Contraceptive continuation rates

Definition:

The cumulative probability that acceptors of a contraceptive method will still be using any contraceptive method offered by the program after a specified period of time (e.g., one year)

This is also known as the “all-method” continuation rate.

When using cross-sectional population data, evaluators calculate the continuation rate for each unit-interval of use (e.g., first, second, third month of use, and so forth) as the complement of the ratio of acceptors who discontinue use of a program method of contraception at that duration to the number of women still using at the beginning of the month (i.e., 1 minus the discontinuation rate). Evaluators then cumulate these continuation rates to obtain the probability that acceptors of a contraceptive method will still be using any program method after the specified period of time.

The indicator (CR_x) is calculated as:

CR_x = x(1-q_x)/π

Where:
x = 1
q_x = T_x/N_x = conditional probability of discontinuing use during a given interval (e.g., one month, one quarter);
T_x= the number of women discontinuing use during the interval; and
N_x = number of women using at the beginning of the interval.

Note: π signifies that (1- q_x ) is multiplied over all intervals from 1 to x.

Illustrative Computation
Continuation rates for reversible methods for durations from 1-12 months, Bangladesh, 1992-97

X	Tx	Nx	qx	CR
1	291.3	4422.5	.0659	93.4
2	160.1	4025.9	.0398	89.7
3	205.9	3803.1	.0541	84.8
4	89.9	3502.6	.0257	82.7
5	61.2	3322.9	.0184	81.1
6	107.5	3173.7	.0339	78.4
7	68.7	2998.3	.0229	76.6
8	57.2	2868.5	.0199	75.1
9	74.6	2757.4	.0271	73.0
10	59.8	2579.8	.0232	71.3
11	55.8	2458.0	.0227	69.7
12	109.5	2350.8	.0466	66.5

Source of data: Bangladesh Demographic and Health Survey, 1996/97

Data Requirements:

Information on contraceptive initiation, duration of use (including method switching), and discontinuation during a given reference period (e.g., the 3-5 years prior to a survey). Based on this information, one can calculate the percentage who have continuously used for a specific duration (e.g., 12 months, 18 months), as well as the median duration of use. Age of the contraceptive user, specific method used, level of education, parity, and other variables can be used to assess differences in contraceptive continuation by method and/or sub-populations of interest.

Data Sources:

Population-based: surveys with retrospective contraceptive use histories or calendars such as the Demographic and Health Survey (DHS)

Program-based: client records accompanied by a follow- up study of program dropouts. This source is rarely used.

Purpose:

Contraceptive continuation rates provide a useful summary measure of the overall effectiveness of program services in enabling clients to sustain contraceptive use even though they may switch from one method to another. However, the calculation of continuation rates from surveys requires knowledge of life table (survival) analysis by those subregions of the country (much less individual facilities), making this indicator more useful at the national than regional or local level.

Although evaluators can calculate continuation rates from either facility-based or population-based data, facility-based data have a number of limitations; thus, researchers tend to use large-scale surveys to provide more valid measurements of continuation among the intended population (e.g., Blanc, Curtis, and Croft, 1999).

Issues:

Obtaining continuation rates at the program level is theoretically possible if evaluators use follow-up studies of new acceptors at a specified period of time after adoption of the method (e.g., 12 months). However, this technique is rarely used (except in clinical trials), given the difficulty and expense of locating these acceptors a year later.

The preferred source of data is the “calendar,” a data collection format used in cross-sectional surveys such as the DHS. However, such surveys have limitations of their own. They (a) depend upon the accuracy of respondent recall, (b) do not allow linking of respondents to specific service delivery points, and (c) may not capture the full contraceptive history (e.g., when five-year calendar is used).

Due to the complexity of using contraceptive calendar data, the DHS Program offers a tutorial and guidance for the calculation of the related indicator, “First-year contraceptive discontinuation rate”. This indicator provides the percentage of contraceptive use episodes that are discontinued within 12 months of initiation, by specific method. Discontinuation rates vary according to method mix, thus there is no ideal rate. However, when combined with information on the reason for discontinuation, programs can seek to reduce discontinuation due to method failure, side effects and health concerns, or wanting a more effective method.

It is important to note the distinction between discontinuation and failure of a contraceptive method. Discontinuation of contraception may occur because the individual chooses to stop using a selected method or because accidental pregnancy intervenes. As such, method failure is a subset of discontinuation. Method failure necessarily results in discontinuation. However, not all discontinuation is attributable to method failure.