Reasons for loss to follow-up of patients with podoconiosis in the Amhara Region, northern Ethiopia

Abstract

Background

Podoconiosis is a non-infectious form of tropical lymphoedema characterised by swelling of the feet and lower legs. Treatment is simple and effective yet evidence indicates that a proportion of patients become lost to follow-up.

Methods

This study was a quantitative questionnaire-based study which aimed to identify the most common reasons for loss to follow-up of patients. A total of 191 participants registered with the International Orthodox Christian Charities (IOCC) who had become lost to follow-up were included in a cross-sectional survey based in the Amhara Region, northern Ethiopia.

Results

The most common reason was distance, stated by 26.7% (51/191). This was significantly associated with living further from the treatment site (p=0.02). Having had podoconiosis for longer was protective against this (p=0.03). For each additional hour of travel time a patient lived from the treatment centre, the odds of them reporting ‘distance’ as the main reason for becoming lost to follow-up increased by 1.61 (95% CI: 1.25–2.08).

Conclusions

The consequences of podoconiosis are exacerbated by walking long distances, but in most areas, this is currently required of patients in order to receive treatment. We recommend expansion of services to widen treatment availability, since provision of transport to and from treatment centres is unlikely to be feasible.

Clinic attendance, Elephantiasis, Ethiopia, Neglected tropical disease, Podoconiosis, Treatment adherence

Background

Podoconiosis is a highly stigmatising, neglected tropical disease caused by long-term barefoot contact with irritant volcanic soil found in parts of tropical Africa, North-west India, and historically in parts of South America.^1–3 Today, the highest prevalence can be found in Ethiopia, primarily affecting poor, subsistence farmers who cannot afford footwear.⁴ With a current population of more than 85 million, podoconiosis is estimated to affect more than one million individuals in Ethiopia, with a further 11 million at risk.⁵ The exact pathogenesis of this non-infectious disease has still to be ascertained, however it is generally thought to involve the absorption of ultrafine silica particles through the skin, which results in ascending, bilateral lymphoedema of the lower legs.^6,7 There is also a strong genetic element with the disease having been found to cluster in families.⁸

Although not associated with an increase in mortality, as a chronic disease podoconiosis causes significant social and economic impact.⁹ Podoconiosis patients are subjected to high levels of stigma, which can reduce their marriage, educational and employment opportunities, and result in their exclusion from places of worship and local meetings.^10,11 People with podoconiosis are about half as productive as unaffected individuals. Loss of economic production and medical costs have been estimated to exceed US$16 million a year.¹²

The disease can be entirely prevented by washing the feet, and wearing good quality shoes as this reduces skin contact with the irritant soil.¹¹ Wearing shoes also reverses early signs of the disease, halts progression and reduces the number of episodes of acute adenolymphangitis that affected individuals suffer from.¹¹ These painful ‘acute attacks’ may occur up to five times a year, causing the individual to develop a fever and require bed-rest for up to five days.^13,14 They are precipitated especially by walking long distances or simply spending a long day working in the fields.¹⁴

Treatment is simple and cheap consisting of several components: foot hygiene, skin care, compression bandaging, socks and shoes, elevation and movement, and (in a minority of cases) nodulectomy.^8,15 However, many endemic communities and health professionals are unaware of this treatment, which is not available in government health facilities.^16,17 Several non-governmental organisations (NGOs) have been set up within Ethiopia in order to fill the gap, including the International Orthodox Christian Charities (IOCC), which started providing treatment for podoconiosis in June 2010.^14,18 The treatment programme consists of three, monthly appointments, with an important dimension of education and social support.¹⁸ Patients are encouraged to join their local patient association branch in their kebele (village). Each branch is headed by an expert patient: a previous patient who has successfully completed the treatment programme, and can read and write. They are the first port-of-call for patients in the community.

For the treatment to be effective, adherence needs to be life-long.⁴ Providing treatment through the IOCC indefinitely would be unsustainable. The general ethos of the podoconiosis programme is ‘to help people help themselves’, and as such, after the third appointment, patients ‘graduate’ and are expected to buy their own locally available treatment supplies. Patients generally become independent after graduation, although further support can be obtained through their local patient association if necessary.

Evidence from IOCC clinic registers suggested that a proportion of patients who register for treatment only attend one or two clinic appointments. Studies show that poor adherence to treatment for chronic conditions can result in disease complications and increased health care costs.^19,20 Two qualitative studies exploring the factors related to this discontinued clinic attendance have been previously conducted, one in southern Ethiopia, and the second, immediately prior to this study, in the north.^21,22 These studies found the following factors to be related to loss to follow-up: ‘remoteness from clinic site’, ‘expectations of special support’, ‘worry about increasing stigma’, ‘illness’, ‘misconceptions about treatment’, ‘economic problems’ and ‘too busy preparing for a religious ceremony’.

Using these results, this research aimed to quantitatively assess the reasons why patients discontinue attending clinic appointments at the IOCC treatment sites in northern Ethiopia. The primary aim was to measure the frequency of each reason given for dropping out; the secondary aim was to assess how each of these reasons was influenced by a range of modifiable and non-modifiable factors. It is hoped that the results of this study will enable targeted improvement of the clinic programme and guide national policy as treatment is expanded to districts identified as endemic by the recent nationwide mapping.

Methods

Regulatory approval and ethical considerations

For each participant, witnessed verbal consent was sought after discussion of the information sheet and consent form. As most participants were illiterate, seeking written consent would have been impractical in this setting. The witness was the accompanying patient association leader, who was present only for the process of obtaining consent. The questionnaire was conducted in a private setting at the patient's home. Patients were reassured that their answers would not be disclosed to any third party and that they could stop the interview or skip any question at any time without reason. In compensation for their time, patients were given a bar of soap to aid with their self-treatment.

Study area and design

A questionnaire-based cross-sectional survey was used to quantitatively assess a range of factors found to be related to specific reasons for loss to follow-up. The questionnaire was developed from the results of the qualitative studies conducted in the area and in southern Ethiopia.^21,22

Participants and data collectors were from four woredas (districts) in the Amhara region of northern Ethiopia: Debre Eliyas, Baso Liben and Machakel in East Gojam, and Dembecha in West Gojam. These areas were chosen as they contained the highest number of individuals who had been lost to follow-up from the IOCC treatment programme. ‘Lost to follow-up’ was defined as anyone who had registered with the IOCC treatment programme and had attended at least one appointment, but had then missed their second or third appointments and currently did not access treatment.

The IOCC is based in Debre Markos, a town 300 km north of Ethiopia's capital, Addis Ababa. The combined population of the woredas it serves is 500 200. At the time of the study (May 2013) there were 400 patients on the IOCC clinic registers, with a further 13 457 on the waiting list. Although over 3900 patients had graduated from the treatment programme, 500 patients had become lost to follow-up. Debre Elias had the greatest loss to follow-up (250 patients) despite having the smallest population (91 400 people). The rest were spread across the remaining areas served by the IOCC.

Sampling

Sampling was performed using registration lists as the sampling frame. Consecutive patients living in the four districts above who were lost to follow-up were selected until the sample size was reached. Inclusion criteria required patients to be over the age of eighteen and to have been lost to follow-up. A total sample population of 350 was used to calculate sample size as this was thought to be the number of patients lost to follow-up at the time. This was done using the freely accessible, online sample size calculation software ‘Sampsize’.²³ Given unknown prevalence of key reasons for loss to follow-up, a sample size of 184 participants was shown to estimate 50% prevalence with 5% precision and confidence level of 95%. Taking into account time, geographical and financial considerations, and potential non-response, 200 participants was considered feasible.

Data collection

Data collection was carried out house-to-house by two clinical officers, seven nurses and six final year public health bachelor students during May 2013. These individuals were used in preference to clinic staff in order to minimise social desirability bias. All data collectors were familiar with podoconiosis, spoke Amharic and English and were known to have good communication skills. They were each assigned a patient association leader, who knew the participants' names and addresses and could guide data collectors to their houses.

All data collectors received training prior to undertaking data collection. Training consisted of the purpose of the study, interviewing techniques, the process of taking informed consent and a refresher course on podoconiosis (cause, prevention, treatment and clinical staging). Data collectors were advised to refer any patients expressing regret towards becoming lost to follow-up back to their local patient association branch for re-inclusion into the IOCC treatment programme.

A pilot study was undertaken immediately after data collector training with nine patients from Debre Markos, an area not included in the main survey. Completed questionnaires were checked and during a final de-briefing, data collectors fed back ideas based on previous experiences and the challenges they faced in interviewing participants. Questionnaires were amended accordingly.

Data collection tools

A structured questionnaire was used to assess patient demographics, understanding of podoconiosis cause, treatment and prevention, history of disease, current disease severity, and the potential factors related to loss to follow-up, as suggested by the qualitative studies. This questionnaire was adapted from previous studies, written initially in English, translated into Amharic and then back-translated into English by a third party to ensure accuracy.

Data analysis

Data entry was undertaken in Ethiopia. Questionnaires from respondents that did not meet inclusion criteria were excluded. Data analysis was conducted in the UK using the Statistical Package for Social Sciences (SPSS version 20.0, IBM SPSS Statistics for Windows, Armonk, NY, USA). The student's t-test was used to compare continuous variables, while Pearson's χ² test was used for analysis of categorical variables. A conventional significance level of p<0.05 was used.

Ethical approval

Ethical approval was obtained from the Research Governance and Ethics Committee of Brighton and Sussex Medical School in the UK and the Debre Markos University Research Ethics Committee in Ethiopia. Support letters were obtained from the East and West Gojam Zonal Health Departments and Woreda Health Offices.

Results

The final data set consisted of 191 adequately completed records. Socio-demographic characteristics, time taken to travel and shoe wearing are presented in Table 1, and clinical characteristics in Table 2. The frequencies of the prime reasons given by participants as to why they stopped attending clinic are presented in Figure 1. ‘Distance’ was the most common barrier, stated by 51 participants (26.7%). This was phrased ‘I cannot visit the treatment centre because it is too far away’ in the questionnaire. The next three most common prime reasons for loss to follow-up were ‘I missed my appointment because I had an acute attack’ (36 participants, 18.8%), ‘I was unwell with other illnesses’ (31 participants, 16.2%) and ‘I was too busy to attend my appointment (preparing for a religious ceremony)’ (25 participants, 13.1%). Worry about stigma, the assumption that treatment wouldn't work and lack of material support from the treatment providers were less commonly cited reasons for non-attendance.

Table 1.

Socio-demographic characteristics of study participants, transport and whether shoes worn

Parameter
Age, years; mean (SD)	50.3 (15.1)
Gender
Male	102 (53.4%)
Female	89 (46.6%)
Religion (n=189)
Orthodox Christian	188 (98.4%)
Muslim	1 (0.5%)
Education level (n=191)
Never attended school	125 (65.4%)
Informal schooling	54 (28.2%)
At least primary	12 (6.2%)
Occupation (n=186)
Farmer	142 (76.3%)
Daily labourer	15 (8.1%)
Merchant	11 (5.9%)
Unemployed	5 (2.7%)
Other	13 (7.0%)
Residence (n=187)
Urban	27 (14.4%)
Rural	160 (85.6%)
Woreda (district) (n=191)
Debre Elias	94 (49.2%)
Baso Liben	36 (18.8%)
Machakel	21 (11.0%)
Dembecha	40 (21.0%)
Income per month: mean (SD), min–max	US$29.5 (23.6), 0–165.3
Travelling time to treatment centre, hours; mean (SD), min–max (n=187)	1.95 (1.32), 0.05–6
Wearing shoes (n=191)
No	59 (30.9%)
Yes	132 (69.1%)

Table 2.

Clinical characteristics of study participants

Parameter
Age of onset, n=177, mean (SD)	25.0 (11.8)
Years before seeking treatment, n=171
0–10	47 (27.5%)
11–20	49 (28.7%)
21–30	43 (25.1%)
31–40	19 (11.1%)
41–50	6 (3.5%)
51–60	3 (1.8%)
61–70	4 (2.3%)
Sought treatment elsewhere, n=190
Yes	112 (58.9%)
No	78 (41.1%)
Frequency of acute attacks, n=191
Every week	26 (13.6%)
Every two weeks	33 (17.3%)
Every month	68 (35.6%)
Every 3 months	22 (11.5%)
Every 6 months	6 (3.1%)
Less often than every year	1 (0.5%)
Never	35 (18.3%)
Days of rest needed, n=155, mean (SD)	5.7 (5.6)
Adenolymphangitis on observation, (n=191)
Yes	48 (25.1%)
No	143 (74.9%)
Wounds visible on feet, n=191
Yes	34 (17.8%)
No	157 (82.2%)

Figure 1.

Frequency of reasons for discontinued clinic attendance, East and West Gojam zones, Amhara Region, May 2013.

On further analysis, participants who said ‘distance’ was the main reason were found, on average, to live further away (p=0.02) according to reported travelling time (Table 3). Logistic regression demonstrated that each additional hour of travel time from the treatment centre was associated with 1.61 (95% CI: 1.25–2.08) greater odds of reporting ‘distance’ as the main reason for discontinuing treatment. Interestingly, a longer duration of illness was found to be ‘protective’ against this reason (p=0.03) (Table 3). For the second most common reason, ‘I missed my appointment because I had an acute attack’, a smaller proportion (55.6%, 20/36) were found to be wearing shoes at the time of interview than of those who did not report this reason, (71.9%, 105/146, p<0.001). A greater proportion of people who said ‘acute attack’ also reported suffering from them more frequently than once a month than those who did not cite this reason.

Table 3.

Comparison of participant characteristics for the four most common reasons for loss to follow-up

Reason for loss to follow-up	Reported this reason	Did not report this reason	p-valuea
Variable
Distance
Age, mean (SD)	50.3 (14.4)	50.3 (15.4)	NS
Time spent travelling, hours, mean (SD)	2.58 (1.29)	1.72 (1.29)	0.02
Income, mean (SD)	US$31.5 (25.5)	US$28.7 (23.0)	NS
Illness duration, years, mean (SD)	20.6 (13.1)	25.4 (13.0)	0.03
Acute attack
Age, mean (SD)	51.1 (15.3)	50.5 (15.2)	NS
Time spent travelling, hours, mean (SD)	1.78 (1.21)	2.00 (1.37)	NS
Wearing shoes, n	20 (55.6%)	105 (71.9%)	<0.001
Illness
Age, mean (SD)	50.7 (13.8)	50.6 (15.5)	NS
Gender, n	19 (61.3%)	67 (44.4%)	<0.001
Time spent travelling, hours, mean (SD)	1.87 (1.42)	1.98 (1.33)	NS
Income, mean (SD)	US$22.7 (13.6)	US$30.7 (25.0)	0.02
Too busy
Time spent travelling, hours, mean (SD)	1.47 (1.12)	2.01 (1.34)	0.02
Income, mean (SD)	US$43.4 (35.1)	US$27.3 (20.6)	0.03
Some education, n	11 (44.0%)	40 (24.2%)	<0.001

NS: Not significant.

^a p-value for χ² test for categorical variables and t-test for continuous variables.

A greater proportion of participants who said ‘illness’ was the reason for discontinuing clinic attendance were female, than of respondents who did not give this reason (19/31, 61.3% vs 67/151, 44.4%, p<0.001) and income was significantly smaller (454 vs 614 Birr [US$22.7 vs US$30.7] per month, p=0.02) (Table 3). Finally, participants who said they were ‘too busy preparing for a religious ceremony’, were found to live significantly closer (1.47 vs 2.01 hours travel, p=0.02) and to have a greater average income than those who did not report this reason (867 vs 545 Birr [US$43.4 vs US$27.3] per month, p=0.03). Interestingly, a greater proportion of people who said they were ‘too busy’ had had some form of education (11/25, 44.0% compared to 40/165, 24.2%, p<0.001).

Discussion

In this predominantly rural district of Ethiopia, the most frequently mentioned reason given for discontinued clinic attendance was ‘distance’, with people walking up to six hours to reach treatment centres. This is consistent with Tora et al.'s qualitative assessment of loss to follow-up in the south of Ethiopia where ‘remoteness from the clinic site’ was a key reason for non-attendance.²¹ It also echoes findings in studies regarding loss to follow-up of patients with HIV in Uganda, Ethiopia and Kenya and was the key finding in a study assessing barriers to eye-surgery in the same region of Ethiopia.^24–27 This factor was significantly associated with reported walking time from participants' home to the treatment centres (p=0.02), but not higher or lower income (p=0.68), suggesting that distance is a structural barrier which cannot be overcome solely by individual-level financial resources. At the time that patients were lost to follow-up, three treatment centres were open, and some patients would have needed to travel up to 45 km across difficult terrain to access these. We did not collect geo-references for patients' households and therefore cannot compare actual distances with reported time spent travelling: this would be valuable future research.

The longer the duration of having podoconiosis, the less likely the participant was to cite distance as their main reason for non-attendance (Table 3). This result was surprising as it was thought that distance would have been perceived as a more profound barrier for these individuals. However, this may reflect secular change in perspectives on distance, or patients becoming accustomed to these distances over time. Further research is needed to investigate this finding.

Also consistent with Tora et al's qualitative study was that patients missed their appointments due to experiencing acute attacks of adenolymphangitis.²¹ Fewer respondents who cited this as their main reason were wearing shoes at the time of the interview than were not. However, more respondents giving this as their prime reason reported experiencing attacks more frequently. This highlights the continuing need for primary and secondary disease prevention in the ‘lost to follow-up’ population, through education and facilitation of shoe wearing. Acute attacks are often triggered by walking long distances, which is required by the majority of patients in order to reach treatment centres.¹⁴ With the two most common reasons for lost to follow-up being inter-linked, alternative models of treatment provision may be necessary. These barriers could perhaps be reduced by the creation of additional satellite or mobile clinics or monthly transportation services. However, the cost and efficacy of providing such services would need to be assessed.

Many participants described illness from other diseases, rather than podoconiosis, as a barrier to clinic attendance. This factor has been a significant finding in many other studies, regarding HIV,^24,25,28,29 leprosy^30,31 and eye surgery.²⁷ Giving illness as a prime reason was associated with being female and having a smaller income, two factors with well-established links to illness through multiple social, structural and cultural pathways.^20,32

Low income, which is further compounded by having podoconiosis,¹² was also demonstrated as a feature in the fourth most commonly cited reason: ‘I was too busy to attend my appointment’. The fact that this reason was more common than financial constraints may be biased by the full wording of this statement: ‘I was too busy to attend my appointment (preparing for a religious ceremony)’. The majority of participants were orthodox Christian (Table 1), and thus social desirability may have encouraged patients to cite religion as a reason for not attending. With many religious festivals in the Orthodox Christian calendar however, this may also be a key reason for loss to follow-up that might prompt the IOCC to change clinic dates. Interestingly, being ‘too busy’ was significantly associated with having a higher income (p=0.03) and having some level of formal education (p<0.001). Perhaps participants with higher incomes and formal education have more scheduled work patterns, and cannot miss a day of work to get treatment. Participants cited this reason despite living significantly closer to the treatment centre, which suggests that being too busy was not associated with the amount of time required to travel there.

In contrast to the study by Tora et al., conducted in the south of the country, this research did not find ‘stigma’ to be a predominant barrier to clinic attendance.²¹ This discrepancy may reflect the cultural differences between patients living in different regions of Ethiopia, or perhaps varying levels of education provided to the community about podoconiosis.

The socio-demographic characteristics of the respondents are consistent with baseline data gathered by Molla et al. in 2012.¹⁴ The average age of patients lost to follow-up (50.3) was greater than the average age of all patients with podoconiosis as described by Molla et al. (44.3). This suggests that patients who drop out are generally older, but the significance or validity of this statement cannot be deciphered without further research. Age of onset (25 years) was consistent with the known natural history of podoconiosis,⁹ however there was a much greater delay in seeking treatment after initial symptoms in the lost to follow-up population (20 years) than found by Molla et al. (5 years). This may reflect a different interpretation of ‘receiving treatment’, rather than the idea that patients who become lost to follow-up also present later. Participants in this study may have been referring to IOCC provision of treatment, which has only been available since 2010.¹⁸

Strengths and limitations of the study

Patients were chosen from areas that had a high frequency of patients who were lost to follow-up, and in Baso Liben, all patients who became lost to follow-up were reached. This method may have introduced systematic bias into the study, but was considered reasonable given time and financial constraints. Recall bias may have been introduced as patients were contacted regardless of when they dropped out, which may have been any time since 2010. This strengthens the study, as results are therefore not season- or time-specific. This study is strengthened by an extremely high response rate making results easily generalisable to the region served by the IOCC if not further.

Conclusions

Podoconiosis is a neglected tropical disease which, if left untreated, imposes significant physical, social and economic burdens on individuals and communities. This research used findings from previous qualitative studies to quantitatively assess the prime reasons for loss to follow-up from a treatment programme.^22,23

Distance was found to be the main reason that patients became lost to follow-up. Podoconiosis may be unique in that the process of trying to access treatment can exacerbate the disease. By walking barefoot for up to six hours to reach the IOCC treatment centres, patients develop attacks of adenolymphangitis, which in turn can accelerate the progression of this disabling disease.³ While studies assessing the reasons for discontinued clinic attendance for other diseases frequently find ‘distance’ to be a barrier to treatment, there may be biological as well as structural and social reasons that patients with podoconiosis cite ‘distance’ as a prime barrier to attendance.

Ultimately, this research provides further evidence as to why podoconiosis treatment should be incorporated into the government health agenda. If efforts to do this are successful, treatment will be available through public health centres, widening access to many more people with podoconiosis. Future provider-related research should monitor treatment offered by these health centres and ensure it is delivered in a way that best facilitates attendance.

Authors' contributions: AC and GD conceived the study; AC and GD designed the study protocol; AC, AT and GT carried out data collection; AC carried out data analysis and interpretation of these data. AC and GD drafted the manuscript. AT and GT critically revised the manuscript for intellectual content. All authors read and approved the final manuscript. GD is the guarantor of the study.

Acknowledgements: Many thanks to the staff at the IOCC office in Debre Markos, the study community and all data collectors.

Funding: This study was supported by The Association of Physicians of Great Britain and Northern Ireland, through their Links with Developing Countries scheme.

Competing interests: None declared.

Ethical approval: Ethical approval was obtained from the Research Governance and Ethics Committee of Brighton and Sussex Medical School in the UK and the Debre Markos University Research Ethics Committee in Ethiopia.