Demand Side Predictors of the Demand and Utilization of Maternal and Family Planning Services of the Philippines’ Conditional Cash Transfer Program: A Quantitative Approach

Abstract

The study explores demand for maternal and family planning services among beneficiaries of the Pantawid Pamilyang Pilipino Program, a conditional cash transfer program in the Philippines. We analyzed 390 households across six provinces in the Philippines through multi-stage sampling. Three models were created: (1) pregnancy during the program, (2) facility-based delivery, and (3) non-utilization of antenatal care visits. The findings show that: (1) For every peso increase in received amounts from the program, the odds of getting pregnant decreased by 0.31%, a negative but weak association, while having more children, living in a high mortality region, having a teenage child, living far (self-perceived) from the health center, marrying at a younger age, and not completing high school also increased the odds of being pregnant. (2) Non-delivery from a health facility was likely for those who lived far from a facility and in a high mortality region, and (3) having a teenager at home increased the likelihood of seeking antenatal visits, while those who have been married longer slacked behind in terms of antenatal visits. The study points to directions for policy and research on access and use of maternal and reproductive health services, and the role of adolescent-aged children.

1. Introduction

The Pantawid Pamilyang Pilipino Program, popularly known as 4Ps, is a flagship anti-poverty program that is both a social welfare and a social development intervention. As a social welfare program, it is a form of conditional cash transfer (CCT) program, providing cash grants to targeted families in extreme poverty. As a social development program, it hopes to break the cycle of intergenerational poverty by ensuring investments in education (PHP 5000 or USD 105 per year) and health (PHP 500 (USD 11) per month for a total of PHP 3000 or USD 63 per year), per household, depending on household composition. The grant is conditional to pregnant women availing themselves of antenatal care and regular health checks for women and children below 5 years old, deworming of school-aged children 6–14 years old, school enrolment of children, and attendance in family development sessions. As of 2016, the Department of Social Welfare and Development (DSWD) reported having 4,393,114 active members, of which 572,809 were from indigenous peoples (IP) households and about 219,149 households have members with disabilities. For the same period, a total of PHP 27.15 billion (USD 572 million) cash grants have been availed of, of which PHP 13.2 billion (USD 278 million) was for education and PHP 13.9 billion (USD 293 million) was for health.

The DSWD, during a 2016 Research Colloquium, showed high compliance rates for Pantawid families. Around 94% attendance was reported for family development sessions and 96% coverage for health visits of pregnant women and children aged 0–5. Earlier research [1] showed that Pantawid families had more members, roughly 6 members, compared to 4 members for non-Pantawid families. The family size for 23% of Pantawid families was at least 8 members. The Pantawid Pamilya 3rd Wave Impact Evaluation [2] showed that women beneficiaries did try the modern FP method at a greater rate than women who were not part of the program. The Contraceptive Prevalence Rate was reported at 52% for Pantawid families compared to 47% for non-Pantawid counterparts. The higher incidence of trial use, however, did not translate into sustained use of modern FP methods. There was a high awareness of modern family planning methods among both Pantawid families and non-Pantawid families, though Pantawid families are aware of more contraceptive options. Compared to an earlier national demographic and health survey (DHS)—sourced average of 84%, an impact evaluation [2] found that 76% of Pantawid mothers made at least four antenatal visits.

The demand for family planning and maternal health services is viewed as part of a decision-making nexus that values health intrinsically. Human capital theory posits that maximization of health is the ultimate objective of seeking health care [3]. Medical care services are inputs in the overall production of health. Consumption of medical care and other goods is determined by the amount of resources available or income, and the price and quantities of these goods. The individual maximizes welfare by choosing among a combination of goods and services that fit within the budget constraint.

Decision-making on sexual and reproductive health is influenced by education, by operating at the cognitive level of imparting information and/or through socialization that happens in these institutions. Attributing to Boongaerts, Darney et al. [4] identified the pathway through proximate determinants like age at marriage, postnatal fecundity (via breastfeeding and postnatal abstinence practices), and contraceptive use. A gender perspective would highlight the access of women to prenatal care and the extent to which their heavy domestic household and farming activities are likely to affect their health. This perspective emphasizes the social and gender roles or structures that influence women’s fertility decisions and health-seeking behavior over reproductive health [5]. This study on demand for FP and MCH services by 4Ps women works around the basic demand function as influenced by a vector of individual and demographic and socio-economic aspects affecting the demand for these services by the women and their households. While not expressed among 4Ps conditions, the quality of maternal health care services forms part of the completeness of services. FP as a reproductive health issue underpins maternal care. Changes in reproductive care attitudes and practices upon participation in the 4Ps program, even if FP is not explicitly part of program conditionalities, are viewed as a consequence of life skills imparted through the family development sessions (FDS; [6]) and health center visits with a trained health worker. The information gleaned from health service encounters is expected to improve demand for reproductive health services.

This study aimed to identify demand side factors in the utilization of maternal health care and family planning services. We hypothesized that the demand side factors that may potentially be predictive for maternal health care and family planning services utilization are the following: individual characteristics of the mother and her spouse or partner such as age, education, employment, and age married, family characteristics such as years of marriage, number of children, presence of teenage child, household monthly income, locality if urban or rural, and if with high maternal mortality rate, distance from health center, transportation costs, and monthly amount received from the CCT program.

2. Methods

2.1 Study design and setting

This study was conducted in the Philippines from 2016 to 2017 and is a national multi-level mixed methods study consisting of household surveys, interviews, and focus group discussions. A regionally representative sample was obtained with a multi-stage purposive sampling design. The three major island groups, Luzon, Visayas, and Mindanao, were the first level of stratification, from which two provinces from each island group were selected: one urban and one rural. The following provinces were selected: Pampanga and Zambales for Luzon, Cebu and Western Samar for Visayas, and Compostela Valley (now Davao de Oro) and South Cotabato for Mindanao. These were selected for having one or more of the following characteristics: vulnerability to disasters, the presence of indigenous people’s communities, and both high and low maternal mortality ratios. Purposive sampling was conducted for the household-level surveys, which sought Pantawid family members in randomly selected villages in the selected cities/municipalities. We defined a household as inclusive of those living under the same roof and sharing the same meals.

We surveyed a total of 390 households. These households received grants from the 4Ps program. Households were interviewed using an interviewer-administered survey tool covering demographic and household composition, socio-economic condition, and health service utilization, including recent experience of childbirth for those with children 5 years old and below. The main respondent was the mother grantee. The questionnaires were encoded daily and interviewers were given one opportunity to go back to the families in case some responses were unclear.

2.2 Qualitative data

Summative qualitative content analysis was conducted with the existing study variables as the initial themes. A coding and categorization tool was developed on MS Excel. Two cycles of coding were conducted to analyze the notes and transcripts, with the first using structural coding and the second using pattern coding. Analytic memos were developed throughout the coding process and then subjected to coding. Two independent coders conducted the coding and thematizing process. Coding sorts were developed and organized into themes.

2.3 Statistical analyses

Several statistical tests were used to analyze the data collected. Univariate analysis was used to describe data and to show frequencies. Bivariate analysis using the t-test and chi-square test was used to show the association between the exposure and the outcome variables. Multivariate analysis through logistic regressions was used to measure the magnitudes of the association. Only p-values of 0.05 or lower were considered significant. The STATA 12 statistics software was used for data analysis.

2.4 Dependent outcomes

Regression analyses examined several layers of women’s decision-making on reproductive health care demand: (1) the factors associated with being pregnant as a 4Ps beneficiary; (2) the demand for facility-based delivery; (3) the demand for antenatal; and (4) the demand for modern contraceptive methods. The first three, the likelihood of being pregnant as a 4Ps beneficiary, facility-based deliveries, and demand for antenatal care, came out to be statistically significant and could explain 49, 8, and 11% of the variability, respectively. The discussion will therefore focus on a reduced form of the 3 demand models.

3. Results and discussion

3.1 Descriptive background of sites, women, and household respondents

The household survey covered 390 Pantawid Pamilya households. Maternal Mortality status of the province was determined in consultation with the DSWD Planning, Monitoring, and Evaluation Division, which identified the study provinces.

3.1.1 Locality and area characteristics

Respondents in Low MMR (Table 1) areas are on average a year older, have on average been married for one fewer year, and have on average one fewer household member than High MMR areas. There are fewer respondents in Low MMR areas with a child below 5 years of age than in High MMR areas, with 46.74% in Low MMR and 53.26% in High MMR. Fewer respondents in low MMR areas also have children between 6 and 18 years of age, with 45.17% in Low MMR areas and 54.83% in High MMR areas. There are far more Service workers (33 in Low MMR, 3 in High MMR) and Laborers (28 in Low MMR, 12 in High MMR) in Low MMR areas. Average income is higher in Rural areas, with an average monthly income of greater than PHP 4000 (USD 84) at 85.13% for Low MMR areas and 67.29% for High MMR areas.

	High MMR; (n = 195)	Low MMR; (n = 195)
Mean age of women respondents	35.65 ± 6.34	36.58 ± 8.04
Year married or cohabiting	14 (2 to 32)	13 (1 to 48)
Mean age when first married
Spouse	24.70 ± 6.02	25.50 ± 6.35
Respondent	21.14 ± 4.98	22.01 ± 5.19
Household size	7 (3 to 12)	6 (2 to 11)
Children per household (18 y/o and below)	4 (1 to 9)	4 (0 to 10)
No. and % of women with children between 6 and 18 years of age	139 (53.26)	122 (46.74)
Educational attainment
No formal education	266 (27.94)	266 (31.22)
Elementary completed	168 (17.65)	24 (2.82)
Elementary not completed	282 (29.62)	316 (37.09)
High school completed	80 (8.4)	64 (7.51)
High school not completed	123 (12.92)	147 (17.21)
Vocational completed	2 (0.21)	4 (0.47)
College completed	10 (1.05)	5 (0.59)
College not completed	21 (2.21)	26 (3.05)
Postgraduate completed	0	0
Occupation of the women respondents (n = 106)
Technicians and associate professionals	1 (1.2)	0
Clerks	2 (2.41)	1 (1.37)
Service workers and shop and market sales owners	3 (3.61)	33 (45.21)
Farmers, forestry workers, and fishermen	4 (4.82)	2 (2.74)
Traders and related workers	0	2 (2.74)
Plant and machine operators and assemblers	0	0
Laborers and unskilled workers	12 (14.46)	28 (38.36)
Special occupations	8 (9.64)	7 (9.59)
Income of women from occupation, earnings for 2015 (mean overall and SD)
< PHP 1000	3 (0 to 4)	4 (0 to 4)
PHP 1000–2000	60 (30.77)	62 (31.79)
PHP 2001–3000	0	0
PHP 3001–4000	0	0
> PHP 4000	3 (1.54)	6 (3.08)

Table 1.

Profile of respondents by low-high MMR status of province.

In terms of the distribution of the 390 women respondents by urban-rural setting (Table 2), there are more respondents in urban areas who have children below 5 years of age, as well as between 6 and 18 years of age. Of the respondents in both settings, the majority were not able to reach or complete High School, with more respondents from Urban areas reaching but failing to complete High School (16.67% Urban, 12.95% Rural) but more respondents from Rural areas completing High School (4.6% Urban, 11.99% Rural). About 46.74% of women in Urban areas and 53.26% of women in Rural areas reported having children below 5 years old, while 45.17% of women in Urban areas and 54.83% of women in Rural areas reported having school-aged (6–18 years of age) children. The average household size reported is 6.

	Urban; (n = 215)	Rural; (n = 175)
Mean age of women respondents	36.08 ± 6.88	36.16 ± 7.69
Year married or cohabiting	13 (1 to 35)	13 (1 to 48)
Mean age when first married
Spouse	25.98 ± 6.28	24.39 ± 6.04
Respondent	21.97 ± 5.25	21.26 ± 4.96
Household size	6 (2 to 11)	6 (2 to 12)
Children per household (18 y/o and below)	4 (0 to 10)	4 (0 to 9)
No. and % of women with children between 6 and 18 years of age	154 (53.10)	136 (46.90)
Educational attainment
No formal education	266 (27.2)	266 (32.2)
Elementary completed	38 (3.89)	154 (18.64)
Elementary not completed	438 (44.79)	160 (19.37)
High school completed	45 (4.6)	99 (11.99)
High school not completed	163 (16.67)	107 (12.95)
Vocational completed	6 (0.61)	0
College completed	1 (0.1)	14 (1.69)
College not completed	21 (2.15)	26 (3.15)
Postgraduate completed	0	0

Table 2.

Profile of respondents by urban-rural setting.

A glimpse of the provincial profile of the respondents shows the following: The richest province Cebu has the most reporting 10 years of marriage while Western Samar (a rural province) respondents reported being married longest at 16 years. Each household, including the respondent and spouse, had an average size of 6 persons. All households had on average one child below 5 years of age. On the overall, Pantawid respondents in the study sites, compared to the National Demographic Health Survey (NDHS) [7] results, can be considered older, married longer, have much less education, have bigger household size having on average two more children, and with less gainful employment.

3.1.2 Spouse/partner’s characteristics

Spouses of respondents are much older in Low MMR areas (36.5 years Low MMR, 31 years High), and were also at least a year older at marriage. Like the respondents, the majority of the spouses did not reach or complete high school, but a good number of spouses completed high school in low MMR areas (32% Low MMR, 23% High MMR). Also similar to the respondents, the largest groups of reported occupation were service workers and unskilled laborers, with almost half of spouses in both Low and High MMR areas reporting to be Laborers.

3.1.3 Family characteristics

Less than half of the households reported having 2 adult earners (41%) but 65.64% still report having an income greater than PHP 4000 (USD 84) per month. Majority report to having monthly food expenditures of less than PHP 2000 (USD 421), and nearly all (99.49%) report to having expenses for cigarettes, alcohol, and games to be less than PHP 1000 (USD 21) per month. A good number (70.77%) report having a television, while only 11.79% report to having a refrigerator. Most use either a wood or coal stove, at 87.69%. The majority source their water from faucets, with water from a spring being a distant second, at 17.95%. More than half, at 59%, have had home renovations since becoming a member of the 4Ps program. The majority, at 82.31%, report having roofs of iron sheets.

3.1.4 Health service characteristics

Health facility staff believe that poor health-seeking behavior and differences in culture inhibit utilization for some beneficiaries. This access issue is limited to those who are far from the health facility, which does not constitute the majority. Health facilities compensate by conducting outreach/missions.

Respondents living in larger villages (barangays) consider that travel to health facilities is costly. This is contrasted in smaller barangays when respondents consider travel costs as affordable. When household survey respondents get sick, they would prioritize going to public health centers (79.8%). It is however notable that despite this high percentage, the same set of respondents would have more regular visits to private clinics. There is also a strong preference to bring their ailing family members to private clinics and hospitals.

About 97.4% of household survey respondents are confident that they will get health care from the health facility that they visited. Only 7.3% of household survey respondents feel that health facilities are far from where they live. For those who take transportation to reach health facilities, the majority prefer tricycles.

Table 3 shows that the average grant received per month was PHP 1400 (US 29), which is roughly the same in both Low and High MMR areas. Respondents have been members of the 4Ps program on average for 4 years, but those living in High MMR areas have been members for longer, at 5 years, while the respondents in Western Samar have been members for 6 years. The single most prevalent mode of grant transfer is the On-site over the counter transaction (41.54%), which is still the method most used in High MMR areas (65%). Provinces differ, with CASH Cards being most prevalent in bigger provinces—Zambales, Cebu, and South Cotabato. A bank-backed prepaid card is a mode of grant transfer in one urbanized province. Urban settings have program members for 4 years with more than a quarter (28.37%) covered as program beneficiaries for 5 years. Those in rural settings are more dispersed, with 32.57% having been members for 6 years, and 29.14% having been a member for only a year.

	Avg amounts received per household in PHP; average (min to max)	Avg years of membership; years	Most prevalent mode of cash transfer; % of respondents
Overall	1400 (250 to 4000)	4	On-site over the counter transaction (41.54%)
High MMR Provinces	1400 (250 to 3600)	3	On-site over the counter transaction (65.13%)
Low MMR Provinces	1400 (500 to 4000)	5	CASH Cards (50.26%)
Urban	—	4	—
Rural	—	6	—
Pampanga	1100 (500 to 1800)	3	Landbank Prepaid card (95.38%)
Zambales	1100 (500 to 1800)	4	CASH Cards (78.46%)
Cebu	1600 (500 to 3600)	2	CASH Cards (60%)
Western Samar	2800 (1000 to 4000)	6	On-site over the counter transaction (100%)
South Cotabato	1400 (250 to 3200)	4	CASH Cards (90.77%)
Compostela Valley (Davao de Oro)	1300 (500 to 3200)	4	On-site over the counter transaction (93.85%)

Table 3.

Profile of pantawid membership.

4. Analytical results and discussion

4.1 Model 1: factors associated with pregnancy during 4Ps

Table 4 highlights the likelihood of being pregnant among 4Ps clientele. These factors explain 49.14% of the variability of the odds of pregnancy during the 4Ps period. The odds of getting pregnant decreased for every peso increase in amount received (OR = 0.9969, p-value<0.001), for every year increase in age when the mother first got married (OR = 0.9157, p-value<0.001), and for those who lived far from the health center (OR = 0.2246, p-value<0.02). These findings show, for instance, a negative association between the cash transfer and odds of getting pregnant, with the increase of PHP 1 (USD 0.02) received by mothers decreasing the odds of getting pregnant by 0.31%. However, with the odds ratio approaching 1 (at .99), the association of the two variables can very well be independent. The weak association may also be due to the lack of variability in the transferred cash, as shown by the narrow confidence intervals.

	P-Value	95% Confidence interval	Adjusted odds ratio
Amount received from Pantawid, per month	0.000	0.9962 to 0.9976	0.9969
No. of children before birthed child	0.000	1.8845 to 3.5762	2.5962
High MMR	0.000	1.8888 to 7.7624	3.8283
With teenage child	0.000	2.2552 to 11.0251	4.9863
Age of women when first married	0.010	0.8565 to 0.9790	0.9157
Far distance to the health center (self-perceived)	0.010	0.0726 to 0.6952	0.2247
Did not finish HS (mother)	0.020	1.1370 to 4.3766	2.2308

Table 4.

Likelihood of getting pregnant as Pantawid members.

P-Value equals to 0.000 means p-value <0.001; R2 = 49.14%.

On the other hand, the odds of getting pregnant are increased for every increase in the number of children prior to 4Ps (OR = 2.59, p-value<0.001), for women who lived in a high MMR province (OR = 3.83, p-value<0.001), for women who had a teenager child (OR = 4.98, p-value<0.001), and for women who were unable to attain at least a high school degree (OR = 2.23, p-value<0.03). These findings show that women with teenage children have almost 5 times higher odds to get pregnant than those who do not have teenage children, reflecting perhaps the lack of vigilance in getting pregnant among relatively older but fecund women.

These results are broadly consistent with the literature. Early marriage and limited education have been shown to reduce the likelihood of pregnancy [8]. The effect of the amount received through a conditional cash transfer is less clear, as a systematic review done on CCT schemes in Latin America shows that receipt of financial incentives does not have a clear effect on fertility [9].

4.2 Model 2: demand for facility-based deliveries

A health conditionality related to maternal health is regular antenatal visits. It is surmised that once women get into the health system before deliveries, they will deliver in health facilities. Table 5 highlights the factors affecting health facility delivery, the final model of which explains under 10 percent or just 8% of the variability in facility-based deliveries. Women in areas with high maternal mortality and living far from a health center have 3–4 times the odds of delivering outside of a health facility.

	P-Value	95% Confidence interval	Adjusted odds ratio
High MMR	0.006	1.4753 to 10.1065	3.3022
Far distance to health center (self-perceived)	0.000	1.6893 to 6.4553	3.8613

Table 5.

Factors for delivery outside of health facility.

P-Value <0.001; R2 = 8.05%.

The low explanatory power of facility access may signify that variability in using health facilities for deliveries can be attributed to other factors. Literature shows how facility-based deliveries in LMICs are strongly influenced by local beliefs about what constitutes “normal” childbirth, perceptions about the quality of care they may receive, the fear of the possible costs of delivering at a facility [10], the comfort they may feel in their own home [11], and the role played by the husband in encouraging or discouraging facility-based delivery [11, 12].

4.3 Model 3: risk factors for non-utilization of ANC, defined as having 3 or fewer ANC visits

The likelihood of ANC non-utilization is shown in Table 6. The final model had an explanatory power of 10.89% and was significant at p-value<0.02.

	P-Value	95% Confidence interval	Adjusted odds ratio
Years married	0.006	1.0468 to 1.3226	1.1766
With teenage child	0.018	0.0343 to 0.7267	0.1578
Far distance to health center (self-perceived)	0.148	0.6443 to 18.4265	3.4455

Table 6.

Factors for non-utilization of ANC.

P-Value = 0.018; R2 = 10.89%.

Findings show that: (1) For every increase in years married, the odds of not availing of antenatal care increases by approximately 17.66%; (2) women with a teenage child have 0.1578 times the odds (meaning 84% less likely) of not availing of antenatal care (i.e. those with a teenage child are more likely to seek antenatal care); and (3) there was insufficient evidence to demonstrate an association between distance from the health center and absence of antenatal care. Having a teen-aged child helps with care for younger children at home, freeing the pregnant woman to have more antenatal visits.

5. Conclusion

The health conditionality, a compulsion provided by the cash grant, is anchored on a human capital perspective that a child’s early years, right from conception, must be appropriately nurtured by a trained health worker or professional; and the proper care and nutrition of neonates and mothers (antenatal care) and young children (immunization and deworming) must be ensured. The standard theory posits that having more children is linked to poverty, not only in terms of direct costs of having children but also in terms of women trading off employment and income to care for children. Our study shows a significant but weak relationship between the cash transfer and the likelihood of pregnancy. Other facilitating factors that lower chances of being pregnant during Pantawid membership include having been married longer and being far (self-perceived) from health centers.

There may be limited understanding of their fecundity and unsystematic use of family planning, since women who already have teenage children were also more likely to get pregnant again. The challenge therefore is inculcating a greater sense of deliberation over fertility control for both couples. The low male attendance in family development sessions points to the need to bring in the husbands into the learning loop, if not through lectures, then possibly to alternative modes of delivery, through cooperatives, workplaces, and the like. The challenge of modern contraceptive uptake remains within the medical establishment to prescribe it and for health professionals to continue advising, anticipating, and being proactive with women’s queries on the side effects of some methods. More work needs to be done to ensure that women deliver in health care facilities, especially in areas with already high MMR. Local government units (LGUs) have been known to send their ambulance service to pick up women about to give birth in farther areas.

There is an unanticipated significant presence in the study, the presence of a teenager or adolescent child in the family. That having an adolescent child in the family can increase chances of pregnancy and improve the likelihood of mothers seeking antenatal care may indicate the relatively long reproductive span of Pantawid members and that there should be no slacking in informing women that they remain fecund well into the years they may expect to be menopausing. The study affirms the need not to slack on Pantawid’s intended outcomes of delaying the age of marriage and completing high school as a way to break the cycle of poverty for many families. There is a need to be relentless in the pursuit of increasing awareness and utilization of family planning and maternal health care services, sustained by family development sessions as a platform for family and community transformations.

Acknowledgments

The currency conversion used was USD 1 = PHP 47.49. This is based on the official exchange rate of the World Bank for 2016, obtained from the World Bank Open Data website.

The study on which this chapter is based would not have been possible without the support of the staff and fieldworkers of the Department of Social Welfare and Development. Appreciation is also extended to the United Nations Population Fund (UNFPA) for co-funding the study and shepherding through the technical aspects of the process. These offices and their personnel are, however, not in any way responsible for the errors and omissions that may be contained in this chapter.

Author contributions

MCGB and RDFE drafted and revised this manuscript, acquired and interpreted the data, and approved the manuscript for publication. MECY and MCGB conceptualized and designed the study. MECY, RDFE, and KVR led fieldwork teams and likewise interpreted the data. VOCR conducted the statistical runs and analyzed the data.

Ethics approval

This study was approved on November 3, 2016, by the St. Frances Cabrini Medical Center and Cancer Institute – Asian Eye Institute Ethics Review Committee (ERC #2016-034).