A survey was carried out among 12,535 children in the age group 0-9 years of 90 villages in Doddballapur sub-division of Bangalore district to study the possible variation in the prevalence of tuberculous infection among the unvaccinated children in a village depending upon the varying prevalence of BCG scars in the same population. In each village, all the children in the age group of 0-9 years were registered and examined for the presence or absence of the BCG scar. Of the 12,535 children, 6269 (50%) who did not have BCG scars were eligible for tuberculin test, while 6045 were actually tested. Each child without BCG scar was tuberculin tested with 1 TU RT 23 with tween 80 and the reaction read between 72 and 96 hours. Two proportions were calculated in each village viz., a) the proportion with BCG scars and b) that of infected children among those without scar and the villages were distributed by these two proportions.

On the basis of distribution of tuberculin reactions, 10 and 12 mm induration was the demarcation between positive and negative reactors. Prevalence of infection among 0-9 years was 4.9%, 2.6% among 0-4 years and 8.9% among 5-9 years. Distribution of villages according to two variables i.e., prevalence of BCG scars and prevalence of infection among unvaccinated children did not show any correlation with the prevalence of infection among the unvaccinated in the same villages.

It is seen from the study that exclusions of various proportions of children with BCG scars did not have any correlation with the prevalence of infection among the unvaccinated in the same villages.

In non-e of the villages any association was seen between these two. In view of this finding, it is felt that the simple method of periodic tuberculin testing of the population in younger age groups could be developed into a method of tuberculosis surveillance even in areas where direct mass BCG vaccination is given. This would appear to be the cheapest, practicable and technically appropriate method of studying the overall tuberculosis situation.

The utility of repeated estimates of prevalence rates of infection in children as a tool for surveillance in tuberculosis is now well recognized. Two prevalence surveys at an interval of 5 years were conducted by National Tuberculosis Institute, Bangalore, with the main objective of studying changes in prevalence rate of infection among children in the age group of 0-9 years. A total population of 42,343 residing in 90 randomly selected villages of Doddaballapur taluk, Bangalore, were registered; of them, 12,535 were children in the age group of 0-9 years. Children were further classified into two sub groups 0-4 and 5-9 years, with or without BCG scars. The unvaccinated children in these two age groups formed the study population.

The population in the study area during the 2nd repeat survey was similar to that of first survey with regard to age, sex distribution, except that a growth rate of 1.1% per year was registered. The BCG scar rate, among children in the age group 0-4, 5-9 years, was 8% & 39% respectively at survey I. All the unvaccinated children below 10 years were given tuberculin test with 1 TU PPD RT 23 and reactions were read 72 to 96 hours after tuberculin testing. In the first survey, level of demarcation to classify the infected children was 10 mm and above, while in II survey it was 12 mm and above. It was observed that the prevalence rate of infection from I survey to II survey was not altered (2.58% & 2.46%) in the 0-4 years of age, while there was an increase in the rate from 8.93% to 12.3% in 5-9 years of age in the II survey. The increase in the infection rate could be attributed to the rising trend of infection, over reading by tuberculin-readers', skills of both tuberculin tester and reader, boosting of tuberculin reaction or scarless BCG vaccination. In conclusion, the study of changes in the prevalence rate of infection in the younger age group is simple, cheap, less time consuming. The data can be used for calculating annual risk of infection as well trend of transmission of infection.

The National Tuberculosis Control Programme is in operation since 1962, and its quantitative achievement is being monitored indirectly through records and reports received from District Tuberculosis Centres. For direct evidence of impact of the programme, tuberculin surveys are useful in reflecting the recent epidemiological situations prevailing in the area. Tuberculosis being a disease of secular nature, a periodic follow up with five years (arbitrary) interval may be preferred over the continuous follow up, for finding the trend of tuberculosis situations in an area.

Keeping in view the importance of tuberculin surveys, National TB Institute (NTI) has evolved a surveillance system which can be adopted by any state in India. The state teams can be trained at NTI in registering population, tuberculin testing & reading, so as to carry out the surveillance in their respective areas. It is essential to create a central organisation for surveillance of tuberculosis using the tuberculin test. The centre would be responsible for technical & administrative support and monitoring. NTI could provide technical expertise in formulating the surveillance system, a training methodology and an in service training to the designated staff.

A methodology in brief about conducting Classical Tuberculosis Sample Survey and Tuberculosis Surveillance is given below:

I. CLASSICAL TUBERCULOSIS SAMPLE SURVEY
A tuberculosis prevalence survey to measure the problem of tuberculosis in the community is a challenging assignment especially so when it is to be conducted by an organisation not created with the specific objectives of carrying out research work e.g., the District Tuberculosis Centre, State Tuberculosis Centre, etc. However, following procedure is described in brief: (1) Selection of district for the study, (2) Collection of basic data like size and distribution of population, number of towns and villages, (3) Selection of sample population for survey by valid statistical methods, (4) Census enumeration of study population by trained census takers - preparation of cards for all the individuals, (5) Tuberculin testing & reading of all subjects under study, (6) BCG scar survey, (7) X- ray examination of the eligible population (> 10 years or > 15 years). Interpretation of X-rays by standard readers, (8) Collection of sputum from chest symptomatics and X-ray abnormal individuals, (9) Transportation of sputum to central laboratory (necessary precaution to be taken during storing and transportation), (10) Sputum to be examined by trained staff, (11) Compilation, analysis and interpretation of data. Number of working teams with full complement of staff depends upon the size of the study population and the time frame of the study. An average survey team may have the following personnel on its strength: Medical Officer - One, Census Takers - Three, Tuberculin Tester & Reader (one each) - Two, Lab Technician - One, Lab Asst. - One, X-ray Technician. - One, Dark room attender - One. Equipment required: Mobile X-ray unit - mounted on a jeep along with the generator mounted on another jeep, Laboratory infrastructure, Vehicles preferably jeep.

Apart from the above, the team may need part time assistance of a Statistician and a few Statistical Computers. In case a state is interested to carry out an epidemiological survey, it may need to create the above infrastructure. Once arranged, it may request the National TB Institute (NTI) to train the required staff on standard survey techniques under field situations which is very essential.

II. TUBERCULOSIS SURVEILLANCE
In contrast to the more complex methodology involved in a classical survey described above, an alternative, much simpler and indirect method to assess the problem of tuberculosis in the community is by finding out the infection rate, through tuberculin surveys. It may be possible to estimate the prevalence of sputum smear positive disease from infection rate. Such survey is conducted by subjecting the age-specific unvaccinated population to tuberculin test periodically. For carrying out the work, one to two teams composed of three to four properly trained tuberculin testers and readers are needed along with at least two vehicles and a standby vehicle per team. Budgetary support for petrol, travelling and daily allowance of staff, and for minor miscellaneous expenditure like stationery, spirit, etc., may be required to be provided. Training could be imparted to such personnel at NTI and their services utilised exclusively for carrying out tuberculin surveys as a regular ongoing surveillance activity. If this methodology is found suitable, one may take action to create posts of tuberculin testers and readers in suitable scales and draft them for training in tuberculin survey methodology. The Institute will be happy to train the required personnel for the purpose, as well as analyse the data so collected for use by the states.

A study was conducted in 1983-84 by the National TB Institute (NTI) in the districts of Dharmapuri (Tamil Nadu) & Ananthapur (Andhra Pradesh). Thirteen health personnel were trained in census taking, tuberculin testing & reading and data keeping etc., at the NTI according to the standard methodology. The trainees were repeatedly assessed and only those who achieved a reasonably high inter-reader correlation with the standard reader were chosen for the field work. Field work was carried out by these health workers and supervised by the team leaders of NTI. Children between 0-9 years were tested with 1 TU RT 23 with Tween 80 in tuberculin testing centres specially set up in each village and the reactions were read between 48 & 72 hours after the test. The tuberculin testing/reading coverage was very high. Of 6702 eligible children, 5904 (97%) were tuberculin test read.

Individual reading assessment carried out at Ananthapur and Dharmapuri for the State Field Workers (SFWs) showed that agreement with Standard Reader (SR) of NTI at three induration levels i.e., 10+ mm, 14+ mm and 18+ mm were very high. The demarcation line between infected and uninfected appeared to be about 18 mm. In Ananthapur, the agreement at 18+ mm was 99% and at Dharmapuri it was 100% for SFW, and 98.4% for SFW-2. The estimates of prevalence rate of infection were 9.3% - SFW V/s 9.7% - SR at Anantapur, 5.2% - SFW V/s 5.2% - SR and 7.2% - SFW.2 V/s 7.2% - SR at Dharmapuri. The study further showed that it was possible to train general health workers, within a period of 3 months to attain a high level of efficiency. The general health services can successfully organise on their own a programme of tuberculin testing in the community with proper liaison and supervision by some nodal agency. The training and field supervision responsibilities may be shouldered initially by NTI or another suitable organisation, till these nodal agencies come up.