Abstract
Objectives
To investigate the impact of lockdown policies on the recruitment of an ongoing cohort study.
Study design
We performed descriptive analyses of recruitment, dropout, and baseline characteristics over time. Oxford Stringency Index was used to assess the impact of regional constraints on recruitment.
Results
Drop in recruitment clearly reflected the Stringency Index within the first months of the pandemic. Unexpectedly, drop-out rates declined in 2020/2021. Baseline characteristics were comparable, yet younger women were recruited more frequently during the pandemic.
Conclusions
There was no strong evidence of recruitment bias due to the pandemic.
Implications
The COVID-19 pandemic is a potential source of bias for ongoing studies and its influence on the study conduct (e.g., recruitment, drop-out) should be thoroughly evaluated to ensure that study results are not biased in this regard. The Oxford's Government Stringency Index can be used to identify pandemic-affected time periods.
Keywords: COVID-19, Europe, Intrauterine devices, IUD, Recruitment, Stringency Index
1. Introduction
On March 11, 2020, the coronavirus disease (COVID-19) was declared a global pandemic by the World Health Organization Emergency Committee [1]. Since then, the pandemic has not only affected people’s lifestyles but may have affected women‘s contraceptive use. Women were faced with service disruptions due to lockdowns and travel restrictions, which led to interrupted supply chains and overwhelmed health facilities [2]. Thus, some women did not get access to their preferred contraceptive method. For women interested in long-acting reversible contraception, such as intrauterine devices (IUDs), it was recommended to use self-administered shorter-acting methods until usual health care access resumed [3], [4], [5].
In 2014, the European Active Surveillance Study on LCS12 (EURAS-LCS12) was initiated, a multinational prospective cohort study investigating the risk of unintended pregnancy with IUDs [6], [7]. The present secondary analysis aimed to investigate whether lockdown restrictions during the pandemic affect the study conduct regarding recruitment intensity, drop-out behavior, and distribution of population’s baseline characteristics.
2. Methods
2.1. Study design
Women with a newly inserted IUD were recruited from 10 European countries (Austria, Czech Republic, Finland, France, Germany, Italy, Poland, Sweden, Spain, and the United Kingdom) via a network of approximately 1200 health care professionals (HCPs) during routine clinical practice and are being followed up for up to 5 years. An informed consent form was signed at recruitment, and ethical approval for the study was acquired following the rules in the respective countries.
2.2. Stringency Index
The Stringency Index is part of the Oxford Covid-19 Government Response Tracker to record the strictness of government policies [8]. It is calculated via nine indicators (school closing, workplace closing, cancel public events, restrictions on gatherings, public transport closing, stay-at-home requirements, restrictions on internal movement, international travel controls, and public info campaigns) and ranges from zero (no restrictions) to 100 (highest restrictions).
2.3. Statistical analyses
The relative number of monthly recruited subjects in 2020 and 2021 was displayed and compared with the Stringency Index per country. Furthermore, we calculated the proportion of women who dropped out (i.e., due to withdrawal of informed consent, going to live abroad, death, or investigator dropout) from 2018 to 2021 for 6-month intervals. We investigated potential recruitment bias by comparing baseline characteristics of women recruited before (i.e., 2019) and during (i.e., 2020) the pandemic and calculated standardized differences to measure the effect size between the two cohorts [9]. Standardized differences of 0.2, 0.5, and 0.8 indicate small, medium, and large effect sizes, respectively [10].
3. Results
3.1. Recruitment
Recruitment numbers and Stringency Index between January 2020 and December 2021 per country are shown in Figure 1. Due to a temporary recruitment stop in 2020 (unrelated to the pandemic), the United Kingdom data were excluded from Figure 1. In the beginning of the pandemic, the number of recruiting HCPs in Italy, Poland, Finland, and Sweden dropped by approx. 13% to 29%. The Stringency Index showed high concordance with recruitment during the first months of the pandemic for all countries except Germany. Recruitment numbers declined when Stringency Index reached its first peak. Afterwards, with decreasing Stringency Index, the recruitment increased. However, recruitment numbers did not substantially change at Stringency Index peaks in Winter 2020 or Spring 2021.
Fig. 1.
Association of recruitment numbers from the EURAS-LCS12 study in 2020–2021 and the Oxford’s Government Stringency Index per country.
3.2. Dropout
The proportion of women who chose to dropout from the study was low before and during the pandemic, with a steady decline from 0.40% in 2018 to 0.16% in 2020 (data not shown). However, at the beginning of 2021, the proportion of active dropouts was 0.23%. No regional differences in drop-out behavior could be detected.
3.3. Baseline characteristics
We recruited 9788 women in 2019 and 8949 women in 2020 ( Table 1). The mean age of women recruited before and during the pandemic was comparable, yet the proportion of younger women (<20 years) was slightly higher in women recruited during the pandemic (10.5% vs 8.8%). Women‘s Body Mass Index, gravidity, parity, smoking status, and number of sexual partners in the past 12 months did not substantially differ between the two cohorts. However, women recruited during the pandemic had a low household income (57.8% vs 51.7%) more frequently than women recruited before the pandemic. In 2020, women were more likely first-time users (69.8% vs 67.4%) and less likely consecutive users of the IUD (13.4% vs 15.3%) than women recruited in 2019. The proportion of women living single was higher in 2019 than 2020 recruitments (27.2% vs 23.8%). Standardized differences for all baseline parameters were below the threshold of 0.2.
Table 1.
Baseline characteristics of EURAS-LCS12 study participants recruited before and during the COVID-19 pandemic (i.e., in 2019 and 2020) across 12 European countries
| Recruitment in 2019 | Recruitment in 2020 | Standardized differencea | |
|---|---|---|---|
| Number of women | 9788 (100%) | 8949 (100%) | |
| Age (years) | −0.04 | ||
| Mean (SD) | 29.3 (6.37) | 29.1 (6.53) | |
| Age category | 0.04 | ||
| <20 years | 863 (8.8%) | 937 (10.5%) | |
| 20 to <30 years | 4120 (42.1%) | 3724 (41.6%) | |
| 30 to <40 years | 4805 (49.1%) | 4288 (47.9%) | |
| Body mass index (kg/m2) | 0.00 | ||
| <30 | 8394 (85.8%) | 7697 (86.0%) | |
| ≥30 | 1318 (13.5%) | 1159 (13.0%) | |
| Missing | 76 (0.8%) | 93 (1.0%) | |
| Gravidity | −0.03 | ||
| Nulligravid | 3581 (36.6%) | 3386 (37.8%) | |
| Gravida | 6207 (63.4%) | 5563 (62.2%) | |
| Parity | −0.02 | ||
| Nulliparous | 4059 (41.5%) | 3797 (42.4%) | |
| Parous | 5729 (58.5%) | 5152 (57.6%) | |
| Education level | 0.17 | ||
| Less than university entrance level | 2731 (27.9%) | 2881 (32.2%) | |
| University entrance level | 3360 (34.3%) | 3017 (33.7%) | |
| More than university entrance level | 3501 (35.8%) | 2727 (30.5%) | |
| Missing | 196 (2.0%) | 324 (3.6%) | |
| Income | 0.17 | ||
| Two lowest categories | 5058 (51.7%) | 5177 (57.8%) | |
| Two highest categories | 3983 (40.7%) | 2918 (32.6%) | |
| Missing | 747 (7.6%) | 854 (9.5%) | |
| IUD user status | 0.16 | ||
| First-time user | 6595 (67.4%) | 6246 (69.8%) | |
| Repeat user | 1668 (17.0%) | 1486 (16.6%) | |
| Consecutive user | 1501 (15.3%) | 1198 (13.4%) | |
| Missing | 24 (0.2%) | 19 (0.2%) | |
| Smoking | 0.04 | ||
| Current | 2155 (22.0%) | 1917 (21.4%) | |
| Ex-Smoker | 1693 (17.3%) | 1400 (15.6%) | |
| Never | 5841 (59.7%) | 5509 (61.6%) | |
| Missing | 99 (1.0%) | 123 (1.4%) | |
| Marital status | 0.08 | ||
| Living single | 2659 (27.2%) | 2130 (23.8%) | |
| Living together with a partner | 6782 (69.3%) | 6339 (70.8%) | |
| Missing | 347 (3.5%) | 480 (5.4%) | |
| Number of sexual partners in the past 12 months | 0.14 | ||
| 0 | 179 (1.8%) | 243 (2.7%) | |
| 1 | 7886 (80.6%) | 7087 (79.2%) | |
| >1 | 1457 (14.9%) | 1279 (14.3%) | |
| Missing | 266 (2.7%) | 340 (3.8%) |
Standardized differences (Stddiff) are calculated acc. to Yang & Dalton [9] and discussions. For continuous variables: Stddiff = (Meangr1 − Meangr2)/Sqrt((Vargr1 + Vargr2)/2). For categorical variables: Stddiff = Sqrt((T − C)′S−1(T − C)), where T and C denote vectors of proportions for the variable levels in groups 1 and 2. S is the covariance matrix with diagonal elements defined as 0.5*(tk(1 − tk) + ck(1 − ck)) and off-diagonal elements defined as −0.5*(tktl + ckcl).
4. Discussion
This secondary analysis investigated the impact of COVID-19 pandemic–related restrictions on recruitment in a large ongoing cohort study. Except for Germany, we observed that recruitment numbers were highly associated with the Stringency Index at the beginning of the pandemic. Similarly, Roland and colleagues reported fewer IUD dispensations shortly after lockdown, but dispensations increased 1 month after lockdown ended [11]. According to the United Nations Population Fund, many countries could restore access to their health services shortly after the beginning of the pandemic, which probably explains why recruitment dropped only in the first months [12]. Stable recruitment in Germany might be attributed to the decentralized character of crisis management [13]. Especially in the first phase of the pandemic, restrictions were dispersedly implemented only by some states and local governments. Furthermore, the capacity and resilience of the German care system were assessed as extraordinarily high compared to other European countries [13]. Most subjects included in the German COVID-19 Snapshot Monitoring study reported not having problems accessing medical care [13], [14].
Unexpectedly, the drop-out rate slightly decreased during the pandemic. We hypothesize that women spent more time filling out study questionnaires or answering follow-up phone calls from the study sites during lockdown phases. Furthermore, our study population consists of mainly healthy younger women who may be less likely to belong to COVID-19 risk groups and, therefore, less likely to drop out from the study due to health issues or death. However, before the pandemic, we already observed a decline in drop-out numbers in 2019. Therefore, the reduction in dropouts may result from reasons unrelated to the pandemic.
Baseline characteristics of women recruited before and during the pandemic were comparable. However, we observed that fewer singles were recruited during the pandemic. This might be due to decreased personal need for long-acting reversible contraception as lockdown policies hampered social contacts, including sexual partners. Women recruited in 2020 indicate lower household income than those recruited in 2019, which may be attributed to short-time work or unemployment due to the pandemic.
In conclusion, there was no strong evidence of bias in recruitment and selection in our study due to the lockdown restrictions during the COVID-19 pandemic. Therefore, it is unlikely that the final results of the EURAS-LCS12 study will be affected in this regard.
Acknowledgments
The authors would like to express their appreciation to the participating HCPs, study participants, and numerous colleagues responsible for the fieldwork in the individual countries and management of the study database.
Footnotes
Conflicts of interest: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this article.
Funding: The EURAS-LCS12 study was funded by an unconditional grant from Bayer AG (Germany). An independent international Advisory Council was responsible for all scientific matters. The funder had no access to the source data and did not participate in analyzing the data or preparing this publication.
References
- 1.World Health Organization. Coronavirus disease (COVID-19) pandemic: overview. 〈https://www.who.int/europe/emergencies/situations/covid-19〉. Accessed 4 Jan 2023.
- 2.Riley T., Sully E., Ahmed Z., Biddlecom A. Estimates of the potential impact of the COVID-19 pandemic on sexual and reproductive health in low- and middle-income countries. Int Perspect Sex Reprod Health. 2020;46:73–76. doi: 10.1363/46e9020. [DOI] [PubMed] [Google Scholar]
- 3.Benson LS, Madden T, Tarleton J, Micks EA. Society of Family Planning interim clinical recommendations: contraceptive provision when healthcare access is restricted due to pandemic response: 2021 update: Society of Family Planning; 2021.
- 4.Ferreira-Filho E.S., de Melo N.R., Sorpreso I.C.E., Bahamondes L., Simões R.D.S., Soares-Júnior J.M., et al. Contraception and reproductive planning during the COVID-19 pandemic. Expert Rev Clin Pharmacol. 2020;13:615–622. doi: 10.1080/17512433.2020.1782738. [DOI] [PubMed] [Google Scholar]
- 5.Wilkinson T.A., Kottke M.J., Berlan E.D. Providing contraception for young people during a pandemic is essential health care. JAMA Pediatr. 2020;174:823–824. doi: 10.1001/jamapediatrics.2020.1884. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.U.S. National Library of Medicine. ClinicalTrials.gov: European Active Surveillance Study of LCS12 (EURAS_LCS12). 〈https://clinicaltrials.gov/ct2/show/NCT02146950〉 (accessed January 4, 2023).
- 7.Boehnke T., Bauerfeind A., Hagemann C., Lange J.A., Heinemann K. Intrauterine device prescribing patterns and types in Europe – a cross-sectional analysis of the European Active Surveillance LCS12 Study. EGO. 2022;4:29–40. [Google Scholar]
- 8.Hale T., Angrist N., Goldszmidt R., Kira B., Petherick A., Phillips T., et al. Variation in government response to COVID-19: BSG-WP-2020/032. 2022. Version 14.0. [Google Scholar]
- 9.Yang D., Dalton J.E. SAS Global Forum – Statistics and Data Analysis. 2012. A unified approach to measuring the effect size between two groups using SAS®. [Google Scholar]
- 10.Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Lawrence Erlbaum Associates; 1988. [Google Scholar]
- 11.Roland N., Drouin J., Desplas D., Cuenot F., Dray-Spira R., Weill A., et al. Effects of the coronavirus disease 2019 (COVID-19) lockdown on the use of contraceptives and ovulation inductors in France. Obstet Gynecol. 2021;137:415–417. doi: 10.1097/AOG.0000000000004281. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.United Nations Population Fund . Impact of COVID-19 on family planning: what we know one year into the pandemic. 2021. [Google Scholar]
- 13.Joyce P., Maron F., Reddy P.S. Good public governance in a global pandemic: between unity and variety: Germany’s responses to the COVID-19 pandemic. 1st ed. Kuhlmann; 2020. [Google Scholar]
- 14.Hajek A., de Bock F., Wieler L.H., Sprengholz P., Kretzler B., König H. Perceptions of health care use in Germany during the COVID-19 pandemic. Int J Environ Res Public Health. 2020;17:9351. doi: 10.3390/ijerph17249351. [DOI] [PMC free article] [PubMed] [Google Scholar]