Call for Action!

This is an ongoing research project, please do not cite since all the information is preliminary

The main problem for the Mexican authorities and the scientific community regarding the recent influenza pandemic is that Mexico has limited capabilites to accumulate information from citizens regarding the spread of the H1NI virus. This is due to two factors 1) the limited physical resources of the Mexican health services, and 2) the limited knowldege about viral diseases in Mexico among the population. This means that the government does not have the necessary staff to detect infected people, and that infected people don't recognize flu symptoms.

More information on the people infected by the virus would give public health officials the possibility of calculating the epidemic curve of the virus, the rate of contagion, and maybe the mortality rate.

Considering these limitations and possibilities, SMS seemed to offer a good method for getting massive information from the population.

One Mexican cellphone company agreed to carry out the project, sending 1 million SMS around the country for the academic purposes of a group of researchers at the Harvard School of Public Health and Mexico's Secretary of Healthy. The technology used was a combination of SMS and USSD.

On May 12, 1 million subscribers received and SMS asking them to participate in a survey conducted by the Secretary of Health by entering a code which opened the USSD session in the company's network. (USSD sessions have no costs for the user)

The questions were the following:

1) Did you have a fever accompanied by cough, and throat ache?
1. Yes, 2. No
2) Please remember the approximate date when the fever started and type a number
(i.e. 16)
3) Did these symptoms incapacitate you to go to work, study or take care of your family?
1. Yes, 2. No
4) Apart from yourself how many people got flu symptoms during the month of April in your family?
(i.e.2)
5) Did you get vaccinated against flu last year?
6) What is your age?

If the first question was answered "No" then the session jumped directly to question 4.

The response rate was around 5.6% (56000 people) which is relatively low, although a large number regarding this type of information. Preliminary analysis of the information shows that it may be prove useful, except for that fact that in many cases the answers regarding dates were the date used by the example (16).