The retention of health human resources in primary healthcare centers in Lebanon: a national survey

The global shortage of health human resources

Being the crucial core of any health system, the presence of an adequate number of well trained and properly experienced HHR is central to the delivery of patient-centered healthcare services across the continuum of healthcare [5, 7]. HHR literature concurs in reporting critical shortages of health workers in many countries around the globe, particularly shortages in physicians and nurses [8–10]. Such shortages are confirmed within the context of the Eastern Mediterranean Region (EMR), particularly in low and middle income countries [11].

Most medical interventions require the services of doctors, nurses and other health professionals. Their performance is the first determinant of the quality, efficacy, efficiency, accessibility and viability of health services [2, 7].Studies that examined the hospital context established a link between HHR densities and health outcomes; including quality of care, morbidity and mortality [7, 12]. Furthermore, HHR account for a high proportion of budgets assigned to the health sector, and health expenditures are strongly linked to the ways in which HHR are deployed and used. Finally, the competences and availability of an appropriate workforce is a critical prerequisite for the growth and development of any organization [5]. Still, HHR has been largely a neglected component of the health-system development; in best cases not given the due importance that they deserve.

The retention of the primary healthcare workforce

Literature over the past two decades has examined the escalating PHC HHR crisis, elaborating on the centrality of recruitment and retention of PHC HHR for building balanced, effective and sustainable primary care services [9, 13–15]. Moreover, the role of HHR has been acknowledged to be even more important in community-based care, which tends to use less advanced equipment and is more dependent on competent personnel [5].

The most well documented cause of turnover among health professionals is related to job dissatisfaction [16, 17], to which various aspects are attributed. Literature presents a plethora of such factors that influence the retention of HHR; they could be summarized under the following three main categories: organizational characteristics, work characteristics, and individual characteristics. Organizational factors documented include salaries and benefits [18, 19], along with organizational commitment and managerial support [20, 21]. Work characteristics revolved around the nature of the job, which includes the workload, work environment, and work group cohesion [17, 20] as well as opportunities for professional development [18, 19, 22]. Individual socio-demographic characteristics associated with turnover included age, education, professional position, and tenure [18, 20, 23].

The primary healthcare context in Lebanon

The first call to the building of a Lebanese PHC system goes back to 1972 [24]. Since then, significant progress has been witnessed on both the public and private fronts as evident by the proliferation and expansion of PHC facilities across the Lebanese territories. On the public front, the Lebanese Ministry of Public Health led the establishment of a national PHC network that includes 140 centers [25]. The Ministry also led multiple initiatives to strengthen the PHC system, including: partnership for delivery with municipalities and private centers, the development of guidelines and health education materials, distribution of vaccines, drugs, supplies and equipment, as well as training activities [25].

Furthermore, the last few years witnessed the strengthening of privately owned and operated networks of PHC center. This astonishingly fast growth made private PHC centers the major provider of PHC services in Lebanon and has motivated the public sector to contract with a number of private centers for the provision of publicly funded PHC.

Although increased interest in the PHC sector in Lebanon has instigated the writing of a number of research reports over the last two decades [26, 27], none have systematically examined factors significantly associated with retention in this vital sector.

This study aims at investigating PHC health providers’ work characteristics, level of burnout and likelihood to quit and identify the factors that are significantly associated with staff retention at PHC centers in Lebanon. To our knowledge, this is the first study to address this subject in Lebanon and the EMR.

Research design

A cross-sectional design was utilized to survey all health providers in 81 PHC centers dispersed in all districts and governorates of Lebanon.

The setting and participants

Six major private networks providing primary healthcare services in Lebanon were identified and invited to participate in this study. Four agreed to participate; the other two did not for administrative and logistic purposes. The four participating networks have a large beneficiary base, running a total of 81 PHC centers dispersed nationally in all 26 districts of Lebanon. These networks are estimated to employ close to half of the primary healthcare providers working in the private sector in Lebanon.

These PHC centers offer health maintenance, promotion and prevention services, as well as a wide range of curative health services including access to generalists, specialists, diagnostic and dental services in various Lebanese areas with service coverage in both rural and urban settings. Although the health promotion and disease prevention services have been expanding over the last few years, it is noteworthy to mention that the focus of PHC centers services in Lebanon has been on delivery of curative and diagnostic services. Furthermore, active public participation remains deficient and indeed needs to be strengthened in the near future.

The target population within these centers was all health providers including: general practitioners, medical specialists (cardiologists, gynecologists, pediatricians, general surgeons and others), nurses, dentists, laboratory and radiation technologists, nutritionists and providers from others specialties (dental assistants, midwives, etc.). In each PHC center, all providers with a minimum of six months of cumulative work experience in the PHC setting were invited to fill the questionnaire. Non-providers, those involved in the management and administration of the PHC centers, were excluded.

Public PHC centers were excluded from this study for several reasons, mainly due to private PHC centers providing the majority of PHC services in Lebanon [26] and the minimal variability in recruitment and retention policies across public PHC centers. Indeed, there has been a freeze on HHR recruitment and very little turnover with respect to current staff at public PHC centers, whereby all staff are salaried governmental employees. In contrast, evidence shows variability across private PHC centers with respect to recruitment and retention policies.

Ethical approval

The study protocol, data collection instruments and consent forms were reviewed and received approval by the Social and Behavioral Sciences Institutional Review Board at the American University of Beirut. Written consent was obtained from all respondents, by ticking on the ‘I agree’ box on the cover page of the questionnaire.

The survey instrument

Data was collected through a questionnaire consisting of four sections: (1) socio-demographic and professional background, (2) organizational and institutional characteristics, (3) likelihood to quit, and (4) level of burnout using the Maslach Burnout Inventory (MBI). The questionnaire was reviewed and approved for content validity by a multi-disciplinary expert panel consisting of a health management and policy researcher, a clinician, nurses and a statistician. The questionnaire was translated to Arabic then back translated to English and compared to the original. No major differences were found. All language versions of the questionnaire were pilot-tested on 40 PHC providers (excluded from the study) to check for clarity of questions and to ensure that all aspects intended to be measured were covered. No major changes were made.

The dependent variable in this study, likelihood to quit, was measured by asking surveyed PHC providers to answer the following question “How likely are you to quit your current job in the next 1–3 years”. Responses were measured on a five-point Likert scale with the following options: Very unlikely, unlikely, not sure, likely, very likely. This measure of likelihood/intention to quit is well documented in literature [21, 28–31].

The maslach burnout inventory

The MBI is the most frequently used instrument to measure occupational burnout. It is a 22-item self-assessment tool that measures burnout syndrome across three subscales: emotional exhaustion-EE (9 questions), depersonalization-DP (5 questions) and personal accomplishment-PA (8 questions). The EE subscale measures workers’ feelings of emotional weariness from one’s work. The DP subscale measures workers’ impersonal manner in responding to recipients of services. The subscale of PA measures workers’ feelings of professional achievement and experience of intrinsic values at work.

The questions were answered according to a seven-point frequency scale ranging from “Never” (given a score of zero) to “Daily” (given a score of six). The scores for the questions related to each of the three subscales were added to compute a subscale score, which was then compared to the provided cut-off points for medical professions to establish the corresponding level of burnout [32].

Data collection

The principal investigators met with the general administration of the four PHC networks in order to explain the objectives of the study. Upon obtaining approval of participation, managers at the selected PHC centers were asked to distribute the questionnaires to all providers fitting the inclusion criteria. Completion of the questionnaire was voluntary. Anonymity of respondents was maintained, with no personal identifiers used and questionnaires returned in sealed envelopes. A total of 755 out of 1247 PHC providers completed the questionnaire (response rate 60.5%).

Data analysis

Upon completion of data collection, a random sample of 10% of the questionnaires was checked for completeness and accuracy to guarantee the quality and integrity of the collected data. Full data was then exported to and analyzed using the Statistical Package for the Social Sciences (SPSS), version 19.0. Sample characteristics were summarized using frequency and percentages. Likelihood to quit was the outcome variable of interest, and a multinomial logistic regression model was used to explain factors associated with the likelihood to quit. All analyses were conducted at a 0.05 significance level.

Sample description

Nurses were the largest professional group surveyed at PHC centers (32.7%), followed by generalists (23.0%) and medical specialists (21.7%). More than a third of surveyed providers (35.9%) had been working in the PHC setting for one to five years. Over half of the respondents (56.7%) were working full-time at the PHC centers and an additional third were working on a part-time basis. Two fifth of them (40.3%) reported less than 15 minutes one-way travel time to work.

Likelihood to quit

Surveyed providers gave a variety of reasons behind willingness to quit their job, the top five of which were: poor salary (54.4%), better job opportunities outside the country (35.1%), lack of professional development (33.7%), job instability (31.6%) and the lack of support from the administration (31.2%). Moreover, when asked about the reasons behind their choice to work in the current PHC center, the following reasons were the most commonly cited by respondents: the nature of the job, potential benefits to the family, proximity to household and area of work.

Also displayed in Table 2 are the findings related to the three subscales of burnout. Analysis reveals that experience of high burnout among providers is related to EE in almost a quarter of the respondents, DP in 13.8% of respondents, and PA in 18.7% of the respondents.

Bivariate analyses with likelihood to quit

Likelihood to quit differed by level of education (p-value=0.045) and number of years since graduation (p-value <0.001). The pattern of correlation was such that providers with more specialized education and more years of work experience were less likely to quit. The number of years working in PHC was also shown to have a significant association with likelihood to quit (p-value=0.003), with 10 years of experience being the cutoff point; providers who had less than 10 years were more likely to quit than those with more years of experience. Among demographic variables, analysis reveals a significant (p-value=0.001) U-shaped relationship between age and likelihood to quit, whereby the younger (18–25 years) and older (more than 55 years) were more likely to quit than other age groups. Both marital status (p-value=0.001) and having children (p-value<0.001) were correlated with likelihood to quit, with married providers and those with children being less likely to quit.

In terms of work position, likelihood to quit was highest among less specialized providers, including allied health professionals (technicians and assistants), as well as “other health professionals”, with proportions of 45.9% and 48.1%, respectively. Nurses expressed the lowest likelihood to quit their job (36.2%, p-value=0.025). Moreover, likelihood to quit was significantly associated with the number of years spent in the current PHC center of employment (p-value=0.003), with providers having worked ‘six months to one year’ being more likely to quit that those who worked for more years.

Sense of belonging (p-value = 0.023) showed to be a significant predictor of likelihood to quit among PHC providers. Moreover, one-way travel time from place of residence to the PHC center was also significantly associated with likelihood to quit (p-value = 0.015); the difference was especially noted at 15 minutes, whereby those who traveled less than 15 minutes were less likely to quit than the rest.

Multinomial logistic regression

Respondents with high level of burnout on the PA subscale had 3.05 times the odds of quitting (95% CI = 1.67-5.56; p-value <0.001) and 2.42 times the odds of being unsure whether to stay or quit (95% CI = 1.06-5.48; p-value 0.035) as compared to respondents with low level PA. With regards to education, respondents who had a university degree had 2.48 times the odds of quitting compared to those with a medical specialty (95% CI = 1.02-6.08; p-value 0.046).

Regarding the number of work hours per week, those who worked 1 to 10 hours had 3.15 times the odds of being unsure of their likelihood to quit as compared to those who worked more than 30 hours (95% CI = 1.22-8.15; p-value 0.018). A couple of findings did not reach the level of significance but are worth noting: Increases in the number of work hours per week was associated with reduced likelihood to quit and full-time employment status served as a protective factor in terms of likelihood to quit.

As for number of years working in PHC, providers who had worked ‘Six months to one year’ had 3.88 times the odds of quitting as compared to those who had worked more than 10 years (95% CI = 1.36-11.10; p-value 0.011). Moreover, those who had worked 6 to 10 years had 1.99 times the odds of being likely to quit as compared to those who had worked more than 10 years at PHC centers (95% CI = 1.03-3.84; p-value 0.041).