|
Field
Abstract
|
Before
In this study, we present evidence from two randomized experiments conducted in the Philippines, where the informal sector remains the population group with the lowest health insurance coverage rate, despite numerous initiatives. One of our objectives is to see how far voluntary enrollment in a government-run SHI program responds to information and premium subsidies, because cost and lack of information about how the program works are obvious reasons why a family may choose not to enroll. Another objective is to see how far non-enrollment among families given information and offered a premium subsidy might be because of the transactions costs – broadly defined – associated with enrollment.
A cluster randomized experiment was undertaken testing two sets of interventions encouraging enrollment in the Individually Paying Program (IPP), the voluntary component of the Philippines’ social health insurance program (also known as PhilHealth). In early 2011, 1037 unenrolled IPP-eligible families in 179 randomly selected intervention municipalities were given an information kit and offered a 50% premium subsidy valid until the end of 2011; 383 IPP-eligible families in 64 control municipalities were not. In February 2012, the 787 families in the intervention sites who were still IPP-eligible but had not enrolled had their vouchers extended, were resent the enrollment kits and received SMS reminders. Half the group also received a ‘handholding’ intervention: in the endline interview, the enumerator offered to help complete the enrollment form, deliver it to the insurer’s office in the provincial capital, and mail the membership card.
Our main experiment tested the effectiveness, in terms of its impact on enrollment in the voluntary government-run Individually Paying Program (IPP), of a combination of information and a 50% premium subsidy. We found that the intervention raised enrollment by three percentage points (37%), but the effect is not quite significant at the 10% level (p = 0.11). We found no evidence of adverse selection: households experiencing an adverse health event in the 12 months before the baseline survey were no more (or less) likely to enroll after receiving the information and being offered the subsidy. We did, however,
find a significantly larger (p = 0.00) effect among city-dwellers (the marginal effect in this group was nine percentage points (p = 0.00)) which we argue is consistent with this group facing lower transport costs to the provincial health insurance agency office where enrollment takes place.
Shortly before the end of the main experiment we planned a sub-experiment on those who had been given the information package and the offer of a 50% subsidy but after 9–10 months had still chosen not to enroll. In the control group we extended the voucher and resent the information package. In the intervention group, we provided a package of measures that drastically reduced the amount of effort a family has to make to enroll. This
group received assistance in person at their home in completing the enrollment form, had their form delivered to the insurance agency office for them, and had their membership card mailed to them.
Compared to the control group, the intervention group recorded a 29 percentage point increase in enrollment, significant at less than the 1% level (p = 0.00). In this case, we found a significantly smaller (p = 0.00) (but still significant (p = 0.00)) impact among city-dwellers. We argue this is consistent with city-dwellers finding this second intervention less valuable than rural households, given their greater proximity to the health insurance agency office, and their greater ability to complete unaided a health insurance enrollment form because of their higher educational attainment (in our sample 19% of urban households have a college-educated household head; only 9% of rural households do).
|
After
In this study, we present evidence from two randomized experiments conducted in the Philippines, where the informal sector remains the population group with the lowest health insurance coverage rate, despite numerous initiatives. One of our objectives is to see how far voluntary enrollment in a government-run SHI program responds to information and premium subsidies, because cost and lack of information about how the program works are obvious reasons why a family may choose not to enroll. Another objective is to see how far non-enrollment among families given information and offered a premium subsidy might be because of the transactions costs – broadly defined – associated with enrollment.
A cluster randomized experiment was undertaken testing two sets of interventions encouraging enrollment in the Individually Paying Program (IPP), the voluntary component of the Philippines’ social health insurance program (also known as PhilHealth). In early 2011, 1037 unenrolled IPP-eligible families in 179 randomly selected intervention municipalities were given an information kit and offered a 50% premium subsidy valid until the end of 2011; 383 IPP-eligible families in 64 control municipalities were not. In February 2012, the 787 families in the intervention sites who were still IPP-eligible but had not enrolled had their vouchers extended, were resent the enrollment kits and received SMS reminders. Half the group also received a ‘handholding’ intervention: in the endline interview, the enumerator offered to help complete the enrollment form, deliver it to the insurer’s office in the provincial capital, and mail the membership card.
Our main experiment tested the effectiveness, in terms of its impact on enrollment in the voluntary government-run Individually Paying Program (IPP), of a combination of information and a 50% premium subsidy. We found that the intervention raised enrollment by three percentage points (37%), but the effect is not quite significant at the 10% level (p = 0.11). We found no evidence of adverse selection: households experiencing an adverse health event in the 12 months before the baseline survey were no more (or less) likely to enroll after receiving the information and being offered the subsidy. We did, however, find a significantly larger (p = 0.00) effect among city-dwellers (the marginal effect in this group was nine percentage points (p = 0.00)) which we argue is consistent with this group facing lower transport costs to the provincial health insurance agency office where enrollment takes place.
Shortly before the end of the main experiment we planned a sub-experiment on those who had been given the information package and the offer of a 50% subsidy but after 9–10 months had still chosen not to enroll. In the control group we extended the voucher and resent the information package. In the intervention group, we provided a package of measures that drastically reduced the amount of effort a family has to make to enroll. This group received assistance in person at their home in completing the enrollment form, had their form delivered to the insurance agency office for them, and had their membership card mailed to them.
Compared to the control group, the intervention group recorded a 29 percentage point increase in enrollment, significant at less than the 1% level (p = 0.00). In this case, we found a significantly smaller (p = 0.00) (but still significant (p = 0.00)) impact among city-dwellers. We argue this is consistent with city-dwellers finding this second intervention less valuable than rural households, given their greater proximity to the health insurance agency office, and their greater ability to complete unaided a health insurance enrollment form because of their higher educational attainment (in our sample 19% of urban households have a college-educated household head; only 9% of rural households do).
|
|
Field
Final Sample Size (or Number of Clusters) by Treatment Arms
|
Before
Main experiment: 1037 treatment participants (given Intervention I), 383 control participants
Sub-experiment: 395 treatment participants (given Interventions I and III), 392 control participants (given Intervention II only)
|
After
Main experiment: 1037 treatment participants (given Intervention I), 383 control participants
Sub-experiment: 395 treatment participants (given Interventions II and III), 392 control participants (given Intervention II only)
|