57
ANNEX 5
TECHNICAL DISCUSSIONS 1. SUBJECT
In accordance with the resolution passed during the fourteenth session
of the Regional Committee, Western Pacific Region, the subject of the Technical Discussions was "The Use of statistics in Public Health Administration" • 2.
ORGANIZATION OF DISCUSSIONS The Technical Discussions took place in three sessions, the first and
third being plenary sessions while for the second the participants met in three groups for group .discussions. The Chairman of the Technical Discuss-
ions was Dr Richard K.C. Lee (United states of America)j General Rapporteur was Dr Amelia S. San Juan (Philippines); Dr W.P.D. Logan (Director of Health statistics Division, WHO, Geneva) and Mr A. Aldama-Contreras (Regional Statistician and Programme Evaluator) served as Secretariat. FIRST SESSION In his opening remarks, the Chairman reminded the Committee that one of the basic responsibilities of health departments is to collect and analyze vital facts bearing on public health. He said that the function of a statistician in a public health department is to collect, interpret, and present statistics that are relevant, reliable, complete, and up to date. With such information at
his disposal, the realth officer could feel the pulse of the community he serves. But while the statistician might get him the information he needs,
the administrator himself has to know how to use this information.
If statistics and statisticians were to be fully utilized, the bureau
of statistical services should be accessible to the health Officer, and the collection and use of health and vital statistics should be tied up with programme planning, research and evaluation. Nowadays, the biostatistician is one of the most important staff officers of the health department. He has become a highly valued and highly priced
health worker, an essential member of operating health agencies, of training programmes, and of research teams. good experienced biostatisticians. Unfortunately, there is a shortage of The re~onsibility
for the collection of
health and vital. statistics among many of our Member countries is divided. Many different agencies of governments are collecting and compiling different a~ects
of health statistics.
Health departments by themselves might not be
able to compile all the statistics needed, but some health departments are not assuming as much of these basic functions as they ought to. An analysis of the number of fellowships, consultantships,and technical
proJects on health and vital statistics over the years of existence of this Regional Office would probably show that compared with other basic health programmes, not enough had been done. Are enough specialists being trained? Are enough funds being allocated These are same of the
to statistics compared with other health programmes?
questions that might be discussed, and they are important questions for all persons concerned with public health administration. Dr Logan stressed the need to focus attention on the uses to be made of statistics rather than on the piling up of inaccurate and possibly useless figures. Statistics ought to be collected with a specific puxpose in view. These should be the kind
Some health statistics are collected routinely.
REPORT OF THE REGIONAL COMMITTEE
59
that are easy to obtain, reasonably accurate, and should be able to contribute to the improvement of the administration of health services. system should be ad fley~ble
The statistical
so that special information can be obtained by
.!!2£
enquiry or survey when there is a real need for it.
Dr Logan also mentioned the various kinds of health statistics needed in public health administration, namely: (1) Statistics about population: other characteristics, etc. (2) ;C
size, distribution, age, sex and
Statistics which describe the health status of the population, which include mortality and morbidity statistics.
(3)
Statistics which describe the health services being conducted within the country. This includes information on preventive and
curative services, their number and distribution, and how they are used. This also includes available personnel and cost of
health service operation.
(4)
Statistics on various conditions which influence the health status of the population, such as socio-economic, occupational, educational and nutritional status.
Countries whose health services are not highly developed do not need elaborate expensive statistical services. At first quite simple statistical
serviceswill suffice; they can be further developed as the health services develop. The important thing is to obtain the kind of statistics that are
really needed and to use them in the right way. In his presentation of the topic "Measurements of Mortality and Morbidity", Dr Dizon (Philippines) stated that while health is not just the absence of infirmity or disease, yet disease is the main force that undermines the state of health. It is therefore essential that disease, in terms
of its morbidity and mortality, must be defined, counted and measured. Statistics are a handy tool for such pu~ose.
60 Statistical data on disease occurrence are required by the health administrator for him to be able to define the illness problem, and, cOlllP~emented
by the
epidemiologic~
method, he is
ab~
to ex;plain such definition and of
disease occurrence.. such epidem1010gic~
On the basis of such
statistic~
characterization, quarantine and control measures may The statistic~
then be instituted.
data on morbidity must be properlY
qualified in terms of either the number of persons ill, the number of illnesses or number of spells of illness. So as to make the definition tot~
more meaningful, this measure of morbidity must be related to the
number ex;posed to the illness and ex;pressed as attack rates, and depending on the time involved, such measures may be in terms of prevalence or incidence. Among the various sources of morbidity data, he mentioned notification of diseases, hospital in- and out-patient records, and special sickness surveys of the who~
population or samples of it.
The limitations of each
source were also presented. Besides the morbidity data and preciselY because of the limitations of morbidity measurements, mortality data are invariablY utilized so as to assess disease problems in a community, analYsis and long-term trends. Although more cOlllPlete and perhaps more accurate than morbidity data, this measurement is dependent on the severity of the disease and therefore, fails to determine the distribution and rising trends of diseases of high infectiousness but of low mortality. The best sources at IWrtality data are death certificates obtained by statutory registration. It is essential that the registration system be He emphasized the importance of accuracy e~eci~
with reference to detailed
firmly established and implemented.
REPORT OF THE RIDIONAL COMMITTEE
61
in the medical certification of cause of death which should be the responsibility of each medical practitioner. On the basis of registration of
deaths, information may be obtained on the number as well as on the characterization of these numbers by various attributes. When these data are
related to the total population or to specific segments of the total population especially exposed to the risk of death, various indicators may be derived and utilized as important measures of mortality. The more
commonly used and more important of these measures are the crude death rate, specific death rates, cause of death rate, infant mortality rate, maternal mortality rate, and proportionate mortality rate. Speaking on "Statistics in the Routine Administration of Health Services",
Dr Andrade (Portugal) cited the utility of statistics as a measure of the
work performance and efficiency of .health services. statistics that are useful in this regard are: (1) (2)
The various kinds of
The number of physicians attached to the health services. The per capita cost of health services. The number and distribution of health establishments, like hospitals and out-patient clinics. It is also necessary to have
(3)
information on bed population ratio, bed capacity, number of admissiOns, length of stay, statistical classification of discharged patients, physicians, nurses and auxiliary nurses employed. For out-patient clinics, relevant data include the
number of patients treated, classified by diagnosis.
(4) Maternal and child services.
Dr Andrade recommended the
registration of pregnanCies for follow up, and classification of their progress and termination into abortions, premature .;,..
labour, normal births,and still-births.
62. •
(5) Imgnm1zation programmes. (6) . Services on special disease problems l:ilte tubercuJ.osis, leprosy, malaria and others. Dr Han (Korea) discussed 'Statistics in National Health Plann:fng and Evaluation ", in relation to the practices in Korea. He mentioned the major
health programmes being conducted in his country, and the reasons why priorities were given to tuberculosis control, of health services, etc. established. fami~
planning, strengthening
The objectives for these programmes have been
He referred to the inadequacy and inaccuracy of available
statistics giving examples of official records and specific survey results. !the programmes had to be implemented before the establishment of baseline data, in view of the magnitude of the problems requiring immediate action. Pilot projects were introduced before the progranu:ms were implemented on a national scale. Techni~
complicated methods of statistical
ana~sis
would not be
practical for health authorities in developing countries.
These could be
better carried out in controlled studies under the guidance of eJg?erts. In the presentation of the fourth topic of the Panel, "Statistics in
Public Health Research", Dr Kennedy (New Zealand) stated that the field of detailed studies, evaluation of public health programmes and operational research could be the three main headings, under which some preliminary canments could be presented. For detailed studies, he stressed that the
medical statistician should be involved from the beginning to ensure proper
collection and utilization of' data.
These studies may be directed to provide
knowledge and understanding of the relationship of' the various factors involved in the incidence of disease. However, when studies conducted in
other more advanced countries can be made use of, there :tnB:y be no need to duplicate them.
REPORT OF THE ROOIONAL COMMITTEE
• Every public health programme, be it polio immunization, mass treatment of hookworm, or the hospital care of some diseases, has to defend itself in terms of human book..lteeping. The statistician has to present a balance sheet He cited as examples, first, the
to see the results of an undertaking.
favourable results obtained from the mass use of Sabin vaccine in 1962, and second, a study demonstrating the value of fluoridation in an area where properly designed baseline studies were made previous to the fluoridation programme.
Dr Kennedy described operational research as being concerned with the ascertainment of ways in which efficiency can be improved.
A few remarks were made on the quality of data, sampling, and the expanding horizon in disease problems. On the quality of data, he said that
all the punch cards and computers cannot make up for the defects resulting from negligence and lack of co-operation in reporting. Sampling is
recommended as it enables us to kno", a lot about a large population with reasonably low cost and great speed. with c~unicable
While we are becoming less concerned
disease control, new diseases and problems especially
degenerative conditions, mental disorders, alcoholism and accidents are coming to the forefront. There is increasing emphasis on medical care,
industrial health and public health engineering. 4. SECOND SESSION
The participants were divided into three groups which met simultaneously in dif:f'erent rooms. Each group had its own chairman and rapporteur. To
facilitate the discussions, ten topics in the form of questions, based on the original guidelines were presented. The discussions were directed to
a limited number of major points and gave rise to varied opiniOns, according to the experience and practices in the participants' own countries.
RIDIONAL COMMITTEE: FIFTEENTH SESSION
5.
THIRD SESSION
(0 ~
"\~.
<:.'>I-. ..... ;~
cJl-~ >_""""-..............b. ...... ,
.... )
At its third session, after consideration of the reports of the three groups and after further discussion, the following conclusions were arrived at: (1) Statistical information is a basic requirement of health departments. Statistics are needed in order to determine the health
status of the population; to guide the operations of the health services; for health planning and programme evaluation; and for public health research. Health statistics, therefore, include
statistics describing the population in a demographic sense, its health status, the health services provided, and conditions influencing health. (2) To describe the health status of the population, mortality and
morbidity statistics are used.
Mortality statistics provide
useful background information for purposes of health administration. Their completeness depends on the efficiency
of civil registration; their accuracy depends on the availability and the co-operation of physicians to certity causes of death.
(3) Amongst morbidity statistics, a system of notification of communicable diseases has existed for many years in most countries. The completeness and accuracy of reporting vary according to the nature of the diseases reported. Those of major public health
importance such as the quarantine diseases are usually reported more or less fully, whereas the relatively unimportant diseases are often grossly under-reported. Periodic revision of the list
of notifiable diseases is desirable so that unnecessary items can be eliminated. Despite their incompleteness and inaccuracy,
statistics of communicable diseases have value in public health
REPORT OF THE REGIONAL COMMITTEE
65
administration in drawing attention to the onset of epidemics, measuring their progress, indicating their geograplucal distribution, and describing the population groups principally affected. (4) Hospital diagnostic statistics, though hi~
selective both as
regards the conditions treated and the population served, can be .~
used to supply general information on the main disease problems of the community. Fuller recognition of the potential value of
hospital statistics for this purpose is advocated.
(5) For special purposes, the routine statistical information at the disposal of the health department may be insufficient. In such
a case, special surveys of scientifically selected samples of the population are useful methods of obtaining the information required. Large, continuous morbidity surveys are relatively expensive, so surveys should be instituted only for particular purposes and for limited time. (6)
Health service statistics are used to describe the various services, preventive and curative, with which the health department is concerned, such as hospitals, dispensaries, maternal and child health services, environmental health services, etc. statistics used are those that: (a) The
describe the resources and
facilities available, e.g., number of hospital beds; (b) the utilization of these facilities, e.g., number of hospital beds occupied, average number of patient days in hospital, etc.; (c) (d) health personnel, e.g., number of physicians, nurses, etc.; the general costs of the health services.
66
REGIONAL COMMITI'EE: FIFTEEN'l'H SESSION
(7)
Statistics are needed also to describe conditions influencing health; for exam;ple,the socio-economic status of the population, nutrition, education, housing., etc., as well as tendencies toward urbanization and industrialization. These will not
usually be collected by the health department but the information should be at its disposal.
(8) Health statistics, like other statistics, are often incomplete and unreliable. The good health. administrator will recognize
their limitations, make the best use he can of the statistics such as they are, and strive to improve them in every way possible.
(9)
The statistical service of a health department must not put all its efforts simply into the amassing of data, regardless. of their quality or usef'ulness. The system must be kept under constant
review so that unnecessary data can be discarded and efforts concentrated upon the collection, interpretation and use of statistics that are really needed and can serve some purpose. The system .should be flexible enough to allow changes of direction and emphasis as required, and allow the adoption of sampling methods and special surveys when these are indicated,· (10) Within the health department organization, the statistician must be so placed that he is constantly in touch with the health administrator, can properly understand his requirements, and can give him the statistical help he needs. The statistical
service cannot work effectively in isolation from the central activities of the health department.
By introducing sound sampling
REGIONAL COMMITTEE: FIFTEENTH SESSIOU
67/68
techniques, the statistician may be able to reduce costs and to obtain inf'ormation more quickly and accurately than can be obtained by routine reporting procedures. (ll) Many of the countries of this region are in need of improvement of their health statistical services. National and international
expansion of vital and health statistical programmes, the training of statistical personnel, the improvement of basic data, and a fuller utilization of statistics for health administration are much needed.