Contextual factors associated with contraceptive utilization and unmet need among sexually active unmarried women in Kenya: A multilevel regression analysis.

BACKGROUND: Unmarried women who report less recent sexual intercourse (>30 days from survey enumeration) are largely excluded from global health monitoring and evaluation efforts. This study investigated level and contextual factors in modern contraceptive utilization and unmet need within this overlooked female subpopulation in Kenya from 2014 to 2019.
METHODS: This study analyzed data from the Performance Monitoring and Accountability (PMA) survey in Kenya, a nationally representative survey of female respondents, to understand the level and contextual factors for family planning utilization and unmet need within female subgroups including married, unmarried sexually active (defined as sexual intercourse within 30 days of survey enumeration), and unmarried with less recent sexual intercourse (defined as sexual intercourse 1-12 months prior to survey enumeration). The analysis included multilevel regression modeling to assess correlates on outcomes of modern contraceptive prevalence rate (mCPR), unmet need, and recent emergency contractive pill (ECP) use, which is a unique PMA question: "Have you used emergency contraception at any time in the last 12 months?".
RESULTS: Cumulatively, the surveys enumerated 19,161 women and this weighted analysis included 12,574 women aged 15-49 from three female subgroups: 9,860 married women (78.4%), 1,020 unmarried sexually active women (8.1%), and 1,694 unmarried women with less recent sexual intercourse (13.5%). In 2019, while controlling for covariates, unmarried women with less recent sexual intercourse exhibited statistically significant differences (p-value<0.02) in current mCPR, mCPR at last sexual intercourse, unmet need for modern contraceptives, and recent ECP use. As compared to an unmarried woman with less recent sexual intercourse (i.e., reported sex 1-12 months prior to survey), the odds of an unmarried sexually active woman (i.e., reported sex within last 30 days of survey) currently using modern contraceptives was 2.28 (95% CI: 1.64, 3.18), using modern contraceptives at last sexual intercourse was 1.44 (95% CI: 1.06, 1.95), and having an unmet need for modern contraceptives was 2.01 (95% CI: 1.29, 3.13) while controlling for covariates. The odds of a married woman using ECP during the last 12 months was 0.60 (95% CI: 0.44, 0.82) as compared to an unmarried woman with less recent sexual intercourse. In 2019, unmarried women with less recent sexual intercourse reported the highest rate of ECP use during the last 12 months at 13.5%, which was similar for unmarried sexually active women at 13.3%. Since 2014, summary measures of unmet need and total demand for modern contraceptives increased for unmarried women with less recent sexual intercourse, but declined for the other female subgroups.
CONCLUSION: In Kenya, unmarried women with less recent sexual intercourse exhibited significantly different contraceptive utilization, unmet need, and recent emergency contraceptive use. Moreover, changes over time in key family planning indicators were asymmetrical by female subgroup. This study identifies an important monitoring gap regarding unmarried women with less recent sexual intercourse. Evidence dissemination by the global measurement community for these unmarried women is exceedingly scarce; therefore, developing an inclusive research agenda and actionable information about these marginalized women is needed to enable targeted planning and equitable service delivery.

Introduction

Access to high-quality family planning services has been identified as one of the most cost-effective strategies to improve health and development outcomes for women and their households [1]. Since the London Summit on Family Planning in 2012, multiple global efforts have sought to strengthen and provide equitable access to family planning services, including high-impact, low-cost methods, such as contraceptive implants and emergency contraceptives [2-5]. Over the last decade, modern contraceptive use has substantively increased in low- and middle-income countries (LMICs); however, the number of women with unmet need–where she wants to avoid pregnancy but is not using a modern contraceptive method—is also rising [6, 7]. Each year, 218 million women have an unmet need for modern contraceptives and approximately 111 million pregnancies are unintended [7]. Differences between married and unmarried women, who were sexually active in the last 30 days, are well documented with married women typically exhibiting lower contraceptive use and unmet need [8, 9]. However, family planning indicators for unmarried women with less recent sexual intercourse (>30 days) are largely unreported.

Women who want to avoid pregnancy, but are not using modern contraceptive methods, account for 77% of unintended pregnancies [7]. Emergency contraception, which is administered within a few days after sexual intercourse, can help prevent pregnancies due to non-use, failure or misuse of contraceptive, or situations of rape or coerced sex [10, 11]. While the copper intrauterine device is considered an emergency contraceptive method; the leading option is emergency contraceptive pills (ECP), which are oral contraceptive pills for women to use as soon as possible (up to 5 days) after sexual intercourse to prevent unwanted pregnancy [11]. ECP has a pregnancy prevention rate ranging from 56% to 95% if promptly used [12-16]. Suitably, ECP was selected as one of 13 high-impact, low-cost commodities by the UN Commission on Life-saving Commodities for Women and Children (UNCoLSC) [17]. ECP use is highest among two groups of women: aged 20–24 years and unmarried sexually active [18, 19]. ECP is safe for over-the-counter sale and often available from a pharmacist or drug seller without a prescription [20].

With a population of approximately 47 million, including 24 million women, Kenya is one of the most populous countries in sub-Saharan Africa and classified as a lower-middle income economy [21, 22]. Approximately 59.7% of women are currently married, which includes married or in union (i.e., living together) with a male partner [23]. Based on Kenya’s most recent Demographic and Health Survey (DHS) in 2014 [23], the modern contraceptive prevalence rate (mCPR) was 39.1% for all women (including the subgroups of married women at 53.2% mCPR and unmarried sexually active women at 60.9% mCPR), which is one of the highest in SSA; however, ECP use was not reported. Correspondingly, unmet need for all women was 12.8% (including 17.5% for married women and 26.4% for unmarried sexually active women), which is relatively low for SSA. Kenya has implemented policies to reduce barriers to access family planning, such as policies enacted in 2013 to effectively eliminate family planning user fees as well as other public outpatient costs [24]. ECP is free at public health facilities and available for purchase without prescription in private pharmacies [25-27]. Since the last DHS was conducted over seven years ago, other data sources are needed to investigate recent changes in family planning practices in Kenya. This analysis utilized data from the Performance Monitoring and Accountability (PMA) survey [28], which was a nationally representative survey on family planning usage, knowledge, and experience of women. In addition, PMA incorporated a unique ECP question: “Have you used emergency contraception at any time in the last 12 months?”. This question has a longer recall period than the traditional ‘current use’ ECP indicator, which underestimates the scale of ECP usage [11].

For family planning indicators, ‘sexually active’ is most commonly defined as a woman having sexual intercourse within one month (four weeks or 30 days) prior to the day of survey enumeration [29-32]. Sexual activity within one month is dramatically different between married and unmarried women. The proportion of married women who were sexually active (within one month) ranges from 50% to 91% across countries in SSA, while the sexual activity of unmarried women exhibits a lower range of 1% to 39% [30]. According to Kenya’s DHS report in 2014, 79.5% of married women and 6.8% of unmarried were sexually active [30]. Research by Dasgupta et al. indicates when extending the definition of sexual activity beyond one month (e.g., 3 months or 12 months) the proportion of unmarried women who are considered sexually active increases drastically, while married women exhibit a modest increase [30]. As compared to married women, sexual encounters for unmarried women can be sporadic and unpredictable [31, 33]. Extending the time interval since last sexual activity for unmarried women can highlight the contraceptive needs of an underreported female subpopulation at risk of unintended pregnancies.

This study aims to evaluate the level in modern contraceptive utilization and unmet family planning needs among female subpopulations in Kenya: married or in union (i.e., living together); unmarried and sexually active within the past 30 days prior to survey (labeled as UA-30days); and unmarried and sexually active between 1–12 months prior to the survey (labeled as UA-12months). The latter, unmarried sexually active women with less recent sexual intercourse (between 1–12 months prior to survey), are underreported by the global health measurement community. Moreover, the analysis utilizes the unique survey design of current PMA questionnaires to investigate the level of recent emergency contraceptive use (within 12 months prior to survey). Lastly, the study assesses the relative effect of contextual factors (e.g., female subgroup, demographics, socioeconomic status) on these family planning outcomes and how these relationships changed over time in Kenya.

Methods

Study setting

PMA/Kenya was a nationally and county-level representative survey in Kenya from 2014 to 2019 that used a multi-stage stratified cluster design with urban-rural classification and geographic county as strata. The survey was enumerated in nine counties in 2014 and 11 counties in 2017 and 2019. Within each county, the sample of enumeration areas (EA) was selected by the Kenya National Bureau of Statistics using its master sample frame to provide a representative estimate of modern contraceptive prevalence rate (mCPR). Within each EA, 42 households were randomly selected for enumeration. Within each household, all eligible females aged 15–49 were designated for interview. Enumeration was conducted by local female residents using mobile technology for rapid data collection and quality control.

The sample of service delivery points (SDPs) include both public and private facilities where the catchment area falls within the EA boundary. Public facilities include health posts, primary health centers and the district hospitals. Up to three private facilities are randomly selected for enumeration if providing adequate maternal, reproductive, or general health services. Female resident enumerators survey private facilities, while enumeration supervisors survey public SDPs. For more details on PMA sampling and methodology see Zimmerman et al. [34].

Data source and measurement

This study used data from three questionnaires: household, female, and service delivery point. The household questionnaire outlines the household roster and socioeconomic measures. The female questionnaire includes marital status, recency of sexual activity, family planning use, contraceptive knowledge, and recent experience with healthcare providers. The SDP questionnaire addresses the type of facility, service offerings, commodity availability (i.e., stock-outs), and fee structure among others.

In Kenya, PMA enumeration began in 2014 with annual or semi-annual cycles of data collection. For this study, data from 2014, 2017, and 2019 were analyzed. The 2017 survey was selected as the midpoint, because PMA introduced the question on emergency contraceptive use within the last 12 months in the 2017 questionnaire. Thus, 2017 will act as the baseline data point for the ECP model. In 2017 and 2019, PMA included an additional two counties, Kakamenga and West Pokot; however, this study restricted the analysis to women in the original nine counties to ensure comparability with 2014 data. Table 1 provides additional enumeration details for each PMA survey.

Table 1

Description of PMA survey enumeration, PMA/Kenya (2014, 2017, 2019).

	2014	2017	2019
Time of Collection	May—July 2014	Nov—Dec 2017	Nov—Dec 2019
PMA Cycle (and citation)	Round 1 [35]	Round 6 [36]	Round 8 [37] [i.e. “Phase 1”]
Counties	9	11	11
Enumeration Areas (EA)	120	151	308
Households (% Response Rate)	4,530 (93.2%)	6,106 (97.8%)	10,378 (98.1%)
Females (% Response Rate)	3,807 (95.9%)	5,876 (99.0%)	9,478 (98.7%)
Facilities (% Response Rate)	263 (nr)	417 (97.2%)	945 (94.6%)

References for description of PMA survey enumeration [35–37]

Nr = Not reported

PMA/Kenya was managed by the Ministry of Health in partnership with International Centre for Reproductive Health Kenya (ICRHK), National Council for Population and Development, and Kenya National Bureau of Statistics. Johns Hopkins University (USA) and Jhpiego provided general direction and technical support. PMA was funded by the Bill & Melinda Gates Foundation.

Study variables

The conceptual framework for analysis is outlined in Fig 1.

Fig 1

Conceptual framework for analysis.

Female subgroups (marital status, sexual activity)

Female respondents are categorized into three subgroups based on marital status and recency of sexual activity:

Unmarried and sexually active between 1–12 months prior to the survey (UA-12months);

Unmarried and sexually active within past 30 days prior to survey (UA-30days); or

Married or in union.

Primary outcome measures

This study modeled four outcome measures at the level of the individual female respondent.

Current mCPR: The first model analyzed modern contraceptive rate (mCPR), which is the proportion of women 15–49 years old who are using (or partner using) a modern method of contraception at the time of the survey (or ‘current’ mCPR). Long-acting modern methods include intra-uterine device (IUD), implant, and sterilization (male and female), while short-acting methods include injectable, pill, emergency contraception, male or female condoms, diaphragm, lactational amenorrhea method (LAM), and the standard days/cycle beads method. Note: traditional methods, such as withdrawal and rhythm, are not considered modern methods.

mCPR at last sex: The second model assessed mCPR at time of last sexual intercourse. According to Fabic et al. [31], using ‘current’ mCPR, which represents contraceptive use at time of survey, likely underreports mCPR for women with less recent sexual activity. A temporal misalignment occurs between ‘current’ use of modern contraceptives and last sexual intercourse, which could have been months prior. Measuring contraceptive use at last sexual activity can remedy this misalignment; however, historically few surveys include this type of question [31]. Data collection for this outcome indicator was only available for the 2019 PMA survey.

Unmet Need: The third model explored unmet need for family planning defined as the proportion of fertile, sexually active women 15–49 years old who are not using contraception and do not want to become pregnant at any time (unmet need for limiting) or within the next two years (unmet need for spacing).

Recent Emergency Contraceptive use: The final model analyzed emergency conceptive (ECP) use within the last 12 months, which is measured by the question (322a) “Have you used emergency contraception at any time in the last 12 months?” or if the female respondent is currently (question 302b) or recently (question 306b) using ECP.

Total Demand: The proportion of female respondents or their partners who are currently using modern contraceptives (current mCPR) plus the female respondents with an unmet need for modern contraceptives (i.e., current mCPR plus unmet need).

Demand satisfied by modern contraceptives: The proportion of women who are currently using modern contraceptives within the population of women demanding modern contraceptives (i.e., current mCPR divided by total demand).

Other explanatory variables

See Table A.1 in S1 Appendix for definitions of additional explanatory variables evaluated including socio-demographic and wealth characteristics; healthcare delivery experience; and measures from the SDP questionnaire.

Statistical analysis

First, univariate analyses were conducted for the explanatory variables across each survey period. Selected explanatory variables were assessed against the contraceptive use and unmet need outcome variables using bivariate regression analysis (not shown). To evaluate the relative effect of explanatory covariates (e.g., sociodemographic, female subgroup), multilevel regression models were built for each of the four outcome variables with EA, household, and female respondent as the respective levels to account for the hierarchical structure of the PMA/Kenya dataset (i.e., women nested within household and households nested within EAs). Relative to single-level regression, multilevel regression models will more accurately estimate standard errors of regression coefficients and properly assess statistical significance when analyzing hierarchical data.

To build the multilevel regression models, explanatory variables were added in a forward stepwise manner. Inclusion of explanatory variables was steered by the conceptual framework to showcase varying categories of influential factors. Variables were retained in the model if they were statistically significant for an outcome variable (p-value < 0.05) or improve goodness of fit (p-value > F value). To support interpretation of each covariate, the odds ratio (OR) was calculated and presented in Tables 4–7, which represents the odds of the outcome given exposure to the covariate. Statistical significance was assessed using 95% confidence intervals and p-values of 0.10, 0.05 and 0.01.

Table 4

Logistic Regression model for Current Modern CPR by Survey Round, PMA/Kenya (2014, 2019).

Characteristic	Categories	KENYA 2014 (Baseline)					KENYA 2019 (Endline)
		OR (95% CI)			P-value		OR (95% CI)			P-value
		Odds Ratio	Lower bound	Upper bound			Odds Ratio	Lower bound	Upper bound
Age Group	15–24 (ref)	1					1
	25–34	1.11	0.87	1.43	0.387		0.72	0.60	0.86	0.000	***
	35–44	0.86	0.60	1.23	0.396		0.47	0.38	0.59	0.000	***
	45–49	0.27	0.17	0.42	0.000	***	0.23	0.17	0.31	0.000	***
Parity	None (ref)	1					1
	1–2	4.30	2.67	6.91	0.000	***	4.65	3.67	5.90	0.000	***
	3–4	6.95	4.12	11.70	0.000	***	7.93	5.78	10.89	0.000	***
	5 plus	6.37	3.72	10.92	0.003	***	8.80	6.15	12.59	0.000	***
Education	Never (ref)	1					1
	Primary	3.17	2.01	4.99	0.000	***	2.31	1.58	3.39	0.000	***
	Secondary or more	3.86	2.50	5.97	0.000	***	2.42	1.63	3.61	0.002	***
Household wealth quintile	Lowest (ref)	1					1
	Middle lowest	1.29	0.99	1.68	0.064	*	0.95	0.76	1.18	0.626
	Middle	1.39	1.05	1.85	0.024	**	1.01	0.82	1.24	0.923
	Middle highest	1.43	1.00	2.03	0.050	*	1.25	0.99	1.57	0.060	*
	Highest	1.48	0.92	2.36	0.104		1.01	0.75	1.37	0.925
Residence	Urban (ref)	1					1
	Rural	0.83	0.62	1.10	0.192		0.76	0.59	0.96	0.025	**
Female subgroup	Unmarried sexually active (1–12 months)	1					1
	Unmarried sexually active (0–30 days)	2.18	1.24	3.82	0.007	***	2.28	1.64	3.18	0.000	***
	Married or in union	1.63	1.08	2.48	0.021	**	1.11	0.87	1.41	0.396
Fees for health provider	No (ref)	1					1
	Yes	1.27	0.95	1.70	0.107		0.90	0.68	1.19	0.452
Recently heard FP message	No (ref)	1					1
	Yes	1.39	0.94	2.06	0.099	*	1.11	0.91	1.36	0.287

P-value

* P < 0.10

** P < 0.05

*** P <0.01

Ref: Reference Category

Table 7

Logistic regression model for recent emergency contraceptive use (in last 12 months) by survey round, PMA/Kenya (2017, 2019).

Characteristic	Categories	KENYA 2017 (Baseline)					KENYA 2019 (Endline)
		OR (95% CI)			P-value		OR (95% CI)			P-value
		Odds Ratio	Lower bound	Upper bound			Odds Ratio	Lower bound	Upper bound
Age Group	15–24 (ref)	1					1
	25–34	1.55	1.00	2.40	0.049	**	1.04	0.78	1.39	0.783
	35–44	1.07	0.53	2.14	0.850		0.85	0.59	1.23	0.391
	45–49	0.48	0.14	1.61	0.230		0.42	0.22	0.79	0.008	***
Parity	None (ref)	1					1
	1–2	0.47	0.31	0.71	0.000	***	0.80	0.53	1.21	0.291
	3–4	0.65	0.41	1.02	0.063	*	0.83	0.52	1.33	0.430
	5 plus	0.58	0.24	1.39	0.216		0.86	0.47	1.57	0.614
Education	Never (ref)	1					1
	Primary	0.48	0.34	0.68	0.002	***	1.42	0.48	4.19	0.526
	Secondary or more	1.00	−	−	−		2.60	0.83	8.19	0.102
Household wealth quintile	Lowest (ref)	1					1
	Middle lowest	1.66	0.88	3.11	0.114		1.11	0.66	1.86	0.697
	Middle	2.08	0.98	4.41	0.056	*	1.20	0.72	1.99	0.481
	Middle highest	1.53	0.67	3.49	0.314		1.89	1.14	3.14	0.014	**
	Highest	2.64	1.10	6.31	0.030	**	1.95	1.14	3.31	0.015	**
Residence	Urban (ref)	1					1
	Rural	1.13	0.54	2.35	0.749		0.77	0.57	1.04	0.091	*
Female subgroup	Unmarried sexually active (1–12 months)	1					1
	Unmarried sexually active (0–30 days)	1.53	0.85	2.74	0.150		0.93	0.60	1.43	0.727
	Married or in union	0.71	0.43	1.18	0.184		0.60	0.44	0.82	0.001	***
Fees for health provider	No (ref)	1					1
	Yes	0.56	0.23	1.36	0.198		1.13	0.74	1.71	0.577
Recently heard FP message	No (ref)	1					1
	Yes	1.13	0.58	2.21	0.711		1.00	0.67	1.50	0.991

P-value

* P < 0.10

** P < 0.05

*** P <0.01

Ref: Reference Category

All analyses were performed in STATA software version 14 (College Station, TX, USA). Survey weights were applied to all estimates and sample sizes to adjust for each woman’s likelihood of selection.

Results

Cumulatively, the weighted analysis included 12,574 women aged 15–49 within the three female subgroups for the three enumeration periods–ranging from 2,956 women in 2014 to 5,957 women in 2019 (Table 2). About two-thirds of the female respondents were 15–34 years old, more than half had parity between 1–4 children, and approximately half completed only primary school or received no education. In 2019, approximately 62% of women lived in rural communities, which is fairly consistent across time periods. Most respondents were married, but a total of 1,694 respondents (13.5%) across the three enumeration periods were unmarried and sexually active between 1–12 months prior to the survey (UA-12months).

Table 2

Background characteristics for women aged 15–49, PMA/Kenya (2014, 2017, 2019).

		2014		2017		2019
Characteristics	Categories	n	%	n	%	n	%
Age Group	15–24	792	26.8	1,028	28.1	1,588	26.7
	25–34	1,246	42.2	1,473	40.2	2,341	39.3
	35–44	710	24.0	903	24.7	1,535	25.8
	45–49	208	7.0	257	7.0	493	8.3
Parity	None	285	9.6	619	16.9	781	13.1
	1–2	1,210	40.9	1,401	38.3	2,272	38.1
	3–4	860	29.1	1,007	27.5	1,743	29.3
	5 plus	601	20.3	634	17.3	1,161	19.5
Education	Never	114	3.9	109	3.0	198	3.3
	Primary	1,637	55.4	1,779	48.6	2,859	48.0
	Secondary or more	1,205	40.8	1,773	48.4	2,900	48.7
Household wealth quintile	Lowest	641	21.8	525	14.3	909	15.3
	Middle lowest	758	25.7	768	21.0	1,225	20.6
	Middle	548	18.6	761	20.8	1,378	23.1
	Middle highest	576	19.6	781	21.3	1,226	20.6
	Highest	423	14.4	826	22.6	1,219	20.5
Residence	Urban	1,145	38.7	1,390	38.0	2,266	38.0
	Rural	1,811	61.3	2,271	62.0	3,691	62.0
Female subgroups	Unmarried sexually active (1–12 months)	291	9.8	558	15.2	845	14.2
	Unmarried sexually active (0–30 days)	154	5.2	338	9.2	528	8.9
	Married or in union	2,511	84.9	2,765	75.5	4,584	77.0
Total Women		2,956		3,661		5,957

Levels of contraceptive use by female subgroup

From 2014 to 2019, current mCPR increased for each female subgroup with the highest rate of change at 16 percentage points among unmarried women, who were sexually active within 0–30 days of the survey (UA-30days) (Fig 2). In 2014, the highest current mCPR was among married women, but UA-30days eclipsed married women by 2019 with an mCPR of 62%. In terms of contraceptive method mix, married women heavily favored implants and injectables, while unmarried women preferred male condoms, implants, and injectables (Table 3). From 2014 to 2019, all female subgroups decreased short-acting modern methods and increased both long-acting modern methods (predominantly contraceptive implants) and traditional techniques (e.g., rhythm, withdrawal). In 2019, mCPR at last sexual intercourse for UA-12months was 12 percentage points higher than their current use of modern contraceptives (Fig 3), while UA-30days exhibited a three percentage point relative increase.

Fig 2

Current modern contraceptive prevalence rate by female subgroup.

Table 3

Contraceptive method mix by female subgroup, PMA/Kenya (2014, 2017, 2019).

Female Population:	Unmarried Sexually Active (1–12 months)				Unmarried Sexually Active (0–30 days)				Married or In Union
PMA2020 Survey Year:	2014	2017	2019	Change	2014	2017	2019	Change	2014	2017	2019	Change
Number of women using any contraceptives	97	232	402		69	201	354		1,350	1,694	2,893
Current Method Mix
Female Sterilization	3%	2%	2%	0%	0%	0%	2%	2%	4%	4%	5%	1%
Male Sterilization	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%
Implant	17%	20%	24%	8%	16%	20%	28%	11%	18%	34%	38%	19%
IUD	5%	1%	2%	‐3%	2%	1%	4%	2%	5%	3%	5%	0%
Injectable: intramuscular (IM)	47%	30%	22%	‐25%	44%	31%	25%	‐19%	58%	45%	33%	‐25%
Injectable: subcutaneous (SC)	0%	0%	2%	2%	0%	0%	3%	3%	0%	0%	3%	3%
Pill	6%	4%	5%	‐1%	11%	8%	9%	‐3%	10%	8%	8%	‐2%
Emergency Contraceptive Pills (ECP)	9%	13%	10%	1%	9%	14%	4%	‐5%	1%	1%	1%	0%
Male Condom	14%	23%	22%	9%	17%	21%	19%	2%	3%	2%	3%	0%
Female Condom	1%	0%	0%	‐1%	0%	1%	0%	0%	0%	0%	0%	0%
Diaphragm	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%
Cycle beads	0%	0%	1%	1%	0%	1%	1%	1%	0%	1%	1%	0%
LAM1	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%
Rhythm	0%	4%	5%	5%	1%	2%	4%	3%	0%	2%	2%	2%
Withdrawal	0%	2%	2%	2%	0%	1%	1%	1%	0%	0%	1%	1%
Other Traditional2	0%	0%	2%	2%	0%	0%	2%	2%	0%	0%	2%	2%
Total	100%	100%	100%		100%	100%	100%		100%	100%	100%
Injectable subtotal
Injectable total (IM + SC)3	47%	30%	24%	‐23%	44%	31%	28%	‐16%	58%	45%	35%	‐22%
Type of Method
Long-acting modern method4	24%	23%	29%	5%	18%	22%	33%	16%	27%	41%	47%	20%
Short-acting modern method5	76%	71%	62%	‐14%	81%	75%	60%	‐21%	73%	56%	48%	‐25%
Traditional method6	0%	6%	9%	9%	1%	3%	6%	6%	0%	3%	5%	5%

1. LAM: Lactational amenorrhea method.

2. Other traditional: all respondent-mentioned other methods.

3. IM: Intramuscular injectable; SC: subcutaneous injectable

4. Long-acting modern methods include intra-uterine device (IUD), implant, and sterilization (male and female)., while short-acting methods.

5. Short-acting modern methods include injectable, pill, emergency contraception, male or female condoms, diaphragm, lactational amenorrhea method (LAM), and the standard days/cycle beads method.

6. Traditional methods include rhythm method, withdrawal, and all respondent-mentioned other methods (labeled ‘other traditional’):.

Fig 3

Modern contraceptive prevalence rate at last sexual intercourse compared to current use mCPR by female subgroup (2019 survey).

Levels of unmet need and total demand by female subgroup

Unmet need for family planning indicated a complementary decline for married and UA-30days over time (Fig 4). Since 2014, unmet need for married and UA-30days decreased by 10 and 24 percentage points, respectively. However, UA-12months more than doubled from 6% to more than 13%. These changes over time were consistent in both unmet need for spacing and limiting (Table A.2 in S1 Appendix). Total demand (mCPR plus unmet need) for UA-30days was highest amongst the female subgroups at 85% and displayed a similarly proportioned decline over time for married women (Fig 5); however, UA-12months increased 16 percentage points to 57%. By 2019, the percent of total demand satisfied by modern contraceptives (Fig 6) was clustered between 73% and 80% for all female subgroups. UA-30days improved dramatically over time, but UA-12months declined since 2014.

Fig 4

Unmet need for modern contraceptives by female subgroup.

Fig 5

Total demand for modern contraceptives by female subgroup.

Fig 6

Proportion of demand satisfied for modern contraceptives by female subgroup.

Levels of recent emergency contraceptive use by female subgroup

In 2019, recent use of emergency contraceptive was highest for UA-12months at 13.5% along with the largest gain of 2.9 percentage points since 2017 (Fig 7). Married and UA-12months increased over time, while UA-30days exhibited a slight decline.

Fig 7

Recent emergency contraceptive use (within the last 12 months) by female subgroup.

Contextual factors influencing modern contraceptive use

Table 4 reports odds ratios of current mCPR while controlling for demographic, socioeconomic, and service delivery factors. In 2019, women who were younger, had more children, had higher levels of education, and lived in urban areas exhibited significantly higher rates of current modern contraceptive use. Since 2014, these factors affecting current mCPR typically became more pronounced, while the influence of household wealth was minimized. While controlling for all other factors, UA-12months reported significantly lower rates of current modern contraceptive use relative to UA-30days. This discrepancy between unmarried groups grew more evident since 2014. Married women had significantly higher current mCPR relative to UA-12months in 2014, but that difference waned over time.

Hearing family planning messages in the community was a modest factor in 2014, but that influence dissipated by 2019. While not statistically significant, having at least one facility not charging fees in the enumeration area increased mCPR and improved the model’s overall performance, so the indicator was retained across models. No other service delivery, patient contact or experience indicators were a significant factor in the regression models.

For mCPR at last sexual intercourse in 2019 (Table 5), there was a significant difference for female subgroups. UA-12months reported significantly higher modern contraceptive use at last sexual intercourse as compared to married women, but significantly lower than UA-30days. Relative to the current mCPR model (Table 4), the magnitude of effect (odds ratio) was smaller for last sexual intercourse when comparing contraceptive use between UA-30days and UA-12months. The other socio-demographic factors (e.g., age, parity, education, residence) exhibited similar influence on mCPR at last sex as compared to current mCPR; however, the odds ratio for parity was almost half the size.

Table 5

Logistic regression model for mCPR at last sexual intercourse, PMA/Kenya 2019.

Characteristic	Categories	KENYA 2019 (Endline)
		OR (95% CI)			P-value
		Odds Ratio	Lower bound	Upper bound
Age Group	15–24 (ref)	1
	25–34	0.74	0.62	0.88	0.001	***
	35–44	0.55	0.44	0.68	0.000	***
	45–49	0.35	0.26	0.47	0.000	***
Parity	None (ref)	1
	1–2	2.54	2.04	3.17	0.000	***
	3–4	4.27	3.15	5.77	0.000	***
	5 plus	4.61	3.33	6.40	0.000	***
Education	Never (ref)	1
	Primary	1.88	1.23	2.90	0.004	***
	Secondary or more	2.25	1.44	3.49	0.000	***
Household wealth quintile	Lowest (ref)	1
	Middle lowest	1.02	0.81	1.27	0.882
	Middle	1.22	0.96	1.54	0.101
	Middle highest	1.43	1.08	1.88	0.012	**
	Highest	1.44	1.03	2.04	0.035	**
Residence	Urban (ref)	1
	Rural	0.68	0.50	0.91	0.009	***
Female subgroup	Unmarried sexually active (1–12 months)	1
	Unmarried sexually active (0–30 days)	1.44	1.06	1.95	0.019	**
	Married or in union	0.67	0.54	0.83	0.000	***
Fees for health provider	No (ref)	1
	Yes	1.27	0.90	1.80	0.178
Recently heard FP message	No (ref)	1
	Yes	1.13	0.93	1.38	0.205

P-value

* P < 0.10

** P < 0.05

*** P <0.01

Ref: Reference Category

Contextual factors influencing unmet need

In 2019, unmet need for contraception was lower for women who were older, had received more education, were in higher income categories, and lived in urban areas (Table 6). Since 2014, the age factor grew more prominent, while the education and economic influence weakened. When controlling for all other factors, unmet need for contraception was significantly different between female subgroups with UA-30days at approximately twice the odds of having unmet need relative to UA-12months in 2019. In 2014, married women had significantly higher unmet need compared to UA-12months, but that differential dissolved by 2014. Over time, the odds ratios relative to UA-12months have decreased, which indicates a narrowing gap between the unmet needs of UA-12months and the other two female subgroups.

Table 6

Logistic regression model for unmet need by survey round, PMA/Kenya (2014, 2019).

Characteristic	Categories	KENYA 2014 (Baseline)					KENYA 2019 (Endline)
		OR (95% CI)			P-value		OR (95% CI)			P-value
		Odds Ratio	Lower bound	Upper bound			Odds Ratio	Lower bound	Upper bound
Age Group	15–24 (ref)	1					1
	25–34	0.88	0.62	1.25	0.479		0.68	0.55	0.85	0.001	***
	35–44	1.12	0.76	1.65	0.558		0.62	0.48	0.81	0.001	***
	45–49	0.90	0.51	1.60	0.714		0.37	0.25	0.54	0.000	***
Parity	None (ref)	1					1
	1–2	0.91	0.62	1.33	0.620		0.79	0.61	1.01	0.055	*
	3–4	0.93	0.62	1.41	0.741		0.96	0.68	1.35	0.801
	5 plus	1.28	0.75	2.21	0.361		1.39	0.96	2.01	0.078	*
Education	Never (ref)	1					1
	Primary	0.35	0.21	0.56	0.000	***	0.66	0.45	0.99	0.031	**
	Secondary or more	0.36	0.22	0.58	0.000	***	0.61	0.40	0.93	0.022	**
Household wealth quintile	Lowest (ref)	1					1
	Middle lowest	0.69	0.52	0.91	0.009	***	0.99	0.77	1.26	0.906
	Middle	0.57	0.41	0.80	0.002	***	0.88	0.68	1.14	0.325
	Middle highest	0.65	0.44	0.95	0.026	**	0.72	0.54	0.95	0.022	**
	Highest	0.56	0.35	0.89	0.016	**	0.70	0.49	1.01	0.057	*
Residence	Urban (ref)	1					1
	Rural	1.04	0.72	1.50	0.920		1.30	1.01	1.67	0.040	**
Female subgroup	Unmarried sexually active (1–12 months)	1					1
	Unmarried sexually active (0–30 days)	13.47	6.30	28.83	0.000	***	2.01	1.29	3.13	0.002	***
	Married or in union	5.12	2.71	9.65	0.000	***	1.16	0.81	1.66	0.416
Fees for health provider	No (ref)	1					1
	Yes	1.11	0.82	1.50	0.496		0.79	0.53	1.19	0.265
Recently heard FP message	No (ref)	1					1
	Yes	0.83	0.57	1.19	0.304		0.84	0.68	1.05	0.127

P-value

* P < 0.10

** P < 0.05

*** P <0.01

Ref: Reference Category

Contextual factors influencing recent emergency contraceptive use

For use of emergency contraceptives within the past 12 months, women aged 45–49 years or living in rural communities were less likely to use, while women in the two highest wealth quintiles were more likely to use ECP recently (Table 7). Since 2017, when this ECP question was first asked, the influence of parity, education and age (between 25–34 years) dissipated. For the female subgroups, married women had significantly lower recent usage of ECP relative to UA-12months in 2019, but not 2017. There was no difference in recent ECP usage between the two unmarried groups.

Discussion

In Kenya, contraceptive dynamics differ by subpopulations of women and the change over time is asymmetric. While controlling for covariates, female subgroups—based on marital status and recency of sexual intercourse—exhibited significant differences for current mCPR, mCPR at last sexual intercourse, unmet need, and recent ECP use during the last 12 months. Since 2014, current mCPR has increased for all female groups, while unmet need, total demand, demand satisfied by modern contraceptives, and recent ECP use are moving in different directions depending on the female subgroup. Unmarried women who were sexually active between 1–12 months prior to survey (UA-12months) notably reported an increase in both unmet need and current mCPR as well as the highest rate of emergency contraceptive use in the last 12 months. Understanding the uniqueness of each female subpopulation will be critical for designing and implementing effective public health programs for all women.

Consistent with global trends [6, 7, 38], modern conceptive use grew across each female sub-population since 2014 with long-acting modern methods chosen over short-acting methods (Table 3). Short-term injectables were swapped for implants–particularly among married women–which matches a broader shift in contraceptive preferences [38]. Female condom use was minimal across all groups despite a comprehensive global push during the time period and the unique disease prevention and empowerment features of female condoms [4, 39]. In terms of accessibility, stock-out at health facilities was not a factor in mCPR, which is consistent with a similar analysis in Kenya [34]. However, women in rural communities had significantly lower mCPR, which may indicate geographic accessibility constraints [40]. While the timing differences between ‘current’ contraceptive use questions and the sexual activity interval of 12 months for unmarried women can generate relatively low CPR and high unmet need [31], mCPR at last sexual intercourse does not include this timing misalignment and this model exhibited similar influence by socio-demographic and economic factors relative to current mCPR models in 2019. However, mCPR at last sexual intercourse for UA-12months was significantly higher than married women and significantly lower than UA-30days, which provides further evidence to consider these unmarried women as a distinct subpopulation for family planning analysis and program design. Developing monitoring systems to track family planning needs and trends in these unmarried women could be the mechanism for understanding how to build and adapt effective universal programming for women.

The female subpopulations diverged on unmet need for contraception. While unmet need doubled for unmarried women with less recent sex (UA-12months), the other female subgroups exhibited a sharp decline over the five-year period. For UA-12months, large increases in both mCPR and unmet need indicate a deep, previously unaddressed demand [8] for contraceptives and represents a critical area for further research and public health programming. The increase over time in unmet need may indicate a growing risk of unintended pregnancy for these unmarried women; however, the increased use of ECP allows for greater agency in response to unexpected sexual encounters. When controlling for female subgroup, the impact of education and wealth weakened since 2014, which may indicate positive effects of Kenya’s 2013 policies to eliminate user fees on family planning services and public outpatient primary care [24]. Unmet need increased for rural and younger women, which are common underserved demographic groups across sub-Saharan Africa [9, 41, 42], and represent a practical target for improvement in family planning programs.

When contraceptive needs are unmet, emergency contraceptives are an important option for women [9-11]. Consistent with other analyses [11], unmarried sexually active women in Kenya have significantly higher rates of recent ECP use (within the last 12 months) relative to married women (Table 7). However, this analysis also indicates a growth in ECP use over time–in particular for unmarried women with less recent sex (UA-12months), where more than one out of every eight women have used ECP in the last 12 months (Fig 7). In this study, unmarried sexually active women shifted away from shorter term modern methods, while increasing use of traditional methods, such as withdrawal, along with ECP. This matches other international reports of unmarried women increasing use of traditional methods [30] and utilizing ECP as back-up for possible method failure [33]. Sexual encounters for unmarried women can be sporadic or unpredictable [31, 33] and restricting recency measurements of sex to 30 days typically biases mCPR upward and unmet need downward [31]. In Kenya, this underreported subgroup of unmarried women with less recent sex exhibited similarly lower mCPR (current and at last sexual intercourse) and higher unmet need relative to their female counterparts–and is increasingly leveraging ECP access to prevent unwanted pregnancies. With ECP free in public health facilities and available without prescription at private pharmacies, approaches to further expand ECP uptake, such as targeted demand generation campaigns, may benefit this female subgroup in managing unintended pregnancies [27, 43].

To reach universal health coverage, public health policy makers and practitioners must understand the behavior and healthcare preferences of varying subpopulations in order to design and manage equitable healthcare delivery systems [5]. This study illustrates how unmarried women with slightly less recent sexual intercourse (0–30 days versus 1–12 months prior to survey) have significantly different contraceptive use and unmet needs for contraceptives. However, this subpopulation of unmarried women with less recent sex (UA-12months)–which constitutes 13% of women in this study–are routinely absent from data analysis and reporting. Effective data-informed decision-making requires iterative cycles of data collection, analysis, dissemination, and corrective action to improve service delivery [44], but the global health measurement community stops at data collection for unmarried women with less recent sex. Without timely information, decisions by public health policy makers and program managers are blinded to the needs of these unmarried women, which effectively marginalizes them through data processing. Moreover, this ‘data marginalization’ further exacerbates existing biases of health workers providing less favorable contraceptives services to young, unmarried sexually active women [8, 11]. While fewer in number, the ramifications of contraceptive access are equally critical for these women, their families, and community.

Beyond Kenya, analysis of unmarried women with less recent sexual intercourse is limited or nonexistent; therefore, it is unclear whether these differences exhibited between female subgroups in Kenya are common in SSA or an aberration. Conducting analyses like this study for other countries or regions is an important topic for future research. Moreover, implementation research is needed on how to design family planning programs for these marginalized female subgroups, who already experience stigma within the health delivery system [8, 11], without limiting the reproductive health improvements for all women. Building a research agenda to better understand unmarried sexually active women can help counteract the systemic bias against these women that pervades all levels of the global healthcare system.

This analysis has unique strengths, such as minimizing temporal bias by matching the time intervals (i.e., 12 months prior to survey) for recent ECP use and sexual activity of unmarried women; however, several limitations must be identified. First, as previously mentioned, there is a timing misalignment between ‘sexually active’ in last 12 months and the calculation of ‘current’ use of contraceptives, which can underestimate contraceptive use and increase unmet need when extending the sexual activity time interval [31]. Analysis of mCPR at last sexual intercourse was incorporated to mitigate this limitation and strengthen interpretation of the results. Second, there is a relatively low sample size for the unmarried groups, particularly in the early surveys, which limits statistical power. Third, there are slight differences in the survey questionnaires between survey years, such as no question on ECP in last 12 months within the 2014 questionnaire. Fourth, the datasets provided by PMA do not allow for linking the respondent to the SDP of actual service delivery, which lessens the validity of service delivery indicators. Also, unmarried women are more likely to underreport sexual activity and contraceptive use, but research to quantify this bias is limited [8]. Lastly, analysis of women in Kenya may not be emblematic or representative of other countries or regions in SSA.

In Kenya, contraceptive use and unmet need are asymmetrically changing among female subgroups. Unmarried women with less recent sexual activity exhibit different modern contraceptive use and unmet need; however, reporting and research by the measurement community is extremely limited for these women. Recognizing this measurement bias and generating targeted information for these marginalized groups is the first step towards inclusive decision-making and equitable service delivery.

(DOCX)

Click here for additional data file.

3 Jan 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Joseph KB Matovu, Ph.D.

Academic Editor

PLOS ONE

Journal requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf  and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability.

Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized.

Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

We will update your Data Availability statement to reflect the information you provide in your cover letter.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: N/A

Reviewer #3: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This is an interesting and generally well written manuscript describing contraceptive demand and utilization in unmarried sexually active women in Kenya. I believe that the manuscript would benefit from some revisions before publication.

Suggested revisions below:

Abstract

- Although the text is generally clear, there are parts where it is very biostats language heavy and could benefit from edits to improve clarity regarding the clinical implications for example "trending in different directions", "survey enumeration" in the background section. Whilst I appreciate that these terms are correct, it may be clearer to say "differences particularly in women not sexually active 1-12 months", or "reporting no recent sexual activity in preceding month".

Background

- It would be helpful to get additional context regarding the Kenyan population. From the manuscript it appears that most women are married, but I'm not certain if this is true generally in Kenya or whether this is true for women who contributed to the data. This would be helpful to understand. This could potentially provide substantial bias as there may be stigma associated with sexual activity outside of marriage and mainly unmarried women may be less likely to contribute. Similarly, there is an implication that unmarried women do not have sexual intercourse which needs to be clarified.

- Do women in Kenya ever use the copper IUD as EC? Only the contraceptive pill EC is discussed.

- In the paragraph: "With a population of ......" in the 4th line it is difficult to understand the breakdown of the 39% of women using contraception (CPR) please edit for clarity

- In the last paragraph "For family planning indicators....." a number of sentences would fit better into methods-suggest removing these and focussing on how this study addresses outstanding issues.

- The last paragraph of the background: additional text here can also be moved to methods

Methods

- Generally clear

Results

- The results are interesting and well represented

- It is particularly interesting that women who are in the unmarried sexually active group are the highest users of EC. In the discussion this is posed as a negative, but perhaps the authors should also emphasise that this is a real positive in terms of women perhaps choosing not to permanently use contraception, but able to access EC when needed. Obviously a longer term choice may be preferable but it is important that women are able to access this option

- Is there a correlation between the sexually active women with unmet contraception need and pregnancy in this study? If women are not becoming pregnant unplanned, then perhaps this is not as much of a concern.

- Also interesting the increase in implant use in this cohort

- The tables are good and clear

- The figures are difficult to understand as no figure headings and legends are included for figure 2 onwards

Discussion

- Well written

- Please see previous comments regarding clarifying this issues and findings, for example the fact that unmarried sexually active women 1-12 months access EC is a plus rather than concerning in my opinion, hopefully averting unplanned pregnancies.

- I'm not sure that the 3rd paragraph is substantiated. "The female subpopulations....". The paper doesn't report pregnancy in the group of unmarried sexually active, so it is possible that using EC is acceptable particularly if the encounter was unplanned/unexpected. It is a positive that women have sufficient agency to choose when they wish to access contraception. Obviously in this cohort if women do have unplanned pregnancies then this is a concern.

- In the paragraph " To reach universal...." whilst it is appreciated that unmarried women who are sexually active 1-12 months may be of concern, I wonder how relevant this specific focus on marriage is, to many women in different countries and regions. There is a danger of focussing separately on this group as this could be stigmatising and essentially messages regarding contraceptive access should reach all women. I realise that is what the authors are saying and perhaps more context regarding marriage in Kenya would be helpful. It seems unlikely that there are so few unmarried sexually active women relative to married women and in itself it seems to perpetuate the message that unmarried women should not be sexually active, which is problematic and likely compounds the issue. Perhaps these points can be discussed.

Reviewer #2: Thank you for the opportunity to review this article. I would only suggest a few changes in this paragraph (2nd para. of Introduction):

(…) The emergency contraceptive pill (EC) is an oral, hormonal contraceptive pill for women to use as soon as possible (up to 5 days) after sexual intercourse to prevent unwanted pregnancy. EC can help prevent pregnancies due to non-use, failure or misuse of

contraceptive, or situations of rape or coerced sex10,11. EC has a pregnancy prevention rate ranging from 56% to 95% if promptly and appropriately administered12–16. Suitably, EC was selected as one of 13 high impact, low-cost commodities by the UN Commission on Life-saving Commodities for Women and Children (UNCoLSC)17. EC use is highest among two groups of women: aged 20-24 years and unmarried sexually active18,19. EC is safe for over-the-counter sale and often available from a pharmacist or drug seller without a prescription20.

1. Remove “hormonal”. Currently there are two types of oral emergency contraption (EC) pills more widely used: one contains levonorgestrel (LNG) which is a hormone; but the other one contains ulipristal acetate (UPA), which is a selective progesterone receptor modulator (SPRM).

2. Emergency contraception refers to pills but also to the use of the IUD. Since the article seems to refer to EC pills I would make it explicit (talk about “emergency contraception pills”)

3. I would refer to “pills” in plural.

4. It is not clear to me what “appropriately administered” refers to. I would rather say “if promptly used”.

5. For reference, WHO’s factsheet on EC can provide further clarity: https://www.who.int/news-room/fact-sheets/detail/emergency-contraception

I hope this is useful. Thank you for this effort to make more visible important subgroups of population that are chronically underrepresented in global health measurements. For the EC community I think the point the paper makes is very valuable too. Congratulations!

Reviewer #3: Review Outcome

Title:

• Suggest the title to be modified

• Suggested title: Trends in Contraceptive Demand and Utilization Among Sexually Active Unmarried Women in Kenya: A Multilevel Regression Analysis

Abstract

• Authors have used the term demand and need interchangeably

• The first statement in the background sub-section is not clear***require modification

• If the primary aim of this paper is to analyze trends in contraceptive utilization and demand, the title should be modified accordingly, as suggested above

• Grammatical error: In the methods, authors stated as: ‘This study analyzed datasets*****’……datasets can’t be analyzed***revise the statements to make it clear and concise.

• Results seems a conclusion statement. Authors should incorporate regression results and 95% CI. They should also describe sample size included in the analysis.

• Conclusions should be drawn based on the aim of the study. No statement in the conclusion referred to trends in contraceptive demand and utilization, and associated consequences

• Finally, authors should avoid using abbreviations

• In general, abstract is not informative and require intense grammar revisions.

Introduction

• Authors have tried to synthesize contraceptive demand and use in the global, regional and study area context including the consequences of non-utilization.

• Authors should revise language***with some statements lacked clarity and coherence of ideas. For instance: ‘Even with this comparatively high performance, Kenya implemented policies to

reduce barriers to access family planning, such as policies enacted in 2013 to effectively eliminate family planning user fees as well as other public outpatient costs

• Authors should discuss the approach or methods followed including data sources in the methods section. For instance: ‘This analysis utilized data from the Performance Monitoring and Accountability (PMA) survey25. Managed by the Kenya Ministry of Health, PMA was a nationally representative survey of female respondents along with service delivery points (e.g., health facilities) to understand family planning usage, knowledge, and experience of women as well as service availability in the community. In addition, PMA incorporated a unique EC question: “Have you used emergency contraception at any time in the last 12 months?”. This question has a longer recall period than the traditional ‘current use’ EC indicator, which underestimates the scale of EC usage’

• The last statement of the introduction (aim statement) is not clear. Authors should clearly specify the aim of the study and aim should be consistent to the one stated in the abstract section and the title of the paper.

• Moreover, authors should conduct language and grammar revisions (for instance: check the 2nd paragraph)

• The conceptual framework should be presented as part of methods (Figure 1)

Methods

• Methods lacks clarity. Authors should clearly describe the method they have used to answer the aim or research question. Suggested sub-sections can be:

o Study setting

o Data source and measurements

o Study variables

o Statistical analysis

• Authors should avoid use of some jargon or non-technical words. For instance: ‘The female questionnaire includes marital status, recency of sexual activity, *****’

• Tables should be self-explanatory, with proper footnote and need to be properly cited inside the document.

• Authors shouldn’t include variable definition as supplementary file.

• The analysis methods used is not clear. Authors should clarify, why and how they have used the multi-level regression model.

Results

• Table/figure titles should be self-explanatory and tables need to be properly cited within the document. Avoid citing like, (see Table 2; see Figure 2); rather (Table 2; Figure 2). For each table, the source of data should be indicated.

• What is the need to include ‘change’ in Table 3? Try to use the proper color whenever presenting figures.

• In Table 3, for each year, include both number and %. All abbreviations should be described as foot note. What do, other traditional included?

• This section is not clear and difficult to follow-up. Authors should organize and briefly present the findings based on objectives of the analysis.

• Table 4.1-4.4 is not clear. What is the aim of this analysis?

• In general, authors should answer previously formulated research questions.

Discussion:

• Authors should properly discuss the theoretical and practical implications of the analysis

• They should adequately discuss the findings in the context of other settings

• Strength and limitations of the analysis need to be explained

• The conclusion should be based on the findings of the analysis

Final Decision: Reject

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Lee Fairlie

Reviewer #2: Yes: Cristina Puig Borràs

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

18 Feb 2022

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This is an interesting and generally well written manuscript describing contraceptive demand and utilization in unmarried sexually active women in Kenya. I believe that the manuscript would benefit from some revisions before publication.

Suggested revisions below:

Abstract

- Although the text is generally clear, there are parts where it is very biostats language heavy and could benefit from edits to improve clarity regarding the clinical implications for example "trending in different directions", "survey enumeration" in the background section. Whilst I appreciate that these terms are correct, it may be clearer to say "differences particularly in women not sexually active 1-12 months", or "reporting no recent sexual activity in preceding month".

Background

- It would be helpful to get additional context regarding the Kenyan population. From the manuscript it appears that most women are married, but I'm not certain if this is true generally in Kenya or whether this is true for women who contributed to the data. This would be helpful to understand. This could potentially provide substantial bias as there may be stigma associated with sexual activity outside of marriage and mainly unmarried women may be less likely to contribute. Similarly, there is an implication that unmarried women do not have sexual intercourse which needs to be clarified.

Authors. This is a good point. We added more background details on the percentage of married women in Kenya (60%) to the third Introduction paragraph and clarified the rates of sexual activity in the fourth paragraph, where 6.8% of unmarried women were sexually active (in last month) in Kenya. Our sample has about 9% of unmarried women who are sexually active (in last month) and another 14% sexually active (1-12 months).

- Do women in Kenya ever use the copper IUD as EC? Only the contraceptive pill EC is discussed.

Authors: Thank you for the comment. We have added a comment on copper IUD to the Introduction and clarified the focus on emergency contraceptive pills.

- In the paragraph: "With a population of ......" in the 4th line it is difficult to understand the breakdown of the 39% of women using contraception (CPR) please edit for clarity

Authors: Thank you. We edited the paragraph for added clarity.

- In the last paragraph "For family planning indicators....." a number of sentences would fit better into methods-suggest removing these and focussing on how this study addresses outstanding issues.

Authors: We appreciate the suggestion. We moved the discussion on temporal misalignment to the Methods section.

- The last paragraph of the background: additional text here can also be moved to methods

Authors: Thank you. We removed one of the sentences for clarity.

Methods

- Generally clear

Results

- The results are interesting and well represented

- It is particularly interesting that women who are in the unmarried sexually active group are the highest users of EC. In the discussion this is posed as a negative, but perhaps the authors should also emphasise that this is a real positive in terms of women perhaps choosing not to permanently use contraception, but able to access EC when needed. Obviously a longer term choice may be preferable but it is important that women are able to access this option

Authors: Thank you for the comment. This was not our intention – we tried to remain neutral by using words such as “shifting”. We edited the last sentence in that paragraph to illustrate that the women are ‘leveraging’ access to ECP to reduce unwanted pregnancies.

- Is there a correlation between the sexually active women with unmet contraception need and pregnancy in this study? If women are not becoming pregnant unplanned, then perhaps this is not as much of a concern.

Authors: This is an interesting research question and something the authors considered, but unfortunately the PMA questionnaire is not well designed to answer and beyond the scope of this study.

- Also interesting the increase in implant use in this cohort

- The tables are good and clear

- The figures are difficult to understand as no figure headings and legends are included for figure 2 onwards

Authors: Apologies, this was an upload error and all tables / figures have now been properly titled and cited.

Discussion

- Well written

- Please see previous comments regarding clarifying this issues and findings, for example the fact that unmarried sexually active women 1-12 months access EC is a plus rather than concerning in my opinion, hopefully averting unplanned pregnancies.

Authors: See our comment above.

- I'm not sure that the 3rd paragraph is substantiated. "The female subpopulations....". The paper doesn't report pregnancy in the group of unmarried sexually active, so it is possible that using EC is acceptable particularly if the encounter was unplanned/unexpected. It is a positive that women have sufficient agency to choose when they wish to access contraception. Obviously in this cohort if women do have unplanned pregnancies then this is a concern.

Authors: Very good point. We added additional commentary on this issue – a decrease in unmet need for ‘current’ mCPR, may indicate women have more agency to access ECP. It is certainly an area for future research.

- In the paragraph " To reach universal...." whilst it is appreciated that unmarried women who are sexually active 1-12 months may be of concern, I wonder how relevant this specific focus on marriage is, to many women in different countries and regions. There is a danger of focussing separately on this group as this could be stigmatising and essentially messages regarding contraceptive access should reach all women. I realise that is what the authors are saying and perhaps more context regarding marriage in Kenya would be helpful. It seems unlikely that there are so few unmarried sexually active women relative to married women and in itself it seems to perpetuate the message that unmarried women should not be sexually active, which is problematic and likely compounds the issue. Perhaps these points can be discussed.

Authors: Thank you for the comment. In short, 13% of the women in the study were unmarried and sexually active 1-12 months prior to survey, so it’s a sizable subgroup. Your question focuses on the ‘how’ to address contraceptive disparities in these subgroups (e.g., without using messaging that further stigmatizes them). The authors felt this type of discussion was complex and nuanced and beyond the scope this section. But, identifying the problem is the first step to a solution. In addition, we added this sentence to the discussion to help address this issue: “Moreover, implementation research is needed on how to design family planning programs for these marginalized female subgroups, who already experience stigma within the health delivery system, without limiting the reproductive health improvements for all women.”

__________________________________________________

Reviewer #2: Thank you for the opportunity to review this article. I would only suggest a few changes in this paragraph (2nd para. of Introduction):

(…) The emergency contraceptive pill (EC) is an oral, hormonal contraceptive pill for women to use as soon as possible (up to 5 days) after sexual intercourse to prevent unwanted pregnancy. EC can help prevent pregnancies due to non-use, failure or misuse of

contraceptive, or situations of rape or coerced sex10,11. EC has a pregnancy prevention rate ranging from 56% to 95% if promptly and appropriately administered12–16. Suitably, EC was selected as one of 13 high impact, low-cost commodities by the UN Commission on Life-saving Commodities for Women and Children (UNCoLSC)17. EC use is highest among two groups of women: aged 20-24 years and unmarried sexually active18,19. EC is safe for over-the-counter sale and often available from a pharmacist or drug seller without a prescription20.

1. Remove “hormonal”. Currently there are two types of oral emergency contraption (EC) pills more widely used: one contains levonorgestrel (LNG) which is a hormone; but the other one contains ulipristal acetate (UPA), which is a selective progesterone receptor modulator (SPRM).

Authors: Thank you for the feedback. This has been edited.

2. Emergency contraception refers to pills but also to the use of the IUD. Since the article seems to refer to EC pills I would make it explicit (talk about “emergency contraception pills”)

Authors: Thank you for the comment. This has been edited.

3. I would refer to “pills” in plural.

Authors: Thank you for the feedback. This has been updated.

4. It is not clear to me what “appropriately administered” refers to. I would rather say “if promptly used”.

Authors: Thank you for the feedback. This has been edited.

5. For reference, WHO’s factsheet on EC can provide further clarity: https://www.who.int/news-room/fact-sheets/detail/emergency-contraception

I hope this is useful. Thank you for this effort to make more visible important subgroups of population that are chronically underrepresented in global health measurements. For the EC community I think the point the paper makes is very valuable too. Congratulations!

__________________________________________________

Reviewer #3: Review Outcome

Title:

• Suggest the title to be modified

• Suggested title: Trends in Contraceptive Demand and Utilization Among Sexually Active Unmarried Women in Kenya: A Multilevel Regression Analysis

Authors: Thank you for the feedback. The title has been edited.

Abstract

• Authors have used the term demand and need interchangeably

Authors: Thank you for pointing this out. We reviewed the entire document and updated terminology to ensure clarity between these two terms.

• The first statement in the background sub-section is not clear***require modification

Authors: Thank you for the feedback. This has been modified.

• If the primary aim of this paper is to analyze trends in contraceptive utilization and demand, the title should be modified accordingly, as suggested above

Authors: Thank you. The title has been modified.

• Grammatical error: In the methods, authors stated as: ‘This study analyzed datasets*****’……datasets can’t be analyzed***revise the statements to make it clear and concise.

Authors: Thank you for identifying this oversight. It has been edited.

• Results seems a conclusion statement. Authors should incorporate regression results and 95% CI. They should also describe sample size included in the analysis.

Authors: Thank you for raising this important point. We added the sample size to the Abstract as well as the regression results. The Abstract length – particularly the Results sub-section - was expanded to accommodate these important points. Initially, we tried to keep the Abstract very short, but we unfortunately it appears we sacrificed clarity and adequate understanding by the reader. Thank you for pointing this out.

• Conclusions should be drawn based on the aim of the study. No statement in the conclusion referred to trends in contraceptive demand and utilization, and associated consequences

Authors: Thank you for this feedback. We restructured the concluding paragraph. The first two sentences refer to the aims of the study. The last two sentences address the implications and forward-looking research needs.

• Finally, authors should avoid using abbreviations

Authors: Yes, we were conscious of this and tried to limit abbreviations as much as possible. However, we included two common abbreviations mCPR and ECP, which were defined and repeated multiple times in the Abstract, to reduce the word count. This is consistent with other PLOS One manuscripts, such as Shiferaw 2017, which is referenced in this manuscript.

• In general, abstract is not informative and require intense grammar revisions.

Authors: Thank you again for this feedback. We conducted a comprehensive review of the Abstract and made large-scale edits to increase clarity and information transfer.

Introduction

• Authors have tried to synthesize contraceptive demand and use in the global, regional and study area context including the consequences of non-utilization.

• Authors should revise language***with some statements lacked clarity and coherence of ideas. For instance: ‘Even with this comparatively high performance, Kenya implemented policies to

reduce barriers to access family planning, such as policies enacted in 2013 to effectively eliminate family planning user fees as well as other public outpatient costs

Authors: Thank you for the feedback. This has been edited.

• Authors should discuss the approach or methods followed including data sources in the methods section. For instance: ‘This analysis utilized data from the Performance Monitoring and Accountability (PMA) survey25. Managed by the Kenya Ministry of Health, PMA was a nationally representative survey of female respondents along with service delivery points (e.g., health facilities) to understand family planning usage, knowledge, and experience of women as well as service availability in the community. In addition, PMA incorporated a unique EC question: “Have you used emergency contraception at any time in the last 12 months?”. This question has a longer recall period than the traditional ‘current use’ EC indicator, which underestimates the scale of EC usage’

Authors: Thank you for the feedback. We shortened this section, but we need to include some mention of PMA and the emergency contraception question in the Introduction, because it provides background for the Study Aims in the last paragraph of the Introduction.

• The last statement of the introduction (aim statement) is not clear. Authors should clearly specify the aim of the study and aim should be consistent to the one stated in the abstract section and the title of the paper.

Authors: Thank you. We updated and simplified the aims statement.

• Moreover, authors should conduct language and grammar revisions (for instance: check the 2nd paragraph)

Authors: Thank you for the recommendation. We have made significant content and grammatic revisions to the second paragraph of the Introduction on ECP. We added more

• The conceptual framework should be presented as part of methods (Figure 1)

Authors: Thank you for the feedback. The conceptual framework has been moved to the Methods section.

Methods

• Methods lacks clarity. Authors should clearly describe the method they have used to answer the aim or research question. Suggested sub-sections can be:

o Study setting

o Data source and measurements

o Study variables

o Statistical analysis

Authors: Thank you for the feedback. The sub-headings have been edited.

• Authors should avoid use of some jargon or non-technical words. For instance: ‘The female questionnaire includes marital status, recency of sexual activity, *****’

Authors: Thank you for the comment. This sentence has been simplified.

• Tables should be self-explanatory, with proper footnote and need to be properly cited inside the document.

Authors: Thank you for the comment. This sentence has been updated.

• Authors shouldn’t include variable definition as supplementary file.

Authors: Thank you for the comment, but we are unclear as to the corrective action. We are happy to remove the variable definition file altogether or move it into the main body of the manuscript. If PLOS One can provide guidance on standard format, that would be helpful.

• The analysis methods used is not clear. Authors should clarify, why and how they have used the multi-level regression model.

Authors: Thank you for the comment. We have added additional description on why and how the multi-level regression models were used.

Results

• Table/figure titles should be self-explanatory and tables need to be properly cited within the document. Avoid citing like, (see Table 2; see Figure 2); rather (Table 2; Figure 2). For each table, the source of data should be indicated.

Authors: Apologies, this was an upload error and all tables / figures have now been properly titled and cited.

• What is the need to include ‘change’ in Table 3? Try to use the proper color whenever presenting figures .

Authors: Thank you for the comment. This is simply a visual aid for the reader. The table includes about 200 figures, so the authors felt a quick reference for the reader was advantageous.

• In Table 3, for each year, include both number and %. All abbreviations should be described as foot note. What do, other traditional included?

Authors: Thank you for the feedback. We added explanations of abbreviations and definitions including for ‘other traditional’ methods. We added the total number of women for each column, which is consistent with standard presentation of method mix in other sources, such as DHS.

• This section is not clear and difficult to follow-up. Authors should organize and briefly present the findings based on objectives of the analysis.

Authors: Thank you for the comment. We clarified the Aims Statement and utilized similar subheaders for the Results section to make the connection more clear.

• Table 4.1-4.4 is not clear. What is the aim of this analysis?

Authors: Thank you for the feedback. We updated the Table titles, narrative and clarified the Aims Statement to more directly link with these Tables and Results section. Hopefully, these changes make the link and utility of the tables more clear.

• In general, authors should answer previously formulated research questions.

Authors: The research questions have been more clearly stated in the Aims Statement and matches the flow of the Results section. The three areas of research are:

1. Level and trend of contraceptive use and unmet need indicators (mCPR, unmet need, demand, etc) by female subgroup.

2. Level and trend of recent ECP use by female subgroup

3. Influence of contextual factors (e.g., demographic, socioeconomic, female subgroup) on the outcome measures - over time

Discussion:

• Authors should properly discuss the theoretical and practical implications of the analysis

Authors: Thank you for mentioning this. It was a catalyst for a deeper review of the Discussion and make it more clear to the reader. We added several sentences (2nd, 3rd, 4th and 5th paragraphs) to more clearly identify the theoretical and practical implications of this analysis.

• They should adequately discuss the findings in the context of other settings

Authors: Thank you – this was useful feedback. Data is limited internationally, so it is hard to make concrete assessments in other settings. We decided to add a paragraph to the Discussion (6th paragraph) to discuss this important issue:

“Beyond Kenya, analysis of unmarried women with less recent sexual intercourse is limited or nonexistent; therefore, it is unclear whether these differences exhibited between female subgroups in Kenya are common in SSA or an aberration. Conducting analyses like this study for other countries or regions is an important topic for future research. Moreover, implementation research is needed on how to design family planning programs for these marginalized female subgroups, who already experience stigma within the health delivery system, without limiting the reproductive health improvements for all women. Building a research agenda to better understand unmarried sexually active women can help counteract the systemic bias against these women that pervades all levels of the global healthcare system.”

• Strength and limitations of the analysis need to be explained

Authors: Thank you for this comment. The 7th paragraph includes the limitations and we added more context about the strength of the analysis here (that wasn’t covered elsewhere in the manuscript, such as the Methods and earlier in Discussion).

• The conclusion should be based on the findings of the analysis

Authors: Thank you for this feedback. We shortened the concluding paragraph. The first two sentences summarize the findings in layperson’s terms, while the last sentence is a more forward looking perspective.

Submitted filename: PMA Paper - Response to Reviewers 2022-02-15.pdf

Click here for additional data file.

20 Apr 2022

1 Jun 2022

Dear Dr. Nemser:

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Abstract & background

1. In the background section of the manuscript, the authors write: “This study aims to explore the level and trend of contraceptive use and unmet family planning needs among female subpopulations in Kenya: married or in union (i.e., living together); unmarried and sexually active within the past 30 days prior to survey (labeled as UA-30days); and unmarried and sexually active between 1-12 months prior to the survey (labeled as UA-12months)”. In the abstract, the authors write: “Unmarried women who report less recent sexual intercourse (>30 days from survey enumeration) are largely excluded from global health monitoring and evaluation efforts”.

A few points to note here:

*In the statement of the objective picked from the background section, the authors refer to ‘… the level and trend’ of contraceptive us – but the reference to ‘level’ is not mentioned in the statement of the objective picked from the abstract. The statement of the overall objective should be consistently presented across the manuscript

>>AUTHORS: Thank you very much for the comment. We edited the background section to incorporate your suggestion and maintain consistency. We removed mention of evaluating trend.

*The statement of the objective picked from the background section uses the verb ‘explore’. I suggest that this verb be revised to a more quantitative term that explains what the authors actually did.

>>AUTHORS: Thank you for the suggestion. We replaced ‘explore’ with the term ‘evaluate’ in the background section.

*In the statement of the objective picked from the background section, the authors refer to ‘female subpopulations in Kenya’ with a list of these sub-populations listed to include: married or in union, unmarried and sexually active within the past 30 days prior to survey and unmarried and sexually active between 1-12 months prior to the survey. Based on these subgroups, I wonder if the use of only ‘unmarried women’ in the title is appropriate.

>>AUTHORS: Thank you for this feedback. As mentioned in the Introduction and Discussion, global health monitoring and evaluation efforts focus specifically on married / in union women or unmarried women with recent sexual intercourse (less than 30 days). This data processing and reporting bias has existed for decades. The data presented in this study on these two subgroups is regularly described at length in other global health reports. However, the unique feature of this analysis is the evaluation of unmarried women with less recent sexual intercourse (between 1-12 months prior to survey). The analysis of this underreported subgroup of women in this study provides the substantive contribution to public knowledge. Therefore, the title focuses on “unmarried women” to reflect the deeper investigation of these women and uniqueness and contribution of this study.

*In the abstract, the authors focus on ‘unmarried women who report less recent sexual intercourse’ but the statement of the objective in the background section refers to three subpopulations. Can the authors clarify on why this is the case?

>>AUTHORS: Thank you for the comment. As mentioned above, the global health monitoring and evaluation community routinely (and almost exclusively) reports on married / in union women or unmarried women with recent sexual intercourse (less than 30 days). This practice is rooted in commonalities between these two sets of women relative to current mCPR indicator. As mentioned in the Methods section, the addition of ‘mCPR at last sex’ and ‘recent use of emergency contraceptive’ data collection by PMA, allows for a more intensive investigate of actual contraceptive use (unbiased by ‘current use of mCPR’) for unmarried women with less recent sexual activity (between 1-12 months prior to survey). This female subgroup makes up about 13% of the women enumerated, but they are systematically eliminated from reporting in the global health literature and thus excluded from influence on decision-making in various context. The unique aspect of this study is the focus on ‘unmarried women who report less recent sexual intercourse’ – otherwise the study would maintain the status quo biases against these women and would not substantively contribute new content to the collective research base.

Note, we reordered the subgroups in the Methods section to focus on unmarried women with less recent sexual intercourse.

2. If the aim of the study was to assess “trends in … contraceptive utilization and unmet need”, then, I would expect to see these trends presented in the abstract. The authors indicate that they used data collected in multiple surveys over the period 2014 to 2019; so, I expected to see some trend analyses presented, and I would be interested to know if there was a significant increasing or decreasing trend or whether there was no change over the years in contraceptive utilization and unmet need. This is not provided. Instead, the authors focus on reporting on 2019 indicators which makes it difficult to tell if the analysis was to assess trends or just contraceptive utilization in 2019. Also, the reporting on the trends in unmet need comes at the extreme end of the results sub-section, and presented in a more generic format.

>>AUTHORS: Thank you for the comment; however, based on the content of the original submission, the previous group of reviewers recommended that the term “Trends” be added to the title. See below. “Trend” was not part of the original title, because analysis of trend was not the primary objective of the study. Given your feedback and our intended focus on level and contextual factors, the Authors decided to remove “Trend” from the title and description of objectives.

Original Title: Contraceptive Demand and Utilization by Unmarried, Sexually Active Women in Kenya: A Multilevel Regression Analysis

From January 3, 2022:

Reviewer #3: Review Outcome

• Suggest the title to be modified

• Suggested title: Trends in Contraceptive Demand and Utilization Among Sexually Active Unmarried Women in Kenya: A Multilevel Regression Analysis

3. I realize that the issue of emergency contraception is singled out in the abstract. Why was this singled out and not considered as one of the contraceptive methods, in the same way the other short-term methods were handled.

>>AUTHORS: Thank you for the feedback. As mentioned in the Abstract and further discussed in the Introduction and Methods, the PMA questionnaire contains a unique question - “Have you used emergency contraception at any time in the last 12 months?”. This is one of the only datasets in the world that includes this question. Emergency contraceptive are the primary post-intercourse contraceptive method; therefore, the typical “current contraceptive use” questions are less applicable and valid. Therefore, PMA added a more appropriate indicator for actual use by women. EC is singled out in the analysis, because PMA cleverly designed its questionnaire to provide a more valid estimate of genuine use by women.

Results

1. Table 3 presents the contraceptive method mix by female Subgroup for the three surveys considered in the analysis (2014, 2017, 2019). Did the authors try to assess if the observed changes, as reported, depicted significant increases or decreases in the outcome over time? This question also applies to Figure 4 on unmet need.

>>AUTHORS: Thank you for the comment. The Authors did consider this analysis, but several overlapping issues led us to simply display the cumulative proportions and point estimates. First, contraceptive method mix is typically reported as cumulative proportions and several contraceptive methods had very small sample sizes; therefore, statistical inference would be limited. Secondly, the primary focus of the study was the level and contextual factors (see discussion above). The Methods Mix and Figures are unadjusted for these contextual factors. The Authors provided the Methods Mix table (and other descriptive indicators / figures) over time to provide background context to readers for better conceptual understand of the logistic models and experience of women in Kenya. [Note, if recommended, we can provide this statistical analysis, but it would require a lot of caveats and add to the length of the Methods section and overall study.]

12 Jun 2022

Contextual Factors associated with Contraceptive Utilization and Unmet Need Among Sexually Active Unmarried Women in Kenya: A Multilevel Regression Analysis

PONE-D-21-35162R2

Dear Dr. Nemser,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Joseph KB Matovu, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

21 Jun 2022

PONE-D-21-35162R2

Contextual Factors associated with Contraceptive Utilization and Unmet Need Among Sexually Active Unmarried Women in Kenya: A Multilevel Regression Analysis

Dear Dr. Nemser:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Joseph KB Matovu

Academic Editor

PLOS ONE