‘It’s a very complicated issue here’:
understanding the limited and declining use of
manual vacuum aspiration for postabortion care
in Malawi: a qualitative study
* Bregje de Kok,
and Maria Lisa Odland
Institute for Global Health and Development, Queen Margaret University, Edinburgh, UK,
Department of Sexual
Health, Cardiff and Vale NHS, Cardiff Royal Inﬁrmary, Cardiff, UK,
Anthropology Department, University of
Amsterdam, Amsterdam, The Netherlands,
Institute for Global Health and Development, Queen Margaret
University, Edinburgh, UK and
Norwegian University of Science and Technology, Trondheim, Norway
*Corresponding author. Department of Sexual Health, Cardiff Royal Inﬁrmary, Newport Road, Cardiff CF24 0SZ, UK.
Accepted on 11 August 2016
Malawi has one of the highest maternal mortality ratios in the world. Unsafe abortions are an im-
portant contributor to Malawi’s maternal mortality and morbidity, where abortion is illegal except
to save the woman’s life. Postabortion care (PAC) aims to reduce adverse consequences of unsafe
abortions, in part by treating incomplete abortions. Although global and national PAC policies rec-
ommend manual vacuum aspiration (MVA) for treatment of incomplete abortion, usage in Malawi
is low and appears to be decreasing, with sharp curettage being used in preference. There is
limited evidence regarding what inﬂuences rejection of recommended PAC innovations. Hence,
drawing on Greenhalgh et al.’s (2004. Diffusion of innovations in service organizations: systematic
review and recommendations. Milbank Quarterly 82: 581–629.) diffusion of healthcare innovation
framework, this qualitative study aimed to investigate factors contributing to the limited and declin-
ing use of MVA in Malawi. Semi-structured interviews with 17 PAC providers in a central hospital
and a district hospital indicate that a range of factors coalesce and inﬂuence PAC and MVA use in
Malawi. Factors pertain to four main domains: the system (shortages of material and human re-
sources; lack of training, supervision and feedback), relationships (power dynamics; expected job
roles), the health workers (attitudes towards abortion and PAC; prioritization of PAC) and the innov-
ation (perceived risks and beneﬁts of MVA use). Effective and sustainable PAC policy must adopt a
broader people-centred health systems approach which considers all these factors, their inter-
actions and the wider socio-cultural, legal and political context of abortion and PAC. The study
showed the value of using Greenhalgh et al.’s (2004. Diffusion of innovations in service organiza-
tions: systematic review and recommendations. Milbank Quarterly 82: 581–629.) framework to con-
sider the complex interaction of factors surrounding innovation use (or lack of), but provided more
insights into rejections of innovations and, particularly, a low- and middle-income country
Key words: Diffusion of innovations, low- and middle-income countries, Malawi, manual vacuum aspiration (MVA), people-
centred health systems, postabortion care
CThe Author 2016. Published by Oxford University Press in association with The London School of Hygiene and Tropical Medicine.
All rights reserved. For permissions, please e-mail: email@example.com 305
Health Policy and Planning, 32, 2017, 305–313
Advance Access Publication Date: 10 September 2016
Malawi has one of the highest maternal mortality ratios in the world
(WHO et al. 2014). Unsafe abortion may contribute to around 18%
of Malawi’s maternal mortality (Bowie and Geubbels 2013), where
abortion is illegal except to save the woman’s life. Levandowski
et al. (2013) estimated 67 300 abortions were performed in Malawi
in 2009; over one-third required treatment for complications. Many
organizations, including WHO (2014) and Ipas (2011) emphasize
that for countries like Malawi to meet global maternal mortality re-
duction targets, unsafe abortion deaths must be reduced. The
Malawian Ministry of Health’s (2009) key strategies for tackling un-
safe abortion-associated mortality and morbidity are family plan-
ning to reduce unintended pregnancies and provision of
postabortion care (PAC). PAC aims to reduce mortality and long-
term morbidity attributable to unsafe abortion through provision of
emergency care, treatment of incomplete abortion by uterine evacu-
ation, counselling and contraception (Corbett and Turner 2003).
Some authors (Descher and Cohen 2003;Grimes 2006) argue that
increasing contraception usage and reducing legal restrictions to safe
abortion should be priorities as PAC only mitigates the conse-
quences of unsafe abortions. However, in Malawi significant abor-
tion law reform is not anticipated soon (Chikuse 2015) and
increasing contraception use will not eliminate unwanted pregnan-
cies ( ˚
hman and Shah 2011). Thus, unsafe abortions will continue
to occur and quality PAC remains essential to reduce morbidity and
mortality (Pattinson et al. 2006).
WHO (2012) recommends vacuum aspiration (electronic or
manual) or misoprostol for first-trimester uterine evacuation as part
of PAC. Manual vacuum aspiration (MVA) is recommended over
dilatation and sharp curettage (D&C) as it is associated with fewer
complications, shorter hospital stays (Tunc¸alp et al. 2010) and can
be performed by midwives or nurses as an outpatient procedure
(Grimes et al. 2006). MVA was introduced as a pilot to Malawi in
1994 (Leme et al. 1997) and rolled out with investments in facilities,
training and equipment in 2001 (Schenck-Ygleslae 2004). It is the
recommended treatment for first-trimester incomplete abortion and
is on the country’s standard equipment list; misoprostol is not cur-
rently recommended in national PAC policy (Ministry of Health
2009). However, in their strategic assessment of unsafe abortion in
Malawi (involving observations of facilities and key informant inter-
views), Jackson et al. (2011) found MVA is used infrequently, with
D&C being used in preference. Reasons suggested for this included:
a lack of MVA equipment, equipment being locked up to prevent its
use in inducing abortions and a lack of trained staff. Moreover, a re-
cent study found declining MVA use in three Malawian hospitals
during 2009–12 from 31.0 to 4.9%, with a corresponding increase
in D&C (Odland et al. 2014). Studies have shown MVA to be more
cost-effective than D&C in Malawi (Benson et al. 2015) and other
low- and middle-income countries (LMIC) (Johnston et al. 2007;
Maonei et al. 2014). Thus, increasing MVA use may improve the
quality of PAC without significantly increasing spending, crucial in
Malawi given its high poverty levels and minimal health budget
(World Bank 2014).
It is therefore important to investigate what contributes to the
limited and declining use of MVA in Malawi, and why Malawi’s
policy, which recommends MVA, does not translate into practice.
Scholars have used various theoretical approaches to attempt to
understand how clinical policies translate into practice. Some studies
from high-income countries have used behavioural theories which
analyse how individual factors affect transfer of clinical policies into
practice (Godin et al. 2008). The widely used ‘diffusion of innov-
ations’ theory (Rogers 2003) goes beyond this individualistic ap-
proach, by focusing on individual decision-making processes but
also recognizing the importance of communication and social net-
works for how and why organizations or groups adopt innovations
(new ideas or technologies) or not. The theory proposes individuals’
use of innovations is the outcome of a staged process: obtaining
knowledge and understanding of the innovation; persuasion; deci-
sion to use the innovation or not; implementation of the decision; re-
inforcement of the decision through observed outcomes. This
process diffuses throughout the organization (Rogers 2003).
Rejection of the innovation, as happened with MVA in Malawi,
may occur at any point and be passive (individuals or organizations
do not consider the innovation) or active (consideration followed by
a decision to not use or discontinue using the innovation).
Greenhalgh et al. (2004) contend that diffusion of innovations
theory still treats behaviours as the outcome of conscious, rational
and individual decision-making processes, rather than as the com-
bined result of individual, communal, organizational, economic and
social factors. Moreover, they argue that diffusion of innovations
theory treats diffusion as an overly simple and linear imitation pro-
cess. In reality, diffusion is more complicated due to the impact of a
myriad of contextual factors, including management structures,
local priorities and power dynamics and result in formal dissemin-
ation processes (e.g. staff training) co-existing with unplanned diffu-
sion processes. Greenhalgh et al. (2004) capture the complexity of
the diffusion process and the interplay of a range of contextual fac-
tors in their diffusion of healthcare innovation framework.
We build on Greenhalgh et al.’s (2004) framework since context-
ual issues will be particularly important for MVA use in Malawi,
given the general importance of social, cultural and political con-
texts for sexual and reproductive health programmes (WHO 2007).
As Greenslade et al. (1994) and Kulczycki (2009) argue, controver-
sies surrounding abortion and women’s reproductive rights
•Numerous interconnecting factors inﬂuence the limited and declining use of MVA for postabortion care in Malawi.
•Effective and sustainable postabortion care policy must adopt a broader, people-centred health systems approach which
considers all these factors, their interactions and the wider socio-cultural, legal and political context of abortion and
•Human factors and perspectives are key to understanding diffusion of innovations.
•This study extends Greenhalgh et al.’s (2004) diffusion of innovation framework to the rejection of innovations and, par-
ticularly, low- and middle-income settings.
306 Health Policy and Planning, 2017, Vol. 32, No. 3
complicate the introduction of PAC innovations. Studies from a
number of LMIC examining implementation of MVA and PAC pro-
grammes found that various health system elements and social and
political issues, in particular abortion-related stigma, create chal-
lenges for the sustained implementation of MVA and PAC pro-
grammes (Tagoe-Darko 2013;Paul et al. 2014;Storeng and
Ouattara 2014). Therefore, the framework developed by
Greenhalgh et al. (2004) is particularly appropriate for the study of
the diffusion (or not) of MVA in Malawi. The framework highlights
that features of the wider context and health system interact
with the innovation adopters (health workers) and their relation-
ships and interactions with one another, and that all these factors to-
gether influence diffusion, dissemination and implementation of
Furthermore, Greenhalgh et al.’s (2004) framework chimes with
the growing body of literature on people-centred health systems,
which emphasize that ‘software’ (ideas, interests, values, norms and
relationships) is at least as important as ‘hardware’ (technical, finan-
cial and material resources) for health system functioning (Sheikh
et al. 2014). Thus, this study will explore how hard- and software
related contextual factors affect PAC and the use of MVA.
Moreover, we explore these issues through the perceptions of
healthcare providers. After all, providers’ perceptions and under-
standings are key for policy implementation (Greenhalgh et al.
2004;Blaauw et al. 2006;Aniteye and Mayhew 2013).
This small-scale qualitative study primarily involved interviews with
health workers conducted by the first author in Chiradzulu District
Hospital and Queen Elizabeth Central Hospital (QECH). We se-
lected these two government hospitals because public hospitals pro-
vide the majority of PAC in Malawi (Kalilani-Phiri et al. 2015) and
the selection of a rural and urban site enabled exploration of poten-
tial differences between rural and urban facilities. Furthermore,
QECH is the largest referral hospital in Malawi and sees the major-
ity of women with abortion complications. Odland et al. (2014)
found Chiradzulu had the lowest use of MVA (7.4%) compared
with two other Malawian hospitals. Interviews were supplemented
by unstructured observations of care practices (but no direct inter-
actions with patients) made during the 4-week study period, to
provide contextual insights concerning organizational set up, re-
sources and staff dynamics. Observations were triangulated with
Through purposeful sampling, we included PAC providing
health workers of different cadres (doctors, nurses and clinical offi-
cers), genders and levels of experience and seniority, to capture a
wide range of views and understandings. We excluded practitioners
who were not currently involved in providing PAC, were not pre-
dominantly based in either hospital or were working in their current
position for <1 month. A total of 11 participants were interviewed
in QECH and 6 in Chiradzulu (Table 1).
We used semi-structured interviews rather than focus groups be-
cause PAC is a potentially sensitive topic and may involve criticizing
colleagues’ practices; individual interviews allowed participants to
speak more freely (Bowling and Ebrahim 2005). Interview topics
included experiences of, and perceptions towards, PAC and MVA.
During the study, small adaptions were made to interview questions
mainly to reflect locally used terms (e.g. ‘evacuation’ rather than
‘D&C’). Acknowledging the importance of reflexivity in qualitative
research, we reflected throughout the study on how participants’ per-
ceptions of the interviewer as female, foreign, white and a doctor and
our own preconceptions and values, might influence the findings. For
instance, our western liberal values according to which abortion is a
human and reproductive right could lead to overly negative interpret-
ations of less liberal attitudes amongst providers. Furthermore, the
interviewer had previous experience of both providing PAC in
Malawi and abortion care within the UK, which may allow personal,
as opposed to participants’ thoughts and experiences, to influence the
findings. Reflection on our own views and experiences led to a more
balanced interpretation of attitudes and preferences.
Informed written consent was obtained from all participants fol-
lowing a written and verbal explanation of the study and the inter-
viewer’s inability to directly improve PAC. They were advised that
their personal responses would remain confidential and anonymous,
that they were not obligated to participate, and could refuse to an-
swer a question or withdraw at any time. Nobody refused to partici-
pate. Permission was requested to digitally record the interviews;
one participant refused, so handwritten notes were taken. The
recordings were transcribed and thematic analysis performed (Braun
and Clarke 2006). The first author coded and arranged the data into
descriptive and analytical themes and sub themes; data and analysis
Table 1. Participants’ characteristics
Nurse Clinical officer Doctor
Number of participants 6 5 6
Hospital QECH: n¼3 QECH: n¼2 QECH: n¼6
Chiradzulu: n¼3 Chiradzulu: n¼3 Chiradzulu: n¼0 (No doctors from
Chiradzulu were interviewed, as they
are rarely involved in PAC)
Gender Female: n¼4 Female: n¼2 Female: n¼2
Male: n¼2 Male: n¼3 Male: n¼4
Seniority Nurse-midwife technician: n¼3
Senior nurse: n¼3
Clinical ofﬁcer (independent practitioner
with 3 years training): n¼3
Senior clinical ofﬁcer: n¼2
Junior intern (doctor in ﬁrst year follow-
ing medical school): n¼1
Senior intern (second year doctor): n¼2
Registrar (doctor post-internship special-
izing in obstetrics and gynaecology):
Length of time in
1 month–15 years (<1 year: n¼3,
2–5 years: n¼2, 15 years: n¼1)
2 months–15 years (<1 year: n¼2, 1–7
years: n¼2, 15 years: n¼1)
2 months–5 years (<1 year: n¼3, 1–5
Health Policy and Planning, 2017, Vol. 32, No. 3 307
were discussed with the second author and compared with other
similar studies to enhance credibility and trustworthiness (Yin
2015). The analysis used a mixed deductive and inductive approach;
some codes (e.g. ‘monitoring and feedback’) were derived from
Greenhalgh et al.’s (2004) framework, whereas others emerged from
Ethical approval was granted from both Queen Margaret
University, Edinburgh and the College of Medicine Research and
Ethics Commitee, Malawi.
The emergent themes are discussed later and displayed in the result-
ant conceptual framework (Figure 1), which builds on Greenhalgh
et al.’s (2004) framework. We categorized findings into four themes:
the system, relationships, health workers’ values and attitudes and
the innovation. The outer context impacts on all these areas; all
these factors together influence the provision of PAC and MVA
usage. One participant described this interaction of interdependent
factors, all influenced by the outer context: ‘it’s a very complicated
issue here’ (P.10, clinical officer, Chiradzulu).
Greenhalgh et al. (2004) emphasize that numerous system features,
including hardware and organizational aspects, are key to the adop-
tion, diffusion and sustained use of innovations. Participants high-
lighted a number of such features as key barriers to MVA use. They
described the challenges of working with limited resources generally
but emphasized MVA instruments are a particular and increasing
problem. After initial donations, there has not been a reliable supply
to replace old and broken equipment.
So the problem we have is these things start as projects, some-
body develops an idea, after possibly looking at how successful it
was somewhere, but then comes incorporated, part of the system,
but how to continue without that capacity, it becomes a chal-
lenge. So that project goes away, there’s no supply of equipment,
so it goes naturally to extinction. (P.4, intern, QECH)
Participants explained that in the absence of MVA equipment, they
would perform D&C rather than delay the procedure. This ap-
peared to sometimes result in health workers automatically resorting
to D&C; they became used to there being no equipment.
In addition to shortage of physical resources, participants
emphasized the problem of staff shortages, particularly staff trained
in PAC. Patients requiring PAC are admitted to the gynaecology ward
in QECH and the female ward in Chiradzulu. Nurse shortages were
observed in both these wards, with often only one or two nurses cov-
ering up to 70 patients. Participants explained how staff shortages
lead to staff feeling overworked and demotivated, thus potentially af-
fecting quality of care. Participants noted that unsafe abortions con-
tribute to a large proportion of gynaecology admissions and some
thought the number of women presenting for PAC may be increasing,
which may be compounding feelings of being overworked.
Staff shortages may also affect MVA use. Nurse shortages were
felt to contribute to nurses trained in MVA not performing it as they
were engaged with other tasks.
We nurses we are in a shortage. So most of time we are concen-
trated in the ward, doing rounds, doing other things. While these
Figure 1. Conceptual framework developed from Greenhalgh et al.’s (2004) framework and the study’s ﬁndings
308 Health Policy and Planning, 2017, Vol. 32, No. 3
interns are assigned to do the assessment of the abortion patients.
So we just leave them to do. (P.1, nurse, QECH)
The junior interns, however, receive little of the practical on-the-job
training they require, again due to staff shortages.
So you find as an intern, it is put on the schedule that you are
going to do an evacuation, but there is nobody to guide... So the
junior, all he has in mind is theoretically this is how we do, in
terms MVA. (P.4, intern, QECH)
Doctors expressed concerns about the quality of on-the-job training;
interns might pass weaknesses down to juniors. Furthermore, senior
interns usually train juniors only in D&C, as they tend to be more
confident in D&C than MVA. Participants said this is leading to
MVA becoming a ‘forgotten skill’, as the next cohort feel even less
confident and are less likely to train others. However, in
Chiradzulu, clinical officers described good on-the-job training and
support. Some nurses had received MVA training, although often
some years ago, during a national programme. Participants from
both hospitals felt it important to have ongoing practical supervision
or regular refresher courses both to train more people and maintain
skills of those already trained.
A lack of follow-up and feedback were highlighted as additional
organizational issues. At QECH, women are not followed-up after
PAC, whereas at Chiradzulu they are advised to return after 2
weeks. Participants at QECH were concerned that without routine
follow-up, they would not know whether women experienced prob-
lems after treatment, and whether ‘what we have done is OK’.(P.4,
intern, QECH). Absence of feedback may limit the use of relatively
new procedures like MVA, about which staff may feel less secure.
Furthermore, a lack of monitoring of health workers’ perform-
ance was felt to contribute to poor care and low MVA use. One
nurse described how performance and MVA use was better when an
non-governmental organization was monitoring their performance
but when this ceased MVA use declined.
In the first maybe because we were being sponsored by
JHIEPEGO and what and they were supervising us... and they
were check listing, so people were doing it. (P.1, nurse, QECH)
Participants also felt that monitoring MVA rates is beneficial and
suggested that MVA usage increased in Chiradzulu following the
dissemination of results demonstrating it had the lowest usage be-
tween 2008 and 2012 (Odland et al. 2014). They attributed this to
appointing a PAC coordinator who organizes a MVA staff rota and
monitors supplies. Thus, leadership may be another organizational
factor affecting MVA rates.
According to Greenhalgh et al. (2004), social networks, that is net-
works of communication and influence between colleagues and
friends, influence the diffusion of innovations. Thus, staff relation-
ships and dynamics need to be examined. At both hospitals, staff
have distinct and set roles and responsibilities. Junior interns pro-
vide PAC in QECH; senior interns described being ‘off the hook’
after completion of their junior internship. Nurses provide initial
and post-procedural care in QECH. Nurses perform the same role at
Chiradzulu but clinical officers and occasionally senior nurses con-
duct procedures. However, nurses perform MVA rarely and may
consider it as the clinicans’ duty, as one nurse indicated.
If we are given a chance to do it [MVA], we can do it, but accord-
ing to scope of practice in Malawi, it is the duty of the clinician.
(P.14, nurse, Chiradzulu)
On the other hand, one doctor described that following wide-
spread MVA training for nurses, doctors started perceiving MVA
as a nurse’s responsibility and D&C as a doctor’s responsibility.
Clear demarcations between nurses and doctors in terms of role
and status may have contributed to doctors discontinuing MVA,
which became seen as a simple innovation of lower standing, de-
signed for nurses.
Strictly delineated roles and responsibilities may also negatively
impact on patient care and teamwork if health workers do not con-
sider a particular task or patient as their responsibility, as one intern
we don’t really work as a team... It’s like when you work, you
do say ‘it’s OK, as long as I do this then it’s fine. If something
happens [to the patient], it won’t be on my side now, I have
cleared on my side, and then they are on that other side’.
The intern points to a distinction between ‘my side’, for which one is
responsible and ‘the other side’, for which one is not. The same intern
explained if a doctor is unable to find the correct senior doctor to help
them, others may refuse, as they do not consider it their responsibility.
So I asked them to assist, they said ‘no I am not on-call, you
need to inform the ones on-call’. I said ‘the patient is not well,
can you just assist?’ They said ‘no I can’t assist’. (P.6, intern,
Whilst poor teamwork will affect the quality of care in general, we
can again see how it may particularly affect use of MVA, a relatively
unknown procedure about which the staff feel less confident.
Collaboration will be further impeded by strong hierarchal power
structures between cadres which were apparent in both hospitals but
particularly QECH. Nurses are perceived to have less power than
The powers are mainly with the College of Medicine as com-
pared with the College of Nursing. So there’s an issue with super-
iority with that one. And the ones who are in charge of the
wards, they are the doctors, the consultants, so they have author-
ity. (P.4, intern, QECH)
This power differential also comes through in the description later; a
senior nurse described how she tries to help and train junior interns
in MVA when they are having difficulties.
If there is a problem then I come in and maybe just suggestion,
you can do this one, so then they do MVA. (P.1, nurse, QECH)
The use of the phrase ‘maybe just suggestion’ points to a lack of
assertiveness and power differentials, despite the nurse having
more experience than junior doctors and displays how MVA use
might be particularly hampered by these power differentials be-
tween cadres if nurses feel unable to speak up. Power differen-
tials between cadres also appeared to influence participants’
views on who should ideally perform MVA. Almost all partici-
pants agreed nurses, clinical officers and doctors should perform
MVA, but they described uterine evacuation as being primarily a
clinical officer or doctor’s responsibility, with nurses only provid-
ing services when doctors or clinical officers are unavailable to
Participants also described a strong hierarchy amongst doctors.
This appears to contribute to juniors’ fears to ask for help, as they
do not want to admit a lack of knowledge or are scared of being
Yeah so instead of teaching, they just shout at you as if, they just
put all the blame on you as if you have completed nothing, so,
Health Policy and Planning, 2017, Vol. 32, No. 3 309
anyway we are used [to it], because even when we are at [med-
ical] school it’s just the same things. (P.4, intern, QECH)
If MVA is expected to be carried out by relatively inexperienced
staff, hierarchical relationships and fears about being scolded rather
than supported may make staff less inclined to carry out MVA.
Health workers’ values and attitudes
Greenhalgh et al. (2004) note that the adopters’ needs, values and
goals matter for diffusion of innovations. They will be especially im-
portant for PAC related innovations given that abortion is a particu-
larly value-laden issue. Participants displayed a range of attitudes
towards providing PAC. They all agreed it should be an emergency
service and patients should be prioritized as they require urgent
treatment. However, participants held differing opinions on whether
PAC is actually prioritized in their hospitals. Some felt it is, as reduc-
ing maternal mortality is a national priority. However, others said
often PAC is not treated as a priority emergency service, resulting in
They are kind of looked at on the lower side of care...emergency
obstetrics is taken like more of a priority than postabortal care.
That’s why even if, that’s why a lady who has signs of miscar-
riage or signs of abortion would come at night but then they
wouldn’t be attended to, but somebody in labour would be at-
tended to there and then. (P.10, clinical officer, QECH)
This suggests that outside the interview context, providers may at-
tribute less importance to PAC. Some health workers described
PAC as a ‘normal duty’, no different from other healthcare ser-
vices, others were very positive and passionate about providing
PAC. One nurse explained PAC was her main motivation to work
in gynaecology. Nurses appeared more enthusiastic about provid-
ing PAC than clinical officers and doctors, reflected in using
words such as ‘passion’, ‘great’ and ‘satisfaction’. Many health
workers found PAC satisfying as they considered it necessary and
life-saving care, and it seemed to provide them with heroic
So I was just like: we saved her life. (P.3, nurse, QECH)
Furthermore, some participants described positive feelings about
helping and providing advice to women who they considered vulner-
able, such as teenagers. Some health workers also described feelings
of sadness or sympathy, towards both women who had spontaneous
and induced abortions, who may have a lack of support and long-
term complications, particularly infertility.
However, many health workers described finding PAC challeng-
ing due to their personal belief that induced abortion is immoral.
These feelings were influenced by religious beliefs and prevailing
community beliefs and norms. Participants explained that PAC is
often only associated with induced abortions by the public and there
tends to be negative attitudes towards these women, with people
considering them ‘sinners’ or ‘criminals’; these words and views
were often reflected by health workers. Many participants described
other health workers exhibiting discriminatory or judgemental atti-
tudes towards patients whom they suspected had induced abortions,
particularly in mission hospitals and amongst health workers who
don’t work directly in PAC.
Participants stated that this stigma surrounding abortion often
leads to women delaying presenting for care or not disclosing
induced abortions (or even pregnancies or miscarriages), resulting
in health workers feeling frustrated with these women. Staff short-
ages and perceived increasing rates of women presenting for PAC
may increase negative feelings towards these patients, who they
feel are increasing already heavy workloads by ‘opting’ to induce
They [staff working in Outpatient Department] can shout at
them ‘why did you do this? You are creating a lot of job here,
you want us to be helping you this month and then next month
you are coming a lot of you! You come here with a lot of induced
abortion!’ (P.12, clinical officer, Chiradzulu)
A nurse also noted other staff feel PAC is facilitating abortions.
Although she did not agree, an intern did and therefore felt very
negatively towards PAC.
People are not afraid of getting pregnant, knowing that if they do
the abortion, someone at the hospital is ready to treat them. (P.5,
It is important to note, however, that negative views of abortion
could go together with a perceived obligation to provide care for all
I believe that doing a abortion is a sin myself, but just because I
believe that it is a sin, it’s not like I say ‘because you did an abor-
tion I’m not going to help you’, I’m supposed to help everybody.
(P.5, intern, QECH)
One participant described this as creating an internal conflict of val-
ues, between their moral or religious views and the care they pro-
vide. Most participants felt their beliefs did not affect their care but
one thought it might.
Knowledge and attitudes towards the innovation (MVA) are con-
sidered highly important for adoption and diffusion (Rogers 2003).
Although the literature has identified various advantages of MVA
over D&C, Greenhalgh et al. (2004) point out that perceived advan-
tages are most important for the uptake of innovations, which are
not fixed and open to negotiation and adjustment.
All participants noted MVA is the first-line treatment for first-
trimester incomplete abortions and curettage should be used for
higher gestations. Some staff members felt for this reason, curettage
will continue to be used more often than MVA, as they perceived
second-trimester abortions to be more common. However, a clinical
officer suggested that some practitioners struggle to determine preg-
nancy gestations and when in doubt, opt for curettage.
Numerous advantages of MVA over curettage were expressed;
most commonly that it is associated with fewer complications.
We can do MVA patient with minimal complications, minimal.
And MVA we cannot perforate unless you are careless. (P.1,
Other advantages cited were that it is cheaper, simpler and quicker
to perform, requires less staff and allows for rapid discharge of pa-
tients resulting in fewer in-patients. Staff noting these advantages
were those mainly responsible for patient care post-procedure:
nurses in QECH and clinical officers in Chiradzulu; these cadres will
feel this benefit more directly to their personal workload.
Yet, many health workers noted that curettage is used more com-
monly in their hospital than MVA. Participants often referred to
other staff preferring curettage; only one participant stated he per-
sonally preferred curettage, as from his experience MVA seems to
have a higher risk of incomplete procedures; a concern that some
other participants shared. Participants suggested junior staff might
prefer curettage as they are more familiar with it and are worried
310 Health Policy and Planning, 2017, Vol. 32, No. 3
about making mistakes or performing incomplete procedures and
disappointing their seniors.
so people are trying to impress and as a result they tend to do
something that they are comfortable with rather than the thing
that they are supposed to do. (P.6, intern, QECH)
Paradoxically, MVA’s perceived simplicity entails a risk; if compli-
cations do occur, they may be attributed more easily to a lack of
competence or being ‘careless’.
It appears, then, that health workers’ attitudes towards using
MVA are affected by perceived personal benefits or risks of MVA,
whether reducing inpatients and workload, becoming personally re-
sponsible for an incomplete procedure or complication or incurring
colleagues’ negative views of one’s professional capacity.
Our findings demonstrate how multiple ‘hardware’ and ‘software’
issues interact and reinforce each other to influence MVA use.
Greenhalgh et al. (2004) stress that few studies consider this inter-
connectedness. The wider context of PAC and abortion influences
many factors contributing to low MVA usage. PAC has been pro-
moted as the least contentious service for reducing unsafe abortions’
adverse consequences (Barot 2014), but it cannot be separated from
the socio-cultural, religious and legal status of abortion. Negative at-
titudes surrounding abortion may lead to delays; women who delay
seeking care are at increased risk of more severe complications such
as sepsis, resulting in being more likely to require further procedures
in theatre rather than MVA as an outpatient. Negative attitudes
may also result in providers’ negative attitudes towards providing
PAC, low prioritization of services and low motivation to improve
services. The idea that large numbers of women choosing to induce
abortions are increasing already heavy workloads may further re-
duce prioritization and motivation to improve PAC. Practical diffi-
culties, such as unreliable equipment supplies, make MVA a more
difficult option than curettage; when motivation and prioritization
are low, staff will likely opt for the easier option, which may not be
what is best for the patient.
Staff shortages, relationships, expected job roles and power dy-
namics appear very important in this context and impact on many
elements. Although social interactions are being increasingly
recognized as important for innovation use (Cain and Mittman
2002), Greenhalgh et al. (2004) note there is a dearth of research
investigating this. The findings of this study support McPake and
Koblinsky’s (2009) and Gilson’s (2012) assertion that health systems
are social in nature; people and their views, behaviours and inter-
actions shape how policies recommending innovations are imple-
mented. Staff relationships impact on training and supervision and
power dynamics may affect staff attitudes towards MVA and their
willingness to use it. Despite widespread acknowledgement of MVA
policies and its benefits, the personal risks and benefits of using
MVA appeared more important. Greenhalgh et al. (2004) emphasize
that it is important that those with more power, such as purchasers,
senior clinicians and managers, see the benefits of creating an ena-
bling environment for staff to use it. Graff and Amoyaw (2009)
found that a lack of prioritizaiton of PAC in Ghana led to purchas-
ing equipment for other services rather than MVA when funds were
limited. Moreover, purchasers may not appreciate MVA’s advan-
tages as equipment is more expensive and does not last as long as
curettage equipment (Zaidi et al. 2014), although it is still more
cost-effective as it can be performed as an outpatient procedure.
Our findings highlight the value of using Greenhalgh et al.’s
(2004) framework as opposed to more individualistic behavioural
theories; rejection of MVA is clearly not only a choice based merely
on individual knowledge and attitudes but also influenced by staff
relationships and broader contextual and organizational issues.
Furthermore, this study adds to the existing literature on diffusion
of innovations. First, this study examines issues from health work-
ers’ viewpoints, which are key as providers’ perceptions inform and
shape practice (Blaauw et al. 2006;Aniteye and Mayhew 2013).
Furthermore, Greenhalgh et al.’s (2004) framework is based largely
on studies conducted in high-income settings. We demonstrate its
applicability to LMIC but here certain factors such as limited mater-
ial and human resources will carry more weight. This study high-
lighted the strong importance that power dynamics can play in
diffusion of innovations. The importance of staff roles and power
dynamics between staff cadres appeared to come through in another
study of MVA usage in another LMIC, with some doctors feeling
midwives always require supervision when performing MVA, des-
pite policies and training to allow for tasking-shifting of MVA to
midwives (Paul et al. 2014). It is therefore possible that whilst power
differentials will affect the provision of care everywhere
(Greenhalgh et al. 2004), they may affect diffusion of innovations
more in LMIC settings but this requires more research.
Furthermore, the study emphasizes the difficulty and complexities of
sustaining the adoption of an innovation, which Greenhalgh et al.
(2004) highlight has been a neglected research area. Rogers (2003)
theorizes that once adoption of an innovation reaches almost 100%,
it will be self-sustaining, as it becomes ‘a normal and taken-for-
granted way of working’ (McEvoy et al. 2014, p. 3). This process is
termed routinization (Greenhalgh et al. 2004), institutionalization
(Billings et al. 2007) or embedding (May et al. 2007). However,
MVA use became briefly widespread in Malawi but then decreased
as facilities struggled to maintain skills and supplies. Routinization
may be less applicable to LMIC health systems, which may be sus-
ceptible to externally driven changes such as finance or supply chain
difficulties. Hence, it is equally, if not only more important, particu-
larly in LMIC, to consider both how to achieve widespread adoption
of innovations but also how to sustain use.
Limitations and future research
This study adds to a limited research base about what influences re-
jection of recommended PAC innovations. However, there are limi-
tations. Some findings may partially be a by-product of the
interview context; e.g. participants may have over-emphasized the
importance of resource issues in the hope the interviewer could help
the situation. Furthermore, as this is a small-scale study involving
only two hospitals, transferability to other situations may be limited.
Greenhalgh et al. (2004), however, note that generally insights re-
garding diffusion of innovations cannot be readily transferred to
other contexts due to the importance of the ‘fit’ between the innov-
ation and the local system or context. Whilst the study provides
some general insights into factors that contribute to rejection of rec-
ommended innovations, it is recognized that MVA and PAC may be
a special case and involves moral sensitivities that other innovations
and policies would not. Thus, we need to be careful with assuming
insights could apply to other health innovations.
A particular limitation in the context of PAC in Malawi is that
no urban or rural health centres were included. As they are often the
first contact for women seeking PAC, yet generally suffer from even
worse resource shortages, further research involving health centres
would be beneficial. Longitudinal ethnographic research could also
Health Policy and Planning, 2017, Vol. 32, No. 3 311
provide further insights and may be particularly useful for under-
standing how best to sustain innovation use in different situations,
as how innovations work out in different contexts is never predict-
able (McPake and Koblinsky 2009).
Implications and recommendations
In spite of the aforementioned limitations, we can make some tenta-
tive recommendations. Recognizing the danger of claims to context-
free knowledge and one-size-fit-all solutions (Sheikh et al. 2014),
these may only be relevant to Malawi and similar LMIC.
Participants emphasized reliable, sustainable equipment supplies
and practical, regular training are necessary to increase MVA use.
Simpson (2005) argues that local ownership of programmes with a
bottom-up approach, i.e. health workers creating strategies them-
selves, improves sustainability. It is also important that practitioners
are involved in equipment purchasing decisions, as they understand
the cost-effectiveness and importance of MVA. On-the-job training
promotes adoption of innovations (Cain and Mittman 2002), but
the current system for doctors in Malawi appears inadequate, and
nurses may not receive on-the-job training at all. Training desig-
nated staff members in teaching others to provide on-the-job train-
ing and supervision may help (RamaRao et al. 2011). However,
since trained nurses currently rarely perform MVA, it is possible
training more nurses may not improve matters.
There is a growing recognition that for successful, sustainable
health programmes whole health systems need to be strengthened,
including all six WHO building blocks (service delivery, human re-
sources, information, medicines and technologies, financing, govern-
ance) (Simpson 2005;Billings et al. 2007;Gilson 2012). This
research has indicated that effective and sustainable PAC policy
likely requires a broader people-centred health systems approach.
Evidence from Hunduras shows that providing MVA equipment
and training alone is unlikely to increase use; Chincilla et al. (2014)
suggest other system factors and human resource issues needed to be
addressed. These include shortages of health workers; a significant
problem in Malawi, particularly in rural areas (WHO 2014), which
have worse and higher rates of abortion complications (Kalilani-
Phiri et al. 2015). However, training and staff shortages are not the
only human resource issues, addressing staff relationships and
power dynamics that negatively impact on PAC and MVA usage is
equally important. Regular team meetings can improve communica-
tion between cadres, team-working and performance (Firth-Colins
2001;Borril and West 2003). A doctor remarked that in QECH
daily team meetings used to occur but stopped due to staff shortages.
Recommencing these may be beneficial but only if different cadres
feel respected and safe to participate (Firth-Colins 2001). Team
meetings may also encourage support and recognition of junior staff
members, which might improve motivation and job satisfaction
(McAuliffe et al. 2009) but only if managers have sufficient training
and awareness about supporting junior staff. Finally, it is important
to consider negative attitudes towards, and low prioritization of,
PAC. Strengthening regular training for all staff may improve staff
attitudes towards women requiring PAC (Shah et al. 2014). Building
on the satisfaction that some staff described from helping vulnerable
women and saving lives, and their recognition of the complexity of
the lives of many women with induced abortions may decrease
judgements, improve care and increase MVA use through increasing
motivation to adopt the innovation that is considered best practice,
even if it is not always the easiest option. However, attempting to
alter attitudes and behaviours, and indeed to improve MVA use and
PAC provision more generally, in a context where abortion is illegal,
highly stigmatized and compounded by multiple gender and power
dynamics will remain ‘complicated’.
Prof. Jon Øyvind Odland, Dr. Ursula Kafulafula, Dr. Gladys Gadama and
Dr. Frank Taulo for their help and support with data collection.
This work was supported by Queen Margaret University’s student develop-
ment fund and the Institute of International Health and Development.
Ethical approval: Ethical approval was granted from both Queen Margaret
University (QMU) and the Malawi College of Medicine Research and Ethics
Conﬂict of interest statement. None declared.
hman E, Shah IH. 2011. Review article: new estimates and trends regarding
unsafe abortion mortality. International Journal of Gynecology and
Obstetrics 115: 121–6.
Aniteye P, Mayhew SH. 2013. Shaping legal abortion provision in Ghana:
using policy theory to understand provider-related obstacles to policy imple-
mentation. Health Research Policy and Systems 11: 1–14.
Barot S. 2014. Implementing PAC programs in the developing world: ongoing
challenges. Guttmacher Policy Review 17: 22–8.
Benson J, Gebreselassie H, Ma~
nibo MA et al. 2015. Costs of postabortion care
in public sector health facilities in Malawi: a cross-sectional survey. BMC
Health Services Research 17:1–7.
Billings DL, Crane BB, Benson J, Solo J, Fetters T. 2007. Scaling-up a public
health innovation: a comparative study of post-abortion care in Bolivia and
Mexico. Social Science & Medicine 64: 2210–22.
Blaauw D, Ambegaokar M, Penn-Kekana L, Hongoro C, McPake B. 2006.
Neither robots nor angels: the ‘dynamic responses’ of health workers and
the unintended effects on health systems functioning London: Health
Systems Development Programme Working Paper, DFID. http://r4d.dﬁd.
gov.uk/Output/179693/, accessed 21 February 2015.
Borril C, West M. 2003. Team Working and Effectiveness in Health Care:
Findings from the Health Care Team Effectiveness Project. Birmingham:
Aston Centre for Health Service Organisation Research. http://homepages.
inf.ed.ac.uk/jeanc/DOH-glossy-brochure.pdf, accessed 18 July 2015.
Bowie C, Geubbels E. 2013. Chapter 5 - epidemiology of maternal mortality
in Malawi. In: Bowie C and Geubbels E.(ed).The Epidemiology of Malawi,
2nd edn. Malawi: Department of Community Health, College of Medicine,
Bowling A, Ebrahim S. 2005. Handbook of Health Research Methods:
Investigation, Measurement and Analysis. Maidenhead: McGraw-Hill
Braun V, Clarke V. 2006. Using thematic analysis in psychology. Qualitative
Research in Psychology 3: 77–101.
Cain M, Mittman R. 2002. Diffusion of Innovation in Healthcare. Oakland,
CA: California Healthcare Foundation.
Chikuse L. 2015. Malawi says no to abortion. Malawi 24. 26 February 2015.
neMF9qAR.dpuf, accessed 8 July 2015.
Chincilla AL, Flores IF, Morales AF, de Gil MP. 2014. FIGO initiative:
changes in the use of manual vacuum aspiration for PAC within the public
healthcare service network in Honduras. International Journal of
Gynecology and Obstetrics 126: S24–7.
Corbett MR, Turner KL. 2003. Essential elements of PAC: origins, evolution
and future directions. International Family Planning Perspectives 29:
Descher A, Cohen SA. 2003. Contraceptive use is key to reducing abortion
worldwide. The Guttmacher Report on Public Policy 6: 7–10.
312 Health Policy and Planning, 2017, Vol. 32, No. 3
Firth-Colins J. 2001. Cultures for improving patient safety through learning:
role of teamwork. Quality and Safety in Health Care 10: 26–31.
Gilson L. 2012. Introduction to health policy and systems research. In: Gilson
L (ed). Health Policy and Systems Research: A Methodology Reader.
Geneva: World Health Organization, 19–38.
Godin G, Belanger-Gravel A, Eccles M, Grimshaw J. 2008. Healthcare profes-
sionals’ intentions and behaviours: a systematic review of studies based on
social cognitive theories. Implementation Science 3: 3–12.
Graff M, Amoyaw DA. 2009. Barriers to sustainable MVA supply in Ghana:
challenges for the low-volume, low-income providers. African Journal of
Reproductive Health 13: 73–80.
Greenhalgh T, Robert G, Macfarlane F, Bate P, Kyriakidou O. 2004.
Diffusion of innovations in service organizations: systematic review and rec-
ommendations. Milbank Quarterly 82: 581–629.
Greenslade FC, Mckay H, Wolf M, McLaurin K. 1994. Post-abortion care: a
women’s health initiative to combat unsafe abortion. Advances in Abortion
Grimes DA. 2006. Reducing the complications of unsafe abortion: the role of
medical technology. In: Warriner K and Shah H (eds). Preventing Unsafe
Abortion and its Consequences: Priorities for Research and Action. New
York: Guttmacher Institute. 73–92.
Grimes DA, Benson J, Singh S et al. 2006. Unsafe abortion: the preventable
pandemic. Lancet 368: 1908–19.
Ipas. 2011. Ensuring Women’s Access to Safe Abortion: A Key Strategy for
Achieving Millennium Development Goals. Chapel Hill, NC: Ipas. http://
accessed 31 March 2015.
Jackson E, Johnson BR, Gebreselassie H, Kangaude GD, Mhango C. 2011. A
strategic assessment of unsafe abortion in Malawi. Reproductive Health
Matters 19: 133–43.
Johnston HB, Gallo MF, Benson J. 2007. Reducing the costs to health systems
of unsafe abortion: a comparison of four strategies. Journal of Family
Planning and Reproductive Health Care 33: 250–7.
Kalilani-Phiri L, Gebreselassie H, Lavandowski BA et al. 2015. Clinical article:
the severity of abortion complications in Malawi. International Journal of
Gynecology and Obstetrics 128: 160–4.
Kulczycki A. 2009. Abortion and PAC. In: Ehiri J and Kulczycki A (eds).
Maternal and Child Health: Global Challenges, Programs, and Policies.
New York: Springer US, 191–201.
Leme VM, Mtimavalye LA, Thole GC, Mvula MT. 1997. The impact of the
manual vacuum aspiration (MVA) technique on health care services at
Queen Elizabeth Central Teaching Hospital, Blantyre, Malawi. South
African Medical Journal 87: 218–22.
Levandowski BA, Mhango C, Kuchingale E et al. 2013. The incidence of
induced abortion in Malawi. International Perspectives on Sexual &
Reproductive Health 39: 88–96.
Maonei C, Miot J, Moodley S. 2014. The cost-effectiveness of introducing
manual vacuum aspiration compared with dilatation and curettage for in-
complete ﬁrst trimester miscarriages at a tertiary hospital in Manzini,
Swaziland. South African Journal of Obstetrics & Gynaecology 20: 27–30.
May C, Finch T, Mair F et al. 2007. Understanding the implementation of
complex interventions in health care: the normalization process model.
BMC Health Services Research 7: 148.
McAuliffe E, Manafa O, Maseko F, Bowie C, White E. 2009. Understanding
job satisfaction amongst mid-level cadres in Malawi: the contribution of or-
ganisational justice. Reproductive Health Matters 17: 80–90.
McEvoy R, Ballini L, Maltoni S et al. 2014. A qualitative systematic review of
studies using the normalization process theory to research implementation
processes. Implementation Science 9: 1–13.
McPake B, Koblinsky M. 2009. Improving maternal survival in South Asia–
what can we learn from case studies? Journal of Health, Population and
Nutrition 27: 93–107.
Ministry of Health. 2009. Malawi Sexual and Reproductive Health and Rights
(SRHR) Policy. Lilongwe: Ministry of Health, April. http://unfpamalawi.
ductive-health-policy.html, accessed 29 June 2015.
Odland ML, Rasm ussen H, Jacobsen GW et al. 2014. Decrease in use of man-
ual vacuum aspiration in PAC in Malawi: a cross-sectional study from three
public hospitals, 2008-2012. PLoS One 9: 1–7.
Pattinson RC, Snyman LC, MacDonald AP. 2006. Evaluation of a strict proto-
col approach in managing women with severe disease due to abortion.
South African Medical Journal 96: 1191–4.
Paul M, Gemzell-Danielsson K, Klingberg-Allvin M, Kiggundu C, Namugenyi
R. 2014. Barriers and facilitators in the provision of post-abortion care
at district level in central Uganda - a qualitative study focusing on task shar-
ing between physicians and midwives. BMC Health Services Research 14:
Ramarao S, Townsend JW, Diop N, Raifman S. 2011. PAC: going to
scale. International Perspectives on Sexual & Reproductive Health 37:
Rogers EM. 2003. Diffusion of Innovations, 5th edn. New York: Free Press.
Schenck-Ygleslae C. 2004. Technical Report: Evaluation of PAC Service
Delivery in Malawi. Lewison D. (ed). Jhpiego Corporation, March.
Shah IH, A
˚hman E, Ortayli N. 2014. Review article: access to safe abortion:
progress and challenges since the 1994 International Conference on
Population and Development (ICPD). Contraception 90: S39–48.
Sheikh K, George A, Gilson L. 2014. People-centred science: strengthening the
practice of health policy and systems research. Health Research Policy and
Systems 12. 1–8.
Simpson L. 2005. Community informatics and sustainability: why social cap-
ital matters. The Journal of Community Informatics 1: 102–19.
Storeng KT, Ouattara F. 2014. The politics of unsafe abortion in Burkina
Faso: the interface of local norms and global public health practice. Global
Public Health 9: 946–59.
Tagoe-Darko E. 2013. “Fear, shame and embarrassment”: the stigma factor in
post abortion care at Komfo Anokye Teaching Hospital, Kumasi, Ghana.
Asian Social Science 9: 134–41.
Tunc¸alp O, Gulmezoglu AM, Souza JP. 2010. Surgical procedures for evacuat-
ing incomplete miscarriage. The Cochrane Database of Systematic Reviews
WHO. 2007. The WHO Strategic Approach to Strengthening Sexual and
Reproductive Health Policies and Programmes. Geneva: World Health
pdf?ua¼1, accessed 9 July 2015.
WHO. 2012. Safe Abortion: Technical and Policy Guidance for Health
Systems, 2nd edn. Geneva: World Health Organization. http://apps.who.
int/iris/bitstream/10665/70914/1/9789241548434_eng.pdf, accessed 8 July
WHO. 2014. MDG 5: Improve Maternal Health. Geneva: World Health
maternal_health/en/, accessed 1 December 2015.
WHO, UNICEF, UNFPA, World Bank, UNPD. 2014. Trends in Maternal
Mortality: 1990 to 2013. Geneva: World Health Organization, May. http://
ity-2013/en/, accessed 8 July 2015.
World Bank. 2014. Health expenditure per capita (current US$). World
Bank. http://data.worldbank.org/indicator/SH.XPD.PCAP, accessed 22 July
Yin RK. 2015. Qualitative Research from Start to Finish, 2nd edn. New York:
Guilford Press Inc.
Zaidi S, Yasmin H, Hassan L et al. 2014. Replacement of dilation and curet-
tage/evacuation by manual vacuum aspiration and medical abortion, and
the introduction of postabortion contraception in Pakistan. International
Journal of Gynecology and Obstetrics 126: S40–4.
Health Policy and Planning, 2017, Vol. 32, No. 3 313