Description of tool

Time-to-default is a measure of how long a defaulter stays in the program before defaulting. This measure distinguishes an early defaulter (i.e. defaults within 4 weeks from admission) from a late defaulter (i.e. defaults after 4 weeks from admission). It is important to distinguish these two classes of defaulters particularly early defaulters because they are most likely current cases who are not covered by the program. A program with a high number of defaults after only one or two visits has a serious defaulting problem1. This additional analysis of defaulting should be done if the routine analysis of defaulting rates found either high or increasing rates of defaulting such as was found in the program described in the section on defaulters over time. Otherwise, there is little benefit to gain from doing this additional analysis of defaulting routinely.

Data requirements

Data on time-to-default can be collected from the outpatient care treatment cards and recorded using a tally sheet.

Analysis of data

Plotting of time-to-default is the most ideal way of analysing this data. A bar plot of the data on time-to-default can be plotted either by hand or using a computer (click here for the time-default spreadsheet). If done by hand, tabulation of time-to-default data from beneficiary cards can be done using a tally sheet approach which also presents the data graphically. The tally sheet is basically a table with the number of visits prior to defaulting . The second column will be the tally column which is filled out using hash counts based on the number of days prior to default that is indicated in the beneficiary cards. Figure 1 is an example of how to collect, tabulate and plot time-to-default data by hand. A template tally sheet can be found in here.

Figure 1: Tally plot of number of visits before defaulting

_{From FANTA technical reference page 27} Using a computer, the tabulated data can be entered into a spreadsheet or a statistical package and a bar plot created with the number of visits prior to default at the x-axis and the frequencies or counts at the y-axis. Figure 2 is an example of a plot of time-to-default of a program with a serious problem of defaulting as majority of defaulters leave the program very early after admissions (i.e. after just one visit). This is a similar pattern of time-to-default as in the example in Figure 1.

Figure 2: Time-to-default in a CMAM program in Gombe State, Nigeria

_{Data courtesy of Gombe State Ministry of Health}

 

Interpretation

  A program with high defaulting rate would benefit from further investigation on time-to-default. A program with a high number of beneficiaries defaulting early similar to that described in Figure 2 has a serious problem with defaulting and most likely has low coverage and should be investigated further with key informant interviews and case histories of defaulters particularly those who have defaulted early (see section on Semi-structured interviews and Simple structured interviews).