Analysis of the Comprehensive Discharge Planning
and Home Follow-up of Hospitalized Elders Research Study

By
 Lora Utech

The purpose of this paper is to critique the research study by Naylor et al. (1999) titled “Comprehensive Discharge Planning and Home Follow-up of Hospitalized Elders: A Randomized Clinical Trial.” This study was performed to determine the effectiveness of an Advanced Practice Nurse-centered discharge planning and home follow-up regimen for elders who were at risk for hospital readmission. 

Problem, Purpose and Hypotheses

To begin, many factors leading to the research problem for this study are mentioned clearly by the researchers in the first two paragraphs of the article. The problems associated with elderly patients being discharged from the hospital sooner despite the presence of many complex health issues, along with the decrease in access to services for the elderly help to set the stage for the pertinence of this study. This problem is significant for nursing from the aspect of hospital nursing and from an advanced practice standpoint. For example, with this type of program, a staff nurse can feel more confident that elderly patients are receiving adequate care after hospital discharge, and the Advance Practice Nurse (APN) has an opportunity to have an incredibly positive impact on the elderly population and the community as a whole.

This study implemented both a statement of study objective (purpose) and a statement of hypothesis. Both were stated clearly and concisely and were placed in a logical and easy-to-find location. The stated research hypothesis is not made in relation to a theory, but is made as a result of the researchers previously conducted research. This directional hypothesis does appropriately predict a relationship between variables.

Literature Review and Framework

The researchers of this study implemented a limited literature review and a brief description of a handful of related research studies. Slightly confusing, the researchers chose not to label the literature review section separately from other sections of the report. However, the information that is present does help to set the stage for why these researchers felt that this type of study was important. This research does seem to include recent literature and does not appear to provide an unnecessary over-dependence on secondary sources. Unfortunately, the literature review does fail to critically appraise and compare the key study contributions and does not mention weaknesses or gaps within the current research. 

The researchers of this study do not mention the use of a theoretical or conceptual framework on which to guide the study. The only mention of a framework is stated as the researchers describe what guided the APNs’ decision making process to achieve individualized care. According to this evaluator, the lack of a theoretical or conceptual framework does not particularly detract from the importance of the research findings.

Design and Sampling

This study is an example of a quasi-experimental design because control and manipulation are present, but true randomization is not. This is because random selection from sample population is not present however, random assignment to experimental and control groups is present. Specifically, this is an example of the post-test only control group design which is appropriate for this study. This study did involve an experimental intervention, which the researchers describe in detail. In order to describe the intervention applied to the experimental group, comparisons were made between this group and the control group. This was appropriate and done in order to determine a relationship between the independent variables (APN-centered discharge and follow-up) and dependent variables (outcomes). Also, this study is considered longitudinal because data was collected over time, which is appropriate for this study.

The researchers for this study did take measures in an attempt to control external and extraneous intrinsic factors. An example of attempting external control was the fact that with the experimental groups, any available family member also was educated by the APN along with the patient. This would help to gain and maintain the support of the family with the plan set out by the APNs.  Also, the five APNs involved in the study all had similar education backgrounds and years of experience.  Attempts to control intrinsic factors included limiting the types of admission diagnoses included in the study, subjects had to be English speaking and be alert and oriented upon hospital admission. All measures were appropriate.

Internal and external validity are also important factors to take into account when critiquing a study. A threat to the internal validity of this study is attrition. Not including deaths, 24 subjects in the control group and 32 subjects in the experimental group dropped out of the study after enrollment. However, it is helpful that almost an equal number dropped from both groups and the researchers do list the reasons for subject loss. Because this study was conducted over a period of almost four years, internal validity is threatened

The fact that this study uses a non-probability convenience sample is a risk to external validity.  However, to increase the level of external validity, this study consists of an equal male/female number of subjects and the two study groups are similar in all sociodemographic and baseline health characteristics. Also, two hospitals were included in this study, which also helps to increase the level of generalizability. It is favorable that this study included 45% black subjects. However, because the study only included black and white subjects, this study can only be generalized to these two ethnic groups. The researchers show support for their level of attrition with a brief mention that this 28% attrition rate is similar to other randomized trials with similar populations. Lastly, it is beneficial that the diagnoses of the subjects met one out of ten of the most common diagnoses among Medicare patients. This is helpful in order to decrease the level of heterogeneity among the subjects and increase the level of validity. Despite the limits of this type of research, it is felt that the researchers took appropriate measures to increase the level of external validity.

As mentioned earlier, this study uses a non-probability convenience sample. This sample group is considered such because the subjects were taken from a population of  available hospitalized patients thus, are not randomly selected from the target population. This is a weakness, but is an aspect that is difficult to control with this type of study

The eligibility criteria for the sample subjects are identified clearly. Researchers attempted to avoid group assignment bias by eliciting the help of research assistants who were blinded to the study groups and hypotheses. Then, a project manager assigned the sample population to study groups with the use of a computer-generated algorithm. This helps to prevent group bias as well. The researchers avoid gender bias because 50% of the sample population was men. However, there is the potential for racial bias because the study mentions only the fact that 45% of the sample population was black and the remaining sample was white. Overall, the results of this study can only be generalized to similar elderly patients with one of the ten predetermined admission diagnoses and who is from a Caucasian or African-American ethnic background.

Data Collection

The majority of the data for this study were collected by assigned research assistants (RAs). The RAs were responsible for collecting the baseline data at the beginning of the study as well as the outcome data at the end of the study. The researchers do not mention however, how the RAs were trained. Were the RAs students of the medical profession? Did the RAs receive adequate training in regards to the tools used in this study? If not, the results of this study could be significantly affected. On a positive note, the researchers did set up the study so that the RAs were blinded to the study groups and hypotheses. This does help to decrease bias during the data collection process.

Data was collected by subject interview and subject medical record review. It is known that dependence of data from medical records can result in inconsistencies related to collector error, record inaccuracies, or the lack of available information. This can lead to inaccurate data. Also, the researchers do not mention under what conditions the data was collected. An inappropriate or distracting environment also can result in inaccurate data.

It does not appear that the study participants were exposed to increased risk or harm by participating in this study. The researchers state that a portion of the data was collected via telephone. With the mean age of the sample being 75 years, an increased amount of stress may possibly be placed on the subjects due to hearing problems or other factors related to age. This can result in inaccurate data.         

Self-Reports

Interviews made up a large portion of the data collection process. Given the fact that the researchers were assessing subjects for satisfaction, functional status and depression, interviews seem appropriate measures for data collection. Minimal information is offered regarding what the RA standardized telephone interviews entailed, therefore it is difficult to assess the reliability of the interview. However, a general improvement of this process would be to conduct some or all of these interviews in person. When assessing information about subjects specific to depression and functional status, person-to-person interaction seems more appropriate and would provide more reliable data, though it is realized that this application would likely be more time consuming.

Scales also were implemented as a part of data collection. Scales for interpreting the level of function and depression of the subjects were incorporated into subject assessment. However, the researchers do not mention the rationale for the use of these particular scales.  

Measurement

When evaluating the data quality, it is important to note that there is strong congruence between the research variables discussed in the introduction and the variables measured. Multiple instruments were implemented in this study in order to measure the study outcomes. These instruments are included for the measurement of functional status, depression status and patient satisfaction. The functional status and depression instruments are mentioned by specific name, however, the researchers fail to specifically mention the reliability and validity of these tools. Also, the instrument implemented to measure patient satisfaction was investigator-developed. No mention is made regarding this tool’s level of reliability or validity. Because this tool is developed by the researchers of this study, and not by a third party, questions must be raised regarding its application in this study. Not yet mentioned, standardized instruments were used by the RAs during collection of baseline data and during collection of information during telephone interviews. The quality of the data improves with this standardization, but again the level of reliability and validity of these instruments is not mentioned by the researchers. Without the reliability and validity information regarding all of these instruments, one must question data quality.

Analysis

The defined outcomes of this study are examined by quantitative analyses. The researchers include descriptive statistics along with inferential statistics when analyzing the data. In both instances, these applications seem appropriate.  The use of simple percentages also is included in the data analysis and does assist with data interpretation and evaluation. In general, the researchers do not offer a rationale for the use of the above statistics. A measure that is helpful with the interpretation process is the use of multiple tables for the summation of data. This is also helpful for data interpretation and evaluation.

Ethical Aspects

Ethical considerations are a concern with any research study. The researchers of this study mention early in the report that this study was approved by the institutional review boards of both participating hospitals in an attempt to ensure ethical study application. There is no written evidence upon review of the study that participants were subjected to intentional physical or psychological harm however, the researchers do not specifically mention any such ethical concerns. No information is offered in regards to what the subjects were told about the study while informed consent was trying to be obtained, so it is difficult to determine if ethical conflicts exist. It is important however, that subjects were given the opportunity to decline or cease participation in this study. Lastly, the researchers do not mention what safeguards were taken to protect the privacy of the participants and if ethical concerns existed within this arena.

Interpretation, Implications, and Recommendations

When interpreting the study findings, the researchers do review all aspects of the stated dependent variables. Also, the researchers’ interpretations are consistent with the results and they do interpret the study findings in terms of the hypothesis stated at the beginning of the report. The researchers interpret and support this study’s findings by comparing to similar studies performed with a focus on discharge planning and home follow-up intervention and studies performed to reduce rehospitalizations. This is done to support their own study while ensuring the reader that this study’s findings are not out of line with other study findings. Lastly, the researchers appropriately state their conclusions to this study because they state their study findings in a positive manner but not conclusive manner.

The implications of this study are that improved patient outcomes for elders at risk can be achieved at a reduced cost. Again, the researchers mention how this APN- directed discharge planning and home care intervention has the potential to promote positive incomes. However, the researchers refrain from strong and conclusive statements. This issue does have future implications for the nursing practice and for nursing research however, these implications are not explicitly mentioned by the researchers. The researchers also imply the improved discharge planning and home care that can be provided by an APN compared to the standard visiting nurse. Lastly, the researchers admit that the intervention process was mostly beneficial from an economic standpoint. Overall, this reviewer feels that the generalizations made the researchers are founded and warranted.

One recommendation mentioned by the researchers was the use of a more sensitive instrument for the measurement of patient satisfaction. The researchers also mention an area of home care that is in need of further study. There were no other recommendations made by the researchers in this report. With a study of this complexity, it is difficult to imagine that the researchers did not have any other recommendations specific to their study or for future studies. However, the recommendations made are supported by the study results.

Presentation

Overall, the research report was well written and concise. The title of the report does adequately capture the key concepts of the study. The abstract is helpful as it summarizes the fine points of the study and indicates the study results. One area of confusion is when the study sample and study patients are discussed. It seems that these two areas should be combined so that all of the information regarding the subjects is located in one area. Lastly, the tables provided in this report were helpful and allowed for a quick reference to most of the statistical study findings.

Suggestions for Merit Improvement

It is felt that there are measures that could improve the merit of the study. First of all, it appears that the majority of the data is collected by RAs. Because no information is offered regarding the RAs training or educational background, it may have been beneficial to include this information or have the APNs to collect a majority of this data. This may have increased the accuracy of the data. Also, it would be advantageous to collect as much data as possible in person instead of over the phone because all of the subjects were elderly and considered high-risk.