Evidence on Pulmonary Rehabilitation and the Challenges Faced in Low-and Middle-Income Countries: A Systematic Qualitative Review

Background In low-income countries (LIC) and low-middle-income countries (LMIC), the burden of chronic obstructive pulmonary disease (COPD) has increased due to the lack of prevention and the presence of barriers to enter rehabilitation programs. The aim of this systematic review is to analyze the evidence on pulmonary rehabilitation (PR) in LIC and LMIC. A systematic literature review was conducted. Four electronic databases were searched for qualitative and quantitative studies that documented the presence of PR in LIC and LMIC. We report our ndings following the Prisma guidelines. In addition, grey literature was also searched. Articles not in English, presenting a point of view and/or not treating an adult population (< 18 years old) were excluded from the review. Data were extracted by one reviewer and synthesized in the form of tables. Tables present individual characteristics of the PR reported within countries, including country of origin, study design, population attending, intervention (kind of program setting), frequency and duration of a program established (if available), with health outcomes. The PICO framework was used for every country with reported PR to assess population, intervention, comparison and outcomes found. This systematic review is registered on Prospero: CRD42020141655. 47 patients was most and frequency of different LMIC. Health outcomes on cardiopulmonary Results found that of LMIC is ongoing. The most important barriers access among a cost-effective rehabilitation (PRP) a multidisciplinary and non-pharmacological hospitalization even (HIC). as well as a symptom-limited exercise testing was done at baseline and at the end of the eight-week program. After an 8-week rehabilitation program, results showed an improvement in exercise tolerance, shortness of breath and muscle fatigue in patients with pulmonary diseases throughout the studied population, according to the Philippines association of cardiovascular rehabilitation 58 aspects of PR in HIC. A 12-question survey was completed by PR representatives of 430 centers from 40 countries from Europe (43.7%) or North America (43.5%). Findings demonstrated large differences among PR across continents in all surveyed aspects, including the setting, case-mix of individuals with a chronic respiratory disease, the composition of the PR team, completion rates, methods of referral and types of reimbursement. Canada, a report conducted an survey in all the identify most important outcomes.

We used the above combinations of search terms in each of the following databases: EMBASE, Medline, PubMed, and ScienceDirect. Only research reported in English was included. Articles were selected if published in the period starting from January 2000 till December 2019 (the date of the nal search). Abstracts and clinical-case studies were not included. The initial screening was based on article titles and abstracts. The second screening was based on the full text. Articles were categorized as eligible for the review if they reported the presence and features of PR, or access to such programs in LIC and LMIC. If the same author(s) reported the study in different forms, e.g. as abstract or point of view, we only consider the form that presented the most complete information on the study.
Unpublished literature (grey literature) was included and limited to the two rst pages on google that cover more recent grey literature for each country.
An additional electronic search was done on selected local websites for countries where literature was unavailable. Speci cally, local associations and international organization websites were visited to nd records on PR availability in the selected countries, where we found reports on the presence of PR according to the initial databases search. The content selected for the review was analyzed qualitatively following the directed content analysis method. Results were organized in the form of tables complemented by a narrative description of the results. The main themes used in the directed content analysis were: research designs, PR features and access-related challenges in LIC and LMIC. Characteristics of the programs were presented based on the PICO tool. The PICO acronym refers to the population, intervention, comparison and outcomes of a (usually quantitative) article. It is commonly used to identify components of clinical evidence for systematic reviews in evidence-based medicine and is endorsed by the Cochrane Collaboration 2 .
We assessed the research designs of the studies included in the review and their ndings in qualitative terms only. We checked the quality of our review using the PRISMA 2009 Checklist (Appendix B).

Results
The initial search in the databases provided 21,694 hits for global access to rehabilitation in addition to a set of 7 records for LIC and 50 for LMIC identi ed through the grey literature search among sources like local organizations of rehabilitation. Hits were then reduced to 351 articles by limiting the search to access to PR speci cally. All articles not in English, published before the year 2000 and not related to PR -which added up to 184 articles -were excluded. Thus, 167 records were used in the screening. Of the aforementioned, 84 articles were excluded. Namely, 62 did not cover access, 18 were about tele-rehabilitation, and 4 were excluded for other reasons. Thus, 83 full-text articles remained in the review.
Of the 83 articles assessed for eligibility, we found 47 research articles on PR, which we included in the review and 36 articles were excluded for not covering pulmonary rehabilitation. Of these studies, 20 articles were related to access to PR, 3 articles were related to home-based PR, and 24 articles were related to PR health outcomes. These articles have been selected for this review.' Figure 1 shows the ow diagram of the selection of the studies.

Study designs reported in the publications reviewed
The articles in the review showed several methodological issues with small sampling along with poor descriptions of PR data collection and analysis 19 . For example, the PRP was not at a multidisciplinary level, or risk of bias appeared during data gathering.

Programs reported in the publications reviewed
All key data gathered through our review have been summarized in two tables. Table 1 presents results for LIC and Table 2 presents results for LMIC. For every country included in these tables, we found studies reporting the absence or presence of PR, including education sessions for healthcare professionals or awareness campaigns about PR for the population. inpatients. The patients treated are COPD, asthma and post TB in Congo and Tanzania. The duration of the PR is about 12 weeks in outpatient cafe and the frequency of the sessions are 3 times a week in Congo. Three studies (2 RCTs and one qualitative study) report a duration of the program of twice a week during 6 weeks in Uganda. In addition, workshops on the future development of PR have been conducted among physicians in Nepal.  Ukraine, Vietnam and Zimbabwe 26-69 . We found 3 types of PR: outpatients, inpatients and home-based. The population treated are mild to severe COPD and post TB patients. Characteristics of the programs presented in Table 1 and 2 are described in the subsequent sub-sections (see Table 3 and 4). We used the PICO framework for every country with the presence of PR to assess the population, intervention, comparison and outcomes of the study found. Table 3 and 4 also present the individual characteristics of PR reported, including country of origin, pathology of patients attending, program setting type, frequency and duration of a program established with health outcomes.   Table 3 and 4, three setting types are reported, which are inpatient, outpatient, and home-based PR.

Inpatient
In India and in the Philippines, inpatient programs were reported. We did not nd the exact number of PR settings and their location. It was necessary to visit the website of every rehabilitation program or medical institution to obtain the essential data for this review. The Philippines has a Lung Center as well as associations for cardiovascular treatment and PR for inpatients and outpatients. In India, we found an "Indian Association of Respiratory Care" (IARC) in Amrita Hospitals in Delhi, as well as a Lung India journal which is the o cial publication of the Indian Chest Society.

Outpatient
According to Table 1, studies conducted in Uganda and Zambia [24][25][26] have shown that the most common PR setting is outpatient care. In Congo, India, Egypt, the Philippines, Nigeria, and Tunisia, most of the programs are available in an outpatient setting.
Speci cally, in Kinshasa, 38 patients in two hospitals were treated who had bronchial asthma (n=14) and COPD (n=24). Patients were treated three times a week for twelve weeks. COPD patients improved their FEV1 signi cantly compared with asthma patients 20 .
In India, one study was carried out with 30 subjects randomly selected based on inclusion and exclusion criteria and assigned to one of two groups; an experimental and a control group. The experimental group underwent two months of PR in addition to medication and the control group patients using medication only. The experimental group showed a signi cant improvement in exercise tolerance and quality of life 64 .
In Tunisia, twenty-six patients with COPD and nine age-matched healthy participants were studied. Exercise measurements included a 6 minutes walking test (6MWT) and an incremental exercise test completed at the beginning and the end of the program. Both patients with COPD and healthy subjects demonstrated functional responses to training but with somewhat different patterns in the quality of improvement of the 6MWD 52 .

Home-based
Home-based PR was reported in Egypt and India. In Egypt, a study evaluated the effect of a two-month, home-based program with outpatient supervision every two weeks, with tolerance exercises and health-related quality of life (HRQOL) using Arabic-translated generic and speci c questionnaires in 39 COPD patients. The two-month home-based PR was found to be an effective non-pharmacological intervention for COPD patients due to health outcomes on the quality of life and better cardiopulmonary function 33 .
An Indian study incorporating 6 weeks of home-based PR was effective in increasing exercise endurance in patients with lung infections. Another Indian study evaluated the effect of home-based PR in COPD patients in rural areas, including 40 patients who completed the PR and 20 who were excluded in PRP, which composed the control group. Results showed that the 6MWT had an average increase of 75.72 meters in the study group, while an average decrease of 2.1 meters in the control group. Results were statistically signi cant (p<0.005) 66,68 .
In another Indian home-based study of forty patients of stable COPD having severe air ow obstructions were divided into control and experimental groups randomly. Exercises of 30 minutes duration were performed at home twice daily for four weeks with supervision. Domiciliary PR for four weeks results in signi cant improvement in the quality of life and exercise tolerance, even without improvement in Force Expiratory Volume in one second (FEV1) 64 .

Patients attending PR
In Uganda, patients-participants included lung tuberculosis (TB), HIV but not COPD [24][25][26] . Results of the studies reviewed suggest most of the patients treated are suffering from respiratory diseases like COPD in Congo, India, Philippine, Nigeria, and Tunisia 20,22,23 In Tunisia, in a study on COPD patients, the intervention group underwent balance training 3 times a week for 6 months in addition to the standard PR. The control group received 6 months of the standard PR only. Balance-training incorporated into a standard PRP signi cantly improved balance test score in COPD patients 53,54 . Frequency and duration of the program In Uganda, PR duration varied from 4-8 weeks for patients who are HIV infected [24][25][26] . The program's frequency varied from 2-4 weeks and the program length of stay diverged from 4-24 weeks.
In Ukraine, the results of a study showed that PR should be applied at all disease stages, starting at the stationary phase and continuing during the outpatient and homecare phases. A recommendation of this study was that the session's exercise duration should be no more than 30 minutes 3-5 times a week for 8-12 weeks [49].
In India, PR was available of different durations and frequencies according to the program setting, phase and medical establishment's policy 60-69 . Also, in Maghreb and in Tunisia, one study done on COPD men who were clinically stable and underwent PR one session per day twice a week for eight weeks reported improvement 55 .

Health outcomes
Evidence on health outcomes established in studies reviewed was based on 6MWT, FEV1, and the Quality-of-Life questionnaire. As previously mentioned in Kinshasa, after the rehabilitation program, COPD patients improved their FEV1 score signi cantly compared with asthma patients 20 .
In Nigeria, anxiety and depression levels have been assessed, and an outpatient program for the rehabilitation of patients with severe ventilatory impairment due to COPD was conducted. Patients entered a 6-week outpatient program during which they attended twice weekly 2.5-hour sessions. There was a signi cant reduction in the depression and patients' anxiety levels after the program. Walking distance also improved signi cantly and was maintained at the improved level for six months 41 . A cohort study was conducted in Philippine, a 6MWT as well as a symptom-limited exercise testing was done at baseline and at the end of the eight-week program. After an 8-week rehabilitation program, results showed an improvement in exercise tolerance, shortness of breath and muscle fatigue in patients with pulmonary diseases throughout the studied population, according to the Philippines association of cardiovascular rehabilitation 58 .

Awareness campaign and education initiatives about PR
In countries where PR is not reported yet, we found awareness programs and/or education sessions for healthcare givers and for the population. Awareness campaigns aimed to increase healthcare providers' education to implement and refer patients to the program when e cient. Results showed that PR is highly recommended in Madagascar, Rwanda, Senegal, Tanzania and Zimbabwe 24 . In addition, some projects on PR's are in process in Africa.
In Nepal, the internal medicine society organized a workshop on PR in August 2016 to educate professional healthcare providers and physicians about the importance and bene ts of PR 21 . In addition, SOLID Nepal, a community-based pulmonary health support a project for 3 years and a COPD control-Promotion of comprehensive PR was established in February 2019. This project was implemented in the Bhaktapur district and the purpose of this project was to improve quality of life of COPD patients by providing PR to the population 24 In Bangladesh, the engagement of stakeholders can improve research prioritization, implementation, and outcomes. The organization of stakeholder engagement aimed to support the implementation of PR for patients with chronic respiratory diseases [27][28][29] . Raising awareness of the bene ts of PR between stakeholders is a step forward in the implementation of PR. Understanding the views of patients, public health o cials, policymakers, politicians, religious leaders help to create support for PR.
In India, a cross-sectional survey was executed to evaluate awareness through a COPD-awareness questionnaire 65 . Also, studies show that PR was going to be implemented in Rwanda, Senegal and Zimbabwe for HIV-infected patients 24 .
Overall, studies concluded that patient-awareness campaigns enabled patients to acquire better self-management skills, helped reduce the severity and frequency of disease's exacerbations, prevents hospitalizations and improved the patients' HRQoL.

Barriers to access PR
In Zambia, a research team developed the international rollout of a larger evaluation trial, which incorporated centers in Uganda, Tanzania and Zambia 24 . In Nigeria, a study conducted by Akinremi et al. sought to investigate physicians' knowledge and perceived barriers to PR in Nigeria. Five top barriers to PR identi ed included unavailable/insu cient rehabilitation specialists, lack of rehab equipment, non-availability of institutionalized documented PR protocols, inadequate training on PR and cost of care to patients 40 .

Discussion
This paper has reviewed the evidence on PRP in LIC and LMIC. Additionally, it has outlined differences in PRP in terms of duration and frequency.
Barriers reported in this study are not limited to one type or context, and extend to individual, health, nancial and environmental factors. PRP are most frequently established under medical supervision in an "outpatient center" design program in LMIC.
In LIC, we observed that the average duration of the programs varied from 4-8 weeks for patients treated, who were HIV infected patients. In LMIC, the frequency of PRP reported, differed from 2 to 4 weeks. A program length of stay (duration) diverges from 4 to 24 weeks in LMIC.
According to one study, PR in HIC and LMIC has a median length of 9 weeks with 2.5 sessions a week. The duration of each session is between 1-2 hours per session. Most of the programs in Europe and the USA are also outpatient 70 .
According to our results depicted in Table 1 and 2, we found health outcomes on HRQOL and 6MWD in 2 LIC as well. Additionally, in LMIC, we found more health outcomes on cardiopulmonary function.
Barriers to access to pulmonary rehabilitation programs We found that PRP reported in LIC and LMIC, vary widely in terms of duration and location 71 . PR faced four major barriers, which include individual, nancial, organization of healthcare and environmental barriers. Individual barriers, including cultural factors, lack of information about diseases and treatments, self-management underuse, over-reliance on pharmacological care, and use of unproven alternative therapies, act as a determinant for the patient's adherence to the program in low and middle-income countries. In addition, no covering of the program by the local government and no funds to implement PR is a major barrier to overcome in LIC and LMIC.
Patients' health status is a major barrier to adherence. Disease severity plays a signi cant role and in uences attendance and adherence to PR in every country. The level of the disease (use of long-term oxygen therapy) and COPD exacerbation are important limitations to their attendance 71 .
Current smokers tend to be less inclined to adopt health promotion behaviors which is a strong predictor of both attendance and adherence to PR 72 .
In LIC, patients who lived alone were less likely to attend PR 73 .
In HIC, PR bene ts are correlated with smoking status (active-smoking), depression, lower levels of social support, female gender, FEV1 and low HRQOL and extremes of age of patients are important determinants of joining the programs 72,73 . The demographic feature of the patient, such as mobility conditions, longer traveling distance to reach hospitals and location to the rehabilitation's clinic are the most important reasons for adherence 71 . Graves et al. found that patients living further than 25 minutes from a PR center were less likely to attend a PR education session while patients were more likely to adhere to a program if they lived closer to PRP in every country 74,75 .
In LIC and LMIC, the most common barriers found are at a nancial level, including poverty, poor education, and infrastructure or no clinics due to a lack of public health priority, implementation costs of PR, and lack of data on morbidity and mortality for every country on pulmonary diseases. Unsustainable generalizations across cultures and healthcare systems explain why management guidelines developed in HIC are di cult to implement in LIC and LMIC. The lack of funds and ineffective governance do not support the implementation of programs and building outpatient rehabilitation programs. Also, no reimbursement is available in LIC and LMIC, which complicates the attendance and adherence to the program if a center does exist.
The nal inherent barrier is the organization of healthcare services. Professional healthcare providers are the main referrals and need awareness and training about PR to limit the use of basic medications and start sending patients to PRP in LIC and LMIC. Poor education of referrals on the management of COPD, non-existence of supply and distribution infrastructure are important barriers to the program.
Comparison with studies in HIC Literature on PR in HIC is more signi cant and much larger compared to that in LIC and LMIC. Plenty of national surveys, international reviews, and meta-analyses have been conducted in HIC. For example, in 2014, one study reviewed differences in the content and organizational aspects of PR in HIC. A 12-question survey was completed by PR representatives of 430 centers from 40 countries from Europe (43.7%) or North America (43.5%). Findings demonstrated large differences among PR across continents in all surveyed aspects, including the setting, case-mix of individuals with a chronic respiratory disease, the composition of the PR team, completion rates, methods of referral and types of reimbursement.
In the US, there are over 230 PR (copdfoundation.org). In Canada, a report from the Canadian Thoracic society conducted an online survey in all the hospitals to identify PR, which also con rms the high availability of PRP. In HIC, HRQOL, 6MWT and dyspnea are identi ed as the three most important outcomes.
An important challenge to add: is the implementation of PR in LMICs during the Covid-19 pandemic. The presence of cost-effective programs will be an added value to Covid-19 survivors. A home-based PRP may overcome barriers in the treatment post coronavirus in LIC and LMIC countries. A call of action is needed to implement PR in LMICs with the help of HIC's experts.

Strength and limitations of the review
Our results are bound by certain limitations as well as supported by strengths. A major limitation is that for a lot of countries, information is unavailable. For 28 LIC and 49 LMIC we were unable to nd any information. Another limitation was the exclusion of non-English literature, which means that some relevant publications from non-English speaking countries and might have been excluded in this review. The nancial barrier in LIC that bounds the implementation of PR centers and training of professional healthcare providers is another limitation. Additionally, PR might be implemented in some countries but not analyzed and reported in the literature. The risk of bias across studies included a lack of researches conducted in LIC, especially in LMIC.

Conclusion
Our review shows that only a few studies are available for LIC and LMIC on PR, especially on access to such programs. Literature is thin or nearly "absent". Evidence on access to PR and barriers faced by patients is also sparse. There is a need for research in this area to provide evidence for policies to decrease the burden of pulmonary diseases and to prevent infections, thus,