Being healthy should be part of an individual’s overall life. Good health is associated with economic progress and prosperity since it determines human happiness. It is because a healthy population saves more, and they are more productive. In an effort to improve and promote better healthcare, bio-statistical studies are conducted to investigate and explore new health phenomena and their effects on human life. Statistics play an essential role in scientific studies by applying different statistical methods to understand and determine the relationship between variables identified in the study. The research findings from bio-statistical studies are essential in aiding the researcher(s) to make evidence-based recommendations for further study or action and precise conclusions on an issue under study. Bio-statistical studies incorporate the use of statistics in various problems affecting health care. A bio-statistical study includes data collection, appropriate measurements, data analysis, interpretation of the results, and conclusion based on the research findings. This paper provides a complete analysis of an article that uses the cause of death and socio-demographic characteristics to analyze current trends in maternal death rate.
The Article: MacDorman, M. F., Declercq, E., & Thoma, M. E. (2017). Trends in maternal mortality by socio-demographic characteristics and cause of death in 27 states and the District of Columbia. Obstetrics and gynecology , 129 (5), 811. https://dx.doi.org/10.1097%2FAOG.0000000000001968
Delegate your assignment to our experts and they will do the rest.
Summary of the Article
MacDorman et al. (2017) conducted a research whose objective is to analyze current maternal mortality trends. The researchers achieved this objective by analyzing the causes of deaths and socio-demographic characteristics of the population of interest. Previous studies had estimated the maternal mortality trends in the US from 2000 to 2014. Therefore, MacDorman et al. (2017) extended the previous analysis by considering the detailed causes of death, race, ethnicity, and maternal age in their analysis of trends in maternal mortality. The study's primary purpose was to identify the vulnerable population and trends in an effort to develop a prevention measure. Besides, the researchers aimed to assess the quality of essential maternal death data. They chose to evaluate this type of data since the data was not widely used in researches. In this article, the researchers present brief background information on maternal deaths, which has affected a significant population in the US. The ministry of health added a new pregnancy question to the typical death certificate in 2003 to enhance the identification of maternal mortality. The new pregnancy question helped discover the pregnancy condition of female decedents at the time of death (MacDorman et al., 2017).
The authors in this article conducted an observational study that compared data from 27 states in the US between 2008 and 2014. The chosen states, including the District of Columbia, had similar reports of maternal death during the time of the research. The researchers used a two-proportion z-test in their study. The test was used to assess the statistical importance of trends and differentials. The data on maternal mortality deaths was obtained from the International Statistical Classification of Diseases and Related Health Problems. The study incorporated the entire population of live births and maternal deaths from states implementing the pregnancy question by 1 st January 2008. The study population involved 7,369,966 live births and 1687 deaths. Findings indicated an increase in maternal death rate from 2008-2009 to 2013-2014. The researchers recorded a rise from 20.6 deaths per 100,000 live births to 25.4 deaths. Most deaths were recorded among pregnant people aged above 40 years, with the cause of death being non-specific. While there was a significant increase in maternal deaths associated with non-specific causes by 48%, mortality rates for specific causes of death only registered a slight increase within the focus period. The authors concluded that there was possible over-reporting of maternal deaths as suggested by the high maternal death rate for non-specific causes of death and those older than 40 years. They recommended that more efforts be put into modifying coding procedures and enhance reporting process. Improving these areas will enhance the reliability of maternal death data applied in monitoring trends.
Biostatistical Analysis
Research findings from biostatistical operations help researchers develop evidence-based operations and make a concrete conclusion about a specific issue in healthcare. Biostatistics refers to a field in health care and research that focuses on using statistics to analyze and recommend solutions for different issues affecting the health of a population. The biostatistical analysis involves collecting health care data, analyzing the collected data, interpreting results, and providing recommendations for future studies.
Available Data and Data Used in the Article
The information available on the topic under consideration include maternity, live births, cause of death, and socio-demographic characteristics of women. In terms of maternity, data is available on the rate of mortality associated with maternity, the emerging issues, associated health complications, and the cumulative case reported. Data on socio-demographic characteristics of the population is available on age, race, ethnicity, and health status ( Hoyert & Miniño, 2020) . In terms of causes of death, the researchers' data include non-specific and specific causes of death. Dividing estimated maternal deaths by total estimated live births provides data on maternal death rate. Thus, the data is presented as the estimated number of mortality per 100,000 live births. Researchers analyzed the data by maternal age by considering women below and above 40 years old and 5-year age groups. They also analyzed the data based on race and ethnicity, including non-Hispanic blacks, Hispanic, and non-Hispanic white.
Measurement Levels, Assumptions, and Statistics
Allanson and Notar (2020) present four measurement levels: ratio, ordinal, nominal, and interval. The authors of this article made measurements relating to the age of participants, the prevalence of maternal mortality, causes of deaths, and other socio-demographic data. The researchers do not directly mention the various types of scale used in their respective measurements in their report. However, the type of data collected and presented in the analysis suggests the types of scales used. Numbers assigned to each observation or variable on a nominal scale are only used to classify the observations or variables ( Allanson & Notar, 2020) . For instance, non-numerical variables in the study include ethnic groups, such as Hispanic, non-Hispanic white, and non-Hispanic blacks. Also, the causes of maternal deaths can be categorized under a nominal scale. Variables such as maternal mortality rate, indirect obstetric causes, and direct obstetric causes assume numerical values, indicating differences between scale values. Therefore, they apply a ratio scale. The study indicates the changes in maternal mortality rate between different periods and in different causes such as specified pregnancy-related conditions and non-specific causes. The researchers present the increased rates of direct and indirect obstetric deaths from 2008 to 2014.
During their study, MacDorman et al. (2017) made some assumptions to ensure the success of their research. Among the assumptions was that the statistical method applied led to equal variance and that the data used in the analysis was evenly distributed. The researchers applied a two-proportion z-test to examine importance of statistical data on trends in the study. They applied a sensitivity analysis model to determine the impact of different possible over-reporting of pregnancy status on maternal mortality rates. Another assumption in the study was that the research used independent and unbiased samples to avoid distortion of results. The authors performed an error measurement to determine the reliability of the collected data and the research findings after analysis. The collected data in this study and those used in the analysis are of good quality since they use all the necessary elements to create an intervention plan. The previous results had only estimated the maternal death trends in the US ( Lu, 2018) . As a result, this study aimed to close the gap by extending the previous analysis by considering the detailed causes of death, race, ethnicity, and maternal age in their analysis of trends in maternal mortality.
Study Design and Biostatistical Study Design
A statistical design in a scientific study is normally a framework that incorporates all the parts and components of a study ( Allanson & Notar, 2020) . A study design enhances data collection, measurement, interpretation, evaluation, organization, and analysis of results. This study applied a longitudinal study design. Longitudinal research design focuses on a group at two or more points in time. It enables the authors to record measurements of an aspect over a given duration of time. The researchers studied the trends in maternal mortality rate from 2008/2009 to 2013/2014 periods. More trends included an increase in indirect and direct obstetric causes within the same study period. The researchers grouped the maternal deaths according to the causes before determining the trends based on the cause-of-death grouping. Finally, the study applied the collected data to determine the relationship between variables and identify the efficiency of a possible intervention plan.
Study Hypothesis and Hypothesis Testing
According to Patwardhan et al. (2016), the hypothesis indicates the author's view of the particular issue under study. The hypothesis can be an alternative or null hypothesis, depending on the research. In this study, the researchers aimed at testing the null hypothesis. The null hypothesis was the effect of over-reporting using checkbox on death rate. The researchers used the identified variables to test their hypothesis. The intervention plan for the issue was modifying coding rules to reduce overreliance on the checkbox. At the end of the observational study, the data collected revealed that maternal mortality rates increased by about 14 to 23% due to 1% over-reporting of pregnancy. This result was more prominent in women in their twenties and early thirties (MacDorman et al., 2017). Furthermore, those between 40 to 54 years recorded a 232% increase in maternal mortality rate due to 1% over-reporting. The table below illustrates the results that support the hypothesis as extracted from MacDorman et al. (2017).
Figure 1: Connection between Maternal Mortality and over-reporting
The research findings agreed with the null hypothesis developed, which indicated that increase in over-reporting increases the rate of maternal mortality.
Statistical Results and Analysis
The study results indicated an increase in maternal maternity rate by 23% in the study population during the last five-year period. The rate rose to 25.4 mortalities per 100,000 live births in 2013-2014 from 20.6 in 2008-2009 (MacDorman et al., 2017). The researchers recorded high death rates in pregnant women above 40 years old, increasing within the five years. While women aged 25-29 recorded, 14.0 deaths per 100,000 live births, those older than 40 years had 141.9 deaths, which is ten more (MacDorman et al., 2017). The study also revealed that the lowest vulnerable group was those between 25 and 29 years, with a p-value less than 0.001. The age group above 40 years recorded an increase in maternal deaths of 90%, accounting for the general increase in death rates between the two study periods. The data below indicate maternal mortality rates by maternal ages from 2008-2009 to 2013-2014 (Data extracted from MacDorman et al. (2017).
Figure 2: Trend in maternal mortality with age
The researcher further conducted a trend analysis using race and ethnicity aspects. According to the findings, non-Hispanic black women recorded the highest maternal deaths within the two periods, with 46.7 and 56.3 deaths in the two consecutive periods (MacDorman et al., 2017). However, non-Hispanic white women had the highest increase in the mortality rate of 28%. To further their research, the authors extended their analysis to include underlying causes of maternal deaths. Research findings revealed a 56.7% increase in deaths for indirect obstetric causes and a 19.7% increase for direct obstetric causes. In addition, diabetes mellitus among pregnant women increased the death rate by about 50% (MacDorman et al., 2017). There was also an increase due to specific and non-specific pregnancy-related conditions. However, these conditions accounted for the majority of maternal deaths in the two study periods. The table below indicates the maternal death rates grouped according to the cause of death.
Despite the rise in maternal death rates due to specific and non-specific causes of death, age also brought the disparity in this cause-of-death analysis. The death rates were highest among women aged above 40 years. This age group also recorded a 26% increase in death rate from non-specific causes from 2008 to 2014. By 2014, the number of maternal mortality from non-specific causes was 20 times higher for pregnant women above forty years than those below forty years (MacDorman et al., 2017). The rates were also high in this age group for specific causes. The sensitive analysis conducted to determine the connection between over-reporting and maternal deaths revealed a direct relationship between the two variables. An increase in over-reporting by 1% led to a rise in the rate of maternal death by about 20%, with more impact on those above 40 years.
Data analysis by the researchers reveals a clear indication of significant change in the maternal mortality rate between 2008 and 2014. The maternal death rate increased in the District of Columbia and all twenty-seven states within the study period. However, only women aged above 40 years indicated a significant increase in maternal deaths. Further analysis revealed that one-third of the death cases were from women aged above 40 years. The researchers attributed this high rate to potential over-reporting of mortality rates among older women or a much greater death risk among this population. The increase in cases over time among older women suggests a condition that might worsen and increase the risk for pregnant women. More analysis revealed that the checkbox might not effectively report maternal deaths since it might be inappropriately checked. Some people might link some deaths to pregnancy even if the woman was not pregnant, leading to biased data. The coding rule would allow for recording deaths as maternal death if the women were pregnant within the past 42 days. The only exception to this coding rule is those who die through homicide, suicide, or accident. The coding rule and the use of checkbox might distort most information leading to over-reporting of maternal deaths.
Substantive Interpretation of the Findings
The research findings from the article by MacDorman et al. (2017) have a substantial impact on the study of human health concerning the effects of a lifestyle on the health of a population. The issue of death rate among pregnant women has attracted more concern over the years due to high death rates among pregnant women in the modern world. The research focuses on the trends in maternal mortality rate according to ethnicity, race, age, and causes of death. The article depicts a new perspective towards health by analyzing maternal deaths with special consideration to age, race, and causes of death. The study findings reveal the five-year trend in maternal rates in the District of Columbia and other twenty-seven states in the US.
The results in the article suggest that older people are at high risk of maternal mortality than other age groups. This is an important analysis since the data can be used as a guide when exploring the topic further. The findings can determine the exact causes of maternal mortality among the vulnerable population, which determines some of the interventions developed for this target group ( Creanga, 2018) . Maternal deaths result from complications during or after pregnancy and childbirth. Most of these complications can be prevented since they develop during pregnancy. Therefore, analyzing the current trends and determining the vulnerable population will help focus on a specific target population to reduce the cases. Identifying vulnerable groups is essential for the design of an intervention program. For instance, the result suggests a high risk of death among older mothers above 40 years ( Rossen et al., 2020) . This statistic associates properly intending to lower childbearing among older mothers to reduce the cases.
This study formed a background for further study, especially from the over-reporting of maternal mortality data. According to the research findings, an increase in false reporting had a significant increase in maternal deaths. In a different study, MacDorman et al. (2020) extended this finding to explore ways to improve maternal death reporting in the US. The public health system relies on vital statistical data since such information is used to monitor public health programs and possible interventions. There have been interests on the quality of maternal death reporting over the last years since the coding rule adopted to report such cases to have several disadvantages. The checkbox implemented in 2003 inquires whether the female decedents were: pregnant at the time of death, pregnant within the last one year before death, or pregnant within the past years ( Joseph et al., 2021) . The checkbox also includes women who were pregnant within 42 days before death. Several studies conducted on states that have implemented the checkbox on their standard death certificates indicated a potential over-reporting of maternal mortality rates by the tool. With such information, finding a connection between over-reporting and the increase in maternal mortality rate, as in the article by MacDorman et al. (2017), helps focus efforts on changing the reporting technique.
The study findings also found a connection between maternal mortality rate and specific causes of deaths such as diabetes mellitus. The combination of pregnancy and type 1 diabetes might lead to maternal death ( Joseph et al., 2021) . The death might result from diabetes, associated diseases, complications of the pregnancy, or other causes not related to neither diabetes nor pregnancy. From the clinical perspective, the greatest reward from Mac Dorman et al. (2017) research would be to determine potential causes of maternal mortality that can be prevented. Placing patients under close surveillance to monitor their pregnancy condition can help prevent some complications related to diseases. The reported cases of maternal death of pregnant diabetes mellitus patients have been approximately 0.5% ( Moaddab et al., 2018) . However, since the article focused on the period between 2008 and 2014, the data might be distorted currently. The current developments in the treatment of diabetes in the past ten years must impact the management of pregnant diabetes patients.
The article by MacDorman et al. (2017) recorded an increase in both direct and indirect obstetric deaths equivalent to 19.7% and 56.7%, respectively. Direct maternal deaths are those resulting from complications during pregnancy. The complications might also result from incorrect treatment, omission, interventions, or several events leading to the causes mentioned above. An increase in the rate of direct obstetric deaths indicates a rise in the causes of direct maternal deaths within the same period. Therefore, this research prompts the need to study the causes of direct and incidental maternal mortalities. About 75% of major complications leading to direct death include eclampsia, pre-eclampsia, infections after birth, and severe bleeding ( Davis et al., 2017) . In addition, deaths due to complications of cesarean section, anesthesia, hypertensive disorders, and obstetric hemorrhage are classified under direct maternal deaths. Indirect causes of death result from pre-existing conditions such as intensification of existing renal disease and cardiac disease. It is easy to develop an intervention when the cause of the problem is known. Therefore, the statistics developed by MacDorman et al. in their article are important when developing an intervention to lower maternal deaths.
The Authors’ Recommendations
The United States has an essential statistic system that provides necessary data on maternal deaths. Nonetheless, the big and increasing rates of maternal deaths, especially for non-specific causes and older pregnant mothers, signify possible problems in the data system. MacDorman et al. (2017) suggest more efforts to be channeled to enhnace the quality of data and reliability of checkboxes. To improve the quality of reporting, the authors recommend implementing data quality checks at the national and state levels and periodic validation of studies. In addition, federal and state agencies should train individuals who fill the death certificates on the importance of providing correct information. Honesty is a virtue that should be planted among every individual involved in state affairs. Proper completion of death certificates would help to eliminate incidences of over-reporting and misrepresentation of the population. MacDorman et al. (2017) further state that it is illogical to rely on the pregnancy checkbox only to obtain information on maternal mortality. Therefore, they suggest changes in the coding rule and pregnancy question to improve the quality of data collected.
It is important to identify and exclude incidental causes of maternal mortality as they distort the quality of data collected. Studies suggest that about 20% of maternal mortality resulting from incidental causes. Examples of pre-existing conditions that pregnancy might intensify include diabetes mellitus, malaria, hypertension, tuberculosis, iron deficiency anemia, and heart disease. Recording such deaths as maternal deaths might lower the quality of data collected since there are incidences of double recording. According to Patwardhan et al. (2016), recording of incidental causes of maternal death as maternal mortality results in a high level of biasness since it provides an inappropriate representation of a population. Currently, there has been an increase in maternal death review committees in several states. MacDorman et al. (2017) recommend using this committee to review essential data on maternal death to increase the quality of data collected. However, they state that reviewing these data would benefit healthcare if the reviews are used to revise important statistics on circumstances and causes of maternal death.
Possible Changes
The research was conducted to explore the trends in maternal death rates in an effort to depict a clear picture of the situation. The study focused on trends according to causes of death, ethnicity, race, and maternal age. The research was observational in nature. The outcomes of the data analysis supported the null hypothesis. However, changing several things would have improved the reliability of the results. The researchers would have included an intervention in their study to evaluate its effectiveness over other methods. It would have enabled them to include a control study population to authenticate the results. Besides, the authors would have divided the population into two to provide an opportunity for comparison between the study population and the control group ( Davis et al., 2017) . By doing so, it is easy to identify if the implemented changes result from the user intervention.
The length of the study duration is also a great concern. The authors used 2008-2009 and 2013/2014 in their observational study. The duration allowed for this study was not enough to produce credible results ( Patwardhan et al., 2016) . For instance, they would have conducted an annual review of maternal mortality rates over ten years. This could allow them to identify the trends every year and potential interventions and policies associated with the trends. The five-year period is not enough to collect data used to conclude about the two variables. There is a connection between maternal mortality rate and technological advancements in healthcare. Therefore, focusing on five years would not depict many changes in healthcare. Extending the study duration would provide more time for analysis of the issue under study, leading to a more concrete conclusion and recommendation. The study did not provide recommendations on ways to lower the high maternal mortality among the older population. Despite the huge disparity in maternal deaths by age, the authors only provided a solution to over-reporting. They would have provided some possible interventions to solve the primary issue, which was an increase in maternal mortality with age ( Patwardhan et al., 2016) . Doing so would have prompted further research on the topic to determine the causes and possible interventions.
Statistical Errors and Data Used
In this bio-statistical article, the authors failed to indicate the validity and reliability of research equipment used to measure and collect data. The outcome of any study depends on the validity and reliability of elements used to collect and analyze data ( Patwardhan et al., 2016) . Determining the variance and standard deviation level in the variables studied in research has a huge impact on the outcome. Therefore, calculating the mean, standard error on the variables would help determine the extent of variance. This calculation is essential in determining the statistical errors in the research and their effects on the validity of the study outcome. Statistics play an essential role in biomedical studies as it presents precise data used to make significant conclusions. Therefore, when analyzing and presenting bio-statistical data, one should know adequate statistical errors and measures. The study population involved 7,369,966 births and 1,687 deaths of pregnant women. The relationship between variables and the results indicate that the changes in the study variables may result in errors in the outcome.
The main goal of this observational research was to analyze the trends in maternal mortality rate and its connection with race, maternal age, and causes of death. To complete the study, the authors needed data on the amount of maternal mortalities per 100,000 births over the two study periods to enhance comparison. In the analysis section, the researchers used relevant data from the study findings. The researchers recorded data on maternal mortality and death rates by age, death rates by cause of death, and impacts of non-specific causes of maternal mortality on the death rate. They also recorded the death rates by maternal age for specific and non-specific causes. The researchers obtained the data from 27 states in the US. The states included Wyoming, Washington, Utah, South Dakota, South Carolina, Rhode Island, Oregon, Oklahoma, Ohio, North Dakota, New York, New Mexico, New Jersey, New Hampshire, Nevada, Montana Nebraska, Michigan, and Kansas. The data used by the researchers were appropriate for the study as they addressed the topic and answered the research question. Besides, the data also helped to make concrete conclusions and recommendations.
Availability of Data, Appropriateness of the Statistics Applied, and Additional Data
MacDorman et al. (2017) used appropriate data in their study, which was readily available and relevant to the study. They retrieved the information used in the study from the US National Vital Statistics System (NVSS). NVSS provides data that is used for international and national comparison of results on the US maternal mortality rate. The researchers used a two-proportion z-test to examine the importance of the statistics on the trend analysis. The inclusion criteria for the suitable data was any information on maternal mortality categorized by age, race, or different causes of death. The authors continued to collect data based on these inclusion criteria. The information collected was essential to making a necessary comparison of maternal death rates based on the pre-determined sub-groups. The article utilized longitudinal research design in their trend analysis and applied a two-proportion z-test. The statistical tests used in the study are appropriate in researches that study two variables to allow for comparison of the results over a certain period.
Statistical Findings and Conclusion
Statistical findings from the study revealed an overall rise in maternal mortality rate over the study periods. The increase was more intense among women aged above forty years and for non-specific causes of maternal mortalities. The high rates among the older population were associated with possible over-reporting of cases. The study found a direct connection between an increase in false reporting and the rise in mortality rates. As a result, the researchers recommended some measures to lower the cases of over-reporting. According to them, changing the coding rule and pregnancy questions would improve the quality of data collected. Racial trend analysis of the issue suggested an rise in maternal death among non-Hispanic blacks than non-Hispanic whites. Statistical analysis of this article reveals important information that can be used to improve the quality of care provided across the country.
References
Allanson, P. E., & Notar, C. E. (2020). Statistics as Measurement: 4 Scales/Levels of Measurement. Education Quarterly Reviews , 3 (3), 375-385.
Creanga, A. A. (2018). Maternal mortality in the United States: a review of contemporary data and their limitations. Clinical obstetrics and gynecology , 61 (2), 296-306. https://doi.org/10.1097/GRF.0000000000000362
Davis, N. L., et al. (2017). Contribution of maternal age and pregnancy checkbox on maternal mortality ratios in the United States, 1978–2012. American journal of obstetrics and gynecology , 217 (3), 352-e1. https://doi.org/10.1016/j.ajog.2017.04.042
Hoyert, D. L., & Miniño, A. M. (2020). Maternal mortality in the United States: changes in coding, publication, and data release, 2018. https://stacks.cdc.gov/view/cdc/84769
Joseph, K. S., et al. (2021). Maternal Mortality in the United States: Recent Trends, Current Status, and Future Considerations. Obstetrics and Gynecology , 137 (5), 763. https://dx.doi.org/10.1097%2FAOG.0000000000004361
Lu, M. C. (2018). Reducing maternal mortality in the United States. Jama , 320 (12), 1237-1238.
MacDorman, M. F., et al. (2020). Improving US maternal mortality reporting by analyzing literal text on death certificates, United States, 2016–2017. PloS one , 15 (10), e0240701. https://dx.doi.org/10.1371%2Fjournal.pone.0240701
Moaddab, A., et al. (2018). Health care disparity and pregnancy-related mortality in the United States, 2005–2014. Obstetrics & Gynecology , 131 (4), 707-712. https://dx.doi.org/ 10.1097/AOG.0000000000002534
Patwardhan, M., et al. (2016). Maternal death: case definition and guidelines for data collection, analysis, and presentation of immunization safety data. Vaccine , 34 (49), 6077. https://dx.doi.org/10.1016%2Fj.vaccine.2016.03.042
Rossen, L. M., et al. (2020). The pregnancy checkbox and misclassification impact maternal mortality trends in the United States, 1999–2017.
