Background

Globally, the nursing profession is facing shortages, high turnover, and inequitable distribution of the nursing workforce [1, 2]. These are particularly acute problems for South East Asia (SEA), where the nurse-to-population ratio is approximately 2–3 nurses per 1000 population [3], ten times lower than that in Europe although higher than in Sub-Saharan African countries [4]. In SEA, the aging population and high incidence of chronic non-communicable diseases is increasing the demand for nursing and other social care personnel. Moreover, an aging nurse population, lower rates of replenishment by young nurses, and attrition through international migration to high income countries has considerably increased the pressure on the diminishing nurse workforce [5, 6]. The Association of South East Asian Nations (ASEAN) Economic Community (AEC) facilitates the free flow of skilled labor across the AEC, and may have major repercussions on nursing demand and supply in the region [7] where capabilities of health system in attracting and retaining their local staff are unknown [8]. A review suggested that a study investigating these issues conducted locally is essential [9].

Another alarming sign is that aging of nursing profession has been reported in many countries [10–12]. These comprise a large component of the current nurse population and gradually approaching retirement. In addition, the decline in younger women choosing nursing as a career is also observed. However, information about the retirement rate of nurses is limited.

Although numerous studies have identified factors influencing both the supply and demand of nurse labour, most of these studies were conducted in Western countries [5, 6, 13]. Furthermore, most studies investigating nursing crisis are cross-sectional studies. Cohort studies have greater potential to address the crisis as well as the emerging issues in women’s health such as breast cancer, cervical cancer, osteoporotic fractures, postmenopausal illness and its consequence, etc. The study of this kind includes the oldest and the largest nurse cohort study, the Nurse Health Study, initiated in 1976 in the Unites States which focused in cancer and other illnesses [14, 15]. The Danish Nurse Cohort Study (DNCS) established in 1993 investigated working conditions and health including hormone therapy use and risk factors of various health outcomes [16]. The Japan Nurses’ Health Study (JNHS), initiated in 2001, investigated effects of lifestyle and healthcare practices on women’s health [17, 18]. The Nurses and midwives e-Cohort Study, established in 2006, aimed to examine factors associated with recruitment and retention of the nursing and midwifery workforce among nurses in Australia and New Zealand [19]. Recently in 2009, the Nurses’ Early Exit Study (NEXT), a cross-cultural and longitudinal project conducted in ten European countries to analyze the relationship between working conditions and nursing workers’ health [20, 21]. Then Thailand established a nurse cohort study.

A longitudinal cohort study on the health and working life of registered nurses in Thailand was launched in 2009. The Thai Nurse Cohort Study (TNCS) is a 20-year prospective cohort study designed to generate evidence on the workforce dynamics and health status of Thai nurses. Specifically, the TNCS aims to: 1) determine rates, patterns, trends and determinants of job transition, and 2) assess the prevalence, incidence, and long term changes of key health problems among Thai nurses.

It is expected that the TNCS data will clarify a number of key policy and research questions for Thailand, a middle income country; particularly as numerous other studies have investigated nurse populations in high-income countries [14, 22]. The cohort design of the TNCS also has potential to address the dynamics of job transition across public and private sectors, migration, determinants of attrition and retention, and to monitor the impact of policy interventions.

Methods

Study design

The TNCS uses a longitudinal design, comprising multiple age cohorts of registered nurses (RN) who hold professional licenses granted by the Thailand Nursing and Midwifery Council (TNC). Participants were selected using age-stratified random sampling using 5-year age groups, from 20 to 64 years (nine age groups in total). The first round, representing the baseline survey, was sampled in September 2009. Data were collected using a self-administered mailed questionnaire.

The TNCS was designed to enroll new members every 3 years, with these new entry cohorts named according to the year of inclusion. New entry cohorts will be sampled from RN newly registered with the TNC and listed in the TNC database. This means the TNCS will have multiple cohorts: the initial cohort collected in 2009, a 2012 new entry cohort, and proposed subsequent new entry cohorts (the 2015, 2018, 2021, 2024, and 2027 cohorts). Cohort members will be followed until 2030, unless they withdraw from participation, are diagnosed with cancer, reach the age of 68 years, or pass away. The second round of the survey, conducted in 2012, served as a follow-up for the initial cohort and provided baseline data for the new entry cohort. The second survey was administered using a mixed mode, either a mailed questionnaire or web-based survey, depending on that member’s preferences. For future surveys, a web-based questionnaire is planned because of the high cost implications of mailed questionnaires.

Measurements

The questionnaire in the first survey comprised three main sections (Table 1): personal data (birth date, address, marital status, education, income, debt, and other burdens); employment characteristics, job transition since graduation, and job strain questions modified from the Job Content Questionnaire short version (JCQ) [23]; and health status, including the EuroQol five-dimension questionnaire (EQ-5D) [24], the short form of the International Physical Activity Questionnaire (IPAQ) [25], history and current illness, and self-care activities. The follow-up survey repeated the employment and health status sections.

Table 1 Parameters and their use in the baseline survey of the Thai Nurse Cohort Study
Full size table

Sample size calculation

A sample size of 30,000 was initially planned, based on the aim of detecting any events with a proportion as low as 1 % with +/−0.1 % precision (i.e., it offered a relative precision of +/−10 % with a two-sided 95 % confidence level). We also allowed 40 % for expected non-responses because of invalid contact addresses, refusal to participate, and loss to follow-up; this excess was slightly better than that used in a study involving nurses in Japan [26]. Therefore, the sample size was inflated to 50,000. However, administrative reasons meant that the actual number of questionnaires mailed out was 50,200.

Sampling techniques

TNCS participants were classified into two categories: the initial cohort and the new entry cohort.

The initial cohort comprised RN who were randomly selected from the TNC database, who signed a consent form, and who responded to the first-wave questionnaire. The TNC database used as the sampling frame included all RN who held nursing licenses granted by the TNC as of 2008, with a total of 142,699 RN forming the sampling frame. We then applied an age-stratified random sampling technique for sample selection. First, RN aged 20–64 years were stratified into nine age groups based on 5-year intervals (< 25, 25–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, and 60–64 years). Based on the sample size of 50,000, the average size of each age group was calculated as 5556 nurses. However, as the oldest two age groups (55–59, and 60–64 years) were smaller than the average size, we selected all members of these groups. Participants from the remaining seven age groups were randomly selected. First, a set of random numbers were generated using Stata statistical software (StataCorp, College Station, TX). This assigned a random number to each nurse’s Citizen Identification Number (CID). CID were then sorted according to the assigned random numbers. Nurses in each age group were listed in the sorted order. Finally, nurses were selected consecutively, starting from the first order, until the specified size was achieved. The sample size was equal for all seven age groups. For administrative reasons, a total of 50,200 nurses were eventually selected for the initial cohort.

The new entry cohort included all RN who held nursing licenses granted by the TNC after 2008 and were not part of the sampling frame of the initial cohort. From a total of 6402 RN in this group, 1600 (25 %) were randomly selected. The selection methods as used for the initial cohort were applied.

Statistical methods

We provided the probability of responses as the inverse proportion to the probability of responses to be used in subsequent papers for weighting in statistical analyses. To describe the cohort profile, we presented the results as number and percentage for categorical variables, and mean and standard deviation (SD) for continuous variables. These descriptive statistics were used for demographic characteristics, employment history and career advancement, current employment status, intention to leave the nursing profession within the next 2 years, and quality of life based on the EuroQol five-dimensions, in both the initial and the new entry cohorts. All analyses were performed using STATA version 13 (StataCorp).

Ethics and good clinical practice

The present study was conducted according to the principles of Good Clinical Practice (International Conference on Harmonization Tripartite Guideline for Good Clinical Practice), the Declaration of Helsinki, and national laws and regulations for clinical studies. The TNCS was approved by the Thai Ministry of Public Health Ethics Committee for Human Research. All participants signed a consent form for participation as a cohort member of the TNCS.

Results

Questionnaire responses and recommended weighting for statistical analysis

In total, 50,200 questionnaires were sent via postal mail for the first survey wave, 18,198 (36.3 %) were returned owing to invalid addresses or no recipient (Fig. 1). There were 18,756 questionnaires returned with signed and dated consent forms. This gave a response rate of 58.6 % for the 32,002 deliverable questionnaires, assuming the questionnaires were received. The first 10,000 responses were received within 2 months of the mail out.

Fig. 1
figure 1

Population, sample, and respondents of the baseline survey of the Thai Nurse Cohort Study

Full size image

Although the overall response rate for the first survey was 58.6 % (Fig. 1), the three younger age groups had the lowest response rates (see Table 2 for the probability of responses). For the new entry cohort, the weight was determined based on the inverse proportion of responses, which was 2.61 (3020/1156).

Table 2 Recommended weight for data analysis of the initial cohort of the Thai Nurse Cohort Study
Full size table

The second round, conducted in August 2012, served as a follow-up for the initial cohort and the baseline survey of the new entry cohort. The follow-up rate for the initial cohort was 60.2 % (11,282 of 18,756 RN who responded in the first round). In the second round, RN who were registered with the TNC in 2009 and 2010, and who were aged between 23–27 and 22–26 years, respectively, formed the sampling frame of the new cohort. Nurses in the new entry cohort were aged 20–24 years when the sampling frame for first round of the study was established. Of the 7050 newly registered RN, 648 were excluded as their age exceeded the required age range. Of the remaining 6402 new RN, 25 % were identified as the final sample size. To allow for 47 % of non-responses based on the previous round survey, a sample size of 3020 was required. Of these, 1156 responded and were enrolled as new entry cohort members. The response rate for the new entry cohort was 38.3 % (Fig. 1).

Demographic characteristics

The majority of the initial cohort members were female (96.7 %), married (60.8 %), lived in the central region of the country (38.7 %), and had attained a bachelor’s degree (57.2 %) (Table 3). About half (48.5 %) were aged older than 44 years. The average age was 43.7 ± 9.7 years. The average age of the new entry cohort was 23.1 ± 3.5 years, and a much higher proportion were single (82.3 %).

Table 3 Demographic characteristics of first-wave survey respondents, presented as number and percentage unless indicated otherwise
Full size table

Employment history and career advancement

A majority of the initial cohort (78.0 %) graduated from nursing colleges owned by the Thai Ministry of Public Health; 79.5 % had been serving mandatory nursing service in public sector after graduation; and 63.1 % were employed as RN when they started working, while the remaining nurses started their career as a Technical Nurse (diploma level) (Table 4). Only 7.9 % had graduated from a public university. The new entry cohort showed similar pattern in their employment history. Around 5833 (31.1 %) of the initial cohort had a diploma-level education at first employment, 75.6 % of these were trained and upgraded (67.1 % to a bachelor’s degree, 8.5 % to a master’s degree, and 0.03 % a doctoral degree). The entire new entry cohort held a bachelor’s degree at their first employment.

Table 4 Employment history and career advancement of the first-wave of the Thai Nurse Cohort Study
Full size table

Current employment status

A majority of the initial cohort were engaged in nursing services (61.2 %), and 21.0 % were mid-level nursing administrators (Table 5). Around 80.8 % were civil servants and 79.0 % worked in hospitals. Only 9.0 % worked in health centers or private clinics. Again, a similar pattern of current employment was found in the new entry cohort.

Table 5 Current employment status of the first-wave of the Thai Nurse Cohort Study
Full size table

Intention to leave nursing career

Of the 16,470 nurses in the initial cohort who were working in the nursing profession at the time of the baseline survey, 2306 (15.4 %) stated that they intended to leave their nursing career. This proportion was slightly lower in the new entry cohort, with 11.2 % indicating an intention to leave the profession in the next two years (Table 6). The number of nurses intending to leave the profession increased slightly with age. More research nurses reported an intention to leave, (16.7 %) compared with other employment categories. In the initial cohort, 18.0 % of nurses who worked in research areas and 20.4 % who worked in the private sector reported an intention to leave the profession.

Table 6 Intention to leave the nursing profession within the next 2 years, as reported in the Thai Nurse Cohort Study
Full size table

Quality of life, health status

Overall, the quality of life of Thai RN, based on EQ-5D scores, was good, with an overall mean score of 0.758 ± 0.191 for the initial cohort and 0.751 ± 0.178 for the new entry cohort (Table 7). In the initial cohort, the mean quality-of-life score remained unchanged at follow-up (3 years after the baseline survey) (mean = 0.764 ± 0.193). Across the EQ-5D dimensions, more than half of the Thai RN reported some degree of pain or discomfort, and about one third reported anxiety or depression.

Table 7 Quality of life based on the EuroQol five-dimensions in the Thai Nurse Cohort Study, presented as number and percentage unless indicated otherwise
Full size table

Discussion

The TNCS was established to help address the Thai nursing workforce crisis, but may also contribute to the understanding of the health of Thai women engaged in a specific profession. To the best of our knowledge, this is the first cohort study with nurses in SEA. Our findings contribute evidence from a middle-income country, where the socioeconomic profile, environment, and health behaviors differ from those of high-income countries. Understanding long-term nursing workforce dynamics and changes in women’s health is essential for relevant policy formulation. The TNCS was also established at a time (2009) of significant change, with the integration of 10 ASEAN countries into a single economic community, the AEC. This means that the TNCS can monitor the corresponding changes in the Thai nursing workforce, including the impact of the free flow of the nurse workforce in the AEC on nursing mobility and shortages, and contribute to the introduction of timely workforce policy. The present paper was prepared during the preparations for the third TNCS survey round (2015), which will yield further information on the impact of the AEC.

In addition, the nine age-groups covered by the initial TNCS cohort will contribute to the understanding of job transition dynamics; while the regular replenishment of the cohort with younger age groups will contribute to a timely understanding of the work and life dynamics of the younger professional group, allowing new questions to be addressed. This design has many advantages over conventional single-age cohort studies, as highlighted in previous research [27].

The baseline survey response rate for the initial cohort (58.6 %) was similar to or greater than some other large studies conducted among the nursing profession [14, 18, 22]. However, we remain concerned about the lower response rate (38.3 %) for the new entry cohort, and more effort may be necessary to ensure a more favorable response rate in future.

A steering committee including key partners and stakeholders provided advisory support to the TNCS secretariat. These include 1) President of the Thailand Nursing and Midwifery Council (Chair), 2) Director, the International Health Policy Program (Vice Chair), 3) President, Thai Nurse Association, 4) President, Consortium of Dean of Nursing Schools, 5) Director, Bureau of Policy and Strategy, Ministry of Public Health, 6) Director, Nursing Systems Research Institute, Thailand Nursing and Midwifery Council, 7) Director, Prabomarajchanok Institute for Health Workforce Development, Ministry of Public Health, 8) Director, Health Systems Research Institute, 9) Director, HRH Research and Development Office, 10) Director, Bureau of Nursing, Ministry of Public Health, 11) Representative from the National Health Security Office, and 12) Representative from the National Health Commission. The Steering Committee is responsible for oversight and support of the work of TNCS, in line with the specified goals and objectives.

Maintaining a professional cohort with a high rate of follow-up responses and low costs was challenging. We designed the TNCS as part of the existing routine data collection system. First, updated postal addresses and other essential follow-up data including job transition and health status were fully embedded in a form that is completed by all RN during the 5-yearly process of professional license renewal, as legally required based on Continued Nursing Education credits. This form is mandatory and was uploaded to the TNC web-based registration system. The professional licensing renewal requirement enables the sustainability of the cohort, maximizes the follow-up rate, and ensures that information such as health status, job transition, and intention to leave the nursing career are regularly monitored. Second, mortality data generated by the National Civil Registration are linked with the TNC database; confidentiality on personal data is strictly observed throughout.

This report presented details of study design and administration of the TNCS. We also provided details of weighting for statistical methods to be used with the TNCS data. We strongly recommended that any analyses should implement weights we provided in the report, or at least, should explore how weighting affects their conclusions. In this report, we found that Thai nurse were aging. About half of them, 48.5 %, were older than 44 years. The average age was 43.7 years old. However, this is partly due to that older nurses participated in the TNCS more than the younger. We then sought to get the average age based on the TNC database- it was 37 years. This was slightly younger when compared to what was reported in a study in the United States which estimated that by 2008 the average age of RNs was 45.4 years, with more than 40 % of the RN workforce were older than 50 years [10].

We found that 15.4 % of Thai RNs intended to leave nursing career someday during their employments. However, the percentage of intention to leave within the next 2 years was 4.3 %. By this proportion, based on the current RN population in Thailand, it is approximately 20,000 RNs who intended to leave their careers in the next two years. This could be an alarmingly high turnover rate.

Conclusions

Health status of the current RNs were generally satisfactory. However, Thailand shares the same nursing crisis as many other countries, with a high proportion of aging nursing workforce, high rate of intention to leave nursing career, and low number of new entry nurses. Further studies are needed to investigate root causes of the crisis so that specific measures can be formulated to minimize the problem.

Availability of data and materials

All original data were archived in a secure database server (http://www.thainursecohort.org). This server archived all study material and served as the portal for cohort members to respond to the survey questionnaire, view the survey summary results in a real-time manner, and participate in an electronic forum for information exchange among cohort members. Cohort members and other Thai researchers are encouraged to use the data for research purposes at no cost. To access data, researchers are required to submit an analysis plan proposal (APP) (via http://tncs.damus.in.th). The Data Archival for Maximize Utilization (DAMUS) Committee is responsible for reviewing and approving research requests. Researchers with an approved APP are required to provide regular progress updates until publication of their work. Currently, all study materials are in the Thai language; international researchers will need to contact the DAMUS Committee to collaborate with the TNCS and access the data.