Background

The aging population leading to the expansion of healthcare infrastructure has contributed to an increase in global demands for healthcare professional workforce. A pool of health care professionals who are better prepared in caring for people with multiple chronic conditions are crucial in addressing current healthcare challenges [1]. A worldwide shortage of nurses has been reported [2, 3]. Compared with other healthcare courses, nursing courses have the largest recruitment target in order to meet the ever-growing demand for a nursing workforce. This recruitment target is met with threat from the increase intake of other healthcare courses.

School-age students are attracted to healthcare related courses due to career attributes such as altruism, job stability, financial remuneration, professional prestige, job autonomy and familial influences [4]. There are differences in factors influencing career choice among healthcare students. While the influence of altruism has a significant influence in the career choice for nursing [5], medicine [6] and pharmacy [7], it has less influence in dentistry [8]. The influence of financial remuneration was important in pharmacy [9] and dentistry [10], but less so in nursing [11] and medicine [6]. The influence of family appeared to be important on career choice in pharmacy [12] and medicine [6], while its influence on career choice in nursing and dentistry is inconclusive. As most of the existing studies examined the influences of career choice on a single healthcare discipline, future studies are needed to compare the influences of career choice between healthcare disciplines.

Among healthcare courses, nursing has been perceived as an unpopular choice and is seldom a first-preference career [13, 14]. With the availability of instruments to measure the perception of nursing, the views of school-age students were extensively explored [15,16,17,18]. Nursing was perceived by school-age students as less intellectual [19], having low job autonomy [20], involving too much hard work [21], holding ‘dirty’ work [22], and receiving low salaries [23]. Although the psychometric properties of the instruments used in these studies require a more in-depth evaluation [24], the outcomes lend further support to the belief that nursing as a career choice holds a low public image.

To further understand which career attributes of nursing were perceived as less ideal, an instrument known as the Indiana Instrument was designed to determine the differences in attitudes between ‘ideal career’ and ‘nursing as a career’. Using two sets of questionnaires with 17-parallel items, this tool was widely used to understand the difference between the ideal career and nursing career among school-age students [25,26,27]. A nursing career was found to match an ideal career in terms of ability to care for others, hard work, busyness and job security [25, 27]. A study by Cohen et al. [25] reported that nursing was less than ideal in terms of respect, appreciation, autonomy and financial remuneration. An earlier study by Mendez and Louis reported nursing career to fall short of an ideal career in terms of knowledge, power and job stability [26].

The utility of the existing instruments has primarily focused on school students’ perception of nursing as a career. It appears to be more important to target on students who are interested in healthcare career, as these students often chose a particular healthcare career after consideration of related health profession. However, existing instruments do not allow for comparison of influences of healthcare career choice with the perception of nursing. There is a need to examine why those who are inclined to choose healthcare courses eventually did not select a nursing career [28]. The comparison of career choice influences and perception of nursing as a career choice among healthcare students would highlight the comparatively stronger attractive factors, which could be used to identify specific recruitment strategies to attract student to choose nursing. Prior to conducting the study, there was a need to develop and evaluate the psychometric properties of scales to compare the influences of healthcare career choice with perceptions of nursing as a career choice.

Methods

A two-phase prospective study was conducted. Phase 1 included the development and content validation of the instrument, (1) Healthcare Career Choice (HCC) scale and (2) Nursing Career Choice (NCC) scale, and phase 2 involved the psychometric evaluation of the instrument.

Phase 1: Development and content validation of the scales

Subscale specification

A qualitative exploratory descriptive study was conducted earlier to explore factors influencing career choice and perception of nursing among healthcare students. The study was conducted with 59 first-year healthcare students from three higher education institutions undertaking health-science related courses including dentistry, dental hygiene, medicine, nursing, pharmacy, physiotherapy and occupational therapy through eight focus group discussions. Each focus group, consisted of six to eight participants and lasted about 60–75 min. Six themes emerged from the thematic analysis: (1) personal interest, (2) prior healthcare exposure, (3) academic performance, (4) perceived nature of work, (5) job prospects, and (6) social influences [29]. These themes were identified as subscales for the HCC-NCC instrument.

Item development

The Indiana Instrument was the most widely used instrument for determining the differences of career attributes or characteristics between nursing and ideal career [24]. Therefore, the HCC-NCC instrument was developed to follow the concept of the tool which comprised of two set of scales with parallel items. These scales enable the comparison between the influences of healthcare career choice and perception of nursing as a career choice. Fifty parallel items were formulated for each scale, with seven to nine items in each subscale. All items were derived from the participants’ words during the focus group discussions, broad literature review including existing instruments, and the developers’ clinical experience and inferential reasoning. A large pool of items was selected and sampled systematically to include all the content that could be potentially relevant to the target construct. This allowed some items to be deleted, considering that subsequent psychometric analyses could potentially identify weak and unrelated items [30]. All items were developed to be rated on a five-point Likert rating scale (1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree, 5 = strongly agree), with higher scores indicating more influential career attribute and lesser scores indicating less influential career attribute.

Content validation

Twelve content experts were invited to evaluate the content validity of the initial 50 instrument. These experts included three nursing educators from three restructured hospitals, seven nursing and two allied health lecturers from two higher educational institutions, and two health administrators from the Ministry of Health. The experts independently reviewed each item using the 4-point rating scale (1 = not relevant, 2 = somewhat relevant, 3 = quite relevant, 4 = very relevant). Content validity index (CVI) was computed for each item by the number of experts giving a rating of either 3 or 4 and divided by the overall number of overall experts. Three items with CVI <0.75 were removed as they were identified as being vague or similar to other items [31]. Five more items were added based on the experts’ recommendations. The revised instrument was sent for a second round of content validation which yielded at least a CVI of 0.78 for each item, a CVI of 0.78 for the HCC, and a CVI of 1.0 for the NCC scale.

Pilot testing

A pilot test was conducted with 15 medical students and 15 nursing students to establish face validity of the instrument, consisting of 52-parellel items. This pilot testing aimed to: (1) evaluate the clarity of each item and instructions provided, (2) receive feedback about the format of the tool, (3) find out the time taken to complete the instrument, and (4) perform a preliminary check on the internal consistency of the scales. The participants did not express any difficulty with the wording and format of the instrument. They reported taking approximately 15–20 min to complete the instrument. The preliminary Cronbach’s alpha of the instrument from the pilot test was reported to be 0.72 to 0.94.

Phase 2: Psychometric evaluation of the instrument

A psychometric testing was conducted to evaluate the psychometric properties of the newly developed HCC-NCC instrument, including factor structure, internal consistency, test-retest reliability, and concurrent validity.

Setting and participants

A convenience sampling method was used to recruit students from one university and one polytechnic institution in Singapore. All first-year students, between 16 and 25 years old and undertaking healthcare related courses at the two institutions, were invited to participate in the study. The initial 52-item HCC-NCC instrument was administered to a total of 300 participants. The sample size of 300 participants was based on Gorsuch’s recommendation of a minimum of 5 respondents per item, and a probability of receiving incomplete questionnaires [32]. There were 283 completed questionnaires.

Data collection and procedure

Following approval from a University Institutional Review Board, email invitations with link to the survey questionnaires were sent out to the potential participants between July to August 2015. The questionnaires included the initial 52-item HCC-NCC instrument and the Indiana Instrument with 17 parallel items. The content validity of the Indiana Instrument was established by a panel of experts and contents from the literature reviews, with the Cronbach’s alpha reported to be 0.84 for the ideal career and 0.81 for the nursing career [18]. About two weeks after the completion of the questionnaires, participants were invited to complete the parallel scales for a second time, to establish the stability of the instrument. The two weeks’ time interval would be lengthy enough for respondents to be unable to remember their original responses, and yet not too long for their attitudes of the material to have changed [33].

Data analysis

Descriptive statistics were computed for demographic variables. Construct validity was assessed by an exploratory factor analysis using the principal component analysis and varimax rotation to examine the factor construct of the instrument. The number of factors were determined by eigenvalues >1. Concurrently validity was tested by Pearson’s correlation coefficients. Internal consistency was evaluated using Cronbach’s alpha and item-to-total correlation. Test-retest reliability was assessed using Intraclass Correlation Coefficient (ICC).

Results

The demographic characteristics of the 283 participants are presented in Table 1. The students were between the age of 16 and 25, with a mean age of 19 (SD = 1.79). The majority of them were female (n = 224, 79.2%) and nursing students (n = 175, 61.8%).

Table 1 Summary of demographic characteristics of healthcare students
Full size table

Construct validity

Exploratory factor analysis was used to examine the factor structure of both HCC-NCC parallel scales. Bartlett’s test of sphericity was statically significant for the HCC (×2 (595) = 5161.30, p < 0.001) and NCC (×2 (595) = 6344.80, p < 0.001). The Kaiser-Meyer-Olkin measure of sampling adequacy was 0.90 for the HCC and 0.91 for the NCC scale, indicating that the sample was large enough to perform factor analysis.

The initial principal component analysis (PCA) showed that all 52 items extracted a eleven-factor solution for HCC and nine-factor solution for NCC using eigenvalue >1. The factor analysis using varimax rotation was then conducted. Items with factor loading <0.4 or items that loads equally on two factors were removed. Any item removed for HCC was also removed for NCC to ensure consistency in keeping the items in both scales parallel to each other. A total of 17 items were removed with this procedure. Using PCA with eigenvalues >1, six factors were extracted from the remaining 36 items which accounted for 59 and 64% of the variance for the HCC and NCC scales respectively. The scree plots in Fig. 1 illustrated the number of factors.

Fig. 1
figure 1

Scree plot for HCC scale (left) & NCC scale (right)

Full size image

Tables 2 and 3 summarises the result of the rotated component matrix using varimax rotation. Using the loading criteria of 0.40 [34], 35 items demonstrated strong factor loadings ranging from 0.48 to 0.82 for the HCC and 0.41 to 0.84 for the NCC scale. Factor 1, personal interest, consisted of five items with factor loadings of 0.57 to 0.84 and accounted for 12.60 and 12.89% of the variance for the HCC and NCC scales respectively. Factor 2, prior healthcare exposure, consisted of six items with factor loadings of 0.55 to 0.78 and accounted for 12.11% (HCC) and 12.52% (NCC) of the variance. Factor 3, self-efficacy, composed of four items with factor loadings of 0.49 to 0.73 and accounted for 10.71% (HCC) and 11.97% (NCC) of the variance. Factor 4, perceived nature of work, consisted of six items with factor loadings of 0.41 to 0.77 and accounted for 10.21% (HCC) and 10.99% (NCC) of the variance. Factor 5, job prospects, consisted of seven items with factor loadings of 0.44 to 0.79 and accounted for 7.78% (HCC) & 8.12% (NCC) of the variance. Factor 6, social influences, consisted of eight items with factor loadings of 0.45 to 0.79 and accounted for 6.32% (HCC) & 7.58% (NCC) of variance. This six factor solutions, therefore, represented the core subscales of the instrument.

Table 2 Principal component analysis with Varimax rotation of the HCC scale
Full size table
Table 3 Principal component analysis with Varimax rotation of the NCC scale
Full size table

Concurrent validity

There was a moderately strong significant positive correlation in the total scores between the Indiana and the HCC-NCC instruments (r = 0.76, p < 0.001). There were significant positive correlations between the Indiana’s nursing career scales and NCC scale (r = 0.73, p < 0.001) and between the Indiana’s ideal career scale and HCC scale (r = 0.61, p < 0.001).

Internal consistency and inter-item correlations

Table 4 present the results for internal consistency and inter-item correlations. The Cronbach’s alpha of the six subscales ranged from 0.71 to 0.89 for the HCC scale and 0.78 to 0.89 for the NCC scale. The correlation coefficients between items and their respective subscales ranged between 0.39 to 0.78 for the HCC scale and 0.46 to 0.80 for the NCC scale. The overall Cronbach’s alpha for all 35 items was 0.93 for the HCC and 0.94 for the NCC scale.

Table 4 Scale item statistics
Full size table

Test-retest reliability

Twenty-eight participants completed the HCC-NCC instrument for the second time after two weeks’ interval. The ICC was 0.63 (95% CI = 0.267–0.813, p = 0.002) for HCC and 0.60 (95% CI = 0.206–0.798, p = 0.005) for NCC.

Discussion

The competition among healthcare courses to attract high quality school leavers is becoming increasingly intense. This is particularly challenging in the nursing course which requires the largest recruitment target to meet the workforce demand [35]. Using the Indiana Instrument, most studies to date have primarily focused on school students about the differences in attitudes between ‘ideal career’ and ‘nursing as a career’ [25, 27]. By applying the parallel scales concept of the Indiana Instrument, we developed and tested the HCC-NCC instrument in order to compare the influences of healthcare career choice and perception of nursing as a career choice among the healthcare students.

The content validity of the HCC-NCC instrument was achieved through a combination of literature review, findings from a previous qualitative study, and experts’ validation. Inclusion of a total of 12 experts from a variety of settings including education institutions, hospitals and the Ministry of Health provided a wide perspective of the tool. Based on the experts’ validation, the CVI achieved Lynn’s (1986) criterion for content validity [36]. Additionally, a pilot test with 30 non-experts, established the face validity.

The construct validity of the HCC-NCC instrument was assessed in a factor analysis by using a principal component analysis. Factor analysis was justified with Bartlett’s test of sphericity, while the calculated Kaiser-Meyer-Olkin measure of sampling adequacy indicated that there was adequate sample for factor analysis [37]. Modifications to the instruments were made following the factor analysis to remove items that were found to be weak or unrelated. Finally, all included items possessed factor loadings of >0.4 and accounted for 59 and 64% of the variance for the HCC and NCC scales respectively. The factor analysis extracted six factors corresponding with five out of the six career factors that emerged from the findings of a qualitative study [29].

The “personal interest” subscale, refers to the students’ personal interest in their chosen professions. According to Holland’s theory of “Career Typology,” individuals choose career environments that best fit their personality and interest [38]. Several studies have shown that students pursuing healthcare careers tend to have similar interests [5, 7, 39]. A personal interest in their chosen professions based on notions of altruism, opportunity to interact with others, as well as an interest for science-related subjects were expressed among the healthcare students in a previous study [29].

The second factor, “prior healthcare exposure”, reflects how healthcare related experiences could influence their choice of a healthcare career, both positively and negatively. The influence in the developmental stage on career choice, spanning from school years to young adulthood, has long been established by a vocational psychologist [40, 41]. The exposure of students to healthcare-related work, including observing a healthcare professional at work has shown to draw students into a healthcare career [42, 43].

The third factor, “self-efficacy”, refers to a set of self-beliefs about one’s personal competence to perform the actions required to produce outcomes in particular domains [44]. Applying the social cognitive career theory, the links between self-efficacy and career choice has been well-established by Lent et al. [45]. Academic ability which often reflects intelligence serves as an important indicator for an individual to evaluate one’s self-efficacy to an academic related career choice. Nursing is often perceived as a course for students with low academic ability which could have deterred academically-abled students from joining the course [46].

The factor, “job prospects”, considers the practical aspects of a healthcare career that could influence the career choice. This includes a desire for job opportunity, job stability, and good income. Healthcare careers are often highly regarded for the ease of getting a job and job stability [10, 11]. Nursing is however often perceived as a poorly paid job [23].

The factor, “perceived nature of work”, relates to how students’ perceived the characteristics of the healthcare careers that influenced their choice of career. The characteristics associated with nursing work, including the involvement of too much hard work [21], and ‘dirty’ work [22] have deterred students from joining the nursing profession.

The final factor, “social influences”, includes social status, gender-type and significant others that have been found to have a significant impact on the students’ career decision-making process. Social influences by significant others was found to affect the career aspirations of Asian students more significantly than the Western students [47, 48].

The concurrent validity of the HCC-NCC instrument was examined by correlating with the Indiana Instrument, both of which were administered at the same time. The Indiana Instrument is one of the most widely used tool for determining the difference between the ideal career and a nursing career with tested reliability and validity [24]. A significant strong positive correlation was found between these two scales, confirming the concurrent validity of the HCC_NCC instrument.

Besides evidence to support the validity of the HCC-NCC scales, the study demonstrated a satisfactory internal consistency as reflected by the Cronbach’s alpha of 0.71 to 0.89 for both scales and its subscale, and the high correlation between the items with their respective subscales. The stability of the HCC-NCC instrument was also demonstrated.

In comparison to the existing 17-parallel items Indiana Instrument, the 35-parallel items HCC-NCC instrument offer a more comprehensive comparison of career choice influences. The instrument has many potential applications for future use. The instrument provides a comparison of factors influencing healthcare career choice and perception of nursing as a career. Such comparison can highlight the differences between career influences in non-nursing careers and a nursing career which has the potential to identify specific strategies to enhance nursing recruitment. Each of the parallel scales can also be used as a stand-alone scale. While the HCC scale can be used for identifying factors influencing healthcare career choice, the NCC scale can be used for examining factors influencing nursing as a career choice. Findings obtained from the instrument may aid in the development of recruitment programmes. The effectiveness of the recruitment interventions can be evaluated using this instrument in a pretest-posttest study or in randomised-controlled trials. Future studies may consider adapting the HCC scale for graduating students to examine influences of career choice on healthcare sub-specialities.

Limitations

Convenience sampling was used to recruit participants in this study, which may lead to potential biasness such as under-representation of different targeted groups within the sample. Future studies with a larger sample size and random sampling may lend added support to the validity and reliability of the instrument. While an acceptable test- retest reliability is achieved, it could be strengthened through a larger sample size of participants attempting the retest after two weeks.

Conclusions

This study has developed and established valid and reliable 35-item instrument with six career-choice factors to compare healthcare career choice and nursing as a career choice (See Additional file 1). The HCC-NCC instrument proves useful for future studies to determine how strongly each of the factors is associated with the healthcare career choice and with nursing as a career, and to examine the differences between students’ healthcare career choice and their perception of nursing as a career choice. The findings may have potential implications for educational institutions and policy-making planners to consider specific recruitment nursing strategies.