Time savings – realized and potential – and fair compensation for community health workers in Kenyan health facilities: a mixed-methods approach

There is an estimated deficit of 2.4 million doctors, nurses and midwives globally. The majority of this deficit lies in sub-Saharan Africa, which accounts for 24% of the global burden of disease yet has only 3% of the world’s healthcare workers (HCWs) [1]. Determining how to expand the healthcare workforce and enhance its quality is necessary for affected countries to achieve their health-related millennium development goals (MDGs) [1‐4]. Task shifting - the reassignment of clinical roles by transferring suitable tasks from higher- to lower-skilled HCWs - is one of the strategies proposed to mitigate the effects of the HCW shortage.

Community health workers (CHWs) are a unique cadre of HCWs and are well suited to help address the HCW gap. Many studies have demonstrated that CHWs can be as effective as more highly skilled HCWs in delivering a variety of services [5‐14] and help achieve MDGs [15]. But until recently, there have been few large-scale, sustainable CHW programs implemented in developing countries. Known barriers to successful implementation of CHW programs include inadequate compensation of CHWs, frequent unreimbursed out-of-pocket expenditures and variability in quality and duration of training and supervision [16‐20].

Kenya’s 2006 Community Strategy (CS) outlines primary health care delivery at the community unit (CU). CUs comprise volunteer CHWs who are linked to the primary health facility through trained Community Health Extension Workers (CHEWs) employed in primary care facilities; each CHEW is meant to supervise 25 CHWs [21]. CHWs are men and women recruited from the community, ‘having demonstrated attitudes valued by the community’ , and, preferably, are literate [21]. Main tasks performed by CHWs are categorized as: 1) disease prevention and control, 2) family health services, and 3) hygiene and environmental sanitation [22]. Further, the CS stipulates that volunteer CHWs should be paid a stipend ‘based on work actually done’ [21]. In a 2011 telegram, Kenya’s Ministry of Public Health and Sanitation stated ‘where funds are available community health workers shall be entitled to a payment of Ksh.2000 per month [approximately USD$ 25] as performance based incentive’ [23]. This statement lacks guidance on where funds should come from and on standardized criteria for performance standards. It remains unclear if and how many CHWs are compensated for their work.

Clinical officers and nurses are also integral to health care at the CU. Clinical officers undergo at least 3 years of university-level training and a year of internship, providing patient care and management at the primary health care level [24, 25]. Nurses undergo 1 to 4 years of training and, accordingly, achieve certificates, diplomas or degrees [26]. Kenya is facing a severe HCW shortage, and attrition among clinical officers is particularly high at the dispensary level [27, 28]. In 2010, the average monthly wage for clinical officers was USD$ 372, for nurses it was USD$ 248, and for public health officers it was USD$ 124 (J. Mwitari, personal communication, March 2010).

In practice, diverse and distinct community health models have emerged since 2010 when Kenya decentralized health care administration [29]. Faith-based and civil society organizations have played an important role in expanding the government’s efforts to establish CUs [30]. As Kenya works toward its One Million Community Health Workers Campaign to standardize CUs, it is important to detail tasks that volunteer CHWs perform. In addition, no studies have looked at the actual time savings realized by task shifting from higher HCW cadres to CHWs [18]. This information is needed to inform implementation of task shifting and help developing country governments, donors and non-governmental organizations (NGOs) better understand the value of CHWs. This mixed-methods study in Kenya was conducted to fill this knowledge gap.

Eight locations incorporating 17 health facilities (7 hospitals, 6 health centers and 4 dispensaries) were visited in Nairobi, Nyanza, Central, Eastern and Western Provinces. Ten sites were in sparsely populated areas, three in mid-populated regions and four in densely populated areas.

Forty-seven interviews were conducted with 28 CHWs and 19 clinicians. No potential participant declined to be interviewed. All 28 CHWs interviewed agreed to complete the task diaries; 20 diaries were returned; and 3 were not included in the analysis due to incorrect recording of time spent performing tasks. Overall, the CHWs spent 60% of their time in facilities and 40% of their time in the community. Tasks performed in the community include defaulter tracing and linking community members to the health facility. Counseling and health education occurred both in the community and facility; all other tasks occurred in the facility.

Multiple studies and a recent meta-analysis support that lay HCWs can be as effective as higher skilled HCWs at providing a wide range of services [8‐11, 34]. In Kenya, CHWs providing care for antiretroviral drug delivery had similar outcomes of HIV control compared to standard clinic visits [12]. Our study’s findings are consistent with these studies in that semistructured interviews and CHW task diaries demonstrated that facility-based CHWs are delivering a significant amount and variety of services, including tasks that have been directly-shifted from clinicians as well as those that are added-value to Kenya’s health system. Tasks performed and the amount of task shifting varies among CHWs and as perceived by clinicians. It is important to note our study did not objectively measure quality of care provided by CHWs, however. In addition, we were unable to consider additional important aspects of the CHW workforce, including attrition, training and supervision.

A study in Rwanda on task shifting from physicians to nurses showed a 78% reduction in HIV-related physician workload as a result of implementing nurse-initiated antiretroviral therapy, saving up to 56 hours of physician time per month [35]. Our interviews revealed that task shifting to CHWs frees up 2.5 hours/day of clinicians’ time on average. This is the first published data that we know of to quantify the time savings realized through CHW task shifting. In addition, our modeling exercise suggests transitioning other selected tasks from clinicians to CHWs could yield further significant time savings for clinicians while maintaining clinic productivity. These findings are consistent with studies showing cost-effectiveness of CHW-based care [5‐7, 13, 36, 37].

Prior research has identified the lack of consistent and appropriate remuneration of CHWs as a significant barrier to the success of CHW programs. Paying CHWs appropriately for their work and providing clear delineation of CHW tasks has been shown to enhance CHW retention and improve sustainability of CHW programs. [2, 7, 18, 38‐40] While a recent study enumerated health system-level costs associated with deploying CHWs using an estimate of current CHW wages [41], we attempted to quantify the monetary value of facility-based CHWs’ work to provide concrete starting data points for the discussion of fair CHW compensation.

CHWs provided higher estimates of a fair monthly salary than clinicians (US$ 178 and US$ 128, respectively). Using the data from the CHW task diaries, we derived an average monthly valuation of current CHW services of US$ 117, consistent with the clinicians’ estimated CHW wage. While this calculation includes added value of CHWs it is likely an underestimate because we were not able to quantify important linkages that CHWs provide to health facilities.

Our modeling exercise suggests that further HCW skill optimization through additional task shifting and the expansion of CHW roles could justify employing 2.5 CHWs at an average-sized health facility at current workloads. Our clinic model is limited by the assumptions made regarding the average completion time for tasks; the proportion of patients in need of each service; the cadre that currently performs a task and which cadre would ideally do it; and the extra time required to complete a task after it has been shifted to a CHW. These assumptions affect the time required of each HCW to provide their services to an individual patient and the total daily cadre times. The sensitivity analysis showed significant variability in the projected clinician time savings and additional CHW time. This highlights the need for future research on these variables in resource-limited clinics in order to better inform models.

Further, in Kenya, CHEWs are facility-based HCWs, but are meant to take on mostly a supervisory role of CHWs. As additional models of CHEWs and CHWs are developed, it may be necessary to create complementary cadres of CHWs: those that are predominately facility-based and take on mostly task-shifted roles in the facility, and those that are community-based and engage mostly in added-value services that CHWs provide in the community.

There are several limitations to our estimates of the value of CHWs’ work. Study subjects may have been more strongly supportive of CHWs given that all who were approached agreed to participate; this selection bias may have influenced results. Misclassification of tasks may also have occurred as they were categorized as directly-shifted or added-value based on semistructured interviews. The valuation of directly-shifted CHW tasks was affected by the uniform time adjustment made for the assumption that clinicians are more efficient than CHWs; there is, however, likely significant variability in the relative efficiency of CHWs and clinicians. The research team chose the value of this discount factor based on input from clinicians and CHWs and their own experiences, not objective data due to an absence of relevant published data. The valuation of added-value tasks was based on a wage figure derived from our interviews with CHWs and clinicians. Because of this linkage, our valuation is less valid than if we had used an estimated CHW salary obtained by other means. CHWs had an incentive to provide a higher estimated CHW wage if they felt they stood to benefit from any future CHW compensation program, although CHWs may also have a better sense of what might be a fair living wage. Lastly, our estimate of clinician time liberated is limited since we did not simultaneously record clinicians’ daily tasks. Nonetheless, we believe our estimates provide useful reference points for policy debates and future research.

In summary, CHWs are an important component of health service delivery in Kenya. They reduce the workload of clinicians, giving them more time to spend with patients, see additional patients and do higher-level job functions. They also provide services that would often not be provided otherwise and provide vital links between communities and health facilities. Our study attempts to put a value on these services to aid those involved with planning, financing, and implementing CHW programs in resource-limited settings. While compensation is just one component of the CHW programs that exist within complex parent health systems, it is important to appropriately address it upfront as these programs are designed, implemented, and scaled up.

CHWs are an important component of healthcare delivery in Kenya. The estimates herein provide starting points for fair compensation of CHWs, so they may be further integrated into health systems faced with a health worker shortage. Optimal utilization of CHWs will reduce the workload of strained health workers and improve access to health care.