CIN: Computers, Informatics, Nursing
& Vol. 33, No. 5, 216–224 & Copyright B 2015 Wolters Kluwer Health, Inc. All rights reserved.
F E A T U R E A R T I C L E
Development of a Brief Instrument to Measure Smartphone Addiction Among Nursing Students SUMI CHO, MSN, RN EUNJOO LEE, PhD, RN
Smartphones are becoming a necessity in our daily lives, and the market share of smartphones is growing rapidly. Although the smartphone was first introduced in South Korea in 2009, its rate of use was 67.6% by 2012, compared with 27% in 2011, more than a twofold increase in 1 year. South Korea has had the most rapid growth rate and highest level of smartphone use, worldwide.1 Smartphones that are built on a mobile computing platform have Web browsers, which provide entertainment, education, banking, games, and GPS navigation services through various applications. In addition, smartphones are used for social networking through Twitter and Facebook and are becoming central to online communication.2 These advantages have brought enormous conveniences to society; however, adverse results can be caused by the overuse of smartphones.2,3 Their use can be more addictive than the use of the computer or Internet because smartphones can be used anywhere and at any time. There is evidence that the use of cellular phones, including smartphones, is a significant source of distraction for driving, classroom learning, and work-related tasks, including those in healthcare settings.4,5 Users constantly check text messages, social media, and e-mail, and they surf the Web. In particular, the use of smartphones by healthcare professionals may lower the performance of tasks requiring mental concentration, resulting in medical errors, and posing serious threats to patient safety.6 Nevertheless, a number of studies have focused on the acceptability and effectiveness of smartphones in healthcare settings.7–10 216
Interruptions and distractions due to smartphone use in healthcare settings pose potential risks to patient safety. Therefore, it is important to assess smartphone use at work, to encourage nursing students to review their relevant behaviors, and to recognize these potential risks. This study’s aim was to develop a scale to measure smartphone addiction and test its validity and reliability. We investigated nursing students’ experiences of distractions caused by smartphones in the clinical setting and their opinions about smartphone use policies. Smartphone addiction and the need for a scale to measure it were identified through a literature review and indepth interviews with nursing students. This scale showed reliability and validity with exploratory and confirmatory factor analysis. In testing the discriminant and convergent validity of the selected (18) items with four factors, the smartphone addiction model explained approximately 91% (goodnessof-fit index = 0.909) of the variance in the data. Pearson correlation coefficients among addiction level, distractions in the clinical setting, and attitude toward policies on smartphone use were calculated. Addiction level and attitude toward policies of smartphone use were negatively correlated. This study suggests that healthcare organizations in Korea should create practical guidelines and policies for the appropriate use of smartphones in clinical practice. KEYWORDS Confirmatory factor analysis & Patient safety & Smartphone addiction & Tool development
Author Affiliations: Doctoral Student (Ms Cho) and Professor (Dr Lee), College of Nursing, Research Institute of Nursing Science, Kyungpook National University, Daegu, Korea. This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (NRF- 2013R1A1A2057977). The authors have disclosed that they have no significant relationship with, or financial interest in, any commercial companies pertaining to this article. Corresponding author: Eunjoo Lee, PhD, RN, College of Nursing, Research Institute of Nursing Science, Kyungpook National University 680, Gukchaebosang-ro, Jung-gu, Daegu, 700-422, Korea ([email protected]). DOI: 10.1097/CIN.0000000000000132
CIN: Computers, Informatics, Nursing & May 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
For example, smartphones can be used to engage students in nursing education7 and to improve patients’ selfmanagement of chronic illnesses.8 In order to prevent interruptions and distractions due to smartphones, research on smartphone use by workers in healthcare settings is essential. An assessment of smartphone use by healthcare workers through reviews of their own behaviors and their realization of its potential risks are important.11–13 Nursing students must be aware of the dangers to patient safety. Therefore, the early detection of smartphone addiction is critical before nursing students and nurses put their patients at risk during clinical practice. Internet addiction criteria have been considered when developing smartphone addiction criteria, as it has been hypothesized that smartphone addiction shares many aspects similar to Internet addiction.14 However, smartphones do have some different characteristics from the Internet. First, they are highly valued by individuals because of portability. In addition, the connectedness of smartphones can help users interact with other people. Second, smartphones are more personalized, as various apps are designed by smartphone users. Therefore, causes or symptoms of smartphone addiction will be different from those of the Internet. There is an urgent need to develop criteria associated with the characteristics of smartphone addiction.15 Hence, the purposes of this study were to investigate smartphone use among nursing students and to develop a smartphone addiction scale that reflects the characteristics pertinent to smartphone use. Moreover, this study analyzed the attitudes of nursing students regarding the use of smartphones in practice. The awareness of attitudes can provide a basis for the prevention of significant risks to patient safety, which can be caused by the overuse of smartphones by workers in healthcare settings.
Background Definition of Addiction and Common Features Addiction is defined as a ‘‘compulsive, uncontrollable dependence on a substance, habit, or practice to such a degree that cessation causes a severe emotional, mental, or physiological reaction.’’16(pp321) Some researchers have argued that the concept of addiction should be applied only to chemical substances, such as drugs or alcohol.17–19 However, others have suggested that problematic behaviors, such as pathological gambling,20 sex,21 computer game playing,22 and Internet addiction,23–26 also belong to the subset of addictive behavior. Peele27 reported that any compulsive or overused activity should be considered an addiction. Akers28 reported that the traditional concept of addiction included tolerance, dependence, and withdrawal and applied them to the physiological demand for a drug. The term psychological dependence currently is used to describe habitual behavior
in the absence of proof of physical addiction. The criteria for dependence are outlined in the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV).29 They include spending a great deal of time and money using the substance; using it more often and longer than one intends; trying to stop using or reducing its use by making repeated unsuccessful efforts to do so; giving up important social, family, occupational, or recreational activities to use it; reporting withdrawal symptoms when one stops using it; and recognizing physical and psychological problems, but continuing its use despite the consequences. Peele27 reported that the major motives for addictive behavior are relief of pain, irritability, anxiety, and other emotional states, such as enhanced control, power, and increased self-esteem (compensation). Addictive behavior can help a person feel more comfortable in his/her surroundings and can serve as a mood stabilizer or way of feeling good as a coping mechanism. The concept of Internet addiction was first introduced by Goldberg,23 based on the DSM-IV criteria.29 Afterward, Young24 continued research on Internet addiction using the DSM-IV criteria29 for psychoactive substance dependence. In her study, many of the participants had symptoms of online dependency, such as tolerance, loss of control, withdrawal, and impairment of academic, social, and occupational functioning. Brenner30 conducted research on self-selected respondents with a mean age of 34 years. The respondents experienced, on average, 10 signs of interference with their daily functioning, such as failure to manage time effectively, cutting down on sleep, missing meals, and other signs, based on the Internet Related Addictive Behavior Inventory. Eighty percent of the respondents reported that they experienced at least five of the signs. Internet addiction research has been conducted in Korea since 1999.31,32 Most of the research on Internet addiction was conducted using the tool developed by Goldberg and others.23–26 Many studies on Internet addiction were published in South Korea after a death from cardiopulmonary arrest related to Internet addiction and an Internet game–related murder.33,34 South Korea considers Internet addiction as one of its most serious health problems.35 The number of persons identified with a high risk of Internet addiction comprises 7.0% of the population who were using the Internet in 2013 and were between 5 and 54 years old, and the number of persons at risk of Internet addiction in Korea was 2 286 000.36 The most popular Internet addiction scale used in Korea is the K-Scale, which was developed by the National Information Society Agency.37 Ko et al38 assessed Internet addiction based on five dimensions: compulsive use, withdrawal, tolerance, interpersonal and health problems, and time management problems. Although many studies have been published on the effects of Internet
CIN: Computers, Informatics, Nursing & May 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
217
addiction, it was not recognized in the DSM-IV or in the International Classification of Diseases, 10th Edition; however, it has been included in DSM-V.39 Recently, most of the research investigations on addiction have shifted from the Internet to the mobile phone, and studies on mobile phone addiction have mostly used measures of Internet or television addiction.40 The first mobile phone addiction study in Korea was conducted as a master’s thesis.41 The author developed a mobile phone addiction tool based on the Internet addiction tool of Young24 and Goldberg23 and proposed four main concepts of tolerance, withdrawal, difficulty maintaining control over one’s use, and dependency. Koo40 developed four concepts through in-depth interviews with adolescents: tolerance, withdrawal, difficulty maintaining daily life, and compulsive-impulsive control. Smartphone addiction can be detected by criteria that are met when an addicted person is separated from the smartphone, for example, a repeated failure to cut back on usage time and feeling happier using it than being with family or friends. However, there have not been many studies on smartphone addiction. Thus, it is necessary to conduct further research to investigate the social and behavioral problems resulting from its abuse or overuse.
METHODS Conceptual and Theoretical Framework The concept development process consisted of three stages. In the first stage, an open-ended survey was sent to nursing students enrolled in a BSN program. In the second stage, a literature and reference search of Web sites and diverse databases were performed. In the third stage, interviews were conducted with six nursing students who had more than 2 years’ experience using a smartphone and felt, based on input from peer nursing students, that they overused the smartphone. After conducting the literature review and interviews, we identified three components common to both Internet and smartphone addiction: withdrawal, tolerance, and interference with daily routines. Smartphones are more individualized and multipurpose in comparison to the Internet.2 Our interviewees noted that they are possessive of their smartphones, feel as though they ‘‘get something special’’ by using them, and feel a connection with other people through them. Therefore, we added a component of positive expectations associated with smartphone use to the scale, to indicate how essential smartphones are in participants’ lives. To ensure face validity, all of the items were evaluated by the nursing students for their accuracy, appropriateness, and relevance to smartphone use. After a review of the first draft and feedback from the nursing students, 218
several items were rephrased. Finally, two experts on tool development were consulted in order to improve the consistency of the items and the structure of the tool.
Measures on Distraction and Policy Development on Smartphone Use In this study, nursing students were asked whether they experienced distractions from smartphone use while participating in their clinical practicum. In addition, they were asked if they agreed with establishing policies on smartphone usage in the workplace to prevent its overuse. The lower score indicates that the students did not agree with policies that restricted smartphone use in the healthcare settings. We measured the relationship between addiction level, distractions during the clinical practicum, and their opinions regarding smartphone use policies in the healthcare setting. Distraction was measured by four items, and students were asked to rate how often they were distracted on a scale ranging from never = 1, rarely = 2, sometimes = 3, usually = 4, to always = 5. Examples of the items were ‘‘I was disturbed by text messages during clinical practice,’’ ‘‘I was disturbed by calls during clinical practice,’’ ‘‘I was disturbed by messages from the social network sites during clinical practice,’’ and ‘‘I was disturbed by the other students who were using their smartphones during clinical practice.’’ The concept of policy development was measured by three items: ‘‘I think healthcare professionals should be prohibited from using smartphones during their shift,’’ ‘‘I think every healthcare organization should implement a policy or regulation that prohibits smartphone use by healthcare professionals during shift,’’ ‘‘I insist that every healthcare professional should turn off his/her smartphone during his/her shift.’’ These items were measured a 5-point Likert scale, and the responses ranged from strongly disagree (1) to strongly agree (5). A total of 428 students who had a clinical practicum during their coursework answered the items on distractions and the development of a policy to ban smartphone use in hospitals.
Data Collection Before data collection, approval was obtained from the institutional review board of Kyungpook National University Hospital (institutional review board 2012-07-016). First, the researcher made a telephone call to the directors of each nursing school to describe the study’s purpose and procedures and to obtain permission to collect data. After permission was obtained for data collection, the researcher visited each school and met with a student representative
CIN: Computers, Informatics, Nursing & May 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
from each class who was designated by the director of the school. The researcher handed over the questionnaires to the student representatives and explained the study’s purpose and procedures. The designated students distributed the questionnaires and return envelopes to the students so that there was no perceived pressure from the schools’ directors and the researcher. The designated students explained the purpose of the study and the procedures directly to the students. The questionnaire also had a cover letter, which explained the study’s purpose and how to complete the questionnaire. The cover letter explained that participation was entirely voluntary and anonymous. Students were informed that they could refuse to participate in the research at any time and that their completion or noncompletion of the questionnaire would not affect any of their course evaluations or grades in other courses. A total of 600 questionnaires, ranging from 100 to 200, were distributed to each nursing school after a discussion with the schools’ directors. Data were collected for 1 week to 10 days at each school according to the school’s class schedule. Completed questionnaires were enclosed in the return envelope and handed to the designated students. The researcher received the returned questionnaires by visiting the designated students or by commercial delivery at the researcher’s address. The return rate of the questionnaire ranged from 72% to 92%. However, five questionnaires were incomplete; therefore, 528 questionnaires were analyzed for the study (Table 1).
RESULTS Reliability Cronbach’s ! for the total scale was .902. The concept reliabilities for the four factors were examined by CFA, as shown in Table 3. Cronbach’s ! coefficients were .854 for withdrawal, .793 for tolerance, .840 for positive expectancy, and .801 for interference with daily routines. These results verified the internal consistency of the instrument.
Exploratory Factor Analysis After calculating Pearson correlation coefficients between the variables, we assessed the pattern of relationships and performed exploratory factor analysis with principal axis factoring with varimax rotation. Four factors were extracted. The overall sampling adequacy of the 19-item scale was assessed using the Kaiser-MeyerOlkin test, and a high value of 0.910 was reported. The P value of the Bartlett test was less than .001, indicating that factor analysis was appropriate. Only one item, T5 (‘‘I spent my break time using my smartphone.’’), was deleted with a factor loading of less than 0.5, and 18 items, each with loadings over 0.5, remained in the final analysis (Table 2).
Data Analysis
Selection of Final Items
Analysis was performed using SPSS 12.0 (IBM, Armonk, NY) and AMOS 18.0 (IBM) for Windows. Cronbach’s ! coefficients were calculated to evaluate internal consistency. Construct validity was first confirmed by exploratory factor analysis. Then, construct validation was tested by confirmatory factor analysis (CFA).
A brief tool consisting of four factors and 18 items to measure smartphone addiction was developed. The instrument’s four factors included withdrawal (six items): irritation or anxiety about not being able to take smartphone messages; tolerance (five items): using the smartphone for longer than intended and feeling the urge to use it again right after using it; interference with daily routines (three items); and positive expectations (five items): the smartphone as something special and beneficial for doing everything the person wants to do.
T a b l e 1 Sociodemographic Characteristics of the Participants
Confirmatory Factor Analysis
Sociodemographic Characteristics Frequency Percentage Sex Age, y
School years
Male Female e20 21–25 26–30 31–35 Freshman Sophomore Junior Senior
60 468 284 215 21 8 106 68 310 44
11.4 88.6 53.8 40.7 4.0 1.5 20.1 12.9 58.7 8.3
To assess the fit of the 18-item, four-factor structure of the smartphone addiction survey, we performed CFA using AMOS 18.0. We used the goodness-of-fit index (GFI), adjusted GFI (AGFI), and normed fit index (NFI) to assess the model’s fit. In this study, GFI = 0.909, AGFI = 0.876, and NFI = 0.907. The value of these indices for good fitting models should be GFI > 0.9, AGFI > 0.8, and NFI > 0.90.42 Thus, the indices reported for this study indicated that the model was a good fit. The # 2 result revealed that the model failed to fit (# 2126 = 441.814, P < .001);
CIN: Computers, Informatics, Nursing & May 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
219
T a b l e 2 Exploratory Factor Analysis and Original 19-Item List No.
Items
Factors
P1 P2 P3 P4 P5 U1 U2 U3 T1 T2 T3 T4 T5 W1 W2 W3 W4 W5 W6
I intend to continue my use of the smartphone in the future. I feel pleasant and have fun while using my smartphone. It is something special for me to have the smartphone service I use. Using a smartphone is useful for my daily life. Using a smartphone is beneficial to what I want to do. I have experienced inability to concentrate in my class because of smartphone messages. I have contacted social network sites during classroom learning. I have used my smartphone during classroom learning. The people around me tell me that I use my smartphone too much. I am always thinking that I should shorten my smartphone usage time. I feel the urge to use my smartphone again right after I stop using it. I have used my smartphone for longer than I had intended. I spend my break time using my smartphone. I feel anxious about not being able to receive important calls or messages. I am lacking adequate sleep due to smartphone use at night. I am irritated when I am not in a hot zone (area unable to use smartphone.) I am always thinking that I have a message on my smartphone. I neglect matters other than smartphone use. I can’t stop using my smartphone even when there are many other things to be done.
1
2 0.609 0.766 0.721 0.756 0.748
3
4
0.622 0.788 0.861 0.650 0.797 0.564 0.589 0.777 0.673 0.783 0.772 0.726 0.695
P: positive expectations; U: interference with daily routines; T: tolerance; W: withdrawal
however, # 2 scores usually result in rejection of the model when large samples are used.43
Convergent Validity Convergent validity can be examined through factor loadings, average variance extracted (AVE), and construct reliability. Each item showed a significant loading onto its respective scale, as shown in the Figure 1. Factor loadings ranged from 0.61 to 0.85. The values of these indices for convergent validity should have a critical ratio greater than 1.965, P < .05.44,45 In this study, critical ratio was greater than 9.894, P = .00. For convergent validity, the AVE should be more than 0.5, and construct reliability should be more than 0.7. Table 3 shows the AVE and construct reliability of this study.
Discriminant Validity We examined correlation coefficients across the latent variables, the confidence intervals of the correlation coefficients across the latent variables, and the # 2 tests on an unconstrained model and a constrained model for discriminant validity. The correlation coefficients across the latent variables ranged from 0.44 to 0.85, as shown in Table 4, and indicated statistical significance. The confidence interval of the correlation coefficients between the latent variables (withdrawal and tolerance) was 0.943–0.755. We compared the # 2 and the degrees of freedom on an 220
unconstrained model and a constrained model, and $# 2/df was 54.893, which is larger than 3.48, indicating a statistically significant difference. For this reason, we were able to confirm discriminant validity.
Addiction and Distraction The Pearson correlation coefficient between addiction level and distraction experience was 0.352 (P < .01). A negative correlation was observed between addiction level and attitude toward smartphone use policies (r = j0.89), as shown in Table 4.
DISCUSSION The primary purpose of this study was to develop and validate a smartphone addiction scale. The candidate items for the smartphone addiction scale were developed by conducting a literature review and obtaining expert opinions in order to enhance the validity of the tool. In addition, an open-ended questionnaire and interviews were combined to recruit realistic opinions from the nursing students. The wording/reading level of the tool was on an elementary school level to ensure that the items were understood. Studies on smartphone addiction are few in number, and most studies have adopted Internet addiction tools.14,15 Kwon et al15 developed the Smartphone Addiction Scale,
CIN: Computers, Informatics, Nursing & May 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
FIGURE 1. Confirmatory factor analysis.
consisting of 33 questions, based on the Korean selfdiagnostic program for Internet addiction (K-Scale).37 However, they overlooked the characteristics of the smartphone, and it has been suggested that a shorter version of this instrument should be created. This study produced an 18-item, four-factor model of tolerance, withdrawal, interference with life routines, and positive expectancy reflecting the characteristics of smartphone usage. The tool developed in this study showed very good internal consistencies of the four factors. The internal consistency values suggest that, overall, the smartphone addiction scale in the study is a reliable measurement tool. This study
examined the factor structure of the smartphone addiction scale. All factor loadings were greater than 0.50 in exploratory factor analysis. In CFA, the GFI and NFI were greater than 0.90, both indicating the model was a good fit with the data. In testing the discriminant and convergent validity of the smartphone addiction scale using CFA with SEM, the model explained approximately 90% of the variance of the data collected, and the variables significantly explained the designated scales.44 We measured the Pearson correlation coefficient between addiction levels, distraction in the clinical practicum, and students’ attitudes about smartphone use policies. The
CIN: Computers, Informatics, Nursing & May 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
221
T a b l e 3 Standardized Regression Weights and AVE No. W6 W5 W4 W3 W2 W1 T4 T3 T2 T1 P5 P4 P3 P2 P1 U3 U2 U1
Factor
Standardized Estimates
P
AVE
Construct Reliability
0.734 0.788 0.786 0.760 0.685 0.750 0.763 0.780 0.630 0.674 0.517 0.616 0.631 0.854 0.630 0.784 0.760 0.658
& Vol. 33, No. 5, 216–224 & Copyright B 2015 Wolters Kluwer Health, Inc. All rights reserved.
F E A T U R E A R T I C L E
Development of a Brief Instrument to Measure Smartphone Addiction Among Nursing Students SUMI CHO, MSN, RN EUNJOO LEE, PhD, RN
Smartphones are becoming a necessity in our daily lives, and the market share of smartphones is growing rapidly. Although the smartphone was first introduced in South Korea in 2009, its rate of use was 67.6% by 2012, compared with 27% in 2011, more than a twofold increase in 1 year. South Korea has had the most rapid growth rate and highest level of smartphone use, worldwide.1 Smartphones that are built on a mobile computing platform have Web browsers, which provide entertainment, education, banking, games, and GPS navigation services through various applications. In addition, smartphones are used for social networking through Twitter and Facebook and are becoming central to online communication.2 These advantages have brought enormous conveniences to society; however, adverse results can be caused by the overuse of smartphones.2,3 Their use can be more addictive than the use of the computer or Internet because smartphones can be used anywhere and at any time. There is evidence that the use of cellular phones, including smartphones, is a significant source of distraction for driving, classroom learning, and work-related tasks, including those in healthcare settings.4,5 Users constantly check text messages, social media, and e-mail, and they surf the Web. In particular, the use of smartphones by healthcare professionals may lower the performance of tasks requiring mental concentration, resulting in medical errors, and posing serious threats to patient safety.6 Nevertheless, a number of studies have focused on the acceptability and effectiveness of smartphones in healthcare settings.7–10 216
Interruptions and distractions due to smartphone use in healthcare settings pose potential risks to patient safety. Therefore, it is important to assess smartphone use at work, to encourage nursing students to review their relevant behaviors, and to recognize these potential risks. This study’s aim was to develop a scale to measure smartphone addiction and test its validity and reliability. We investigated nursing students’ experiences of distractions caused by smartphones in the clinical setting and their opinions about smartphone use policies. Smartphone addiction and the need for a scale to measure it were identified through a literature review and indepth interviews with nursing students. This scale showed reliability and validity with exploratory and confirmatory factor analysis. In testing the discriminant and convergent validity of the selected (18) items with four factors, the smartphone addiction model explained approximately 91% (goodnessof-fit index = 0.909) of the variance in the data. Pearson correlation coefficients among addiction level, distractions in the clinical setting, and attitude toward policies on smartphone use were calculated. Addiction level and attitude toward policies of smartphone use were negatively correlated. This study suggests that healthcare organizations in Korea should create practical guidelines and policies for the appropriate use of smartphones in clinical practice. KEYWORDS Confirmatory factor analysis & Patient safety & Smartphone addiction & Tool development
Author Affiliations: Doctoral Student (Ms Cho) and Professor (Dr Lee), College of Nursing, Research Institute of Nursing Science, Kyungpook National University, Daegu, Korea. This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (NRF- 2013R1A1A2057977). The authors have disclosed that they have no significant relationship with, or financial interest in, any commercial companies pertaining to this article. Corresponding author: Eunjoo Lee, PhD, RN, College of Nursing, Research Institute of Nursing Science, Kyungpook National University 680, Gukchaebosang-ro, Jung-gu, Daegu, 700-422, Korea ([email protected]). DOI: 10.1097/CIN.0000000000000132
CIN: Computers, Informatics, Nursing & May 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
For example, smartphones can be used to engage students in nursing education7 and to improve patients’ selfmanagement of chronic illnesses.8 In order to prevent interruptions and distractions due to smartphones, research on smartphone use by workers in healthcare settings is essential. An assessment of smartphone use by healthcare workers through reviews of their own behaviors and their realization of its potential risks are important.11–13 Nursing students must be aware of the dangers to patient safety. Therefore, the early detection of smartphone addiction is critical before nursing students and nurses put their patients at risk during clinical practice. Internet addiction criteria have been considered when developing smartphone addiction criteria, as it has been hypothesized that smartphone addiction shares many aspects similar to Internet addiction.14 However, smartphones do have some different characteristics from the Internet. First, they are highly valued by individuals because of portability. In addition, the connectedness of smartphones can help users interact with other people. Second, smartphones are more personalized, as various apps are designed by smartphone users. Therefore, causes or symptoms of smartphone addiction will be different from those of the Internet. There is an urgent need to develop criteria associated with the characteristics of smartphone addiction.15 Hence, the purposes of this study were to investigate smartphone use among nursing students and to develop a smartphone addiction scale that reflects the characteristics pertinent to smartphone use. Moreover, this study analyzed the attitudes of nursing students regarding the use of smartphones in practice. The awareness of attitudes can provide a basis for the prevention of significant risks to patient safety, which can be caused by the overuse of smartphones by workers in healthcare settings.
Background Definition of Addiction and Common Features Addiction is defined as a ‘‘compulsive, uncontrollable dependence on a substance, habit, or practice to such a degree that cessation causes a severe emotional, mental, or physiological reaction.’’16(pp321) Some researchers have argued that the concept of addiction should be applied only to chemical substances, such as drugs or alcohol.17–19 However, others have suggested that problematic behaviors, such as pathological gambling,20 sex,21 computer game playing,22 and Internet addiction,23–26 also belong to the subset of addictive behavior. Peele27 reported that any compulsive or overused activity should be considered an addiction. Akers28 reported that the traditional concept of addiction included tolerance, dependence, and withdrawal and applied them to the physiological demand for a drug. The term psychological dependence currently is used to describe habitual behavior
in the absence of proof of physical addiction. The criteria for dependence are outlined in the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV).29 They include spending a great deal of time and money using the substance; using it more often and longer than one intends; trying to stop using or reducing its use by making repeated unsuccessful efforts to do so; giving up important social, family, occupational, or recreational activities to use it; reporting withdrawal symptoms when one stops using it; and recognizing physical and psychological problems, but continuing its use despite the consequences. Peele27 reported that the major motives for addictive behavior are relief of pain, irritability, anxiety, and other emotional states, such as enhanced control, power, and increased self-esteem (compensation). Addictive behavior can help a person feel more comfortable in his/her surroundings and can serve as a mood stabilizer or way of feeling good as a coping mechanism. The concept of Internet addiction was first introduced by Goldberg,23 based on the DSM-IV criteria.29 Afterward, Young24 continued research on Internet addiction using the DSM-IV criteria29 for psychoactive substance dependence. In her study, many of the participants had symptoms of online dependency, such as tolerance, loss of control, withdrawal, and impairment of academic, social, and occupational functioning. Brenner30 conducted research on self-selected respondents with a mean age of 34 years. The respondents experienced, on average, 10 signs of interference with their daily functioning, such as failure to manage time effectively, cutting down on sleep, missing meals, and other signs, based on the Internet Related Addictive Behavior Inventory. Eighty percent of the respondents reported that they experienced at least five of the signs. Internet addiction research has been conducted in Korea since 1999.31,32 Most of the research on Internet addiction was conducted using the tool developed by Goldberg and others.23–26 Many studies on Internet addiction were published in South Korea after a death from cardiopulmonary arrest related to Internet addiction and an Internet game–related murder.33,34 South Korea considers Internet addiction as one of its most serious health problems.35 The number of persons identified with a high risk of Internet addiction comprises 7.0% of the population who were using the Internet in 2013 and were between 5 and 54 years old, and the number of persons at risk of Internet addiction in Korea was 2 286 000.36 The most popular Internet addiction scale used in Korea is the K-Scale, which was developed by the National Information Society Agency.37 Ko et al38 assessed Internet addiction based on five dimensions: compulsive use, withdrawal, tolerance, interpersonal and health problems, and time management problems. Although many studies have been published on the effects of Internet
CIN: Computers, Informatics, Nursing & May 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
217
addiction, it was not recognized in the DSM-IV or in the International Classification of Diseases, 10th Edition; however, it has been included in DSM-V.39 Recently, most of the research investigations on addiction have shifted from the Internet to the mobile phone, and studies on mobile phone addiction have mostly used measures of Internet or television addiction.40 The first mobile phone addiction study in Korea was conducted as a master’s thesis.41 The author developed a mobile phone addiction tool based on the Internet addiction tool of Young24 and Goldberg23 and proposed four main concepts of tolerance, withdrawal, difficulty maintaining control over one’s use, and dependency. Koo40 developed four concepts through in-depth interviews with adolescents: tolerance, withdrawal, difficulty maintaining daily life, and compulsive-impulsive control. Smartphone addiction can be detected by criteria that are met when an addicted person is separated from the smartphone, for example, a repeated failure to cut back on usage time and feeling happier using it than being with family or friends. However, there have not been many studies on smartphone addiction. Thus, it is necessary to conduct further research to investigate the social and behavioral problems resulting from its abuse or overuse.
METHODS Conceptual and Theoretical Framework The concept development process consisted of three stages. In the first stage, an open-ended survey was sent to nursing students enrolled in a BSN program. In the second stage, a literature and reference search of Web sites and diverse databases were performed. In the third stage, interviews were conducted with six nursing students who had more than 2 years’ experience using a smartphone and felt, based on input from peer nursing students, that they overused the smartphone. After conducting the literature review and interviews, we identified three components common to both Internet and smartphone addiction: withdrawal, tolerance, and interference with daily routines. Smartphones are more individualized and multipurpose in comparison to the Internet.2 Our interviewees noted that they are possessive of their smartphones, feel as though they ‘‘get something special’’ by using them, and feel a connection with other people through them. Therefore, we added a component of positive expectations associated with smartphone use to the scale, to indicate how essential smartphones are in participants’ lives. To ensure face validity, all of the items were evaluated by the nursing students for their accuracy, appropriateness, and relevance to smartphone use. After a review of the first draft and feedback from the nursing students, 218
several items were rephrased. Finally, two experts on tool development were consulted in order to improve the consistency of the items and the structure of the tool.
Measures on Distraction and Policy Development on Smartphone Use In this study, nursing students were asked whether they experienced distractions from smartphone use while participating in their clinical practicum. In addition, they were asked if they agreed with establishing policies on smartphone usage in the workplace to prevent its overuse. The lower score indicates that the students did not agree with policies that restricted smartphone use in the healthcare settings. We measured the relationship between addiction level, distractions during the clinical practicum, and their opinions regarding smartphone use policies in the healthcare setting. Distraction was measured by four items, and students were asked to rate how often they were distracted on a scale ranging from never = 1, rarely = 2, sometimes = 3, usually = 4, to always = 5. Examples of the items were ‘‘I was disturbed by text messages during clinical practice,’’ ‘‘I was disturbed by calls during clinical practice,’’ ‘‘I was disturbed by messages from the social network sites during clinical practice,’’ and ‘‘I was disturbed by the other students who were using their smartphones during clinical practice.’’ The concept of policy development was measured by three items: ‘‘I think healthcare professionals should be prohibited from using smartphones during their shift,’’ ‘‘I think every healthcare organization should implement a policy or regulation that prohibits smartphone use by healthcare professionals during shift,’’ ‘‘I insist that every healthcare professional should turn off his/her smartphone during his/her shift.’’ These items were measured a 5-point Likert scale, and the responses ranged from strongly disagree (1) to strongly agree (5). A total of 428 students who had a clinical practicum during their coursework answered the items on distractions and the development of a policy to ban smartphone use in hospitals.
Data Collection Before data collection, approval was obtained from the institutional review board of Kyungpook National University Hospital (institutional review board 2012-07-016). First, the researcher made a telephone call to the directors of each nursing school to describe the study’s purpose and procedures and to obtain permission to collect data. After permission was obtained for data collection, the researcher visited each school and met with a student representative
CIN: Computers, Informatics, Nursing & May 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
from each class who was designated by the director of the school. The researcher handed over the questionnaires to the student representatives and explained the study’s purpose and procedures. The designated students distributed the questionnaires and return envelopes to the students so that there was no perceived pressure from the schools’ directors and the researcher. The designated students explained the purpose of the study and the procedures directly to the students. The questionnaire also had a cover letter, which explained the study’s purpose and how to complete the questionnaire. The cover letter explained that participation was entirely voluntary and anonymous. Students were informed that they could refuse to participate in the research at any time and that their completion or noncompletion of the questionnaire would not affect any of their course evaluations or grades in other courses. A total of 600 questionnaires, ranging from 100 to 200, were distributed to each nursing school after a discussion with the schools’ directors. Data were collected for 1 week to 10 days at each school according to the school’s class schedule. Completed questionnaires were enclosed in the return envelope and handed to the designated students. The researcher received the returned questionnaires by visiting the designated students or by commercial delivery at the researcher’s address. The return rate of the questionnaire ranged from 72% to 92%. However, five questionnaires were incomplete; therefore, 528 questionnaires were analyzed for the study (Table 1).
RESULTS Reliability Cronbach’s ! for the total scale was .902. The concept reliabilities for the four factors were examined by CFA, as shown in Table 3. Cronbach’s ! coefficients were .854 for withdrawal, .793 for tolerance, .840 for positive expectancy, and .801 for interference with daily routines. These results verified the internal consistency of the instrument.
Exploratory Factor Analysis After calculating Pearson correlation coefficients between the variables, we assessed the pattern of relationships and performed exploratory factor analysis with principal axis factoring with varimax rotation. Four factors were extracted. The overall sampling adequacy of the 19-item scale was assessed using the Kaiser-MeyerOlkin test, and a high value of 0.910 was reported. The P value of the Bartlett test was less than .001, indicating that factor analysis was appropriate. Only one item, T5 (‘‘I spent my break time using my smartphone.’’), was deleted with a factor loading of less than 0.5, and 18 items, each with loadings over 0.5, remained in the final analysis (Table 2).
Data Analysis
Selection of Final Items
Analysis was performed using SPSS 12.0 (IBM, Armonk, NY) and AMOS 18.0 (IBM) for Windows. Cronbach’s ! coefficients were calculated to evaluate internal consistency. Construct validity was first confirmed by exploratory factor analysis. Then, construct validation was tested by confirmatory factor analysis (CFA).
A brief tool consisting of four factors and 18 items to measure smartphone addiction was developed. The instrument’s four factors included withdrawal (six items): irritation or anxiety about not being able to take smartphone messages; tolerance (five items): using the smartphone for longer than intended and feeling the urge to use it again right after using it; interference with daily routines (three items); and positive expectations (five items): the smartphone as something special and beneficial for doing everything the person wants to do.
T a b l e 1 Sociodemographic Characteristics of the Participants
Confirmatory Factor Analysis
Sociodemographic Characteristics Frequency Percentage Sex Age, y
School years
Male Female e20 21–25 26–30 31–35 Freshman Sophomore Junior Senior
60 468 284 215 21 8 106 68 310 44
11.4 88.6 53.8 40.7 4.0 1.5 20.1 12.9 58.7 8.3
To assess the fit of the 18-item, four-factor structure of the smartphone addiction survey, we performed CFA using AMOS 18.0. We used the goodness-of-fit index (GFI), adjusted GFI (AGFI), and normed fit index (NFI) to assess the model’s fit. In this study, GFI = 0.909, AGFI = 0.876, and NFI = 0.907. The value of these indices for good fitting models should be GFI > 0.9, AGFI > 0.8, and NFI > 0.90.42 Thus, the indices reported for this study indicated that the model was a good fit. The # 2 result revealed that the model failed to fit (# 2126 = 441.814, P < .001);
CIN: Computers, Informatics, Nursing & May 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
219
T a b l e 2 Exploratory Factor Analysis and Original 19-Item List No.
Items
Factors
P1 P2 P3 P4 P5 U1 U2 U3 T1 T2 T3 T4 T5 W1 W2 W3 W4 W5 W6
I intend to continue my use of the smartphone in the future. I feel pleasant and have fun while using my smartphone. It is something special for me to have the smartphone service I use. Using a smartphone is useful for my daily life. Using a smartphone is beneficial to what I want to do. I have experienced inability to concentrate in my class because of smartphone messages. I have contacted social network sites during classroom learning. I have used my smartphone during classroom learning. The people around me tell me that I use my smartphone too much. I am always thinking that I should shorten my smartphone usage time. I feel the urge to use my smartphone again right after I stop using it. I have used my smartphone for longer than I had intended. I spend my break time using my smartphone. I feel anxious about not being able to receive important calls or messages. I am lacking adequate sleep due to smartphone use at night. I am irritated when I am not in a hot zone (area unable to use smartphone.) I am always thinking that I have a message on my smartphone. I neglect matters other than smartphone use. I can’t stop using my smartphone even when there are many other things to be done.
1
2 0.609 0.766 0.721 0.756 0.748
3
4
0.622 0.788 0.861 0.650 0.797 0.564 0.589 0.777 0.673 0.783 0.772 0.726 0.695
P: positive expectations; U: interference with daily routines; T: tolerance; W: withdrawal
however, # 2 scores usually result in rejection of the model when large samples are used.43
Convergent Validity Convergent validity can be examined through factor loadings, average variance extracted (AVE), and construct reliability. Each item showed a significant loading onto its respective scale, as shown in the Figure 1. Factor loadings ranged from 0.61 to 0.85. The values of these indices for convergent validity should have a critical ratio greater than 1.965, P < .05.44,45 In this study, critical ratio was greater than 9.894, P = .00. For convergent validity, the AVE should be more than 0.5, and construct reliability should be more than 0.7. Table 3 shows the AVE and construct reliability of this study.
Discriminant Validity We examined correlation coefficients across the latent variables, the confidence intervals of the correlation coefficients across the latent variables, and the # 2 tests on an unconstrained model and a constrained model for discriminant validity. The correlation coefficients across the latent variables ranged from 0.44 to 0.85, as shown in Table 4, and indicated statistical significance. The confidence interval of the correlation coefficients between the latent variables (withdrawal and tolerance) was 0.943–0.755. We compared the # 2 and the degrees of freedom on an 220
unconstrained model and a constrained model, and $# 2/df was 54.893, which is larger than 3.48, indicating a statistically significant difference. For this reason, we were able to confirm discriminant validity.
Addiction and Distraction The Pearson correlation coefficient between addiction level and distraction experience was 0.352 (P < .01). A negative correlation was observed between addiction level and attitude toward smartphone use policies (r = j0.89), as shown in Table 4.
DISCUSSION The primary purpose of this study was to develop and validate a smartphone addiction scale. The candidate items for the smartphone addiction scale were developed by conducting a literature review and obtaining expert opinions in order to enhance the validity of the tool. In addition, an open-ended questionnaire and interviews were combined to recruit realistic opinions from the nursing students. The wording/reading level of the tool was on an elementary school level to ensure that the items were understood. Studies on smartphone addiction are few in number, and most studies have adopted Internet addiction tools.14,15 Kwon et al15 developed the Smartphone Addiction Scale,
CIN: Computers, Informatics, Nursing & May 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
FIGURE 1. Confirmatory factor analysis.
consisting of 33 questions, based on the Korean selfdiagnostic program for Internet addiction (K-Scale).37 However, they overlooked the characteristics of the smartphone, and it has been suggested that a shorter version of this instrument should be created. This study produced an 18-item, four-factor model of tolerance, withdrawal, interference with life routines, and positive expectancy reflecting the characteristics of smartphone usage. The tool developed in this study showed very good internal consistencies of the four factors. The internal consistency values suggest that, overall, the smartphone addiction scale in the study is a reliable measurement tool. This study
examined the factor structure of the smartphone addiction scale. All factor loadings were greater than 0.50 in exploratory factor analysis. In CFA, the GFI and NFI were greater than 0.90, both indicating the model was a good fit with the data. In testing the discriminant and convergent validity of the smartphone addiction scale using CFA with SEM, the model explained approximately 90% of the variance of the data collected, and the variables significantly explained the designated scales.44 We measured the Pearson correlation coefficient between addiction levels, distraction in the clinical practicum, and students’ attitudes about smartphone use policies. The
CIN: Computers, Informatics, Nursing & May 2015 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
221
T a b l e 3 Standardized Regression Weights and AVE No. W6 W5 W4 W3 W2 W1 T4 T3 T2 T1 P5 P4 P3 P2 P1 U3 U2 U1
Factor
Standardized Estimates
P
AVE
Construct Reliability
0.734 0.788 0.786 0.760 0.685 0.750 0.763 0.780 0.630 0.674 0.517 0.616 0.631 0.854 0.630 0.784 0.760 0.658
