Medico-social pathways of obstetric fistula patients in Mali and Niger: an 18-month cohort follow-up

Authors


Corresponding Author Nathalie Maulet, Institute of Health and Society, Université Catholique de Louvain, Clos chapelle-aux-champs, 30 B1 30.16, 1200 Brussels, Belgium. Tel.: +32 (0)764 39 36; E-mail: nathalie.maulet@uclouvain.be

Abstract

Objectives

To gain understanding of care pathways and induced mobility patterns of obstetric fistula patients in French-speaking West African fistula repair centres.

Methods

We followed prospectively during 18 months a cohort of obstetric fistula patients in Mali and Niger (2008–2009). 120 patients were recruited at different stages of their care process in 5 reference fistula repair centres. Follow-up was carried out in repair centres and communities. Quantitative data were collected through close-ended questionnaires at three time points. Medico-social pathways were explored through a longitudinal analysis focusing on six indicators: fistula duration, care process duration, patients' mobility, marital status, number of surgeries and continence status.

Results

Patients' pathways were characterised by their length: fistula duration (median 4 years), care process duration (median 2.7 years), aggregated time spent in repair centres during study (median 7 months). Patients developed four mobility patterns (homebound, itinerant, institutionalised and urbanised). Adverse marital status change continued over time. Sample stratification according to number of previous surgeries revealed differences in care process duration and outcome: 23/31 new cases (≤1 surgery) gained continence with a mean of 1.5 surgeries in a median of 0.6 year while only 17/78 old cases (≥2 surgeries) became continent with a mean of 4 surgeries in a median time of 4.9 years.

Conclusion

The quest for continence does not end with admission to a fistula repair centre. Analysing fistula care experience across time within the varying settings highlights the twofold population and mixed medico-social outcomes that should prompt new development in obstetric fistula care management and research.

Abstract

Objectifs

Acquérir une compréhension sur les parcours de soins et les profils de mobilité induits, des patientes souffrant de fistules obstétricales dans des centres de traitement de la fistule en Afrique de l'ouest francophone.

Méthodes

Nous avons suivi prospectivement pendant 18 mois une cohorte de patientes avec une fistule obstétricale au Mali et au Niger (2008–2009). 120 patientes ont été recrutées à différents stades de leur processus de soins dans 5 centres de référence de la fistule. Un suivi a été effectué dans les centres de réparation et/ou dans les communautés. Les données quantitatives ont été recueillies par questionnaires fermés à 3 moments. Les parcours médico-sociaux ont été explorés au moyen d'une analyse longitudinale se concentrant sur 6 indicateurs: la durée de la fistule, la durée du processus de soins, la mobilité des patientes, l’état matrimonial, le nombre de chirurgies et le control de l'incontinence.

Résultats

Les parcours des patientes ont été caractérisés par leur longueur: durée de la fistule (médiane de 4 ans), durée du processus de soins (médiane de 2,7 ans), temps total passé dans des centres de réparation au cours de l’étude (médiane de 7 mois). Les patientes ont développé 4 modèles de mobilité (confiné à la maison, itinérant institutionnalisée et urbanisé). Des changements défavorables d’état civil ont continué au fil du temps. La stratification de l’échantillon selon le nombre de chirurgies antérieures a révélé des différences dans la durée du processus de soins et dans les résultats: 23/31 nouveaux cas (≤1 chirurgie) ont récupéré le control de l'incontinence après une moyenne de 1,5 interventions chirurgicales et une durée médiane de 0,6 année alors que seuls 17/78 cas anciens (≥2 chirurgies) ont retrouvé le control de l'incontinence après une moyenne de 4 chirurgies et une durée médiane de 4,9 ans.

Conclusion

La quête du control de l'incontinence ne s'arrête pas à l'admission dans un centre de traitement de la fistule. L'analyse de l'expérience dans les soins de la fistule au cours du temps dans les différents cadres, souligne les résultats doubles médico-sociaux mixtes et de population qui devraient inciter à de nouveaux développements dans la prise en charge des fistules obstétricales et dans la recherche.

Abstract

Objetivos

Entender la trayectoria de cuidados sanitarios y los patrones de movilidad inducida de pacientes con fístula obstétrica en centros de reparación de fístula en países francófonos de África Occidental

Métodos

Hemos realizado un seguimiento prospectivo durante 18 meses de una cohorte de pacientes con fístula obstétrica en Mali y Níger (2008–2009). Se reclutaron 120 pacientes en diferentes momentos de su proceso sanitario en 5 centros de referencia de reparación de fístula. El seguimiento se hizo en los centros de reparación y/o en la comunidad. Se recogieron datos cuantitativos en 3 momentos en el tiempo mediante cuestionarios de pregunta cerrada. Se exploró la trayectoria médico-social mediante un análisis longitudinal centrado en 6 indicadores: duración de la fístula, duración del proceso de cuidados, movilidad del paciente, estatus marital, número de cirugías y estatus de continencia.

Resultados

Las trayectorias de los pacientes se caracterizaron por su duración: duración de la fístula (media de 4 años), duración del proceso de cuidados (media 2.7 años), tiempo agregado gastado en los centros de reparación durante el estudio (media 7 meses). Los pacientes desarrollaron 4 patrones de movilidad (ligado al hogar, itinerante, institucionalizado y urbanita). El cambio del estatus marital adverso continuó a lo largo del tiempo. Una estratificación de las muestras según el número de cirugías previas reveló diferencias en la duración del proceso de cuidados y en el resultado: 23/31 nuevos casos (≤1 cirugía) consiguieron la continencia con una media de 1.5 cirugías en una media de 0.6 años mientras que solo 17/78 casos antiguos (≥2 cirugías) alcanzaron la continencia con una media de 4 cirugías en un tiempo medio de 4.9 años.

Conclusión

La búsqueda de continencia no termina con la admisión a un centro de reparación de fístula. El análisis de la experiencia de los cuidados de fístula, a lo largo del tiempo y en varios emplazamientos, pone de manifiesto la división de la población y resultados médico-sociales mixtos. La variedad de los resultados médico sociales que debería promover nuevos desarrollos en el manejo sanitario de y la investigación en la fístula obstétrica.

Introduction

Obstetric fistula is the generic name standing for vesico/recto-vaginal fistula caused by obstructed labour without timely skilled medical intervention. Besides chronic incontinence, obstetric fistula has a range of intertwined health and social consequences (Wall et al. 2005). Lack of evidence base and paucity of fistula data has long been deplored. For decades, scientific literature focused on fistula aetiology and surgical treatment (Wall et al. 2005; Kelly & Winter 2007). In 2003, an international campaign to end fistula helped escape circular epidemiology, boosting both research and care programs (Donnay & Ramsey 2006). Consequently, fistula knowledge base has broadened to new outcomes such as psychological status (Gharoro & Agholor 2009; Nielsen et al. 2009), quality of life (Pope et al. 2011; Umoiyoho et al. 2011) and post-fistula reproductive health (Bangser et al. 2011; Benfield et al. 2011; Wilson et al. 2011). Nevertheless, to date, understanding of fistula victims' harsh reality remains fragmented. Studies tend to hinge on singular events before or after repair. Very few authors have documented fistula patients' actual experience while in care (Ndiaye et al. 2009; Cam et al. 2010; Yeakey et al. 2011).

Getting rid of the preconceived idea that fistula repair comes down to one single short-lived care episode is a challenge. Indeed, after obstetric trauma and a stigmatising journey, admission to a fistula repair centre is seen as the near end of women's pain and torment. However, considering the estimated fistula backlog and surgeons' workload – especially in hospitals not exclusively dedicated to fistula surgery – the waiting period is usually counted in weeks and treatment completion can take several months (Wall et al. 2005). Several surgeries might be needed and not all women will gain continence nor have the opportunity to resume their former life. Our research aims to gain understanding of the medico-social pathways of obstetric fistula patients during their care process. In doing so, we highlight mixed prospects of care and ensuing specific needs.

Materials and methods

In the framework of a large mixed methods study on medico-social future ahead of obstetric fistula patients, a prospective follow-up of a 120 fistula patients' cohort was organised in Mali and Niger from February 2008 until November 2009. This paper presents the quantitative data collected through three repeated interviews by questionnaire applied verbally. It provides a quantitative analysis of obstetric fistula patients' subjective health and social experience.

Participants were recruited in the five hospitals recognised as reference fistula repair centres with a permanent in-house local fistula surgeon specialist (WHO 2006). In Mali, these were the Teaching Hospital of Point G/Bamako and the Regional Hospital of Somine Dolo/Mopti; in Niger, the National Hospital of Niamey, the National Hospital of Lamordé/Niamey and the Regional Maternity Hospital of Zinder.

Inclusion criteria were age (minimum 16), diagnosis (any type of obstetric fistula), language (Bambara/Fulfulde in Mali and Haussa/Zarma in Niger) and residence (Koulikouro, Segou and Mopti regions in Mali and Tillabery, Dosso, Maradi and West-Zinder regions in Niger). Participants were recruited regardless of their previous fistula care management: some just arrived with a recent fistula trauma; others had suffered for many years and/or had already experienced several surgeries.

Selection and first data collection (Phase A) were organised consecutively in the five fistula repair centres. Follow-ups took place, respectively, at 6 (Phase B) and 18 months (Phase C) after inclusion, either in health structures or at participants' homes when patient had already returned to her family or workplace. A set of questions on participants' family, housing, occupation and village with several crosschecking was prepared to that end. Periods of data collection are detailed in Table 1.

Table 1. Data collection periods
Data collectionaNigerMali
  1. a

    Mean time between Phase A and B was 182 days (SD: 27) and 323 days between Phase B and C (SD: 32).

  2. b

    One last participant was included or found back on mentioned date.

  3. c

    Due to the rain season, four participants were unreachable during the last follow-up and had to be visited later.

Phase A (N = 120)28/02/08–10/04/0819/05/08–21/07/08
Recruitment 28/08/08b
Phase B (N = 115)01/08/08–18/10/0811/11/08–08/01/09
+6 months23/12/08b 
Phase C (N = 109)07/07/09–18/08/0919/10/09–26/11/09
+18 months09/10/09–11/10/09c25/01/10b

All interviews were conducted by the authors with the help of two graduate local female assistants. The interviews lasted approximately 1 h included an informed consent procedure and were held in a private place at the best convenience of participant. Research was approved by Ministry of Public Health in Niger and University of Bamako's Ethical Committee in Mali. It was subsequently authorised by each surveyed health structure authorities. Our research exclusively deals with data provided by patients during interviews. We could not make use of clinical files due to their non-existence or incompleteness in some repair centres. An ad hoc questionnaire was produced and pre-tested by the research team. It was revised between each follow-up with minor additions aiming to confirm or detail specific information. The questionnaire comprised 25 close-ended questions focusing on basic socio-demographic and obstetrical variables, and on care and social pathways. Socio-demographical variables included ethnicity, age, literacy status, living area (rural/urban) and gravidity.

In our descriptive and non-clinical approach, care and social pathways refer to the changing life and care trajectories experienced by women during their fistula care process. Pathways were described through six patient-focused indicators emphasising complexity of overall trajectories development: (i) fistula total duration, (ii) care process duration, (iii) continence, (iv) number of surgical repairs, (v) marital status, (vi) mobility (Definitions in Box 1).

Continence was considered as the sole factor for repair success and recorded as a binary variable. The continence rate was linked to our new/old case ratio (Box 1) to present two indicators recommended by WHO: proportion of women continent after a first repair attempt and proportion of women operated on twice or more without success.

Participants who were lost to follow-up or deceased were excluded from the final analysis to enable comparison with the aim to highlight medico-social pathway evolution. Their profiles did not significantly differ from the other participants.

In accordance with our descriptive purpose, statistics were restricted to uni- and bivariate analysis. Differences in independent groups were checked using Pearson or Fisher's exact test for proportions and the Mann–Whitney test for distributions. Wilcoxon signed rank or binomial tests allowed comparisons between subsample and total sample. Mc Nemar chi² or Wilcoxon sign test was used to detect differences between 2 phases and the Friedman test or Cochrane Q test for comparison between 3 phases. Relative risks were used to determine associations. Confidence interval was fixed at 95%. Data were analysed using STATA 11.

Box 1. Definitions of medico-social pathways' indicators

IndicatorsDefinitions
  1. a

    As per patients own definition.

  2. b

    Cut-off ≥ 1 year was determined to exclude case of seasonal exodus usually lasting <1 year.

  3. c

    Our sample counted two cases of urinary diversion and two cases of fistula reoccurrence. In case cases of fistula reoccurrence at study inclusion, care pathway data solely started from the second event onwards.

Fistula total durationPeriod going from obstetric fistula occurrence to study inclusion
Care process durationPeriod starting with first admission to a fistula repair centre and ending with continence gain, or with a diagnosis of incurabilitya
Continence

The fact of ‘being dry’a.

A closed fistula without any residual or stress incontinence

Women with catheterContinence status was registered as unknown
Women with urinary diversionc Continence status was registered as incontinenta
Women with reoccurent fistulac When complete continence of obstetric fistula trauma was regained through surgery(es) in a reference fistula repair centre, and a subsequent pregnancy caused a second fistula or incontinence episode
Number of surgical repairsTotal number of obstetric fistula surgery endured by a woman after a fistula trauma
New/old cases ratioNumber of patients with ≤1 obstetric fistula surgery reported to number of patients with ≥2 surgeries at study baseline
Marital statusWomen's marital status including mention of polygamy
Marital abandonWomen without any sign of their husband during ≥1 yearb
MobilityAny change of place of stay for a minimum period of 1 month during the 18-month study period
Mobility patternsDeductive sorting of mobility according to participants' main place of stay (cut-off ≥9 months aggregated time period in the same place) and their number of movements (cut-off < 3 changes of place)

Results

All 324 patients admitted in the five fistula repair centres were contacted. Figure 1 details exclusion at recruitment, refusal to participate and dropout during study. Our initial sample of 120 patients was reduced to 109 participants due to 8 (7%) lost to follow-up and three deaths. Cause of death was investigated through relatives, health staff at fistula repair centre or local health centre and NGO in charge of women reinsertion. Despite source triangulation, two of the three deaths' causes remained doubtful (because of contradictory information) and a relation with fistula co-morbidity complications could not be excluded. The third death was due to HIV/AIDS opportunist tuberculosis. The deceased profile did not differ from the rest of the sample except for a significantly higher fistula duration (P = 0.009). All deceased had endured several surgeries (2, 4 and 13) but only one gained continence a few weeks before dying. Only one was married and 2/3 had lived several years in a fistula repair centre before settling in its neighbourhood.

Figure 1.

Recruitment process.

The number of participants visited in their communities was 27/109 (25%) at phase B and 65/109 (60%) at phase C; all others were found back in fistula wards. The 109 women followed during the whole study represented 17 ethnic groups. All but four came from a rural and generally remote environment. Cohort education level was very low: only 9/109 (8%) participants were literate.

The median fistula duration being 4 years (IQR: 8.50), the global profile at baseline data collection had significantly evolved (Table 2). Alterations of marital status illustrate how time affected social status. The most radical change was divorce immediately after incontinence diagnosis or later, when the care process took longer than expected, and irreversibility of incontinence was feared (11/73 women still married at recruitment were divorced during study).

Table 2. Socio-obstetrical profile, surgeries and continence status at fistula occurrence and during 18-month follow-up (n = 109)
OutcomesAt fistula occurrenceRecruitment – Phase A+6 months – Phase B+18 months – Phase C
n %Me (IQR) n %Me (IQR) n %Me (IQR) n %Me (IQR)
  1. a

    Median age at Phase A is significantly higher than at fistula occurrence (P = 0.000).

  2. b

    Marital status in phase B and C varies with participants getting married, divorced or widowed. Moreover, multiple changes also happened with the following patterns in A-B-C: 1 case of marriage-divorce-marriage, 2 case of divorce-marriage-divorce, 1 case of marriage-widowhood-marriage.

  3. c

    Median gravidity is significantly lower at trauma than at Phase A (P = 0.0000). Gravidity distributions are statistically different at each phase: From trauma occurrence to A (P = 0.002). From A to B (P = 0.000). From A to B (P = 0.000).

  4. d

    Cases of spontaneous fistula closure and continence recovery by early insertion of catheter without surgical intervention.

  5. e

    Thirty-one surgeries were performed in the 3 months before recruitment and resulted in nine continence gain.

Age (years)a
10–19494521 (11)141328 (12)131228 (11)131228 (13)
20–294440444045414440
30–391312312831283128
40–4933141314131514
≥5000666666
Marital statusb
Married10495 7367 6963 7367 
Divorced00282630282927
Widowed00335533
Single54555544
Gravidity c
Primigravida (1)64591 (3)41382 (4)41382 (4)36333 (4)
Gravida (2–4)2624373437344239
Gravida (≥5)1917312831283128
Number of surgeries
0/  13122 (2)2d23 (1)2d23 (2)
1/ 181723211817
2/ 252319171716
3–4/ 353242393734
≤5/ 181723213532
Incontinence status
Urinary10495 7367 6560 6560 
Urinary and faecal55443344
Unknown (Catheter)0023218700
Continent00 9e833304037

Another change in marital status is related to the common practice of polygamy in research countries: 22/76 participants remaining married after trauma had to accept an additional co-spouse in their household at fistula occurrence (three of them later became widows and never remarried) and 8/73 during study. Another way fistula affected marriage is by reducing it to social convenience only. Husbands of 10/73 married participants had absconded for more than 1 year (Me: 2, IQR: 1) in phase C.

Nevertheless, participants' marital prospects continued while in care: 13 women got married during the study, of whom seven were still incontinent and five still living in fistula repair centres or their neighbourhoods.

Median time between trauma and first contact with a fistula repair centre was 2 months (IQR: 11.96), with a minimum of 1 day and a maximum of 29 years. Our cohort consisted of 28% (31/109) new cases and 72% (78/109) old cases (new/old case ratio: 0.4). Consequently, at recruitment, a total of 299 surgeries (Me: 3, IQR: 2) had already been endured by our participants. An additional 101 surgeries were performed during the study on 65 participants.

Table 3 shows continence outcome of these 101 surgeries taking into account the number of surgeries endured before our study. The proportion of women continent after the first repair attempt was 75% (15/18). Twenty-three other participants needed surgery twice or more to gain continence. Two cases of spontaneous fistula closure by early insertion of catheter without surgical intervention were also registered.

Table 3. Final continence status disaggregated by number of surgeries before and during follow-upThumbnail image of

The new/old ratio among the 40 continent participants at the end of follow-up was 1.3 (23/17). New cases needed a mean of 1.5 surgeries (SD: 1.3) to gain continence while old cases needed significantly more surgeries (mean: 4, SD: 1.6, P = 0.0000). Continence gain was statistically associated with the number of fistula surgeries already undergone (P = 0.000). Old cases had a 3.4 times smaller chance to gain continence than new ones, and the total duration of care to gain continence was significantly higher (P = 0.0000) for old cases than for new ones: a median of 4.9 (IQR: 7.1) years vs. 0.6 years (IQR: 1.5). Up to 72% (82/109) of our cohort were operated on twice or more without success. Figure 2 highlights that most continence gain occurred within 5 years of care and with five or fewer surgeries. In view of this fact, 22% (19/86) of the old cases could be considered as deemed incurable.

Figure 2.

Relation between elapsed time since first admission in fistula repair centre and total number of surgeries according to continence status at the end of study. Continence is generally gained during a care process of maximum 5 years with a maximum of 5 surgeries (34/40 participants). The future lying ahead of the 35/69 incontinent women represented outside the grey rectangle raise questions concerning advisability of repeated surgeries and adapted medico-social care to be provided to deemed incurable patients.

Length of stay in a fistula repair centre and return home depends on a set of intertwined clinical, socio-economical and health system factors. Median time spent in a fistula repair centre during our study period was 7 months (IQR: 10). Stay duration was statistically shorter for patients gaining continence than for those who did not (P = 0.004) but duration did not differ between new and old patients (P = 0.2). Nine participants (all old cases, 8/9 incontinent, 2/9 deemed incurable) did not move at all from the centre during the whole 18 months. The 100 others all changed their place of stay from 1 to 7 times. A median of 2 (IQR: 2) movements per participant was registered.

Four mobility patterns emerged from monthly movement monitoring: (i) homebound, (ii) itinerant, (iii) institutionalised, (iv) urbanised. Table 4 highlights medico-social specificities associated with each pattern.

Table 4. Patients' profile per mobility pattern at end of study period (Phase C)
OutcomesAll (N = 109) Profile 1 Homebound (n = 42) Profile 2 Itinerant (n = 21) Profile 3 Institutionalised (n = 23) Profile 4 Urbanised (n = 23)
N Me (IQR) n Me (IQR) n Me (IQR) n Me (IQR) n Me (IQR)
  1. a

    At recruitment (Phase A).

  2. b

    Time elapsed since first contact with a reference fistula centre and study baseline.

Age (years) 28 (13) 26.50 (9) 30 (13) 30 (15) 30 (14)
Marital status
Married73 36 14 10 13 
Divorced2957107
Widowed31020
Single40013
Gravidity
Primigravida (1)363 (4)103 (4)111 (4)82 (3)73 (3)
Gravida (2–4)4217 2 1112
Gravida (≥5)3115 8 44
With living children5023 8 910
Number of surgeries
024 (3)12 (3)14 (2)04 (3)04 (2)
11812213
21710052
3–4371011610
≥53597118
New/old case (ratio)33/86 (0.38) 17/25 (0.68) 6/15 (0.40) 5/18 (0.27) 3/20 (0.15) 
Incontinence status
Urinary65 20 11 18 16 
Urinary & faecal40022
Continent40221035
Durationa (years)
Fistula 3.99 (8.47) 3.49 (4.98) 1.99 (6.98) 4.98 (7.98) 7.98 (7.98)
Fistula careb 2.74 (6.96) 0.99 (3.78) 1.91 (4.98) 3.98 (7.48) 6.48 (7.25)
In fistula centre during study 7 (10) 4 (3) 9 (4) 17 (7) 2 (7)

Homebound pattern participants aim to go back home as soon as possible. The archetypal pathway is a woman returning to her family right after a successful surgery. With 20/42 women still incontinent, this pattern also hides the harsher reality that some have accepted their fate resignedly. Homebound presents the highest proportion of married women (86%) and with children (55%). Differences to other patterns are significant for marital status (P = 0.001), but not for the child factor (P = 0.14).

The itinerant pattern includes participants with multiple comings and goings between repair centre and community. With a median of 4 (IQR: 2) movements per participant, these women moved significantly more than the rest of the cohort (P = 0.0000). However, the total aggregated time of their episodic stays in repair centre is higher than the median stay duration of all other participants (P = 0.001). Moreover, the proportion of women who underwent multiple operations was significantly higher in this pattern than in the rest of the cohort (P = 0.03).

The institutionalised pattern is characterised by a significantly longer continuous stay in fistula repair centre than all others (P = 0.0000). At first, institutionalised women were determined to settle until continence recovery. Over time, if the situation was not psychologically unbearable, women ended up acknowledging the incurability of their condition. Some stayed because they had nowhere to go, others because the repair centre routine gradually became their only reality. Notably, two participants who got continent in phase B were still living in the repair centre 1 year later with no intention to leave.

Urbanised pattern regroups women who stayed purposefully in the vicinity of repair centre or settled in town either for medical reasons (awaiting a new surgery, afraid of recurrence or new pregnancy) or for psycho-economical reasons (fistula victims sisterhood or work opportunity in/near fistula repair centre or in town).

Total lifetime with fistula was significantly shorter for homebound and itinerant participants than for institutionalised and urbanised ones (P = 0.000). Likewise, women in homebound and itinerant patterns have a significantly higher continence rate than those in institutionalised and urbanised patterns (P = 0.000). New cases were found in each pattern but their proportion decreased under the four successive patterns. There is a moderate positive correlation between duration of care process and the successive patterns (P = 0.0003).

Discussion

This paper reports results of an obstetric fistula patient cohort followed-up in space and time throughout their care. So far, this topic has been rarely studied. Duration and difficulties in fistula care were only incidentally reported in few articles (Turan et al. 2007; Yeakey et al. 2009). Until now the before/after study design prevailed (Browning et al. 2007; Browning & Menber 2008; Muleta et al. 2008; Weston et al. 2011). Consequently, it is assumed that fistula patients are either in the repair centre before surgery or at home after it. Our analysis of care process duration and the resulting ‘patient wandering’ helps overcome this simplistic time and location dichotomy.

Our cohort's socio-demographic profile at fistula occurrence is similar to that of other studies (Wall et al. 2005). Patients' profile evolution echoes qualitative findings of the existence of a wide variety of lives experiences with fistula (Yeakey et al. 2009). Participants' medico-social pathways point out several persisting problems already subject to research, such as residual incontinence (Browning 2006; Gutman et al. 2007) or reproductive health issues (Browning 2009; Wilson et al. 2011). Others issues are still unexplored: experience of those deemed incurable, mobility during care and institutionalisation in fistula repair centres.

The inverse correlation between continence and number of surgeries is known (Wall et al. 2005; WHO 2006). Previous surgeries and inoperable cases are mentioned occasionally in literature (Cam et al. 2010).

Our results reveal that the obstetric fistula population in care is twofold: new cases (≤1 surgery) standing a good chance of rapid continence gain and old cases (≥2 surgeries) clinging to hope. New/old ratio might vary between type of fistula repair centres (stand-alone fistula hospital, fistula repair centre within general hospital, repair camp, etc.). Old cases might be less noticeable, notably because of their mobility patterns. Yet, they do exist and have long and/or recurrent periods of stay in fistula repair centre. Conversely, a high patients' turnover characterised the group of new cases. The surgery success rate of new patients was as high as those previously reported in literature (Wall et al. 2005). Nevertheless, care duration of several months among new cases in our study deserves to be further investigated.

Patients' clinical files would have sharpened our new/old ratio (clinical fistula typology) and other outcomes (closure/continence). Our study is limited by the type of information collected: subjective and not checked with clinical or familial sources. Notable efforts were made to reduce interview biases. We believe growing trust of participants towards researchers increased quality of our data and reduced potential adverse effects inherent in any quantitative- and community-based research.

We intentionally opted for a patient-centred research and strived to be representative of all fistula patients in care. Spatial triangulation (five centres/seven regions/two countries) and investigation in four languages aimed to increase representativeness. Regional and/or language selection bias may yet exist.

Our sample size (N: 120) was larger and had a smaller dropout rate (7%) than other community-based follow-up studies (Nielsen et al. 2009; Wilson et al. 2011), but was still too small to highlight all statistical significance.

Our results emphasise the negative outcomes of some women who underwent fistula repair. Negative outcomes are not frequently reported in the literature. We believe this is due to a strong bias towards positive outcome reports. The fact that fistula repair tends to be used as catalyst for maternal morbidity reduction efforts may be one of the reasons behind this bias. However, publication of negative results should be encouraged to grasp reality of fistula repair process and outcomes.

It is understood that obstetric fistula needs holistic management (Arrowsmith et al. 1996). Given the length and the mixed socio-medical outcomes of care process, it would also require customised care. An acute care approach for new fistula cases should be combined with a chronic care approach for patients with a long obstetric fistula or care pathway.

With care often spreading over several months, patients' profiles and their bonds with fistula repair centres change. Insufficient and isolated solutions are currently proposed for women deemed incurable. Issues such as institutionalisation or other specific problems resulting from long-term care should receive more attention from health planners. The new/old cases ratio is a straightforward indicator that could be used for public health planning as well as for awareness raising.

Despite the increasing number of publications, research struggles to come closer to obstetric fistula patients' reality. Long-term, multisettings and mixed methods research might be strategic orientations to improve understanding of obstetric fistula care needs.

Acknowledgements

Sincere thanks to all fistula patients participating in this study and/or welcoming us in fistula repair centres and their communities. This study could not have happened without Ms. Assanatou Thiam's (Mali) and Ms. Rabi Hassane Bako's (Niger) tireless translations and empathic contact facilitation with patients. We would like to thank fistula surgeons, medical and social staff of the five participating fistula repair centres, the Fistula Network Eradication (REF)/Ministry of Public Health and NGO Dimol in Niger, NGO Delta Survie, Iamaneh and Doctors of the World in Mali.

Ancillary