Absenteeism among family planning providers: a mixed-methods study in western Kenya

Abstract

Public-sector healthcare providers are on the frontline of family planning service delivery in low- and middle-income countries like Kenya, yet research suggests public-sector providers are frequently absent. The current prevalence of absenteeism in Western Kenya, as well as the impact on family planning clients, is unknown. The objective of this paper is to quantify the prevalence of public-sector healthcare provider absenteeism in this region of Kenya, to describe the potential impact on family planning uptake and to source locally grounded solutions to provider absenteeism. We used multiple data collection methods including unannounced visits to a random sample of 60 public-sector healthcare facilities in Western Kenya, focus group discussions with current and former family planning users, key informant interviews (KIIs) with senior staff from healthcare facilities and both governmental and non-governmental organizations and journey mapping activities with current family planning providers and clients. We found healthcare providers were absent in nearly 60% of unannounced visits and, among those present, 19% were not working at the time of the visit. In 20% of unannounced visits, the facility had no providers present. Provider absenteeism took many forms including providers arriving late to work, taking an extended lunch break, not returning from lunch or being absent for the entire day. While 56% of provider absences resulted from sanctioned activities such as planned vacation, sick leave or off-site work responsibilities, nearly half of the absences were unsanctioned, meaning providers were reportedly running personal errands, intending to arrive later or no one at the facility could explain the absence. Key informants and focus group participants reported high provider absence is a substantial barrier to contraceptive use, but solutions for resolving this problem remain elusive. Identification and rigorous evaluation of interventions designed to redress provider absenteeism are needed.

Key messages.

• Provider absence exacerbates the acute shortage of healthcare workers in low- and middle-income countries (LMIC) health systems, yet the impact of provider absenteeism on family planning use has not been previously investigated.

• Nearly 60% of family planning providers sampled in our Western Kenya study were absent. Providers were often absent because they arrived late to work, did not return to the facility after their lunch break or because they attended personal errands during work hours.

• High provider absence is a substantial barrier to contraceptive use in Western Kenya.

• Future research is urgently needed to identify and rigorously evaluate public health interventions with the potential to reduce provider absenteeism.

Introduction

Many low- and middle-income countries (LMICs) face an acute shortage of healthcare providers, jeopardizing the ability of health systems to provide quality care, achieve Universal Health Coverage (UHC) and meet the Sustainable Development Goals. Adjusted for population, healthcare provider shortages are most severe in the African region (World Health Organization, 2016). Provider attrition, ageing of the workforce, brain drain and population growth have all been identified as contributing factors (Chankova et al., 2009; World Health Organization, 2016). Yet researchers are also increasingly finding that provider absence greatly exacerbates the acute shortage of healthcare workers in LMIC health systems (Chaudhury et al., 2006; Goldstein et al., 2013; Tumlinson et al., 2013; Ackers et al., 2016).

In a landmark study of public-sector absenteeism in six countries, Chaudhury and colleagues found that, on average, healthcare providers were absent for 35% of unannounced visits, with proportions ranging from 25% to 40% across different country contexts (Chaudhury et al., 2006). For primary care facilities, which are often staffed with a single provider, such high levels of absenteeism can mean that no provider is available to provide services. Such high levels of absenteeism represent an inefficient use of government expenditures on health and undermine national efforts to promote UHC (Chaudhury et al., 2006; World Health Organization, 2016).

Further, provider absences have a pernicious effect on health outcomes. According to a 2013 study conducted in Kenya, nurse absence during a woman’s first antenatal care visit was associated with a 25-percentage-point reduction in the probability of facility delivery and also reduced the probability of breastfeeding, receiving antiretroviral medication and enrolling in a free treatment programme among women who self-reported expecting to test HIV-positive (Goldstein et al., 2013). In 2016, Ackers and colleagues found provider absenteeism to be the single largest factor contributing to delays in the provision of care in Ugandan healthcare facilities which resulted in adverse birth outcomes, including stillbirth and maternal death (Ackers et al., 2016). Provider absenteeism has also been reported to be associated with mistreatment in maternal care (Mannava et al., 2015; Warren et al., 2017).

A key knowledge gap in the study of this behaviour is the analysis of the impact of absenteeism on family planning use, which plays a critical role in reducing maternal, child and infant mortality in LMICs (Cleland et al., 2012). In Kenya, the setting for this study, approximately 43% of all women of reproductive age are using a modern contraceptive method and 60% of these women obtain their method from public-sector healthcare facilities (KDHS, 2015; FP2020, 2020). Kenya has committed to increasing contraceptive use among vulnerable populations and to addressing barriers to family planning, particularly in remote areas (FP2020, 2020). Provider absenteeism may thwart such efforts, yet little research is available to describe the current extent and impact on contraceptive use of public-sector provider absenteeism in Kenya. A 2008 study from southeastern Kenya estimated an absenteeism rate of 19% among public-sector nurses (Muthama et al., 2008) and a 2019 qualitative study conducted in Kisumu and Nairobi indicates provider absenteeism may function as an obstacle to family planning use (Tumlinson et al., 2019). However, a sizeable knowledge gap still surrounds the nature and extent of provider absenteeism in Kenya and the potential impact on contraceptive uptake and continuation.

Further, an exploratory study of facility-level barriers to family planning in Kenya using simulated clients suggests that the nature of provider shortages may be more complex than the term ‘absenteeism’ suggests (Tumlinson et al., 2013). In particular, women’s inability to access providers may be due to long queues, competing priorities (e.g. patients with other clinical needs) or providers’ failure to attend to patients during posted hours (due to lateness, failure to return after lunch or leaving early). However, more research is needed to understand the extent to which these factors may impact family planning use.

Thus, the purpose of this study is to quantify the prevalence of absenteeism in a representative sample of public-sector facilities in Western Kenya, better understand provider absenteeism in terms of its nature and the effect on contraceptive access, and to surface locally generated solutions to the problem. We also explore provider and facility-level characteristics that may influence absenteeism, under the hypothesis that providers stationed in unfamiliar locations, those who live far from their workplace or those who report high levels of professional dissatisfaction or burnout may be more likely to regularly miss work. To our knowledge, this is the first study of this kind to be conducted in Western Kenya.

Materials and methods

We investigated the issue of provider absenteeism as part of a larger project designed to identify facility-level barriers to family planning in Western Kenya using multiple data collection methods: unannounced visitors (UAVs, focus group discussions (FGDs), KIIs and journey mapping activities. In this analysis, we present a descriptive analysis on the scope and effect of absenteeism on contraceptive access. Across data collection and analysis, we employ a concurrent mixed-methods approach to provide a holistic and rich description of provider absenteeism in Western Kenya (Creswell and Clark 2006). All data were collected between October 2018 and March 2019.

UAVs, facility audit and provider interview

Unannounced visits, a brief facility audit and a short provider interview were conducted within a random sample of 60 public-sector facilities located in 5 of 10 counties comprising Western Kenya. To ensure inclusion of all facility types across each of the five counties, prior to random selection, all public facilities were stratified by county and facility type, which included three categories: Level 2 facilities (i.e. smaller facilities often characterized as dispensaries or clinics), Level 3 facilities (i.e. medium-sized facilities often characterized as health centres) and Level 4 and 5 facilities (i.e. larger facilities, often characterized as sub-county or county hospitals). Facilities were selected at random from the Kenya Master Health Facility List, a publicly available and annually updated list of all public-sector facilities in Kenya. We randomly selected six Level 2 facilities (i.e. dispensaries), three Level 3 facilities (i.e. health centres) and three Level 4 or 5 facilities (i.e. medium to large hospitals) in each of the five counties. Facility managers for all selected facilities provided written informed consent for participation in the unannounced visit portion of this study and 100% of selected facilities agreed to participate.

An ‘unannounced visitor’ is an enumerator who arrives at a healthcare facility without any advance notice to the facility manager or staff. Such visits are a key component of measuring provider absenteeism, as alerting a facility manager in advance of the visit may alter the attendance of facility staff. The UAV arrives at the facility at a specific time and physically verifies the presence or absence of each provider on a pre-obtained provider roster. Providers are considered absent if the visitor does not find the provider physically present at the facility at the time they check attendance. If a provider is present at the facility, regardless of whether they are seeing patients or otherwise engaged (such as attending a meeting), they are counted as present. Providers are only considered absent if they are not on the premises.

In each of the 60 participating facilities, we first created a roster of all current healthcare providers who routinely offer family planning services, with assistance from facility staff. Each participating facility contained between 1 and 10 providers who routinely offer family planning services. Prior to conducting unannounced visits, we conducted a brief facility audit to check for electricity, running water and distance to several amenities. At this visit, the enumerator also conducted a brief provider interview to capture information on provider demographics and job satisfaction; if the provider was not present at the time of the short facility audit, we conducted the provider interview at a subsequent visit.

We conducted two unannounced visits in each of the 60 facilities. Approximately five business days before each unannounced visit, an enumerator contacted each facility to confirm the employee roster was correct and up-to-date. Across all 60 facilities, enumerators checked the attendance of 212 providers and 205 of these providers (97%) provided written informed consent to participate in a short provider interview. Public-sector facilities in each participating county are officially open Monday through Friday from 8:00 a.m. to 5:00 p.m. and officially provide family planning services during these hours (with the exception of the lunch hour from 1:00 p.m. to 2:00 p.m.). Each facility was visited once at 9:00 a.m. (1 h after the official opening time) and once at 3:00 p.m. (1 h after the end of the official lunch break). If the provider was physically present in the facility when the enumerator arrived, they were marked as present. Otherwise, they were marked absent. The two unannounced visits occurred between 2 and 4 weeks apart. Additionally, certain times of year may contribute to higher or lower rates of absenteeism. For example, the primary religious season (mid-December through early January) may incur more pronounced absenteeism as clinicians may travel to their rural homes for an extended period. Additionally, during the rainy season, absenteeism may increase as road conditions worsen. We intentionally chose to measure the prevalence of provider absenteeism during a period that would avoid the holiday season and would also straddle the dry and rainy seasons to avoid the effects of seasonality. The UAV portion of data collection, therefore, occurred from mid-January to mid-March 2019.

Qualitative data collection

We conducted eight FGDs across four counties in Western Kenya. The four counties participating in the FGDs were selected at random from the initial group of five counties, to streamline FGD logistics. For these discussions, we recruited current and former family planning clients between the ages of 18 and 49. Participants were identified by community health volunteers, who approached 240 women, of whom 88 agreed to participate; of these 88, 55 women arrived on the day of the FGD. FGDs averaged 103 min, took place in a private and convenient location, and contained six to eight participants ranging in age from 18 to 46. FGDs were led by a trained female moderator using a semi-structured questionnaire of 18 predetermined questions designed to explore barriers women face in accessing contraception. Participants were encouraged to communicate in the language with which they felt most comfortable, with moderators able to interpret as needed.

We also conducted 19 interviews with key informants, selected via a snowball sampling technique that began with the Head of Reproductive Health within the County Health Department for each county. We initially contacted 27 key informants, all of whom agreed to participate, but eight became unavailable due to repeated scheduling conflicts. Our final sample included senior staff from public- and private-sector healthcare facilities and non-governmental public health organizations, as well as county-level government officials. Each KII was conducted by an experienced and trained enumerator using a semi-structured questionnaire of 19 predetermined questions to explore feasible and promising solutions to the barriers women face in accessing contraception. Interviews averaged 55 min and were conducted in English.

Finally, we synthesized all data collected within the larger project into two journey maps (Trebble et al., 2010): a patient journey map and a provider journey map. The goal of the journey maps was to visually represent the process of seeking or providing family planning services. We vetted each map using a workshop. Eligibility criteria and recruitment procedures for the ‘patient’ journey mapping workshop were identical to those used for the FGDs (described above) and resulted in nine married, current contraceptive users aged between 27 and 41 years. For the ‘provider’ journey map, we recruited using snowball sampling in which a public-sector nurse well-known to the study PI was the first point of contact, resulting in 12 participants: 9 were females, aged between 27 and 52 years, and years providing family planning services ranged from 2 to 12.

Audio recordings of the FGDs, KII and journey mapping workshops were professionally transcribed and, if needed, translated to English. We used a qualitative description approach to conduct conventional content analyses on the transcripts. We managed the data with NVivo 11.0 (QSR International). An American and a Kenyan member of the research team with qualitative methods training read each transcript holistically, generated a codebook with definitions and examples, and then both coded every transcript. They reviewed coding daily by Skype to talk through disagreements until they reached a consensus. To enhance trustworthiness, they kept an audit trail, bracketed their bias in reflective practice and triangulated the findings with the other forms of data collected (Lincoln and Guba 1986).

Additional details of our overall sampling strategy and the methodology for each of the study components described above are provided in a previous publication (Tumlinson et al., 2021). Ethical approval for the study protocol was provided by both (information removed for blinded review).

Results

Facility characteristics

Characteristics of the 60 participating public-sector healthcare facilities are presented in Table 1. The facilities were primarily characterized as rural (88%), with between one and three providers on duty at any one time. Most facilities (63%) reported not having a staff duty roster. The large majority of participating facilities had amenities such as electricity, running water, a telephone and—on average—were located within 4 to 6 km of other points of interest such as a main road, a bus station, a market or an internet café. Across the 212 observed providers, participants were approximately evenly distributed among the three facility types, with 32% working in a dispensary (Level 2 facility), 31% in a health centre (Level 3 facility) and 37% in sub-county or county hospital (Level 4 facility) (data not shown).

Table 1.

Characteristics of 60 public-sector healthcare facilities in Western Kenya, 2019

	N = 60	%
Rural versus urban location
Rural	53	88%
Urban	5	8%
Peri-urban	2	3%
Total number of providers in the facility
One	4	7%
Two	28	47%
Three	7	12%
Four	5	8%
Five to nine	16	27%
Does this facility currently have a duty roster in place?
Yes	22	37%
No	38	63%
How often does the facility create a new roster?	(n = 38)
Every other week	2	5%
Monthly	34	89%
Quarterly	1	3%
Yearly	1	3%
Does this facility currently have a manager?
Yes	59	98%
No	1	2%
Is this facility monitored by the community?
Yes	58	97%
No	2	3%
Does this facility currently have electricity?
Yes	52	87%
No	8	13%
Does this facility have a generator?
Yes	8	13%
No	52	87%
Does this facility have running water?
Yes	57	95%
No	3	5%
Does this facility have drinking water?
Yes	59	98%
No	1	2%
Does this facility have soap?
Yes	58	97%
No	2	3%
Does this facility have a refridgerator?
Yes	56	93%
No	4	7%
Does this facility have a telephone?
Yes	45	75%
No	15	25%
Distance to other points of interest
Road
Less than 1 km	16	27%
1–5 km	20	33%
6–10 km	18	30%
More than 10 km	6	10%
Mean distance (range)	6 km (0–50 km)
Bus stand
Less than 1 km	16	27%
1–5 km	23	38%
6–10 km	15	25%
More than 10 km	6	10%
Mean distance (range)	6 KMs (0–50 KMS)
Market
Less than 1 km	20	33%
1–5 km	29	48%
6–10 km	7	12%
More than 10 km	4	7%
Mean distance (range)	4 km (0–30 km)
Bank
Less than 1 km	2	3%
1–5 km	11	18%
6–10 km	12	20%
More than 10 km	35	58%
Mean distance (range)	17 km (0–80 km)
School
Less than 1 km	43	72%
1–5 km	16	27%
6–10 km	0	0%
More than 10 km	1	2%
Mean distance (range)	1 km (0–25 km)
Café
Less than 1 km	18	30%
1–5 km	26	43%
6–10 km	10	17%
More than 10 km	6	10%
Mean distance (range)	6 km (0–50 km)

Provider characteristics

Table 2 describes the 205 providers participating in this study. Most providers were nurses¹ (90%), females (69%) and currently married (87%); among those married, just 69% reported that they lived with their spouse. Participants had an average of one or two children, and two-thirds were below 40 years of age. Twenty-two percent of participating providers were the facility managers. A few (11%) held a bachelor’s degree and more than 80% reported good or excellent health. On average, participants had worked as healthcare providers for approximately 12 years, with an average of 4 years at their current facility. Over 70% of providers had a daily one-way work commute of 30 min or less and 20% lived within a 5-min walk from the facility where they worked. Seventy-one percent received training in the previous 12 months while 67% received a supervisory visit in the previous three months. Sixty-nine percent were from the county in which they worked and 78% were fluent in the local language commonly spoken by most of their patients. About two out of five participants (43%) reported earning less than 75 000 KSH per month and a similar percentage (45%) reported that they were not always paid on time. Just over a third (34%) reported that the necessary commodities and supplies were often out of stock and only 56% would rate their facility as a four or five on a five-point scale of cleanliness.

Table 2.

Provider characteristics and distribution of absenteeism—defined as absent at one or both visits—Western Kenya, 2019

	Provider Present		Provider Absent		Total
	N = 40	%	N = 165	%	N = 205	%
Provider type
Nurse	34	85%	150	91%	184	90%
Clinical Officer	5	13%	15	9%	20	10%
Other	1	3%	0	0%	1	0%
Gender
Female	30	75%	111	67%	141	69%
Male	10	25%	54	33%	64	31%
Age
21–30	9	23%	27	16%	36	18%
30–34	9	23%	57	35%	66	32%
35–39	6	15%	28	17%	34	17%
40–44	5	13%	17	10%	22	11%
45–49	1	3%	13	8%	14	7%
50–54	2	5%	11	7%	13	6%
55 and older	8	20%	12	7%	20	10%
Education—Completed a bachelor’s degree
Yes	3	8%	20	12%	23	11%
No	37	93%	145	88%	182	89%
Health
Excellent	6	15%	16	10%	22	11%
Good	23	58%	123	75%	146	71%
Fair	11	28%	25	15%	36	18%
Poor	0	0%	1	1%	1	0%
Marital status
Currently married	35	88%	143	87%	178	87%
Not currently married	5	13%	22	13%	27	13%
Children under 5 years of age
No children under 5	1	3%	17	10%	18	9%
One child under 5	20	50%	47	28%	67	33%
Two children under 5	18	45%	79	48%	97	47%
Three children under 5	1	3%	20	12%	21	10%
Four or more children under 5	0	0%	2	1%	2	1%
How many years as a provider?
1–5 years	10	25%	37	22%	47	23%
6–10 years	13	33%	60	36%	73	36%
11–15 years	4	10%	27	16%	31	15%
16–20 years	2	5%	12	7%	14	7%
21–25 years	1	3%	8	5%	9	4%
More than 25 years	10	25%	21	13%	31	15%
How long at this facility?
Less than 1 year	6	15%	14	8%	20	10%
1–2 years	8	20%	41	25%	49	24%
3 years	3	8%	30	18%	33	16%
4 years	5	13%	32	19%	37	18%
5–9 years	10	25%	44	27%	54	26%
10 or more years	8	20%	4	2%	12	6%
Commute from work to home
0–15 min	17	43%	51	31%	68	33%
16–30 min	16	40%	62	38%	78	38%
31–35 min	2	5%	21	13%	23	11%
46–60 min	4	10%	23	14%	27	13%
1–2 h	1	3%	7	4%	8	4%
More than 2 h	0	0%	1	1%	1	0%
Received training in last 12 months
Yes	24	60%	122	74%	146	71%
No	16	40%	43	26%	59	29%
Supervise visit in the last 3 months
Yes	28	70%	110	67%	138	67%
No	12	30%	55	33%	67	33%
	N = 40	%	N = 165	%	N = 205	%
From this county
Yes	23	58%	118	72%	141	69%
No	17	43%	47	28%	64	31%
Fluent in local language
Yes	28	70%	131	79%	159	78%
No	12	30%	34	21%	46	22%
Lives within a 5-min walk
Yes	9	23%	33	20%	42	20%
No	31	78%	132	80%	163	80%
Always paid on time
Yes	23	58%	89	54%	112	55%
No	17	43%	76	46%	93	45%
Participant is the facility manager
Yes	9	23%	36	22%	45	22%
No	31	78%	129	78%	160	78%
How often are necessary commodities stocked out?
Often	16	40%	54	33%	70	34%
Sometimes	18	45%	77	47%	95	46%
Rarely	6	15%	32	19%	38	19%
Never	0	0%	2	1%	2	1%
Cleanliness of facility on scale of 1–5
1 (least clean)	0	0%	3	2%	3	1%
2	2	5%	11	7%	13	6%
3	15	38%	60	36%	75	37%
4	21	53%	77	47%	98	48%
5 (most clean)	2	5%	14	8%	16	8%
Current wages
Less than 35 000 KSH/month	6	15%	12	7%	18	9%
35 000–49 000 KSH	5	13%	10	6%	15	7%
50 000–74 000 KSH	11	28%	45	27%	56	27%
75 000 KSH or more	10	25%	58	35%	68	33%
Declined to state	8	20%	40	24%	48	23%
Lives with spouse, if married
Yes	25	71%	97	68%	122	69%
No	10	29%	46	32%	56	31%
Job satisfaction
‘My job is usually worthwhile’
Yes	37	93%	155	94%	192	94%
No	3	8%	10	6%	13	6%
‘My job is usually interesting to me’
Yes	39	98%	161	98%	200	98%
No	1	3%	4	2%	5	2%
‘My job suits me very well’
Yes	38	95%	162	98%	200	98%
No	2	5%	3	2%	5	2%
‘If given a chance (at the same rate of payment) I would get a different job’
Yes	5	13%	16	10%	21	10%
No	35	88%	149	90%	184	90%
‘Feels emotionally drained at the day’s end’
Yes	23	58%	95	58%	118	58%
No	17	43%	70	42%	87	42%
‘ Feels working with people the whole day is a strain’
Yes	8	20%	29	18%	37	18%
No	32	80%	136	82%	168	82%
‘Feels fatigued when wakes up in the morning and has to face another day’
Yes	16	40%	68	41%	84	41%
No	24	60%	97	59%	121	59%
‘Feels is becoming harsh towards people since taking this job’
Yes	5	13%	13	8%	18	9%
No	35	88%	152	92%	187	91%
‘Feels this job is hardening them emotionally’
Yes	10	25%	65	39%	75	37%
No	30	75%	100	61%	130	63%
‘treats some patients as if they were impersonal objects’
Yes	2	5%	1	1%	3	1%
No	38	95%	164	99%	202	99%

Across multiple measures of job satisfaction, the majority (90% or more) of providers reported their job is worthwhile, interesting and suitable and that they would not switch to another job for the same pay. Few reported becoming harsh towards patients (11%) or treating them as impersonal objects (1%). However, 58% reported feeling emotionally drained at the end of the day, 18% reported that working with people the whole day is a strain and 41% reported feeling fatigued upon waking. More than a third (37%) felt the job is hardening them emotionally.

Nature of absenteeism

Key informants largely confirmed that they knew that public-sector staff were often not on duty serving patients during the hours of operation. As seen in the data collected by the UAVs, key informants discussed provider absenteeism as a range of different behaviours, including providers who are late, take long lunches, do not return from lunch, close the facility early or are absent from the facility for the entire scheduled workday. Providers might also be in the facility but not attending to patients.

In the focus groups, women recounted many forms of absenteeism they encountered while trying to obtain family planning services, ‘You can arrive, you start queueing, wait for a while then afterwards you are being told that the provider is not around’ (current contraceptive user, patient journey map workshop). Women also observed providers chatting with each other, instead of seeing patients, ‘They open, shortly then they close and go telling stories to each other’ (current contraceptive user FGD, Kisumu County, Urban).

Women participating in focus groups were accustomed to providers leaving for lunch with patients in the queue, who were expected to wait for the providers to return to receive services. Women expressed frustration that providers at times failed to return or would not provide services when they did.

I went there at 8…When it was already 8, they did cleaning and other things and when they finished and began treatment at 10, we were in queue…. at about 1 going to 1:30, he tells me it’s already his lunch time and that I should go back outside for him to go to lunch and when he comes back he will attend to me. He left at 1 and came back at 3:30. And when he came back, he refused to attend to me saying I should come back the next day. It seems when he came back at 3:30, it was to close for the day. (current contraceptive user FGD, Kisumu County, Urban).

Women in focus groups observed that attending off-site meetings was an official duty for providers, but also noted that providers had discretion about how they fit those excused absences into their workday and whether or not they returned to work after meetings to attend to clients: ‘They have been called for a meeting. So, it will force them to attend that meeting first. If they finish early, they will come and attend to you but if they finish, they don’t’ (current contraceptive user, Busia County, Urban).

Key informants suggested a range of possible work-related rationales for absences. For example, providers may be required to attend clinical trainings and official meetings. Providers may also be late or absent because they were obtaining supplies from other facilities to address stockouts at their own facility. In the journey mapping workshops, providers confirmed that official work-related issues, such as fulfilling documentation requirements, may interfere with their ability to see patients while they are in the facility. Additionally, providers indicated that organizational partners made unscheduled visits to facilities and requested reporting. When that occurred, they felt obligated to stop seeing patients and attend to those representatives.

Key informants also discussed more personal reasons why a provider may be absent from work. One key informant expressed that funerals were only excused for immediate family members, but staff would often co-operate to cover the shift of colleagues attending other funerals. Funeral attendance was not cast as irresponsible. Transportation was also cited as a reason providers arrived late to work. However, punctuality was characterized as a ‘moral value’ (public-sector provider, KII), and unexplained absences were judged negatively. Some key informants generally attributed absenteeism to a bad attitude or a lack of commitment to the work on the part of the provider.

In the provider journey mapping workshop, public-sector providers affirmed that they might be late or absent for work because of official leave, official meetings, transportation difficulties, personal errands, taking care of sick children, being sick themselves and attending funerals. They also recognized that burnout, motivation and wages could affect engagement in their work, but denied that those were legitimate reasons for absenteeism: ‘That wages are insufficient – of course. It’s not enough but it does not make us not wake up in the morning and go to work… About the unmotivated or burnt out—I don’t think it’s right. Right now, we are going back to work unmotivated and it cannot make someone not go to work’ (public-sector provider, provider journey map workshop).

Providers in the journey mapping workshop offered a range of perspectives on lunch breaks. One described providers as often skipping lunch breaks to keep seeing patients, working without breaks. Others were sceptical and suggested that providers often lacked ‘discipline’ about their breaks. One participant asserted that, ‘Our lunch break should be one hour and our tea session, thirty minutes’ (public sector provider, provider journey map workshop). Providers also describe experiencing hostility from patients when they take their lunch break:

We just work throughout even up to two. You find you are exhausted, dehydrated. So you compensate. Sometimes you don’t even spend that one hour. You come back early but the type of clients we have if you tell them you’re going for lunch, they start clicking at you. They don’t understand that you’ve been working since morning and so you get annoyed (public-sector provider, provider journey map workshop).

Scope/Prevalence of absenteeism

The percentage of providers absent at the first and second unannounced visits was 57% and 59%, respectively (Table 3). Thirty-five percent of providers were absent at both unannounced visits and less than one out of five providers were present at both visits. Among those present at one or both visits (n = 178), 19% were not visibly working at the time of the visit (Table 4), meaning that, although physically present, the provider was not attending to patients, completing paperwork or otherwise engaged in any work related to the functioning of the facility or service delivery. Absenteeism calculated from the morning unannounced visit (9:00 a.m.), at 69%, was substantially higher than the percent absent calculated from the afternoon visit (3:00 p.m.), at 47% (Table 5); this may indicate that a portion of absent providers are arriving more than 1 hour late to work rather than being absent from the facility for the entire day. There were no meaningful differences in the percent of providers absent by rural versus urban location or across the different types/sizes of facilities (data not shown). Very few meaningful differences in absenteeism were seen across these provider characteristics. Those providers with just one child and those who had worked at their facility for more than 10 years were notably more often present than absent (Table 2).

Table 3.

Provider absenteeism across both unannounced visits (N = 212) in 60 public-sector facilities in Western Kenya, 2019

		Provider absent at second visit
		Yes	No	Total
Provider absent at first visit	Yes	74	51	125 (59%)
	No	47	40	87
	Total	121 (57%)	91	212

Table 4.

Among those providers present at one or both visits (N = 178), the percent of providers offering services in 60 public-sector facilities in Western Kenya, 2019

Offering services	N	%
Yes	144	81%
No	34	19%
Total	178	100%

Table 5.

Provider absenteeism, stratified by morning versus afternoon visit in 60 public-sector facilities in Western Kenya, 2019

	9 a.m./N	9 a.m./%	3 p.m./N	3 p.m./%	Total N	Total %
Absent	147	69%	99	47%	246	58%
Present	65	31%	113	53%	178	42%
Total	212	100%	212	100%	424	100%

During the first round of unannounced visits, at 11 of the 60 facilities there were no providers at all on site when the UAV arrived. In six of these facilities, at least one provider came within an hour of the start of the unannounced visit (meaning they arrived before 10:00 am); in the other five facilities, the facility remained closed for the entire day. During the second round of unannounced visits, there was no one present in 13 of the 60 facilities; in all 13 of these facilities, one or more providers arrived sometime between 9:00 a.m. and 12:00 p.m. Across both visits, enumerators found no providers present in 20% of the visits (24 out of 120 visits). Enumerators shared feedback with the study team to describe their experience of finding no medical staff present at the facility, for example: The facility was opened at 9:10 a.m. by the cleaner. The in charge [manager] arrived at the facility at 9:56 am.  The cleaner then said one provider who is the in charge [manager] has been away for 3 days and today she was still not coming, the other provider she was not sure of her coming because yesterday she was unwell.  When we arrived at the facility, it was open but only the cleaner was present. The first provider came in at 10:00hrs. She claimed she had gone to look for some vaccines at the sub-county hospital which she didn’t find.

Reasons for absenteeism

Across the 246 instances in which a provider was absent, a little more than half of absences (54%) were reportedly sanctioned (Table 6). For example, 38% of absences were reportedly due to authorized personal leave such as vacation, sick leave or a personal day, while 12% were due to official off-site duties such as a workshop or training event. Another 4% were absent during the unannounced visit due to a variety of valid reasons such as being transferred to another facility or being assigned to the night shift. However, the remaining 46% of absences could not be categorized as a sanctioned absence; in nearly one out of four absences (23%), no one at the facility could explain why the provider was absent. In 12% of absences, the provider was reported to be attending to personal business and in 7% of absences, the provider was observed or reported to be arriving more than 90 min late to work. In nine instances, it was not possible to know why the provider was late because there was not one person present at the facility to answer any questions from the enumerator. In the majority of instances of absence (81%), another healthcare provider supplied the reason why their colleague was absent, while occasionally (12% of the time) it was a staff person, such as a cleaner or lab technician, who provided this information (Table 6). In 98% of reports, the person supplying the information was ‘very certain’ that the information they provided was accurate; in the remaining 2%, the informant was ‘somewhat certain’ (data not shown).

Table 6.

Reported reason for provider absence, among those providers not present (n = 246 visits)

Reported reason for absence	N	%
Sanctioned absence
Authorized personal leave such as vacation, sick leave or a personal day	93	38%
Official off-site duties such as a training, workshop, outreach or other	29	12%
Staff was transferred, working elsewhere today or on night duty	11	4%
Possibly unsanctioned absence
No reason could be provided	56	23%
Family or personal reasons such as personal errands, a funeral, etc.	30	12%
The provider was more than one hour late to work in the morning or did not return to work after lunch	18	7%
There was no one present at the facility to indicate why the provider was absent	9	4%
Total	246	100%
Who provided the reason
Another healthcare worker	199	81%
Another staff person such as a cleaner	29	12%
The enumerator observed to provider arriving more than 90 min late to work (morning visit only)	9	4%
There was no one present at the facility to indicate why the provider was absent	9	4%
Total	246	100%

Impact of absenteeism

In every FGD, women described how provider absenteeism led to long queues, which resulted in a range of subsequent outcomes, described above. Some women described waiting until the provider returned to the facility in order to obtain family planning. When women could not wait or the provider did not return, they left without family planning.

Women described attempting to obtain family planning while fulfilling their other responsibilities, but feeling thwarted when they encountered long waits or had to make repeat visits:

So when you get there when it’s like one hour to twelve, they want to go for lunch. And you have just reached at two and know it’s still working hours and the doctors are still there. And they will tell you, ‘Where were you for you not to come early? It’s lunch time and we cannot inject you. Go and look for somewhere else or come back tomorrow. (former contraceptive user FGD, Bungoma County, Urban).

Key informants characterized provider absenteeism as harming patients by creating long queues and long waits, ultimately reducing family planning utilization and discouraging women from seeking services. Late providers were characterized as more likely to treat patients poorly, such as not greeting them or behaving with hostility. Women confirmed that they encountered hostile responses when they tried to expedite their visits.

Key informants noted that absenteeism had the most dramatic effect on facilities with a single provider, because the facility would have to close. However, even in settings with multiple providers, absenteeism may negatively impact family planning clients. For example, when senior providers attend trainings, they leave less experienced colleagues at the facility who might offer a limited method mix as they lack the skill or confidence to insert intrauterine devices or implants.

Addressing absenteeism

In the FGDs, women articulated that absenteeism was a problem that needed to be fixed, characterizing the behaviour as inappropriate.

The doctors should know that during working hours, they should be attending to everyone and not making stories. If they are working, they should leave behind chatting so that they can attend to everyone instead of us waiting and they tell us they are busy. They should tell us when they are busy so that we do not go to the hospital, instead of telling us to go back to the hospital, only to find them busy. They will really delay us. So, the doctors should do their work because they are being paid for it. (current contraceptive user FGD, Busia County, Rural).

Some women viewed the unpredictability of provider availability as a problem to be fixed: ‘Let them know that 2:00 pm is time to resume work. At 2:00 pm, they should be in so that when the client comes, they can find you.’ (former contraceptive user FGD, Kisii County, Rural). In the provider journey mapping workshop, a participant suggested that provider lunch breaks be staggered to ensure uninterrupted coverage.

One key informant described attempts to address provider lateness at their clinic by booking appointment times. Some key informants described systems of ‘local arrangements’ for personal absences although these did not consistently assure coverage and queues still resulted (senior government official, KII).

Increased supervision and accountability also emerged as possible solutions, with one key informant suggesting that absenteeism is a result of ‘weakness of leadership’ (private sector/NGO high level staff, KII). However, key informants framed absenteeism as requiring a degree of close supervision that was challenging to achieve because employees could be deceptive, staff may be too busy to monitor supervisees and some supervision would require travel, which was particularly burdensome at rural or remote facilities. It was noted that providers with personal relationships with their supervisors did not expect to be held accountable. When describing absenteeism in the public sector, the private sector was invoked as a setting where monetary gain motivated punctuality. Furthermore, participants indicated that private-sector providers could be fired, while public-sector employees saw no consequences to absenteeism as they were neither promoted when they performed well nor punished for absenteeism.

Those providers participating in the brief structured interview with the UAV offered several suggestions for ways to minimize absenteeism among public-sector healthcare providers (Table 7). The most popular provider-sourced suggestion included increasing the workforce and instituting more frequent salary increases. About one in five providers posited registers for signing in and out would reduce absenteeism while an equal number complained that a lack of supplies and commodities reduced the motivation for providers to come to work, with one provider commenting, ‘The government should supply drugs so that providers can be motivated to come to work.’ Others suggested provider training and sensitization around absenteeism, recommending a ‘Change of positive attitude towards work by providers through training so that they can know the importance of their job and the value of their patients.’ Residential housing (i.e. housing within the facility compound, to remove time spent in transit) and regular promotion were additional suggestions offered by many participants, with one enumerator noting: ‘The provider felt like his job at this facility is not worthwhile since he has 8 years of experience but has never been promoted.’ Of note, many providers cited the need for solutions that address the intrinsic motivation and attitude of public-sector healthcare providers. For example, 13% of participants suggested an important role for non-financial incentives such as words of appreciation or occasional tea and 6% of providers called for a shift in negative attitudes among providers. Another 9% called for better communication and relationships across staff and with their managers. Rarely mentioned solutions included increased supervision, discipline and overtime or hardship pay.

Table 7.

Provider suggestions for ways to minimize absenteeism

Ways to reduce absenteeism	N	%a
Increase the number of healthcare workers	98	48%
Reliable/regular salary review and increment	79	39%
Introduce sign-in/sign-out registers	38	19%
Adequate supplies and commodities	37	18%
Training and sensitization around absenteeism and job commitment	35	17%
Residential houses	34	17%
Regular promotion	29	14%
Institute non-financial incentives (such as on-site meals, appreciation and training)	26	13%
Better communication and relations with staff and management	18	9%
Commuting allowances	18	9%
Discipline absentee workers	17	8%
Better supervision and attention to absence	16	8%
Shift negative attitude & increase intrinsic motivation	12	6%
Accommodation allowances	12	6%
Institute a duty roster	9	4%
Post to a facility closer to family	5	2%
Pay for overtime	4	2%
Hardship and risk allowance	4	2%

^aParticipants could name more than one suggestion; total does not sum to number of participants; percent column has 205 in denominator.

Discussion

Nearly two-thirds of providers were absent at each unannounced visit and one-third of providers were absent for both visits conducted in this study of 60 public-sector facilities in Western Kenya. Nearly half of absences were unsanctioned—a result of providers arriving late, conducting personal business, taking an extended lunch or being missing in action for reasons unknown to their peers. Accordingly, just over half of absences were sanctioned. Sanctioned absences primarily included official leave, attending meetings, seminars and sourcing clinic supplies. Whether sanctioned or not, it could be argued that absent providers make advanced arrangements for colleagues to cover their shifts, yet our investigation found this is not always the case: in approximately 20% of unannounced visits, we found no providers present at the time of the visit and for five different facilities we found the facility closed for the entire day on the day of our first unannounced visit. A complete absence of providers or a facility closure could pose an especially large barrier for potential new contraceptive users who may be discouraged from initiating contraceptive use as a result of not finding any providers present when seeking contraceptive care.

We hypothesized that absenteeism may be higher among providers who are not from the community where they work (thus lowering their incentive to serve their community) or among those with long commute times, which may impede timely attendance. However, most providers participating in this study were from the area, fluent in the local language and live close to the facilities where they work, which were often situated within a reasonable distance from other points of interest. Additionally, we hypothesized that living apart from a spouse, being in poor health or poor job satisfaction may contribute to the high rates of absenteeism found in our study. Yet providers in our sample overwhelmingly reported being in good health and finding their job worthwhile/interesting and more than two-thirds reported living with their partners. Notably, even though large numbers of providers reported they were not paid on time, found their jobs emotionally draining and found necessary supplies often out of stock, these facility-level characteristics did not vary meaningfully between providers who were present versus those who were absent.

This study sheds light on the frequency with which family planning clients in Kenya face a critical—and in some cases, insurmountable—barrier in the form of absent providers. While the field of international family planning frequently calls for improvements in provider training and knowledge, our findings suggest such investments are unlikely to translate into improved family planning service delivery if public-sector family planning providers simply do not show up for work. The Kenyan government is committed to increasing contraceptive prevalence over the next decade. These commitments, formalized at the 2012 London Summit on Family Planning, were updated in the 2017 Kenyan Family Planning Costed Implementation Plan (CIP). As the Kenyan government has devolved their public healthcare system and as UHC is implemented across the country, protecting against continued disruption to family planning service delivery resulting from high provider absenteeism will be critical for achieving the goals of the CIP, protecting the health of Kenyan women and children and safe-guarding reproductive justice.

Key informants in the healthcare sector, as well as women participating in focus groups, confirmed that absenteeism occurs and has a negative impact on family planning uptake, regardless of the reason. In the absence of another provider to offer coverage, women experience an excused absence as a barrier much the same as an unexcused absence. Notably, in the provider journey mapping workshop, participants who themselves were public-sector providers emphasized how absenteeism occurs in part because the working conditions diminish motivation, which triangulated with the provider survey results about the frustration of working without sufficient supplies or promotions.

Healthcare providers are a key pillar of the health system and are integral to the realization of UHC. Provider absenteeism, therefore, has the potential to affect the quality of family planning service delivery and may also derail Kenya’s aspirations of achieving UHC. The level of provider absenteeism documented in this study is much higher than reported in a study conducted in 2008 in southeastern Kenya, for which nurses were absent in 19% of unannounced visits (Muthama et al., 2008). Our findings also differ from publicly available data collected by the World Bank in 2012 in Kenya, which found public-sector nurses absent in nearly 30% of visits (Martin and Pimhidzai 2013), but are more closely aligned with more recent World Bank data (collected in 2018) which found public-sector nurses absent in 59% of visits; in the five specific counties included in our study, the absence rate among all nurses ranged from 55 to 59% (Das and Omollo 2019). Notably, although similar percentages of nurses were absent due to authorized personal leave (42% in the World Bank study compared with 38% in our study) only 10% of nurse absenteeism in the World Bank 2018 study was classified as unauthorized, compared with 46% in our study. This discrepancy could result from our decision to classify as ‘unauthorized’ any provider absence that could not be explained by other providers or staff on-site at the time of the unannounced visit.

In seeking to compare our results with those of prior studies on provider absenteeism in Kenya, it should be noted that our methodology differs slightly from these prior assessments, in that we included all providers who routinely offer family planning in our denominator (including those with scheduled leave), allowing us to fully document all sanctioned absences among family planning providers. This slightly different methodology may explain part of the difference between our results and prior studies. However, it is also possible that provider absenteeism in Kenya is growing, as a reflection of mounting burnout among healthcare providers or a growing need to seek supplemental income during working hours as public-sector wages are stagnant or frequently late (Tumlinson et al., 2020). In contrast to our study, the 2008 study in southeastern Kenya found job satisfaction and stress were important determinants of provider absenteeism (Muthama et al., 2008). Similar to our study, the authors found that providers with longer service at their current facility were less likely to be absent, compared with those more recently posted to their current position.

The negative impact of absenteeism noted by participants in our study is confirmed by a larger literature. Absenteeism is known to reduce available person-hours, thereby exacerbating the shortage of health care workers plaguing health systems in many LMICs like Kenya (Dovlo, 2005; Miseda et al., 2017; Tumusiime, 2017). Additionally, absenteeism is known to increase the workload of providers who do show up, to contribute to poor quality of care and to increase patient dissatisfaction with health services (Rowe et al., 2005; Duclay et al., 2015). These negative impacts may be especially problematic across all types of service delivery—not just family planning—in the context of UHC. If UHC is to succeed in the Kenyan health system, it is imperative that skilled healthcare providers are in place and attending to those individuals seeking care during all hours of operation, not just a few select hours a day or select days during the week.

Due to the diversity of reasons for high provider absenteeism found in our study, multiple strategies may be needed to deal with and mitigate the effects of provider absenteeism in Kenya. Authorized absenteeism could be mitigated by rationing staff leave time and attendance at off-site trainings; rational approval of staff leave will be essential given the large percentage of absenteeism—54%—resulting from authorized leave. Unauthorized absenteeism could be addressed through regular supportive supervision, unannounced visits to health facilities by health leaders and developing mechanisms for tracking staff availability in health facilities. Late arrivals or prolonged lunch breaks could be attributed to lack of staff housing within health facilities especially in rural areas (Chaudhury et al., 2006). Yet health system deficiencies such as those reported in this study—inadequate supervision, shortage of essential health products and technologies, low wages and irregular salary payment—cultivate an environment ripe for provider absenteeism and may frustrate efforts to improve attendance (Willis-Shattuck et al., 2008; Mbindyo et al., 2009; Mukasa et al., 2019).

We note two important limitations of this analysis. First, of necessity, we contacted facilities in advance to obtain informed consent from facility managers and to verify the current roster of providers scheduled to work on the day of the unannounced visit. While we do not believe facilities had any advanced knowledge of the exact date and time of our unannounced visit, we cannot guarantee all facilities had no advanced knowledge of each visit. In the event that one or more facilities or providers anticipated an unannounced visit, the provider(s) may have increased attendance, and this would have resulted in a downwardly biased (or more conservative) estimate in the prevalence of absenteeism among providers in Western Kenya. Secondly, we did not group FGD participants based on characteristics such as age, education or marital status and FGD participants skewed slightly older. This may have resulted in lack of generalizability among adolescent women and discussions may have been dominated by married women with more education.

Conclusion

As previous quality experts have noted, quality begins with showing up (Chaudhury et al., 2006). Our study finds that a sizeable portion of family planning providers in Western Kenya are missing in action, resulting in wasted resources and contributing to poor reproductive health outcomes in the communities they serve. Current supervisory efforts do not ensure that public-sector healthcare providers show up to work on time—or sometimes at all. These findings have dire implications as UHC continues to roll out in Kenya. Key informants made recommendations related to improved supervision and accountability of family planning providers while simultaneously acknowledging that such steps are difficult to achieve in the current Kenyan health system, particularly in the context of busy supervisors, difficult to reach facilities, delayed promotions and the cultural expectation that personal relationships between managers and providers prevent appropriate sanctions for absent workers. Future research is urgently needed to identify and rigorously evaluate public health interventions with the potential to reduce provider absenteeism.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Funding

Support for this research was provided in part by a career development grant (R00 HD086270) to the lead author and an infrastructure grant for population research (P2C HD047879) to the Carolina Population Center at the University of North Carolina at Chapel Hill. The Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) of the National Institutes of Health (NIH) awarded both these grants. The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the NIH/NICHD.

Author contributions

K.T.: conception and design of the work, data collection, data analysis and interpretation, drafting the article, critical revision of the article, project administration and funding acquisition. L.B.: data analysis and interpretation, drafting the article and critical revision of the article. C.W.: drafting the article and critical revision of the article. D.W.: data collection, drafting the article, critical revision of the article and project administration. D.O.: drafting the article and critical revision of the article. All authors approved the paper prior to submission.

Ethical approval.

Ethical approval for the study protocol was provided by both the University of North Carolina (#18-0783) and the Kenya Medical Research Institute (#634). Written informed consent was provided by study participants.

Conflict of interest statement.

None declared.