Cancer
registration in
low-resourced settings: practice and recommendations
Paul Opoku1,
Baffour Awuah1,
Kofi Nyarko2
1National
Center for Radiotherapy
and Nuclear Medicine, Komfo Anokye
Teaching Hospital, Kumasi Ghana.
2Cancer
Control
Corresponding
author: Paul
Opoku
Agyemang; P.O. Box KS 7738, Kumasi, Ghana; Tel: +233208850420; e-mail: pokuism@yahoo.com
Afr J Haematol Oncol
2010;1(4):129-137
SUMMARY
A Cancer Registry is
an information system designed for the collection, storage, management,
and
analysis of data on persons with cancer, usually covering a defined
population.
Cancer registration is the
process of continual, systematic
collection
of data on the occurrence and characteristics of cancers with the
purpose of
helping to assess and control the impact of cancer on the community.
Compilation of worldwide standardised cancer
rates allows
the identification of countries and regions where particular tumour
types are most common. Such differences
usually
reflect exposure to distinct causative environmental factors. In
addition to
providing data on the distribution of cancer disease, descriptive
epidemiology
provides the basis for prevention, health service planning and resource
allocation.
The purpose of
this paper is to discuss the design and implementation of a
population-based
cancer registry in resource limited settings. The discussion is based
on our
experience in Ghana.
Keywords: Ghana; Developing
countries; Registries; Africa;
Epidemiology; Cancer
INTRODUCTION
Countries require cancer surveillance programmes to collect and analyze data on the scale of the cancer burden. These are urgently needed in Africa as cancer data sources are scarce. Data can help to evaluate the impact of prevention, early detection/screening, treatment and palliative care programmes. Reducing Ghana’s cancer burden is a crucial priority that involves many different professional groups including physicians, researchers, epidemiologists, public health planners, nurses, legislators and medical students. All however require and rely on cancer data in order to best tackle cancer. Cancer registries can provide this type of data and are valuable research tools for those interested in the etiology, diagnosis, and treatment of cancer. Cancer data may also point to environmental risk factors or high risk behaviors, so preventive measures can be taken to reduce the number of cancer cases and resulting deaths. Local and national cancer agencies and cancer control programs can also use registry data from defined areas to make important public health decisions that maximize the effectiveness of limited public health funds, such as the implementation of screening programs. Lifetime follow-up is another important aspect of the cancer registry. Patient follow-up serves as a reminder to physicians and patients to schedule regular clinical examinations and provides accurate survival information. Compilation of worldwide standardised 1 cancer rates allows the identification of countries and regions where particular tumour types are most common. In brief, the importance of cancer registries lies in the fact that they collect accurate and complete cancer data that can be used for cancer control and epidemiological research, public health program planning, and patient care improvement. Ultimately, all of these activities reduce the burden of cancer. The two common cancer registry types are departmental based cancer registry and population based cancer registry.
DEPARTMENTAL
BASED CANCER REGISTRY
The National Center for Radiotherapy and Nuclear Medicine – an Oncology Directorate (OD) of the Komfo Anokye Teaching Hospital (KATH) in Kumasi, Ghana, initiated a departmental-based cancer registry just after its commencement of operations in 2004, using analytic case finding reportability. This reports cancer cases diagnosed and treated at the reporting facility.
The success of the
registry
compelled the World Health Organisation (WHO) International
Agency for Research
on Cancer (IARC) to train a cancer
registrar (first
author) on several occasions to help establish a population-based
cancer
registry in the country to provide meaningful data for healthcare
planning,
implementation of intervention programs and development of a national
cancer
control program which Ghana currently does not have.
POPULATION-BASED
CANCER REGISTRY (PBCR)
PBCR records
information on
cancer patients occurring in a given population (most frequently a
defined
geography) with an emphasis on epidemiology and public health. The aim
is to
produce statistics on cancer occurrence; incidence, survival and
mortality and
also to support research and cancer control programs. Unlike
hospital-based
cancer registries (HBCR) which focus on individuals and are used for
treatment,
evaluation and administrative purposes, PBCR focuses on populations and
helps
to act
as a driver for policy development
and program
evaluation as
recommended by WHO.
The following are the
preliminary
steps in setting up PBCRs: (1) define population size; (2) obtain
denominators
from official census data (age-specifics); (3) define variables and
datasets;
(4) define sources of information; (5) establish method of data
collection -
passive or active, analytic or non-analytic; (6) decide on mode of data
collection - electronic versus paper recording; (7) establish quality
control
procedures; (8) establish administrative framework; (9) establish a
legal
framework for registration.
DEFINING
POPULATION SIZE
In our
case, Kumasi has been the first region to have a PBCR (Figure 1).
The
feasibility and gradual implementation of the Kumasi
Cancer Registry (KCR) has compelled us to strategically replicate the
design in
additional geographical areas of the country. The PBCR is now being
further extended to cover the geography of the
Ashanti Region of Ghana.
OBTAINING
DENOMINATORS FROM
OFFICIAL CENSUS DATA
This
is an important step for a PBCR. However, we are yet to resolve
issues with census data ourselves. Ghana is working on population
census and we
hope to get accurate statistics in due course. We are currently
focusing on
data collection.
DEFINING VARIABLES
AND DATASETS
The basis of
criteria for cancer incidence in five continents are
all
included in the
PBCR. We use the International Classification of Diseases for Oncology,
Third
Edition (ICD-O-3) coding scheme. The
ICD-O is designed to promote international comparability in the
collection,
processing, classification, and presentation of cancer statistics.
ICD-O is the
conventional way to translate a medical diagnosis and text into
standard codes.
This includes providing a format for reporting cancer diagnosis and
causes of
death on the death certificate. The reported conditions are then
translated
into medical codes through use of the classification structure and the
selection
and modification rules contained in the applicable revision of the
ICD-O,
published by WHO.
These coding rules
improve the
usefulness of cancer statistics by giving preference to certain
categories, by
consolidating conditions, and by systematically selecting a single
cause of
death from a reported sequence of conditions. The
datasets used for registration cover the following items: patient
demographics,
tumour data, source of data collection, and follow-up details (Tables 1-4).
|
|
Figure
1. Strategic Structure of Kumasi
Population-Based Cancer Registry |
DEFINING SOURCES OF INFORMATION
Casefinding is the system used to locate every new case of cancer that comes from the area covered by the registry. The identification of all cancer cases is essential to ensure completeness of coverage; therefore cancer cases should be available to the cancer registry databases. Examples of casefinding sources include hospital inpatient departments like oncology, haematology; pathology departments; hospices; and death certificates (Tables 5-6).
In our situation the cancer registry started from the OD - which could be considered as part of a HBCR, and later projected to cover other departments of the hospital thereby making it a HBCR except that only cases within the defined geography were abstracted. The registry is yet to capture cases from other health facilities within the Ashanti Region. The teaching hospital where the registry is located is a tertiary teaching hospital and serves as a referral center for almost all cancer cases diagnosed. OD also serves the middle and northern sectors of the country.
ESTABLISHING METHOD OF
DATA
COLLECTION - PASSIVE OR ACTIVE, ANALYTIC OR NON-ANALYTIC
The three ways in which cancer cases are found and reported are: (1) Active, where registry staff go out to collect cancer data; (2) Passive, where cancer cases are reported to the registry; and (3) a combination of active and passive.
The cancer registry must be complete in terms of the extent to which all the incident cancer cases occurring in the target population are included in the registry database. If the degree of completeness of registration is low, the rates will be underestimated and there is a possibility of unknown bias. The use of all possible sources is the ideal and the concept of sources and percentage each source contributes to the database must be specified.
KCR currently uses a combination of active and passive analytic case reporting. An analytic cancer case is defined as a case that is diagnosed and treated at the reporting facility; whereas a non-analytic case is diagnosed and treated before admission to reporting facility.
DECIDING ON MODE OF DATA COLLECTION - ELECTRONIC VERSUS PAPER RECORDING
The mode of data collection is an essential part of cancer registration. Our registry uses both hard (paper) and soft copies of notification forms for data management. The cancer data is abstracted on a paper cancer notification form-which is normally done at the casefinding facility. The abstracted data is then coded (mostly done at the cancer registry office) and later entered into an electronic database. Old cases are abstracted and the data updated periodically when new information is available during follow up of cases.
Ideally, notification forms are usually filed after indexing with a cancer registration identification number and stored in a secure room under lock with no unauthorized access. Only registry personnel have access to the electronic database and filing systems. Alternatively, the notification form could be scanned and stored electronically and then shredded as employed in some developed settings.
In short, the registration order is as follows: (1) Abstract; (2) Code; and (3) Record
ABSTRACTION
Abstract is the summary of a medical record for a specific cancer patient. The abstract relates the events leading to the diagnosis of cancer through initial treatment of the identified cancer.
An abstract primarily contains four parts: (1) Demographic; (2) Tumour; (3)
Facility/Source: and (4) Follow-up
A complete abstract contains both text and codes. The text summarizes the cancer patient medical experience and the codes collapse the text into assigned data fields. It is highly recommended that the abstractor or cancer registrar records the text section of the abstract first. Coding should be done from the recorded text.
The abstractor or cancer registrar must have knowledge of: Registry operations; Medical knowledge of cancer (diagnosis and treatment); Knowledge of anatomy; Understand medical terminologies used by clinicians, radiologists, oncologists, surgeons and pathologists; General characteristics of cancer i.e. the ability to identify and code treatment, topography, morphology, behaviour, grade and staging; How cancer spreads to organs and tissues; and How lymphatic system of major organs drain.
Complete and accurate abstract depends on the knowledge and experience of the abstractor, therefore the abstractor’s work should be reviewed.
The order of cancer abstraction is as follows: (1) Pathology report; (2) Surgical report; (3) Radiology report; (4) Medical record report
In some registries, national or local legislation determine which, how and when cancer cases will be reported. In order to collect data that are comparable among registries, standards are needed and registry procedures need to be consistent (when defining): Residents and non-residents; Reference data; Diagnostic confirmation; Multiple primaries; and Staging.
All abstractors need to know the standards and definitions that pertain to the registry. The purpose of these standards is to increase the accuracy, quality and comparability of data and its utility.
CODING
After data abstraction with a data notification form, cancer sites are coded for topography (primary sites), morphology (cell type), grade (cancer aggression) and behaviour (tumour behaviour) according to the International Classification of Diseases for Oncology.
Geographic location of case addresses are also coded using a list of towns and villages of districts in the Ashanti Region where the KCR covers, using custom developed standard codes featured in the CanReg4 electronic software. These addresses are editable in the “Dictionary” of CanReg4.
STORAGE
Once the data has been coded, it is captured into an electronic database. The registry uses CanReg4 (an acronym for Cancer Registry software version 4), a customized database software designed by the International Agency for Research on Cancer (IARC) for storing cancer data. It is developed and distributed by International Association of Cancer Registries (IACR) in close collaboration with IARC. CanReg4 can be used to code, store, analyze and generate cancer reports for local and international use.
The software requires the use of passwords to allow different levels for users to access, modify or analyse the data. It incorporates data entry, editing, import, export, analysis and backup facilities. It features a search engine used to track down duplicate records and multiple primaries using a probability matching, and consistency checking for impossible or rare cases.
In much more developed settings like the Thames Cancer Registry in the UK, complex software developed by an independent vendor providing detailed characteristics on data items are used. Additionally Picture Archiving and Communication System (PACS), Heath Episode Statistics (HES) and Geographical Information System Software (GIS) etc, are employed from data entry through reporting. Flagging and automated online notification systems are also used where cancer information is securely shared among various facilities and registries over the internet.
CanReg4 on the other hand provides a “one-size-fit-all” cancer data management. It is best suited for countries starting new cancer registries because of its simplicity, flexibility and ease of use. It does however, require official population data and geographical details of area of coverage to establish reports on incidence of cancer.
ESTABLISHING
QUALITY CONTROL PROCEDURES
Quality control is the mechanism by which the quality of data can be assessed. To ensure data validity, KCR performs routine checks for accuracy using CanReg4 Edits feature. The registry supervisor performs data quality checks after routine data storage. It is recommended that the Backup file is routinely updated and stored under strict security protected against loss, fire and interference with other files. The following recommendations should guide cancer registry managers:
a) Data Collection: Cancer notification forms should be locked away immediately after case abstraction. Data abstracted should not be disclosed to other parties other than the data source and the registry. Notification forms should not be left in a place where an unauthorised person might have access. It is highly recommended that data actively collected should be kept under lock and key preferably in a lockable case until they reach the registry.
b) Access To And Storage Of Data: Cancer notification forms should be locked away and limited access only to authorised persons. Identifiable information is never taken outside the registry. If at all taken out, this is only by the registry manager and strict measures to protect data should be adhered to. Access to the datasets by staff should be at different levels and is controlled by the use of pin codes. Each staff member should adhere to his/her level of access. Only the registry manager has the authority to make changes to the datasets. Information on a paper should be shredded into small pieces before being disposed into a paper bin.
c) Use And Release Of Data: Matters regarding the release of confidential data should be referred to the registry director. No information should be provided to medical funds, insurance companies, pension schemes, etc. Data sent abroad for international publication should not permit individual identification. A registry code number may be used which should not permit personal identification. Annual reports should be prepared in tabular form or graphs or histograms making individual identification impossible.
ESTABLISHING ADMINISTRATIVE FRAMEWORK
This is yet to be
established in
our situation, but casefinding
facility
managers are
all considered as administrators and are given the privilege to request
for
information and also contribute to future decision making of the
management of
the registry.
The cost of running
cancer
registries increases with time. However a well planned strategic budget
should
be allocated to allow sustainability, continuity and publications.
PBCRs should
be simple to present all the essential and mandatory variables,
flexible and
must be subject to change (ICDO codes and software updates),
representative and
resourceful in health policy planning, epidemiology and basic research.
Cancer
Registry budget depends on: (1) size
of the population; (2) number of data sources; (3) passive or active
data
collection; (4) Number of variables to be recorded; (3) number and type
of
staff – registrars, epidemiologists, etc; (4) equipment; (5)
projects and
publications.
CANCER REGISTRARS
Cancer registrars are the data management experts who collect cancer data from a variety of sources and report the resulting cancer statistics to various healthcare professionals and agencies.
Cancer registrars' work goes far beyond simply collecting and analysing cancer data. In order to accomplish the goal of preventing and controlling cancer, they also work closely with physicians, administrators, researchers, and health care planners to provide support for cancer program development, ensure compliance of reporting standards, and serve as a valuable resource for cancer information. Cancer registrars should access: local health facilities; available medical care to the population; all sources of information on cancer diagnosis; cooperation and support of medical community; and Support from policy makers and other stakeholders
ESTABLISHING A LEGAL
FRAMEWORK FOR
REGISTRATION
This
is yet to be established in our situation. Data notification is
currently
voluntary.
But to ensure effective operations and sustainability of the registry, policies and protocols must be developed in order to protect the privacy of individuals while at the same time maximizing the completeness and accuracy of the information reported. A reasonable balance between the protection of privacy and the enhanced use of information for the betterment of the population is needed.
CONCLUSION
The Kumasi
PBCR is in the interim being used to evaluate treatment and outcomes;
and to
assist
public health and screening
intervention
programmes in collaboration with other departments of the teaching
hospital. It is worth noting
that not all of the steps required to
set up a PBCR are completed. Administrative and legal frameworks are
yet to be
established as cancer notification is voluntary; but as a Chinese
proverb goes
“keep doing and you will improve as you do”.
With standardised and consistent definitions being put in place, data from several registries would be comparable and can be aggregated for an effective central registry which can then act as a driver for policy development and program evaluation for the urgent need of a National Cancer Control Program in Ghana. Cancer registration is a ‘pillar’ of cancer control programs. 2-3.
FOOTNOTES
Conflicts
of interest: The authors declare
no
competing conflicts of interest
REFERENCES
1. Stewart BW, Kleihues
P. (2003). WHO
World Cancer Report. IARC Press
2. Greenwald P, Sondik
EJ, Young JL. Emmerging
roles for cancer registries in cancer control. Yale Journal of Biology
and
Medicine 1986;59:561-566
3. Armstrong BK. The role
of the cancer registry in cancer control. Cancer Causes and Control.
1992;3:569-579