A pregnant woman in rural Bihar receives a phone call she did not request. It plays a short audio message, in her language and in a voice she finds warm and trustworthy, about what to expect in her seventh month of pregnancy. The message covers warning signs, nutrition and when to visit the health centre, and it lasts about ninety seconds. Next week she will receive another call, timed to her eighth month.
That is the whole service. The technology behind it was simple, and the difficult work was in the design.
How it works
The programme is an automated phone call service. A pregnant woman is registered in the system with her phone number and expected due date. From then on, the system sends her a weekly audio message timed to her stage of pregnancy. The messages continue through pregnancy and into the first year of the child’s life, covering breastfeeding, immunisation, complementary feeding and danger signs.
The calls are outbound. The woman does not need to dial a number, download an app or have internet access. She needs a phone that can receive calls. In rural India that phone is often shared and used by the whole household, and the woman does not own it herself. The calls are designed with this in mind, so that whoever answers can learn something relevant.
Why this is harder than it sounds
Building it meant designing under several constraints.
Language was the first. India has hundreds of languages and dialects, and the messages need to be in the language the woman speaks, which is often a district dialect rather than Hindi or English. Each message is recorded separately for each language variant. The content stays the same, and the voice and phrasing are local.
Timing was the second. The value of a message depends on when it arrives. A warning about pre-eclampsia is useful at 28 weeks. At 12 weeks it is premature and frightening, and at 40 weeks it is too late. Matching messages to the week of pregnancy is where most of the system’s clinical value lies.
Completion was the third. An outbound call that goes unanswered reaches nobody, so the system needs retry logic: call back at a different time of day, and try again the next day. The call completion rate is the operational measure that decides whether the programme reaches anyone at all.
Trust was the fourth. In most settings, an unsolicited call from an unknown number is spam. The calls need to be introduced through the health system, ideally by a frontline health worker during a registration visit, so that the woman expects the call and does not hang up.
The scale
The programme became one of the largest mobile maternal health services in the world. It ran across several Indian states and delivered millions of calls a week. That scale was possible because the cost per call was low. Each call cost a fraction of a rupee in airtime, and each message was recorded once and delivered millions of times.
The programme was evaluated through large randomised trials. Women who received the calls showed improvements in health knowledge and in reported health behaviours compared with women who did not. The effect sizes were modest, as one would expect from an information intervention delivered through a phone speaker. Across millions of women, modest effects still add up to real change in a population.
What we carry from this
This work taught us three things about working at scale.
The design has to be as simple as the user’s environment. The woman receiving the call may not be able to read, may not own the phone, and may live in a household where someone else makes her health decisions. Every design choice, including language, timing, retry logic and message length, has to allow for this.
During implementation, the operational measure matters more than the outcome measure. Call completion rate is not a health outcome, but if the calls are not completed there is no health outcome to measure. The dull operational work of raising call delivery by two percentage points is what makes the programme reach people.
Scale also changes the evidence question. At small scale, the question is whether the programme works. At large scale, the question is whether it works well enough, cheaply enough and reliably enough to justify the cost of the system. That answer is harder to get, and it matters more.