 36 (17.0) | 212 (100.0) |
| HPV 16/18 | 78 (57.3) | 28 (20.6) | 30 (22.1) | 136 (100.0) |
| HPV 16 | 52 (53.1) | 20 (20.4) | 26 (26.5) | 98 (100.0) |
| HPV 18 | 26 (65.0) | 8 (20.0) | 6 (15.0) | 40 (100.0) |
| HPV 52 | 18 (50.0) | 13 (36.1) | 5 (13.9) | 36 (100.0) |
| HPV 58 | 19 (63.3) | 5 (16.7) | 6 (20.0) | 30 (100.0) |
| HPV 33 | 7 (53.8) | 2 (15.4) | 4 (30.8) | 13 (100.0) |
| Cytological diagnosis | ||||
| NILM | 97 (70.8) | 23 (16.8) | 17 (12.4) | 137 (100.0) |
| ASC-US | 126 (90.0) | 10 (7.1) | 4 (2.9) | 140 (100.0) |
| AGC | 11 (73.3) | 4 (26.7) | 0 (0.0) | 15 (100.0) |
| LSIL | 102 (68.0) | 45 (30.0) | 3 (2.0) | 150 (100.0) |
| ASC-H | 34 (44.7) | 34 (44.7) | 8 (10.5) | 76 (100.0) |
| HSIL | 22 (55.0) | 12 (30.0) | 6 (15.0) | 40 (100.0) |
| Total | 392 (70.3) | 128 (22.9) | 38 (6.8) | 558 (100.0) |
*Multiple infection was counted by each type.
AGC, atypical glandular cells; ASC-H, high-grade squamous intraepithelial lesion; ASC-US, atypical squamous cells of undetermined significance; CIN, cervical intraepithelial neoplasia; HPV, human papillomavirus; HR, high-risk; HSIL, high-grade squamous intraepithelial lesion; LSIL, low-grade squamous intraepithelial lesion; NILM, negative for intraepithelial lesion or malignancy.
Comparison among three screening strategies to identify high-grade CIN with 3 years follow-up
When identifying CIN2+ as end point, strategy 1 had sensitivity rate of 95.9% and specificity rate of 86.8%, PPV of 10.2%, NPV of 99.9%, its colposcopy referral rate was 7.8% and the needed number of performed colposcopies to detect one case was 9.8. Strategy 2 had the highest specificity (90.4%) and colposcopy referral rate was 9.4%, but it had the lowest sensitivity (48.0%) and NPV (99.1%). Strategy 3 showed the highest sensitivity (98.0%), NPV (99.9%) and colposcopy referral rate (11.9%), but it had the lowest specificity (79.8%) and PPV (7.0%). The needed number of performed colposcopies to detect one case was 14.3 (table 3).
Table 3.
Clinical performance and cost analysis of different screening strategies for CIN2+ and CIN3+ diagnoses after 3 years follow-up
| Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | Colposcopy referral rate (%) | NNR* | |
| ≥CIN2 | ||||||
| Strategy 1 | 95.9 (90.1 to 100.0) |
86.8 (85.7 to 88.0) |
10.2 (9.2 to 11.2) |
99.9 (99.7 to 100.0) |
7.8 (6.8 to 8.7) |
9.8 (8.9 to 10.9) |
| Strategy 2 | 48.0 (34.2 to 61.8) |
90.4 (89.3 to 91.4) |
7.3 (5.5 to 9.7) |
99.1 (98.8 to 99.3) |
9.4 (8.4 to 10.4) |
13.7 (10.3 to 18.2) |
| Strategy 3 | 98.0 (94.0 to 100.0) |
79.8 (78.4 to 81.2) |
7.0 (6.5 to 7.5) |
99.9 (99.7 to 100.0) |
11.9 (10.8 to 13.0) |
14.3 (13.3 to 15.4) |
| ≥CIN3 | ||||||
| Strategy 1 | 100.0 (100.0 to 100.0) |
86.4 (85.2 to 87.5) |
6.3 (5.8 to 6.8) |
100.0 (100.0 to 100.0) |
7.8 (6.8 to 8.7) |
15.9 (14.7 to 17.2) |
| Strategy 2 | 56.7 (38.9 to 74.4) |
90.2 (89.2 to 91.2) |
5.2 (3.8 to 7.1) |
99.5 (99.3 to 99.7) |
9.4 (8.4 to 10.4) |
19.2 (14.1 to 26.3) |
| Strategy 3 | 100.0 (100.0 to 100.0) |
79.3 (77.9 to 80.7) |
4.2 (4.0 to 4.5) |
100.0 (100.0 to 100.0) |
11.9 (10.8 to 13.0) |
23.8 (22.2 to 25.0) |
CIN, cervical intraepithelial neoplasia; NNR, the number of performed colposcopies to detect one case; NPV, negative predictive value; PPV, positive predictive value.
When CIN3+ as end point, both strategy 1 and strategy 3 had 100% of sensitivity rate and NPV, but strategy 1 had a higher specificity (86.4% vs 79.3%) and a higher PPV (6.3% vs 4.2%) than strategy 3. In addition, strategy 1 had a lower needed number of performed colposcopies to detect one case (15.9) than strategy 3 (23.8). Strategy 2 had the lowest sensitivity (56.7%) and NPV (99.5%), while it had the highest specificity rate (90.2%) and the needed number of performed colposcopies to detect one case was 19.2 (table 3). Moreover, sensitivity analysis was conducted to test the stability of the results. The adjusted results are shown in table 4, which show similar results to those in table 3.
Table 4.
Clinical performance and cost analysis of different screening strategies for CIN2+ and CIN3+ diagnoses after 3 years follow-up after adjustment
| Adjusted 1 | Adjusted 2 | |||||||
| Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | |
| ≥CIN2 | ||||||||
| Strategy 1 | 88.7 (80.1 to 97.2) |
86.8 (85.6 to 88.0) |
10.2 (9.0 to 11.4) |
99.8 (99.5 to 99.9) |
96.2 (91.1 to 100) |
86.9 (85.8 to 88.1) |
11.0 (10.0 to 12.1) |
99.9 (99.7 to 99.9) |
| Strategy 2 | 46.2 (32.6 to 59.7) |
90.1 (89.3 to 91.4) |
7.3 (5.5 to 9.7) |
99.0 (98.8 to 99.2) |
50.0 (36.4 to 63.4) |
90.4 (89.4 to 91.5) |
7.9 (6.0 to 10.3) |
99.1 (98.8 to 99.3) |
| Strategy 3 | 92.3 (85.1 to 99.6) |
79.8 (78.4 to 81.2) |
7.0 (6.3 to 7.7) |
99.8 (99.6 to 99.9) |
98.1 (94.3 to 100.0) |
79.9 (78.5 to 81.3) |
7.4 (6.9 to 8.0) |
99.9 (99.7 to 99.9) |
| ≥CIN3 | ||||||||
| Strategy 1 | 93.5 (84.9 to 100.0) |
86.3 (85.1 to 87.5) |
6.3 (5.6 to 7.1) |
99.9 (99.7 to 100.0) |
100.0 (100.0 to 100.0) |
86.4 (85.2 to 87.6) |
6.7 (6.2 to 7.3) |
100.0 (100.0 to 100.0) |
| Strategy 2 | 54.8 (37.3 to 72.4) |
90.2 (89.2 to 91.2) |
5.2 (3.8 to 7.1) |
99.5 (99.3 to 99.7) |
58.1 (40.7 to 75.4) |
90.2 (89.2 to 91.3) |
5.5 (4.1 to 7.4) |
99.5 (99.3 to 99.7) |
| Strategy 3 | 93.5 (84.9 to 100.0) |
79.3 (77.9 to 80.7) |
4.2 (3.8 to 4.7) |
99.9 (99.7 to 100.0) |
100.0 (100.0 to 100.0) |
79.4 (78.0 to 80.8) |
4.5 (4.2 to 4.8) |
100.0 (100.0 to 100.0) |
Adjusted 1: all cases found by other strategies were considered as missed cases. Adjusted 2: all cases found by other strategies were considered as detected cases.
CIN, cervical intraepithelial neoplasia; NPV, negative predictive value; PPV, positive predictive value.
Discussion
According to the guidelines from the USA and Europe, individuals were recommended to start cervical cancer screening at 25 years old and undergo primary HPV testing every 5 years through 65 years old (preferred). However, the guideline in China recommended HPV or cytology alone or both strategies for women aged 30–64 years, but there is no preferred strategy.6 23 24 This real-world study combines the baseline and the 3-year follow-up data, and provides evidence on the clinical performance of the three strategies with consideration of the cost of screening examination technology. Our results showed that HPV primary screening for women aged 30–64 years appears to be the optimal strategy.
Overall, HR-HPV prevalence in the study population at baseline was 14.4%, which was lower than 18.0%, as concluded by pooled analysis by Zhao et al from 17 population-based studies in rural China in 2010.25 In our study, the prevalence of HPV slightly declined in middle ages (40–49 years), yet increased among the oldest people (60–64 years), which are consistent with other studies in China26 27 or in some Latin America/Caribbean populations.28 A meta-analysis of 1 million women from five continents showed that age-specific HPV distribution presented with a first peak at younger ages (<25 years) and, in the Americas and Africa, a rebound at older ages (≥45 years).29 Such a bimodal pattern could be due to changes in the sexual behaviour or the reactivation of latent viral infections.30 However, studies have shown that the relationship between HPV prevalence and age factors varied by regions. It seems that in Southern Europe, Western Africa, South America, the HPV prevalence shows a decreased trend before about 50 years, then it has an increased trend among 50 years to 65 years.4 6 Compared with some studies in China,26 27 31–33 the ‘V’ trend of HPV prevalence among age groups is more common among the Chinese population. Moreover, the prevalence of ≥ASC-US is also lower at 13.9% compared with 17.0%, as concluded from a pooled analysis of 13 population-based studies on rural women throughout China.34 Notably, the observed rate of ASC-US+ in China is very high, which was higher than other countries, including 6.7% for the USA, 5.47% for Italy, 3.57% for Switzerland, 9.60% for France. In additoin, the lower rate of HR-HPV prevalence and cytological abnormality rate could potentially be explained by the cervical cancer screening and promotion campaigns in this area throughout the past 20 years.35 36 In addition, we found that among women with CIN2+, the most frequent HPV types were 16, 58, 18 and 52, which was consistent with the cross-sectional population-based study in rural Northern China.37
For women aged 30–64 years, primary HPV with triage by 16/18 genotyping and cytology (strategy 1) is cost-effective, compared with co-testing, as shown in other studies as well.38–40 To our knowledge, there is no evidence that co-testing can significantly improve the accuracy of screening than primary HPV. The main difference between primary HPV and co-testing is that in co-testing, women who are HPV(−) are triaged with cytology, whereas the triage does not occur in primary HPV.41–43 When cytologists are inadequately trained, co-testing, compared with primary HPV, has similar accuracy, and worse colposcopy rate and NNR. For CIN2+, its specificity, PPV and NPV are much higher, its colposcopy referral rate and NNR are lower, although its sensitivity is lower (not statistically significant). Because participants of screening are generally healthy people, it is of great public health significance to reduce unnecessary examination and psychological anxiety related to overdiagnosis, which is reflected by colposcopy referral rate and NNR.44 In addition, it is possible for individual with CIN2 to return to normal,12 39 40 thus we also consider CIN3+ to be the disease end point. For CIN3+, we found all accuracy indicators of primary HPV to be equal to or better than co-testing.
Although the total and annual costs of cytology for all participants are only one-third of the costs of co-testing, the accuracy of cytology combined with triage by HPV testing (strategy 2) was far less than the accuracy of co-testing and primary HPV, which is already well-established in other studies.45 46 Furthermore, our accuracy of cytology is lower than other studies, which shows that the accuracy of cytology is largely dependent on the level of expertise of the cytologist.47 48 Although there are limitations in accuracy and quality assurance, the widespread application of cytology for screening in populations has successfully reduced the burden of cervical cancer, especially squamous cell cervical cancer, since the middle of the last century.37 The implementation of HPV testing in all healthcare setting worldwide will take a long time, if possible at all, due to the strict laboratory conditions and large investment required. Therefore, in order to achieve the goal of worldwide elimination of cervical cancer in a timely manner, primary cytology followed by triage by HPV testing is recommended for areas with limited resource where cervical cancer screening otherwise would be infeasible.49 However, level of cytologist is the most important factor affecting the accuracy and cost-effectiveness in the general population, therefore special attention should be paid to train cytologists.
Several studies have shown that co-testing increases the screening cost compared with primary HPV.25–27 And, co-testing is most suitable for well-developed areas with high levels of cytologists, but it is difficult to carry out in most areas of China and other LMICs.15 34 Primary HPV requires HPV(−) individuals to be recalled every 5 years, while co-testing requires those who are HPV(−) with cytology ASC-US to be recalled every 3 years. Currently, it is unclear whether this shorter screening interval is justified, especially considering its increased costs and resource uses. There is no prospective cohort study on this topic, and our study is expected to provide data needed with follow-up in the future.
Our results supported that primary HPV testing was the preferred screening strategy among women aged 30–65 years, which was consistent with the evidence from the American Society for Colposcopy and Cervical Pathology.6 A decision analysis showed that cytology alone resulted in the lowest benefit about life-years gained and number of CIN2 or CIN3 cases detected, and it has lower sensitivity for precancer than primary HPV testing or co-testing.38 On the other hand, a retrospective study concluded that the combination of cytology and HPV testing (co-testing) offered very little incremental benefit in detection but increased the number of procedures and the risk for harms.47
This screening cohort has low loss to follow-up rate (11.9%), which increases the stability of our results. In addition, a rigorous diagnostic procedure performed by cytologists and pathologists (a diagnostic panel of three senior doctors) ensure the accuracy of the end point. However, our study also have limitations; first, the single-centre study limits the extrapolation of our results. Second, three strategies based on one population might have cross-contamination, and rigorous randomised controlled trial is preferred. Finally, although there is a strict quality control process for cytology and pathology results, HPV testing is based on only one technique of kit and may be biased.
In conclusion, guidelines from Europe or the USA recommend cervical cancer screening strategies in the general population. However, there is still a lack of real-world data on cervical cancer screening in China, and our study provides real-world data on cervical cancer screening. Our results show that primary HPV screening strategy should be adapted whenever resources permit. In resource-poor areas, primary cytology is an acceptable alternative. However, cytologists should be trained adequately to ensure acceptable cytology accuracy. Co-testing is costly, requires well-trained cytologists and it does not show improvement in clinical performance over primary HPV.
Supplementary Material
Acknowledgments
We acknowledge the significant contributions of all the investigators in CICAMS and local health providers for their efforts in conducting the study and women who have participated in this study.
Footnotes
Contributors: YQ was responsible for conception and design;ZL drafted the article; ZL, XJ analyzed and interpreted the data;ZL, XF, SZ, YH and XS performed investigations and HPV testing; XZ, QP carried out cytological and histological examinations. XZ, XJ, YQ participated in revising the manuscript. All authors read and approved the final manuscript.
Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests: None declared.
Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review: Not commissioned; externally peer reviewed.
Supplemental material: This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
Data availability statement
All data relevant to the study are included in the article.
Ethics statements
Patient consent for publication
Consent obtained directly from patient(s).
Ethics approval
This study was approved by the Institutional Review Board of Affiliated Cancer Hospital of Zhengzhou University (No. 2017007). Each participant was informed and signed an informed consent.
References
- 1.WHO . WHO Director-General calls for all countries to take action to help end the suffering caused by cervical cancer; 2018.
- 2.Zhao F, Qiao Y. Cervical cancer prevention in China: a key to cancer control. Lancet 2019;393:969–70. 10.1016/S0140-6736(18)32849-6 [DOI] [PubMed] [Google Scholar]
- 3.Jie-Chen L, Lin-Lin W, Qing-Rui B. Surveillance of human papillomavirus vaccination rates and suspected abnormal vaccination reactions in Shanghai from 2017-2019. China Vaccine and Immunity 2020;26:322–5. doi:1006-916X（2020）03-0322-05 [Google Scholar]
- 4.Chen H, Xia CF, T T Y. Causes of the rapidly rising burden of cervical cancer in Chinese women and its response. Chinese Journal of Epidemiology 2022;43:761–5. 10.3760/cma.j.cn112338-20211004-0076 [DOI] [PubMed] [Google Scholar]
- 5.Shanying H, Xuelian Z, Yong Z. Prevention of cervical cancer: interpretation of who guidelines for screening and treatment of cervical precancerous lesions (second edition). Chin J Med 2021;101:2653–7. 10.3760/cma.j.cn112137-20210719-01609 [DOI] [Google Scholar]
- 6.Wang W, Arcà E, Sinha A, et al. Cervical cancer screening guidelines and screening practices in 11 countries: a systematic literature review. Prev Med Rep 2022;28:101813. 10.1016/j.pmedr.2022.101813 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Underwood M, Arbyn M, Parry-Smith W, et al. Accuracy of colposcopy-directed punch biopsies: a systematic review and meta-analysis. BJOG 2012;119:1293–301. 10.1111/j.1471-0528.2012.03444.x [DOI] [PubMed] [Google Scholar]
- 8.Wentzensen N, Walker JL, Gold MA, et al. Multiple biopsies and detection of cervical cancer precursors at colposcopy. J Clin Oncol 2015;33:83–9. 10.1200/JCO.2014.55.9948 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Ronco G, Dillner J, Elfström KM, et al. Efficacy of HPV-based screening for prevention of invasive cervical cancer: follow-up of four European randomised controlled trials. Lancet 2014;383:524–32. 10.1016/S0140-6736(13)62218-7 [DOI] [PubMed] [Google Scholar]
- 10.Ronco G, Giorgi-Rossi P, Carozzi F, et al. Efficacy of human papillomavirus testing for the detection of invasive cervical cancers and cervical intraepithelial neoplasia: a randomised controlled trial. Lancet Oncol 2010;11:249–57. 10.1016/S1470-2045(09)70360-2 [DOI] [PubMed] [Google Scholar]
- 11.Frank AP. Human papillomavirus DNA versus papanicolaou screening tests for cervical cancer. N Engl J Med 2008;358:642. [PubMed] [Google Scholar]
- 12.Schiffman M, Doorbar J, Wentzensen N, et al. Carcinogenic human papillomavirus infection. Nat Rev Dis Primers 2016;2:16086. 10.1038/nrdp.2016.86 [DOI] [PubMed] [Google Scholar]
- 13.Cho HW, So KA, Lee JK, et al. Type-specific persistence or regression of human papillomavirus genotypes in women with cervical intraepithelial neoplasia 1: a prospective cohort study. Obstet Gynecol Sci 2015;58:40–5. 10.5468/ogs.2015.58.1.40 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cuzick J, Mayrand MH, Ronco G, et al. Chapter 10: new dimensions in cervical cancer screening. Vaccine 2006;24 Suppl 3:S90–7. 10.1016/j.vaccine.2006.05.122 [DOI] [PubMed] [Google Scholar]
- 15.Castle PE, Katki HA. Benefits and risks of HPV testing in cervical cancer screening. Lancet Oncol 2010;11:214–5. 10.1016/S1470-2045(09)70385-7 [DOI] [PubMed] [Google Scholar]
- 16.Zhang Y-zhen, QIAO N, Ling CAO. Cancer incidence and mortality in registration areas of Shanxi province, 2013. China Cancer 2017;26:925–31. 10.11735/j.issn.1004-0242.2017.12.A001 [DOI] [Google Scholar]
- 17.Ma Z, Gao Q, Cao L. Analysis of cancer incidence and mortality in registration areas of Shanxi province in 2014. Cancer Research and Clinic 2020;32:186–91. 10.3760/cma.j.cn115355-20191030-00492 [DOI] [Google Scholar]
- 18.Niu L, Jiang N. Analysis of the situation of cervical cancer screening from 2018 to 2020 in Lucheng district of Changzhi City, Shanxi Province. Cancer Research and Clinic 2021;33:938–41. 10.3760/cma.j.cn115355-20210429-00200 [DOI] [Google Scholar]
- 19.Women's Health Branch, Chinese Preventive Medicine Association . Expert panel interpretation:Comprehensive prevention and control guidelines for cervical cancer. BEIJING: People’s Medical Publishing House, 2017. [Google Scholar]
- 20.Castle PE, Kinney WK, Xue X, et al. Effect of several negative rounds of human papillomavirus and cytology Co-testing on safety against cervical cancer: an observational cohort study. Ann Intern Med 2018;168:20–9. 10.7326/M17-1609 [DOI] [PubMed] [Google Scholar]
- 21.Newcombe RG. Two-sided confidence intervals for the single proportion: comparison of seven methods. Stat Med 1998;17:857–72. 10.1002/(SICI)1097-0258(19980430)17:8<857::AID-SIM777>3.0.CO;2-E [DOI] [PubMed] [Google Scholar]
- 22.Wilson EB. Probable inference, the law of succession, and statistical inference. J Am Stat Assoc 1927;22:209–12. 10.1080/01621459.1927.10502953 [DOI] [Google Scholar]
- 23.Fontham ETH, Wolf AMD, Church TR, et al. Cervical cancer screening for individuals at average risk: 2020 guideline update from the American cancer Society. CA Cancer J Clin 2020;70:321–46. 10.3322/caac.21628 [DOI] [PubMed] [Google Scholar]
- 24.Women CPMAf . Comprehensive prevention and control guidelines for cervical cancer. People's Medical Publishing House, 2017. [Google Scholar]
- 25.Zhao F-H, Lewkowitz AK, Hu S-Y, et al. Prevalence of human papillomavirus and cervical intraepithelial neoplasia in China: a pooled analysis of 17 population-based studies. Int J Cancer 2012;131:2929–38. 10.1002/ijc.27571 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Luo L-P, He P, Liu Q-T, et al. Prevalence and genotype distribution of HPV infection among 214,715 women from southern China, 2012-2018: baseline measures prior to mass HPV vaccination. BMC Infect Dis 2021;21:328. 10.1186/s12879-021-06019-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Zhu X, Wang Y, Lv Z, et al. Prevalence and genotype distribution of high-risk HPV infection among women in Beijing, China. J Med Virol 2021;93:5103–9. 10.1002/jmv.27013 [DOI] [PubMed] [Google Scholar]
- 28.Franceschi S, Herrero R, Clifford GM, et al. Variations in the age-specific curves of human papillomavirus prevalence in women worldwide. Int J Cancer 2006;119:2677–84. 10.1002/ijc.22241 [DOI] [PubMed] [Google Scholar]
- 29.Bruni L, Diaz M, Castellsagué X, et al. Cervical human papillomavirus prevalence in 5 continents: meta-analysis of 1 million women with normal cytological findings. J Infect Dis 2010;202:1789–99. 10.1086/657321 [DOI] [PubMed] [Google Scholar]
- 30.Gravitt PE, Rositch AF, Silver MI, et al. A cohort effect of the sexual revolution may be masking an increase in human papillomavirus detection at menopause in the United States. J Infect Dis 2013;207:272–80. 10.1093/infdis/jis660 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Zhu Y, Qian F, Zou W, et al. Prevalence and genotype distribution of human papillomavirus infection in Huzhou city, eastern China, 2018-2019. Trans R Soc Trop Med Hyg 2021;115:30–7. 10.1093/trstmh/traa077 [DOI] [PubMed] [Google Scholar]
- 32.Su Y, Yuan Z, Xu C, et al. Prevalence and genotype distribution of human papillomavirus infection among women: a population-based study in Dali BAI autonomous prefecture, Yunnan province, China. J Med Virol 2019;91:1553–61. 10.1002/jmv.25477 [DOI] [PubMed] [Google Scholar]
- 33.Luo Q, Lang L, Han N, et al. Prevalence and genotype distribution of high-risk human papillomavirus infection among women with cervical cytological abnormalities in Chongqing, China, 2014-2020. Diagn Cytopathol 2021;49:1237–43. 10.1002/dc.24891 [DOI] [PubMed] [Google Scholar]
- 34.Pan Q-J, Hu S-Y, Guo H-Q, et al. Liquid-based cytology and human papillomavirus testing: a pooled analysis using the data from 13 population-based cervical cancer screening studies from China. Gynecol Oncol 2014;133:172–9. 10.1016/j.ygyno.2014.03.008 [DOI] [PubMed] [Google Scholar]
- 35.Qiao Y-L, Sellors JW, Eder PS, et al. A new HPV-DNA test for cervical-cancer screening in developing regions: a cross-sectional study of clinical accuracy in rural China. Lancet Oncol 2008;9:929–36. 10.1016/S1470-2045(08)70210-9 [DOI] [PubMed] [Google Scholar]
- 36.Zhao F-H, Lin MJ, Chen F, et al. Performance of high-risk human papillomavirus DNA testing as a primary screen for cervical cancer: a pooled analysis of individual patient data from 17 population-based studies from China. Lancet Oncol 2010;11:1160–71. 10.1016/S1470-2045(10)70256-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Zhao S, Zhao X, Hu S, et al. Distribution of high-risk human papillomavirus genotype prevalence and attribution to cervical precancerous lesions in rural North China. Chin J Cancer Res 2019;31:663–72. 10.21147/j.issn.1000-9604.2019.04.10 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Kim JJ, Burger EA, Regan C, et al. Screening for cervical cancer in primary care: a decision analysis for the US preventive services task force. JAMA 2018;320:706–14. 10.1001/jama.2017.19872 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Matsumoto K, Oki A, Furuta R, et al. Predicting the progression of cervical precursor lesions by human papillomavirus genotyping: a prospective cohort study. Int J Cancer 2011;128:2898–910. 10.1002/ijc.25630 [DOI] [PubMed] [Google Scholar]
- 40.McCredie MRE, Sharples KJ, Paul C, et al. Natural history of cervical neoplasia and risk of invasive cancer in women with cervical intraepithelial neoplasia 3: a retrospective cohort study. Lancet Oncol 2008;9:425–34. 10.1016/S1470-2045(08)70103-7 [DOI] [PubMed] [Google Scholar]
- 41.Schiffman M, Kinney WK, Cheung LC, et al. Relative performance of HPV and cytology components of cotesting in cervical screening. J Natl Cancer Inst 2018;110:501–8. 10.1093/jnci/djx225 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Malinowski DP, Broache M, Vaughan L. Cotesting in cervical cancer screening. Am J Clin Pathol 2020. 10.1093/ajcp/aqaa169 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Schiffman M, Solomon D. Clinical practice. Cervical-cancer screening with human papillomavirus and cytologic cotesting. N Engl J Med 2013;369:2324–31. 10.1056/NEJMcp1210379 [DOI] [PubMed] [Google Scholar]
- 44.Melnikow J, Henderson JT, Burda BU, et al. Screening for cervical cancer with high-risk human papillomavirus testing: updated evidence report and systematic review for the US preventive services task force. JAMA 2018;320:687–705. 10.1001/jama.2018.10400 [DOI] [PubMed] [Google Scholar]
- 45.Cox JT, Castle PE, Behrens CM, et al. Comparison of cervical cancer screening strategies incorporating different combinations of cytology, HPV testing, and genotyping for HPV 16/18: results from the ATHENA HPV study. Am J Obstet Gynecol 2013;208:184.e1–184.e11. 10.1016/j.ajog.2012.11.020 [DOI] [PubMed] [Google Scholar]
- 46.Liang LA, Einzmann T, Franzen A. Cervical cancer screening: comparison of conventional pap smear test, liquid-based cytology and human papillomavirus testing as stand-alone or co-testing strategies. Cancer Epidemiol Biomarkers Prev 2020. 10.1158/1055-9965.EPI-20-1003 [DOI] [PubMed] [Google Scholar]
- 47.Ronco G, Cuzick J, Pierotti P, et al. Accuracy of liquid based versus conventional cytology: overall results of new technologies for cervical cancer screening: randomised controlled trial. BMJ 2007;335:28. 10.1136/bmj.39196.740995.BE [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Nandini NM, Nandish SM, Pallavi P, et al. Manual liquid based cytology in primary screening for cervical cancer--a cost effective preposition for scarce resource settings. Asian Pac J Cancer Prev 2012;13:3645–51. 10.7314/APJCP.2012.13.8.3645 [DOI] [PubMed] [Google Scholar]
- 49.Goldie SJ, Gaffikin L, Goldhaber-Fiebert JD, et al. Cost-Effectiveness of cervical-cancer screening in five developing countries. N Engl J Med 2005;353:2158–68. 10.1056/NEJMsa044278 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
bmjopen-2022-063622supp001.pdf (71.3KB, pdf)
Data Availability Statement
All data relevant to the study are included in the article.