Critiquing Nursing Research - Antenatal Depression

Research is diligent, systematic inquiry or investigation to validate and refine existing knowledge and generate new knowledge that directly and indirectly influence nursing practice, Burns & Grove(2007). To judge limitations, significance, strength and weakness of any study or research, we need to review all the aspects of that study comprehensively and that is critique. The research article, on which I am going to critique, has been published in Acta Psychiatr Scand. June 1, 2007. The title of this article is “Association between antenatal depression and low birthweight in a developing country”. In my point of view the title explicitly defines the purpose and setting of the study but it should have mentioned which area the authors want to focus in developing country.

According to Norwood (2010), a research problem is the problematic issue that is the focus of a study. The description of a research problem identifies who and what is being studied, and the issue’s relevance and significance, or why it needs to be studied. In this article the author have mentioned the problem statement in the abstract section under the heading of objective that the prevalence of depression in south Asian women is very high. This statement captures the readers’ attention and then the authors have explicitly explained that they intend to study the association between antenatal depression and low birth weight. Furthermore, the authors have identified that the problem is under-researched in developing countries. Therefore the problem they have highlighted is researchable. Variables are well defined and data can be easily collected and analyzed on the problem. Moreover, although their previous studies have shown the strong association between maternal depression and later growth retardation in infants in Pakistan but the association with low birth weight was not studied before. In addition to this, authors have mentioned a recent study in India that self-reported maternal psychological illness is independently associated with low birth weight (LBW) so it’s a thought provoking issue that needs to be studied in Pakistan. The problem is surely significant to nursing. According to Burns & Grove (2007), the research in nursing must address the questions relevant to nurses and must develop the unique body of knowledge for practice. Based on identified problem, health care personnel may conduct studies with various methodologies to describe, explain, predict and control phenomena that are essential to nursing. Therefore, in this article, although the researcher has not explicitly defined problem’s significance to nursing but it gives an overview to health care personnel to monitor antenatal women for depression ultimately this will help nurses to reduce incidence of LBW and provide specific attention to them as a part of routine prenatal care. Moreover, a solid research base will provide evidence of the nursing actions that are effective in promoting patients health.The background information of the problem is well presented in introduction paragraph. The background of any study gives researcher an opportunity to identify problem and the need for further study.The authors have clearly stated that in the developed countries, antenatal psychological morbidity is associated with poor birth outcomes. Whereas in developing country no researches has been done. In addition to that, as stated earlier that they have mentioned a recent study from India found that maternal psychological illness during pregnancy can lead to LBW.

Review of the literature is relevant to research topic and it is done extensively. According to Burns & Grove(2007), the good review usually begins with an introduction which identifies the key words used to conduct the search and information about which databases were used. In this article the authors have begun the literature review from the introduction part and related the previous studies findings which help to identify any gaps in the literature relating to the problem. For instance, researcher has stated that the study has shown in developed countries that the maternal psychosocial stress and anxiety during pregnancy are associated with poor birth outcomes. Further they identified the need for this study to be done in the developing countries because nobody had done this before. Secondarily the prevalence of depression is significantly high in South Asian women. Hence, in order to identify the association between antenatal depression and low birth weight in developing countries they conducted this study. The majority of studies are of recent origin and the sources which were used in this study are relevant and current according to the publishing year. Although, the author has used some old sources too but there may be exceptions to this, for example, in areas where there is a lack of research, is still relevant to current practice, Burn & Grove(2007).The documentation of sources in this article is clear and complete in reference section, although in text citation of sources is not evident because author has not followed APA style. The relationship of the problem with previous research is clear. It is discussed earlier that the authors have identified the association of antenatal depression with low birth weight while they were studying maternal depression and later growth retardation in infants in Pakistan.The appropriate summary statement is not clearly added in literature review but it ends with its implications for the problem. The authors have discussed the implication of problem under the heading of discussion that the low birth weight babies are more prone to get growth retardation early in their live and poor intellectual development. Besides this they are at risk of developing depression in adolescence. According to Pilot & Beck (2006), literature review can instigate new research ideas and help to set the foundation for further studies.

As said by Polit and Beck (2006), the purpose of the study should at least broadly indicate to the reader what is to be studied. The purpose should be clear through the title of the research. As already mentioned earlier, the title of this article explicitly indicates what the researcher wants to study but the setting where the study is going to be conducted is not covering the major part of the developing country. In my point of view, the purpose is clearly stated and congruent with the data presented in the literature. The researchers stated under the heading of study population that the data is collected in the rural sub district of Rawalpindi, Pakistan by the trained, government-employed Lady Health Workers (LHWs) working in that area. In this article the formal hypothesis statement is clearly mentioned which should be there in cohort study. The authors have stated under the heading of aim of the study that in developing country the antenatal depression is associated with low birth weight. As Pilot & Beck (2006) declare that hypothesis is a statement of a predicted relationship between two or more variables. A testable hypothesis states the anticipated association between one or more independent and one or more dependent variable. Cohort study comes under the umbrella of analytical studies which helps to generate and test hypothesis.The researchers have mentioned the conceptual definition of relevant terms and variablesat various places in the article.According to Pilot & Beck (2006), a conceptual definition provides the theoretical meaning of a variable. For example, authors have used low birth weight (weight at birth of less than 2500g) for antenatal depression they have mentioned maternal depression during pregnancy but they have not given any specific definition for that. However, authors have not cited anywhere in the article neither in the references that from where they have taken these definitions and what are the sources. Simultaneously, authors have provided operational definitions of variables and difficult terms. The operational definition of variables is defined in the procedure part. They have assessed mental state in the third trimester of pregnancy by using Schedules for Clinical Assessment in Neuropsychiatry (SCAN). This tool was developed by World Health Organization to assess Depressive disorder. Similarly, for theassessment of low birth weight they used spring balance Salter Scale. Furthermore, the authors have used likert scale to assess the economic and social status of the sample.

According to Coughlan et al(2007), theresearch design refers to the nuts and bolts of how a research study is undertaken. In this article the study design is clear in abstract that is prospective cohort study.In cohort study we see basically exposures which lead to certain outcomes. This design is good for assessing the exposure for certain outcomes in terms of relationship, cost wise it is expensive, changes can take place over time but it can loss the follow-up as changes can take place between exposure and outcome assessment. Moreover, they have not explained the reason for choosing this study design. A prospective cohort study is an epidemiological study in which a group of people are identified who are at risk for experiencing a particular event. The study design is appropriate to the problem as cohort study also referred as follow up study that requires monitoring of the participants continuously and it is appropriately done by the researcher. In addition to that according to Burns & Grove (2007), some of the research design are causality, multicausality,probability, bias, control, manipulation and validity.Causaility basically defines causes that lead to effects and examine relationships among varaiables. The authors have discussed under the heading of introduction that maternal psychological illness during pregnancy can lead to poor birth outcomes.A number of interrelating variables can be involved in producing a particular effect.The presence of multiple causes for an effect is reffered to as multicausility like poverty, educational status, antenatal care etc are mentioned in table 2in this article. The authors have very well controlled these confounding factors.

Additionally, the target population is described explicitly in the article. The inclusion and exclusion criteria are well defined under the heading of study population. The sample requirement was 154 infants of depressed and same number of non-depressed mothers. The researchers have taken 701 women in their last trimester of pregnancy and 632 out of them were followed. This shows that the authors have provided enough space for attrition.Furthermore, the researchers have mentioned response rate in percentages which is a good part of the article. According to Burns & Grove (2007), the adequacy of the sample size is evaluated using a power analysis. Power is the capacity to correctly reject a null hypothesis. The minimum acceptable level of power for a study is 0.8 or 80%. The researchers have clearly mentioned power in statistical analysis part. In this article the level of significance (alpha) is 0.05 and CI is 95%, therefore there is only 5% chance of a type II error which evaluates that sample is adequate. The sample size and major characteristics are appropriate but to generalize from the sample to the population, the sample has to be representative of the population. The safest way to ensure that is representative is to use a random selection procedure which is not done by the researchers. I believe that author has done non probability sampling i.e. convenient sampling. Although non probability sampling could be done in quantitative research but when the sample is not the representative of the population, sampling bias is a possibility as random community is not selected. Therefore, it should be evident how the sample was selected which is not explicitly defined in the article.According to Burns & Grove (2007), convenient sampling is including subjects in the study who happened to be in the right place at the right time, with addition of available subjects until the desired sample size is reached. Furthermore, the authors have paid an additional incentive to LHWs for every subject identified this could create a selection bias in the study. Besides that the data was verified only in some villages so the accuracy of data is questionable.

As we move forward to the instruments part, there are three instruments used in the study to collect and analyzed the data that are Schedules for Clinical Assessment in Neuropsychiatry(SCAN) to assess depression, Spring Balance weighting scales to obtain weight and likert scale to assess socioeconomic status.However author was unable to highlight the content, strengths and weaknesses of instruments in detail. The SCAN tool was from WHO which gives the authenticity to the tool but the strengths of rest of the two are not mentioned. The rationale for choosing spring balance Salter scale was mentioned that it was similar to the one that was routinely used by the LHWs. The author used SCAN because it is internationally validated and used to diagnose the depressive disorder. The rationale for using likert scale was not mentioned in the article. Validity of spring balance weighing scale is also not acknowledged in the article. How the reader would be able to know that the zero of weighing scale equipment was properly checked before measurement or not. This could be one of the measurement error i.e. systemic errors for measurement tool. Similarly, the validity for likert scale is not mentioned. According to Polit and Beck 2006, the researcher should indicate how the validity and reliability of the adapted instrument was established. (p. 365). The reliability of SCAN was mentioned as Kappa 0.90 which is nearest to 1 so this tool is reliable but the rest of two are doubtful in terms of reliability.

In quantitative study any number of strategies can be adapted when collecting data. In this article, the data collection procedure is appropriate for the study. According to Polit and Beck (2006), questionnaires are the most commonly used data gathering instruments and consist mainly of close ended questions. In this study, the authors have used structural means closed ended questionnaire for the collection of data which is appropriate for the study.According to Burns & Grove 2007, the actual steps of collecting the data are specific….that threatens to disrupt the study. Similarly, in this article the researchers focus was on pregnant women, two trained and experienced clinicians and government employed LHWs were there who took interview and collected data in a consistent way, maintained research control by setting inclusion and exclusion criteria, confounding factors and tried to maintain reliability but they were not able to define validity which gives the credibility to any research. Furthermore, the subjectivity in data collection about socioeconomic status is there as the LHWs who rated the house hold on likert scale may be in real the households are not that poor. This can create an observational bias to the study as the households were rated on the LHW’s assumption.Chances of recall bias is also there, as mentioned by the author in discussion part that their measures of relative poverty are biased as the depressed women might recall debt due to their poor mental state.Protection of human rights was a concern for researchers. It is evident by following considerations. Firstly, ethical approval was given from Research Ethics Committee for the study. Secondly, women participated voluntarily and written informed consent was taken from them which are mentioned under the heading of procedures. Local health workers obtained consent on behalf of non-literate participants. The comprehension of the components of consent was nowhere mentioned in the article.According to Polit & Beck (2006), informed consent means that participant have adequate information regarding the research, comprehend the information and have power of free choice enabling them to consent voluntarily to participate in the research or decline participation. In my point of view researcher should acknowledge what they have communicated to the participants while taking consent. As per my opinion the study is replicable. The researchers have described the setting, methods and findings in details under the heading of statistical analysis and results. In future if another investigator would like to carry out investigation using similar participants, setting and variables then similar findings can be used.The best part of the article is that appropriate limitationsand assumptions of study are clearly mentioned. The researchers have explicitly stated under the heading of limitation that because of the small sample size the generalizability is not guaranteed and the data was not collected on maternal anemia which can be a potential confounder. The significant assumption is also mentioned that maternal depression during pregnancy can lead to low birth weight.

The choice of statistical procedure is appropriate. The researchers have performed chi square, Mann Whitney U-test and logical regression for the distribution. In my point of view, the researchers have used inferential statistical test to identify the relationship or difference between variables. Mann-Whitney U test is also equivalent to a chi square test that is non-parametric means which is not normally distributed.Another test which is applied in the study is logical regression which is used for categorical outcomes i.e. relative poverty, maternal empowerment etc. The importance of regression in a study is to control the confounding or extraneous factors, those factors are well managed by the researchers in the paragraph of statistical analysis. Moreover, table 1 illustrates the comparison between depressed and non-depressed mothers in terms of sociodemographic factors. The table is clear but author could have showed mean in the table to make it clearer. In addition to that authors have explicitly mentioned in results that they have found significant association between antenatal depression and low birth weight after controlling all confounders. They have further explained that the relative poverty is the second significant factor after antenatal depression and these two factors are showing the significance in table # 2 with P value which is less than 0.01 for both factors. According to me, statistical procedures are appropriate to the study variables and the statistical procedures are correctly applied. Moreover the authors have calculated the relative risk in table 2 which is the measure of association for cohort studies. If we see the results the antenatal depression relative risk is 1.9 and relative poverty is 1.6 which again shows the positive significant association with LBW. In table 3 the authors have calculated the odds ratio which is usually done in case control studies but it can be done with relative risk in the cohort studies. The odds ratio findings also remained significant for antenatal depression. According to me the tables are clear and the findings mentioned are similar to findings mentioned in results.

Generalization is the extension of the implication of the research findings from the sample to a larger population. The generalizability is questionable, because the authors have not randomly selected the community as it has been discussed earlier. Researchers have well interpreted the findings and explained the effects of antenatal depression and low birth weight according to the findings of the research.Conclusion is not stated in the article except abstract. The significance of findings should reflect in the conclusion. Authors should infer study findings in the conclusion which is not incorporated. I believed that conclusion should be stated in a separate heading so that the reader would be able to find the reflection of study findings in it. Recommendations in this article are very feasible and appropriate. The authors have explicitly mentioned the recommendations in the discussion part in concluding paragraphs that diagnosis and treatment of depression during pregnancy can reduce burden on mother as well as promote physical and mental health of off-spring.