The Utility of Hayashi's Quantification Theory Type 2 for The Rapid Assessment of The Epidemiological Survey in The Developing Countries - In a Case of The Vaccine Coverage Survey in Yunnan Province , China -

It is difficult to conduct the epidemiological survey with accurate statistical theory in the developing countries due to the problems especially emerged in the developing countries such as no availability of the accurate population data, limitation of the time for survey, geographical obstacle and financial problems. In the consideration of the health needs in those countries , results of the epidemiological survey should be reflected in the present health projects as rapidly as possible. The authors conducted the survey for the immunization coverage of oral poliomyelitis vaccine( OPV) in China by the questionnaire and interview. The data was analyzed by using the Hayashi's quantification theory type 2 and made clear about the relationship between the health behaviors for immunization among residents and several factors . It is considered that this theorem can provide the obvious numeral materials about multi-factorial cause effect relationship for the decision making and is useful for the rapid assessment or initial survey in the developing countries. J Epidemiol, 1998 ; 8 : 24-27 .

It is difficult to conduct the epidemiological survey with accurate statistical theory in the developing countries due to the problems especially emerged in the developing countries such as no availability of the accurate population data, limitation of the time for survey, geographical obstacle and financial problems.In the consideration of the health needs in those countries , results of the epidemiological survey should be reflected in the present health projects as rapidly as possible.
The authors conducted the survey for the immunization coverage of oral poliomyelitis vaccine( OPV) in China by the questionnaire and interview.The data was analyzed by using the Hayashi's quantification theory type 2 and made clear about the relationship between the health behaviors for immunization among residents and several factors .It is considered that this theorem can provide the obvious numeral materials about multi-factorial cause effect relationship for the decision making and is useful for the rapid assessment or initial survey in the developing countries.
quantification theory, international health, China, Poliomyelitis It has been conducted the nation-wide immunization program of the oral poliomyelitis vaccine( OPV ) by mass campaign method, which is called as National Immunization Day 1.2.1) (NID) and targeted the population less than 4 years old since 1993 in China as one program of the poliomyelitis eradication project.It is consisted by 2 rounds immunization of OPV.Last year it was organized in 5th December 1996 and 5th January 1997 in 23 provinces out of 30 in China.The authors carried out the survey for vaccine coverage by NID during the first round immunization in December 1996 in two prefectures of Yunnan Province China which bordered with Laos and Myammar.

MATERIALS AND METHODS
The target population of the survey was children under 5 years old in Xishuangbanna Dai nationality autonomous prefecture and Simao prefecture.Target population for NID was expanded to under 5 years in those prefectures because they bordered with Laos and Myanmar and were considered as high risk area.Accompanied persons of the target population were interviewed by using questionnaire which contains questions about present and past history of the immunization of oral poliomyelitis vaccine, age, sex and style of inhabitation.Fifty five cases were sampled by non-probability method, namely, accidental sampling.Mandarin Chinese was used for survey and there was no problem on mutual understanding.
Hayashi's quantification theory type 2 was adopted for analysis as present immunization history was applied for the criterion variable ( dependent variable ) and age, sex, past immunization history and style of inhabitation were for the explanation variables ( independent variables ).Age was categorized into five groups.About the past immunization history, if the subjects received OPV by past NID or routine EPI program they were categorized as "immunized".Style of inhabitation was classified into three groups, namely , the settled, intraprovincial floating population and inter-provincial floating population.About the floating population, the former was defind as the group of persons who moved to the place of the survey from the other place in the same province and the latter was from other provinces.It depended on the answer of interviwee.The children who were born in the place in the family of floating population were included to the floating population.

RESULTS
Table 1 shows the distributions of personal characters by age, sex, past immunization history, style of inhabitation and place of survey.Chi square tests were used to see the difference in the items for the immunization by present history and significant differences were revealed in the items of past immunization history, style of inhabitation and place of survey, that is, the categories of 0 dose ( p=.000 ), intra-provincial floating population (p=.042) and Mengla ( p=.030 ) were seemed to be less received the service of present MD in each groups respectively.
In the second step, Hayashi s quantification theory type 2 was adopted for analysis as present immunization history was applied for the dependent variable and age, sex, past immunization history and style of inhabitation were for the indepen-dent variables( Table 2 ).In the comparison of the assigned weights, the categories of 0 dose ( -1.246 ), intra-provincial floating population ( -.685 ), Mengla ( -.478 ), unknown sex ( -.315 ) and Daluo ( -.147 ) showed low value and Simao (.841), the immunized ( .394 ), the settled ( .378 ), age 0 to 1 years old ( .240 ), male children ( .078 ) revealed high value ( Table 3 ).The range, which was defined as the difference of maximum and minimum of the assigned weights in each factors was computed to infer which factor attributed to the criterion variable more than others ( Table 2 ).As the result, the factors of past immunization history, place of survey and style of inhabitation showed relatively large value ( 1.640, 1.319 and 1.063 ).This result corresponded with that by chi square tests.

DISCUSSION
It is analyzed the relationship between the situation of the immunization of OPV by present MD and several factors, that is, age, sex, past immunization history, style of inhabitation and place of survey.About the factor of age, the assigned weights were likely to decrease according to the increase of age.Two categories more than 3 years old revealed minus value.It was considered that the parents of the children less than 3 years old had more chance to be involved in the events which provided them more concern to the immunization like EPI and it might raise the vaccine coverage among those.
It was shown in sex that male was the positive factor toward the present immunization while female was negative.Though the absolute value of those two categories was small, there appeared the tendency of the difference by sex.
The assigned weight of the category of OPV zero dose children indicated the largest absolute value ( -1.246) among those of all categories.It became obvious that OPV zero was the risk factor which affected to the immunization of the present NID mostly.
It was found that intra-provincial floating population was the strong negative factor.On the other hand, the assigned weight of inter-provincial population showed the positive value and it was close to that of the settled.It might be included the persons who had already settled in the place of the survey more than decades among interviwee who answered as the interprovincial floating population.Not few children of those persons might be bom in the place and those figures showed that there seemed to be few difference between those persons and settled as the receiver of health services.The standerdized category of the floating population should be taken into consider for the health services.
There was the significant difference about the coverage of immunization by present NID among the places of survey.This survey had conducted in the period of NID, that is, 7th December in Mengla, 8th December in Daluo, 10th December in Ximeng and 10th December in Simao.To see the coverage rate, there was the tendency that it increased according to the passing of dates.It was considered that the difference might not be attributed by the places but dates.The assigned weights of Simao, which was the capital of the prefecture against to the fact that other 3 places were located near to the border revealed the positive value.Its health infrastructure and knowledge of the people about immunization were better.It was infered that the geographical factor also reflected to the result.Hayashi's quantification theory 4), which was one of the methods of multivariate analysis was developed in 1950s in Japan.It is characterized that qualitative variables are usable both for criterion variable and explanation variables, and used widely in the field of the study of psychology and the humanities.It provides the numerical data to realize which and how much the category attributes to the criterion variable.Cross tabulation tables and chi square tests are likely to be used wide-ly to analyze the data in the survey with qualitative data.In this survey, it is also possible to find by cross tabulation tables that there is the tendency that the coverage rate of OPV becomes higher among the subjects who have the past immunization history, are the settled and lives in the large city ( Simao ).But the priority of attribution among categories are not clear.It is possible to make clear by the analysis of quantification theory.
It is difficult to conduct the epidemiological survey with accurate statistical theory in the developing countries due to the problems especially emerged in the developing countries such as no availability of the accurate population data, limitation of the time for survey, geographical obstacle and financial problems.In the consideration of the health needs in those countries, results of the epidemiological survey should be reflected in the present health projects as rapidly as possible.To answer the demands for needs taking into accout of such special situation, data analysis by using Hayashi's quantification theory type 2 is useful to provide more obvious numerical materials about multi-factorial cause effect relationship for the decision making.Application of this method is considered to be useful for the rapid assessment or initial survey in the developing countries.

Table 3 .
Categories and Assigned weights in order.