Introduction
This paper is a critique on an epidemiological study on cancer survival rates published by the Middle Eastern Journal of Cancer in 2011. The title of the report was, “Epidemiology and Survival Analysis of Jordanian Female Breast Cancer Patients Diagnosed from 1997 to 2002”. It used publicly available records from institutions carrying out breast cancer surveillance, and death records from government departments to determine the survival rate of female breast cancer patients diagnosed from 1997 to 2002. The sponsor of the study was the Non-communicable Disease Directorate, Ministry of Health in Amman, Jordan. The study originally sought to use the entire population of breast cancer patients in Jordan but had to exclude some potential participants because unavailability of records relating to some of the patients.
Research Question
The primary objective of the study was “to measure the observed five-year survival rate of female breast cancer patients diagnosed from 1997 to 2002 in Jordan” (Tarawneh, Arquob, & Sharkas, 2011, p. 81). In addition, the study sought “to investigate the impact of a wide range of factors on breast cancer survival” (Tarawneh, Arquob, & Sharkas, 2011, p. 81). This objective came from the realization that the best way to test the efficacy of breast cancer treatment and to uncover intervening factors influencing the efficacy of these treatments was to investigate the rates of survival of people diagnosed with it. Women carry a greater disease burden due to breast cancer compared to men (National Health Priority Action Council, 2006). Jordan had the records needed for the study hence the choice of the country.
There were no secondary objectives listed for the study. Rather, the researchers indentified the specific factors that contributed to the survival of cancer. The factors included, “age of patient at diagnosis, histopathology, laterality, grade, and stage of the tumor, and treatment modalities” (Tarawneh, Arquob, & Sharkas, 2011, p. 72). However, by looking at the results, it is possible to infer that undeclared objectives of the study included a comparison of the survival rates of breast cancer in women in other countries such as Malaysia, Saudi Arabia, and Oman (Tarawneh, Arquob, & Sharkas, 2011). This comparison also sufficed as a means of data validation. In addition, the study also aimed at comparing the rates with a similar study conducted earlier in Jordan to determine the changes in the mortality trends attributable to medical intervention and improvement in breast cancer therapy in the country.
Just like the secondary study, there was no clearly stated hypothesis for this study. The study sought to measure survivability of breast cancer. This is a mathematical measure and not a qualitative measure lending itself to the development of a hypothesis. The study simply wanted to measure how many survived and why.
Study Design
The study design used for the research project was observational study. An observational study does not include the intervention of the researcher in the same way as an experimental study. In other words, the researcher does not intervene or isolate the subjects in any way but simply studies their characteristics to decipher the patterns under investigation. All the subjects in this study pursued their own lines of therapy based in their personal physician advice. Secondly, the researchers used “historical cohort” study to collect the data needed to meet the objectives of the study (Tarawneh, Arquob, & Sharkas, 2011, p. 73). It is an application of the observational study method. Another name for historical method of data collection is longitudinal study.
At the macro level, the researchers had a choice between experimental and observational research methods. Experimental research tends to be expensive because of the need to maintain certain conditions for the duration of the experiment. However, since the research in this case wanted to establish a naturally occurring trend with several interventions, it was best to use observational study and deduce the efficacy of the treatments and associated breast cancer therapy at a later stage. Another important reason for this choice of design was that there it was not in the researchers place to decide who gets cancer, or to predetermine the treatment that the patients got, or to control its management in any way.
The main advantages of historical cohort study, as a study design option is that it is cost effective and yields information that is difficult to collect in any other way. The researcher does not have to spend time creating the required research conditions. The researcher simply defines the variables at play in a natural setting and studies the specific variables required by the research objectives (Corson, Heath, & Bryant, 2000). The major disadvantage of this design method is the relative difficulty of finding suitable research samples because of the degree of variance in the research environment. In this case, there were differences in the types of cancer such as laterality, and differences in the individual patients such as preexisting conditions and subsequent complications. This diversity of conditions complicates the design of cohort studies.
The selection of the subjects for the study was rather simple. The researchers chose to use the entire population of women cancer patients in Jordan because of the availability of information from the Jordan Cancer Registry (JCR) (Tarawneh, Arquob, & Sharkas, 2011). There only criteria required for inclusion in the study was a breast cancer diagnosis in any of the health centers in the country. By studying the entire population, it was not necessary for the researchers to find any other inclusion criteria. However, some subjects were not fit for study because of absence of records the researchers needed to determine the extent of the other variables studied.
The main factors influencing the choice of subjects included in the study were sex and a positive breast cancer diagnosis. The study only sought to identify women with a positive breast cancer diagnosis. The time factor that the researchers used to identify the suitable subjects in the study was by determining a diagnosis date and an exit date. The researchers wanted to study suitable subjects whose diagnosis came in between 1997 and 2002. Any diagnosis before or after this time frame did not qualify for inclusion in the study. The place factor was that the diagnosis took place in Jordan. It is unclear whether the research excluded non-Jordanian women residing in Jordan or whether it included women diagnosed in another country and then moved to Jordan.
The same factors that defined the source population also served as the inclusion and exclusion criteria. The inclusion criteria required that the subject be a woman diagnosed with breast cancer between 1997 and 2002. Importantly, the subject also had to be in the JCR registry. The exclusion criteria included being male, suffering from any other type of cancer apart from breast cancer, and a diagnosis outside the 1997 to 2002 range. It was also important for the researchers to have access to the subject’s medical records to help determine the extent of the influence of the other variables. It is unclear whether the researchers included or excluded people with multiple cancers including breast cancer.
The researchers did not sample the source population. They took the whole population of women suffering from breast cancer in Jordan for the entire period of the study. This decision came from the fact that the records were already available through the JCR. Therefore, it did not require the researchers to collect the data on their own or to establish data management methods for the research project. On the other hand, it gave them a platform to establish credible and influential results because their sample size reflected the total population of women diagnosed with cancer. They also used data from patients’ health records and the data from the Department of Civil Status.
By relying on government agencies and institutions dealing with health records, the researchers achieved a high rate of enrollment and retention. This case is special because the subjects may not have been aware that they were part of the breast cancer surveillance that constituted the project. It appears that the subjects had no say in appearing in the JCR, just as the hospitals they visited had to keep a record of the treatments offered. In addition, the Department of Civil Status also required no permission to keep records of deaths since it is a government agency hence high retention of subjects for the project.
Explanatory Variable
The main exposure variable in this study was the presence of breast cancer proved by a positive diagnosis at a medical facility. The measurement of the exposure variable did not matter much to the researchers because of their reliance on data from JCR. The cancer management program in Jordan requires all cancer patients to register with JCR. For this study, the registration with JCR served the purpose needed for the measurement of the exposure variable. Otherwise, it was the responsibility of medical doctors to diagnose the disease and send the information to JCR. It was in health centers that the measurement of the exposure variable took place. The researchers played no part in this stage.
The researchers used the conventional classification of cancer progression. The disease progression is in four stages, denoted Stage I, Stage II, Stage III, and Stage IV. There were no study specific classifications of the exposure variable. The need to identify these levels came from the fact that the subjects seek medical attention at various stages. These stages influence the survival chances of the subject hence the need to determine the stage at which the cancer was as at the time of the diagnosis. This information influenced the reporting of the findings further proving the importance of registering the classification as at the time of inclusion into the study. Other factors used by the researchers were laterality, morphological type, and tumor grade. These were the measures used to determine which patients had the highest chance of survival.
The accuracy of measurement of the exposure did not depend on the researchers hence it is not attributable to them. However, the systems used to collect this information are credible. The fact that medical doctors diagnosed the cancer patients is enough to assure the researchers and peer-reviewers that the data received from JCR is credible. In case of any doubts, it would have been easy for the researchers to account for it based on the recorded percentage of wrong cancer diagnosis. Moreover, the length of study gave sufficient time for correction of any diagnosis. In addition to these facts, the researchers used medical records to track the progress of the patients. It would have been easy for them to find inconsistencies and correct them.
Study Outcome
The health outcome studied by this research was the mortality rate of women diagnosed with breast cancer. The research followed the progress of the disease in the women over a period of five years from the day of diagnosis, and stopped the process exactly five years later. As such, the first lot of women admitted into the project in 1997 exited in 2002, while those admitted in 2002 exited in 2007. In this period, the researchers were able to measure various elements of breast cancer presentation and treatment that helped them to track the progress of the disease.
The definition of the cases was not complex. Using the exclusion and inclusion criteria, the researchers identified the subjects that they could use for the experiment. In this sense, the definition of the cases depended on the health progress of the subjects and their eventual mortality status. The main issue the researchers were looking for is the survivability of breast cancer. The researchers found variables to use to account for death by other causes. Since they used official death reports from the Department of Civil Status, determining the cause of death was straightforward.
Measures of Frequency and/or Association
Incidence is the rate of occurrence of the health issue while prevalence refers to the total population affected by it (Helme & Gibson, 2001). This study did not look at the incidence of breast cancer. It recruited women who already had breast cancer for the study hence it was not concerned with the rate of occurrence of breast cancer. However, the measure of deaths associated with the disease can offer some measure of incidence in the study. On the other hand, prevalence of deaths due to breast cancer was a core issue in the study. The study’s core objective was the determination of the deaths associated with breast cancer within five years of diagnosis.
The two commonly used measures of association are Relative Risk (RR) and Odds Ratio (OR) (Akhtar, 2008). Relative Risk is the ratio between the cumulative incidence of the disease in both a group exposed to a disease and one not exposed to the disease (Akhtar, 2008). Odds ratio on the other hand measures the correlation between exposures and occurrence of the disease (Akhtar, 2008). In this case, it is the measure between a positive cancer diagnosis and death. This study did not include these measures explicitly in its computations. The use of Cox regression analysis and univariate analysis seemed to suffice for the researchers (Tarawneh, Arquob, & Sharkas, 2011).
The interpretation of these results depends on the stated objectives. These results apply in as far as the measurement of the mortality rates of the women breast cancer patients.
Sources of Error
Selection bias in this study could have come from inclusion of women misdiagnosed as breast cancer patients while they have another type of disease. The possibility that this happened is rather low. Therefore, the risk of selection bias is very low for this particular study. In addition, the fact that the study relied on data from JCR made it easier for the researchers to avoid selection bias. Another potential source of selection bias was the risk that some of the cancer suffers in the population may not have been diagnosed as such.
The experiment had little chance of suffering from information bias because of the data collection methods. The actual data collection took place in health centres, and credible government agencies with little influence from the subjects. The main source of confounding bias was that there were other factors influencing the mortality rate of women suffering from breast cancer. Some women died from accidents or from other health complications. It is difficult in this sense to isolate the impact of other causes of death because of the large number of potential combinations of causes of death.
In order to determine the degree of generalization possible from the results of the study, it is important to look at the sampling techniques used (Ulmer, 2010). In this study, the entire population was also the sample space in use for the study. As such, the results apply to the entire population hence no further need to generalize. It may be possible to generalize the results to countries with similar social-economic characteristics and comparable health systems (Jacobs, Rapoport, & Jonsson, 2009). However, the chance of finding another country that fits well in this criterion is difficult. Therefore, it is more likely that the result can only offer a comparison but not a generalization of the trends in question.
Statistical Analysis
There main results from the study were as follows. The worst survival rates of breast cancer patients were in women under the age of 30. Their cumulative survival rate was 51.7%. The second worst survival rate was in the 70 years and over age group where the survival rate was 58%. Women in the 30-39 years age group had a survival rate of 61.0%, while the 40-49 years age group had the highest survival rate at 69.3%.
The 50-59% age group had a 64.9% survival rate and the 60-69 years age group had a 63.3% survival rate. The study also had results showing that under tumor morphology, the best survival rates were in women with Medullary carcinoma. Under tumor grade, the best survival rates were in women with well-differentiated tumors. They had a 73.8% survival rate. Finally, the worst survival rate was in women with bilateral tumors at 46.15%. Those with left and right tumors had survival rates of 64.08% and 65.01%.
There were several other statistical measures used in the study after the presentation of the results. The researchers used univariate analysis and Cox regression analysis to process the basic data. The univariate analysis revealed that the “stage, grade, and laterality of breast cancer influenced cancer survival rate” (Tarawneh, Arquob, & Sharkas, 2011, p. 77). Cox regression revealed that the, “stage, grade and age factors correlated with prognosis, while laterality showed no significant effect on survival” (Tarawneh, Arquob, & Sharkas, 2011, p. 72). The statistics appear credible because of their credible extraction.
Conclusion
The causality criteria in this study revolved around the determination of the relationship between the cancer diagnosis and mortality. In simple terms, the study looked at cancer as the cause of death among women with a positive diagnosis. The strength of association between these two issues is that the experimenters determined that 62% of cancer patients survived in the first five years after diagnosis. The researchers presented several other statistics such as the rate of mortality among specific age groups, the influence of laterality, tumor morphology and tumor grade, showing that they had an influence on the mortality rate.
As observed earlier, the researchers did not have an explicit hypothesis in regards to the results of this study. Since they were looking for a relationship not measured before by any other means, their study had a baseline role in order to establish the parameters needed to compute the mortality rates of breast cancer in women in Jordan.
The bottom line of the study was the determination of the survivability of breast cancer in women. From this study, it is clear that in Jordan, 62% of women survive breast cancer in the first five years after diagnosis. The long-term trend is that because of improving health standards in Jordan, the rate of breast cancer survival is on the increase.
References
Akhtar, S. (2008). Epidemiologic Measures of Association. Karachi: Division of Epidemiology and Biostatistics, Aga Khan University.
Corson, D., Heath, R. L., & Bryant, J. (2000). Human Communication Theory and Research: Concepts, Context, and Challenges (2nd ed.). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
Helme, R. D., & Gibson, S. J. (2001). The Epidemiology of Pain in Elderly People. Clinics in Geriatric Medicine , 17 (3), 417-31.
Jacobs, P., Rapoport, J., & Jonsson, E. (2009). Cost Containment and Efficiency in National Health Systems: A Global Comparison. Weinheim: Wiley Verlag.
National Health Priority Action Council. (2006). National Chronic Disease Strategy. Canberra: Australian Government Department of Health and Ageing.
Tarawneh, M., Arquob, K., & Sharkas, G. (2011). Epidemiology and Survival Analysis of Jordanian Female Breast Cancer Patients Diagnosed from 1997 to 2002. Middle East Journal of Cancer , 71-80.
Ulmer, C. (2010). Future Directions for the National Healthcare Quality and Disparities Reports. Washington DC: National Academies Press.