Family Planning Performance in India: 1992

Abstract

This article analyses family planning performance in India based on a composite performance index which is newly developed for this purpose. The analysis suggests that there has been an improvement in family planning performance in the country over time. However, the performance remains poor, and there is substantial scope for further improvement. The analysis also reveals significant variation in family planning performance across India’s states/union territories and districts. The classification modelling exercise undertaken in this article reveals that districts can be grouped into eight clusters in terms of family planning performance. The districts are essentially different in different clusters on both the demand met for family planning and the composition of the demand met. This article emphasises that family planning in the country must be promoted as a development strategy rather than an intervention to limit births.

Keywords

India states union territories family planning performance composite index

Introduction

Organised family planning efforts in India date back to 1952 when the country launched the first official family planning programme of the world. In its initial phase, the main appeal of family planning was based on considerations of health and welfare of the family, especially children and women. Family planning was deemed necessary to secure better health of the mother and better care and upbringing of the child (Chaurasia & Singh, 2014). As the programme progressed, it shifted its attention towards fertility reduction to control population growth through a strong focus on birth limitation. In the early days of organised family planning efforts in India, the only source of family planning-related information was the programme service statistics and indicators like equivalent sterilisations and couples effectively protected were used to measure family planning performance (Chaurasia, 1985; Government of India, 1990). The first nationally representative survey of family planning practices in the country was conducted in 1970 which revealed that, despite almost two decades of organised family planning efforts, only around 14% currently married women aged 15–44 years or their husband were using a family planning method, while less than 10% were using a modern family planning method (Operations Research Group, 1970). The second all India survey, carried out in 1980, however, revealed that 35% of currently married women aged 15–44 years or their husband were using a family planning while more than 28% were using a modern family planning method (Khan & Prasad, 1980). In 1992, Government of India launched the National Family Health Survey (NFHS) Programme and the contraceptive prevalence rate (CPR), defined as the proportion of currently married women aged 15–49 years or their husband using a family planning method, became the basis for measuring family planning performance. The NFHS 1992–1993, revealed a CPR of 40% with substantial inter-state variations (Government of India, 1995) whereas NFHS 1998–1999 revealed that the CPR in the country had increased to almost 45% (Government of India, 2000). The NFHS 2005–2006 estimated a CPR of 55% (Government of India, 2007) but the NFHS 2015–2016 estimated that the CPR in the country had decreased to around 53% (Government of India, 2017). Figure 1 shows the trend in all methods and modern methods CPR in India.

Figure 1.

Source: ORG (1970); Khan and Prasad (1980); GoI (1997; 2000; 2007; 2017).

The popularity of CPR as a measure of family planning performance rests with its strong inverse relationship with total fertility rate (TFR) based on cross-country data (Bongaarts, 1978; Bongaarts & Potter, 1983; Jain, 1997; Ross & Mauldin, 1996; Stover, 1998; Tsui, 2001; United Nations, 2020). Tsui (2001) has shown that an increase of 15 percentage points in CPR reduces TFR by one birth per woman. There are, however, studies that show inconsistency between CPR and TFR in many countries (United Nations, 2020). Using state-level data from India, Srinivasan (1988) has argued that as one goes down the level of aggregation, variation in family planning use explains less and less of the variation in TFR. Chaurasia (2004), using development block-level data from Madhya Pradesh, has observed that variation in family planning use explains only around 20% of the variation in total marital fertility rate. Chaurasia (2000) has also observed that, at the individual level, use of family planning methods is positively related to the number of children ever born.

The CPR, however, is not an appropriate indicator to measure family planning performance for many reasons. First, the upper limit of CPR is difficult to establish as a substantial proportion of currently married women aged 15–49 years do not practice family planning because they either want a child or they are pregnant or they or their husband are sterile, and this proportion varies across different population groups. The CPR, for all practical purposes, therefore, can never be 100%. It is, therefore, not possible to construct a universal scale to measure family planning performance based on CPR. Second, CPR is based on the number of family planning users only and does not take into account the composition of family planning use by different family planning methods. Considering the composition of family planning use by different family planning methods is important for measuring family planning performance as family planning is an integral component of the family-building strategy. Family planning needs of a couple vary by different phases of the family-building process and are conditioned by such factors as personal circumstances, individual knowledge, family size desires and preferences and availability of different family planning methods as well as effectiveness of the method in regulating fertility. Family planning allows couples to anticipate and attain their desired number of children by spacing and limiting births and, therefore, has implications for the health and well-being of women and outcome of pregnancy. The composition of family planning use also reflects the basic orientation of family planning efforts, particularly, organised family planning efforts, availability of different family planning methods and preferences and choices of potential family planning users. It has, therefore, been emphasised that family planning performance measurement should not be limited to counting how many currently married women aged 15–49 years or their husband are using a family planning method. Rather, it must also take into account the range and types of family planning methods being used (United Nations, 2019).

In this article, we propose a new composite performance index that takes into account both the extent of the use of family planning methods and composition of family planning use to measure family planning performance in India and in its constituent states, union territories and districts during the period 1992 through 2016. Analysis of family planning performance at the district level is important as family planning services in India, especially, under the official family planning efforts, are essentially planned and delivered at the district level, although, family planning-related policies and family planning priorities are set at the national level and customised at the state level. Measuring family planning performance at the district level is, therefore, critical to improving family planning performance in the context of meeting the very diverse and dynamic family planning needs of couples.

The article is organised as follows. The next section develops the composite family planning performance index. Section 3 analyses family planning performance in India and in its states/union territories during the 25 years period between 1992 and 2016 based on data from different rounds of NFHS. Inter-district variation in family planning performance, based on data from 2015–2016 round of NFHS, are analysed in Section 4. The fifth section presents findings of the classification modelling exercise that classifies districts in terms of family planning performance based on the met demand for family planning and composition of the met demand. The last section summarises findings of the analysis and discusses their policy and programme implications.

Family Planning Performance Index

Family planning efforts are essentially directed towards meeting family planning needs of couples. The family planning needs of couples are very diverse because couples need different family planning methods at different stages of their family-building process and to achieve their reproductive goals. Because of the diversity in the family planning needs of couples, the cafeteria approach is adopted for family planning services delivery. In this approach, a range of family planning methods are made available so that couples can chose and practice family planning method according to their choice. The diversity in family planning needs also implies that family planning performance should be measured not only in terms of the number of couples practising family planning but also in terms of the composition of family planning use by different family planning methods or the method mix. This means that family planning performance should be measured on a two-dimensional space. One dimension of this space should measure performance in terms of number of couples practising family planning while the other should measure performance in terms of the method mix. Since different couples need different family planning methods to achieve their reproductive goals, it is argued that the method mix should be a balanced one. If the method mix is not balanced or family planning use is dominated by one- or two-family planning methods, then it suggests that family planning needs of a substantial proportion of couples remain unmet.

Let c denotes an index that reflects the extent of family planning use by couples and q denotes an index that reflects the method mix. Then the above argument implies that an index p reflecting family planning performance should be conceptually linked to c and q through an aggregation function f. Or

p = f ​ (c, q)

(1)

Equation (1) describes a two-dimensional framework which is an improvement over the existing one-dimensional approaches of measuring family planning performance such as CPR and met demand for family planning (FP2020, 2013). It highlights the importance of an appropriate method mix to ensure meeting the diverse family planning needs of couples.

Equation (1) requires specifying the aggregation function f and defining indexes c and q for its operationalisation. The indexes c and q should be constructed in a manner that variation in the index c is independent of the variation in the index q, but both c and q must covary with the index p. Mutual independence of c and q is important as it ensures that improvement in the index p is contingent upon the improvement in both the indexes.

The simplest and the commonly used aggregation function to combine two indexes is the simple arithmetic mean or average. However, an undesirable feature of the simple arithmetic mean as the aggregation function is its additive compensability which means that the low value of an index is compensated fully by the high value of the other index (OECD, 2000). It is argued that any measure of performance involving more than one dimension should be more sensitive or biased to that dimension in which the performance is poorer than the dimension in which the performance is better. This biasedness of the performance measure provides an impetus to focus on poor performing dimensions to improve overall performance.

Given the limitation of simple arithmetic mean as an aggregation function, the other alternative is to use the weighted power or generalised mean (Bullen, 2003; Stanislav, 2009). Anand and Sen (1997) have used unweighted power mean to construct a measure of multidimensional poverty whereas, Chaurasia (2018) has used weighted power mean to measure the level of development in Indian villages.

Using the weighted power mean as the aggregation function, a composite family planning performance index, p_α, may be constructed by calculating the weighted power mean of order α of the index c and the index q. Or

p_{α} = {(w_{c} c^{α} + w_{q} q^{α})}^{1 / α}

(2)

where w_c is the weight assigned to index c, and w_q is the weight assigned to the index q, such that w_c + w_q = 1. If equal importance is given to both the indexes, then w_c = w_q = 0.5 and Equation (2) reduces to

p_{α} = {(c^{α} + q^{α}) / 2}^{1 / α}

(3)

The parameter α in Equation (3) decides the relative importance of indexes c and q in determining the index p_α. When α = 1, p_α is nothing but the simple arithmetic mean of c and q. When α < 1, relatively more weight is given to that of the two indices which is lower than the index which is higher. For example, when c = 0.64 and q = 0.27, then p₁ = 0.455 when α = 1, but when α = 1/3, p_1/3 = 0.423. The relative contribution of c in p₁ is 70% but only 57% in p_1/3. On the other hand, the relative contribution of q in p₁ is only 30% but 43% in p_1/3 which means that p_1/3 is more sensitive or biased to the index which is low than the index p₁. There is, however, an inescapable arbitrariness in selecting α. Theoretically, α can be any value <1. When α = 1/3, the index p_α accords four times more importance to that index which has a lower value than the index which has a higher value. When, α = 1/5, the relative importance of the component having the lower performance is 16 times the importance of the component having the higher performance. For α < 1/5, the difference turns out to be even wider. However, as performance improves, the difference in the performance of the two dimensions narrows down and when performance in the two dimensions becomes the same, the difference in the relative contribution of the two dimensions in determining the composite performance index turns out to be zero.

Choosing α = 1/3, the composite family planning performance index, p, may now be defined as

p = {(c^{1 / 3} + q^{1 / 3}) / 2}^{3}

(4)

The properties of the index p are the same as the properties of the human poverty index proposed by Anand and Sen (1997). For example, it can be shown that p always lies between the minimum and maximum values of indexes c and q. Similarly, it can be shown that if indexes c and q are doubled, the index p will also be doubled. The index p also increases monotonically with the increase in either c or q but with a diminishing rate of improvement.

It now remains to define the index c that measures the performance in terms of the use of family planning methods and the index q that measures the composition of family planning use by different family planning methods. The index c may be defined in terms of the met demand for modern family planning methods which is one of the core indicators identified by the FP2020 initiative to track family planning progress (FP2020, 2013). The met demand for modern family planning methods is defined as the proportion of currently married women or their partner, desiring either to have no additional child or to postpone the next child, using a modern family planning method (FP2020, 2013). This definition assumes that women or their partner using a traditional family planning method have an unmet need for modern family planning methods.

The met demand for modern family planning methods may be divided into the met demand for modern spacing methods and the met demand for permanent methods of family planning. This distinction is important from the performance measurement perspective as the context of using a modern spacing method is essentially different from the context of using a permanent method. Permanent methods of family planning, like female and male sterilisation, are irreversible so that they cannot be used when the family building process is incomplete or when couples want to delay or postpone their next child.

An index of the met demand for modern spacing methods, c_s, may be constructed from the prevalence of modern spacing methods such as intra-uterine devices (IUD), condom, and pill and the unmet need for spacing births. Let PR_s denotes the prevalence of modern spacing methods, PR_t denotes the prevalence of traditional methods and UN_s denotes the unmet need for spacing births. Then, then the index c_s is defined as

c_{s} = \frac{P R_{s}}{P R_{s} + P R_{t} + U N_{s}}

(5)

Here, it is assumed that the use of traditional methods reflects the unmet need for modern spacing methods to space births. It is obvious that the index c_s ranges between 0 and 1 and the higher the index the higher the met demand for modern spacing methods.

Arguing in the same manner, an index of the met demand for permanent methods c_p, may be constructed from the prevalence of permanent methods such as female and male sterilisation and the unmet need for limiting births. If PR_p denotes the prevalence of permanent methods and UN_l denotes the unmet need for limiting births, then the index c_p is defined as

c_{p} = \frac{P R_{p}}{P R_{p} + U N_{l}}

(6)

Like the index c_s, the index c_p also varies between 0 and 1 and higher the index the higher the met demand for permanent methods.

The indexes c_s and c_p may now be combined using the power mean of order 1/3 and assigning equal weights to the two indexes to obtain the index c of the met demand for family planning:

c = {(c_{s}^{1 / 3} + c_{p}^{1 / 3}) / 2}^{3}

(7)

It is obvious that index c varies between 0 and 1 and the higher the index c the higher the met demand for family planning and vice versa.

On the other hand, the method mix serves the basis for constructing the index q. The method mix is also one of the core indicators identified by the FP2020 initiative to track family planning progress (FP2020, 2013). There is, however, no ideal method mix but, it is argued, that the method mix should not be skewed or dominated by one family planning method as providing access to a wide range of family planning methods is an important principle of rights-based family planning and an important component of the quality of family planning efforts (Ross & Stover, 2013). A skewed method mix may also reflect user preference for a particular method, but it is hard to believe that a skewed method mix can meet diverse family planning needs of couples at different stages of their family building process (Bertrand et al., 2014). User preference may also be influenced by the information provided and advice given by service providers so that method skew reflects provider bias. In case of organised family planning efforts, method skew may reflect the orientation of these efforts. When family planning efforts are directed primarily towards fertility reduction through birth limitation, for example, the method mix may be skewed towards permanent methods. Method skew may also reflect poor performance of family planning efforts in terms of counselling or in overcoming a host of exogenous factors that have a strong impact on the use of different methods. It may, therefore, be argued that the higher the method skew the poorer the family planning performance in meeting diverse family planning needs of couples and vice versa.

Recently, Chaurasia (2020) has developed a method skew index, s, based on the proportionate distribution of different family planning methods among family planning users. This index is defined as

s = \sqrt{\frac{\sum_{i = 1}^{n} x_{i}^{2} - (1 / n)}{1 - (1 / n)}} f o r n > 1. s = 1 w h e n n = 1

(8)

where x_i is the ratio of the users of method i to total family planning users and n is the total number of family planning methods available. The index is independent of the number of family planning methods available and ranges from 0 to 1, the higher the index the higher the method skew. Since method skew is an undesired feature of family planning performance, the index q reflecting the balance in the method mix may be defined as

q = 1 - s = 1 - \sqrt{\frac{\sum_{i = 1}^{n} x_{i}^{2} - (1 / n)}{1 - (1 / n)}}, f o r n > 1. q = 0 w h e n n = 1

(9)

The index q also varies from 0 to 1 and the higher the index the more balanced the method mix and hence the better the family planning performance in meeting the family planning needs of couples. When q = 0, entire family planning use is confined to one method only so that family planning performance may be rated as the poorest in meeting the diverse family planning needs.

Using district-level data available through NFHS 2015–2016, the simple zero order correlation coefficient between the index c and the index q is found to be 0.116. This means that inter-district variance in the index c or in the index q accounts for only about 1.3% of the inter-district variance, respectively, in the index q or in the index c. On the other hand, the simple zero order correlation coefficient between the index p and the index q is 0.844 while that between the index p and the index c is 0.615. These correlations provide empirical justification of the index p as a measure of family planning performance. The definition of indicators used in the construction of the index p is given in Table 1.

Table 1.

Definition of Indicators Used in the Analysis.

Indicator	Definition
Prevalence of modern spacing methods (PR_s)	Proportion of currently married women aged 15–49 years or their husband using either intra-uterine devices (IUD) or pill or condom.
Prevalence of permanent methods (PR_p)	Proportion of currently married women aged 15–49 years or their husband using either female or male sterilisation.
Prevalence of traditional methods (PR_t)	Proportion of currently married women aged 15–49 yeas or their husband using a traditional method.
Unmet need for spacing births (UN_s)	Proportion of currently married women aged 15–49 years who (a) are not pregnant and not postpartum amenorrhoeic, are considered fecund, and want to postpone their next birth for 2 or more years, but are not using a contraceptive method, or (b) have a mistimed or unwanted current pregnancy or (c) are postpartum amenorrhoeic and their last birth in the last two years was mistimed or unwanted.
Unmet need for limiting births (UN_s)	Proportion of currently married women aged 15–49 years who (a) are not pregnant and not postpartum amenorrhoeic, are considered fecund, and want to stop childbearing altogether, but are not using a contraceptive method, or (b) have a mistimed or unwanted current pregnancy or (c) are postpartum amenorrhoeic and their last birth in the last two years was mistimed or unwanted.

Source: Government of India (2017).

Based on the composite performance index p, family planning performance may be classified as very poor if 0 ≤ p < 0.300. Similarly, family planning performance can be rated as poor if 0.300 ≤ p < 0.550; average if 0.550 ≤ p < 0.750; good if 0.750 ≤ p < 0.900 and very good if p ≥ 0.900. The same scale may be used to rate the performance in terms of the met demand for family planning and in terms of the composition of the met demand. The index p, in combination with indexes c and q, constitutes a comprehensive family planning performance measurement framework that can be instituted to monitor family planning performance (Table 1).

Family Planning Performance in India

Estimates of the prevalence of different family planning methods along with estimates of unmet need for spacing and limiting births are available from different rounds of the NFHS (Table 2). Estimates available through the first three rounds of the survey are limited to the country and its constituent states/union territories. The fourth round of the survey, however, provides these estimates for 640 districts of the country as they existed at the time of the 2011 population census. These method-specific prevalence rates and estimates of unmet need for spacing and limiting births constitute the database for the present analysis (Table 2).

Table 2.

Family Planning Performance in India, 1992–1993 through 2015–2016.

Particulars	Period
Particulars	1992–1993	1998–1999	2005–2006	2015–2016
CPR	40.7	48.2	56.3	53.5
CPRm	36.5	42.8	48.5	47.8
Method-specific prevalence
Female sterilisation	27.4	34.1	37.3	36.0
Male sterilisation	3.5	1.9	1.0	0.3
IUD	1.9	1.6	1.7	1.5
Oral pill	1.2	2.1	3.1	4.1
Condom	2.4	3.1	5.2	5.6
Other modern methods	0.1	0.0	0.2	0.3
Traditional methods	4.2	5.4	7.8	5.7
Unmet need for spacing	12.2	8.3	6.1	5.6
Unmet need for limiting	8.1	7.8	7.8	7.2
Quality index (q)	0.295	0.242	0.272	0.288
Coverage index (c_s)	0.257	0.338	0.429	0.502
Coverage index (c_p)	0.792	0.822	0.831	0.834
Coverage index (c)	0.475	0.545	0.608	0.654
Performance index (p)	0.378	0.373	0.417	0.446

Source: The author.

Data available through NFHS suggest that, although there is an improvement in family planning performance in the country during 1992–1993 through 2015–2016, the performance remains poor. The performance index c increased throughout the period under reference, but the performance index q decreased sharply between 1992–1993 and 1998–1999 and then increased only gradually so that it was lower in 2015–2016 compared to that in 1992–1993. This means that the preference for a particular method over other available methods has increased with time. Moreover, there remains a big gap between the met demand for modern spacing methods and the met demand for permanent methods, although the gap has narrowed down over time (Figure 2).

Figure 2.

Source: The author.

Family planning performance is found to vary widely across states and union territories, being the best in Sikkim but the poorest in Andhra Pradesh during 2015–2016. There is, however, no state/union territory where family planning performance may be rated as good or very good whereas, in 28 of the 36 states/union territories, the performance remains either poor or very poor in terms of the performance index p. The inter-state/union territory variation in both met demand for modern family planning methods and the composition of the met demand has contributed to the variation in family planning performance. There is, however, no state/union territory where performance in both met demand for modern family planning methods, measured in terms of the index c, and composition of the met demand, measured in terms of the index q, is found to be very good. There are 17 states where the performance in terms of the composition of the met demand for modern family planning methods is rated as very poor whereas there is only one state where the met demand for modern family planning method is very poor (Table 3).

Table 3.

Indicators of Family Planning Performance in States/Union Territories of India.

State/Union Territory	1992–1993	1998–1999	2005–2006	2015–2016
	Index p
Andaman and Nicobar Islands	na	na	na	0.376
Andhra Pradesh and Telangana	0.314	0.279	0.235	0.166
Arunachal Pradesh	0.469	0.46	0.505	0.525
Assam	0.371	0.448	0.459	0.492
Bihar and Jharkhand	0.285	0.235	0.315	0.264
Chandigarh	na	na	na	0.628
Dadra and Nagar Haveli	na	na	na	0.339
Daman and Diu	na	na	na	0.350
Delhi	0.659	0.639	0.635	0.606
Goa	0.365	0.364	0.439	0.442
Gujarat	0.385	0.402	0.450	0.408
Haryana	0.478	0.480	0.562	0.577
Himachal Pradesh	0.520	0.455	0.569	0.510
Jammu and Kashmir	0.474	0.468	0.555	0.589
Karnataka	0.299	0.273	0.290	0.223
Kerala	0.384	0.306	0.346	0.264
Lakshadweep	na	na	na	0.352
Madhya Pradesh and Chhattisgarh	0.364	0.334	0.374	0.365
Maharashtra	0.422	0.405	0.460	0.437
Manipur	0.508	0.453	0.522	0.395
Meghalaya	0.403	0.422	0.415	0.495
Mizoram	0.374	0.425	0.488	0.555
Nagaland	0.430	0.454	0.483	0.499
Odisha	0.294	0.313	0.439	0.507
Puducherry	na	na	na	0.262
Punjab	0.555	0.604	0.615	0.613
Rajasthan	0.303	0.361	0.426	0.443
Sikkim	na	0.519	0.626	0.678
Tamil Nadu	0.341	0.288	0.294	0.250
Tripura	0.44	0.490	0.534	0.484
Uttar Pradesh and Uttarakhand	0.383	0.375	0.443	0.480
West Bengal	0.383	0.459	0.488	0.595
Andhra Pradesh	na	na	na	0.133
Bihar	na	na	0.295	0.236
Chhattisgarh	na	na	0.368	0.370
Jharkhand	na	na	0.369	0.339
Madhya Pradesh	na	na	0.373	0.363
Telangana	na	na	na	0.207
Uttar Pradesh	na	na	0.433	0.476
Uttarakhand	na	na	0.591	0.570
	Index q
Andaman and Nicobar Islands	na	na	na	0.205
Andhra Pradesh and Telangana	0.206	0.124	0.073	0.034
Arunachal Pradesh	0.505	0.406	0.443	0.539
Assam	0.449	0.453	0.486	0.439
Bihar and Jharkhand	0.236	0.170	0.229	0.153
Chandigarh	na	na	na	0.521
Dadra and Nagar Haveli	na	na	na	0.191
Daman and Diu	na	na	na	0.224
Delhi	0.597	0.552	0.538	0.529
Goa	0.26	0.261	0.343	0.353
Gujarat	0.238	0.230	0.283	0.257
Haryana	0.384	0.319	0.388	0.403
Himachal Pradesh	0.438	0.302	0.356	0.366
Jammu and Kashmir	0.420	0.387	0.479	0.518
Karnataka	0.159	0.105	0.097	0.063
Kerala	0.271	0.161	0.186	0.106
Lakshadweep	na	na	na	0.335
Madhya Pradesh and Chhattisgarh	0.298	0.192	0.194	0.176
Maharashtra	0.28	0.226	0.251	0.220
Manipur	0.587	0.483	0.679	0.657
Meghalaya	0.431	0.582	0.522	0.501
Mizoram	0.186	0.243	0.316	0.476
Nagaland	0.583	0.499	0.594	0.566
Odisha	0.217	0.188	0.297	0.405
Puducherry	na	na	na	0.074
Punjab	0.442	0.495	0.482	0.456
Rajasthan	0.214	0.227	0.268	0.272
Sikkim	na	0.508	0.628	0.686
Tamil Nadu	0.199	0.121	0.099	0.072
Tripura	0.464	0.407	0.476	0.382
Uttar Pradesh and Uttarakhand	0.425	0.366	0.429	0.448
West Bengal	0.346	0.366	0.388	0.472
Andhra Pradesh	na	na	na	0.019
Bihar	na	na	0.209	0.133
Chhattisgarh	na	na	0.203	0.181
Jharkhand	na	na	0.288	0.202
Madhya Pradesh	na	na	0.19	0.175
Telangana	na	na	na	0.059
Uttar Pradesh	na	na	0.427	0.450
Uttarakhand	na	na	0.446	0.448
	Index c
Andaman and Nicobar Islands	na	na	na	0.623
Andhra Pradesh and Telangana	0.454	0.529	0.545	0.469
Arunachal Pradesh	0.434	0.519	0.572	0.510
Assam	0.303	0.444	0.433	0.548
Bihar and Jharkhand	0.34	0.315	0.420	0.419
Chandigarh	na	na	na	0.749
Dadra and Nagar Haveli	na	na	na	0.547
Daman and Diu	na	na	na	0.516
Delhi	0.724	0.734	0.743	0.690
Goa	0.494	0.491	0.551	0.546
Gujarat	0.584	0.645	0.674	0.609
Haryana	0.586	0.687	0.782	0.795
Himachal Pradesh	0.611	0.652	0.854	0.688
Jammu and Kashmir	0.532	0.560	0.639	0.667
Karnataka	0.503	0.562	0.645	0.542
Kerala	0.526	0.520	0.578	0.532
Lakshadweep	na	na	na	0.368
Madhya Pradesh and Chhattisgarh	0.439	0.532	0.640	0.655
Maharashtra	0.606	0.661	0.761	0.763
Manipur	0.436	0.424	0.391	0.214
Meghalaya	0.376	0.295	0.324	0.489
Mizoram	0.660	0.682	0.714	0.641
Nagaland	0.306	0.411	0.387	0.438
Odisha	0.387	0.483	0.619	0.625
Puducherry	na	na	na	0.639
Punjab	0.686	0.727	0.770	0.802
Rajasthan	0.415	0.538	0.638	0.673
Sikkim	na	0.529	0.624	0.669
Tamil Nadu	0.539	0.566	0.652	0.599
Tripura	0.417	0.583	0.596	0.602
Uttar Pradesh and Uttarakhand	0.344	0.384	0.456	0.513
West Bengal	0.423	0.566	0.605	0.738
Andhra Pradesh	na	na	na	0.427
Bihar	na	na	0.401	0.380
Chhattisgarh	na	na	0.605	0.659
Jharkhand	na	na	0.464	0.527
Madhya Pradesh	na	na	0.648	0.655
Telangana	na	na	na	0.501
Uttar Pradesh	na	na	0.439	0.503
Uttarakhand	na	na	0.764	0.713
	Index c_s
Andaman and Nicobar Islands	na	na	na	0.444
Andhra Pradesh and Telangana	0.18	0.253	0.267	0.182
Arunachal Pradesh	0.306	0.381	0.510	0.464
Assam	0.135	0.304	0.298	0.564
Bihar and Jharkhand	0.149	0.134	0.255	0.237
Chandigarh	na	na	na	0.673
Dadra and Nagar Haveli	na	na	na	0.365
Daman and Diu	na	na	na	0.339
Delhi	0.701	0.681	0.709	0.726
Goa	0.277	0.304	0.379	0.462
Gujarat	0.360	0.430	0.483	0.475
Haryana	0.397	0.512	0.711	0.719
Himachal Pradesh	0.394	0.447	0.801	0.610
Jammu and Kashmir	0.346	0.437	0.546	0.557
Karnataka	0.255	0.303	0.426	0.286
Kerala	0.264	0.260	0.328	0.284
Lakshadweep	na	na	na	0.158
Madhya Pradesh and Chhattisgarh	0.202	0.315	0.453	0.477
Maharashtra	0.393	0.461	0.622	0.637
Manipur	0.307	0.283	0.334	0.287
Meghalaya	0.168	0.246	0.236	0.47
Mizoram	0.437	0.488	0.570	0.589
Nagaland	0.291	0.329	0.424	0.422
Odisha	0.158	0.247	0.180	0.506
Puducherry	na	na	na	0.409
Punjab	0.551	0.601	0.707	0.705
Rajasthan	0.192	0.354	0.495	0.514
Sikkim	na	0.427	0.610	0.720
Tamil Nadu	0.309	0.339	0.455	0.372
Tripura	0.222	0.460	0.522	0.532
Uttar Pradesh and Uttarakhand	0.222	0.269	0.359	0.422
West Bengal	0.189	0.353	0.404	0.620
Andhra Pradesh	na	na	na	0.135
Bihar	na	na	0.228	0.203
Chhattisgarh	na	na	0.389	0.472
Jharkhand	na	na	0.323	0.343
Madhya Pradesh	na	na	0.468	0.479
Telangana	na	na	na	0.226
Uttar Pradesh	na	na	0.341	0.410
Uttarakhand	na	na	0.725	0.695
	Index c_p
Andaman and Nicobar Islands	na	na	na	0.844
Andhra Pradesh and Telangana	0.921	0.955	0.971	0.964
Arunachal Pradesh	0.594	0.688	0.638	0.560
Assam	0.573	0.621	0.603	0.533
Bihar and Jharkhand	0.649	0.612	0.644	0.674
Chandigarh	na	na	na	0.830
Dadra and Nagar Haveli	na	na	na	0.783
Daman and Diu	na	na	na	0.746
Delhi	0.748	0.790	0.778	0.656
Goa	0.803	0.742	0.769	0.639
Gujarat	0.887	0.923	0.908	0.766
Haryana	0.827	0.897	0.859	0.876
Himachal Pradesh	0.895	0.913	0.910	0.772
Jammu and Kashmir	0.775	0.706	0.741	0.789
Karnataka	0.873	0.939	0.928	0.917
Kerala	0.922	0.912	0.931	0.895
Lakshadweep	na	na	na	0.713
Madhya Pradesh and Chhattisgarh	0.813	0.831	0.874	0.874
Maharashtra	0.886	0.911	0.919	0.904
Manipur	0.597	0.605	0.453	0.155
Meghalaya	0.709	0.349	0.432	0.508
Mizoram	0.949	0.923	0.881	0.696
Nagaland	0.322	0.506	0.352	0.455
Odisha	0.771	0.834	0.782	0.761
Puducherry	na	na	na	0.943
Punjab	0.842	0.870	0.838	0.907
Rajasthan	0.766	0.778	0.806	0.861
Sikkim	na	0.648	0.638	0.621
Tamil Nadu	0.861	0.878	0.899	0.903
Tripura	0.702	0.726	0.675	0.678
Uttar Pradesh and Uttarakhand	0.504	0.529	0.569	0.616
West Bengal	0.796	0.849	0.864	0.870
Andhra Pradesh	na	na	na	0.979
Bihar	na	na	0.644	0.639
Chhattisgarh	na	na	0.889	0.890
Jharkhand	na	na	0.642	0.769
Madhya Pradesh	na	na	0.870	0.870
Telangana	na	na	na	0.939
Uttar Pradesh	na	na	0.554	0.608
Uttarakhand	na	na	0.805	0.732

Source: The author.

Note: na: not available.

The trend in family planning performance has also varied across states/union territories. There are 25 states including undivided states of Andhra Pradesh, Bihar, Madhya Pradesh and Uttar Pradesh for which estimates of the prevalence of different modern family planning methods are available for both 1992–1993 and 2015–2016. In 8 of these 25 states, family planning performance appears to have decreased in 2015–2016 compared to 1992–1993 as is reflected through the decrease in the index p. In the remaining states, improvement in family planning performance has been the most marked in Odisha closely followed by West Bengal whereas there has been virtually no improvement in family planning performance in the undivided Madhya Pradesh, Maharashtra and Gujarat. Moreover, improvement in family planning performance has, at best, been marginal in eight states.

District-level estimates of method specific prevalence rates are available from the fourth round of NFHS conducted during 2015–2016. These estimates suggest that there is no district in the country where family planning performance may be rated as good or very good as the index p is found to be less than 0.800 in all districts. On the other hand, in more than 85% districts of the country, family planning performance may be rated as either poor or very poor as the index p is found to be less than 0.400 in these districts (Table 4). This leaves only around 15% districts in the country where family planning performance may be rated as average as recent as 2015–2016. There are 19 states/union territories in the country where family planning performance is found to be either poor or very poor in all districts whereas in eight states/union territories, family planning performance is found to be very poor in at least 50% districts. By contrast, in 16 states/union territories of the country, there is no district where family planning performance is found to be very poor as of 2015–2016 (Table 5).

Table 4.

Distribution of Districts by the Level of Family Planning Performance and Its Components, 2015–2016.

Performance Level	Modern Spacing Methods Coverage Index (c_s)		Permanent Methods Coverage Index (c_p)		Coverage Index (c)		Quality Index (q)		Performance Index (p)
Performance Level	N	%	N	%	N	%	N	%	N	%
Very poor	163	25.5	31	4.8	32	5.0	338	52.8	173	27.0
Poor	283	44.2	67	10.5	228	35.6	263	41.1	373	58.3
Average	146	22.8	143	22.3	296	46.3	39	6.1	97	15.2
Good	27	4.2	227	35.5	82	12.8	0	0.0	0	0.0
Very good	1	0.2	178	27.8	2	0.3	0	0.0	0	0.0
	Summary measures
Minimum	0		0.031		0.077		0		0.015
Q1	0.298		0.667		0.472		0.133		0.289
Median	0.435		0.808		0.585		0.277		0.403
Q3	0.584		0.91		0.68		0.448		0.509
Maximum	0.917		0.989		0.949		0.71		0.745
IQR	0.285		0.243		0.208		0.314		0.219
CV	0.402		0.266		0.277		0.599		0.356
N	640		640		640		640		640

Source: The author.

Table 5.

District Family Planning Performance by States/Union Territories.

Country/State/Union Territory	Family Planning Performance			Number of Districts
Country/State/Union Territory	Very Poor	Poor	Average	Number of Districts
Andaman and Nicobar Islands	33.3	66.7	0.0	3
Andhra Pradesh	100.0	0.0	0.0	13
Assam	6.3	56.3	37.5	16
Arunachal Pradesh	0.0	96.3	3.7	27
Bihar	86.8	13.2	0.0	38
Chandigarh	0.0	0.0	100.0	1
Chhattisgarh	22.2	77.8	0.0	18
Dadra and Nagar Haveli	0.0	100.0	0.0	1
Daman and Diu	50.0	50.0	0.0	2
Delhi	0.0	33.3	66.7	9
Goa	0.0	100.0	0.0	2
Gujarat	23.1	76.9	0.0	26
Haryana	0.0	38.1	61.9	21
Himachal Pradesh	0.0	58.3	41.7	12
Jammu and Kashmir	0.0	40.9	59.1	22
Jharkhand	37.5	62.5	0.0	24
Karnataka	86.7	13.3	0.0	30
Kerala	78.6	21.4	0.0	14
Lakshadweep	0.0	100.0	0.0	1
Madhya Pradesh	40.0	60.0	0.0	50
Maharashtra	8.6	88.6	2.9	35
Manipur	11.1	88.9	0.0	9
Meghalaya	14.3	71.4	14.3	7
Mizoram	0.0	62.5	37.5	8
Nagaland	0.0	100.0	0.0	11
Odisha	0.0	90.0	10.0	30
Puducherry	75.0	25.0	0.0	4
Punjab	0.0	5.0	95.0	20
Rajasthan	0.0	97.0	3.0	33
Sikkim	0.0	0.0	100.0	4
Telangana	90.0	10.0	0.0	10
Tamil Nadu	78.1	21.9	0.0	32
Tripura	0.0	100.0	0.0	4
Uttar Pradesh	8.5	85.9	5.6	71
Uttarakhand	0.0	76.9	23.1	13
West Bengal	0.0	31.6	68.4	19
India	27.0	57.8	15.2	640

Source: The author.

Classification of Districts

The index p is the composite of indexes c and q. The index c, in turn, is the composite of the index of the met demand for modern spacing methods, c_s and index of the met demand for permanent methods, c_p. This means that family planning performance in a district can be characterised in terms of indexes c_s, c_p and q. The contribution of the three indexes to the index p is, however, not additive as different weights are assigned to the three indexes in the construction of the index p. We have carried out the classification modelling exercise (Han et al., 2012; Tan et al., 2006) to classify districts in terms of index p based on the characterisation of districts in terms of c_s, c_p and q. The classification and regression tree (CRT) method (Brieman et al., 1984) was used for the purpose. CRT is a non-parametric recursive partitioning method that divides districts into mutually exclusive clusters in such a way that within-group homogeneity in the index p is the maximum. A cluster in which all districts have the same value of the index p is termed as ‘pure’. If a cluster is not pure, the impurity in the cluster can be measured through the Gini index. If the dependent variable is a categorical one, the method provides cluster-specific distribution of the dependent variable. If the dependent variable is a scale variable, the method provides estimates of arithmetic mean and standard deviation of the dependent variable within the cluster (Chaurasia, 2018). In the present case, the dependent variable, the index p, is a scale variable and the three explanatory variables, q, c_s, c_p, are also scale variables. The classification modelling exercise, therefore, provided mean and standard deviation of the inter-district distribution of the index p in each cluster. The TREE routine of the SPSS software package was used for classification modelling.

Results of the classification modelling exercise are summarised in Table 6 and the classification tree is depicted in Figure 3. The 640 districts of the country, as they existed at the time of 2011 population census, can be grouped into eight mutually exclusive clusters which differ in family planning performance, on average. The most important classification variable is the index q while the least important classification variable is the index c_p. Family planning performance is relatively the best in cluster 14 comprising of 54 (8.4%) districts. In these districts, the index q is at least 0.398, index c_s is more than 0.605 and index c_p is less than or equal to 0.925. On the other hand, family planning performance is relatively the poorest in cluster 5 comprising of 75 (11.7%) districts. In these districts, the index q is less than or equal to 0.153 while the index c_p is more than 0.925. Family planning performance is also very poor in cluster 7 comprising of 68 (10.6%) districts. In these districts, both the index q and the index c_s are very low.

Table 6.

Results of the Classification Modelling Exercise.

Node	Parent Node	Primary Independent Variable	Improvement	Split Values	Mean	Standard Deviation	N	Remarks
0						0.397	0.141	640
1	0	Met demand permanent methods (c_l)	0.001	<= 0.925	0.414	0.124	521
2	0	Met demand permanent methods (c_l)	0.001	> 0.925	0.322	0.184	119
3	1	Family planning services capacity index (q)	0.007	<= 0.227	0.289	0.067	183
4	1	Family planning services capacity index (q)	0.007	> 0.227	0.483	0.089	338
5	2	Family planning services capacity index (q)	0.005	<= 0.153	0.198	0.081	75	Terminal
6	2	Family planning services capacity index (q)	0.005	> 0.153	0.532	0.100	44	Terminal
7	3	Met demand modern spacing methods (c_s)	0.001	<= 0.289	0.221	0.036	68	Terminal
8	3	Met demand modern spacing methods (c_s)	0.001	> 0.289	0.329	0.046	115	Terminal
9	4	Met demand modern spacing methods (c_s)	0.001	<= 0.462	0.417	0.068	129	Terminal
10	4	Met demand modern spacing methods (c_s)	0.001	> 0.462	0.523	0.074	209
11	10	Family planning services capacity index (q)	0.001	<= 0.398	0.465	0.049	89	Terminal
12	10	Family planning services capacity index (q)	0.001	> 0.398	0.567	0.058	120
13	12	Met demand modern spacing methods (c_s)	0.000	<= 0.605	0.534	0.042	66	Terminal
14	12	Met demand modern spacing methods (c_s)	0.000	> 0.605	0.607	0.049	54	Terminal

Source: The author.

Figure 3.

Source: The author.

Table 7 presents the distribution of districts in different clusters by the index p. Family planning performance is average (0.550 ≤ p < 0.750) in more than 90% districts of cluster 14. There are only five districts in this cluster where the family planning performance is poor (p < 0.550). By contrast, family planning performance is very poor (p < 0.300) in more than 97% districts of cluster 7 and almost 90% districts of cluster 5. There are only two districts in cluster 7 and only eight districts in cluster 5 where the family planning performance is poor (0.300 ≤ p < 0.550). On the other hand, there is no district in clusters 6, 11, 13 and 14 where the family planning performance is very poor (p < 0.300).

Table 7.

Distribution of Districts by Family Planning Performance Index in Different Clusters Identified Through Classification Modelling Exercise.

	Cluster
	5	6	7	8	9	11	13	14
	Defining characteristics
q	≤0.153	>0.153	≤0.227	≤0.227	>0.227	>0.227 ≤0.398	>0.398	>0.398
c_s			≤0.289	>0.289	≤0.462	>0.462	>0.462 ≤0.605	>0.605
c_p	>0.952	>0.952	≤0.925	≤0.925	≤0.925	≤0.925	≤0.925	≤0.925
	Frequency
Very poor	89.3	0.0	97.1	27.8	6.2	0.0	0.0	0.0
Poor	10.7	45.5	2.9	72.2	93.0	97.8	68.2	9.3
Average	0.0	54.5	0.0	0.0	0.8	2.2	31.8	90.7
	Summary measures
Mean	0.198	0.532	0.221	0.329	0.417	0.464	0.534	0.607
SD	0.081	0.100	0.036	0.046	0.068	0.049	0.042	0.049
CV	0.409	0.188	0.163	0.140	0.163	0.106	0.079	0.081
Minimum	0.015	0.359	0.138	0.231	0.236	0.326	0.434	0.461
Median	0.200	0.568	0.220	0.320	0.419	0.470	0.535	0.599
Maximum	0.358	0.677	0.309	0.435	0.551	0.580	0.671	0.745
IQR	0.126	0.174	0.050	0.066	0.090	0.060	0.048	0.037
N	75	44	68	115	129	89	66	54

Source: The author.

Regional pattern of family planning performance at the district level is also very much evident from the classification modelling exercise (Table 8 and Figure 4). More than 63% districts of cluster 14 are located in the northern region while another 35% are located in the eastern region of the country. There is only one district in this cluster which is located in the central region of the country. There is no district in this cluster which is located in either the western or the southern regions of the country. On the other hand, 80% districts of cluster 5 are located in the southern region and the remaining ones are located in the central and western regions of the country. There is no district in this cluster which is located in the northern and eastern regions of the country. Figure 4 depicts the geographical continuity of districts of different clusters as regards family planning performance.

Table 8.

Regional Patterns of Family Planning Performance in India, 2015–2016 (Proportion of Districts in Different Clusters by Region).

Region	Clusters								All
Region	14	13	6	11	9	8	7	5	All
North Jammu and Kashmir, Himachal Pradesh, Punjab, Chandigarh, Uttarakhand, Haryana, Delhi	63.0	19.7	50.0	20.2	7.0	1.7	0.0	0.0	15.3
East Sikkim, Arunachal Pradesh, Nagaland, Manipur, Mizoram, Tripura, Meghalaya, Assam, West Bengal, Odisha	35.2	56.1	9.1	27.0	37.2	2.6	0.0	0.0	21.1
West Rajasthan, Gujarat, Dadra and Nagar Haveli, Daman and Diu, Maharashtra	0.0	1.5	34.1	32.6	8.5	30.4	4.4	4.0	15.2
South Andhra Pradesh, Telangana, Karnataka, Goa, Kerala, Lakshadweep, Tamil Nadu, Puducherry, Andaman and Nicobar Islands	0.0	1.5	0.0	1.1	1.6	22.6	27.9	80.0	17.0
Central Uttar Pradesh, Bihar, Jharkhand, Chhattisgarh, Madhya Pradesh	1.9	21.2	6.8	19.1	45.7	42.6	67.6	16.0	31.4
Number of districts in the cluster	54	66	44	89	129	115	68	75	640

Source: The author.

Figure 4.

Source: The author.

Discussions and Conclusions

This article employs a composite index to measure family planning performance in terms of meeting the diverse family planning needs of couples. The family planning performance index, used in the present analysis, may serve as the basis to convey summary information about family planning progress and may signal priorities to improve the delivery of family planning services to meet the family planning needs of couples. The index offers a more rounded assessment of family planning performance and presents the ‘big picture’ in a simple, yet, convincing manner that appeals to policymakers, programme managers and even to the common people. An advantage of the index is that it may be calculated from the already available data. The index provides the evidence to make policy-level changes and programme-level modifications in the delivery of family planning services.

The application of the composite performance index to India suggests that India’s family planning performance remains poor despite sustained official efforts spanning almost seven decades. The delivery of family planning services in India has almost entirely been a prerogative of the official family planning efforts. Poor family planning performance in the country, as revealed through the present analysis, therefore, reflects, largely, the poor performance of the official family planning efforts in the context of meeting the very diverse family planning needs of couples to achieve their reproductive goals. There is, obviously, a need to reinvigorate official family planning efforts in the country to improve family planning performance.

The analysis also reveals very strong, inter-district, variation in family planning performance that appears to have persisted over time. It appears that family planning performance at the district level is strongly influenced by district-specific factors both endogenous and exogenous. Endogenous factors are related to the organisation of family planning services—planning, implementation and monitoring. At present, very little is known about the organisation of family planning services at the district level, but an understanding of the organisational efficiency of family planning services is necessary to improve family planning performance. On the other hand, exogenous factors like rural–urban distribution of the population and its religious and social class composition, level of education and standard of living also have a strong impact on family planning performance. It may, however, be emphasised that the impact of exogenous factors on family planning performance can be minimised by improving the organisational efficiency of family planning services.

In conclusion, the evidence available through NFHS suggests that family planning efforts in India need comprehensive reinvigoration to improve its performance. It is important that any reinvigoration of family planning efforts must be based on a family building approach that emphasises proper birth spacing rather than the existing birth limitation approach. Such a strategic shift in family planning efforts is necessary because an increasing number of states and union territories of the country has now achieved replacement fertility and many more will be achieving replacement fertility in the near future. There is already evidence to suggest that a major share of future population growth in India will be the result of the momentum built-in the young age structure of the population (Chaurasia, 2016; Chaurasia & Gulati, 2007). It is well-known that momentum effects of population growth cannot be eliminated. They can be only minimised by either lowering the completed fertility below the replacement level or by increasing the mean age at childbearing through promoting appropriate birth planning practices—delaying entry into the marital union and increasing spacing between successive births (Bongaarts, 1994). Birth planning will also contribute to accelerating the reduction in infant, child and maternal mortality which remain high by international standards. There is, therefore, a need to treat family planning as a development strategy rather than a fertility reduction intervention. It needs to be integrated in the national development agenda. At the policy level, the need for such a shift has been recognised in India’s ‘Vision FP 2020’ (Government of India, 2014). The challenge, however, remains to translate this vision into actual practice.

Footnotes

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship and/or publication of this article.

References

Anand

, & Sen

A. K.

(1997). Concepts of human derivation and poverty: A multidimensional perspective. Poverty and human development: Human development Papers. United Nations.

Bertrand

J. T.

, Sullivan

T. M.

, Knowles

E. A.

, Zeeshan

M. F.

, & Shelton

J. D.

(2014). Contraceptive method skew and shifts in method mix in low- and middle-income countries. International Perspectives on Sexual and Reproductive Health, 40(3), 144–153.

Bongaarts

(1978). A framework for analyzing proximate determinants of fertility. Population and Development Review, 4(1), 105–132.

Bongaarts

(1994). Population policy options in the developing world (Working Paper No. 59). The Population Council. Research Division.

Bongaarts

, & Potter

(1983). Fertility, biology, and behavior: An analysis of the proximate determinants. Academic Press.

Brieman

, Friedman

J. H.

, Olshen

R. A.

, & Stone

C. J.

(1984). Classification and regression trees. CRC Press.

Bullen

P. S.

(2003). Handbook of means and their inequalities (pp. 175–177). Kluwer.

Chaurasia

A. R.

(1985). Organisational aspects of family welfare programme in India. Indian Journal of Community Medicine, X(3&4), 181–190.

Chaurasia

A. R.

(2000). The institutional context of fertility and reproductive health in Madhya Pradesh, India. University of Washington, Center for Demography and Ecology.

10.

Chaurasia

A. R.

(2004). Fertility in the development blocks of Madhya Pradesh: An analysis of the impact of contraceptive use. Social Change, 34(2), 26–39.

11.

Chaurasia

A. R.

(2016). Demographic drivers of future population growth in India. Demography India, 44(1&2), 83–95.

12.

Chaurasia

A. R.

(2018). The state of development in villages of India: An analysis of 2011 census data. Indian Journal of Human Development, 12(3), 305–325.

13.

Chaurasia

A. R.

(2020). Contraceptive method skew in India 1992–2016. https://doi.org/10.1101/2020.07.14.20154013

14.

Chaurasia

A. R.

, & Gulati

S. C.

(2007). India: The state of population. National Population Commission and Oxford University Press.

15.

Chaurasia

A. R.

, & Singh

(2014). Forty years of family planning programme in India. Journal of Family Welfare, 60(2), 1–16.

16.

FP2020 (2013). Progress report 2012–2013: Partnership in action. FP2020.

17.

Government of India (1990). Family welfare programme in India. Yearbook 1988–89. Ministry of Health and Family Welfare.

18.

Government of India (1995). National Family Health Survey (MCH and family planning), India 1992–93. International Institute for Population Sciences.

19.

Government of India (2000). National Family Health Survey (NFHS-2), 1998–99: India. International Institute for Population Sciences.

20.

Government of India (2007). National Family Health Survey (NFHS-3), 2005–06: India: Volume I. International Institute for Population Sciences.

21.

Government of India (2014). India’s vision FP 2020. Ministry of Health and Family Welfare.

22.

Government of India (2017). National Family Health Survey (NFHS-4), 2015–16: India. International Institute for Population Sciences.

23.

Han

, Kamber

, & Pei

(2012). Data mining: Concepts and techniques. Elsevier.

24.

Jain

A. K.

(1997). Consistency between contraceptive use and fertility in India. Demography India, 26(1), 19–36.

25.

Khan

M. E.

, & Prasad

C. V. S.

(1980). Family planning practices in India: Second all India survey. Operations Research Group.

26.

Operations Research Group (1970). Family planning practices in India: First all India survey. Operations Research Group.

27.

Organisation of Economic Cooperation and Development (OECD) (2000). Handbook on constructing composite indicators. Methodology and user guide. Organisation for Economic Cooperation and Development.

28.

Reynolds

, Janowitz

, Homan

, & Johnson

(2006). The value of contraception to prevent perinatal HIV transmission. Sexually Transmitted Diseases, 33, 350–356.

29.

Ross

J. A.

, & Mauldin

W. P.

(1996). Family planning programs: Efforts and results. Studies in Family Planning, 27(3), 137–147.

30.

Ross

, & Stover

(2013). Use of modern contraception increases when more methods become available: Analysis of evidence from 1982–2009. Global Health, Science and Practice, 1(2), 203–212.

31.

Stanislav

(2009). Mathematical means and averages: Basic properties. Stan’s Library: Castano Primo, Italy. https://doi.org/10.3247/SL3Math09.001

32.

Srinivasan

(1988). Modernization, contraception and fertility change in India. International Family Planning Perspectives, 14(3), 94–102.

33.

Stover

(1998). Revising the proximate determinants of fertility framework: What have we learned in the past 20 Years? Studies in Family Planning, 29(3), 255–267.

34.

Stover

, Fuchs

, Halperin

, Gibbons

, & Gillespie

(2003). Costs and benefits of adding family planning to services to prevent mother-to-child transmission of HIV (PMTCT). The Futures Group.

35.

Tan

P.-N.

, Steinbach

, & Kumar

(2006). Introduction to data mining. Pearson Education.

36.

Tsui

A. O.

(2001). Population policies, family planning programs and fertility: The record. Population and Development Review, 27(5), 184–204.

37.

United Nations (2019). Contraceptive use by method. Data booklet. United Nations, Department of Economic and Social Affairs.

38.

United Nations (2020). World fertility and family planning 2020: Highlights. United Nations, Department of Economic and Social Affairs, Population Division.

Family Planning Performance in India: 1992–2016

Abstract

Keywords

Introduction

Family Planning Performance Index

Family Planning Performance in India

Classification of Districts

Discussions and Conclusions

Footnotes

Declaration of Conflicting Interests

Funding

References