Report of an ENCR consultant visit
10-12 August 2000 The Georgian National Cancer Registry Tbilisi, GeorgiaDr. Risto Sankila Unit of Descriptive Epidemiology International Agency for Research on Cancer 150, cours Albert Thomas 69372 Lyon Cedex 8 France The Georgian National Cancer Registry, Tbilisi, Georgia Report of an ENCR consultation visit, 10-12 August 2000
1. Background and purpose of the consultation visit
Following an invitation by the National Association of Cancer Control (Georgia), a European Network of Cancer Registries consultation visit was made to the Georgian National Cancer Registry (GNCR), Tbilisi, Georgia on 10-12 August 2000. The purpose of the visit was to increase knowledge of the Georgian cancer registration system. Further, information was obtained about the infrastructure of other health related registries. The future developments in health information sector were discussed and potential national collaborators were identified. Finally, based on these discussions, an effort was made to highlight areas where improvements in the registration processes were needed. My hosts were Dr Vasil Tkeshelashvili MD, PhD; President of the National Association of Cancer Control; Director of the Georgian National Cancer Registry; Head, Department of Cancer Epidemiology and Control of the National Cancer Center; and Dr Levan Bakanidze, MD; Coordinator of International Relations and Development Programs, National Association of Cancer Control. The National Association of Cancer Control is an NGO and it was registered in September 1999. I also had the opportunity to discuss with Professor Revaz Vepkhvadze, MD, PhD; President, National Cancer Center; Chief Radiologist of Georgia; Professor Rezo Gagua, MD, PhD; Director, National Cancer Center; Head, Thoracic department of National Cancer Center; Chief Oncologist of Georgia; and Dr Rusudan Klimiashvili, MD; WHO Liaison Officer in Georgia. There were plans that I would meet the Minister of Health, but he was on duty travel on the day of the supposed meeting. There were further plans that I would have met with the Vice Minister of Health, but this meeting was not organised. However, Professor Vepkhvadze is the father-in-law of the Minister of Health and they live in the same house. Thus, an inofficial but direct link to the highest executive level was established.2. Developments leading to the current Georgian National Cancer Registry
I was informed that a nation-wide and population-based GNCR was founded in 1950 by the Ministry of Health for administrative purposes and for descriptive epidemiology. The registration was performed according to the old 'Soviet style'. The registration was compulsory right from the beginning, meaning that an obligatory notification of each cancer patient was required from every physician to be sent to the regional oncological centres or dispensaries. The dispensaries collected the notifications, and they were then sent to the GNCR. This system functioned from 1950 until approximately 1989. There was increasing political instability during the last years of Georgia being part of the Soviet Union in 1989 to 1991. During the early years of independence in Georgia, there was severe political instability and a short civil war at the end of 1991. The internal regional conflicts with South Ossetia and Abkhazia as well as those related to the activities of the supporters of Gamsakhurdia in western Georgia added to the troubles between 1990 and 1993. Despite the continuing tensions, the Shevardnadze Government has been able to remain in power since 1992. There is some sense of increasing stability, although the collapse of the Russian economy and that of the value of the rouble combined with the need to restructure the industrial basis of the economy have severely hurt the economy. I was told that officially the average monthly salary was about USD 5 (there are no reliable estimates concerning the existing underground economy). The Government has not paid any salaries to medical researchers during the last 6 months. Hospital staff are being paid regularly. In winter months from approximately October to March, electricity is only available for a couple of hours in the morning and in the evening. Further burdens are the approximately 200 000 internally displaced people of Georgian origin from Abkhazia and approximately another 200 000 refugees from the neighbouring Chechnya. Concerning the Cancer Registry, I was informed that all the data collected during the Soviet era in 1950 to 1989 was lost during the turmoil in the early 1990s. Probably tabulated data exists in Moscow, but this is an unknown issue. The National Cancer Center was established in 1977 on the outskirts of Tbilisi by the Lisi Lake high above the city. A separate effort to collect cancer data was started at the Center, first as a hospital registry and later incorporating some data from other hospitals. Most of the data collected prior to 1992 has been lost. Since 1992 some partially collected data exist in the National Cancer Center Registry, which now calls itself the Georgian National Cancer Registry. Other specialised cancer centers in Telavi, Gori, Utaisi and Batoumi send notifications (unknown completeness, however). There are several other clinics that treat cancer patients, but do not send any notifications. Besides the obvious underreporting, there were considerable 'overdiagnostics' during the last five years. The clinicians write a cancer diagnosis to patients with benign diseases, as only in this way the patients will be compensated by the state for their treatment. This 'overdiagnosis' only concerned certain primary sites. However, according to a new law the patients will have to pay for 50% of the hospital expenses, but 100% for their drugs (chemotherapy). Thus, the 'overdiagnoses' do not exist anymore. It appears that there is no formal direct link between the GNCR itself and the Ministry of Health or the Population Office. There is no administrative or scientific board for cancer registration purposes.3. Population data and population coverage of the Cancer Registry
The last population census was performed in 1989. Thereafter only official estimates exist. These estimates have minor discrepancies, but factually there are no reliable population data. The estimates are further hampered due to the hundreds of thousands of internally displaced people and the varying inclusion or exclusion (in international estimates) of the populations of Abkhazia and South Ossetia. There are no reliable estimates of emigrations out of Georgia, nor of the population effects (other than internally displaced people) of the armed conflicts during the 1990s. Basically, all births and deaths should be registered in Georgia, but this is not necessarily happening in the rural areas with low population densities. The 1989 official population figure was 5,400,800 (1.25 million in Tbilisi) and the 1997 official estimate was 5,423,000. A population census is under planning but it has been repeatedly postponed due to the lack of funds. The next census is supposed to be performed in 2001, but local sources are not confident with this plan. With some underreporting from the National Cancer Center, unknown proportion of cases being reported from other cancer centers and no reporting from any other hospitals in combination with the unreliable population data, the population coverage of the GNCR can not be estimated.4. Staff and facilities
Dr Tkeshelashvili is the Director of the GNCR. He has a background in gynaecology, but has been trained in epidemiology in Leningrad and in Moscow (IARC course in 1988). There are no other senior staff or research personnel at the GNCR. The GNCR is situated at the National Cancer Centre (address: Lisi Lake, 380077 Tbilisi, Georgia). The GNCR is tightly linked with the archives of the National Cancer Center. Thus, it is not straightforward how the staff (six clerks) is linked to the activities of the GNCR and those of the archive. The same applies to the computer facilities, which appear to be used for hospital administrative registrations as well as for cancer registration. Three office rooms comprise the GNCR: one for Dr Tkeshelashvili: one for the hospital archives and one for the registry clerks using computers. Currently, the computing hardware fulfils minimum criteria for the registration needs. The three computers were purchased with a one-time grant from the Soros Foundation. There is no CD-ROM drive at the GNCR. The Internet appears to be a source for references in their publications, but I am not certain if the GNCR computers have Internet access. Dr Tkeshelashvili uses the Association's e-mail address (nacc@nacc.org.ge) and I would suppose that there is no separate e-mail address to the GNCR itself.5. Methods of registration, coding, data management and quality control
Dr Tkeshelashvili has created a Georgian version of the cancer registration form based on the past Soviet form. These forms are filled in by the registrars based on information from the hospital records of each patient. The information is then fed to a computer file using a registration programme developed by Dr Tkeshelashvili based on Microsoft Windows FoxPro software. These computerised records exist from the last five years. However, the data are practically from the National Cancer Center only and of varying completeness. The primary site is coded with ICD-10 codes, but morphology is given as text only in Georgian. This is due to the fact that no one can translate the ICD-O-2 codes into Georgian. This further leads to the conclusion that the pathologists at the Cancer Center do not use any standard system for classification and nomenclature. Finally, there is no separate reporting system from any pathological laboratory, but all data are incorporated in the hospital records and extracted from there. I understood, that search for duplicates was not systematically performed. However, multiple neoplasms in a single patient are supposed to be identifiable, and updating of patient information was possible, although - again - not systematically performed. Only patients dying in hospitals would be likely to have a medical death certificate, but the information in these is not stored with the population register files. For patients dying outside of the hospital, only date of death would be filed in the population register. Autopsies on (cancer) patients are not performed, as the relatives traditionally do not give their permissions for them. Quality control analyses were not performed at the GNCR. The coding program does not contain checks for internal validity.6. Results
The GNCR produced annual reports in a simple tabular format between 1952 and 1989 for administrative purposes. Since 1990, the annual reports have not been published. The Cancer Center data are in the format of hand written tabulations of ICD-9 codes by age groups. No age-standardised rates are available. The Cancer Center provides numbers of cases and crude rates to the Statistical Office. The Population Office provides the number of 'cancer' deaths, but there is no link between diagnosed patients and (cancer) deaths. For 1998, following numbers were available: - The number of new cases was 4579 in the whole country. (Dr Tkeshelashvili told me that this was an underestimate, and that the correct number would be about 7000). - Based on this number the crude incidence per 100 000 was estimated to be approximately 89. (In 1989 the crude incidence was estimated to be about 130, and in the neighbouring Armenia and Azerbaijan, it was about 223 and 225, respectively.) - The official number of cancer deaths was 3868 in the whole country (with crude mortality of about 75/ 100 000). The official statistics of the Soviet Union published in Moscow for 1993 estimated the incidence to be 107 / 100 000 (World Standard Rate) and the corresponding mortality to be 82.7.7. Training
Basic epidemiology is taught at the Medical School. There is no Institute of Public Health at the Medical Faculty, but there are separate departments of hygiene and preventive medicine. Cancer epidemiology as such is not in the curriculum and there are no related PhD programmes. There are two Medical Universities in Tbilisi (State Medical University and State Medical Academy). There are medical schools in other major cities, but there are also 45 private medical universities in the country. These private institutions were started some 4-5 years ago. However, due to new practice of licensing, the students will not receive a medical license. There are no PhD students at the GNCR. The State has not paid any scientific salaries during the last 7 months. On the contrary, young scientists must pay for working at the Cancer Center. The scientific publications from the GNCR until 1989 were published in the Soviet Union in Russian. Several of Dr Tkeshelashvili's publications are based on CI5 data and published in Georgian and Russian (one in English) some with Summaries in English.8. Course on population-based cancer registration
During my visit, great interest was expressed in organising an ENCR cancer registration course in Georgia in the future. After my return, I was informed that the Soros Foundation in Georgia has announced the program for development of Epidemiology and Medical Statistics. The National Association of Cancer Control has prepared the project: 'Conduction of training courses on: Modern Methods of Cancer Epidemiological Research and Medical Statistics at Georgian Republican Cancer Centers'. A recommendation from an international organisation was considered to be of potentially substantial value. However, as no structure of the organisation, no draft programme and no list of potential teachers and their qualifications was provided, a blank recommendation on behalf of the ENCR will be formulated.9. Specific problems and future developments
9.1. National infrastructure and legislation concerning cancer registration As in so many nations in transition, the infrastructure of the country does not provide relevant support for nationwide, population-based cancer registration. Thus, a feasible start could be to first organise the data collection in the Tbilisi area fully utilising all available data sources. As this process is gradually developed into a functioning system, it can be used as a model for the rest of the country. The development of this regional registration system should be encouraged. After an initiation period (possibly of several years), the system should be expanded gradually into the rest of the country. I was informed of the possibility of the Minister of Health issuing a decree making cancer registration compulsory on a national level and obliging the physicians to send cancer notifications the National Cancer Registry. A letter on behalf of the ENCR will be sent to the President of the Cancer Center to be used as a reference in his discussions with the Minister of Health stressing the importance of making cancer a notifiable disease, thus improving the legislative grounds for cancer registration. In the future, obligatory cancer registration could be expanded to the entire country. This could, however, wait until the experience from the Tbilisi area indicates that the methods and regional coverage have reached reasonable standards. 9.2. Coordinated collaboration of different health care related government offices During the consultation visit, I tried to stress the importance of the GNCR being active in the process of improving the efficiency of the government organisations with which it needs to collaborate. During the discussion with Dr R. Klimiashvili, the WHO Liaison Officer in Georgia, I understood that several different approaches to improve the situation were already started, some of them in collaboration or after consultations with the IMF and the World Bank. 9.3. Population and mortality data Obviously, a cancer registry without proper population data cannot achieve its goals of producing reliable age-standardised rates of the disease. If possible, the GNCR should be in contact with the officials planning to perform the next Census. The GNCR should try to ensure that the methods (and potential weaknesses) of the data collection at the Census are fully understood for future use by the GNCR. Further, good relationships should be created between the Population Office and the GNCR to allow an easy access to any relevant population data after the Census. Further, Dr R. Klimiashvili told me about the possibility of future collaboration between the two offices with mortality data. Currently, the Population Office collects data on all deaths, but without cause of death information. Another office (in the Ministry of Health?) collects data on the causes of death (when available). For the GNCR, it would be most important to be able to merge data from these two sources on individual level allowing linkage with the GNCR files. This operation would create opportunities for receiving death certificate notifications to the GNCR and for the follow-up of registered patients for death. 9.4. Collaboration with pathologists The GNCR should consider contacting 'the National Association of Pathologists in Georgia' to start a close collaboration for improving the quality of the histological data to be entered to the GNCR files. The items to be discussed should include 9.4.1. Increasing the proportion of cancer patients with histological confirmation of the diagnosis; 9.4.2. Use of standard classification and nomenclature in cancer diagnoses (i.e., ICD-O-3); 9.4.3. Creating computerised pathology registers, which would include relevant data from all the biological samples examined at the pathological laboratory. These data would be periodically submitted to the GNCR in a standard format. These procedures would create the possibility of tracing back to the clinical records cases that were only notified from a pathological laboratory. I have undertaken to locate a copy of the WHO International Classification of Diseases for Oncology, version 2 (ICD-O-2) in Russian to be sent to the National Cancer Center and the GNCR. This classification should be used by all pathologists and clinicians in the cancer field in order to increase the accuracy of the histological nomenclature and to improve the comparability of the diagnoses. The uptake of this classification should also encourage the pathologists to provide data to the GNCR in a standard format (a special form needs to be created by the GNCR) and to increase the proportion of cancer patients with a histological diagnosis. 9.5. Publications, training and research The GNCR should aim at publishing numbers of cases and age-standardised rates on a regional level as the coverage and completeness improves. Meanwhile, model-based estimates are probably the most feasible figures available for publication. The GNCR personnel should be separated from those working for the hospital archives. The GNCR personnel should be locally trained to understand the need to follow the standard procedures and classifications in order to improve the quality of the GNCR data. Automated and extra, periodical quality control exercises should be part of the training. Devoted young scientists should be recruited to work on PhD projects utilising GNCR data in the fields of cancer epidemiology, biostatistics and public health. In-depth research projects should be planned (once the data quality allows) in order to convince the international scientific community of the quality of the data and the capabilities of the scientific staff.10. Conclusions
Currently, there is remarkable interest and activity in cancer registration in Georgia. However, unless some major obstacles are overcome (see above), compulsory cancer registration on national level will not produce satisfactory results. However, after making cancer a notifiable disease, a regional registration system should be encouraged, starting with the Tbilisi area. After an initiation period (possibly of several years), the system could be gradually expanded into the rest of the country.