Definitions

In this study, contraceptive prevalence was defined as the percentage of women who report themselves or their partners as currently using at least one contraceptive method of any type. For analytical purposes, contraceptive methods are classified as either modern or traditional. Modern methods of contraception include female and male sterilization, the intrauterine device (IUD), the implant, injectables, oral contraceptive pills, male and female condoms, vaginal barrier methods (including the diaphragm, cervical cap, and spermicidal foam, jelly, cream, and sponge), the lactational amenorrhoea method (LAM), emergency contraception, and other modern methods not reported separately (e.g., the contraceptive patch or vaginal ring). Traditional methods of contraception include rhythm (e.g., fertility awareness-based methods, periodic abstinence), withdrawal, and other traditional methods not reported separately. Unmet need for family planning is the percentage of women who want to stop or delay childbearing for at least 2 years but are not using any contraceptive method. Unmet-need calculation is based on the definition and computation used by the Demographic and Health Surveys (DHS) [19] (Section 2.2 in S1 Appendix). Demand for family planning satisfied by modern methods (SDG indicator 3.7.1 ‘Proportion of women who have their need for family planning satisfied with modern methods’) is modern contraceptive prevalence divided by the sum of contraceptive prevalence and unmet need (also referred to as total demand for family planning). We provide estimates for all WRA, UWRA, and MWRA.

MWRA pertains to women who are married (defined in relation to the marriage laws or customs of a country) and to women in a union, which refers to women living with their partner in the same household (also referred to as cohabiting unions, consensual unions, unmarried unions, or ‘living together’). UWRA pertains to women who are not married and not in a union and is a complement to MWRA. To the extent possible, this distinction was applied to all survey-based observations, or a bias was assigned to data points that had different definitions of marital or union status.

While family-planning indicators among MWRA are commonly used and readily interpreted, more consideration needs to be given to the calculation and interpretation of the family-planning indicators among UWRA and all WRA.

The major differences in the approaches used for estimating family-planning indicators among UWRA were related to whether the universe of the population was all UWRA or only UWRA who are deemed to be sexually active (defined by sexual activity in past 28 days, past 3 months, past 1 year, or ever sexually active) and so potentially exposed to pregnancy. In this paper, we report indicators for the population of all UWRA. Contraceptive users among UWRA refers to all UWRA who report in a survey using a contraceptive method, irrespective of sexual activity. For unmet need among all UWRA, it is necessary to determine the timing of their most recent sexual activity. UWRA who are not pregnant or postpartum amenorrhoeic are considered currently at risk of pregnancy (and thus could potentially be included in the numerator as having unmet need) if they have had intercourse within the 4 weeks prior to the survey interview. Unmet need for UWRA who are pregnant or postpartum amenorrhoeic is determined regardless of their most recent sexual activity either as pregnant UWRA whose pregnancies were unwanted or mistimed at the time of conception or postpartum amenorrhoeic UWRA who are not using family planning and whose last birth was unwanted or mistimed. Using this approach, estimates of family-planning indicators for all women are directly comparable to those published generally in the survey reports.

Data compilation

The data compilation uses data from multiple sources and employs unified definition and computation of survey-based estimates of family-planning indicators for UWRA. We reviewed all nationally representative household surveys that provided information on contraceptive prevalence. The starting point was the data sources used for estimating contraceptive use for MWRA [18], complemented by queries to national statistical offices, research institutes, and international survey programs. For each survey, we determined whether the question on current use of contraception was asked among all women and whether it was possible to estimate contraceptive prevalence and unmet need for family planning among UWRA. The data compilation includes 1,247 observations of contraceptive prevalence for 195 countries or areas for the period from 1950 to 2019, 540 observations of unmet need for family planning for MWRA across 143 countries, 551 observations across 136 countries or areas for contraceptive prevalence, and 250 observations across 72 countries or areas for unmet need for UWRA (Section 2.3 in S1 Appendix) [5]. For 361 surveys—including DHS; Multiple Indicator Cluster Surveys (MICS); Performance, Monitoring and Accountability 2020 (PMA2020) surveys, Gender and Generations Survey (GGS); and other surveys—we obtained estimates, including sampling errors, from micro-datasets. For other surveys, we used estimates derived from published tabulations, or we obtained specific tabulations from the institutions responsible for data collection. Many of these survey-based estimates had not been previously published. The input data are provided as Supporting Information (S1, S2 and S3 Data).

Ethics approval was not required for this study.

Statistical model

For MWRA, we used existing statistical models [8,18] on the updated data compilation. We expanded the model to UWRA (see Section 3 in S1 Appendix). Both models were fitted to the data described earlier, and we report the model-based estimates and projections. The model provides an assessment of uncertainty in the estimates based on the availability and quality of input data. It allows for greater precision when more and better data are available and indicates the extent of uncertainty in cases in which the data are insufficient or are from sources more susceptible to systematic differences as captured by the biases assigned to each survey-based data point. We describe some key aspects of it subsequently; for a full description, see Section 3 in S1 Appendix. For comparisons between model outputs and the underlying data obtained from surveys for each country, see S2 Appendix.

For each country, the systematic trends over time in total contraceptive prevalence and the ratio of modern-method use to use of any method were modelled with logistic growth curves. The logistic function is appropriate for representing social diffusion processes such as the adoption of contraceptive methods, such that the pace of adoption is slow at first and then is expected to increase at a faster rate, after which it slows down once prevalence approaches higher levels [20,21]. Autocorrelated error processes were added to account for deviations, including reversals, from these smooth trends. Unmet need among UWRA was similarly modelled as a systematic trend plus autocorrelated error but as a function of total contraceptive prevalence rather than time. This function also captures the trend in countries with low sexual activity among UWRA where both contraceptive prevalence and unmet need are very low.

Country-specific estimates of model parameters were obtained by fitting a Bayesian hierarchical model [22], which is fully described in Section 3.6 in S1 Appendix. The output of a Bayesian model is a ‘posterior’ distribution, a high-dimensional probability distribution for the parameters of interest. It is the result of combining information about the parameters contained in the data (encapsulated in the ‘likelihood’ component of the model) with existing prior knowledge (encapsulated by the ‘prior distribution’ component). Standard practice is to summarise the posterior distribution with quantiles of key parameters that provide point estimates and ranges representing the magnitude of uncertainty. We report posterior medians for point estimates and uncertainty intervals (UIs) constructed from the 2.5th and 97.5th percentiles. Under the model, and given the available data, there is a 95% probability that these UIs contain the true value of the parameter.

Many countries had only a few data points, and some countries had none. For example, 59 countries with data on contraceptive prevalence for MWRA had no such data for UWRA. Of the 136 countries that did have data, only 26% of them had any data before 1990 (compared with 63% for MWRA). To improve the precision and accuracy of estimates and projections, a hierarchical structure was added to the Bayesian model. Hierarchical structures cluster countries together based on common characteristics and allow ‘borrowing of strength’ among observations within the same cluster. Estimates for countries with little, biased, or no data are based on data for other countries in the cluster—partly if they have some data, entirely if they have none. The model implicitly weights data from other countries based on the hierarchical structure positions of the countries.

The model for MWRA used a purely geographic clustering [18]; in other words, countries geographically near each other were clustered together. To improve estimation for UWRA in countries with very few—or no—data points and to account for the cross-country variations in reproductive behaviour and sexual activity among UWRA that have an impact on family-planning indicators, we developed a 2-category sexual activity classification and combined it with geographical clusters. The 2 categories were (i) countries with very low levels of sexual activity (Group 0) and (ii) all other countries (Group 1). Eighty-one countries were assigned using information about proportion sexually active from the most recent DHS or MICS; countries with less than 2% sexually active (defined as having sexual intercourse in the past 28 days) among UWRA were assigned to Group 0 and the rest to Group 1. An additional 43 countries were classified using information about the acceptance of sex between unmarried adults reported in the Pew 2013 Global Attitudes Survey (29 countries) or the World Values Survey Wave 6 (14 countries), and the remaining 71 countries (all in Asia and Northern Africa) were assigned on the basis of the proportion religious in the population (Section 2.4 and Fig D in S1 Appendix).

Estimates for countries and indicators with no data were obtained from weighted averages of parameter estimates for certain sets of countries with data. The sets and weights were determined by the hierarchical structure of the model, such that averages based on countries most proximate in the hierarchy were given higher weights (Sections 3.4 and 3.7 in S1 Appendix). We included additional parameters in the model to account for surveys that used different target populations (e.g., nonstandard age groups, omission of subnational geographies) or a different categorization of contraceptive method use (the bias and misclassification parameters are described in Section 3.5 in S1 Appendix). For instance, some surveys include women of different age groups compared to the baseline population of women aged 15 to 49 years (e.g., the GGS, which surveyed only women aged 18 and older). To account for differences in data quality, we estimated total errors by data source type.

Estimates for years outside the periods of data availability were obtained from the fitted Bayesian model by interpolation and extrapolation of the systematic logistic trends and the autocorrelated error processes. The parameters of the systematic trend component are time invariant (Section 3.3 in S1 Appendix) and thus provide estimates of the trend for all years. The autocorrelated error process is also parameterised by time-invariant parameters. Once these were obtained from the model fit, extrapolations beyond the period of data availability were obtained by sequentially sampling from the error process conditional distributions. We labelled all estimates for years 2020 and beyond as ‘projections.’

We used a Markov chain Monte Carlo (MCMC) algorithm to generate samples from the joint posterior distribution of model parameters [22,23,24]. The MCMC sampling algorithm was implemented using JAGS 4.2.0 Open Source software [24], and the analysis was carried out in R version 3.5.2 [25]. The source code is available at https://github.com/FPcounts/FPEMglobal.

The samples from the posterior were transformed to produce probabilistic results for other indicators, such as need for family planning satisfied by modern methods. Results on the count scale, such as the number of contraceptive users or women with unmet need, were produced by multiplying posterior sample elements by country-specific estimates and projections of the number of UWRA or MWRA, as appropriate [26,27]. For instance, the posterior distribution of the number of UWRA using any method of contraception was obtained by multiplying each element of the posterior sample for contraceptive prevalence by the number of UWRA in the appropriate year, within country. Estimates for country aggregates (e.g., sub-Saharan Africa) were obtained by summing the posterior distributions of the counts across the countries in the aggregate. Estimates for UWRA and MWRA were combined to provide WRA estimates in the same way. That is, for each country at each year, UWRA and MWRA estimates were summed on the count scale to produce WRA estimates.

A key contribution of this article is the data compilation and statistical model for estimating contraceptive prevalence and unmet need for family planning among UWRA itself, therefore a prespecified analysis plan was not prepared, and all analyses are non-prespecified. The statistical model for MWRA [8,18] is applied to a larger and more recent dataset. As in Alkema and colleagues’ work [18], model development began with substantive reasoning about the prevalence of contraceptive use and needs and how they might differ between MWRA and UWRA. We identified sexual activity among UWRA as a major factor to be included in the hierarchical structure of the model for UWRA.

The validity of the model was assessed with 3 cross-validation exercises [22]. These entailed leaving out certain sets of observations (depending on the exercise), refitting the model, and checking model results against the left-out observations for calibration and predictive accuracy. The results indicated that the model was robust and adequately calibrated to generate the estimates for family-planning indicators among UWRA (Section 4.3 in S1 Appendix). We also tested the sexual activity classification for UWRA. Because there were 2 levels (‘low’ and ‘other’), we identified countries with data that were close to the threshold of 2% sexually active among UWRA. The country closest to the threshold was Indonesia, which was assigned to the ‘low’ sexual activity group. A sensitivity analysis was conducted, assigning Indonesia to the ‘other’ group, but its relatively abundant data resulted in no meaningful change in the results.

Indicators used

We computed the medians and 95% UIs for all indicators of interest using the 2.5th and 97.5th percentiles of the posterior distributions for the following 7 indicators: contraceptive prevalence (any, modern, traditional), unmet need for family planning, unmet need for modern methods, total demand for family planning, and demand for family planning satisfied with modern methods.

We generated posterior estimates of change in family-planning indicators over periods of years at the trajectory level so as to obtain probabilistic estimates of change with posterior probabilities of an increase (PPIs) or decrease (PPDs; such that PPD = 1 − PPI). These probabilities reflect the certainty regarding the reported change: a higher posterior probability corresponds to greater certainty about the result. For example, PPI > 0.9 means that there is more than 90% probability that the actual change was greater than 0. These posterior estimates of change and PPIs account for correlation between estimates of the start and end years and are more accurate than, for example, comparing the separately generated posterior estimates (including uncertainty) at the start and end years of the period.

We produced national, regional, and global results for the period from 1990 to 2030 and discuss the changes in estimates from 2000 to 2019 and projections to 2030 in line with the requirements for SDG monitoring. Results quoted for 2030 are model-based predictions, and we use different visual elements to distinguish these in plots. We present results for the 185 countries with at least 90,000 inhabitants in 2017 [26]. Estimates from an additional 10 countries and areas with total population below 90,000 inhabitants are included in the aggregated results. The classifications used to form regions, subregions, and other country aggregates are defined in S1 Appendix Table G. The results are reported according to the Guidelines for Accurate and Transparent Health Estimates Reporting (GATHER) statement [28] (S1 Checklist).

Family-planning trends among MWRA

Modern contraceptive prevalence among MWRA increased worldwide between 2000 and 2019 from 55.0% (95% UI 53.7%–56.3%) to 57.1% (95% UI 54.6%–59.5%), or 2.1 percentage points (95% UI −0.7 to 4.8, PPI > 0.93) (Table 1 and Fig 1; Section 5 in S1 Appendix, S4 and S5 Results for country-specific results; Table 2 for corresponding numbers of women). Prevalence in 2019 was highest in Eastern Asia (80.8% [95% UI 74.8%–85.4%]). Median estimates for Western Europe, South America, and Northern Europe were above 70%, while those for Northern America, Central America, and Australia and New Zealand were above 60%. In contrast, prevalence was lowest in 2019 in Middle Africa (13.7% [95% UI 10.5%–17.9%]) and Western Africa (20% [95% UI 17.8%–22.5%]). These were, however, increases over the 2000 levels by, respectively, 7.1 percentage points (95% UI 3.5–11.5) and 11.7 percentage points (95% UI 9.3–14.4), with PPI = 1 in both cases. The increases were also large in Eastern Africa, greater by 23.7 percentage points (95% UI 20.9–26.7, PPI = 1), and Eastern Europe, greater by 10.5 percentage points (95% UI 0.6–19.9, PPI > 0.98).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Global estimates of the proportion (%) of WRA using a modern contraceptive method, having unmet need for modern methods, and having demand for family planning, and the proportion of need for family planning satisfied by modern methods, for 1990, 2000, 2010, 2019, and 2030.

https://doi.org/10.1371/journal.pmed.1003026.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Global estimates of the number (millions) of WRA using a modern contraceptive method, having unmet need for modern methods, and having demand for family planning, and the proportion unmarried among all women, for 1990, 2000, 2010, 2019, and 2030.

https://doi.org/10.1371/journal.pmed.1003026.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Percentage of WRA who used a modern contraceptive method by marital group for the period 1990–2030, by World Bank income group and subregion.

The country classification by income level is based on June 2019 gross national income per capita from the World Bank. Estimates are indicated for the years 1990, 2000, 2010, and 2019 by solid circles and interpolated by solid lines. Projections for the year 2030 are indicated by open circles and interpolated by dashed lines. Shaded ribbons indicate 95% UIs; intervals for projection years are shaded lighter. UI, uncertainty interval; WRA, women of reproductive age (15–49 years).

https://doi.org/10.1371/journal.pmed.1003026.g001

Over the same period, the proportion of the need for family planning satisfied by modern methods among MWRA worldwide changed from 74.3% (95% UI 73.0%–75.4%) to 75.8% (95% UI 73.4%–78.0%) (Table 1 and Fig 2; Table 2 for corresponding numbers of women). The median estimate of change was 1.5% (95% UI −1% to 3.9%, PPI = 0.88). Eastern Asia and Western Europe had the highest proportions of need satisfied (92.9% [95% UI 88.9%–95.6%] and 89.7% [95% UI 83.5%–93.6%], respectively). Middle Africa (27.3% [95% UI 21.9%–33.6%]) and Western Africa (42.3% [95% UI 38.2%–46.6%]) had the lowest estimated proportions of need satisfied, although these were increases on 2000 levels—as for modern contraceptive prevalence—of 12.4 percentage points (95% UI 6.2–19.4, PPI = 1) and 20.1 percentage points (95% UI 15.3–24.9, PPI = 1), respectively. Eastern Africa had the third lowest estimated proportions of need satisfied in 2000 (32.7% [95% UI 31.1%–34.4%]) but a substantial increase of 29.8 percentage points (95% UI 26.3%—33.3%, PPI = 1) in the 2000–2019 period, which resulted in an estimate of 62.5% (95% UI 59.3%–65.6%) in 2019.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Proportion of need for family planning satisfied by modern methods among WRA by marital group for the period 1990–2030, by World Bank income group and subregion.

The country classification by income level is based on June 2019 gross national income per capita from the World Bank. Estimates are indicated for the years 1990, 2000, 2010, and 2019 by solid circles and interpolated by solid lines. Projections for the year 2030 are indicated by open circles and interpolated by dashed lines. Shaded ribbons indicate 95% UIs; intervals for projection years are shaded lighter. UI, uncertainty interval; WRA, women of reproductive age (15–49 years).

https://doi.org/10.1371/journal.pmed.1003026.g002

Family-planning trends among UWRA

Between 2000 and 2019, the proportion of UWRA using modern contraceptive methods increased worldwide from 15.4% (95% UI 12.0%–23.1%) to 20.1% (95% UI 16.3%–26.3%), or 4.6 percentage points (95% UI 0.2–8.7, PPI > 0.98) (see Section 5 in S1 Appendix, S2 and S3 Results for country-specific results). Contraceptive prevalence and the scale of its increase varied widely across geographic subregions. In 2019, more than a third of UWRA were using a modern method in Australia/New Zealand, the Caribbean, Northern America, Northern Europe, South America, Southern Africa, Southern Europe, and Western Europe (Table 1 and Fig 1; Table 2 for corresponding numbers of women). The 2 subregions with the largest increase since 2000 were Central America and South America, greater by 12.7 percentage points (95% UI 3.0–23.4, PPI > 0.99) and 16.8 points (95% UI 6.3–28.7, PPI > 0.99), respectively (Fig 3). Although UWRA contraceptive use generally remained low in sub-Saharan Africa (except Southern Africa), the proportion of UWRA using a modern contraceptive method increased with posterior probability of 99% in Eastern Africa by 6.7 percentage points (95% UI 4.6–9.1), in Middle Africa by 8.9 percentage points (95% UI 4.0–14.7), and in Western Africa by 7.0 percentage points (95% UI 4.2–10.0).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Percentage-points change between 2000 and 2019 in the proportion of WRA who used a modern contraceptive method, by marital status.

The 95% UIs are displayed by horizontal lines. UI, uncertainty interval; WRA, women of reproductive age (15–49 years).

https://doi.org/10.1371/journal.pmed.1003026.g003

The proportion of UWRA using a modern contraceptive method remained very low in Northern Africa, South-Eastern Asia, and Western Asia (median estimates for the subregions are less than 3%) (Fig 4A). These are subregions in which overall need for family planning and sexual activity among UWRA is low.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. 2019 estimates for family-planning indicators among UWRA.

Median estimates of (A) modern contraceptive prevalence, (B) unmet need for a modern method, and (C) the proportion of need for family planning satisfied by modern methods. For countries marked with ‘*,’ no survey-based observations are available for UWRA. The base map was obtained from Natural Earth (https://naturalearthdata.com). The boundaries and names shown and the designations used on this map do not imply official endorsement or acceptance by the United Nations. UWRA, women of reproductive age (15–49 years) who are unmarried and not in a union.

https://doi.org/10.1371/journal.pmed.1003026.g004

The use of traditional methods is not common among UWRA. Globally, 1.6% (95% UI 1.0%–3.3%) of UWRA used a traditional method in 2019. Middle Africa is an exception (5.4% [95% UI 2.6%–10.3%]) with the highest prevalence of traditional methods in Congo (11.5% [95% UI 3.1%–25.1%]).

The proportion of UWRA with unmet need for modern methods (including traditional-method users) remained unchanged since 2000 and stood at 6.6% (95% UI 4.8%–10.1%) globally in 2019. High proportions of UWRA continued to experience unmet need in Middle and Western Africa, exceeding 15% in Angola, Benin, Congo, Côte d’Ivoire, Democratic Republic of the Congo, Gabon, Guinea-Bissau, Liberia, Madagascar, and Sierra Leone (Fig 4B).

The proportion of need for family planning satisfied by modern methods among UWRA in 2019 was estimated at 75.4% (95% UI 66.1%–82.6%) globally, increasing from 68.6% (95% UI 56.2%–78.6%) in 2000. It remained low in Middle Africa (48.1% [95% UI 37.5%–57.9%]), Western Africa (58.1% [95% UI 52.0%–63.6%]), and Eastern Africa (63.0% [95% UI 58.4%–67.5%]), though it has increased since 2000 by more than 10 percentage points in all 3 subregions. Estimates of the proportion of need for family planning satisfied by modern methods among UWRA is presented for 140 countries (excluding countries classified as having low sexual activity among unmarried women) and among them, the proportion of need satisfied was below 50% in 11 countries and below 75% in an additional 74 countries in 2019 (Fig 4C, Section 5 in S1 Appendix).

UWRA contraceptive users in the context of all users

UWRA account for an increasing proportion of modern users (from 12.1% [95% UI 9.7%–17.1%] in 2000 to 15.7% [95% UI 13.4%–19.4%] in 2019; Table 2). In Northern Africa, Southern Asia, South-Eastern Asia, and Western Asia, less than 5% of all modern users were UWRA in 2019, compared to the highest share of 58.5% (95% UI 52.2%–63.7%) in Southern Africa (Fig 5). More than a third of modern contraceptive users are UWRA in Australia/New Zealand, Middle Africa, Northern America, Northern Europe, and Southern Europe. Globally, UWRA accounted for 16.0% (95% UI 12.9%–22.1%) of women with unmet need for modern methods in 2019.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Proportion of UWRA among WRA using modern contraception by World Bank income group and by subregion in 2000 and 2019.

The 95% UIs are displayed by horizontal lines. In the United Nations classification, the more developed regions comprise all countries of Europe, Northern America, Australia and New Zealand, and Japan. The less developed regions comprise all other countries. The country classification by income level is based on June 2019 gross national income per capita from the World Bank. UI, uncertainty interval; UWRA, women of reproductive age (15–49 years) who are unmarried and not in a union; WRA, women of reproductive age (15–49 years).

https://doi.org/10.1371/journal.pmed.1003026.g005

Contraceptive prevalence and need for family planning among WRA

Among all WRA, use of modern contraceptives increased worldwide between 2000 and 2019 from 42.0% (95% UI 40.5%–44.6%) to 44.3% (95% UI 42.1%–47.0%) (PPI > 0.97) and in absolute number of users from 660 million (95% UI 637–701 million) to 842 million (95% UI 800–893 million) (Tables 1 and 2, Section 5 in S1 Appendix, S6 and S7 Results for country-specific results).

Global unmet need for modern methods declined marginally from 15.1% (95% UI 14.1%–16.8%) in 2000 to 14.2% (95% UI 12.9%–15.8%) in 2019. In absolute numbers, 270 million (95% UI 246–301 million) women have an unmet need for modern methods in 2019, up from 237 million (95% UI 222–264 million) in 2000. The subregion of Southern Asia has the highest number of women with unmet need for modern methods in 2019 at 87 million (95% UI 71–109 million), with 4 other subregions above 15 million: 22 million (95% UI 18–21 million) of Eastern Africa, 20 million (95% UI 18–22 million) of Western Africa, 23 million (95% UI 20–27 million) of South-Eastern Asia, and 28 million (18–48 million) of Eastern Asia (Fig 6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Number of WRA having unmet need for modern methods by subregion in 2000 and 2019.

The 95% UIs are displayed by vertical lines. UI, uncertainty interval; WRA, women of reproductive age (15–49 years).

https://doi.org/10.1371/journal.pmed.1003026.g006

The proportion of need for family planning satisfied by modern methods among WRA globally increased from 73.5% (95% UI 71.3%–75.3%) in 2000 to 75.7% (95% UI 73.2%–78.0%) in 2019, or 2.2 percentage points (95% UI −0.3 to 4.7, PPI > 0.95) (Fig 7, Table 1). Of 76 countries that in 2000 had less than half of the demand satisfied by modern methods, 34 countries had demand satisfied above 50% in 2019, with the highest increases in Rwanda, from 13.3% (95% UI 11.4%–15.6%) in 2000 to 66.8% (95% UI 56.1%–75.9%) in 2019, and in Ethiopia, from 15.7% (95% UI 13.7%–18.2%) in 2000 to 62.5% (95% UI 55.5%–69.4%) in 2019 (PPI = 1.0) (Fig 8). Our probabilistic projections indicate that countries and regions with currently small proportions of need satisfied would achieve higher levels by 2030 as projected based on country-specific experiences together with experiences in other countries in the region and the world.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Proportion of need for family planning satisfied by modern contraceptive methods among WRA, median and 95% UIs, by subregion.

Estimates are shown for the years 2000 and 2019. The 95% UIs are displayed by solid horizontal lines and median estimates by the solid circles. Projections for 2030 are shown using open circles and dashed lines. In United Nations classification, the more developed regions comprise all countries of Europe, Northern America, Australia and New Zealand, and Japan. The less developed regions comprise all other countries. The country classification by income level is based on June 2019 gross national income per capita from the World Bank. Note: results for all countries presented in S2 Appendix. UI, uncertainty interval; WRA, women of reproductive age (15–49 years).

https://doi.org/10.1371/journal.pmed.1003026.g007

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Proportion of need for family planning satisfied by modern contraceptive methods among WRA, median and 95% UIs, by country, among countries where proportion of need for family planning satisfied by modern contraceptive methods was less than 50% in 2000.

The 95% UIs are displayed by solid horizontal lines and median estimates by the solid circles. Projections for 2030 are shown using open circles and dashed lines. UI, uncertainty interval; WRA, women of reproductive age (15–49 years).

https://doi.org/10.1371/journal.pmed.1003026.g008

The proportion of need satisfied by modern methods remains smallest in countries with low income, despite the large increase from 32.1% (95% UI 30.5%–34.0%) in 2000 to 52.6% (95% UI 50.1%–55.2%) in 2019 (PPI > 0.99) (Fig 7), with the smallest proportions of need satisfied by modern methods in Chad 23.6% (95% UI 15.9%–33.5%), Somalia 17.7% (95% UI 4.6%–44.8%), and South Sudan 17.8% (95% UI 9.5%–35.4%). Among 43 countries with less than half of their demand satisfied by modern methods in 2019, 32 are low and lower-middle income.