You are here

  • Home
  • Archive
  • Volume 10, Issue 10
  • Availability of equipment and medications for non-communicable diseases and injuries at public first-referral level hospitals: a cross-sectional analysis of service provision assessments in eight low-income countries

Article Text

Article menu
XML
Global health
Original research
Availability of equipment and medications for non-communicable diseases and injuries at public first-referral level hospitals: a cross-sectional analysis of service provision assessments in eight low-income countries
  1. Neil Gupta1,2,3,
  2. Matthew M Coates1,2,
  3. Abebe Bekele4,
  4. Roodney Dupuy5,
  5. Darius Leopold Fénelon6,
  6. Anna D Gage7,
  7. Theodros Getachew4,8,
  8. Biraj Man Karmacharya9,
  9. Gene F Kwan1,3,10,
  10. Aimée M Lulebo11,
  11. Jones K Masiye12,
  12. Mary Theodory Mayige13,
  13. Maïmouna Ndour Mbaye14,
  14. Malay Kanti Mridha15,
  15. Paul H Park1,2,3,
  16. Wubaye Walelgne Dagnaw3,16,
  17. Emily B Wroe1,2,3,
  18. Gene Bukhman1,2,3,17
  1. 1Program in Global Noncommunicable Disease and Social Change, Harvard Medical School Department of Global Health and Social Medicine, Boston, Massachusetts, USA
  2. 2Division of Global Health Equity, Brigham and Women’s Hospital, Boston, Massachusetts, USA
  3. 3Partners In Health, Boston, Massachusetts, USA
  4. 4Health System and Reproductive Health Research Directorate, Ethiopian Public Health Institute, Addis Ababa, Ethiopia
  5. 5Le Ministère de la Santé Publique et de la Population, Port-au-Prince, Haiti
  6. 6Zanmi Lasante, Port-au-Prince, Haiti
  7. 7Department of Global Health and Population, Harvard T.H. Chan School of Public Health, Boston, Massachusetts, USA
  8. 8College of Medicine and Health Science, Institute of Public Health, University of Gondar College of Medicine and Health Sciences, Gondar, Ethiopia
  9. 9Department of Community Medicine, Kathmandu University School of Medical Sciences, Kathmandu, Nepal
  10. 10Section of Cardiovascular Medicine, Boston University School of Medicine, Boston, Massachusetts, USA
  11. 11Kinshasa School of Public Health, University of Kinshasa, Kinshasa, Democratic Republic of Congo
  12. 12Malawi Ministry of Health, Lilongwe, Malawi
  13. 13National Institute for Medical Research, Dar es Salaam, Tanzania
  14. 14Department of Internal Medicine, University Hospital Center of Dakar, Cheikh Anta Diop University, Dakar, Senegal
  15. 15Center for Non-communicable Diseases and Nutrition, BRAC James P Grant School of Public Health, BRAC University, Dhaka, Bangladesh
  16. 16Ethiopia Ministry of Health, Addis Ababa, Ethiopia
  17. 17Division of Cardiovascular Medicine, Brigham and Women’s Hospital, Boston, Massachusetts, USA
  1. Correspondence to Gene Bukhman; Gene_Bukhman{at}hms.harvard.edu

Abstract

Context and objectives Non-communicable diseases and injuries (NCDIs) comprise a large share of mortality and morbidity in low-income countries (LICs), many of which occur earlier in life and with greater severity than in higher income settings. Our objective was to assess availability of essential equipment and medications required for a broad range of acute and chronic NCDI conditions.

Design Secondary analysis of existing cross-sectional survey data.

Setting We used data from Service Provision Assessment surveys in Bangladesh, the Democratic Republic of the Congo, Ethiopia, Haiti, Malawi, Nepal, Senegal and Tanzania, focusing on public first-referral level hospitals in each country.

Outcome measures We defined sets of equipment and medications required for diagnosis and management of four acute and nine chronic NCDI conditions and determined availability of these items at the health facilities.

Results Overall, 797 hospitals were included. Medication and equipment availability was highest for acute epilepsy (country estimates ranging from 40% to 95%) and stage 1–2 hypertension (28%–83%). Availability was low for type 1 diabetes (1%–70%), type 2 diabetes (3%–57%), asthma (0%–7%) and acute presentations of diabetes (0%–26%) and asthma (0%–4%). Few hospitals had equipment or medications for heart failure (0%–32%), rheumatic heart disease (0%–23%), hypertensive emergencies (0%–64%) or acute minor surgical conditions (0%–5%). Data for chronic pain were limited to only two countries. Availability of essential medications and equipment was lower than previous facility-reported service availability.

Conclusions Our findings demonstrate low availability of essential equipment and medications for diverse NCDIs at first-referral level hospitals in eight LICs. There is a need for decentralisation and integration of NCDI services in existing care platforms and improved assessment and monitoring to fully achieve universal health coverage.

  • cardiology
  • diabetes & endocrinology
  • health policy
  • tropical medicine
  • asthma
  • surgery
http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/bmjopen-2020-038842

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Strengths and limitations of this study

    • To our knowledge, this is the first analysis with cross-country comparisons of readiness at first-referral level hospitals for acute and chronic presentations of a broader range of non-communicable diseases and injuries in low-income countries using practical and well-defined clinical criteria.

    • Valid cross-country analysis was possible by comparing facilities at analogous levels of the health system and using Service Provision Assessment data, which is largely standardised across countries.

    • The Service Provision Assessment surveys lack longitudinal data, and our analysis does not include information about supply chains, limiting the nature of the description about the availability of medications and equipment.

    • While we analysed data from eight countries representing a variety of low-income countries geographically, there are many more countries excluded from our analysis.

    Introduction

    Non-communicable diseases (NCDs) and injuries (NCDIs) are major drivers of the disease burden in low-income countries (LICs), accounting for 41% of mortality and morbidity in terms of disability-adjusted life years in 2017.1 In many LICs, the risk factors, epidemiology and disease conditions that comprise the burden of NCDIs differs from that seen in higher income countries.2 In these countries, harmful environments, infectious diseases and poor access to timely and high-quality health services are important factors contributing to the burden of NCDs.3 4

    Health sector interventions to address this burden have been increasingly recognised as both cost-effective and equitable, particularly for severe NCDIs affecting individuals early in life.2 5 In many LICs, availability of services to diagnose and manage NCDs is low and most often found primarily in urban higher-level hospitals.4 However, several NCDI interventions may be optimally delivered at first-referral level hospitals, which have been recognised as an essential component of the primary healthcare system.5 These first-referral hospitals, called district hospitals in some health systems, provide an opportunity to decentralise care, as they are more accessible to patients than tertiary referral hospitals and more capable of providing advanced services than health centres.6 7 Populations in rural areas, which tend to have higher rates of poverty in LICs,8 often face challenges accessing healthcare at distant facilities.9 Although one study has shown low readiness of health facilities in five LICs to deliver general services for cardiovascular disease, diabetes and chronic respiratory diseases,4 there has been limited multicountry assessment of hospital capacity to deliver a broader range of priority NCDI interventions. Some facility surveys assessing readiness and quality of care for other types of care, such as for maternal health, have found lower quality in facilities located in areas with higher rates of poverty.10

    In this study, we evaluated the availability of equipment and medications for management and diagnosis of the acute and chronic presentations of a broad range of NCDIs at first-referral level hospitals in eight LICs: Bangladesh, the Democratic Republic of the Congo (DRC), Ethiopia, Haiti, Malawi, Nepal, Senegal and Tanzania. We selected specific NCDIs with potentially severe presentations early in life, including asthma, hypertensive emergencies, heart failure, rheumatic heart disease, type 1 and 2 diabetes, epilepsy, injuries and minor surgical conditions and chronic pain. Given previous findings linking poverty and healthcare quality, we examined whether there were associations between subnational prevalence of extreme poverty and availability of equipment and medications. To the best of our knowledge, the countries we included in this study are the only LICs with comparable, openly available, nationally representative data on NCDI service provision recently collected via a standardised survey.

    Methods

    Study setting and data sources

    We used publicly available data from all Service Provision Assessment (SPA) surveys conducted in LICs through 2018. The SPA surveys are nationally representative health facility assessments administered as part of the Demographic and Health Survey (DHS) programme.11 These surveys were designed to assess human resources, infrastructure, equipment and medications available for maternal and child health (MCH) and priority infectious diseases.11 In 2012, the SPA questionnaires were updated to include indicators for some NCDIs, including infrastructure, human resources, medications, equipment and guidelines. Survey collectors indicate medications and equipment as available if they directly observe these items on the day of the survey.

    Since the initial inclusion of questions on NCDs, SPA surveys had been completed as of 2018 in eight LICs representing a broad range of geography, population size, economic productivity, healthcare expenditure and health system capacity: Bangladesh (2014), the DRC (2017–2018), Ethiopia (2014), Haiti (2013), Malawi (2013–2014), Nepal (2015), Senegal (2016–2017) and Tanzania (2014–2015) (see online supplementary appendix table 1).12 Bangladesh subsequently graduated to lower-middle-income status in 2015, and Senegal was moved from a lower-middle-income country to an LIC in 2017 and back to a lower-middle-income country in 2020. These countries, excluding Senegal and Bangladesh, together represent 19% of countries classified as LICs by the World Bank for the 2020 fiscal year and 44% of the global population living in LICs.13 14 The surveys in Haiti and Malawi were facility censuses, intended to capture all health facilities in the country. In Nepal, all public facilities were in the sampling frame and almost all public hospitals were surveyed. All hospitals in Ethiopia were included in the survey collection, along with a representative sample of private clinics and health centres. In Tanzania, all types of facilities were in the sampling frame, and 99% of hospitals were selected for the sample. In the DRC, the survey was done using a stratified random sample to obtain results by province and type of health facility. In Bangladesh, the surveys were conducted on a stratified random sample of facilities to obtain representative estimates by seven administrative divisions and by facility types (including a census of public district hospitals but a sample of public upazila health complexes). The combined 2016 and 2017 surveys in Senegal essentially includes a census of hospitals. Full details on the data from each country can be found in online reports, along with survey instruments.12 Data from these surveys were obtained from the DHS programme (www.dhsprogram.com).

    We build on methods previously developed to assess the quality of primary healthcare using similar datasets.15 16 The datasets were cleaned and standardised across the countries, categorising facilities as hospitals, health centres and clinics, or other facilities such as dispensaries. Facility weights used in analysis accounted for survey design and nonresponse to ensure representativeness, as oversampling is often done for certain facility types in these surveys. Our assessment focused on public first-referral hospitals. We regarded first-referral level hospitals as the first point of care for patients requiring referral from a primary health centre level, and the names for these facility types varied across countries. We limited our analysis to public sector facilities to optimise evaluation of health system investment and capacities provided from government sources for the poorest segment of the population, though data are not available within SPA surveys to specify payment source and mechanism for each commodity or service provided. In Nepal, we classified public district hospitals as first-referral level hospitals. In Bangladesh, we classified upazila health complexes as first-referral level hospitals because they are described as hospitals with in-patient beds and surgical care, occupy a lower level than district hospitals in Bangladesh, and have population catchment areas similar to district hospitals in other countries (1 facility per 375 000 population, compared with a range across other countries of 1 facility per ~175 000–560 000). In Malawi and Tanzania, we combined public district hospitals with public rural/community hospitals (Malawi) and district designated hospitals (Tanzania) as first-referral level hospitals. In Haiti, we used the community referral hospital classification. In Senegal, we reported on all public hospitals because more specific categories were not available from the SPA data. In the DRC, we used public hospitals below the provincial/tertiary level, though we could not differentiate between additional categories from the available data. In Ethiopia, we used both primary hospitals and general hospitals given the relatively recent introduction of primary hospitals and similarities in service delivery standards. Additional details can be found in online supplementary appendix table 1.

    Data analysis

    We analysed the availability of essential equipment and medications required for diagnosis and treatment of nine chronic disease states and four acute presentations of eleven NCDI conditions (table 1). We defined the minimum set of essential equipment and medications for the diagnosis and treatment of each condition using existing guidelines and iterative expert review from a group of public health practitioners, researchers and clinicians familiar with the local contexts. Acute conditions are those that require urgent procedures or hospitalisation, whereas chronic conditions were those requiring longitudinal follow-up for ongoing monitoring and disease management. The availability of the full essential set of functioning equipment and unexpired medications was determined for each facility. We considered equipment available if it was present in general outpatient, NCD or minor surgical areas. Medications were considered available if they were observed present and unexpired. In most cases, the data necessary to create these sets were available. In cases in which a survey question about one of the components of the equipment and medication set for a condition was not answered but the rest of the components were present at the facility, the facility received an ‘unknown’ classification for that set. If any one of the components was unavailable, then the essential set was classified as unavailable. Surveys in some countries did not contain questions about all of the relevant medications and equipment. In these cases, as well as in countries with >10% missing data for a particular variable, the country was excluded from analysis (see online supplementary appendix table 2 for list of missing variables). Missing variables and ‘unknown’ classifications for a set of equipment and medications were rare, resulting most frequently from surveys in particular countries excluding certain pain medications or surgical equipment. A total of 21 out of 797 public first-referral hospitals (less than 3%), 20 of these in Bangladesh, did not provide NCD services according to the survey. If these particular facilities were missing data for particular medications or equipment, we assumed the medications or equipment were unavailable. The availability of the essential sets of equipment and medications and their component items were tabulated by geographic units (both by country and by subnational units within countries). We also compared the proportion of facilities that reported diagnosing and managing chronic respiratory diseases, cardiovascular diseases and diabetes with the availability of essential equipment and medications for asthma, diabetes, hypertension (HTN), heart failure and rheumatic heart disease at those same facilities. We reported 95% CI for estimates using standard survey tabulation methods for countries that surveyed a sample of public first-referral hospitals (Bangladesh and the DRC) but not for countries where surveys were intended as a facility census. The surveys from Bangladesh and the DRC sampled a relatively large proportion of the total number of hospitals, so we calculated the 95% CI incorporating a correction for finite population size.

    View this table:
    Table 1

    Assigned essential equipment and medications for acute presentations of and chronic care for NCDI conditions at first-referral level hospitals

    To examine a potential association between the availability of NCDI medications and equipment with the prevalence of extreme poverty in subnational regions, we used a modified version of the Multidimensional Poverty Index from the Oxford Poverty and Human Development Initiative (see online supplementary appendix tables 3 and 4).8 We counted the number of the individual components across our disease-related sets of medications and equipment (table 1) that were available on the day the survey was conducted at each facility, deduplicating items in multiple sets. We assessed the association between the logit-transformed proportion of the total items available in a public first-referral level hospital and the prevalence of extreme poverty in the subnational unit (district or region) where the hospital was located using linear regression. We conducted regressions separately for each country to account for likely differences in governance and health systems. We used different regression specifications to assess the association between the availability of equipment and medications and the prevalence of extreme poverty. In one model specification, we used the prevalence of extreme poverty as a continuous variable, assuming a linear association. For possible non-linear association, we additionally used model specifications categorising extreme poverty prevalence into categorical groups by quartiles and by evenly spaced ranges of prevalence in each country. We also examined the association between the density of public first-referral level hospitals and hospitals per population in subnational areas with the prevalence of extreme poverty using linear regressions, accounting for country differences. Full details for these analyses are described in the appendix.

    Data cleaning, formatting and preparation were conducted using Stata/IC V.15.1 (StataCorp), and tabulations and regressions were conducted using R V.3.5.1 (the R Foundation for Statistical Computing).

    Patient and public involvement

    This research study was conducted without patient or public involvement in the design, execution or dissemination of the study.

    Results

    Overall, of the 9375 heath facilities across the eight countries which were surveyed, we identified 797 public first-referral level hospitals. Table 2 shows the availability of sets of equipment and medications for condition-specific acute care services, including surgery. The availability at these facilities of the complete set of essential equipment and medications needed for diagnosis and chronic care of specific conditions is shown in table 3.

    View this table:
    Table 2

    Availability of complete essential equipment and medications for acute presentations of NCDIs at public first-referral level hospitals in eight low-income countries

    View this table:
    Table 3

    Availability of complete essential equipment and medications for chronic care of NCDI conditions at public first-referral level hospitals in eight low-income countries

    Overall, medication and equipment availability was highest for acute management of epileptic seizures with diazepam (ranging between 40% and 95% in countries), followed by chronic care of stage 1–2 HTN (ranging between 28% and 83%), although this declined with HTN requiring more classes of medications (ranging between 0% and 34% for 4 classes of medications) (figure 1). Medication and equipment availability was low both for type 2 diabetes (requiring only oral medications) as well as type 1 diabetes requiring insulin. Availability was particularly low for management of acute presentations of diabetes such as diabetic ketoacidosis requiring intravenous fluids and monitoring of blood chemistries. Availability of essential equipment and medications for both acute and chronic presentations of asthma was extremely low in part due to the absence of beclomethasone inhalers at most facilities. Few hospitals had equipment or medications needed to diagnose and manage heart failure and rheumatic heart disease, which required ultrasound equipment. Essential surgical supplies were missing at most hospitals. Most countries had insufficient data to report on the availability of adequate medications to provide treatment of chronic pain.

    Figure 1
    Figure 1

    Availability of complete essential equipment and medications for acute presentations and chronic care of NCDI conditions at public first-referral level hospitals in eight low-income countries. CRD, chronic respiratory disease; CVD, cardiovascular disease; HTN, hypertension; NCDI, non-communicable diseases and injury; RHD, rheumatic heart disease.

    Notably, there was much lower observed availability of essential medications or equipment for NCDIs than the self-reported availability of services for these conditions by the facility (table 4). For chronic respiratory diseases, across six countries (not collected in Bangladesh or Malawi), over 75% of public first-referral hospitals reported diagnosis and management services, though fewer than 7% had the essential medications and equipment available for chronic asthma care and fewer than 4% for care of acute asthma exacerbations. Similarly, at least 66% of public first-referral hospitals in each country reported availability diagnostic and management services for diabetes, with the exceptions of Bangladesh (43%) and Malawi, where diagnosis or management was reported in 84% of these hospitals. Compared with this reported service provision, availability of essential medications and equipment were lower for type 1 diabetes (1.3%–70.1%), type 2 diabetes (3.3%–75.2%) and acute care for diabetic ketoacidosis (0%–25.6%). Between 48.8% and 94.3% of the hospitals reported availability of diagnostic and management services for cardiovascular diseases, though availability of essential medications and equipment were lower for HTN (27.7%–82.9%), heart failure (0%–31.6%) and rheumatic heart disease (0%–22.9%).

    View this table:
    Table 4

    Observed equipment and medication availability for selected NCDIs compared with self-reported service availability at public first-referral hospitals

    There was inconsistency in associations between availability of medications and equipment in public first-referral hospitals and prevalence of extreme poverty in corresponding sub-national units, both across countries and between regression approaches (see online supplementary appendix figure 1, tables 5 and 6). For example, there was a negative association in Bangladesh and the DRC, while there was a positive association in Ethiopia and Haiti. These findings also varied across model specifications. The method for grouping hospitals based on prevalence of extreme poverty in sub-national areas affected the estimates of association. Full regression results are shown in online supplementary appendix tables 5 and 6. The density of public first-referral level hospitals per population did not vary by poverty prevalence, though there was evidence that the density of hospitals overall as lower in poorer subnational areas.

    Discussion

    Our findings demonstrate that availability of essential equipment and medications for acute and chronic services for NCDI conditions across LICs remains extremely low at public first-referral level hospitals. Furthermore, the availability of essential equipment and medications for NCDI services is much lower than facility-reported management and diagnosis of chronic respiratory diseases, diabetes and cardiovascular diseases. Although we found some evidence of associations, both positive and negative, between availability of essential equipment and medications at a public first-referral level hospital and the prevalence of extreme poverty in the corresponding subnational area, results were inconsistent.

    Facilities included in this study were more equipped to provide services for the treatment of stages 1 and 2 HTN as compared with other more complex conditions. Previous facility assessments have demonstrated similar levels of medication availability for HTN treatment in LICs,17 18 and significantly lower availability as compared with high-income countries.19 The low availability of essential medications and equipment likely contributes to low overall coverage of services for HTN. The Lancet Commission on HTN reported that for countries in sub-Saharan Africa (SSA) with household level surveys, over half of hypertensive adults had not been diagnosed, and treatment coverage was low, ranging from 7% to 61%.20 Effective coverage was even lower, ranging from 1% to 31%.20

    Systematic reviews of heart failure in LICs have identified a high burden of non-ischaemic heart failure and high utilisation of diuretic therapy, though have not assessed systems readiness or medication availability.21 22 One study utilising a health facility assessment of hospitals in Kenya and Uganda reported higher availability of beta-blockers (98% and 92%, respectively) and furosemide (98% and 94%, respectively), but similar availability of ultrasound equipment and ACE-inhibitors, as compared with that found in our study.23 The availability of ultrasound has remained a limiting factor for the diagnosis and monitoring of heart failure and has been a focus of health systems diagnostic improvements and training in LICs.23 Availability of equipment and medications was similarly low for outpatient management of RHD, a common cause of heart failure in LICs. Availability of benzathine penicillin for primary and secondary RHD prophylaxis was highly limited across most countries. Although no detailed health facility reports exist for the availability of benzathine penicillin at primary health facilities across LICs, there have been global concerns for the availability and quality of this essential medication.24

    Our findings of availability of medications and equipment for diabetes services are consistent with previous reports of low availability of services and low coverage of diabetes services in LICs. In a group of 12 countries in SSA, only 22% of eligible individuals had received blood glucose measurement.25 Of those meeting biochemical criteria for diabetes, only 36% had previously received blood glucose measurement, 27% had been previously diagnosed, 25% were taking oral diabetes therapy, and 11% were taking insulin.25

    The availability of medications or services for asthma and other chronic respiratory diseases has not been well studied in low-income settings. Inhaled beta agonists, inhaled corticosteroids and systemic corticosteroids have more recently been included on lists for essential medicines in LICs, and price fluctuations have created a wide range of affordability for these medications.26

    Care for chronic epilepsy has traditionally been lacking in resource-poor settings.27 Despite the availability of benzodiazepines for acute management of seizures at the inpatient level, the management of chronic epilepsy and seizure prophylaxis is highly lacking. The SPA survey collects information on a limited number of medications that are used for daily prophylaxis of seizures in epileptic disorders and most countries did not include this in their surveys (only Nepal collected data on carbamazepine and phenobarbital). Others have reported the lack of coverage data, and the highly vulnerable characteristics of patients with epilepsy contribute to poor access to care, utilisation of available services, and poor overall coverage, resulting in high morbidity and mortality from epilepsy.28 The availability of second-line medications or intensive care for refractory seizures was not available.

    The extremely low availability of surgical equipment and medications for care of injuries and acute minor surgical conditions is consistent with previous reports.29 Recent modelling estimates have suggested that coverage of adequate surgical services in SSA and South Asia is less than 5% with sporadic and variable availability of individual tracer items.30 The ability of surgical interventions to offer one-time curative treatment leads to the high cost-effectiveness and financial risk protection offered by surgical interventions in low-income health systems.31 Additionally, there are limited data collected in the SPA survey regarding the availability of palliative care services or essential medications, and the data collected were limited only to medications for pain, such as analgesics and opiates. The availability of injectable morphine was collected in only three of the eight countries and availability of oral morphine in only two of the eight countries. This is consistent with the dearth of development in palliative care policies and implementation in LICs, particularly in SSA and a dramatic gap in the availability of morphine and other essential medications for palliative care.32 33

    Countries did vary in their level of availability of essential equipment and medications for the conditions analysed. Senegal displayed consistently lower rates of availability of equipment and medications than the other countries. In contrast, Ethiopia and Tanzania displayed considerably higher rates than other countries across many conditions, including most notably for HTN, diabetes care and heart failure. In these two countries, there was a relatively early recognition and coordinated approach to NCDs, including strategic and costed operational planning, strong civil society engagement and leadership, progressive outreach from well-established tertiary centres and a strong and well-organised primary care network.34 35

    This study has several limitations. First, the essential equipment and medications for NCDI services defined here does not include presence of well-trained and supervised human resources, a cornerstone for healthcare delivery and one that has been well established to be lacking for NCDs in SSA.36 The availability of essential equipment and medications presented in this analysis may, therefore, overestimate the overall service availability for the corresponding condition. Second, certain equipment and medications may have been observed within the health facility, but these items may not necessarily be accessible to the unit providing NCDI services within the facility (ie, ultrasound may be reserved for obstetrics) or affordable by the patients. We minimised this possibility by including equipment and medications from general outpatient, NCD and minor surgical areas. Additionally, medication availability may fluctuate; however, the date of data collection is randomly assigned and should not bias our findings. Third, the components of these sets of equipment and medications may not be comprehensive of all items needed for care associated with each disease condition, but rather represent a core number of elements measured within the available survey tools. Fourth, availability of supplies and equipment is not always associated with quality care.37 While there are no data available capturing nationally representative observations of the quality of NCDI services, previous studies in the field of MCH suggest that poor quality care can exist even in the presence of necessary supplies.38 Fifth, there is some incomplete data for specific disease conditions in certain countries reflecting adaptation of the SPA questionnaire by country teams. Sixth, the year of survey data collection varied among countries, which may limit direct comparison, and the results may underestimate current levels of availability if substantial improvements have been made following the data collection period, particularly in countries with older surveys. Finally, the level of geographical specificity for extreme poverty prevalence estimates limited the design of the analysis examining associations between poverty and the availability of medications and equipment. Further study is warranted, both with more specific data about populations in de facto catchment areas of facilities and process information about how factors like supply chains and planning affects variation in availability across facilities.

    Our findings have several implications to improve service availability of NCDIs in LICs. There is a need to prioritise decentralisation of a broad set of cost-effective and equitable interventions at first-referral level hospitals to increase care availability. Such interventions may include those typically confined to referral and university teaching hospitals and densely populated urban areas, such as chronic care delivery for severe NCDIs including heart failure, type 1 diabetes, advanced asthma and palliative care. Our finding that the density of hospitals per population is lower in poorer subnational areas, while the distribution of first-referral hospitals is more equitable, suggests that decentralisation of services can create more equitable access. Collaborations to develop such service packages have recently been launched.39 Several studies have reported promising outcomes for task-shifting and task sharing of essential NCD services to support decentralisation and increase availability of such services.40–42 Integration of NCD services with existing HIV and MCH services has been suggested as a cost-effective and important step towards increasing the availability of services in a universal health coverage (UHC) package, particularly at the primary care level.43 Coordination of governance and policy making for NCDI health sector interventions would also provide opportunities for integration of staff, training, guidelines, and supply chains required for adequate service delivery.44 Health financing for integrated platforms of NCDI service delivery within a UHC framework will be essential to improve basic availability of services. A high priority package for essential interventions for NCDIs within UHC has been proposed, and may provide reasonable cost estimates required to increase coverage of health sector services for NCDIs, including cross-cutting approaches to mental health, surgery, palliative care and rehabilitation.5

    The strengthening of health facility monitoring and extension of core NCDI indicators is highly needed.45 Although some readiness indices exist within standardised surveys, such as SPA and the Service Availability and Readiness Assessment (WHO), these are typically limited to outpatient management of HTN, type 2 diabetes and asthma. These surveys do not measure readiness for acute complications or inpatient needs, rely in part on reported (rather than observed) measures and do not identify minimum requirements for service delivery. An appropriate monitoring framework for NCDIs will need to include a greater number of tracer items that reflect current clinical guidelines, other NCDI conditions currently not represented in existing monitoring frameworks but represent a large burden of disease (ie, cancer, epilepsy, mental health, renal failure, liver cirrhosis, palliative care surgical services) and domains such as organisation, management, access, availability and quality of effective services.46 Additional examination of costing data, procurement processes and the supply chain could identify the drivers of stockouts.

    Our findings demonstrate variable but overall low availability of the minimum required equipment and medications to provide adequate diagnostic and therapeutic interventions for nine chronic conditions and four acute presentations of chronic conditions at first-referral level hospitals in eight LICs in three different regions. This observed availability of medications and equipment is substantially lower than self-reported diagnosis and management of chronic diseases by facilities in these countries. The provision of cost-effective and equitable health sector interventions for the diagnosis and management of both acute and chronic presentations of NCDIs are highly needed at first-referral level hospitals in LICs. The strengthening of these services through the public sector can help to keep patients from facing high costs for medicines and procedures. There is a need for progressive decentralisation of services for these conditions to first-referral level facilities, integration of such services in existing platforms of care, and improved assessment and monitoring of delivery of services in LICs to fully achieve targets established for UHC.

    Acknowledgments

    The authors would like to thank Gisela Robles Aguilar for estimation of subnational poverty prevalence, as well as Mamka Anyona for assistance with initial data analysis and Chris Noble for support with manuscript preparation. Gene F. Kwan was supported in part from the National Institutes of Health (1K23HL140133).

    References

    1. ↵
      1. Kyu HH,
      2. Abate D,
      3. Abate KH, et al
      . Global, regional, and national disability-adjusted life-years (DALYs) for 359 diseases and injuries and healthy life expectancy (HALE) for 195 countries and territories, 1990-2017: a systematic analysis for the global burden of disease study 2017. Lancet 2018;392:1859–922.doi:10.1016/S0140-6736(18)32335-3pmid:http://www.ncbi.nlm.nih.gov/pubmed/30415748
      OpenUrlCrossRefPubMed
    2. ↵
      1. Bukhman G,
      2. Mocumbi AO,
      3. Atun R, et al
      . The Lancet NCDI Poverty Commission: bridging a gap in universal health coverage for the poorest billion. Lancet 2020;396:991–1044.doi:10.1016/S0140-6736(20)31907-3pmid:32941823
      OpenUrlPubMed
    3. ↵
      1. Ogoina D,
      2. Onyemelukwe GC
      . The role of infections in the emergence of non-communicable diseases (NCDS): Compelling needs for novel strategies in the developing world. J Infect Public Health 2009;2:14–29.doi:10.1016/j.jiph.2009.02.001pmid:http://www.ncbi.nlm.nih.gov/pubmed/20701857
      OpenUrlCrossRefPubMed
    4. ↵
      1. Moucheraud C
      . Service readiness for noncommunicable diseases was low in five countries in 2013-15. Health Aff 2018;37:1321–30.doi:10.1377/hlthaff.2018.0151pmid:http://www.ncbi.nlm.nih.gov/pubmed/30080459
      OpenUrlPubMed
    5. ↵
      1. Jamison DT,
      2. Alwan A,
      3. Mock CN, et al
      . Universal health coverage and intersectoral action for health: key messages from disease control priorities, 3rd edition. Lancet 2018;391:1108–20.doi:10.1016/S0140-6736(17)32906-9pmid:http://www.ncbi.nlm.nih.gov/pubmed/29179954
      OpenUrlCrossRefPubMed
    6. ↵
      1. WHO Expert Committee on the Role of Hospitals at the First Referral Level & World Health Organization
      . Hospitals and health for all: report of a who expert Committee on the role of hospitals at the first referral level. World Health organization technical report series 744. Geneva: World Health Organization, 1987. https://apps.who.int/iris/handle/10665/38113
    7. ↵
      1. van LW,
      2. Lafort Y, World Health Organization
      . The role of the hospital in the district: delivering or supporting primary health care? Geneva: World Health Organization, Division of Strengthening of Health Services, 1990. https://apps.who.int/iris/handle/10665/59744
    8. ↵
      1. Robles Aguilar G,
      2. Sumner A
      . Who are the world’s poor? a new profile of global multidimensional poverty. Working paper 499. Washington, DC: Center for Global Development, 2019. https://www.cgdev.org/publication/who-are-worlds-poor-new-profile-global-multidimensional-poverty
    9. ↵
      1. Gabrysch S,
      2. Campbell OMR
      . Still too far to walk: literature review of the determinants of delivery service use. BMC Pregnancy Childbirth 2009;9:34. doi:10.1186/1471-2393-9-34pmid:http://www.ncbi.nlm.nih.gov/pubmed/19671156
      OpenUrlCrossRefPubMed
    10. ↵
      1. Sharma J,
      2. Leslie HH,
      3. Kundu F, et al
      . Poor quality for poor women? inequities in the quality of antenatal and delivery care in Kenya. PLoS One 2017;12:e0171236. doi:10.1371/journal.pone.0171236pmid:http://www.ncbi.nlm.nih.gov/pubmed/28141840
      OpenUrlCrossRefPubMed
    11. ↵
      1. Service Provision Assessments (SPA)
      . The DHS program. Available: https://dhsprogram.com/What-We-Do/Survey-Types/SPA.cfm [Accessed 14 Feb 2018].
    12. ↵
      1. The DHS Program - Survey Search
      . The DHS program. Available: https://dhsprogram.com/What-We-Do/survey-search.cfm?pgtype=main&SrvyTp=type [Accessed 14 Feb 2018].
    13. ↵
      1. United Nations
      2. Department of Economic and Social Affairs
      3. Population Division
      . World population prospects: the 2019 revision. United nations, 2019. Available: https://population.un.org/wpp/ [Accessed 10 Apr 2020].
    14. ↵
      1. The World Bank
      . World bank country and lending groups. Available: https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country-and-lending-groups [Accessed 11 Jun 2020].
    15. ↵
      1. Kruk ME,
      2. Chukwuma A,
      3. Mbaruku G, et al
      . Variation in quality of primary-care services in Kenya, Malawi, Namibia, Rwanda, Senegal, Uganda and the United Republic of Tanzania. Bull World Health Organ 2017;95:408–18.doi:10.2471/BLT.16.175869pmid:http://www.ncbi.nlm.nih.gov/pubmed/28603307
      OpenUrlPubMed
    16. ↵
      1. Gage AD,
      2. Leslie HH,
      3. Bitton A, et al
      . Assessing the quality of primary care in Haiti. Bull World Health Organ 2017;95:182–90.doi:10.2471/BLT.16.179846pmid:http://www.ncbi.nlm.nih.gov/pubmed/28250531
      OpenUrlPubMed
    17. ↵
      1. Peck R,
      2. Mghamba J,
      3. Vanobberghen F, et al
      . Preparedness of Tanzanian health facilities for outpatient primary care of hypertension and diabetes: a cross-sectional survey. Lancet Glob Health 2014;2:e285–92.doi:10.1016/S2214-109X(14)70033-6pmid:http://www.ncbi.nlm.nih.gov/pubmed/24818084
      OpenUrlPubMed
    18. ↵
      1. Mendis S,
      2. Al Bashir I,
      3. Dissanayake L, et al
      . Gaps in capacity in primary care in low-resource settings for implementation of essential noncommunicable disease interventions. Int J Hypertens 2012;2012:584041 doi:10.1155/2012/584041pmid:23251789
      OpenUrlPubMed
    19. ↵
      1. van Mourik MSM,
      2. Cameron A,
      3. Ewen M, et al
      . Availability, price and affordability of cardiovascular medicines: a comparison across 36 countries using WHO/HAI data. BMC Cardiovasc Disord 2010;10:25. doi:10.1186/1471-2261-10-25pmid:http://www.ncbi.nlm.nih.gov/pubmed/20534118
      OpenUrlCrossRefPubMed
    20. ↵
      1. Olsen MH,
      2. Angell SY,
      3. Asma S, et al
      . A call to action and a lifecourse strategy to address the global burden of raised blood pressure on current and future generations: the Lancet Commission on hypertension. Lancet 2016;388:2665–712.doi:10.1016/S0140-6736(16)31134-5pmid:http://www.ncbi.nlm.nih.gov/pubmed/27671667
      OpenUrlCrossRefPubMed
    21. ↵
      1. Damasceno A,
      2. Mayosi BM,
      3. Sani M, et al
      . The causes, treatment, and outcome of acute heart failure in 1006 Africans from 9 countries. Arch Intern Med 2012;172:1386–94.doi:10.1001/archinternmed.2012.3310pmid:http://www.ncbi.nlm.nih.gov/pubmed/22945249
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Callender T,
      2. Woodward M,
      3. Roth G, et al
      . Heart failure care in low- and middle-income countries: a systematic review and meta-analysis. PLoS Med 2014;11:e1001699. doi:10.1371/journal.pmed.1001699pmid:http://www.ncbi.nlm.nih.gov/pubmed/25117081
      OpenUrlCrossRefPubMed
    23. ↵
      1. Carlson S,
      2. Duber HC,
      3. Achan J, et al
      . Capacity for diagnosis and treatment of heart failure in sub-Saharan Africa. Heart 2017;103:1874–9.doi:10.1136/heartjnl-2016-310913pmid:http://www.ncbi.nlm.nih.gov/pubmed/28490619
      OpenUrlAbstract/FREE Full Text
    24. ↵
      1. Wyber R,
      2. Taubert K,
      3. Marko S, et al
      . Benzathine penicillin G for the management of RhD: concerns about quality and access, and opportunities for intervention and improvement. Glob Heart 2013;8:227–34.doi:10.1016/j.gheart.2013.08.011pmid:http://www.ncbi.nlm.nih.gov/pubmed/25690500
      OpenUrlPubMed
    25. ↵
      1. Manne-Goehler J,
      2. Atun R,
      3. Stokes A, et al
      . Diabetes diagnosis and care in sub-Saharan Africa: pooled analysis of individual data from 12 countries. Lancet Diabetes Endocrinol 2016;4:903–12.doi:10.1016/S2213-8587(16)30181-4pmid:http://www.ncbi.nlm.nih.gov/pubmed/27727123
      OpenUrlPubMed
    26. ↵
      1. Babar Z-U-D,
      2. Lessing C,
      3. Mace C, et al
      . The availability, pricing and affordability of three essential asthma medicines in 52 low- and middle-income countries. Pharmacoeconomics 2013;31:1063–82.doi:10.1007/s40273-013-0095-9pmid:http://www.ncbi.nlm.nih.gov/pubmed/24127259
      OpenUrlPubMed
    27. ↵
      1. Meyer A-C,
      2. Dua T,
      3. Ma J, et al
      . Global disparities in the epilepsy treatment gap: a systematic review. Bull World Health Organ 2010;88:260–6.doi:10.2471/BLT.09.064147pmid:http://www.ncbi.nlm.nih.gov/pubmed/20431789
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Ba-Diop A,
      2. Marin B,
      3. Druet-Cabanac M, et al
      . Epidemiology, causes, and treatment of epilepsy in sub-Saharan Africa. Lancet Neurol 2014;13:1029–44.doi:10.1016/S1474-4422(14)70114-0pmid:http://www.ncbi.nlm.nih.gov/pubmed/25231525
      OpenUrlPubMed
    29. ↵
      1. Spiegel DA,
      2. Droti B,
      3. Relan P, et al
      . Retrospective review of surgical availability and readiness in 8 African countries. BMJ Open 2017;7:e014496. doi:10.1136/bmjopen-2016-014496pmid:http://www.ncbi.nlm.nih.gov/pubmed/28264832
      OpenUrlAbstract/FREE Full Text
    30. ↵
      1. Alkire BC,
      2. Raykar NP,
      3. Shrime MG, et al
      . Global access to surgical care: a modelling study. Lancet Glob Health 2015;3:e316–23.doi:10.1016/S2214-109X(15)70115-4pmid:http://www.ncbi.nlm.nih.gov/pubmed/25926087
      OpenUrlPubMed
    31. ↵
      1. Meara JG,
      2. Leather AJM,
      3. Hagander L, et al
      . Global surgery 2030: evidence and solutions for achieving health, welfare, and economic development. Lancet 2015;386:569–624.doi:10.1016/S0140-6736(15)60160-X
      OpenUrlCrossRefPubMed
    32. ↵
      1. Rhee JY,
      2. Garralda E,
      3. Torrado C, et al
      . Palliative care in Africa: a scoping review from 2005-16. Lancet Oncol 2017;18:e522–31.doi:10.1016/S1470-2045(17)30420-5pmid:http://www.ncbi.nlm.nih.gov/pubmed/28884701
      OpenUrlPubMed
    33. ↵
      1. Knaul FM,
      2. Farmer PE,
      3. Krakauer EL, et al
      . Alleviating the access abyss in palliative care and pain relief-an imperative of universal health coverage: the Lancet Commission report. Lancet 2018;391:1391–454.doi:10.1016/S0140-6736(17)32513-8pmid:http://www.ncbi.nlm.nih.gov/pubmed/29032993
      OpenUrlCrossRefPubMed
    34. ↵
      1. Ministry of Health, Community Development, Gender, Elderly and Children, Government of the United Republic of Tanzania
      . Strategic plan and action plan for the prevention and control of non communciable diseases in Tanzania 2016-2020, 2016. Available: https://extranet.who.int/ncdccs/Data/TZA_B3_NCD%20Stategic%20Plan%202016%20-%202020v0.3.pdf [Accessed 14 Feb 2018].
    35. ↵
      1. Ministry of Health, Federal Democratic Republic of Ethiopia
      . National strategic action plan (NSAP) for prevention & control of non-communicable diseases in Ethiopia. Addis Ababa, Ethiopia: Ministry of Health, Federal Democratic Republic of Ethiopia, 2014-2016. https://www.iccp-portal.org/system/files/plans/ETH_B3_National%20Strategic%20Action%20Plan%20%28NSAP%29%20for%20Prevention%20and%20Control%20of%20Non-Communicable%20Diseases%20-%20Final.pdf
    36. ↵
      1. World Health Organization
      . Health workforce requirements for universal health coverage and the sustainable development goals, 2016. Available: https://www.who.int/hrh/resources/health-observer17/en/ [Accessed 14 Feb 2018].
    37. ↵
      1. Leslie HH,
      2. Sun Z,
      3. Kruk ME
      . Association between infrastructure and observed quality of care in 4 healthcare services: a cross-sectional study of 4,300 facilities in 8 countries. PLoS Med 2017;14:e1002464. doi:10.1371/journal.pmed.1002464pmid:http://www.ncbi.nlm.nih.gov/pubmed/29232377
      OpenUrlPubMed
    38. ↵
      1. Kruk ME,
      2. Gage AD,
      3. Arsenault C, et al
      . High-quality health systems in the sustainable development goals era: time for a revolution. Lancet Glob Health 2018;6:e1196–252.doi:10.1016/S2214-109X(18)30386-3pmid:http://www.ncbi.nlm.nih.gov/pubmed/30196093
      OpenUrlCrossRefPubMed
    39. ↵
      1. WHO Regional Office for Africa
      . PEN-Plus meeting in Kigali – the management and treatment of non-communicable diseases at primary levels strengthened. Available: https://www.afro.who.int/news/pen-plus-meeting-kigali-management-and-treatment-non-communicable-diseases-primary-levels [Accessed 16 Sep 2019].
    40. ↵
      1. Joshi R,
      2. Alim M,
      3. Kengne AP, et al
      . Task shifting for non-communicable disease management in low and middle income countries-a systematic review. PLoS One 2014;9:e103754. doi:10.1371/journal.pone.0103754pmid:http://www.ncbi.nlm.nih.gov/pubmed/25121789
      OpenUrlCrossRefPubMed
    41. ↵
      1. Anand TN,
      2. Joseph LM,
      3. Geetha AV, et al
      . Task sharing with non-physician health-care workers for management of blood pressure in low-income and middle-income countries: a systematic review and meta-analysis. Lancet Glob Health 2019;7:e761–71.doi:10.1016/S2214-109X(19)30077-4pmid:http://www.ncbi.nlm.nih.gov/pubmed/31097278
      OpenUrlPubMed
    42. ↵
      1. Eberly LA,
      2. Rusingiza E,
      3. Park PH, et al
      . 10-year heart failure outcomes from nurse-driven clinics in rural sub-saharan Africa. J Am Coll Cardiol 2019;73:977–80.doi:10.1016/j.jacc.2018.12.028pmid:http://www.ncbi.nlm.nih.gov/pubmed/30819366
      OpenUrlFREE Full Text
    43. ↵
      1. Atun R,
      2. Jaffar S,
      3. Nishtar S, et al
      . Improving responsiveness of health systems to non-communicable diseases. Lancet 2013;381:690–7.doi:10.1016/S0140-6736(13)60063-Xpmid:http://www.ncbi.nlm.nih.gov/pubmed/23410609
      OpenUrlCrossRefPubMedWeb of Science
    44. ↵
      1. Gupta N,
      2. Bukhman G
      . Leveraging the lessons learned from HIV/AIDS for coordinated chronic care delivery in resource-poor settings. Healthc 2015;3:215–20.doi:10.1016/j.hjdsi.2015.09.006pmid:http://www.ncbi.nlm.nih.gov/pubmed/26699346
      OpenUrlPubMed
    45. ↵
      1. Nickerson JW,
      2. Adams O,
      3. Attaran A, et al
      . Monitoring the ability to deliver care in low- and middle-income countries: a systematic review of health facility assessment tools. Health Policy Plan 2015;30:675–86.doi:10.1093/heapol/czu043pmid:http://www.ncbi.nlm.nih.gov/pubmed/24895350
      OpenUrlCrossRefPubMed
    46. ↵
      Primary health care performance initiative (PHCPI), 2018. Available: https://improvingphc.org/ [Accessed 20 May 2019].

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Twitter @NCDSynergies, @NCDIpoverty

    • Contributors GB initiated the study. NG, MMC and GB designed the analysis. NG conducted the literature search. ADG and MMC conducted data cleaning. MMC conducted data analysis. NG, MMC and GB drafted the manuscript. All authors (GB, NG, MMC, AB, RD, DLF, ADG, TG, BMK, GFK, AML, JKM, MTM, MNM, MKM, PHP, WWD and EW) contributed to data interpretation and critical evaluation and revision of the study manuscript.

    • Funding This study was funded by the Leona M. and Harry B. Helmsley Charitable Trust.

    • Competing interests None declared.

    • Patient consent for publication Not required.

    • Ethics approval The study protocol was reviewed and determined non-human subject research by the Institutional Review Board of the Harvard Faculty of Medicine.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data availability statement Data from the Service Provision Assessment surveys are available in a public, open access repository through the DHS Program: http://dhsprogram.com/data/available-datasets.cfm.