Research Critique of Quantitative Research on Nursing Practice

Nurses are expected to provide the best standards of care possible for their patients and clients, and in order to do this, they are required to provide evidence-based practice wherever possible. Part of this process of providing care based on the best available evidence involves appraising primary research (Elliott, 2001, p 555). If nurses are to improve their practice, and apply evidence to improve their clinical and theoretical knowledge and skills, they must be able to assess the quality of the available research which is relevant to their practice (Freshwater and Bishop, 2003k p23; Hek, 2000, p 19). According to Hek (2000 p 19-21), evidence based practice incorporates professional expertise, patient need and preference, and the best available evidence. But in order to identify this ‘best evidence’, the nurse must undertake an evaluation and critical review of research studies, to see if the research is useful and of sufficient quality to be applied to their practice (Fink, 2005).

This essay evaluates a quantitative research article which has relevance to nursing practice, because it deals with a chronic condition and one which is prevalent in worldwide populations. The author will review Tangkawanich et al (2008 p 216) ‘Causal model of health: health-related quality of life in people living with HIV/AIDS in the northern region of Thailand.’ This article is published in a reputable journal, The Journal of Nursing and Health Sciences which addresses issues of interest and concern to nurses internationally, and as such, offers specific insight into HIV/AIDS and nursing within a global context.

There are a number of appraisal tools available to healthcare practitioners to assist them in analysing and critiquing primary research articles. Such models are developed over time in relation to standardised conceptualisations of what constitutes quality and rigour in research and in its reporting. For the purposes of this essay, the primary tool used is that proposed by Cormack (2000), but the author will make reference to other critiquing guides and information, including the popular CASP tool (PHRU, 2009, online)

Discussion

1. Title

The title is concise (Cormack, 2000) and describes the focus of the research itself. While it clearly indicates what the purpose of the study was (Cormack, 2000), it could be clearer and more indicative of the nature of the study. While the nature of the research in setting out to ‘examine the causal relationships between age, antriretroviral treatment, social support, symptom experience, self-care strategies and health related quality of life’ (Tangkawanich et al, 2008 p 216) in the chosen sample and population, is apparent in the abstract, there is little indication of these particular variables in the title, although the description of the ‘causal model’ (Tangkawanich et al, 2008 p 216) does indicate the nature of the research. The use of the word ‘causal’ (Tangkawanich et al, 2008 p 216) also suggests that this is a quantitative research article. It does not clearly indicate the research approach used (Cormack, 2000). To the less research-aware reader, it would be difficult to divine this information from the title alone, and it could be argued that it would be better to include in this title a clearer indication of the nature of the study. This would then help the reader to identify if this is the type of research study that would be applicable to their own practice or learning.

2. Authors

The authors’ qualifications are provided, and they all are registered nurses, all of which have postgraduate degrees, and two of which have doctorates. They all work within nursing education within a University context. This would suggest that they have the research skills and expertise to carry out such a study. However, there is no indication in the author list whether or not any of them have the statistical expertise for the study.

3. Abstract

Tangkawanich et al, (2008 p 216) provide an abstract, which is identified by Cormack (2000) as an important introduction to the article. The study effectively summarises the research, by identifying the variables being tested. The authors do not, however, present the hypothesis in the abstract (Cormack, 2000). The abstract contains a summary of the study sample, and also identifies the research tools that have been used. They include the results and a summary of conclusions from these findings. As such, the abstract does represent the article itself (PHRU, 2009, online), and for the reader, it does make it easy to identify whether or not the article is relevant to their interest. In particular, it does indicate clearly that it is a quantitative paper which uses recognisable data collection tools.

4. Introduction and Literature Review

Although Cormack (2000) separates these two into distinct subheadings, within this article, the introduction and literature review are contiguous. The author has noted that this is often the case in the reporting of such studies, but this may simply be a convention of the publication itself, and not the preference of the authors of the study. The authors use the introduction to contextualise the problem in relation to published research, stating the importance of health-related quality of life (HRQL) for people living with HIV/AIDS (PLWHA), because of the impact of the disease on these individuals’ daily lives (Tangkawanich et al, 2008 p 216). They discuss changes and advances in treatment options for this condition, and relate this to HRQL, and then discuss the disease itself, and how these impact upon HRQL (Tangkawanich et al, 2008 p 216). They summarise some research about this topic, and also look at self-care strategies, symptom management and treatment (Tangkawanich et al, 2008 p 216). There is some exploration of HRQL and its relationship to nursing and to existing literature on this topic, which aids understanding of the concept prior to reading the rest of the article. They highlight some important topics in relation to the focus of the article, including treatment, social support, and other issues (Tangkawanich et al, 2008 p 216).

Although the introduction/literature review contextualises and introduces this study, this author believes that a more detailed critical analysis of the literature would be warranted here. It is not enough to cite previous research as a means of establishing the credentials of the study, as it were. A wider range of research could have been included Gerrish and Lacey, 2006, 38; Fontana, 2004, p 93), and this research could have been evaluated to identify its quality. It is left to the reader to pursue this matter and determine the quality of the research upon which they base the premise and justification for this study. This could be considered a limitation in the reporting of this research.

5. The Hypothesis

The authors do not state a hypothesis (Cormack, 2000), as such, but instead present a research question. ‘The purpose of this study was to examine the causal relationships between age, antiretroviral treatment, social support, symptom experience, self-care strategies, and the HRQL in Thai PLWHA’ (Tangkawanich et al, 2008 p 217). The nature of this statement would suggest that it is not an experimental study, but that it is within a quantitative research paradigm.

6. Operational definitions

Cormack (2000) suggests that people appraising research question whether operational definitions are clearly presented. In this study, operational definitions are explained within the introduction but in language that would make it inaccessible to the less experienced or less knowledgeable reader.

7. Methodology

The methodology section does not clearly state or discuss the choice of a quantitative approach (Cormack, 2000). The focus of the section on Methods is rather on the instruments that are being used. The quality of this study seems to rest in the choice of a quantitative approach, and the choice of data collection instruments. Quantitative research approaches offer a better standards of evidence, with generally greater ability for replication and greater rigour (Kitson et al, 2000 p 149; Duffy, 2005, p 233). As far as research for healthcare practice is concerned, quantitative studies hold better status than those based with a qualitative paradigm (Hek, 2000 p 19; Newman et al, 1998 p 231; Pepler et al, 2006, p 23). There is however no real discussion of the underpinning principles of quantitative research (Parahoo, 2006).

It is good that the research instruments are explained in such detail, because it helps overcome one of the limitations of quantitative research, that of not asking the right questions to elicit answers that relate to person al experience (Johnson and Onwuegbuzie, 2004 p 14).

8. Subjects

There is very limited detail given of sample selection, save stating that eight hospitals were randomly selected using a ‘lottery method without replacement’ and that the 422 participants were randomly sampled (Tangkawanich et al, 2008 p 217). This could constitute a limitation of this research, as it is not possible to identify if there was any sampling bias, how participants were recruited, who recruited them, and any ethical issues in relation to participant recruitment (Hek, 2000, p 20; PHRU, 2009, online, Bowling, 2002). This author would argue that this is a weakness of the study, as these are crucial elements of quality measurement in primary research within healthcare (Austin, 2001 p 1; Cooper, 2006, p 439; Nuremberg Code, 1949, online).

9. Sample selection

Sample selection is not discussed in any detail, which could be a weakness of the study, as mentioned above (Cormack, 2000). Sample size is stated, but it is not stated whether this was statistically determined, which could also be considered a weakness, as achieving a statistically sound sample size is important within quantitative research (Daggett et al, 2005, p 255; Donovan, 2002).

10. Data Collection

Data collection procedures are not adequately described in this study (Cormack, 2000). This would be problematic for replication and for rigour. There is no indication who administered the questionnaires, whether or not they were self-report in the participant’s own time and convenience, or whether a researcher was present at the time of completion. This could be considered a weakness of this study’s reporting. The issue of researcher bias is important in the completion of data collection tools, and while questionnaires may be considered a way of avoiding this, if they are remotely administered, it is not always possible to check they are full, or honest, or completed by the intended target (Gillham, 2000, p 48). Having the researcher present, however, could introduce bias or influence of some kind, particularly in vulnerable people (Bowling, 2001). As these are vulnerable adults attending clinics for their chronic condition, not discussion how the data was collected from them is a serious failing and may also constitute an ethical issue.

However, a strength of this study could be considered to be the use of multiple data collection instruments, and the detail with which they are described, and their provenance accounted for. They have used the Social Support Questionniare, the Symptom Experience Questionnaire, the Self-Care Strategies Questionnaire, and the Health-Related Quality of Life Questionnaire (Tangkawanich et al, 2008 p 217). These are described in detail, and where they are based on previously developed research or derived directly from previous studies, this is also described. As these are established research data collection tools, this would suggest they have been previously validated, which adds to the quality of the research (Yu and Cooper, 1983 p 36; Oppenheim, 1992; McDowell and Newell, 1996; McColl et al, 2001 p 1). The demographics of the sample are addressed to some extent, and the use of multiple tools also helps to address potential confounding variables or factors (PHRU, 2009, online).

11. Ethical Considerations

It is clearly stated within the study that ethical approval was derived from an appropriate body, and that informed consent was achieved, both of which signify good ethical consideration here (Cormack, 2000). However, there is no explanation of what information was given to the participants, how informed consent was achieved, or if there was any issue with communication or accessibility for people with different communication needs. It does not specifically address issues of anonymity and confidentiality (Cormack, 2000), but instead seems to focus on safety (Tangkawanich et al, 2008 p 218).

12. Results

The results of this study are presented in tables, diagrams, and in great detail within the text of the paper. The key results and demographic issues are presented, but data presentation is summarised in a results section and then in more detail in the discussion. The findings are not very accessible, but p values are clearly stated, which is important in a study of this kind. PHRU (2009, online), within the CASP tool, poses the question of ‘do you believe the results?’. This is an important question. The results seem plausible, and relate to established statistical analysis procedures (see below). But because of the lack of detail about the sample, and the selection method, it is not possible to eradicate the doubts about these findings, in relation to potential bias. But in the context of the author’s wider knowledge and understanding about people living with HIV/AIDS, the results seem believable. However, the issue of bias cannot be overlooked. More transparency in reporting of key elements of this study would have made it easier to determine whether these results constitute good evidence for practice (Rosswurm and Larrabee, 1999 p 317; Pepler et al, 2006, p 23).

12. Data analysis

The authors carried out the statistical analyses using SPSS, which is an established statistical programme, and LISREL, which is not a programme this author is familiar with. They describe generating descriptive statistics for each of the variables under consideration, but do not present these in any detail (Tangkawanich et al, 2008 p 218). They then describe the use of multivariate analysis for specific model development, and using other tests, some of which are familiar to the author, some of which are not. While the multivariate analysis and X2 tests are acceptable tests of inference or relationships between variables (Duff, 2005 p 234), anyone who does not have the specialist knowledge of the other tests would find it hard to determine their appropriateness here. The level of specialist statistical knowledge required to understand this would be significant. More transparency could have been achieved by including an explanation of these tests.

13. Discussion

The discussion appears quite balanced (Cormack, 2000), and is very detailed, which is good, given the complexity of this study, involving as it does multiple data collection instruments. The study relates the findings back to a wide range of other research studies, which is a strength of this part of the report, showing congruence with many other findings in relation to quality of life, age, socioeconomic status, social support, antiretroviral treatment, symptoms, and self-care (Tangkawanich et al, 2008 p 219). Limitations of the study are also acknowledged (Cormack, 2000). However, the implications for practice are presented in a separate section.

14. Conclusions

The conclusion supports the results obtained (Cormack, 2000), although it is not a very detailed summation of the complexity of the findings. However, this is referred to in the ‘Implications and Recommendations’ section. (Tangkawanich et al, 2008 p 217).

15. Recommendations

The authors of this study suggest that ‘social support’ would have the strongest effect on HRQL(Tangkawanich et al, 2008 p 220). They recommend a wide range of strategies to support this, and also look at all the other factors, making recommendations for improving care for this patient population. These are all consistent with the study. However, the only recommendation for future research is that it should be longitudinal rather than cross-sectional (Tangkawanich et al, 2008 p 217).

Conclusion

It is apparent from this study that social and other factors have a serious impact upon the lived experience and quality of life of the target population. It is also apparent that social factors may outweigh medical factors, other than in the provision of antiretroviral therapies for these patients. It would appear, therefore, that understanding these factors and the relationships between them could improve nursing and healthcare practice for people living with HIV/AIDS. However, this author would also conclude that because of some methodological and reporting weaknesses in this study, it would be better to find other research confirming these findings before using it as evidence for practice.