Alarm fatigue may endanger the safety of patients by negatively affecting nurses' concentration and ability to provide effective care. Identifying alarm fatigue and taking appropriate measures are critical in preventing medical errors and for nurses to work with high motivation. This study aimed to test the psychometric properties of the Turkish version of the 23-item Alarm Fatigue Assessment Questionnaire (AFAQ).
MethodThe study was conducted between February 2022 and April 2022 and included nurses with at least one year of clinical or intensive care experience. The data were collected via a web-based questionnaire. During the adaptation of AFAQ, language, content, and construct validity were evaluated; reliability was examined by internal consistency analysis.
ResultsThe item and scale content validity index of AFAQ were found to be high (>0.80). The Kaiser–Meyer–Olkin measure of sampling adequacy indicated an adequate sampling (0.85); Bartlett's test of sphericity χ2 was 1935.074, p<0.001. Exploratory factor analysis (EFA) showed that the 21-item scale had a five-factor structure, explaining 51.606% of the total variance, and the factor loadings of the items were >0.30 (0.422−0.803). Confirmatory factor analysis (CFA) showed that the five-factor model had a good fit index (χ2/df=1.855, SRMR=0.039, RMSEA=0.048, CFI=0.915, and TLI=0.908) and appropriate factor loadings (>0.30). The internal consistency of AFAQ (Cronbach's alpha coefficient) was 0.85, and the corrected item-total correlations were between 0.32−0.55.
ConclusionThe results indicated that the Turkish version of the Alarm Fatigue Assessment Questionnaire was sufficiently valid and reliable to measure alarm fatigue in nurses.
La fatiga de las alarmas puede poner en peligro la seguridad de los pacientes al afectar negativamente la concentración y la capacidad de las enfermeras para brindar una atención eficaz. Identificar la fatiga de las alarmas y tomar las medidas adecuadas es fundamental para prevenir errores médicos y para que las enfermeras trabajen con una gran motivación. Este estudio tuvo como objetivo probar las propiedades psicométricas de la versión turca del Cuestionario de Evaluación de Fatiga de Alarma de 23 ítems.
MétodoEl estudio se realizó entre febrero de 2022 y abril de 2022 e incluyó enfermeras con al menos un año de experiencia clínica o de cuidados intensivos. Los datos fueron recogidos a través de un cuestionario basado en la web. Durante la adaptación del cuestionario se evaluó la validez de lenguaje, contenido y constructo; la fiabilidad se examinó mediante análisis de consistencia interna.
ResultadosEl índice de validez de contenido de los ítems y las escalas del AFAQ resultó ser alto (>0,80). La medida de adecuación muestral de Kaiser–Meyer–Olkin indicó un muestreo adecuado (0,85); La prueba de esfericidad χ2 de Bartlett fue 1935.074, p<0.001. El análisis factorial exploratorio mostró que la escala de 21 ítems tenía una estructura de cinco factores, explicando el 51,606% de la varianza total, y las cargas factoriales de los ítems eran >0,30 (0,422−0,803). El análisis factorial confirmatorio mostró que el modelo de cinco factores tenía un buen índice de ajuste (χ2/df=1,855, SRMR=0,039, RMSEA=0,048, CFI=0,915 y TLI=0,908) y cargas factoriales apropiadas (>0,30). La consistencia interna del cuestionario (coeficiente alfa de Cronbach) fue de 0,85 y las correlaciones ítem-total corregidas oscilaron entre 0,32 y 0,55.
ConclusionesLa versión turca del cuestionario era lo suficientemente válida y fiable para medir la fatiga de alarma en las enfermeras.
What is known/what it contributes
Invasive and noninvasive technologies used in intensive care units and clinics are designed with multiple alarm sounds to evaluate and monitor the patient. These alarm devices notify healthcare staff of changes in vital signs and any equipment failure. Especially nurses are exposed to alarm fatigue, as they spend most of their time with the patients 24h a day. In view of the importance of recognizing alarm fatigue in nurses and its consequences for nurses and patients alike, this study provide a valid ve reliable measurement to evaluate nurses’ alarm fatigue.
Implications of the study
The results of the study show that the validity and reliability of the Turkish version of Alarm Fatigue Assessment Questionnaire are satisfactory. This questionnaire is efficient enough for measuring nurses’ alarm fatigue and can be used in the development of programs for reducing alarm fatigue-related problems and issues.
;1;
IntroductionWith technological advancements, various medical devices and tools have emerged to facilitate patient monitoring and treatment as part of care management in healthcare services. These devices have many warning systems that report the status of the patient or the device.1 These devices, with both audible and visual alerts, provide clinicians and nurses with critical information about the functional status of the device as well as the physiological status of the patient.2,3
Medical errors remain a major concern for healthcare organizations despite advances in information technology and monitoring systems.4 In this context, one of the major risk factors that can jeopardize patient safety is alarm fatigue.5 Alarm fatigue is defined as the desensitization and indifference of healthcare workers to a large number of repetitive or simultaneous alarm sounds.6 The presence of multiple alarms linked to the use of a large number of devices in healthcare settings, combined with the potential for misinterpretation of false alarms, leads to alarm fatigue among healthcare professionals, especially nurses.1,2 Another important cause of alarm fatigue is the high sensitivity of patient monitoring systems to capture all possible clinical events, resulting in many false positives and non-actionable alarms.7 As a consequence, healthcare professionals may lose confidence in these systems, and their responsiveness to alarms may decrease. However, this unresponsiveness may lead to an acute change in the patient's condition going unnoticed.7,8
Indifference and insensitivity due to alarm fatigue can cause irreversible harm to patients, resulting in prolonged hospitalization, increased cost, and even death.4,5 It is critical to detect alarm fatigue to ensure patient safety and to reduce the stress and dissatisfaction caused by the burden of alarms in healthcare professionals, especially nurses who provide one-to-one care to patients.9,10 The first tool for assessing alarm fatigue was developed by Torabizadeh et al. (2017),11 and the scale became one of the most widely used tools in studies on the subject.12,13 The scale was also adapted to the Turkish population.14 Another tool was developed by Ashrafi et al. (2017), but the validity and reliability of this tool have not been demonstrated in the Turkish population.15 Therefore, this study aimed to evaluate the psychometric properties of the Turkish version of the Alarm Fatigue Assessment Questionnaire (AFAQ) and provide an alternative measurement tool for researchers in the field.
MethodStudy designThis study was conducted methodologically to evaluate the validity and reliability of the Turkish version of AFAQ.
Setting and sampleThe study was conducted between February and April 2022 through a web-based survey including the clinical and intensive care nurses working in various regions and cities of Turkey.
The criteria recommended in validity and reliability studies were taken into consideration in determining the sample size for the study, which required the study sample to be ten times the number of items in the tool.16 Since AFAQ has 23 items, it was planned to include at least 230 nurses in the study. The study included those who had been working as clinical or intensive care nurses for at least one year and who did not have any problems in reading and understanding Turkish.
Data collection tools and proceduresThe data collection tools consisted of a 10-item Nurse Identification Form, which consisted of questions about the socio-demographic (age, gender, educational status, etc.) and institutional characteristics (type of hospital, department, duration of practice, number of patients per nurse, weekly working hours, shifts worked, etc.) of the nurses, and the 23-item AFAQ.
The AFAQ, developed by Ashrafi et al. (2017), consists of 23 items with a five-point Likert-type. The items in the scale are graded as "Always (4 points)", "Usually (3 points)", "Sometimes (2 points)", "Rarely (1 point)" and "Never (0 points)". The score ranges from 0 to 92; low scores indicate increased alarm fatigue in nurses. In the study describing the original scale, the content validity index (CVI) was reported to be 0.70 and Cronbach's alpha (α) internal consistency to be 0.80.15
In this study, the data were collected through the "Google Forms" platform, which allowed volunteer nurses, who were reached through the nursing groups and communities in which the researchers are involved, to access the data collection tools with a direct link. The web-based survey platform consisted of three sections. In the first section, information about the method, importance, conditions of participation, and researchers conducting the study were provided to the participants, a statement was made that the information obtained in the study would be kept confidential, and the informed consent of the participants was obtained. In this section, the participant was asked to respond to the statement "I agree / I do not agree to participate in the research in question", and the system was adjusted so that only those who agreed to participate in the study were directed to the second section. After completing the second section, the Nurse Identification Form, the participants were directed to the third section, which includes the items of AFAQ. In the system settings, the "mandatory" option was enabled for each question so that participants answered all questions. The "Send another response link" tab was not enabled in the system settings to prevent a respondent from entering the survey multiple times. Data were obtained only from nurses who voluntarily agreed to participate in the study.
Adaptation steps of the scaleDuring the adaptation of the scale, language, content, and construct validity were utilized to demonstrate the validity, and internal consistency was utilized to demonstrate reliability.
ValidityLanguage and content validityAn expert panels were formed to assess the language and content validity of the scale. Members of this expert panels were especially required to have a good command of the English language, to be consisted with experience in questionnaire adaptation and who were specialized in at least one of the following fields: intensive care, nursing, medicine, public health, healthcare management, and work ergonomics.
A total of six members, two professors (M.D.) and four nursing faculty members (Ph.D.), who met the above-mentioned qualifications, translated the original AFAQ form from English to Turkish. Researchers selected the most appropriate expressions by comparing the translations provided by the members of the expert panel. Then, the Turkish version of AFAQ was translated back to English by an independent linguist specializing in medical and health sciences who was not familiar with the scale. It was observed that there was a high level of similarity between the original version and the back-translated version; thus, the semantic equivalence of the scale was confirmed within the scope of language validity.
For the content validity of the scale, a second separate expert panel was formed. Two rounds of online content validity assessments were conducted. An invitation letter was sent via email to six expert panel members with detailed explanations to participate in the round 1, which included two professors (M.D.) with expertise in the field, two nursing faculty members (Ph.D.), one intensive care, and one clinical nurse (with ≥10 years of experience). This panel was asked to examine the appropriateness of the items in the Turkish version of AFAQ according to the content validity index (CVI).17 Accordingly, the experts rated each item of questionnaire according to how clear, relevant, and important it was using a 4-point CVI scale where 1=not clear/ relevant/ important, 2=somewhat clear/ relevant/ important, 3=quite clear/ relevant/ important, and 4=highly clear/ relevant/ important. Panel members were also asked to provide additional comments or feedback on the parts they deemed necessary and were given 2-weeks to complete the round 1. For each item, Item-CVI (I-CVI) was computed as the number of experts giving a rating of 3 or 4, divided by the total number of experts.17 Since the content validity included more than five experts, CVI values of 0.80 and above were accepted for the items. Items that scored ≥0.80 for each of relevance, importance and clarity (without suggestions for revisions) were retained for the final round. Items that scored ≥0.80 for each of relevance, importance and clarity (with suggestions for revisions), or ≥0.80 for each of relevance and importance but <0.80 for clarity were revised by the research team based on expert feedback and included in the round 2.
The second round was a priori determined to be the final round, and commenced 1-week after the completion of round 1. Experts were emailed a second invitation and asked to rate the revised items relative to clarity, relevance, and importance using the 4-point CVI scale; and to suggest item revisions. Experts had 2-weeks to complete the round 2. A reminder email was sent to participants on days 5 and 12 of the round 2 period if they had not yet participated. Analysis of the round 2 survey results followed the same format as round 1. A scale-level CVI (S-CVI/Ave) score was also calculated by averaging the I-CVIs by summing them and dividing by the number of items for all items included in the final round. The minimum acceptable S-CVI is considered to be any value between 0.80 and 0.90.17
Construct validityConstruct validity evaluates the extent to which a measurement tool can accurately measure the abstract concept or behavior to be measured.18 Explanatory and confirmatory factor analyses were utilized to demonstrate the construct validity of the scale. Exploratory factor analysis (EFA) was used to determine whether the factor underlying the data was sufficient for a hypothesis and the number of factors. Confirmatory factor analysis (CFA) was used to investigate whether there was an agreement between the factors determined after EFA and the theoretical factors.19 The expected criteria for EFA and CFA were explained in the statistical methodology of the study.
ReliabilityInternal consistencyCronbach's alpha (α), the most widely used index of the reliability of a scale, was utilized to assess its internal consistency.20 Its values range from 0 to 1. A commonly accepted rule for describing internal consistency when using Cronbach's alpha is: α≥0.9=excellent, 0.9>α ≥ 0.8=good, 0.8>α ≥ 0.7=acceptable, 0.7>α ≥ 0.6=questionable, 0.6>α ≥ 0.5=poor, 0.5>α=unacceptable.21
Ethical considerationsThis study was approved by the Bursa Uludağ University Ethics Committee for Health Sciences Research and Publication (Decision no: 2022-02/08). Data were collected after the participants marked "I agree to participate in the research in question" in the informed consent section of the web-based form. All data were kept confidential and set to be accessible only by the responsible investigator. The questionnaire did not contain any sensitive personal information.
Statistical analysisThe data were analyzed with Statistical Package for Social Sciences (SPSS version 28, SPSS Inc, Chicago, USA). Before the data analysis phase, the normality of the data was checked by assessing the skewness and kurtosis values. Descriptive statistics were used to summarize the demographic and institutional characteristics of the participants. Summaries of quantitative variables were presented as mean, standard deviation, minimum and maximum; qualitative variables were shown as frequency and percentage. The level of statistical significance was set as p<0.05.
Before EFA, CFA, and other analyses, the data sets were checked for missing, erroneous or extreme data; no outliers that could cause problems were found. Kaiser–Meyer–Olkin (KMO) and Bartlett sphericity tests were used to determine whether the study sample was suitable for EFA. The KMO value was expected to be 0.60 and above; Bartlett’s sphericity test was expected to have p<0.05.22 The EFA rotation method was determined by examining the correlation coefficients between the factors. Since there was a low correlation between the factors, the varimax method, an orthogonal rotation method, was applied.23 Then, the number of factors in the tested structure was determined, and items included in different factors were determined; the factor loadings were expected to be 0.30 and above.23
CFA was applied with Amos 28.0 to test the accuracy of the structure obtained through EFA. The current structure was examined in line with the goodness-of-fit indices and the criterion values for these indices as recommended in the literature (Table 1).24 Cronbach's α was considered as the reference for reliability analysis.
Fit indices used for confirmatory factor analysis.
| Goodness-of-fit index | Criteria |
|---|---|
| χ2/df | 1≤ χ2/df ≤5 |
| SRMR | <0.08 |
| RMSEA | <0.10 |
| CFI | >0.90 |
| TLI | >0.90 |
χ2/df, chi-square divided by the degrees of freedom; SRMR, Standardized Root Mean Square Residual; RMSEA, Root Mean Square Error of Approximation; CFI, Comparative Fit Index; TLI, Tucker-Lewis Index.
A total of 379 nurses working in various cities and hospitals in Turkey participated in the study (Table 2). The mean age of the nurses was 28.30±5.57 years; the majority were female (88.7%) and had bachelor's degrees (67.8%). The mean work experience was 6.12±5.71 years. Most nurses reported working in university and state hospitals (30.3% and 30.1%, respectively), working in internal medicine areas (42.5%), and caring for 1–3 patients per nurse (52.2%).
Socio-demographic and institutional characteristics of the nurses.
| Characteristics | n (%) or M±SD |
|---|---|
| Age (years) | 28.30±5.57 |
| Work experience (years) | 6.12±5.71 |
| Gender | |
| Female | 336 (88.7) |
| Male | 43 (11.3) |
| Education | |
| Associate degree | 84 (22.2) |
| Undergraduate degree | 257 (67.8) |
| Postgraduate degree | 38 (10.0) |
| Type of hospital | |
| University | 115 (30.3) |
| Public | 114 (30.1) |
| City | 63 (16.6) |
| Private | 81 (21.4) |
| Other | 6 (1.6) |
| Hospital setting | |
| Clinic | 119 (31.4) |
| ICU | 242 (63.9) |
| Other | 18 (4.7) |
| Type of ward | |
| Internal medicine/medical clinic | 67 (17.7) |
| Surgery clinic | 33 (8.7) |
| Pediatric/neonatal clinic | 13 (3.4) |
| Obstetrics clinic | 6 (1.6) |
| Emergency | 18 (4.7) |
| Internal medicine/medical ICU | 94 (24.8) |
| Surgery ICU | 21 (5.5) |
| General ICU | 48 (12.7) |
| Cardiovascular/cardiology ICU | 30 (7.9) |
| Pediatric/neonatal ICU | 49 (12.9) |
| Nurse: patient ratio | |
| 1:1–1:3 | 198 (52.2) |
| 1:4–1:6 | 90 (23.7) |
| 1:7–1:9 | 33 (8.7) |
| ≥1:10 | 58 (15.3) |
| Working hours (per week)a | 50.82±10.04 |
| The most alarming medical devices in your ward | |
| Mechanical ventilator | 214 (56.5) |
| Infusion pump | 282 (74.4) |
| Patient monitor | 293 (77.3) |
| Dialysis | 59 (15.6) |
| Enteral pump | 93 (24.5) |
| ECMO | 4 (1.1) |
| Pulse oximeter | 141 (37.2) |
M, mean; SD, standard deviation; ICU, Intensive care unit; ECMO, extracorporeal membrane oxygenation.
Based on the results obtained from the panel of six experts for content validity in round 1, 19 items scored ≥0.80 for each of clarity, relevance, and importance; 4 items scored ≥0.80 for two of clarity, relevance, and importance but <0.80 for one of the criteria. Four items were revised and included in the round 2. Of the 4 items included in the round 2, all scored ≥0.80 for each of clarity, relevance, and importance. The S-CVI scores for the 23-item AFAQ were above 0.80 (Table 3).
Content validity index scores for each AFAQ item in rounds 1 and 2.
| Items | Round 1 | Decision | Round 2 | Decision | ||||
|---|---|---|---|---|---|---|---|---|
| Clarity I-CVI | Relevance I-CVI | Importance I-CVI | Clarity I-CVI | Relevance I-CVI | Importance I-CVI | |||
| 1 | 0.83 | 0.83 | 0.83 | A | – | – | – | – |
| 2 | 0.83 | 1.00 | 1.00 | A | – | – | – | – |
| 3 | 0.83 | 0.83 | 0.83 | A | – | – | – | – |
| 4 | 0.83 | 1.00 | 1.00 | A | – | – | – | – |
| 5 | 0.83 | 0.83 | 0.83 | A | – | – | – | – |
| 6 | 1.00 | 1.00 | 1.00 | A | – | – | – | – |
| 7 | 0.83 | 1.00 | 0.83 | A | – | – | – | – |
| 8 | 1.00 | 1.00 | 0.83 | A | – | – | – | – |
| 9 | 0.67 | 0.83 | 1.00 | R2 | 0.83 | 1.00 | 1.00 | A |
| 10 | 1.00 | 1.00 | 1.00 | A | – | – | – | – |
| 11 | 1.00 | 0.83 | 0.83 | A | – | – | – | – |
| 12 | 0.67 | 0.83 | 0.83 | R2 | 1.00 | 1.00 | 1.00 | A |
| 13 | 0.83 | 0.83 | 0.83 | A | – | – | – | – |
| 14 | 0.83 | 0.83 | 0.83 | A | – | – | – | – |
| 15 | 1.00 | 1.00 | 1.00 | A | – | – | – | – |
| 16 | 1.00 | 1.00 | 0.83 | A | – | – | – | – |
| 17 | 0.83 | 0.83 | 0.83 | A | – | – | – | – |
| 18 | 0.67 | 0.83 | 0.83 | R2 | 0.83 | 0.83 | 1.00 | A |
| 19 | 0.83 | 1.00 | 1.00 | A | – | – | – | – |
| 20 | 1.00 | 1.00 | 0.83 | A | – | – | – | – |
| 21 | 0.83 | 0.83 | 0.83 | A | – | – | – | – |
| 22 | 1.00 | 1.00 | 1.00 | A | – | – | – | – |
| 23 | 0.67 | 1.00 | 0.83 | R2 | 0.83 | 1.00 | 1.00 | A |
| S-CVI | 0.88 | 0.93 | 0.91 | |||||
A, Accepted and retained for the final questionnaire; R2, Revised and included in round 2; –, Not included in round; I-CVI, Item content validity index; S-CVI, Scale-level content validity index.
To determine the construct validity, the data set was randomly divided into two groups: 189 (for EFA) and 190 (for CFA) participants. The KMO coefficient of the data set (0.850), and Bartlett’s sphericity test (χ2=1935.074, p<0.001) yielded favorable results for construct validity (Table 4). According to the EFA in the first attempt, based on the Varimax rotation method, the 23-item scale was found to have a six-factor structure with an eigenvalue above 1. The item statistics were examined based on the criteria of loading on more than one factor, being ambiguous, and not loading on any factor (<0.30); the procedures were repeated until the most appropriate structure was achieved. As a result of this process, two items (4 and 22) in the scale did not meet these criteria, and they were excluded from the analysis. Thus, the remaining 21 items in the scale were grouped into five factors with eigenvalues above 1 and explained 51.606% of the total variance (Table 4). These five factors and the items grouped in these factors were examined and named to best represent the group as follows: Factor 1, barriers (items 23, 6, 5, 20, 11); Factor 2, indifference (items 17, 18, 19, 21, 16, 9); Factor 3, attention (items 10, 3, 1); Factor 4, specific devices (items 13, 14, 15); Factor 5, relevant reaction (items 2, 8, 7, 12). According to EFA, the factor loadings of the items ranged from 0.422 to 0.803 (Table 4).
Factor loading of the questionnaire after varimax rotation with five factors.
| Factors | Items | Factor loadings | h2 (communality) | Variance |
|---|---|---|---|---|
| 1: Barriers | 23. I become confused with successive sounds of alarms. | 0.803 | 0.670 | %14.579 |
| 6. I get nervous when I hear an alarm. | 0.734 | 0.584 | ||
| 5. Alarms hinder my focus on professional duties. | 0.728 | 0.580 | ||
| 20. Multiplicity and concurrence of alarms confuse me in making decisions. | 0.668 | 0.594 | ||
| 11. In the morning shift, the crowd hinder my immediate reaction to alarms. | 0.596 | 0.420 | ||
| 2: Indifference | 17. I am indifferent to the alarms. | 0.602 | 0.491 | %12.150 |
| 18. During a CPR in a patient, I become indifferent to the alarms of other patients. | 0.699 | 0.569 | ||
| 19. By repetition of alarms, I become indifferent to them. | 0.636 | 0.432 | ||
| 21. I do not pay attention to the alarm when I do not feel well. | 0.589 | 0.502 | ||
| 16. In the course of time, my sensitivity to alarms decreases. | 0.449 | 0.463 | ||
| 9. I stop as I hear the alarm, maybe it is settled by itself. | 0.424 | 0.448 | ||
| 3: Attention | 10. I pay more attention to the alarms in night shifts. | 0.677 | 0.479 | %9.395 |
| 3. I go to the patients bed immediately after I hear alarms. | 0.619 | 0.586 | ||
| 1. I pay attention to the changes of alarm source immediately after hearing the alarm. | 0.568 | 0.458 | ||
| 4: Specific Devices | 13. I have an immediate reaction to the ventilator alarms. | 0.625 | 0.446 | %8.190 |
| 14. I have an immediate reaction to the infusion pump alarms. | 0.517 | 0.501 | ||
| 15. I have an immediate reaction to cardiac monitoring alarms. | 0.738 | 0.581 | ||
| 5: Relevant Reaction | 2. I am sure that the alarms are true. | 0.439 | 0.409 | %7.291 |
| 7. I have a proper professional reaction toward alarms. | 0.560 | 0.477 | ||
| 8. I try to distinguish the informing alarms (yellow) and warning alarms (red). | 0.759 | 0.654 | ||
| 12. At the beginning of each shift, I pay more attention to the alarms. | 0.422 | 0.493 | ||
| Kaiser-Meyer Olkin=0.850; Barlett Sphericity χ2=1935.074 (p<0.001) | Total variance | %51.606 | ||
CFA was applied to confirm the factor structure obtained through EFA. The model fit indices of the scale were found to be acceptable: χ2/df=1.855, standardized root mean square residual (SRMR)=0.039, root mean square error of approximation (RMSEA)=0.048, CFI=0.915, and Tucker-Lewis index (TLI)=0.908. As a result of CFA, the factor loadings were found to be between 0.55 and 0.79 for the first subscale, between 0.36 and 0.65 for the second subscale, between 0.38 and 0.75 for the third subscale, between 0.47 and 0.53 for the fourth subscale, and between 0.35 and 0.51 for the fifth subscale (Fig. 1).
Reliability
In the internal consistency analysis, Cronbach's α was found to be 0.839 for the overall 21-item scale and between 0.602 and 0.795 for the subscales. The corrected item-total correlations of the scale ranged from 0.32 to 0.55 (i.e., >0.30 for all items) and the items were found to be homogeneous (Table 5).
Alarm fatigue assessment questionnaire item mean scores, item-total correlations, and Cronbach's α coefficient if item deleted.
| Subscales | Items | Mean (SD) | Item-total correlation | Cronbach's α if item deleted |
|---|---|---|---|---|
| Barriers | 1 | 2.59 (1.11) | 0.527 | 0.826 |
| 2 | 1.95 (1.18) | 0.519 | 0.827 | |
| 3 | 2.15 (1.21) | 0.460 | 0.830 | |
| 4 | 2.43 (1.04) | 0.572 | 0.824 | |
| 5 | 2.22 (1.15) | 0.455 | 0.830 | |
| Indifference | 6 | 3.68 (0.56) | 0.451 | 0.832 |
| 7 | 2.97 (1.06) | 0.327 | 0.837 | |
| 8 | 3.00 (0.98) | 0.558 | 0.825 | |
| 9 | 3.62 (0.65) | 0.516 | 0.829 | |
| 10 | 2.63 (1.11) | 0.551 | 0.825 | |
| 11 | 2.91 (0.94) | 0.423 | 0.831 | |
| Attention | 12 | 3.37 (0.77) | 0.492 | 0.840 |
| 13 | 3.49 (0.62) | 0.465 | 0.831 | |
| 14 | 3.67 (0.52) | 0.427 | 0.833 | |
| Specific Devices | 15 | 3.63 (0.69) | 0.483 | 0.837 |
| 16 | 3.40 (0.68) | 0.467 | 0.834 | |
| 17 | 3.73 (0.55) | 0.511 | 0.836 | |
| Relevant Reaction | 18 | 3.03 (0.87) | 0.439 | 0.835 |
| 19 | 3.12 (0.72) | 0.399 | 0.833 | |
| 20 | 3.45 (0.72) | 0.450 | 0.838 | |
| 21 | 3.08 (0.921) | 0.467 | 0.838 |
SD: standard deviation.
Defining and identifying alarm fatigue is critical because of its implications in many adverse impacts on both nurses and patients. The tools for the assessment of alarm fatigue in the literature are very limited.11,15 Thus, this study aimed to determine the psychometric properties of the Turkish version of the Alarm Fatigue Assessment Questionnaire (AFAQ), which has not been verified in Turkey. The results indicated that the Turkish version of the scale had appropriate and acceptable features within the scope of validity and reliability analyses.
Within the framework of the language and content validity of the scale, translation, back translation, and cultural adaptation procedures were performed by using expert panels consisting of physicians, nurses, and academics. The scale had acceptable CVI values (≥0.80) for all items. In the original study of the scale, content validity was evaluated with the Lawshe technique and it was reported to have an acceptable CVI (0.70), similar to our study.15
The EFA was performed in the first stage of determining the construct validity of the scale, and two items were removed from the scale. In the EFA performed after the removal of these items, 21 items were categorized under five factors (barriers, indifference, attention, specific devices, relevant reaction) and explained 51.606% of the total variance. This result underlines an important difference between the Turkish version and the original scale because no factor analysis was performed in the original study.15 Therefore, the lack of information about possible sub-dimensions in the original scale prevents the comparison of the results presented here with the results for the original scale in terms of factor loadings. Nevertheless, the EFA in our study demonstrated that each item had a factor loading higher than the acceptable value of 0.30.23
In the second stage, CFA was used to validate the factors obtained in the EFA. The CFA showed that the model had a good fit index (1≤χ2/df≤5, CFI and TLI>0.90, SRMR<0.08, and RMSEA<0.10) and a five-subscale structure with appropriate factor loadings (>0.30) consistent with the EFA results. The findings suggest that the Turkish version of the scale can reflect the level of alarm fatigue among nurses in five dimensions. However, in the current study of the Polish adaptation of the questionnaire, it was reported that the single dimension CFA model obtained by removing three items had a good fit index and the factor loadings for 20 items ranged between 0.203 and 0.769.25 We believe that the five sub-dimensions identified in our study will aid in distinguishing the elements contributing to alarm fatigue in nurses and in planning interventions by designing strategies based on these characteristics.
The reliability analyses indicated that Cronbach's α value for the overall scale was 0.839 (i.e., above the acceptable value of 0.70), which was consistent with the value reported in the original (0.80) and the Polish adaptation (0.88) studies.15,25 However, Cronbach's α values of sub-dimensions three, four, and five were below the acceptable level (0.602, 0.636, and 0.672, respectively). This may have been because these sub-dimensions consisted of four or fewer items. The fact that the corrected item-total correlation values were higher than 0.30 indicated that the relationships between the items and the theoretical construct being measured were substantial and that the items were effective and sufficient in measuring the intended behavior.
Nurses are the primary healthcare provider responsible for close patient monitoring and therefore are at the forefront for developing alarm fatigue. The primary intervention to reduce alarm fatigue is through alarm management.26 According to Turmell et al. (2017), implementing an alarm management strategy has the potential to save hundreds of thousands of healthcare dollars.7 Additionally, it can collectively save nurses hundreds of hours of time, so they may focus more on direct patient care. The Alarm Fatigue Questionnaire, which has been effectively adapted into Turkish, can be used in healthcare settings to improve patient safety, assist hospital managers in developing an alarm management strategy, and create relevant training programs, policies, and procedures. The use of the Alarm Fatigue Assessment Questionnaire in clinical risk management can bring tangible benefits to the employer and medical staff.
LimitationsDespite its contributions to the field, this study has some limitations. The fact that the data was collected through an online survey and voluntary participation may have led to selection bias and limited the performance of other analyses such as test-retest reliability. The fact that factor analyses were not performed for the original scale prevented the comparison of the scales in terms of construct validity. Another limitation of the study is that the concurrent validity, which allows the scale to be compared to a similar scale with previously known validity, was not investigated.
ConclusionThe findings of the study demonstrated that the validity and reliability of the Turkish version of the Alarm Fatigue Assessment Questionnaire were satisfactory. Thus, the scale represents a tool that may facilitate studies to identify and measure alarm fatigue in nurses in Turkey and help develop programs to reduce associated problems.
FundingThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Conflict of interestsThe authors declare there is no conflict of interest.
