Table 1.
Design of P4P schemes in LMICS (41 schemes in 29 LMICs)
| Design features | Proportion of schemes [Number of schemes] | |
|---|---|---|
| Measures of performance incentivised (N = 41) a | ||
| Healthcare visits | 83% [34] | |
| Quality of care (process) | 66% [27] | |
| Health outcomes | 17% [7] | |
| Quality of care (structural) | 27% [11] | |
| Management practices | 22% [9] | |
| Efficiency | 5% [2] | |
| Whose performance measured (N = 41) a | ||
| Individuals | 26% [10] | |
| Groups of health workers | 3% [1] | |
| Health facility | 76% [31] | |
| Health system managers | 5% [2] | |
| Who (ultimately) receives the payment (N = 38) a | ||
| Individuals | 86% [32] | |
| Groups of health workers | 3% [1] | |
| Health facility | 46% [17] | |
| Health system managers | 10% [4] | |
| Payment attributes | ||
| (N = 23) | Frequency | |
| Monthly or weekly | 44% [10] | |
| Bi-monthly or quarterly | 30% [7] | |
| Every 6 months | 13% [3] | |
| Annual or one-off | 13% [3] | |
| (N = 11) | Median size | 10% of monthly income |
| (N = 0) | Lag time | – |
| (N = 41) | Reward versus penalty | |
| Rewards | 98% [38] | |
| Penalties | 2% [1] | |
| (N = 3) | Coupled payments | |
| Yes | 67% [2] | |
| No | 33% [1] | |
| (N = 34) | Use of money | |
| Staff income | 56% [19] | |
| Operating budget | 6% [2] | |
| Both | 38% [13] | |
| Basis for payment | ||
| (N = 32) | Each action (e.g. visit) | |
| Yes | 72% [23] | |
| No | 28% [9] | |
| (N = 32) | Threshold target (single target) | |
| Yes | 28% [9] | |
| No | 72% [23] | |
| (N = 32) | Threshold target (multiple targets) | |
| Yes | 9% [3] | |
| No | 91% [29] | |
| (N = 41) | Type of ranking | |
| Relative ranking (tournament) | 5% [2] | |
| Own performance (absolute) | 76% [31] | |
| Own performance (improvement) | 19% [8] | |
| (N = 41) | Payment adjustment | |
| Equity | 22% [9] | |
| Quality | 19% [8] | |
| None reported | 59% [24] | |
| Gaming safeguards | ||
| (N = 41) | Performance audit | |
| Yes (without penalties reported) | 61% [25] | |
| Yes (with penalties) | 5% [2] | |
| None reported | 34% [14] | |
Note: a = multiple options possible