Critical of a Research Article - Literature review Example

Critical Review of a Research Article 2008 Critical Review of a Research Article Aiken, L. H., Clarke, S. P., Sloan, D. M., Sochalski, J., & Silber, J. H. (2002). Hospital Nurse Staffing and Patient Mortality, Nurse Burnout, and Job Dissatisfaction. Journal of American Medical Association, 288(16), pp. 1987-1993. I. Review of Literature 1. The authors provide a very concise review of literature: instead of analyzing and synthesizing the most credible research in the field they briefly summarize the results of some surveys and analytical works. The selection of sources Aiken and colleagues refer to to support their argument is also not perfect: the list mostly includes newspaper and journal articles though the authors also provide the results of at least two surveys carried out by highly credible organizations such as Kaiser Family Foundation and Joint Commission on Accreditation of Healthcare Organizations. 2. The lack of detailed literature review barely allows to somehow link Aiken’s findings to a wider framework of information available in the field. Although the study’s objective (to determine the association between the patient-to-nurse ration and patient mortality, deaths following complications, and factors related to nurse retention) clearly suggests that the authors’ intention is to reveal and detail the root causes of the recent nurse shortage, absence of a serious literature review does not make it possible to find out whether this objective is an actual extension of the existing knowledge base. By contrast, a solid literature enables the reader to identify gaps in the relevant research and properly justify the need for another study in this field. 3. Although the authors do not state any specific rationale for the study there is simply no need to do so: the rationale is absolutely clear. The issue of nurse shortage is not new to the scholarly literature though Aiken and colleagues provide only a very brief insight into the existing publications. A growing body of empirical evidence is available to support the notion that effectiveness of the health care system largely depends upon the nurses. At the same time, there are few studies that attempt to clarify the root relationships between nurse retention rates and job satisfaction and burnout. Therefore, expanding the body of knowledge on this phenomenon is an important task building opportunities for better performance in hospitals. II. Definitions 1. The authors provide no operational definitions of the independent and dependent variables under study. However, the lack of definitions does not negatively affect the study: the variables are simple and do not necessarily require definitions. Furthermore, the overall purpose of the study and the study’s design (cross-sectional research) does not require to clarify/disprove/validate any definitions. Aiken and colleagues simply identify the three control variables, namely size, teaching status and technology of the hospitals included in the study. In order to adjust the results on mortality and failure-to-rescue dimensions Aiken and colleagues have used 133 variables. They list only the key ones among them: age, sex, surgery types, and dummy variable indicating the presence of chronic preexisting health conditions reflected in the ICD. The final set of control variables was produced after a selection process that paralleled. III. Hypothesis 1. In the introductory paragraphs the authors identify the key purpose of their study, namely to determine the association between the patient-to-nurse ratio and patient mortality, failure-to-rescue (deaths following complications) among surgical patients, and factors related to nurse retention. This purpose gives a clear notion about the assumption/hypothesis the study is built on, although no clear statement of it is provided by the authors. Evidently, the hypothesis is that that higher patient-to-nurse ratio positively correlates with patient mortality, failure-to-rescue among surgical patients and negatively correlates with factors related to nurse retention. 2. The hypothesis logically stems from the growing concern that overall hospital understaffing (and nurse shortage in particular) negatively affects the quality of health care in the US. Although the literature review is rather concise, even those few publications cited by Aiken and colleagues provide the overall picture of the problem. 3. The authors state no null hypothesis, but the default null hypothesis is that higher patient-to-nurse ratio does not correlate with patient mortality, failure-to-rescue and nurse retention rate. IV. Methodology A. Design 1. The authors employed non-experimental research as the most appropriate design for their study. The choice is fully justified given the specific purpose of their research. 2. Cross-sectional research has been chosen among the non-experimental designs. 3. Since the authors do not employ any experimental design many of the majority of internal validity threats is not relevant. The relatively low response rate (52%) among the participants, coupled with the exclusion of vascular surgery patients and non-licensed nursing staff were the most essential factors that might undermine validity of the study. The authors believe that neither of these limitations is serious enough to deteriorate the findings. The comparison of their results with the findings reported by other credible agencies substantiates this claim. For example, the level of nurses’ dissatisfaction revealed by Aiken and colleagues coincides with results of the National Sample Survey of Registered Nurses 2000 on this dimension. The use of multidimensional statistical adjustments (size of hospitals, technological level, teaching status, etc) and other statistical instruments (e.g. dummy variables) reduces the effects of the potential internal validity threats to minimum. Statistical adjustments made by Aiken and colleagues helps them consider each and every of the factors that might otherwise confound the results. Thus, failure-to-rescue for patient characteristics and comorbidities was accomplished by using 133 variables, including age, sex, surgery types, and dummy variables, and final set of control variables was selected using the previously reported and tested method B. The Sample 1. Since the authors adopt cross-sectional design which is fast and can study a large number of patients at relatively little cost and short time the key purpose is to cover the whole population. Therefore, the sample is random and its impressive size makes is fully representative. 2. The only population to which findings can be generalized is the nurse population: the purpose of research does not leave any other option. 3. The sample was selected via mail survey of registered nurses who were on the Pennsylvania Board of Nurses rolls and resided in the state. However, no specific inclusion criteria were applied. 4. The overall sample included 10184 staff nurses and 232342 general, orthopedic, and vascular surgery patients discharged from the hospital between April 1, 1998, and November 30, 1999. Administrative data from 168 nonfederal adult general hospitals in Pennsylvania was analyzed. Evaluate sample size and how representative is was of the population. 5. Aiken’s sampling strategy ensured correct unbiased selection of participants. Impressive size of the sample which included 10184 nurses and 232342 patients contributed substantially to the study’s representativeness of the total nursing population in relation to the factors being studied. Therefore the authors’ claim that their findings can be generalized to the overall nurse population are absolutely justified. Probably the most problematic issue in terms of internal validity is exclusion of non-licensed nurses from the study. Since this population constitutes an essential part of the US nurse population further studies may be required to fully understand the relation between this population and the factors examined by Aiken and colleagues. C. The instruments 1. Aiken and colleagues describe the process of data collection comprehensively so that the reader gets full information about every instrument and factor involved. The authors used data provided by the 1999 American Hospital Association (AHA) Annual Survey and the 1999 Pennsylvania Department of Health Hospital Survey. 2. The authors used a standard set of instruments to collect the data. On the stage of data collection, job satisfaction was measured using a 4-point scale from very dissatisfied to very satisfied while the burnout parameter was measured with the help of the Maslach Burnout Inventory. Both these instruments have a history of successful use. 3. The reliability and validity of tools employed by the authors can barely be questioned. The Maslach Burnout Inventory has been designed specifically to measure burnout in a variety of human services occupations including nursing and has been repeatedly validated as a reliable instrument (Maslach et al. 1996). Standard errors are corrected using the Huber/White procedure traditionally used in statistical analysis for such purposes. D. Ethical considerations 1. The author’s fully protect the rights of participants by maintaining absolute confidentiality. No personal data is disclosed in the study. 2. The nature of the study that seeks to identify the overall tendency which means the staff was required to provide only very general data. Therefore, the probability of any psychological or other discomfort sustained by subjects due to the necessity to reveal some very personal information is absent. 3. Nothing could have been done that wasn’t done to maintain ethical standards: the expertise and huge experience of the researchers suggests the ethical aspect is on the highest level possible. V. Analysis of Data 1. On the stage of data analysis, logistic regression models were used to assess the effects of staffing on job dissatisfaction and burnout; Huber-White procedures were used to estimate the regression models (to account for the clustering of patients within hospitals and to correct the standard errors). 2. The authors provide only three tables, but each of them perfectly summarizes the textual information provided in the Results analysis section of the article. 3. The findings generated from the data using the reliable and tested methods of statistical analysis fully supports the hypothesis that higher patient-to-nurse ratio positively correlates with patient mortality, failure-to-rescue among surgical patients and negatively correlates with factors related to nurse retention. 4. The researchers use the level of statistical significant standard for such type of study designs, namely 0.05. VI. Discussion 1. Aiken and colleagues found out that each additional patient per nurse was associated with a 7% increase in the likelihood of dying within 30 days of admission and a 7% increase in the odds of failure-to-rescue. Also, each additional patient per nurse was associated with a 23% increase in the odds of burnout and a 15% increase in the odds of job dissatisfaction. Such findings are fully congruent with the assumption/hypothesis implied by the authors in the introductory section. Therefore, the conclusion that higher patient-to-nurse ratio positively correlates with patient mortality, failure-to-rescue among surgical patients and negatively correlates with factors related to nurse retention is fully based on the data collected and analyzed within the framework of study. Besides, the authors used data captured by third parties. Although credibility of those parties can hardly be questioned, in this case the authors rely on third-party data on the initial stage of their study, namely grouping the hospitals into three different classes. Thus, no information is available regarding the methods of classification used by those third parties (e.g. AHA, PDHHS), no definitions of technological level and teaching status (the key variable used by Aiken) is provided. Since these three variables were used as control variables this potential consideration might be crucial. However, the authors’ qualifications and professional background practically exclude any possibility of non-reliable information being used in the study. 2. Aiken and colleagues did not use any specific literature to explain their findings. However, such decision is probably justified by the fact that their study is rather innovatory and there is simply no serious publications that can be used to support and fully explain their findings. Nonetheless, the do cite several sources their findings are congruent with. 3. Evidently, the authors are fully aware of the limitations pertaining to their study, particularly in terms of its design. Cross-sectional study is highly effective at revealing various correlations, but this design is not appropriate for identifying causality: it is always difficult to find out what is the cause and what is the effect in a cross-sectional study. The authors reasonably believe that longitudinal data sets will be needed to exclude the possibility that low hospital nurse staffing is the consequence, rather than the cause, of poor patient and nurse outcomes. 4. The authors strongly believe that their findings can be generalized to the whole population of US nurses. This belief is reasonable granted the number of participants involved in the research: such number guarantees full representativeness of the study. The choice of sampling strategy, reliable collection and analysis of data, smart use of statistical analysis techniques fully fits the research objectives. Consequently, Aiken and colleagues managed to address probably the most common point of criticism levied at such studies, namely poor generalizability of findings to a larger population. 5. The implications of the study are essential. First of all, Aiken and colleagues have developed a distinct measure of staffing that might be the best up to date and superior to those traditionally used in administrative databases. This measure was used to address one of the most common challenges associated with such studies, namely data missing in the national surveys. Evidently, the authors do not address the whole scope of issues associated with the problem of nurse staffing in US hospitals, but the do highlight several highly important aspects of this problem, namely that low nurse staffing may be one of the key reasons for low nurse retention levels. Unfortunately, there is no absolute clarity regarding the issues (moreover, low hospital nurse staffing observed several years ago in the US might be the consequence, rather than the cause, of poor patient and nurse outcomes) further longitudinal research is highly recommended to either confirm or disprove the link. VII. Abstract: How will the abstract represented the study under invest. The abstract represents a finely structured piece that provides every key aspect of the study. Division into six section (context; objective; design, setting and participants; main outcome measures; results; and conclusions) enable the reader to understand the essence of the study carried out by Aiken and colleagues. However, the authors apparently exceed the word limit normal for such abstracts (200-250 words). References Maslach, C., Jackson, S. E., & Leiter, M. P. (1996). Maslach Burnout Inventory. (3rd ed.). Palo Alto, CA: Consulting Psychologists Press. Read More