An Empirical Assessment of Health Workers ’ Organisational Commitment in Ghana : A Comparative Analysis

Organizational commitment has been empirically found to be poor among health workers in Ghana. In this paper, we tried to compare health workers’ organizational commitment in the private and public sectors to understand if organizational commitment remains the same across the two sectors. We collected and used data on healthcare institutions in the two sectors using self-administered questionnaires. Factor analysis was used to analyse data. Results suggest that workers in private healthcare institutions have a favourable organizational commitment, whereas their counterparts in the public sector have poor organizational commitment. The validity of our analysis is based on meeting the Measure of Sampling Adequacy criterion at the levels of individual indicator variables and all variables put together with respect to both the private and public sectors. The Bartlett’s test of sphericity was also significant at the chosen level of significance for both sectors [(Private: Chi-square = 173.68; p = .000); (Public: Chi-square = 235.44; p = .000)]. We therefore provisionally accept the hypothesis that organizational commitment is not the same among private and public health workers. The study recommends that managements of the public health institutions must increase attention on the organizational commitment of health workers by enhancing compensation, ensuring fairness in organizational justice and offer satisfactory job designs.


Introduction
Healthcare is a basic need of every society, a reason for which countries, organizations, and individuals are constantly making efforts to improve its quality.Moreover, though organizations such as World Health Organization (WHO) continue to expend efforts towards delivery of quality healthcare, healthcare quality is still a problem in many jurisdictions.The issue of quality healthcare is even worse in developing African countries such as Ghana, where healthcare quality is confirmed to be low and below the required standard (Bonenberger et al., 2014).A remedy to poor healthcare quality could be increasing health workers organizational commitment.
A good number of researchers (Tiwari & Mishra, 2008;Akanbi & Itiola, 2013;Bonenberger et al., 2014) have confirmed that quality of healthcare is positively influenced by the motivation given to health workers.Moreover, the positive effect of motivation on healthcare quality could be confounded by job satisfaction.More importantly, healthcare quality has been empirically confirmed to be positively related to health workers' organisational commitment (Tiwari & Mishra, 2008).Consequently, some authors such as Al-Aameri (2000) and Altindis (2011) have argued that poor healthcare is largely as a result of health workers' low organisational commitment.
Organisational Commitment (OC) is the extent to which workers identify with their organisation and are committed to contributing to achieving organisational goals and objectives (Akanbi & Itiola, 2013).OC could also be defined as the degree to which a person recognises himself as an employee of an organisation, and how much he is dedicated to meeting his job roles.On the basis of this definition, employees with high OC highly identify themselves with their firms and strive to make the best possible contribution to desired corporate performance and growth.OC is therefore the basis of employee enthusiasm towards meeting or exceeding their performance quotas.OC, by definition, is the basis of the amount of effort health workers contribute to quality healthcare.Invariably some researchers (e.g.Al-Aameri, 2000;Khan & Jan, 2015) observed that healthcare quality is driven by the level of health workers' OC.Efforts to improve the quality of healthcare must therefore be based on a good understanding of health workers' OC.This paper therefore assesses the OC of health workers in selected health institutions in Ghana.
Management's uncertainty about health workers' OC in three health institutions in the Secondi-Takoradi metropolis of Ghana is increasingly high.Moreover this situation is coupled with an observed poor healthcare quality in these institutions, which is believed to be severer relative to the national situation.It is hoped that this paper will contribute to managements' knowledge on what constitutes health workers' OC in these institutions, which are Effia Nkwanta Regional Hospital, UQ Hospital and Sycamore Medical Centre.To enhance the importance of results and the study's contribution to knowledge, we are interested in comparing OC of health workers in private and public health institutions.Findings of this paper are also expected to influence the course of our future research on the impact of health workers' OC on healthcare quality, employee performance and job satisfaction.

Literature Review
Organisational commitment reflects attitudes towards the entire organisation and is typically seen as broader than job satisfaction and is more consistent than job satisfaction over time (Bonenberger et al., 2014).Hence satisfied employees do not necessarily have organisational commitment, but every employee who has organizational commitment is more likely to be satisfied.Organizational commitment is important because a committed employee contributes to the success of the organisation and is less likely to quit that organization.OC has been categorised as affective, normative and continuance commitment (Meyer & Allen, 1991).
Affective commitment is an emotional attachment to the organization that an employee works for (Meyer & Allen, 1991).They added that normative commitment is a sense or feeling of moral obligation to stay with the organization.Continuance commitment is characterized by a more rational analysis of the costs of staying vis-à-vis leaving the organization (Meyer & Allen, 1991;Adekola, 2012).Any of these categories is driven by factors such as how comfortable the employee is at work, opportunity for career development at the workplace and the organization having the welfare of the employee at heart, and fairness of an organisational system (Meyer & Maltin, 2010).
The fairness of any organizational system that drives organisational commitment is related to two main components.The first of these components is Distributive Justice (DJ), which is an individual's perception about their reward in relation to their contributed effort and comparison with others' efforts (Meyer & Allen, 1991;Meyer & Maltin, 2010).The influence of this component is extended to employee compensation and consequently job satisfaction and performance.The other component is Procedural Justice (PJ), which is an individual's perception about the fairness of the procedures used to make decisions about rewards (Meyer & Allen, 1991).
From the viewpoint of the Job Characteristics Theory (JCT) developed by Hackman and Oldham (1976), job design attributes are aspects of the individual employee's job tasks that shape how the individual perceives his or her particular role in the organisation.According to Hackman and Oldham (1976), five core job characteristics including Skill variety, task identity, task significance, autonomy and feedback prompts certain psychological states which in turn results some essential personal and organisation outcomes.Thus job characteristics, which are essential tenets of job design influences whether or not an employee would be motivated and committed to his or her job and for that matter the organisation.For instance, a new employee who is offered a satisfactory pay but has little time for recreation and for family members is likely to lack organisational commitment.Another new employee who has satisfactory pay and is allowed sufficient time to recreate and expend time with relations is more likely to be committed to the organisation.
Organisational commitment of employees is also underpinned by the Social Exchange Theory (SET), which was introduced by George Homans in 1958.Homans defined social exchange as the exchange of activity, which could be tangible or intangible, and more or less rewarding or costly, between at least two persons.His SET similarly posits that human relationships are formed by the use of a subjective cost-benefit analysis and the comparison of alternatives.In the context of the foregoing discussion, SET implies that employees exchange their competencies (intangibles) with a mixed of tangible and intangible benefits (e.g.recognition, salaries, allowances, opportunity to develop one's career, etc.).Depending on how this exchange meets or exceeds employees' expectations, they become committed to their organisation (Al-Aameri, 2000;Kan & Jan, 2015).
It is therefore evident that organisational commitment and its relationship with such variables as job satisfaction and job performance are recognised by the JCT, SET and possibly other theories.On the basis of this theoretical recognition, the dimensions of the construct of Organisational Commitment have been explored by several researchers (e.g.Al-Aameri, 2011;Akanbi & Itiola, 2011;Akanbi & Itiola, 2013), purposely to understand this construct with respect to various groups of employees (e.g.bankers, health workers, construction workers, etc.) and in connection with other variables such as job satisfaction and job performance.However current contributions by researchers in this regard do not address the information needs of some organisations, decision makers and researchers.
The organisational commitment of three health institutions in Ghana (i.e.Effia Nkwanta Regional Hospital, UQ Hospital and Sycamore Medical Centre) is suspected to be poor, with uncertainty about whether or not health workers in these institutions have the expected organisational commitment.The seriousness of this situation is worsened by deteriorating health service quality in these institutions.While their managements wound want to understand how the organisational commitment of health workers in these institutions relates to some variables such as job performance and job satisfaction, their understanding of the construct of OC, based on the assertion of Obi-Nwosu et al. ( 2013) is a prerequisite.
Obi-Nwosu et al. ( 2013) has stressed that it is important to assess constructs such as OC in sufficient depth before making an attempt to empirically examine their relationship to other constructs, making it possible to know their dimensions and items.If, for instance, OC is understood in this regard, the constituent items of employees' organisational commitment would be known, leading to ample knowledge about whether interventions are needed to improve or maximise this construct in practice.We also believe that this construct-specific assessment is important in situations where managements or decision makers lack knowledge on the nature of the construct in real life situations, as the case is in the selected health institutions in Ghana.In this paper therefore, we attempted to assess health workers' organisational commitment in the three health institutions to equip their managements, researchers and decision makers with empirical evidences on health workers' organisational commitment, which is critical to quality healthcare delivery.As mentioned earlier, this assessment is relevant to further evaluations of the impact of OC on job satisfaction and job performance.
Whereas researchers have used less robust statistical tools such as descriptive statistics in assessing health workers' organisational commitment outside the three chosen health institutions, we contribute to the literature and knowledge by using a supposedly robust statistical tool, precisely Factor Analysis (FA) to assess OC in the three health institutions.The research methods employed are discussed in the next section.

Methodology
The study was conducted in the three selected healthcare institutions in Sekondi-Takoradi area in the Western Region of Ghana.Effia Nkwanta Regional Hospital represented the public health institution, and UQ Hospital and Sycamore Medical Centre represented the private health institutions.The population of this study was all permanent workers in the health institutions who had worked for at least 2 years.This population spanned all departments, which are Paediatrics, Internal Medicine, Obstetrics and Gynaecology, Surgery, Health Information, Information Communication Technology, and Health Administration and Support Services.This population was made up of 629 health workers for Effia Nkwanta Regional Hospital; 40 for UQ Hospital; and 49 for Sycamore Medical Centre.
Considering the fact that the population sizes of the private health institutions were small, we used every member in them.Thus data was collected on all 89 health workers of the private institutions.To ensure that an equal number of respondents participated in the private and public hospitals for the purpose of sectorial comparison, we decided to sample 89 health workers from Effia Nkwanta Regional Hospital using simple random sampling by stratification.The stratification was done with respect to the departmental affiliation of each worker.We used this method of sampling to make the sample representative and random.
Many previous studies (Obi-Nwosu et al., 2013;Bonenberger et al., 2014) have empirically shown that health workers' organisational commitment in developing African countries is poor for reasons such as poor job design, occupational stress and poor remuneration.In the context of FA, this evidence means that variables ultimately making up health workers organisational commitment are negative statements (items) that reveal health workers' refusal to identify themselves with their institutions.In the context of the three health institutions selected, we therefore hypothesise that: Health workers' organisational commitment is ultimately made up of negative statements or items revealing health workers' refusal to identify themselves with the institutions.
We put forth this hypothesis in view of existing empirical evidences (Obi-Nwosu et al., 2013;Bonenberger et al., 2014) indicating poor job design, a high level of occupational stress among workers and unsatisfactory remuneration in Ghanaian healthcare institutions.However private health institutions are better organised administratively and offer better job designs, compensation, and motivational schemes that enable employees to control their stress level (Bonenberger et al., 2014).We therefore further hypothesise that: Workers' organisational commitment in the public hospital (Effia Nkwanta Regional Hospital) is ultimately made up of a higher number of negative statements relative to that of health workers in the private healthcare institutions (UQ Hospital and Sycamore Medical Centre).
To control for response errors and ensure that this hypothesis is validly tested, we measured OC by pairing each positive statement of the OC scale used with a negative one as recommended by Creswell (2003).We collected and used primary data using a cross-sectional research method alongside a quantitative research approach.Data was collected using a self-administered questionnaire.This type of instrument was used because, as argued by Creswell (2003), it is self-explanatory and is therefore easily completed by respondents to avoid response errors.Creswell (2003) also observed that the use of this type of instrument is more suitable for collecting quantitative research data.The questionnaire primarily contained standard but adjusted items of OC shown in Table 1.These items were borrowed from Altindis' (2011) and measured using a Likert scale, which allowed the participants to respond on a scale of 1 to 5 indicating their extent of agreement or disagreement to each OC item.The scale of the Likert scale is represented by: 1 = strongly disagree; 2 = disagree; 3 = not sure; 4 = agree; and 5 = strongly agree.In coding however, not sure was corresponded to 0, since it represents neutrality and uncertainty.
We made sure the questionnaire was valid and reliable.To ensure that the instrument used was valid, we submitted it to experts who reviewed it and made practical suggestions for improving its validity.The reliability of the instrument was verified by computing the Cronbach's alpha in SPSS Version 21.The computation showed that OC has a Cronbach's coefficient of 0.901.Morse (2002) has indicated that the lower cut-off value for a reliable questionnaire is 0.700.Since each Cronbach's alpha calculated is greater than this value, we deemed the questionnaire reliable.
Before data was collected, the Business School of Kwame Nkrumah University of Science and Technology (KNUST) approved this study.Managements of the health institutions were also served with formal permission letters endorsed by the KNUST Business School.Individual respondents were given informed consent forms to endorse their participation.The consent form indicated the purpose of the study, its timelines, benefits to respondents, risks, confidentiality of information provided, the fact that a respondent can withdraw from the study any time before the study's results are published, among others.After participants had signed their informed consent form, questionnaires were administered by hand delivery.A 96% response rate was achieved in data collection.Thus 85 questionnaires were correctly completed and analyzed.
We analyzed data using SPSS Version 21.Factor Analysis (Principal Axis Factoring, PAF) was used to present results after data was checked for outliers using descriptive statistics (i.e.means, skewness, and kurtosis).Some researchers (Suhr, 1999;Ringner, 2008) observed that FA is traditionally a data reduction statistical tool used for several purposes such as:  Testing dimensionality of constructs;  Identifying the main variables that make up a construct;  Eliminating indicator variables which are not strongly or significantly part of a construct leading to a retention of the ultimate construct; and  Assessing the overall strength of a construct on the basis of the variation explained by its ultimate indicator variables.
In this paper, we assess the OC construct on the basis of these four purposes of using FA.We therefore expect that health workers' ultimate organisational commitment would be variables retained after all theoretical iterations of the FA.Iteration is a session of FA in which variables which are weakly related to the ultimate variables of the construct are removed or extracted.The final iteration is the session of FA in which the final set of variables weakly related to the ultimate variables is removed.After the last iteration, all variables are sufficiently related based on the theoretical benchmark of each variable having an extraction value of at least 0.5 (Suhr, 1999).
FA-PAF was also used to double-check data for outliers.Other important assumptions were also tested via the use of FA-PAF.In data analysis, FA-PAF was used in an attempt to reduce the dimension of OC through as much iteration as possible.This was done for both public and private health institutions.In the analysis, the ultimate OC of health workers in each sector is the list of items remaining after several iterative FA-PAFs.Data are analyzed and presented as follows.

Analysis
In this analysis, we try to identify a difference in the organisational commitment of health workers of the chosen private and public hospitals.We approach this analysis by using FA to screen items of an OC scale using as many iterative FA sessions as possible.For each hospital, we deemed items of the final iteration as the ultimate measures of health workers' organisational commitment.The first iteration of the FA starts with Table 2. Table 2 shows the correlation matrix of all items of the OC scale employed, with the matrices of private and public hospitals clearly shown.In this table, many of the correlation coefficients are greater than 0.3, as theoretically required (Ringner, 2008).This situation suggests that one or more indicator variables would be retained in the first iteration of the FA for each hospital.The FA is also theoretically required to satisfy the MSA (i.e.Measures of Sampling Adequacy) criterion at two levels.At the first level, each indicator variable must be associated with an anti-image correlation of at least 0.5 (Ringner, 2008), for both public and private hospitals.These correlations are shown in Table 3 (in bold case).Evidently, MSA criterion is satisfied at the first level for the private and public hospitals.The MSA criterion at the second level requires all indicator variables to produce an MSA value of at least 0.5 (Suhr, 1999).The MSA value, which satisfies the criterion at the second level, is shown in Table 4. Also in Table 4, the Kaiser-Meyer-Olkin (KMO) test is significant at 5% significance level [(Private: Chi-square = 173.68;p = .000); (Public: Chi-square = 235.44;p = .000)].As a result of satisfying this criterion at the two levels, the FA is validated.(Suhr, 1999;Ringner, 2008).On the basis of this rule of thumb, we removed all variables with communalities less than 0.5 (i.e.those in bold case) for private and public hospitals.Therefore for the private and public hospital, four out of ten variables are retained, though the set of variables retained are different.For the private hospital, OC 2 , OC 3 , OC 6 , and OC 10 are retained, whereas for the public hospital G 3 , G 4 , G 5 and G 8 are retained.
Please see Table 1 for variables represented by OC 1 , OC 2 … OC 10 and G 1 , G 2 … G 10 .The only retained variable that is common to the private and public hospitals is OC 3 /G 3 .The second iteration starts with Table 6.Source: Field survey, 2015.
In Table 6, one out of four variables is retained in the second iteration for the private hospital.The remaining variable is OC 2 , which represents "I really feel as if this organisation's problems are my own".The FA then ended in two iterations for the private hospital since it does not support a single variable.Therefore OC 2 is the ultimate measure of health workers organizational commitment in the private hospital.In Table 8, this variable explains the largest amount of the variation of 66.7% of the total variation.With respect to the public hospital, three variables are retained, whilst OC 6 is removed.
Table 6 shows the communalities of the variables taken into the third iteration with respect to the public hospital.It can be seen that all three variables are retained in the third iteration.These variables read: G3: I do not feel a strong sense of belonging to my organisation G4: I do not feel emotionally attached to this organisation G8: I do not feel any obligation to remain with my current employer.
These three variables retained in the FA are the ultimate measures of health workers' organisational commitment in the public hospital.In Table 8, the three variables account for 76% of the total variation.The study found that the three ultimate measures of health workers' organisational commitment are negative measures, which reveal poor organisational commitment among health workers of the public hospital.On the other hand, though a single variable ultimately measures organisational commitment in the private hospital, this variable is positive and expresses favourable organisational commitment.The hypothesis of the study is therefore supported.

Discussion of Results
Out of 10 items of the OC scale analysed, three items are retained after three iterations for the public hospital.The items retained (I do not feel a strong sense of belonging to my organisation; I do not feel emotionally attached to this organisation; and I do not feel any obligation to remain with my current employer) in this regard are negative.For the private health institutions, a single positive item is retained (I really feel as if this organisation's problems are my own).Some writers (Suhr, 1999;Ringner, 2008) have observed that variables retained in a Factor Analysis ultimately reflect the construct being measured.On the basis of this argument, we are of the view that health workers have poor and unfavourable organisational commitment in the public hospital.
Health workers in the private health institutions however have a more favourable organisational commitment.This study's evidence of poor and unfavourable organisational commitment among health workers in the public hospital is backed by many researchers.In a foreign country context, some researchers (Tiwari & Mishra, 2008;Akanbi & Itiola, 2013) found that organisational commitment among health workers in developing countries is poor.Locally in Ghana Bonenberger et al. (2014), have revealed that compensation, fairness of organisational system, nature of organisational environment and other attributes leave much to be desired in health institutions in Ghana and since these variables have recorded significant impact on organisational commitment, its stands verified that the unfavourable organisational commitment among health workers in the public hospital as found in this study is theoretically backed.
On the contrary, this not the case for workers in Private health institutions in Ghana.Whiles Bonenberger et al. (2014) found that organisational commitment among health workers in Ghana is poor and unfavourable, this study's result for private health institutions proves otherwise.This is likely to be as a result of private health institutions having satisfactory compensation, fairness of organisational system and organisational environment.
It is however admitted that the study found little and insufficient empirical support for our result, a reason for which more research work is needed on this topic.

Conclusions and Implications
What constitutes health workers' organisation commitment is ultimately unfavourable for the public health institution, as all three items retained in the FA are variables indicating workers' unwillingness to identify themselves with the health institution.We therefore confirm the hypothesis that health workers' organisational commitment is ultimately made up of negative statements or items revealing health workers' refusal to identify themselves with the institution.On the contrary, a positive item was retained in the FA for the private health institutions, confirming that organisational commitment among health workers in the private health institutions is more favourable or better relative to the public institution.
On the basis of findings and our conclusion, there is the need for managements of the health institutions to increase attention on the organisational commitment of health workers in the public hospital.Steps must be taken to incentivise health workers to up their commitment to their respective institutions.This can be done by enhancing compensation, ensuring fairness of the organisational system, offer satisfactory job designs, and enforcing organisational justice.
This study hopefully provides insights into the organisational commitment of health workers in the selected institutions.It therefore reveals the urgency of taking actions to improve organisational commitment to enhance quality of healthcare and organisational performance.This study also contributes to academic debate on organisational commitment of health workers in a developing country context.
The limitation of this study is defined by a non-representative sample of health workers from one of the health institutions (Effia Nkwanta Regional Hospital).This limitation was encountered as a result of the need to incorporate equal numbers of health workers from private and public institutions.This situation has minimised the external validity of the study.

Table 1 .
Variables and their symbols

Table 5
shows communalities (i.e.extraction values) of the FA.These values indicate which indicator variables would have to be removed in the iteration.The general rule of thumb is to remove variables with values less than 0.5

Table 8 .
Variance explained in final iteration