Study selection
Searches were performed on 4th May, 2022 with a repeat search performed on 21st August, 2023. A total of 3,049 articles were reviewed after duplicates were removed, with 2,970 being excluded on initial screening. Seventy-nine texts were eligible for full review, and 10 studies were included in the final review. Common reasons for exclusion were lack of original data inclusive of the COVID-19 pandemic era and wrong article type (i.e, commentary). The PRISMA flowchart can be seen in Fig. 1.
Figure 1: PRISMA flowchart
Figure legen: PRISMA flow diagram depicting the number of papers identified, screened and excluded.
The characteristics of the included studies can be seen in Table 1. The risk of bias scores for each included study can be seen in Table 2. All studies were graded as having high risk of bias. The certainty of evidence (GRADE [9]) can be seen in Table 3.
Table 1
Characteristics of included studies
Authors | Study type | Study Region | Country | Study sites | Time period of study | Study outcomes | Age of participants | Number of adolescents included | Data collected | Overall result | Study results |
Shikuku et al [10] | Retrospective cohort | African Region (AFR) | Kenya | Urban and Rural | 2019–2020 | Adolescent pregnancy | 10–14, 15–19, 10–19 years | 253,218 | Adolescent pregnancy rate | Decrease in adolescent pregnancy | Proportion of adolescent pregnancy reduced from 0.4 to 0.3 in 10–14 years olds (p < 0.0001) and from 8.4 to 7.0 in 15–19 year olds (p < 0.0001) |
Kassie et al [11] | Retrospective cohort | African Region (AFR) | Ethiopia | Urban and Rural | March–June 2019, March–June 2020 | Proportion of teenage pregnancy, proportion of teenage abortion | Not defined beyond "teenage" | 644 | Pregnancies in those < 19 years | Increase in adolescent pregnancy | The proportion of teenage pregnancy increased from 7.5–13.1% |
Forum for African Women Educationalists (FAWE) Uganda Chapter [12] | Retrospective cohort | African Region (AFR) | Uganda | Urban and Rural | March to October 2020 | Adolescent pregnancy | 10 to 14 subgroup | 237892 | 1st ANC attendance | Increase in adolescent pregnancy | In 10-14-year olds, the number of young girls who attended first ANC increased by 366.5% from 290 in March 2020when the country entered into a lock down to 1,353 in September 2020. |
Barron et al [13] | Retrospective cohort | African Region (AFR) | South Africa | Urban and Rural | 2017–2021 | Numbers of deliveries in the public sector to adolescent girls aged 10–14 years and 15 - 19 years | 10–14, 15–19, 10–19 years | 686487* | Adolescent pregnancy rate | Increase in adolescent pregnancy | Pregnancy rate 10–19 years, per 1 000: 2017: 25.8, 2018: 27.9, 2019: 28.7, 2020: 28.4, 2021: 30.0 |
Zulaika et al [14] | Prospective cohort | African Region (AFR) | Kenya | Rural | 2018–2021 | Incident pregnancy in girls who became pregnant during Forms 3–4 (est. age 17–18) | Enrollment via school year. Mean age 17.2 in pre-COVID cohort, and 17.5 in COVID-19 cohort | 910 | Adolescent pregnancy rate | Increase in adolescent pregnancy | Incident pregnancy between the start of Form 3 and completion of examinations was 10.9% among COVID-19 cohort girls versus 5.2% in the pre-COVID-19 cohort. |
Martin et al [15] | Retrospective cohort | Region of the Americas (AMR) | United States of America | Urban and Rural | 2021 | Birth rate by age | All age groups, sub-group 15–19 | Not reported, total of 3,664,292 births in whole population in 2021, 3,613,647 in 2020 | Adolescent birth rate. Age-specific birth rates: Births per 1,000 females in the specified age group. | Decrease in adolescent pregnancy | The birth rate for teenagers aged 15–19 declined 7% from 2020 to 2021 (15.0 to 13.9 per1,000 births); this rate declined by 10% from 2019 to 2020 |
Alunyo et al [16] | Retrospective cohort | African Region (AFR) | Uganda | Unsure | March 2019 to March 2021 | First access of antenatal care (ANC) services for adolescent girls aged 10–19 | 10 to 19 | 4,122 | first access of antenatal care (ANC) services for adolescent girls aged 10–19 | Increase in adolescent pregnancy | Increased by 1.53 pregnancies per month [95% CI: -3.4 to 6.0] after lockdown commenced Overall, 8% increase in teenage pregnancies in the district during COVID-19 lockdowns |
Paudel et al [17] | Retrospective cohort | Region of the Americas (AMR) | United States of America | Urban and Rural | January - June 2019, January - June 2021 | Adolescent pregnancy | Not defined beyond "adolescent" | 244 | Adolescent pregnancy rate | No change in adolescent pregnancy | 5.8% of 4151 pregnancies in Pre-COVID era (4.9–6.9 9%CI) in pre COVID 19 era, 5.9% of pregnancies (5.0–7.1 95% CI ) |
Musinguzi et al [18] | Cross-sectional | African Region (AFR) | Uganda | Unsure | December 2021 to January 2022. | Pregnancy among teenage girls aged 13–19 years | 13–19 years | 314 | Ever pregnant during the COVID-19 pandemic | Increase in adolescent pregnancy | 30.6% of teenage girls had gotten pregnant during the COVID-19 pandemic. 2019 adolescent pregnancy rate 29% in this region. |
Monteiro et al [19] | Retrospective cohort | Region of the Americas (AMR) | Brazil | Urban and Rural | 2001–2022 | Adolescent pregnancy | 10 to 19 | 381653 | Adolescent pregnancy rate | Decrease in adolescent pregnancy | Adolescent pregnancy in Brazil in 2020 represented 14% of total LB, which is equivalent to a reduction of 8.4% in relation to 2019. |
Table 2
Risk of Bias scores of included papers
Study author | Country | Bias score |
Monteiro et al | Brazil | high risk of bias |
Kassie et al | Ethiopia | high risk of bias |
Zulaika et al | Kenya | high risk of bias |
Shikuku et al | Kenya | high risk of bias |
Barron et al | South Africa | high risk of bias |
Alunyo et al | Uganda | high risk of bias |
Forum for African Women Educationalists (FAWE) Uganda Chapter | Uganda | high risk of bias |
Musinguzi et al | Uganda | high risk of bias |
Martin et al | USA | high risk of bias |
Paudel et al | USA | high risk of bias |
Table 3
GRADE assessment for certainty of evidence
GRADE Assessment for the effect of lockdown measures on rate of adolescent pregnancies (studies n = 10) |
GRADE criteria | Rating | Comment | Certainty of evidence |
Study Designs | Observational | Downgraded 2 for lack of controlled setting | ⮾○○○ Very low |
Risk of Bias | Very serious | Downgraded 2. All studies were assessed as having a high risk of bias due as data came from different study groups and no adjustment for confounding was performed |
Inconsistency | Very serious | Downgraded 2.Half of included studies indicated an increase in adolescent pregnancy and half indicated a decrease. There was a large variation in the magnitude of effect. |
Imprecision | Some concern | Confidence intervals for baseline or outcome data not provided. Most studies had sufficient size of information extracted from national registries. A synthesis of data was not possible due to heterogeneity of study settings, study design, and type of outcome measurements. |
Indirectness | Very serious | Downgraded 2. The ouctome of interest was by definition measured in different study groups, before and during the pandemic respectively but always in the same setting. Four studies used proportion of adolescent pregnancies and not pregnancy rate as outcome measure. |
Publication bias | None detected | |
Other considerations (upgrading factors) | None detected | |
Result of individual studies
A total of 10 studies were included in the final analysis (Table 1). All studies came from two regions: Africa (7/10) and the Americas (3/10). Most studies (8/10) were performed in low- or middle-income settings.
A mix of study designs were reported. We noted many authors self-defined their studies as cross-sectional when retrospective analysis of regional or national datasets had been performed. These studies were classified as retrospective cohorts by reviewers as it was felt to be a more accurate definition of study design. The predominant study type involved retrospective analysis of regional or national datasets or retrospective cohorts (8/10), with cross-sectional studies (1/10) and prospective cohort studies (1/10) being the minority. Increase in adolescent pregnancy was noted in just over half of the studies (6/10, 54.5%), all of which were undertaken in Africa. Other studies took place in Brazil (1/10) and the United States of America (2/10). The results of individual studies by region are discussed below. A brief description of school closures in each country is also provided (see Fig. 2).
Figure 2, School Closure status by country, source: UNESCO map on school closures and UIS, March 2022 [1].
Africa
Uganda
In Uganda, all schools closed in March 2020 with a partial reopening in October 2020. Schools were closed again in June 2021 and reopened fully in January 2022. Uganda had the longest school closure globally, with a total of 66 weeks of full closure and 23 weeks of partial closure [1].
Three studies were performed in Uganda across different regions of the country. Musinguzi et al [18] performed a cross-sectional study between December 2021 to January 2022 in Hoima district, Western Uganda. Of the 319 girls aged 13–19 enrolled via multi-stage sampling across the district, there was a rate of adolescent pregnancy of 30.6% during the study period. The authors noted a preceding adolescent pregnancy rate of 29% in the region, though no statistical analysis was performed to examine the significance of this variation.
An interrupted time series analysis examining first access of antenatal care (ANC) for girls aged 10–19 was undertaken between March 2019 to March 2021 in Pakwach district, Northern Uganda, by Alunyo et al [16]. The records of 4,122 adolescents were included in the study. After the implementation of COVID-19 lockdowns the adolescent pregnancies increased by 1.53 pregnancies per month [95% CI: -3.4 to 6.0], with an overall 8% increase in teenage pregnancies in the district. The increase was not statistically significant. Analysis of risk and protective factors was not undertaken.
A report by the Forum for African Women Educationalists (FAWE) in Uganda reviewed results of the National Health Management Information System (HMIS) [12]. Between March and October 2020, a total of 234,839 pregnancies in 10–14-year-olds and 434,939 pregnancies in 15–19-year-olds were recorded. Amongst the youngest girls, aged 10–14, the number attending first ANC rose from 290 in March 2020 to 1,353 in September 2020, a rise of 366.5%. Kikuube district, Western Uganda, reported the highest number of cases (n = 1,106). Some districts reported no cases (Buhweju, Ibanda, Karenga, Kisoro, Rubanda, and Nabilatuk), though pregnancies in more rural areas with limited antenatal services had not been captured. In girls aged 15–19, the highest increase in ANC attendance were recorded between March and June 2020 with an increase of 25.5% between these months. By October 2020, cases had fallen back to pre-lockdown numbers.
Kenya
In Kenya, all schools closed on the 16th March 2020. Guidelines were developed by June 2020 to facilitate school reopening and a phased opening of certain grades commenced in October 2020. In January 2021, there was full resumption of in-person instruction [1].
Two studies were performed in Kenya. Shikuku et al [10] performed a retrospective cohort study examining the Kenyan Health Information System (KHIS) utilisation data for 2019 and 2020. Data for adolescents aged 10–14 and 15–19 were reviewed. Hospital data for the first four months of the pandemic (March to June 2020) were compared with the same period in 2019. For 10–14-year-olds, there were 4,971 pregnancies in the 2020 time period (proportion of all pregnancies, 0.3%) compared to 6,872 in the same period in 2019 (proportion of all pregnancies 0.4%). A similar trend was seen in girls aged 15–19 (107,667 pregnancies in 2020 compared to 133,708 pregnancies in 2019; proportion of all pregnancies, 8.4%). Significance testing in both age groups suggests a true reduction in adolescent pregnancies in March-June 2020 compared to the same months in 2019
(p < 0.001). Adolescent maternal deaths (10–19 years) rose as a proportion of all maternal deaths from 6.2% in 2019 (23/373) to 10.9% (45/412) in 2020 (p = 0.009).
In contrast to this, a prospective cohort study of schoolgirls performed by Zulaika et al [14] in Siaya County, Western Kenya, performed between May 2018 and March 2021 reported that school disruptions caused by lockdowns doubled the risk of adolescent pregnancy amongst girls in the COVID-19 cohort (Relative Risk = 2.11,; CI: 1.13 to 3.95). The rate of pregnancy in the COVID-19 cohort was 10.9% compared to 5.2% in the pre-COVID group. Pregnancy incidence increased from 5.2 per 100 person-years at risk between January 2019 to March 2020 (CI: 3.64 to 7.37) to 8.8 per 100 person-years at risk during the 7 months in which schools were closed (March to October 2020) (CI: 5.91 to 13.16, p = 0.046).
Ethiopia
Schools in Ethiopia closed in March 2020, with a phase re-opening commencing in October 2020 [1].
A retrospective cohort study was performed in South-West Ethiopia by Kassie et al [11], comparing the pregnancies of girls under the age of 19 in 2019 and 2020 (March – June in both time periods). Data were extracted from the medical records at selected governmental medical facilities. The proportion of adolescent pregnancy increased from 7.5% in 2019 (n = 285) to 13.1% in 2020 (n = 359) .
South Africa
Schools in South Africa closed in March 2020 for a total of 10 weeks, with a subsequent phased reopening. Schools were closed again in July 2020 amidst a rising wave of infections. There were, then, phases of easing restrictions following trends in infection. Schools re-opened in February 2021 [1].
In South Africa, Barron et al [13] performed a time-series analysis on public health data for the years 2017–2021. Data were presented in girls aged 10–14 and in those aged 15–19. Overall, the pregnancy rate for 10–19 year-olds rose annually from 27.9 per 1000 girls in 2018 to 30 per 1000 girls in 2021. The upward trend in adolescent pregnancy occurred year on year, but occurred at higher rates in more rural provinces. No further analysis of data from the COVID-19 era to the pre-COVID-19 era was performed.
The Americas
Brazil
Brazilian state and municipal governments first adopted non-pharmaceutical interventions to mitigate COVID-19 in March 2020. A total of 26 states and the Federal District closed schools. Remote learning programmes were developed, but had limited accessibility. Schools re-opened for in-person learning in the second half of 2020 and in 2021 [1].
In Brazil, Monteiro et al [19] examined the Live Births Data System (SINASC) of the Brazilian Health Ministry. Data were obtained on the number of live births per region and by age groups (10–14 and 15–19 years) in 2019 and 2020. Age specific fertility rate per 1,000 adolescents in the same age group was calculated. In 2020, the adolescent mothers accounted for 14% of all live births, a reduction of 8.4% compared to 2019. Age specific fertility rates decreased in both age groups by 8% and 8.4%, respectively. No statistical testing was performed.
USA
The first COVID-19 school closure occurred in late February 2020 in Washington state. By the end of March all but one US public school district was closed,with learning shifting to online platforms. State policies varied on return to in-person learning [1].
Two studies were performed in the USA. Paudel et al examined adolescent pregnancy in urban and rural Appalachian populations by reviewing electronic medical records between January-June 2019 and January-June 2021 [17]. There was no significant difference in the rate of adolescent pregnancies between the two time points (5.8–5.9%).
Martin et al examined the National Centre for Health Statistics to review birth records from 2020 and 2021 to assess pregnancy rates in girls aged 15–19 [15]. The birth rate for this group fell by 10% between 2019 to 2020 and fell again by 7% between 2020 and 2021 (15.0 to 13.9 per 1,000 births).