Abstract
Nursing home (NH) nurses play a key role in hospital transfer decisions. In case of unavailability of physicians, they have to decide alone. In doing so, they consider potential threats to residents’ quality of life (QoL), but they also fear the consequences of wrong decisions. The present study examines the causal influence of these deliberations on nurses’ transfer decisions and emotional strain. In all, 241 NH nurses participated in the experiment. The vignette describing a resident with deteriorating health state elicited either thoughts on residents’ QoL, on legal consequences, or on QoL and legal consequences. In the QoL condition, the likelihood of a hospital transfer was lower compared with all other conditions. Emotional strain was stable across experimental conditions. When thoughts on QoL and legal consequences were induced at the same time, strain was positively correlated with the inclination to transfer the resident. Promoting QoL considerations can help to reduce avoidable transfers.
Despite the high physical and emotional burden for residents (Dwyer, Gabbe, Stoelwinder, & Lowthian, 2014), hospital transfer rates from nursing homes (NHs) are high. A growing number of studies focuses on the problem of avoidable hospital transfers (Renom-Guiteras, Uhrenfeldt, Meyer, & Mann, 2014) and end of life transfers from NHs threatening residents’ quality of life (QoL; for example, Goddard, Stewart, Thompson, & Hall, 2013). NH staff plays a key role in the transfer decision-making process. They often have to make transfer decisions alone without physician presence as general practitioners (GPs) are only fragmentary available (Kada & Janig, 2016; Laging, Ford, Bauer, & Nay, 2015).
According to two recently published literature reviews of qualitative studies (Laging et al., 2015; O’Neill, Parkinson, Dwyer, & Reid-Searl, 2015), nurses’ transfer decisions are complex and have to be made under uncertainty (e.g., lack of resources for diagnostics in the NH, no physician seeing the patient, lack of clinical knowledge and skills). Nurses take into account the potential hazards of hospitalizations for older people (Laging et al., 2015; O’Neill et al., 2015). For example, NH nurses reported that hospital transfers are burdensome for residents, especially those suffering from dementia, and that residents’ condition often deteriorates during hospitalization (Kada et al., 2011). Nurses see themselves as advocates for residents’ wishes (Laging et al., 2015), but they also fear the (legal) consequences of wrong transfer decisions (O’Neill et al., 2015); the inclination to send residents to hospital seems to be greater if nurses fear risking their license by working beyond their scope (McCloskey, 2011). These conflicting values in addition to the complexity of the situation may cause moral distress in nurses (Goethals, Gastmans, & Dierckx de Casterle, 2010; Pijl-Zieber et al., 2018). According to moral distress theory (Corley, 2002), nurses suffer psychologically when they (think they) are unable to advocate for their patients or residents; moral distress is reinforced by decisional options with unknown or unpredictable outcomes (which applies to transfer decisions).
The present study seeks to understand the causal influence of these deliberations on nurses’ transfer decisions and emotional strain. It was hypothesized that thoughts on residents’ QoL reduce the likelihood of a hospital transfer whereas considerations on legal consequences of the decision increase the likelihood of a hospital transfer. Regarding emotional strain, the analysis was exploratory in nature as no hypothesis could be deduced from the current state of research. Furthermore, effects of work experience and professional group were tested.
Method
Sample
A convenience sample of 241 nurses from 13 NHs in Austria (province of Carinthia) participated in the study; this corresponds to a response rate of 91.6% in relation to the nurses available at the time of the survey and 62.4% in relation to the total number of nurses working in the respective NHs (information provided by NHs). One third of the study subjects that provided information on their professional background were registered nurses (RNs) (34%) and 66% were nurse assistants (NAs). NAs, whose primary duty is to assist RNs and physicians, were included in the experiment as NHs are not required to have RNs onsite during night shifts (Kada & Janig, 2016). Thirty-one percent had geriatric work experience less than 5 years, 29% reported a work experience between 5 and 10 years, and 40% had work experience in geriatric care of more than 10 years (missing values: n = 8).
Procedure
The present vignette experiment was conducted onsite in NHs in the context of a larger survey on NH care and supervised by a trained student. Respondents answered the questionnaire voluntarily and anonymously. This study was approved by the ethics committee of Carinthia (A 37/16).
Design and Methods of Data Collection
The vignette experiment was used because it has proven to be a valid and reliable approach to study causal effects on clinical decision making (Evans et al., 2015). Each questionnaire comprised a case vignette, which was based on the moral distress theory (Corley, 2002) in accordance with recommendations for vignette construction (Evans et al., 2015). Hence, as there are no agreed criteria on what constitutes potentially avoidable transfers (Renom-Guiteras et al., 2014) and nurses often have to decide under uncertainty, the vignette intentionally described a situation where optimal choice is controversial. Vignette content was derived from clinical experience and research on hospital transfers from NHs (Kada & Janig, 2016; Laging et al., 2015; Lamb, Tappen, Diaz, Herndon, & Ouslander, 2011). The vignette (and its different versions) was pilot tested during a research methods course with 21 nursing management students: The students, most of them working in NHs, were introduced to the vignette and asked to provide feedback on how relevant and plausible the described situation was.
In the present study, respondents were instructed as follows: “Please read through the following description and try to put yourself in the situation. The subsequent questions refer to the described situation.”
Misses S. is 87 years old and has finally acclimatized to the NH. She suffers from multiple illnesses and her health is bad. Suddenly, her health state deteriorates and she complains of pain. Because of her progressive dementia, she is limited in her ability to give detailed information on her health state. It is not clear if she would benefit from remaining in the NH or being transferred to hospital. You tried to reach the GP, but you were not successful. Now, you have to decide alone. [Different supplements according to vignette type]
By between-subjects manipulation of the independent variable, four versions of the vignette were created, each triggering different considerations in the respondent. The QoL vignette was meant to trigger thoughts concerning the physical and psychological well-being of the resident by adding the following supplement to the vignette: “You remember that a transfer to the hospital can be a big emotional and physical burden for people like misses S.” The legal consequences vignette was meant to elicit deliberations on potential legal consequences for the decision maker: “You remember that wrong decisions often have legal consequences.” The interaction vignette combines thoughts on QoL and on legal consequences to test which trigger is more powerful. The fourth vignette is the neutral one (no supplement to the vignette).
Transfer decision (Dependent Variable 1) was measured using a single item that was answered on a 7-point Likert scale (1 = very unlikely to 7 = very likely): “How likely is it that you decide to send the resident to hospital?”
The emotional burden for the decision maker (Dependent Variable 2) was assessed using the Short Questionnaire for Current Strain (SQS; Müller & Basler, 1993)—a test to measure short-term changes in subjective strain in real-life situations as well as experimentally induced stress situations—which comprises six pairs of bipolar emotional states to be answered using a 6-point rating scale (e.g., tensed vs. calm). Hence, the SQS is adequate to be applied in a vignette study (see also Evans et al., 2015). The instruction was slightly changed from “Please mark with a cross, how you feel now” in the manual to “Please mark with a cross, how you feel in this situation.” In line with the manual, an example was provided on how to answer the items. The SQS has shown satisfying reliability and validity (Müller & Basler, 1993). After recoding the negatively poled items, the mean value across all six items was calculated to yield the final composite score with values ranging from 1 to 6, where higher scores indicate more subjective strain.
One out of the four versions of the questionnaire was randomly assigned to each study subject (vignettes: QoL n = 63, legal consequences n = 62, interaction n = 64, neutral n = 52).
The vignettes are equally distributed across professional groups, χ2(3, N = 233) = 2.432, p = .488, and work experience, χ2(6, N = 235) = 6.351, p = .385. All vignettes were perceived as equally realistic as measured with one item (“How often do you experience comparable situations in your everyday work?”) that was answered on a 7-point Likert scale (1 = very seldom to 7 = very often), Kruskal–Wallis H test: H = 1.966, p = .580.
Methods of Data Analysis
Data were not normally distributed (Kolmogorov–Smirnov test, transfer decision: p < .001, emotional strain: p = .005). Mann-Whitney U tests were used to test differences between two groups. The Kruskal–Wallis H test, the non-parametric alternative to the analysis of variance (ANOVA), was used to analyze differences between more than two groups. The Jonckheere–Terpstra test was used in case of ordinal independent variables (here: work experience). Pearson correlation and partial correlation analysis were used to describe relations between metric variables. It should be noted that there was a surprisingly large number of missing values for the SQS: 49 respondents (20.3%) only answered five out of six items (neutral vignette: 17.3%, QoL vignette: 27%, legal consequences vignette: 19.4%, interaction vignette: 17.2%).
Results
The mean likelihood for a hospital transfer was M = 4.08 (SD = 1.86, range = 1.0-7.0, n = 241). Respondents reported a mean current strain of M = 3.53 (SD = 0.70, range = 1.33-5.0, n = 192). Nurses rated the mean occurrence of comparable situations in their everyday work as M = 3.89 (SD = 1.68, range = 1.0-7.0, n = 237).
Regarding the effect of different deliberations on decision making, a marginally significant difference was observed, H = 6.862, p = .076 (no post hoc tests performed because of non-significant result of overall test). As shown in Figure 1, eliciting considerations on residents’ QoL reduced the likelihood of a hospital transfer; a hospital transfer was equally likely in all other experimental conditions.
Effect of vignette type on transfer decision (n = 241).
There was no effect of professional group (RNs vs. NAs) on transfer decision making, z = −0.338, p = .735. A marginally significant effect of work experience could be detected, H = 4.889, p = .087 (Jonckheere–Terpstra test, n.s.). Nursing staff with a work experience of 5 to 10 years were less likely to decide to send the resident to hospital as compared with nursing staff with a work experience of less than 5 or more than 10 years (descriptive result; no post hoc tests conducted because of non-significant result of overall test).
Emotional strain did not differ between experimental conditions as shown in Figure 2, H = 2.508, p = .474. Emotional strain was lower in RNs versus NAs, z = −3.268, p = .001. Emotional strain was significantly higher in nursing staff with low work experience (less than 5 years) as compared with nurses working in geriatric care for 5 to 10 or more than 10 years, z = −2.503, p = .012.
Effect of vignette type on current emotional strain (n = 192).
The overall correlation between transfer decision and emotional strain was low (r = .140, p = .052); controlling for the experience with comparable situations in everyday work, the correlation was somewhat higher (r = .162, p = .026). As shown in Figure 3, correlations differed between experimental conditions. A moderate positive correlation between transfer decision and emotional strain was observed only for the interaction vignette (see Figure 3d, r = .388, p = 007, 13.4% variance explained).
Correlations between transfer decision and emotional strain.
Discussion
Qualitative studies have shown that nurses consider potential benefits of hospital transfers such as quicker diagnosis or more effective pain treatment. On the other hand, they take into account potential negative consequences of hospital transfers such as physical and psychological distress, unfamiliar personnel and environment, and the risk of deteriorating health; furthermore, nurses’ own legal safety is another important aspect influencing transfer decisions (Laging et al., 2015; O’Neill et al., 2015). The present vignette experiment shows that eliciting considerations on residents’ QoL may have an effect on nurses’ transfer decision making favoring the decision to avoid sending the resident to the hospital albeit only marginally significant. Furthermore, results indicate that considerations on legal consequences may neutralize the effect of QoL induction (interaction vignette). If interventions aim at the reduction of avoidable hospital transfers of NH residents trainings focusing on QoL could be useful. Clarifying the role of nursing staff might help to reduce the fear of legal consequences (e.g., Laging et al., 2015).
Emotional strain was stable across vignettes; RNs and nursing staff with higher work experience suffered from less strain compared with NAs and nursing staff with lower work experience, respectively. One result warrants further attention: While transfer decision and emotional strain were generally not correlated, a moderate positive correlation could be observed for the interaction vignette. If respondents were made weighing considerations on QoL and legal consequences at the same time emotional strain increased with the likelihood of sending the resident to hospital. This might be interpreted in terms of moral distress (Goethals et al., 2010) if deciding in favor of own legal safety at the expense of residents’ QoL. Given the fact that nurses are often forced to make transfer decisions without physician presence (Kada & Janig, 2016), interventions to reduce moral distress caused by decision dilemmas should be developed. Clear transfer guidelines, such as the ones developed and tested in the INTERACT project (Ouslander et al., 2011), can help to reduce ambiguity and moral distress (Goethals et al., 2010). Interventions on the individual level such as improving nurses’ geriatric knowledge and skills alone may not be sufficient to improve transfer decisions given the complexity of the problem (Lamb et al., 2011). In addition, interventions on the organizational level (Corley, 2002) reducing constraints like staff shortage—a major source of moral distress in dementia care (Pijl-Zieber et al., 2018)—seem to be necessary. Furthermore, NHs have to care for residents with complex care needs with increasing frequency. Current standards for NH care in Carinthia such as fragmentary presence of physicians, restricted availability of RNs during night shifts, and the resulting limited interdisciplinary collaboration are no longer adequate (Kada & Janig, 2016). It seems necessary to change these standards to improve transfer decisions and reduce moral distress in decision makers.
There are several limitations of the present study. The small sample size and resulting low power is one of the main restrictions. The large number of missing values regarding the SQS is a further restriction to be kept in mind. Even though the randomization check showed that professional group and work experience were equally distributed across vignettes and all versions of the vignette were perceived as equally realistic, other systematic differences between groups might exist. The vignettes were developed based on current research results to establish internal validity (Evans et al., 2015). On the other hand, the present vignette represents a simplified situation—usually nurses know their residents very well and might use this background information to inform their decision making—which might reduce external validity. Future factorial surveys should also test the effects of the characteristics of the residents (e.g., level of dependency), his/her wishes, and the severity of symptoms on transfer decision making and moral distress. Nevertheless, the use of vignette studies in medical and nursing research has proven useful (Bachmann et al., 2008) and yields similar results as compared with the “gold standard” of using standardized patients (Evans et al., 2015).
Nonetheless, the present study is the first one to analyze causal effects of nurses’ transfer decision making and related stress, and hence adds to the body of research. Replications and extensions of the present study are needed. Mixed methods designs (e.g., Kada & Janig, 2016) could help to gain further insight into this complex phenomenon and help to develop interventions to reduce potentially avoidable hospital transfers in NHs residents and reduce strain in nurses.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
