Ngalinda, Innocent: Age at First Birth, Fertility, and Contraception in Tanzania

187

Chapter 6. Contraception in Tanzania

6.1. Introduction

In discussing proximate determinants of age at first birth and fertility (Chapter 2), the most frequent proximate determinant was contraceptive use. In this study contraception means the use of modern methods of family planning. Contraception is important because it can delay the first birth and reduce fertility by either spacing or limiting births. In Chapter 4, we found that 14 percent of sexually active women in Tanzania were doing something to delay their first births. In Chapter 5 we found that in Tanzania fertility has started declining. This was confirmed by the results of decomposition of change in fertility in so far as the proportion of married women is decreasing and the duration of breastfeeding is shortening while contraceptive practice is slightly compensating natural birth control. Therefore it is necessary to understand the levels and determinants of contraceptive use in order to formulate policies supporting proper strategies for raising contraceptive prevalence.

An important term that needs to be introduced at this point is contraceptive transition. Contraceptive transition is a process that takes place when individuals or couples become able to conceptualise and accept logically and emotionally their own ability to control their fertility. They must perceive that it is in their own self-interest and within their power to control births by spacing or limiting them. Moreover, they must have reasonable access to an acceptable and reliable means of controlling fertility. These conditions mean that the couple is ‘ready, willing and able’ to reduce their fertility, and actually use a contraceptive. Freedman and Freedman (1989) also provide a similar description of the process, stating that demand for fertility control has two components, namely the desire to limit the number of births or to space them widely, and the readiness to use fertility control measures for either purpose.

There may be a wide range of cultural, socio-familial, economic or programmatic factors that make it difficult for an individual to implement the decision to contracept. These factors include the financial and cultural costs of using contraception. These


188

factors will influence the decision to attempt the use of contraceptive. Thus the decision to initiate contraception is, ‘partly a function of the particular mix of means available to the decision-maker and a subjective evaluation of the cost. These costs may be high at the outset of the transition, but they erode more rapidly than method-specific costs’ (Robinson and Cleland, 1992: 118).

If an individual decides to attempt to practice family planning, the question then becomes a choice of the method, and method-specific costs become relevant. Such factors as financial costs of various methods, various sources, method-specific side effects, and personal preferences become more important. In this regard, it is clear that certain methods will have much higher psychosocial costs than others and this will influence their adoption.

Finally, once a method has been adopted, the user’s experience with that method (and the perceived cost) will determine whether the user continues or discontinues use. The information and counselling that the user received at the time of the initiation of the use will very likely influence subsequent continuation or discontinuation. Similarly, the other methods available will play an important role in whether the dissatisfied user becomes a ‘switcher’ or a ‘discontinuer’.

Unmet need may arise as the result of factors at any one of these decision points. It may result from imprecise or weak fertility control goals on the part of the individual or couple, from high generalised ‘costs’ of attempting regulation, from a limited method mix or high method-specific costs, or from limited or inaccurate information (Robinson and Cleland, 1992). Unmet need may also arise from discontinuation of the method use.

It seems likely that the point at which unmet need emerges shifts as the contraceptive transition progresses. In the pretransition and early stages of transition, the precision and intensity of fertility reduction goals as well as the general costs of undertaking regulation are likely to be the major factors leading to unmet need. Later on in the transition, method-specific costs, availability, and information become more important.


189

6.2. Levels and Determinants of Contraceptive Knowledge and Use

The knowledge of levels and determinants of contraceptive use is important in any effort to reduce fertility within a country. As Caldwell and Caldwell (1977) point out, the path of fertility will be determined by the extent to which modern contraception substitutes abstinence, and ultimately by the extent to which it is more efficient than periodical abstinence as a means of fertility regulation.

Several studies have been carried out to investigate factors influencing the use of contraceptives. In Tanzania such studies include that of Muna (1987), who focused on the Morogoro urban district; Madihi (1988) who did a case study of married Dar es Salaam women; Nanyaro (1992), whose work was based on the Dodoma region; Ghuhiya (1993), who focused on Zanzibar; and Mbago (1996), who did a case study of eight regions. All these studies with exception of Mbago’s, analysed factors influencing contraceptive use based on simple cross tabulations. No rigorous statistical analysis was carried out. Mbago (1996) focused on the following eight regions: Dar es Salaam, Pwani, Kilimanjaro, Dodoma Rukwa, Ruvuma, Mwanza, and Shinyanga. In his study, Mbago carried out multivariate analysis to ascertain the relative importance of factors in a multiple logistic regression model. But his study cannot be generalised as Tanzania has 25 different regions.

The major focus of the following sections is the examination of the extent of contraceptive use and the identification of sub-groups of women who are particularly unlikely to use contraception. The chapter examines the individual-level factors related to contraceptive use only. Both tabular analysis and logistic regression analysis are used for these purposes. The background variables used throughout this chapter are similar to those used in previous chapters. However, before studying contraceptive use, it is important to have an idea about the knowledge of contraceptive methods among women. As explained earlier in the introduction of this chapter it is an important pre-requisite of contraceptive use.

In this study we will study currently married women and never married women separately because theoretically family planning services in Tanzania are not provided for unmarried women. This is particularly important because premarital sexual activity is very common as seen in Chapter 4. The only way to avoid premarital pregnancies of


190

which most are unwanted, is to take proper measures to avoid premarital conception (fertilisation).

6.2.1. 5.2.5. Knowledge of Contraception

The 1996 TDHS collected information about the knowledge of contraception for all respondents. The interviewer asked the respondent to name all the family planning methods of which she had heard. For methods not mentioned by the respondent, the interviewer gave a one-line description and asked again whether the respondent had ever heard of it. This is known as ‘probing’. The respondent is considered to have known a method if she said that she had heard of it either spontaneously or after probing. Table 56 presents the percentage distribution of the respondents according to their knowledge of contraception and selected background characteristics. A woman is considered to know a modern method if she stated that she knew at least one of the following methods: the pill, intrauterine device (IUD), injection, foam or jelly, barrier methods (diaphragm, condom), female or male sterilisation. A woman is considered to know only traditional methods if she stated that she knew only one of the following methods: periodic abstinence, calendar (safe period), Billings (temperature and or cervical mucus method), withdrawal (coitus interuptus), or any other traditional method known as herbs and strings. The ’none category’ includes all women who said they did not know any method at all.

Table 56 indicates a high level of knowledge (88.5 percent) among currently married women in Tanzania; 87.8 percent know at least one modern method. Only 0.7 percent know only a traditional method, while 11.5 percent of currently married women do not know any method of family planning. However, the level of knowledge of never married women is lower than of currently married women. Only 69.8 percent of the never married women claimed to be aware of a contraceptive method. A major challenge for Tanzania therefore is to work out how to introduce knowledge of family planning methods to the 30.2 percent of never married women and 11.5 percent of currently married women who have never heard of a contraceptive method.


191-192

Table 56: Percentage distribution of women by knowledge of contraception by selected background characteristics

Currently married

Never married

Characteristics

None

Tradit.

Modern

N

None

Tradit.

Modern

N

Place of Residence

Urban

2.5

0.1

97.4

1130

14.8

85.2

560

Rural

13.9

0.9

85.2

4282

36.7

0.1

63.2

1327

Education

None

24.3

1.8

73.9

1841

55.8

44.2

217

Incomplete primary

9.4

0.4

90.2

908

47.0

53.0

538

Complete primary

3.7

0.1

96.1

2439

19.7

0.1

80.2

912

Secondary +

100

223

7.7

92.3

221

Complete primary

6.5

0.4

93.2

2428

Secondary +

2.9

0.3

96.8

657

Religion

Moslem

5.8

0.5

93.8

1697

28.6

71.4

539

Catholic

8.5

0.4

91.1

1621

29.9

0.1

70.0

676

Protestant

8.8

0.7

90.5

1330

25.2

74.8

548

Other

35.4

2.1

62.4

748

62.3

37.7

122

Polygyny

Monogamous

9.3

0.7

90.0

3853

Polygamous

17.0

0.9

82.1

1525

Age at first birth

<15

16.0

2.0

82.1

351

10.0

90.0

10

15-17

11.8

0.5

87.7

1582

10.1

89.9

109

18-19

7.9

0.6

91.5

1437

5.5

94.5

128

20-21

9.4

0.5

90.1

852

3.8

96.2

78

22-24

13.9

1.0

85.2

512

100.0

36

25 +

15.0

1.7

83.3

234

17.9

82.1

28

Age at first intercourse

At first union

14.7

1.3

84.1

1587

<15

20.4

0.8

78.8

707

18.1

0.7

81.2

138

15-17

10.6

0.4

89.0

1646

13.5

86.5

453

18-19

5.0

0.5

94.5

599

4.2

95.8

143

20-21

5.0

95.0

218

6.3

93.8

64

22-24

4.8

1.6

93.5

62

100.0

17

25 +

3.4

3.4

93.1

29

13.3

86.7

15

Age at first marriage

<15

19.2

1.5

79.4

756

15-17

11.5

0.7

87.8

2073

18-19

8.7

0.6

90.7

1171

20-21

8.3

0.6

91.1

689

22-24

12.2

0.5

87.3

426

25 +

10.1

0.7

89.2

297

Children ever born

None

18.5

0.7

80.9

444

36.3

0.1

63.6

1499

1-2

10.4

0.4

89.2

1612

7.6

92.4

341

3-4

7.7

0.8

91.5

1343

100.0

37

5-6

11.8

0.8

87.4

924

14.3

85.7

7

7+

14.9

1.1

84.0

1089

100.0

4

Current age

15-19

20.0

0.2

79.8

401

39.8

0.1

60.1

1292

20-24

9.3

0.4

90.3

1131

9.8

90.2

410

25-29

7.4

0.5

92.1

1184

7.5

92.5

106

30-34

9.0

0.3

90.7

947

10.0

90.0

50

35-39

11.5

0.9

87.6

740

20.0

80.0

15

40-44

14.8

1.1

84.1

561

100.0

10

45-49

21.7

2.5

75.8

447

100.0

4

15-49

11.5

0.7

87.8

5411

30.2

0.1

69.7

1887

Source: calculated from 1996 TDHS

It is evident from Table 56 that a majority of the currently married women know at least one modern method. This is true even for women, who never attended formal schooling, and those residing in rural areas. However, only women with traditional beliefs or no beliefs at all do not follow the same trend. A very large minority of these women (35.4 percent) has never heard of any contraceptive method. This is not surprising as this sub-group of women is likely to have a combination of the factors (e.g. reside in rural areas, and have little or no education). Therefore, they are disadvantaged in terms of gaining modern ideas.

The results for never married women are mixed. Although the percentages are lower than those for currently married women, in many categories a majority of unmarried women knows at least one contraceptive method. However, there are two categories in which the majority of never married women do not know any contraceptive method, namely those with no schooling or low levels of schooling and those categorised to have another religion than Catholic, Protestant and Moslem. As an initial step in raising contraceptive prevalence in Tanzania, these groups should be targeted.

6.2.2. 5.2.6. Current Use of Contraception

Women, who were not pregnant at the time of the interview, were asked if they were currently doing something or using a method to delay or to avoid getting pregnant. This information is very useful as a measure of one of the proximate determinants of fertility as well as a measure of the coverage of family planning programmes (Bertrand et al.,


193

1993). In order for the results to be comparable with other studies, we computed contraceptive prevalence (the percentage of all women of reproductive age, currently married or living in a union, using some type of contraception), according to their background characteristics. The percentages of current users among never married women were also computed. The results are presented in Table 57 .

Table 57 shows that 13.3 percent of the currently married women were using modern contraception at the time of the interview. Among the never married women, only 5.6 percent were using a modern contraceptive method. These rates indicate that the use of contraception is very low in Tanzania. Nevertheless there is a rising trend of contraceptive users as in the 1991/92 TDHS it was estimated to be 6.6 percent for currently married women and 2.6 percent for never married women (Mturi, 1996).


194-196

Table 57 further reveals that currently married Tanzanian women residing in urban areas have a contraceptive prevalence of about 17 percentage points higher than those residing in rural areas. However, the difference between urban and rural areas is smaller for never married women (8.4 percentage points). This is due to the tendency of never married women to shy away from attending family planning clinics. Only those who are already mothers can freely attend clinics. It is considered a taboo for a never married woman to attend family planning clinics. The highest contraceptive prevalence (20 percent or more) is observed for those women who had their first sexual intercourse at age 20 or older. Interesting results are shown for those women at the age of 15-19 and 45-49. They have a low contraceptive prevalence. The reason may be that adolescents are newly married, and marriage is looked upon as an institution of producing children as well as they have no access to family planning services. For older women have reduced their coital frequency and most of them rely on other methods like string tie and are afraid to talk about them in an interview. However, a good number of older women might be not sexually active. It is also interesting to note that women who had their first sexual intercourse, marriage and birth below the age of 15 years have a low percentage of contraceptive use. Never married women, who gave birth to their first child when they were younger than 15 years, seem not to use contraceptives at all. This suggests an interaction between not using contraception and early age at birth. On the other hand it reflects the real situation where as family planning services are thought to be for only married women. Hence never married women would shy away even to acknowledge the use of contraceptives.

Table 57: Percentage distribution of women currently using contraceptive methods by selected background variables

Currently married

Never married

All

Non-pregnant

All

Non-pregnant

Total (15-49)

13.3

15.3

5.6

5.8

Number of cases

5411

4717

1887

1823

Place of residence

Urban

26.6

29.5

11.6

12.0

Rural

9.8

11.3

3.2

3.3

Education

None

5.2

5.9

0.9

1.0

Incomplete primary

13.1

14.8

1.7

1.7

Complete primary

17.9

20.9

7.2

7.6

Secondary +

31.4

35.9

13.1

13.5

Partner's education

None

5.3

6.0

Incomplete primary

9.2

10.3

Complete primary

14.8

17.5

Secondary +

29.1

32.0

Religion

Moslem

17.0

19.0

7.6

7.9

Catholic

11.9

13.6

6.2

6.4

Protestant

17.4

20.2

4.2

4.3

Other

0.8

1.0

-

-

Polygyny

Monogamous

14.8

17.1

Polygamous

9.5

10.8

Age at first birth

<15

10.9

11.8

-

-

15-17

14.3

16.1

16.7

17.5

18-19

14.7

16.7

18.1

19.3

20-21

15.9

18.3

16.7

17.6

22-24

14.9

16.5

19.4

21.2

25+

12.8

14.2

-

-

Age at first sexual intercourse

At union

10.7

12.2

-

-

<15

8.5

9.9

7.3

7.9

15-17

14.5

16.7

13.7

14.9

18-19

15.7

18.2

18.9

19.9

20-21

21.1

24.1

9.4

10.3

22-24

25.8

31.4

-

-

25+

17.2

18.5

-

-

Age at first marriage

<15

8.7

9.8

15-17

13.3

15.4

18-19

13.9

16.1

20-21

16.6

19.0

22-24

15.0

17.3

25+

12.4

13.8

Current age

15-19

4.5

5.8

2.4

2.5

20-24

12.7

15.6

12.2

12.9

25-29

14.4

17.0

14.2

15.3

30-34

14.0

15.8

17.6

19.1

35-39

15.9

17.7

6.7

6.7

40-44

17.1

17.7

-

-

45-49

9.2

9.3

-

-

Children ever born

None

0.5

0.6

2.9

3.0

1-2

12.5

14.5

14.7

15.6

3-4

16.1

18.3

29.7

32.4

5-6

15.2

17.0

16.7

16.7

7+

14.8

15.9

-

-

Children surviving

None

0.5

0.7

3.0

3.1

1-2

12.8

14.9

15.7

16.6

3-4

16.3

18.3

33.3

37.0

5-6

15.4

16.9

-

-

7+

16.0

17.3

-

-

Children dead

None

14.1

16.3

5.3

5.4

1

12.5

14.5

15.0

17.3

2

10.9

12.3

33.3

33.3

3

10.4

11.2

-

-

4+

13.9

14.7

-

-

Source: calculated from 1996 TDHS

Never married women display a different pattern of contraceptive use. Among never married women, the highest contraceptive prevalence is observed for women who had their first sexual intercourse and first birth above the age of fifteen and those who gave birth to more than one child. Surprising enough are those who had their first birth and intercourse below 15 years. Their contraceptive prevalence is very low although they were expected to have more prevalence than others as they started sexual activities earlier. Due to the experience of very early and potentially difficult motherhood, they are expected to contracept more than those who started sexual relationships and motherhood later. However, this might reflect the interaction between early child bearing and none usage of contraception as low contraceptive prevalence leads to earlier birth. Somehow the expectation was that the early motherhood prompts them to prevent or space births in the future. However, these expectations could not be verified.

Let us take a look at women’s educational level, which is cited as the most important variable associated with contraceptive use in many countries. It has been observed that better educated women are more likely to use contraception (Rutenberg et al., 1991; Robey et al., 1992; Bertrand et al., 1993). The Tanzanian situation reflects this pattern. The percentage of currently married women using contraception increases consistently


197

with the level of education. The gap between users who attended at least secondary school and those who never attended any type of schooling is enormous (26 percentage points). This is also true when the education of their partners is considered.

The extent of the use of contraception for currently married women, does not vary much between Moslems and Protestants. In the category ‘never married’ women, Moslems have the highest contraceptive use followed by Catholics. Protestant women, who are currently married, are more likely to use contraceptives than other currently married women, while Protestant never married women are less likely to use contraception than Moslem and Catholic never married women. Whether married or not, women who declared themselves not to belong to one of these three religions are less likely to use contraceptives than other women.<28> In traditional African societies, people believe that God has control over the human reproductive system or that children are a gift from God. Therefore, no one should prevent a child from coming into the world (Omari, 1989). Most women with traditional faiths are likely to advocate this ideology. Omari (1989) too has argued that Tanzanian women who follow traditional belief systems are less likely to use contraception than other women.

Women bound in a polygamous marriage are less likely to use contraceptives than women who live in monogamous marriages. A lower frequency of intercourse for women in polygamous marriages can discourage them from using contraception. Also, these women are likely to adhere to traditional values and customs that encourage large families. Judging from the Table 57 , it seems that contraceptive use does not directly relate to ages at first birth, intercourse, and marriage.

Contraceptive use is higher among currently married women aged 40-44 years than among women either younger or older than that. Figure 22 presents the contraceptive prevalence by five-year age groups of women. The shape is approximately an inverted U-shape. Contraceptive prevalence is lowest for the age group 15-19, increases gradually to reach a maximum at the age group 30-34, after which it decreases consistently to the age group 45-49. This pattern has been found almost everywhere in


198

sub-Saharan Africa (Rutenberg et al., 1991; Robey et al., 1992; Bertrand et al., 1993; Curtis and Neitzel, 1996).

Figure 22: Percentage distribution of currently married women using contraception by current age

Source: calculated from 1996 TDHS

Finally, it is important to examine the association between the number of children ever born to a woman and contraceptive use. The number of children ever born to a woman has been found to be associated with the use of contraception (Rutenberg et al., 1991; Robey et al., 1992). The TDHS data show that the percentage of women who never had a live birth and who are using contraception is very small (particularly for currently married women). Among currently married women with at least one live birth, the extent of use varies much with the number of children ever born. The use of contraceptives increases with the increasing number of children ever born. The percentage of never married women with no live birth using contraception is 2.9


199

percent, 14.7 percent for those having 1 to 2 children, and 29.7 percent for those having 3 to 4 children. Those who had 7 children stated not to use contraceptives due to the fact that they are either not currently married, have no regular partners, or have reached menopause (44-49 years).

Table 58: Contraceptive prevalence in sub-Saharan Africa (currently married women 15-49 who know and use any modern method in percentage) and other selected developing countries

Country

Year

Knowledge

Use

All

Non-pregnant

Benin

1996

76.2

3.4

4.0

Central African Republic

1994

68.6

3.2

3.8

Comores

1996

98.1

11.4

12.8

Côte d’Ivoire

1994

71.5

4.3

5.0

Ghana

1993

90.7

10.1

11.4

Kenya

1993

96.9

27.3

30.8

Mali

1995

64.6

4.5

5.3

Tanzania

1996

87.7

13.3

15.3

Uganda

1995

91.6

7.8

9.6

Zambia

1996

97.7

14.4

17.2

Zimbabwe

1994

98.5

42.2

48.0

Egypt -North Africa

1995

99.8

45.5

50.9

Bangladesh-Asia

1996

99.8

36.2

39.7

Brazil - South America

1996

99.9

70.3

74.3

Dominican Republic - Latin America

1996

99.7

59.2

65.2

Source: calculated from DHS III

The DHS-III surveys included only eleven countries in sub-Saharan Africa ( Table 58 ), other countries are representatives of their regions. In all countries analysed, nearly all currently married women know at least one modern method of contraception. A high degree of awareness of at least one modern method is observed in all countries in sub-Saharan Africa except Mali, Central African Republic, and Côte d’Ivoire, where such knowledge is less prevalent. In these three countries more than 25 percent of currently married women admited that they are unaware of any modern method of contraception. Contraceptive use follows the same trend. Since women in these countries are to have low knowledge of contraceptives (below 75 percent), the percentage of women who use contraceptives is below 5 percent. This type of measurement is a crude measure as the denominator includes those women who are not at risk of becoming pregnant (already pregnant). We tried to refine our denominator by eliminating those women who acknowledged that they are pregnant. After refining our denominator, the percentages of


200

contraceptive users changed. The percentage of non-pregnant married women using contraceptives in Benin, Central African Republic was still under 5 percent of all married non pregnant women. However a good number of women in these countries are believed to sterile.

In sub-Saharan Africa the percentage of non-pregnant married women who use a modern contraceptive method ranges from 4 percent in Benin and the Central African Republic to 31 percent in Kenya and 48 percent in Zimbabwe. Kenya and Zimbabwe are countries with exceptionally high contraceptive prevalence in sub-Saharan Africa as most of the countries in this region have less than 20 percent contraceptive prevalence of all non-pregnant married women contracepting. Countries on other continents have more than 50 percent of non-pregnant women using contraceptives, for example Egypt (51 percent), the Dominican Republic (65 percent), and Brazil (74 percent). In general contraceptive use is still very low in sub-Saharan Africa.

6.2.3. 5.2.7. Contraceptive Method Mix

It is important to examine the specific contraceptive methods used, since different methods have different implications for family planning programmes. In Tanzania family planning services are provided free of charge (except for selected methods in some private clinics). Therefore, the financial costs of contraceptives are unlikely to be a deterrent to the use of contraceptives for Tanzanian women. The major constraints for women who know and want to use a modern method are perhaps the high transport costs and the lack of facilities in clinics.

Table 59 presents the percentage distribution of women using contraception at the time of the interview according to the method they used and their marital status. Tanzanian women use modern methods more frequently than traditional methods. The pill is the most popular method for married women followed by injection. For singles the situation is different. Although the pill is the most popular, it is followed by periodic abstinence. For widowed women all three methods i.e. pill, injection, and condom are popular. In general, the pill is most popular followed by injection. The main reason for choosing the pill as the major contraceptive method for most women might be due to the history of contraceptives in Tanzania. Pills were the first contraception method introduced in


201

Tanzania. It is interesting to note that family planning is nicknamed ‘kidonge/majira’ which means ‘pill’. ‘Pill’ is used as an overall-term to refer to any kind of contraceptive. One cannot be sure which method a woman uses if she speaks of using pills. The use of condoms has serious implications since nowadays it is almost exclusively associated with HIV/AIDS. People do not consider it a ‘normal’ method of contraception unless somebody is afraid of being infected with HIV. The association with HIV/AIDS causes couples not to use them in order not to imply that the partner is not reliable in terms of sexual fidelity.

Table 59: Percentage distribution of women currently using a contraceptive method according to the methods used and marital status

Never married

Married

Widowed

Divorced

Not living together

In total

Number of Women

157

997

50

72

30

1,306

Pill

29.3

29.9

14.0

34.7

40.0

29.7

IUD

1.9

3.2

6.0

2.8

3.3

3.1

Injections

12.7

24.2

20.0

25.0

30.0

22.8

Diaphragm/Foam/Jelly

0.6

0.1

Condom

22.9

4.5

20.0

15.3

10.0

8.0

Female Sterilisation

10.1

20.0

6.9

3.3

9.0

Male Sterilisation

0.1

0.1

Norplant

0.2

1.4

0.2

Modern methods

67.5

72.2

80.0

86.1

86.7

73.0

Periodic Abstinence

28.0

11.1

14.0

5.6

6.7

12.9

Withdrawal

3.2

14.2

2.0

4.2

3.3

11.6

Other

0.6

0.4

4.0

1.4

0.6

Strings

0.6

1.0

2.8

1.0

Herbs

1.0

3.3

0.8

Traditional methods

32.5

27.8

20.0

13.9

13.3

27.0

Sum of percentages

100

100

100

100

100

100

Source: calculated from 1996 TDHS

It has also been noted that the availability and promotion of the pill has much to do with its popularity in many countries (Bulatao, 1989). Injection proves to be popular due to the fact that they can be used without the agreement of the husband or partner. Many men do not want their sexual partners to use any contraceptive methods especially in rural areas. Men would like many children as a future labour force or future social security. The other hidden reason for men to dislike modern contraceptives might be due to culture that permits men to have sexual activities outside their marriage as long as their wives are either pregnant or breastfeeding. During that time, and only then, they can engage in sexual relationships with other women. In many of the various cultures in


202

Tanzania, men have an interest in keeping their wives either pregnant or breastfeeding since they want to keep their privilege of having sexual partners outside their marriage. Therefore, they do not to accept modern contraception, which implicitly restricts their sexual freedom or give some sexual freedom to their women.

A married woman usually does not have sexual intercourse with another man for the fear of having a child that is not her husband’s. This could lead to divorce. In this sense, men believe that contraceptive use is a warrant for women to commit adultery, and are therefore against it. For these reasons injections are now popular to countercheck these cultural factors as a husband would not know if a woman were using contraceptives. The convenience of injections will in future make them popular since instead of taking a pill every day it is convenient for a woman to get an injection once in three months.

Many people in Tanzania dislike barrier methods, especially condoms, because they are considered to reduce sexual pleasure. However, as an HIV-preventive measure condoms are highly used nowadays. As explained earlier, the use of condoms among married couples is very low due to the implication of HIV/AIDS protection. This can be seen in Table 59 where the percentage of condom users among married women is very low compared to widows, divorcees and singles. Widows, divorcees and singles use condoms much more. The result for divorcees is interesting in a way. In the Wanyambo tradition <29>, for example, divorce has a different meaning than in other cultural contexts. Culturally, the divorcee remains the wife of the former husband unless she is remarried to another man. In fact, divorce means temporary separation after which the couple could reunite even years later. Hence, divorcees are regarded by society as married women. The use of condoms could lead us to draw a conclusion that the Wanyambo traditions might be universal to many Tanzanian. It shows that although divorcees are not allowed by the society to be sexually active, in reality they are. From this fact divorcees may shy away to visit family planning clinics. Then the concept of family planning services to be provided for only married women is a misleading concept.

Most staff members in the MCH/FP and UMATI clinics follow the unwritten rule that family planning advice can only be given to married clients with a child. This situation reflects the opinion voiced by politicians and religious leaders, that family planning


203

should be used for family spacing only not for delaying the first birth or preventing premarital births. Therefore mainly married women with at least one child have easier access to contraceptives. This situation might be changing gradually. But it is still awkward for an unmarried schoolgirl, for instance, to visit a family planning clinic. It is therefore convenient for them to use traditional methods that do not require them to go to clinics. Unmarried women with the courage to go to a clinic, or with other means of getting contraception (for example sending a person to obtain contraception on their behalf), prefer the condom because of its ‘non-clinical’ nature. The other reason for the condom being the most popular modern method among singles has to do with the protection it offers against sexually transmitted diseases such as HIV/AIDS (Mpangile et al., 1993). This is particularly important since never married women are likely to have multiple partners or may change their partners after a short period.

6.2.4. Multivariate Analysis of the Determinants of Current Contraceptive Use

Binary logistic regression is the multivariate analysis technique used to predict the presence or absence of a characteristic or outcome based on values of a set of predictor variables. It is similar to linear regression model but it is suited to models where the dependent variables are dichotomous. In this chapter, logistic regression will be used to examine the relative importance of the determinants of contraceptive use. The response variable is use or non-use of contraceptives at the time of the survey. Women, who were currently pregnant, or who were unsure about being pregnant, were excluded from the analysis. Table 57 describes the data used for the multivariate analysis (in the columns labelled ‘non-pregnant’) for currently married women and never married women. Women for whom some information on the explanatory variables is missing were dropped from the analysis. The final sample for currently married women includes 4673 women, and for never married women the sample is 1834. That is 1.2 percent of currently married women who were not pregnant had some missing information on at least one of the explanatory variables and hence were excluded from the multivariate analysis. The analysis of ever-married women produced results similar to that for


204

currently married women. Therefore the results are presented only for currently married and never married women.

6.2.4.1. Results for Currently Married Women

Table 59 presents the model for the determinants of current contraceptive use for women currently married or living with a partner. Six variables were found to influence the use of contraceptives significantly after keeping the other explanatory variables constant. The analysis indicates that women's education is the strongest predictor of the use of contraceptives in Tanzania. Women with ‘incomplete primary’ education were 2 times more likely to use contraceptives than women who had never attended school. The likelihood of using contraceptives increases further as the educational level increases beyond ‘lower primary’. Women, who had at least some secondary school education, were 3.37 times more likely to use contraceptives than women without schooling. It is interesting to note that the education of a woman's partner also has a significant effect independently of her own educational level. The direction of this effect is the same, although the odds ratio is weaker. The fact that both the education of women and of their partners were significant indicates that these two variables have separate effects in determining contraceptive use.

Protestant women are 1.1 times more likely to use contraceptives than Moslem women but the difference is not significant. However, Catholic women are 1.6 times less likely to use contraceptives than Moslems (1/0.64); the difference is significant. The odds ratio of religious groups other than three dominant groups is negligible although significant. This means that they are 11 times less likely to use contraceptives than Moslems. In general Moslems are more likely to use contraceptives than any other religious group in Tanzania.

The prevalence of contraception depends to a large extent on the type of the place of residence. Tanzanian women residing in rural areas are 2.4 times less likely to use contraception than their counterparts residing in urban areas. However, other studies in sub-Saharan African countries had different findings. Brass and Jolly (1993), for example, found that Kenyan women residing in urban centres are less likely to use contraception than women residing in rural areas (after controlling for other factors). They argue that once other variables that are known to influence contraceptive use are


205-206

controlled, ‘there is nothing about urbanisation itself that is significant in increasing contraceptive use and in fact it can be a negative influence’ (Brass and Jolly, 1993: 160). Furthermore, in a multivariate analysis for the pooled WFS and DHS data for Kenya, Ghana, Senegal, and Sudan, Bertrand et al. (1993) found that living in a major city had little or no direct effect on the knowledge of at least one modern method and on modern contraceptive use.

The analysis also suggests that the number of living children influence the use of contraception. Women with 1 to 2 children are 28.7 times more likely to use contraceptives than women without surviving children. The odds ratio increases with an increasing number of surviving children. Women with 7 and more surviving children are 58.67 times more likely to use contraceptives than women without any surviving child. Women with 1 or 2 children were less likely to use contraception, while the chance of using a method increased as a woman's family becomes bigger than 3 children.

Table 60: Odds ratio associated with the determinants of current contraceptive use for currently married women

Variable

Odds Ratio

95% Confidence Interval

Education

None

1.00

Primary incomplete

2.00***

1.470

2.713

Primary complete

3.13***

2.369

4.144

Secondary and above

3.37***

2.185

5.203

Partner Education

None

1.0

Incomplete primary

1.14

.809

1.619

Complete primary

1.89***

1.367

2.606

Secondary +

2.93***

2.048

4.197

Polygyny

Monogamous

1.0

Polygamous

.77**

.621

.949

Place of residence

Urban

1.0

Rural

.42***

.342

.505

Religion

Moslem

1.0

Catholic

.64***

.515

.797

Protestant

1.07

.866

1.329

Other/none

.09***

.039

.214

Children ever born

None

1.0

1-2

.90

.035

23.293

3-4

1.05

.039

28.467

5-6

1.30

.046

36.842

7+

1.50

.049

45.587

Children surviving

None

1.0

1-2

28.70**

1.432

575.432

3-4

40.01**

1.904

840.982

5-6

40.21**

1.822

887.686

7+

58.67**

2.522

1364.720

Children loss

None

1.0

1

.88

.681

1.136

2

.89

.580

1.350

3

1.08

.600

1.927

4+

1.61

.790

3.262

Constant

-2.62***

*** Highly significant at level .001

** Significant at level .01

* Significant at level .05

Source: calculated from 1996 TDHS

Although age at first birth did not show any significance ( Table 61 ), age at first intercourse, age at first marriage, and current age are highly related to contraceptive use. Women who had the first sexual experience after attaining age 15 to 17 were 1.5 times more likely to use contraceptives than those who had the same experience below age 15. The odds ratio increases as the age at first intercourse rises. Women who got married between age 15 and 17 were 1.6 times more likely to use contraceptives than those who were married before attaining age 15. The odds ratio increases as the age at first marriage rises. This shows that women who experience sexual intercourse or marry at an early age do not do something to plan their families. This was also found in bivariate analysis. Otherwise, they lack information on modern contraceptives as Table 56 shows. Modern contraceptive awareness increases with rising age at first intercourse, marriage, and birth. At the same time, knowledge of traditional methods of contraceptives decreases with increasing age at first intercourse, marriage, and birth. This means that adolescents should be given information about contraception at an early age.


207

Table 61: Odds ratio associated with the retrospective determinants of current contraceptive use for never and currently married women

Variable

Currently married

Never married

Current age

15-19

20-24

2.25**

6.67***

25-29

2.29**

8.24***

30-34

2.14**

14.99***

35-39

2.61***

2.82

40-44

2.59***

0.00

45-49

1.30

0.00

Age at first intercourse

<15

15-17

1.45**

1.32

18-19

1.59**

1.92

20-21

2.40***

0.47

22-24

3.18***

0.00

25+

1.13

0.00

Age at first marriage

<15

15-17

1.59**

18-19

1.74**

20-21

2.24***

22-24

2.04**

25+

1.51

Age at first birth

<15

15-17

1.09

5493.83

18-19

0.96

3101.47

20-21

0.86

2289.16

22-24

0.69

3567.88

25+

0.71

0.91

*** Highly significant at level .001

** Significant at level .01

* Significant at level .05

Source: calculated from 1996 TDHS

6.2.4.2. Results for never married women

The logistic regression model for the determinants of contraceptive use obtained for never married women is presented in Table 62 . All the variables found to influence contraceptive use for currently married women also influence it for never married women. The only exceptions are variables referring to the partner's characteristics (which are, of course, not observed for never married women). In addition, a woman's


208

age is an important determinant of contraceptive use for never married women. It seems that unmarried women are more likely to use contraception as they become older. Moreover, the relationship is non-linear. After age 35, the relationship is not significant, but never married women aged between 30 to 34 are 15 times more likely to use contraceptives than adolescents who are not married yet. However, the majority (more than 90 percent) of these women (35 years) are aware of family planning methods ( Table 57 ) but the sample size is too small to have meaningful regression results.

Table 62: Odds ratio associated with the determinants of current contraceptive use for never married women

Variable

Odds ratio

95% Confidence Interval

Education

None

1.0

Primary incomplete

1.72

.362

8.167

Complete primary

4.59*

1.093

19.314

Secondary +

9.34**

2.097

41.560

Place of residence

Urban

1.0

Rural

.36***

.229

.568

Religion

Moslem

1.0

Catholic

.65

.393

1.071

Protestant

.44**

.247

.781

Other/none

.003

.000

46688.004

Children ever born

None

1.0

1-2

5.39

.724

40.054

3-4

6.91

.117

409.499

5-6

16.68

.051

5420.387

7+

18.83

.000

20.000

Children surviving

None

1.0

1-2

1.11

.155

7.916

3-4

2.80

.055

143.848

5-6

.00

.000

5.375E+37

7+

.00

.000

4.695E+88

Child loss

None

1.0

1

1.01

.265

3.856

2

4.14

.112

153.526

Constant

-5.36

*** Highly significant at level .001

** Significant at level .01

* Significant at level .05

Source: calculated from 1996 TDHS


209

The number of living children is a very strong predictor of contraceptive use for never married women. Women with 1 or 2 living children are 8 times more likely to use a family planning method than women without a living child. Women with 3 to 4 children are 144 times more likely to use contraceptives than singles without any child surviving. Women with seven or more children are more likely to use contraceptive than singles without any surviving child. Urban women use contraception more frequently than rural women. Rural singles are 2.8 times (1/0.36) less likely to use a contraceptive than urban never married women; the relationship is statistically significant. A never married woman needs to have at least secondary education in order to be significantly more likely (9.34 times) to use contraceptives than a woman with no schooling. The results for women's religion show that Moslems use contraceptives more frequently than believers of any other denomination. While Catholic singles are 1.5 less likely to use contraceptives than Moslems, Protestants are 2.3 less likely to use contraceptives than Moslems. Non-believers are 333 times less likely to practice contraception than Moslem singles.

6.3. Unmet need for Contraception

It has been shown (see 6.1) that the majority of women in Tanzania know at least one modern method of family planning but very few actually use any method. This situation can arise in many ways as seen in Table 58 . Among the issues that have received attention in the literature is what is called the ‘unmet need’ for contraception (Mturi, 1996). That is the proportion of women who are exposed to the risk of pregnancy but are not using contraception despite the fact that they want to limit or space births.<30> The proportion of women with unmet need along with the proportion of women currently using contraception provide information on the magnitude of the potential demand for contraceptives and family planning services. Therefore, it is important for family planning programmes to utilise the information about unmet need so as to understand the specific needs of women in their reproductive years.


210

Women with an unmet need for family planning are classified into two groups: those who would like to postpone the next birth (spacing) and those who would not like to have further children (limiting). The unmet need for limiting births refers to women who state that they do not want any more births. Women who state that they want another birth in a period exceeding two years are considered to have unmet need for spacing if they do not use any contraceptive method. The definition of unmet need for contraception has been expanded recently to include pregnant and amenorrheic women, who became unintentionally pregnant because they had been unable to use contraception (Westoff, 1988; Westoff and Ochoa, 1991; Westoff and Moreno, 1992). If a pregnant or amenorrheic woman states that her current pregnancy or the pregnancy for her most recent birth was not intended, it means that access to contraception could have delayed or prevented that pregnancy. Thus the woman had unmet need for limiting births if she wanted no more births, or unmet need for spacing births if she wanted to postpone her next birth.

This section will estimate the magnitude of unmet need for contraception in Tanzania by using the 1996 TDHS data. The demographic and social variables associated with unmet needs for contraception will be examined using bivariate analysis as well as multivariate. Finally, the estimates of the total demand for family planning will be obtained by adding the proportion of women with an unmet need for contraception to the proportion of current contraceptive users. The analysis in this section will help to determine the actual and potential demand for family planning services in Tanzania.

As already said, women have an unmet need for contraception if they are not using a contraceptive method but are capable of conceiving, are exposed to the risk of pregnancy, and wish to avoid or to postpone pregnancy. Therefore, women currently using some form of contraception or being sterilised are assumed to have no further need. Dixon-Mueller and Germain (1992) have identified three groups of current users who can be said to have an unmet need: women who definitely want to avoid or postpone pregnancy but who are using an ineffective method; women who definitely want to avoid or postpone pregnancy but are using a theoretically effective method incorrectly or sporadically; and women who regardless of their reproductive intentions are using a method that is unsafe or unsuitable for them. However, the effect of including the latter two groups of women in the overall estimate of unmet need is


211

minimal, particularly if the contraceptive prevalence of the country under investigation is low. No attempt is made in this analysis to include any of the current users in the unmet group.

Women not using contraception are classified into two categories: pregnant or amenorrheic, and not pregnant or amenorrheic. For current pregnancy status, the TDHS asked: ‘Are you pregnant now?’ The respondents were supposed to give one of the three answers, ‘yes’, ‘no’ or ‘unsure’. Only those who answered ‘yes’ are regarded as pregnant at the time of the interview. Pregnant women are further classified according to whether or not their current pregnancy was intended. The question which is used in this classification is: ‘At the time you became pregnant, did you want to become pregnant then, did you want to wait until later, or did you not want to become pregnant at all?’ The responses to this question give the fraction of pregnant women who were unintentionally pregnant. If the pregnancy had occurred earlier than desired then it is unmet need for spacing births and if pregnancy was not wanted at all it is unmet need for limiting births. The major problem about the current pregnancy status classification is that many women in the early months of gestation do not know whether or not they are pregnant.

Currently amenorrheic women were identified by the response to the question ‘Has your period returned since the birth of (name)?’ where name refers to the latest child. Women who answered ‘no’ are considered amenorrheic and classified according to the intention of their most recent pregnancy. Amenorrheic women who would have wanted to postpone their last child for two or more years are considered to have an unmet need for spacing and those who stated they did not wanted another child are considered to have an unmet need for limiting. Currently pregnant or amenorrheic women, whose pregnancy was intended at the time at which it happened, are excluded from the unmet need group.

Women who were neither pregnant nor amenorrheic are classified separately. The first thing is to distinguish the fecund from the infecund. Non-pregnant women who have


212

been in a union for at least five years (without using contraception)<31> and who have not had a child are classified as infecund. The infecund group also includes non-pregnant women who have not menstruated in the past twelve weeks. Infecund women are excluded from the unmet need category. All women who have had a child in the last five years or who have not had a child but who have been married for fewer than five years are assumed to be fecund. We then classify fecund women who are not using contraception according to their reproductive intentions. Women who report wanting to postpone the next birth by at least two years from the time of the interview are considered to have an unmet need for spacing, and those who report wanting no more births at all are considered to have an unmet need for limiting. Women who want another births soon are excluded from the unmet need group.

In this analysis, we assume that all currently married women who are not pregnant or amenorrheic but who are fecund are sexually active. It is possible that the levels of unmet need are slightly over-estimated as some of these women are not actually sexually active. One way of solving this problem is to exclude all women who stated that they were sexually inactive from the unmet need group. But the reliability of data on sexually activity has been questioned (Westoff, 1988). Finally it is possible to combine all women with an unmet need for spacing with those with an unmet need for limiting; these women together form the overall estimate of unmet need for contraception in Tanzania.

6.3.1. 5.2.8. Estimates of Unmet Need and Demand for Contraception

The 1996 TDHS included 5411 currently married women of reproductive age (15-49 years). These women were used as a denominator in the calculations of unmet need. Figure 23 shows the distribution of currently married women according to the components of unmet need for family planning. The figure shows that 72.5 percent of currently married women were considered not to be having an unmet need of contraception due to the following facts:


213

  1. they are currently using contraceptive methods: 13.3 percent,
  2. they are infecund women: 11.7 percent,
  3. their current or most recent pregnancy was intended: 28.5 percent,
  4. they want a child soon: 19.1 percent.

The total unmet need for family planning among currently married women in Tanzania is 27.5 percent of which 15.7 percent have unmet need for spacing births and 11.8 percent want to limit births. These results are expected as spacing is the predominant family planning concern in sub-Saharan Africa. DHS data from other African countries have shown that estimates of unmet need range from 22 percent (Zimbabwe) to 40 percent (Togo), and that in all countries the unmet need for spacing exceeds that for limiting by a considerable margin (Westoff and Ochoa, 1991; Westoff and Moreno, 1992).

Figure 23: Estimates of unmet need for family planning

6.3.2. Socio-demographic Differentials in Unmet Need for Family Planning

Several variables have been noted to be associated with unmet need for contraception. Among the commonly mentioned covariates in the literature are age, education,


214-215

religion, the number of children, and place of residence (Westoff and Ochoa, 1991; Westoff and Moreno, 1992). Table 61 shows a list of ten covariates hypothesised to have a relationship with unmet need in Tanzania,<32> along with the percentages of currently married women with an unmet need for spacing, and for limiting. Women who had their first birth below 15 years of age, women who were married for the first time below 15 years of age, women over 35 years of age and women with five or more live births have an unmet need for limiting which is higher than that for spacing. It seems that many of these women have achieved their desired family size which is why they do not want any more children. For all other categories of women, the unmet need for spacing is greater than that for limiting (except for women whose partners have incomplete primary education). This study shows that number of surviving children is a main determining variable.

Table 63: Percentage of currently married women with unmet need for family planning

Characteristics

Unmet need to

Total unmet need

Space

Limit

Total

15.7

11.8

27.5

Place of Residence

Urban

13.2

7.3

20.5

Rural

16.0

8.8

24.8

Education

None

11.8

10.7

22.5

Incomplete primary

15.3

11.9

27.2

Complete primary

18.3

5.7

23.9

Secondary +

14.4

8.1

22.5

Partner’s education

None

13.2

9.7

23.0

Incomplete primary

12.4

13.1

25.5

Complete primary

18.3

6.3

24.6

Secondary +

13.2

6.8

20.1

Religion

Moslem

16.2

8.0

24.2

Catholic

15.3

8.9

24.2

Protestant

15.0

7.9

22.9

Other

14.6

10.1

24.7

Polygyny

Monogamous

15.9

8.0

23.9

Polygamous

14.0

9.6

23.6

Age at first birth

<15

10.5

13.7

24.2

15-17

15.7

10.6

26.2

18-19

18.3

8.0

26.3

20-21

18.5

7.9

26.4

22-24

16.2

9.6

25.8

25 +

11.2

5.2

16.4

Age at first sexual intercourse

<15

15.3

8.3

23.6

15-17

16.4

8.4

24.7

18-19

17.0

9.1

26.1

20-21

9.8

9.0

18.8

22-24

14.3

2.9

17.1

25 +

14.6

8.7

23.3

Age at first marriage

<15

11.3

12.7

24.0

15-17

16.3

8.8

25.1

18-19

18.7

7.2

25.9

20-21

14.9

8.6

23.5

22-24

13.7

5.4

19.1

25 +

10.4

5.7

16.1

Current age

15-19

18.5

1.7

20.2

20-24

22.5

3.5

26.1

25-29

20.5

4.1

24.6

30-34

14.0

8.9

22.9

35-39

11.6

15.3

26.9

40-44

5.3

20.3

25.6

45-49

2.9

12.1

15.1

Surviving children

None

4.3

0.4

4.6

1-2

19.9

2.9

22.8

3-4

17.8

7.3

25.1

5-6

13.5

15.2

28.7

7+

9.7

24.3

34.0

Source: calculated from 1996 TDHS

The highest levels of total unmet need (over 27.5 percent) are observed for women at the age of 35, and for those having more than 5 surviving children. In order for the sub-group of women with five or more children in Tanzania to decrease in future as a step towards further fertility decline, it is important to put emphasis on assisting these women to avoid unwanted births. One strategy to achieve this would clearly be to reduce the number of women in this category with an unmet need for family planning.


216

Women in rural areas have a high percentage of unmet need compared to their counterparts in the urban areas. Since Tanzanian family planning services are concentrated in urban areas, there is a great need for disseminating information about family planning. Associated services need to be organised so that they reach women in rural areas. Table 63 further shows that the unmet need is relatively high for women with some formal education. This pattern is very similar to that observed in other African countries which participated in the DHS (Westoff and Ochoa, 1991).

Younger women have a higher unmet need for spacing whereas older women have a higher unmet need for limiting. The levels of unmet need do not vary much in the various religious groups, and types of marriages. Finally, the bivariate analysis shows that women having late marriage, late first birth, and late first sexual intercourse have a lower level of unmet need than their counterparts in early ages. This may be a calendar effect given the time span between those events and the time of the interview.

6.3.3. 5.2.9. Multivariate Analysis of the Determinants of Unmet Need

The findings presented in Table 63 give the general picture of determinants of unmet need when each covariate is examined by itself. For a better understanding of the determinants of unmet need, it is necessary to look at all covariates hypothesised to affect unmet need in a multivariate perspective. As in the bivariate analysis, it is interesting to examine the determinants of unmet need for spacing separately from the unmet need for limiting. This implies that the response variable has three categories: unmet need for spacing, unmet need for limiting, and no unmet need. The multinomial Logit model is therefore the best statistical technique to apply. However, the coefficients from a multinomial Logit model are expected to be equal to coefficients obtained when performing two binary Logit models with the same reference category for the response variable (Begg and Gray, 1984).

In this analysis two binary Logit models have been fitted. The response variable for the first model has two categories: unmet need for spacing (coded 1) and no unmet need (coded 0). The response variable for the second model has also two categories: unmet need for limiting (coded 1) and no unmet need (coded 0). The ten predictor variables


217-218

used are the same in both models. The covariates used are listed in Table 61 along with the categories as introduced in the logistic regression.

The results for the binary logistic models are given in Table 64 . The type of place of residence, education of a woman, polygyny, and religion did not influence unmet need in either model. Two variables, age at first marriage, and the number of surviving children were found to be significant in both models, while partner’s education had an impact only on the unmet need for limiting. Age at first birth, age at first intercourse, and current age of a woman were found to affect only the unmet need for spacing. Two-way interactions were found to be highly significant.

Table 64: The coefficients of logistic regression models for the determinants of unmet need

Variable

Space

Limit

Place of Residence

Urban

RC

Rural

0.162

-0.094

Education

None

RC

Primary incomplete

0.093

0.242

Primary Complete

0.118

-0.102

Secondary +

0.196

0.433

Partner’s Education

None

RC

Primary incomplete

-0.128

0.108

Primary complete

-0.129

0.068

Secondary +

-0.167

-0.411*

Polygyny

Monogamous

RC

Polygamous

0.023

0.059

Religion

Moslems

RC

Catholics

-0.138

-0.116

Protestants

-0.194

-0.092

Other

-0.160

-0.112

Age at first birth

<15

RC

15-17

0.291

-0.246

18-19

0.415

-0.158

20-21

0.429

0.002

22-24

0.698**

0.222

25 +

0.478

0.019

Age at first intercourse

<15

RC

15-17

0.026

0.039

18-19

-0.139

0.302

20-21

-0.520*

0.319

22-24

-0.181

-0.173

25 +

0.073

-0.200

Age at first marriage

<15

RC

15-17

0.028

-0.163

18-19

0.303

-0.314

20-21

0.099

-0.248

22-24

0.088

-0.661*

25 +

0.514*

-0.464

Current age

15-19

RC

20-24

-0.295

0.273

25-29

-0.807***

-0.351

30-34

-1.499***

-0.076

35-39

-1.873***

0.188

40-44

-2.650***

0.654

45-49

-3.375***

0.139

Surviving children

None

RC

1-2

1.296***

1.311

3-4

1.590***

1.940*

5-6

2.053***

2.088*

7+

2.109***

2.235*

Constant

-1.982***

-2.302***

*** Highly significant at level .001

** Significant at level .01

* Significant at level .05

Source: calculated from 1996 TDHS

For a better understanding of the results given in Table 64 , it is helpful to construct a multiple classification analysis (MCA) table which shows the probabilities of having an unmet need for spacing, and for limiting. The details of the procedure used to calculate such a table are given in Retherford and Choe (1993) and are reproduced in Chapter 3. The results obtained are presented in Table 65 . The probabilities of having an unmet need for spacing and limiting births are estimated for all covariates used in the logistic regression. The probability of having unmet need for spacing is much higher than that for limiting across most of the covariates with nine exceptions; women with 0-7 years of education, women with partners who completed primary education, women who were first married before age 15, women of age 35 and more, and women with more than two living children. It is not surprising to find that women approaching the end of


219

their child bearing age have a higher probability of having an unmet need for limiting than for spacing (refer also to the bivariate analysis). Currently married women without a living child have negligible probabilities of having any unmet need as most of them want to have a child as soon as possible.

Women residing in rural areas have a higher unmet need for spacing and limiting than those living in urban areas. A woman residing in a rural area has a chance of having unmet need for spacing and limiting 2 percentage points greater than a woman living in an urban. The same pattern of probabilities is observed across educational levels of women. The highest chance of having unmet need for spacing was registered for women in the category ‘complete primary schooling’ although they had the lowest percentage of unmet need for limiting. Although the probabilities for age at first birth, marriage, and intercourse do not differ by a big margin, they do show a similar trend in so far as women’s age at first intercourse, marriage and birth goes up and the probability of having unmet need for either spacing or limiting goes down.

The age of women gives interesting results: In both models, the relationship between age and unmet need is found to be significant and linear. The probability of a woman to have an unmet need for spacing decreases with age while the probability of having unmet need for limiting increases with age. Women aged 15-19 years have the highest levels of unmet need for spacing births (25 percent) as well as the lowest unmet need for limiting births (5 percent). By age 35, the probability of having an unmet need for spacing is reduced to 0.13 percent, and for limiting to 0.17 percent. These results suggest that young women are highly in need of family planning services for spacing. In reality, they are disadvantaged in terms of family planning provision. They either shy away from clinics or they are turned away due to their young age and their marital status as explained earlier.


220-221

Table 65: Multiple classification analysis for the estimated percentages having unmet need

Characteristic

Probability of unmet need for

N

Spacing

Limiting

 

Place of Residence

Urban

15

10

994

Rural

17

12

3884

Education

None

13

15

1677

Primary incomplete

15

17

835

Primary Complete

20

7

2176

Secondary +

16

11

191

Partner’s Education

None

15

14

1067

Primary incomplete

13

17

1059

Primary Complete

20

8

2157

Secondary +

14

9

595

Polygyny

Monogamous

17

11

3471

Polygamous

15

13

1408

Religion

Moslem

17

12

1520

Catholic

16

12

1489

Protestant

16

11

1194

Other

16

12

675

Age at first birth

<15

12

19

344

15-17

17

12

1549

18-19

18

10

1413

20-21

17

10

835

22-24

16

11

506

25 +

10

9

232

Age at first intercourse

<15

17

12

713

15-17

17

11

1734

18-19

17

12

620

20-21

11

11

211

22-24

14

5

60

25 +

16

12

1541

Age at first marriage

<15

13

17

694

15-17

17

12

1896

18-19

20

10

1048

20-21

16

11

598

22-24

14

8

378

25 +

14

9

264

Current age

15-19

25

4

223

20-24

24

6

953

25-29

22

5

1103

30-34

15

10

911

35-39

13

17

705

40-44

7

27

547

45-49

4

20

436

Surviving children

None

6

1

117

1-2

20

4

1782

3-4

18

10

1434

5-6

14

18

880

7+

10

28

664

*** Highly significant at level .001

** Significant at level .01

* Significant at level .05

Source: calculated from 1996 TDHS

The probability of having unmet need for limiting increases as the number of living children rises. Unmet need for spacing decreases as the number of surviving children increases. Women with 7 and more children have a very high probability of having an unmet need for limiting (0.28), and for spacing (0.1). From the data it seems that after having between 3 and 4 surviving children, a Tanzanian woman finds there are enough and she would like to limit the number of children.

6.3.4. 5.2.10. Total Demand for Family Planning

As noted earlier, the total demand for family planning is the sum of those women currently using contraception and those with an unmet need. This produces the prevalence rate that would be achieved if all women who need family planning could use a method. Table 66 presents the total demand for family planning along with the percentage of women currently using a method (satisfied demand), and the percentage of women who want either to space or limit their births (unsatisfied demand). Overall, the demand for family planning in Tanzania is 40.8 percent. It seems that the demand for family planning in Tanzania is in line with other sub-Saharan African countries. Westoff and Ochoa (1991) have demonstrated that Kenya, Zimbabwe, and Botswana have the highest demand in the region (over 60 percent), and Mali has the lowest demand (28 percent).


222-223

Table 66: Percentage of currently married women according to demand for family planning and background characteristics

Characteristics

Current users<33>

Unmet need

Total<34>

Total

13.3

27.5

40.8

Place of Residence

Urban

26.6

20.5

47.1

Rural

9.8

24.8

34.6

Education

None

5.2

22.5

27.7

Incomplete primary

13.1

27.2

40.3

Complete primary

17.9

23.9

41.8

Secondary +

31.4

22.5

53.9

Partner’s Education

None

5.3

23

28.3

Incomplete primary

9.2

25.5

34.7

Complete primary

14.8

24.6

39.4

Secondary +

29.1

20.1

49.2

Religion

Moslem

17

24.2

41.2

Catholic

11.9

24.2

36.1

Protestant

17.4

22.9

40.3

Other

0.8

24.7

25.5

Polygyny

Monogamous

14.8

23.9

38.7

Polygamous

9.5

23.6

33.1

Age at first birth

<15

10.9

24.2

35.1

15-17

14.3

26.2

40.5

18-19

14.7

26.3

41.0

20-21

15.9

26.4

42.3

22-24

14.9

25.8

40.7

25 +

12.8

16.4

29.2

Age at first sexual intercourse

<15

9.6

23.6

33.2

15-17

14.6

24.7

39.3

18-19

15.8

26.1

41.9

20-21

21.6

18.8

40.4

22-24

22.9

17.1

40

25 +

11.2

23.3

34.5

Age at first marriage

<15

8.7

24

32.7

15-17

13.3

25.1

38.4

18-19

13.9

25.9

39.8

20-21

16.6

23.5

40.1

22-24

15

19.1

34.1

25 +

12.4

16.1

28.5

Current age

15-19

4.5

20.2

24.7

20-24

12.7

26.1

38.8

25-29

14.4

24.6

39

30-34

14

22.9

36.9

35-39

15.9

26.9

42.8

40-44

17.1

25.6

42.7

45-49

9.2

15.1

24.3

Surviving children

None

0.5

4.6

5.1

1-2

12.8

22.8

35.6

3-4

16.3

25.1

41.4

5-6

15.4

28.7

44.1

7+

16.0

34.0

50.0

Source: adopted from Table 57 and Table 63

Women residing in urban areas, have higher demand than their counterparts residing in rural areas. The demand for family planning increases as the education of the woman or her partner's education rises. It is interesting to note that women with secondary education or above have a demand of more than 26 percentage difference (53.9 percent) by comparison with women without formal education (27.7 percent). There is also a positive relationship between the number of living children and the demand for family planning. The demand is lowest (5.1 percent) for women without a child and highest (50 percent) for women with 7 children.

It can be concluded therefore that the analysis indicates that the type of place of residence, a woman's education, her partner's education and the number of living children a woman has are highly associated with the demand for family planning in the expected direction. The demand for family planning is observed to be highest for Moslems, women in monogamous marriages, women who had their first birth between age 20-21, first sexual intercourse between 18-19, first marriage between age 20-21, and women in the age group 35-39 years.


224

6.4. Concluding Remarks

Knowledge of contraception is relatively high in Tanzania, particularly among currently married women. The TDHS data show that 88 percent of currently married women know at least one method. However, the level of knowledge is lower for never married women (only 70 percent know at least one method). Whilst almost all sub-groups of currently married women demonstrate a high level of knowledge, a majority of women in some sub-groups of never married women do not know any family planning method. Specifically, unmarried women with no education, and without any living child have low levels of knowledge.

Use of contraception is very low in Tanzania. According to 1996 TDHS data, the percentages of women age 15-49 currently using any method are 13.3 and 5.6 for currently married women and never married women respectively. The percentage of users is as low as 4.5 for some sub-groups of currently married women, and 2.4 for some sub-groups of never married women. It seems that contraception in Tanzania is mainly used for spacing purposes. Methods most widely used are the pill and injections. The common traditional methods are periodic abstinence and withdrawal. It is interesting to note that the percentage of never married women using the pill is somehow equal to those using periodic abstinence. This has been suggested to be a consequence of the lack of services that modern methods require. In most cases, young unmarried women cannot get services from governmental family planning clinics. This prompts these women to choose traditional methods, which are considered to have low use-effectiveness. However, in general the main reasons for not using contraceptives include a need for more children, and opposition by the respondent to contracept.

A multivariate analysis of the determinants of current contraceptive use among non-pregnant women showed that for currently married women, 8 out of 11 variables chosen influenced the use of contraception significantly. Currently married Catholic women residing in rural areas, with no schooling, married to a man with no schooling, involved in a polygamous marriage, or without a living child are less likely to use contraception than other women.

The multivariate analysis for never married women showed that the use of contraceptive is less frequent for those residing in rural areas, with no schooling or incomplete


225

primary education, and no living child. The variables influencing the use of contraception for never married women were also found to influence currently married women. It is interesting to note that unmarried women are less likely to use contraception as they become older, unlike currently married women who are more likely to contracept as they become older. The association is non-linear. It should be stressed that the majority of never married women were adolescents. A study conducted in Dar es Salaam referral hospital (Muhimbili Medical Centre) found that the majority of illegally induced abortions occurred to young and unmarried women (Justesen et al., 1992).

These women are still not accessible to family planning information and services. This problem needs to be addressed urgently by the National Family Planning Programme because premarital sexuality is very prevalent in Tanzania, and premarital fertility and abortion rates are rising as found in Chapter 4.

The total unmet need is estimated to be 27.5 percent among currently married women. Most of these have an unmet need for spacing births (15.7 percent) rather than for limiting births (11.8 percent). The sum of the unmet need and current contraceptive use (13.3 percent) gives a total demand for contraception of 40.8 percent. It seems, therefore, that Tanzania has a significant demand for family planning according to sub-Saharan African standards.


Fußnoten:
<28> This is due to a lack of knowledge caused by little or no education and the fact that these women are residing in rural areas.
<29> Wanyambo is a tribe in Karagwe district in the Kagera region.
<30>

include women who are currently pregnant who did not intended to be pregnant at the time thay became pregnant.

<31> Data on the timing of contraceptive use was not collected in the 1996 TDHS. Since contraceptive prevalence is still very low in Tanzania, we assume all ever-users of contraception have used a method in the five years prior to the interview date.
<32> The variables used here are similar to those used in the analysis of current users of contraception presented in section 6.2.
<33> From Table 57
<34> From Table 63


[Titelseite] [Abkürzungsverzeichnis] [1] [2] [3] [4] [5] [6] [7] [8] [Bibliographie] [Anhang] [Selbständigkeitserklärung]

© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.

DiDi DTD Version 1.1
a subset from ETD-ML Version 1.1
Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML - Version erstellt am:
Mon Jan 17 19:31:05 2000