health service provision in tamil nadu · ihme institute for health metrics and evaluation iphs...

Assessing Facility Capacity, Costs of Care, and Patient Perspectives

HEALTH SERVICE PROVISION IN TAMIL NADU

CCESS,OTTLENECKS,OSTS, ANDQUITY

A B CE

PUBLIC HEALTH FOUNDATIONOF INDIA

INSTITUTE FOR HEALTH METRICS AND EVALUATIONUNIVERSITY OF WASHINGTON


A B CE

5 Acronyms6 Termsanddefinitions8 Executivesummary11 Introduction13 ABCEprojectdesign18 MainfindingsHealth facility profiles Facility capacity and characteristics Patient perspectives Efficiency and costs48 Conclusionsandpolicyimplications52 Annex

HEALTH SERVICE PROVISION IN TAMIL NADU

Table of Contents

Assessing Facility Capacity, Costs of Care, and Patient Perspectives

PUBLIC HEALTH FOUNDATIONOF INDIA

INSTITUTE FOR HEALTH METRICS AND EVALUATIONUNIVERSITY OF WASHINGTON

32

About IHME

About Public Health Foundation of India Collaborations

About this report

The Public Health Foundation of India (PHFI) is a public-private initiative to build institutional capacity in India for strengthening training, research, and policy development for public health in India. PHFI adopts a broad, integrative approach to public health, tailoring its endeavors to Indian conditions and bearing relevance to countries facing similar challenges and concerns. PHFI engages with various dimensions of public health that encompass promotive, preventive, and therapeutic services, many of which are often lost sight of in policy planning as well as in popular understanding.

The Institute for Health Metrics and Evaluation (IHME) is an independent global health research center at the University of Washington that provides rigorous and comparable measurement of the world’s most important health problems and evaluates the strategies used to address them. IHME makes this information freely available so that policymakers have the evidence they need to make informed decisions about how to allocate resources to best improve population health.

This project has immensely benefitted from the key inputs and support from Dr. K. Kolanda Swamy, Director of Public Health and Preventive Medicine, Government of Tamil Nadu, and from Dr. Thamma Rao. Approvals and valuable support for this project were received from the Tamil Nadu state government and district officials, which are gratefully acknowledged.

Assessing Facility Capacity, Costs of Care, and Patient Perspectives: Tamil Nadu provides a comprehensive assessment of health facility performance in Tamil Nadu, including facility capacity for service delivery, efficiency of service delivery, and patient perspectives on the service they received. Findings presented in this report were produced through the ABCE project in Tamil Nadu, which aims to collate and generate the evidence base for improving the cost-effectiveness and equity of health systems. The ABCE project is funded through the Disease Control Priorities Network (DCPN), which is a multiyear grant from the Bill & Melinda Gates Foundation to comprehensively estimate the costs and cost-effectiveness of a range of health interventions and delivery platforms.

54

Acronyms Acknowledgments

We especially thank all of the health facilities and their staff in Tamil Nadu, who generously gave of their time and facil-itated the sharing of facility data that made this study possible. We are also most appreciative of patients of the facilities who participated in this work, as they too were giving of their time and were willing to share their experiences with the field research team.

At PHFI, we wish to thank Rakhi Dandona and Lalit Dandona, who served as the principal investigators for the ABCE project in India. We also wish to thank Anil Kumar for guidance with data collection, management, and analysis. The quantity and quality of the data collected for the ABCE project in India is a direct reflection of the dedication of the field team. We thank the India field coordination team, which included Md. Akbar, G. Mushtaq Ahmed, and S.P. Ramgopal. We also recognize and thank Venkata Srinivas, Sagri Negi, and Sheetal Bishnoi for data management and coordination with field teams.

At IHME, we wish to thank Christopher Murray and Emmanuela Gakidou, who served as the principal investigators. We also recognize and thank data analysts and Post-Bachelor Fellows at IHME: Roy Burstein, Alan Chen, Emily Dansereau, Katya Shackelford, Alexander Woldeab, Alexandra Wollum, and Nick Zyznieuski for managing survey programming, survey updates, data transfer, and ongoing verification at IHME during fieldwork. We are grateful to others who contrib-uted to the project: Michael Hanlon, Santosh Kumar, Herbie Duber, Kelsey Bannon, Aubrey Levine, and Nancy Fullman. Finally, we thank those at IHME who supported publication management, editorial support, writing, and design.

This report was drafted by Marielle Gagnier, Lauren Hashiguchi, and Nikhila Kalra of IHME and Rakhi Dandona from PHFI.

Funding for this research comes from the Bill & Melinda Gates Foundation under the Disease Control Priorities Network (DCPN).

ABCE Access, Bottlenecks, Costs, and EquityANC Antenatal careANM Auxiliary nurse midwifeCHC Community health centreCI Confidence intervalDCPN Disease Control Priorities NetworkDEA Data envelopment analysisDH District hospitalDOTS Directly observed treatment, short-courseIHME Institute for Health Metrics and EvaluationIPHS Indian Public Health StandardsNCD Non-communicable diseasesOR Odds ratioPHC Primary healthcare centrePHFI Public Health Foundation of IndiaSDH Sub-district hospitalSFA Stochastic frontier analysisSHC Sub health centreSTI Sexually transmitted infectionTN Tamil NaduWHO World Health Organization

76

A B C E I N TA M I L N A D U

Table 1 defines the types of health facilities in Tamil Nadu; this report will refer to facilities according to these definitions.

Table1 Health facility types in Tamil Nadu1

1 Directorate General of Health Services, Ministry of Health & Family Welfare, and Government of India. Indian Public Health Standards (IPHS) Guidelines. New Delhi, India: Government of India, 2012.

T E R M S A N D D E F I N I T I O N S

Districthospital(DH)These facilities are the secondary referral level for a given district. Their objective is to provide comprehensive secondary health care services to the district’s population. DHs are sized according to the size of the district population, so the number of beds varies from 75 to 500.

Sub-districthospital(SDH)These facilities are sub-district/sub-divisional hospitals below the district and above the block level hospitals (CHC). As First Referral Units, they provide emergency obstetrics care and neonatal care. These facilities serve populations of 500,000 to 600,000 people, and have a bed count varying between 31 and 100.

Communityhealthcentre(CHC)These facilities constitute the secondary level of health care and were designed to provide referral as well as specialist health care to the rural population. They act as the block-level health administrative unit and as the gate-keeper for referrals to higher-level facilities. Bed strength ranges up to 30 beds.

Primaryhealthcentre(PHC)These facilities provide rural health services. PHCs serve as referral units for primary health care from sub- centres and refer cases to CHC and higher-order public hospitals. Depending on the needs of the region, PHCs may be upgraded to provide 24-hour emergency hospital care for a number of conditions. A typical PHC covers a population of 20,000 to 30,000 people and hosts about six beds.

Subhealthcentre(SHC)Along with PHCs, these facilities provide rural health care. SHCs typically provide outpatient care, which includes immunizations, and refer inpatient and deliveries to higher-level facilities.

Health facility types in Tamil Nadu

Terms and definitions

Definitions presented for key technical terms used in the report.

Constrainta factor that facilitates or hinders the provision of or access to health services. Constraints exist as both “supply-side,” or the capacity of a health facility to provide services, and “demand-side,” or patient-based factors that affect health-seeking behaviors (e.g., distance to the nearest health facility, perceived quality of care received from providers).

DataEnvelopmentAnalysis(DEA)an econometric analytic approach used to estimate the efficiency levels of health facilities.

Efficiencya measure that reflects the degree to which health facilities are maximizing the use of the resources available in producing services.

Facilitysamplingframethe list of health facilities from which the ABCE sample was drawn. This list was based on a 2012–2013 facility inventory published by the Tamil Nadu state government.

Inpatientvisita visit in which a patient has been admitted to a facility. An inpatient visit generally involves at least one night spent at the facility, but the metric of a visit does not reflect the duration of stay.

Inputstangible items that are needed to provide health services, including facility infrastructure and utilities, medical supplies and equipment, and personnel.

Outpatientvisita visit at which a patient receives care at a facility without being admitted.

Outputsvolumes of services provided, patients seen, and procedures conducted, including outpatient and inpatient care, laboratory and diagnostic tests, and medications.

Platforma channel or mechanism by which health services are delivered.

StochasticFrontierAnalysis(SFA)an econometric analytic approach used to estimate the efficiency levels of health facilities.

9

E X E C U T I V E S U M M A R Y

W

Executive summary

ith the aim of establishing universal health coverage, India’s national and state gov-ernments have invested significantly in expanding and strengthening the public

health care sector. This has included a particular com-mitment to extending its reach to rural populations and reducing disparities in access to care for marginalized groups. However, in order to realize this goal it is neces-sary for the country to critically consider the full range of factors that contribute to or hinder progress toward it.

Since its inception in 2011, the Access, Bottlenecks, Costs, and Equity (ABCE) project has sought to compre-hensively identify what and how components of health service provision – access to services, bottlenecks in de-livery, costs of care, and equity in care received – affect health system performance in several countries. Through the ABCE project, multiple sources of data, including facility surveys and patient exit interviews, are linked together to provide a nuanced picture of how facili-ty-based factors (supply-side) and patient perspectives (demand-side) influence optimal service delivery.

Led by the Public Health Foundation of India (PHFI) and the Institute for Health Metrics and Evaluation (IHME), the ABCE project in Tamil Nadu is uniquely positioned to inform the evidence base for understanding the country’s drivers of health care access and costs of care. Derived from a state-representative sample of 168 facilities, the findings presented in this report provide governments, international agencies, and development partners alike with actionable information that can help identify areas of success and targets for improving health service provision.

The main topical areas covered in this report move from an assessment of facility-reported capacity for care to quantifying the services actually provided by facilities and the efficiency with which they operate; tracking facil-ity expenditures and the costs associated with different types of service provision; and comparing patient per-spectives of the care they received across different types of facility. Further, we provide an in-depth examination

and comparison of facility-level outputs, efficiency, capac-ity, and patient experiences. It is with this information that we strive to provide the most relevant and actionable in-formation for health system programming and resource allocation in Tamil Nadu.

Facilitycapacityforserviceprovision

Whilemostfacilitiesreportprovidingkeyhealthservices,significantgapsincapacitywereidentifiedbetweenreportedand functionalcapacityforcare.

• Health facilities generally reported a high availability of a subset of key services. Services such as antenatal care, routine deliveries, general medicine, and phar-macies were widely available across facilities.

• Services for non-communicable diseases (NCDs) were limited. While significant numbers of district hospitals reported providing psychiatry (77%) and car-diology (77%), very few provided chemotherapy (8%). Availability decreased markedly at lower levels of the health system.

• Basic medical equipment such as scales, stethoscopes, and blood pressure apparatus were widely available at all health facility levels, but laboratory equipment such as glucometers, blood chemistry analyzers, and incubators were less readily available. For example, only 69% of district hospitals had glucometers, drop-ping to 45% at the sub health centre level. This shows limited capacity for testing throughout the health sys-tem, with particular implications for diagnosing and treating NCDs.

• While ECGs and ultrasounds were widely available, gaps also emerged with regard to imaging equipment, particularly at lower-level health facilities. While 92% of district hospitals had X-rays, this figure was just 38% for community health centres. CT scans were available in just 38% of district hospitals and 4% of sub-district hospitals.

• A service capacity gap emerged for the majority of health facilities across several types of services. Many facilities reported providing a given service but lacked full capacity to properly deliver it, for instance lacking functional equipment or medications. For example, while all primary and community health centres re-ported providing routine delivery care, none were fully equipped to do so. This discordance has substantial programmatic and policy implications for the health system in Tamil Nadu, highlighting continued chal-lenges in ensuring facilities have all the supplies they need to provide a full range of services.

Physicalinfrastructureofhealthfacilities hasimproved,butgapsintransportand communicationremain.

• Functional electricity was available at all hospitals, community health centres, and primary health centres. Ninety-two percent of sub health centres had electric-ity, showing substantial improvement on figures from past government surveys.

• Access to piped water was relatively high at district hospitals (85%), sub-district hospitals (88%), commu-nity health centres (88%), and primary health centres (87%), though it is notable that the figure is lowest for district hospitals. Piped water was limited at sub health centres (59%). Similarly, access to flushed toilets was markedly lower at sub health centres (59%) than other facility types (85%–91%). These figures do reflect in-vestments into improving physical infrastructure at health facilities, though discrepancies remain between high- and low-level facilities.

• There was nearly universal access to phones and computers across facility types. However, only 22% of primary health centres had any access to vehicles. These findings have serious implications for the timely transportation of patients to receive higher levels of care.

Nursescomposedthemajorityofstaffatdistricthospitals,whileatotherfacilitylevelsparamedicalstaffoutnumberedbothdoctorsandnurses.

• Staff numbers were concentrated at district hospitals with an average of nearly 211 personnel. Sub-district hospitals had the second highest number of person-

nel, but this was a quarter of that at district hospitals, while health centres averaged between one and 32 staff. While some of this variation is a result of service provision and population size, this also demonstrates relative shortages in human resources for health.

Facilityproductionofhealthservices

Healthfacilitiessawincreasesinboth outpatientandinpatientvisitsovertime.

• Between 2007 and 2011, outpatient numbers in-creased, with the highest patient volumes at district hospitals. Outpatient visits accounted for the large ma-jority of patients seen per staff member per day across all facility types. Inpatient visits increased for all facility types between 2007 and 2011. The average number of immunization doses administered remained stable over the five years, with slight declines in sub-district hospitals and community health centres.

Facilitiesshowedcapacityforlargerpatientvolumesgivenobservedresources.

• In generating estimates of facility-based efficiency, or the alignment of facility resources with the num-ber of patients seen or services produced, we found a wide range of efficiency levels within facility types, suggesting that a substantial performance gap may exist between the average facility and facilities with the highest efficiency scores. Efficiency scores were rela-tively low across all health facilities, with 74% being the highest mean across platforms.

• If they operated at optimal efficiency, district hospitals could provide 249,706 additional outpatient visits with the same inputs (including physical capital and per-sonnel), while primary health centres could produce 21,906 additional outpatient visits.

• These efficiency scores indicate that there is consid-erable room for health facilities to expand service production given their existing resources. Future work on pinpointing specific factors that heighten or hinder facility efficiency, and how efficiency is related to the quality of service provision, should be considered.

8

1110

Costsofcare

Trendsinaveragefacilityspendingbetween2007and2011variedbetweenfacilitytypes,thoughallplatformsrecordedhigherspend-ingin2011than2007.

• Spending on personnel accounted for the vast major-ity of annual spending across facility types. Compared to other facility types, sub-district hospitals and primary health centres put a slightly greater propor-tion of their total expenditure toward personnel, while district hospitals put the greatest proportion toward medical supplies.

Patientperspectives

Travelandwaittimeswereshorterfor patientsvisitinglower-levelfacilitiesthanhigher-levelones.

• Nearly all patients receiving care at sub health centres, and just under 80% of patients at primary health cen-tres, reported traveling less than 30 minutes to receive care. In contrast, nearly half of patients at district hos-pitals had travel times of over 30 minutes, reflecting the greater distances people travel to receive special-ist treatment from facilities of this type.

• The large majority of patients waited less than 30 minutes to receive care across all facilities. Nearly all patients seeking care at sub health centres received care in less than 30 minutes. Wait times were longer at hospitals, but overall less than 6% of patients waited more than one hour to receive care.

Patientsgavehigherratingsofhealthcareprovidersthanfacilitycharacteristics.

• Across all facility types except community and sub health centres, patients receiving care from doctors re-ported slightly higher levels of satisfaction than those treated by nurses. Satisfaction with staff interactions, for both doctors and nurses, were lowest at district hospitals and generally higher at health centres.

• Facility characteristics, such as cleanliness and privacy, received generally low ratings from patients. Cleanli-ness at hospitals received particularly low marks. As with staff interactions, patient satisfaction with facility characteristics was higher at health centres.

• Most patients received all drugs that they were pre-scribed during their visits. Proportions of patients receiving all prescribed drugs ranged from 99% of patients at sub-district hospitals and primary health centres to 80% at sub health centres.

With its multidimensional assessment of health ser-vice provision, findings from the ABCE project in Tamil Nadu provide an in-depth examination of health facility capacity, costs of care, and how patients view their in-teractions with the health system. Tamil Nadu’s health provision landscape was markedly heterogeneous and will likely continue to evolve over time. This highlights the need for continuous and timely assessment of health service delivery, which is critical for identifying areas of successful implementation and quickly responding to service disparities or faltering performance. Expanded analyses would also allow for an even clearer picture of the trends and drivers of facility capacity, efficiencies, and costs of care. With regularly collected and analyzed data, capturing information from health facilities, recipi-ents of care, policymakers, and program managers can yield the evidence base to make informed decisions for achieving optimal health system performance and the eq-uitable provision of cost-effective interventions throughout Tamil Nadu.

Introduction

The performance of a country’s health sys-tem ultimately shapes the health outcomes experienced by its population, influencing the ease or difficulty with which individuals

can seek care and facilities can address their needs. At a time when international aid is plateauing1 and the gov-ernment of India has prioritized expanding many health programs,2,3 identifying health system efficiencies and promoting the delivery of cost-effective interventions has become increasingly important.

Assessing health system performance is crucial to opti-mal policymaking and resource allocation; however, due to the multidimensionality of health system functions,4

comprehensive and detailed assessment seldom occurs. Rigorously measuring what factors are contributing to or hindering health system performance – access to services, bottlenecks in service delivery, costs of care, and equity in service provision – throughout a country provides crucial information for improving service delivery and popula-tion health outcomes.

The Access, Bottlenecks, Costs, and Equity (ABCE) project was launched globally in 2011 to address these gaps in information. In addition to India, the multi-pronged, multi-partner ABCE project has taken place in seven other countries (Bangladesh, Colombia, Ghana, Kenya, Lebanon, Uganda, and Zambia). In India, the ABCE project was undertaken in six states: Andhra Pradesh and Telangana, Gujarat, Madhya Pradesh, Odisha, and Tamil Nadu.

The ABCE project, with the goal of rigorously assessing the drivers of health service delivery across a range of set-tings and health systems, strives to answer these critical

1 Institute for Health Metrics and Evaluation (IHME). Financing Global Health 2015: Development assistance steady on the path to new Global Goals. Seattle, WA: IHME, 2016. 2 Planning Commission Government of India. Eleventh Five Year Plan (2007-12). New Delhi, India: Government of India, 2007. 3 Planning Commission Government of India. Twelfth Five Year Plan (2012-17). New Delhi, India: Government of India, 2012.4 Murray CJL, Frenk J. A Framework for Assessing the Performance of Health Systems. Bulletin of the World Health Organization. 2000; 78 (6): 717-731.

questions facing policymakers and health stakeholders in each country or state for public sector health care service delivery:

• What health services are provided, and where are they available?

• What are the bottlenecks in provision of these services?

• How much does it cost to produce health services?

• How efficient is provision of these health services?

Findings from each country’s ABCE work will pro-vide actionable data to inform their own policymaking processes and needs. Further, ongoing cross-country analyses will likely yield more global insights into health service delivery and costs of health care. These eight countries have been purposively selected for the overar-ching ABCE project as they capture the diversity of health system structures, composition of providers (public and private), and disease burden profiles. The ABCE project contributes to the global evidence base on the costs of and capacity for health service provision, aiming to de-velop data-driven and flexible policy tools that can be adapted to the particular demands of governments, de-velopment partners, and international agencies.

The Public Health Foundation of India (PHFI) and the Institute for Health Metrics and Evaluation (IHME) com-pose the core team for the ABCE project in India, and they received vital support and inputs from the state Ministry of Health and Family Welfare for data collection, analysis, and interpretation. The core team harnessed information from distinct but linkable sources of data, drawing from a state-representative sample of health facilities to create a large and fine-grained database of facility attributes, ex-penditure, and capacity, and patient characteristics and outcomes. By capturing the interactions between facility characteristics and patient perceptions of care, we have been able to piece together what factors drive or hin-der optimal and equitable service provision in rigorous, data-driven ways.

1312


We focus on the facility because health facilities are the main points through which most individuals interact with the health system or receive care. Understanding the ca-pacities and efficiencies within and across different types of public sector health facilities unveils the differences in health system performance at the level most critical to patients – the facility level. We believe this information is immensely valuable to governments and development partners, particularly for decisions on budget alloca-tions. By having data on what factors are related to high facility performance and improved health outcomes, pol-icymakers and development partners can then support evidence-driven proposals and fund the replication of these strategies at facilities throughout India.

The ABCE project in India has sought to generate the evidence base for improving the cost-effectiveness and equity of health service provision. In this report, we ex-amine facility capacity across platforms, as well as the efficiencies and costs associated with service provision for each type of facility. Based on patient exit interviews, we

consider the factors that affect patient perceptions of and experiences with the state’s health system. By considering a range of factors that influence health service delivery, we have constructed a nuanced understanding of what helps and hinders the receipt of health services through facilities in the state of Tamil Nadu.

The results discussed in this report are far from ex-haustive; rather, they align with identified priorities for health service provision and aim to answer questions about the costs of health care delivery in the respective state in India. This report provides an in-depth examina-tion of health facility capacity across different platforms, specifically covering topics on human resource capacity, facility-based infrastructure and equipment, health ser-vice availability, patient volume, facility-based efficiencies, costs associated with service provision, and demand-side factors of health service delivery as captured by patient exit interviews.

Table 2 defines the cornerstone concepts of the ABCE project: Access, Bottlenecks, Costs, and Equity.

AccessHealth services cannot benefit populations if they cannot be accessed; thus, measuring which elements are driving improved access to – or hindering contact with – health facilities is critical. Travel time to facilities, user fees, and cultural preferences are examples of factors that can affect access to health systems.

BottlenecksMere access to health facilities and the services they provide is not sufficient for the delivery of care to popula-tions. People who seek health services may experience supply-side limitations, such as medicine stockouts, that prevent the receipt of proper care upon arriving at a facility.

CostsHealth services cost can translate into very different financial burdens for consumers and providers of such care. Thus, the ABCE project measures these costs at several levels, quantifying what facilities spend to provide services.

EquityVarious factors influence how populations interact with a health system. The nature of these interactions either facilitates or obstructs access to health services. In addition to knowing the cost of scaling up a given set of services, it is necessary to understand costs of scale-up for specific populations and across population-related factors (e.g., distance to health facilities). The ABCE project aims to pinpoint which factors affect the access to and use of health services and to quantify how these factors manifest.

Access, Bottlenecks, Costs, and Equity

Table2 Access, Bottlenecks, Costs, and Equity

F

ABCE project design

or the ABCE pro jec t in Ind ia , we conducted primary data collection through a two-pronged approach:

1. A comprehensive facility survey administered to a representative sample of health facilities in select states in India (the ABCE Facility Survey)

2. Interviews with patients as they exited the sampled facilities

Here, we provide an overview of the ABCE sur-vey design and primary data collection mechanisms. All ABCE survey instruments are available online at http://www.healthdata.org/dcpn/india.

ABCEFacilitySurveyThrough the ABCE Facility Survey, direct data collec-

tion was conducted from a state-representative sample of health service platforms and captured information on the following indicators for the five fiscal years (running from April to March of the following year) prior to the survey:

• Inputs: the availability of tangible items that are needed to provide health services, including in-frastructure and utilities, medical supplies and equipment, pharmaceuticals, personnel, and non-medical services.

• Finances: expenses incurred, including spending on infrastructure and administration, medical supplies and equipment, pharmaceuticals including vaccines, and personnel. Facility funding from different sources (e.g., central and state governments) and revenue from service provision were also captured.

• Outputs: volume of services and procedures pro-duced, including outpatient and inpatient care, emergency care, and laboratory and diagnostic tests.

• Supply-sideconstraintsandbottlenecks: factors that affected the ease or difficulty with which patients received services they sought, including bed avail-ability, pharmaceutical availability and stockouts, cold-chain capacity, personnel availability, and service availability.

Table 3 provides more information on the specific indicators included in the ABCE Facility Survey.

1514

A B C E I N TA M I L N A D U A B C E P R OJ E C T D E S I G N

Figure2 Sampling strategy for health facilities in a district in the ABCE survey in India

Selected facilities are in blue; unselected facilities from the sampling frame are in grey.DH: District hospital; SDH: Sub-district hospital; CHC: Community health centre; PHC: Primary health centre; SHC: Sub health centre

SURVEY MODULE SURVEY CATEGORY KEY INDICATORS AND VARIABLES

Module1:Facilityfinances andinputs

Inputs Input funding sources, managing authority, and maintenance information

Availability and functionality of medical and non-medical equipment

Finances Salary/wages, benefits, and allowances

Total expenses for infrastructure and utilities; medical supplies and equipment; pharmaceuticals; administration and training; non-medical services, personnel (salaries and wages, benefits, allowances)

Performance and performance-based financing questions

Revenue User fees; total revenue and source

Personnel characteristics Total personnel by cadre

Funding sources of personnel

Health services provided and their staffing; administrative and support services and their staffing

Module2:Facilitymanagementanddirectobservation

Facility management and infrastructure characteristics

Characteristics of patient rooms; electricity, water, and sanitation

Facility meeting characteristics

Guideline observation

Direct observation Latitude, longitude, and elevation of facility. Facility hours, characteristics, and location; waiting and examination room characteristics

Module3:Lab-based consumables, equipment, andcapacity

Facility capacity Lab-based tests available

Medical consumables and equipment

Lab-based medical consumables and supplies available

Module4:Pharmaceuticals

Facility capacity Drug availability and stockout information

Module5:Generalmedical consumables, equipment, andcapacity

Medical consumables and equipment

Availability and functionality of medical furniture, equipment, and supplies

Inventory of procedures for sterilization, sharp items, and infectious waste

Inventory of personnel

Module6:Facilityoutputs

Facility capacity Fund and vehicle availability for referral and emergency referral

General service provision Inpatient care and visits; outpatient care and visits; emergency visits; home or outreach visits

Laboratory and diagnostic tests

Module7:Vaccines

Facility procedures for vaccine supply, delivery and disposal

Source from vaccine obtained

Personnel administering vaccine

Procedures to review adverse events

Disposal of vaccines

Vaccine availability, storage, and output

Stock availability and stockouts of vaccines and syringes

Types and functionality of storage equipment for vaccines

Temperature chart history; vaccine inventory and vaccine outputs; vaccine outreach and home visits

Vaccine sessions planned and held

Table3Modules included in the ABCE Facility Survey in India Figure1 Sampled districts in Tamil NaduSampledesignA total of 13 districts in Tamil Nadu were selected for

the ABCE survey (Figure 1). The districts were selected us-ing three strata to maximize heterogeneity: proportion of full immunization in children aged 12–23 months as an in-dicator of preventive health services; proportion of safe delivery (institutional delivery or home delivery assisted by skilled person) as an indicator of acute health services; and proportion of urban population as an indicator of overall development. The districts were grouped as high and low for urbanization based on median value, and into three equal groups as high, medium, and low for the safe delivery and full immunization indicators. Twelve districts were selected randomly from each of the various combi-nations of indicators, and in addition the capital district was selected purposively.

Within each sampled district, we then sampled pub-lic sector health facilities at all levels of services based on the structure of the state health system (Figure 2).

1716


Table5 Facility sample, by platform, for the ABCE project in Tamil Nadu

FACILITY TYPE FINAL SAMPLE

Districthospital 13

Sub-districthospital 26

Communityhealthcentre 24

Primaryhealthcentre 54

Subhealthcentre 51

Totalhealthfacilities 168

Table4Types of questions included in the Patient Exit Interview Survey in India

SURVEY CATEGORY TYPES OF KEY QUESTIONS AND RESPONSE OPTIONS

Directobservationofpatient Sex of patient (and of patient’s attendant if surveyed)

Directinterviewwithpatient Demographic questions (e.g., age, level of education attained, caste)

Scaled-response satisfaction scores (e.g., satisfaction with medical doctor)

Open-ended questions for circumstances and reasons for facility visit, as well as visit characteristics (e.g., travel time to facility)

Reporting costs associated with facility visit (user fees, medications, transportation, tests, other), with an answer of “yes” prompting follow-up questions pertaining to amount

In each sampled district, one district hospital (DH); all sub-district hospitals (SDH, from a total of zero to three) for each sampled DH; two community health centres (CHC, from a total of two to five) for each sampled SDH; two primary health centres (PHC, from a total of two to four) for each sampled CHC; and one sub centre (SHC, from a total of one to four) for each sampled PHC were randomly selected for the study.

PatientexitinterviewsurveyA fixed number patients or attendants of patients were

interviewed at each facility, based on the expected out-patient density for the platform. A target of 16 patients were interviewed at district hospitals, 16 at SDH, 12 at CHC, 10 at PHC, and five at SHC. Patient selection was based on a convenience sample. The main purpose of the Pa-tient Exit Interview Survey was to collect information on patient perceptions of the health services they received and other aspects of their facility visit (e.g., travel time to facility, costs incurred during the facility visit, and sat-isfaction with the health care provider). Table 4 provides more information on the specific indicators included in the exit survey. This information fed into quantifying the “demand-side” constraints to receiving care (as opposed to the facility-based, “supply-side” constraints and bottle-necks measured by the ABCE Facility Survey).

DatacollectionfortheABCEsurvey inTNData collection took place from October 2012 to Jan-

uary 2013. Prior to survey implementation, PHFI and the data collection agency hosted a two-week training work-shop for 30 interviewers, where they received extensive training on the electronic data collection software (Dat-Stat), the survey instruments, the Tamil Nadu health system’s organization, and interviewing techniques. Following this workshop, a one-week pilot of all survey in-struments took place at health facilities. Ongoing training occurred on an as-needed basis throughout the course of data collection.

All collected data went through a thorough verification process between PHFI and IHME and the ABCE field team. Following data collection, the data were methodically cleaned and re-verified, and securely stored in databases hosted at PHFI and IHME.

A total of 168 health facilities participated in the ABCE project in Tamil Nadu. Eleven facilities were replaced (one DH, one SDH, two CHCs, one PHC, and six SHCs) due to data being unavailable for the years considered; the re-porting chain of the sampled facility being incorrect; or the facility having been functional for less duration.

A B C E P R OJ E C T D E S I G N

1918

M A I N F I N D I N G S : H E A LT H FA C I L I T Y P R O F I L E S

Main findingsHealth facility profiles

The delivery of facility-based health ser-vices requires a complex combination of resources, ranging from personnel to phys-ical infrastructure, that vary in their relative

importance and cost to facilities. Determining what fac-tors support the provision of services at lower costs and higher levels of efficiency at health facilities is critical in-formation for policymakers to expand health system coverage and functions within constrained budgets.

Using the ABCE TN facility sample (Table 5), we analyzed five key drivers of health service provision at facilities:

• Facility-based resources (e.g., human resources, infrastructure and equipment, and pharmaceuticals), which are often referred to as facility inputs.

• Patient volumes and services provided at facilities (e.g., outpatient visits, inpatient bed-days), which are also known as facility outputs.

• Patient-reported experiences, capturing “de-mand-side” factors of health service delivery.

• Facility alignment of resources and service production, which reflects efficiency.

• Facility expenditures and production costs for service delivery.

These components build upon each other to create a comprehensive understanding of health facilities in TN, highlighting areas of high performance and areas for improvement.

Facilitycapacityandcharacteristics

ServiceavailabilityAcross and within district hospitals, sub-district hospi-

tals and community health centres in TN (Table 6), several notable findings emerged for facility-based health service provision. While fundamental services such as antenatal care, routine deliveries, general medicine, and pharmacy and laboratory services were nearly universally available,

fewer facilities reported available services for non-com-municable diseases, such as cardiology, psychiatry, and chemotherapy, particularly at the sub-district and com-munity levels. District hospitals reported a wide range of services such as blood banks, surgical services, den-tistry, and emergency obstetrics. Sub-district hospitals generally offered fewer services than district hospitals but still reported high coverage of services like obstetrics and gynecology and accident and emergency services. One-quarter of community health centres reported that they did not provide surgical services or emergency obstetrics, while only around half provided dentistry and anesthesiology.

HumanresourcesforhealthA facility’s staff size and composition directly affect

the types of services it provides. In general, a greater availability of health workers is related to higher service utilization and better health outcomes.1 India has a severe shortage of qualified health workers and the workforce is concentrated in urban areas.2 The public health system has a shortage of both medical and paramedical per-sonnel. The number of primary and community health centres without adequate staff is substantially higher if high health-worker absenteeism is taken into consider-ation.3 The Indian Government is aware of the additional requirements and shortages in the availability of health workers for the future. The National Rural Health Mission, for instance, recommends a vastly strengthened infra-structure, with substantial increases in personnel at every tier of the public health system.4

Based on the ABCE sample, we found substantial het-

1 Rao KD, Bhatnagar A, Berman P. So many, yet few: Human resources for health in India. Human Resources for Health. 2012; 10(19). 2 Rao M, Rao KD, Kumar AK, Chatterjee M, Sundararaman T. Human resources for health in India. The Lancet. 2011; 377(9765): 587-98.3 Hammer J, Aiyar Y, Samji S. Understanding government failure in public health services. Economic and Political Weekly. 2007; 42: 4049–58.4 National Rural Health Mission. Ministry of Health and Family Welfare, Govern-ment of India. Mission Document (2005-2012). New Delhi, India: Government of India, 2005.

Table6Availability of services in health facilities, by platform

DISTRICT HOSPITAL (DH)

SUB-DISTRICT HOSPITAL (SDH)

COMMUNITY HEALTH CENTRE (CHC)

Total obstetrics and gynecology services 100% 100% 100%

Routine births 100% 100% 100%

Emergency obstetrics 100% 88% 75%

Antenatal care 100% 100% 100%

Surgical services 100% 92% 75%

Cardiology 77% 38% 8%

Psychiatric 77% 19% 0%

Accident, trauma, and emergency 100% 100% 88%

Ophthalmology 100% 35% 75%

Pediatric 100% 92% 88%

General anesthesiology 100% 69% 54%

Blood bank 100% 27% 8%

Dentistry 100% 77% 46%

DOTS treatment 92% 81% 82%

STI/HIV 100% 69% 83%

Immunization 100% 77% 92%

Internal/general med 100% 100% 96%

Mortuary 100% 85% 0%

Burns 100% 58% 21%

Orthopedic 100% 31% 0%

Pharmacy 100% 100% 100%

Chemotherapy 8% 4% 4%

Dermatology 69% 23% 4%

Alternative medicine 100% 100% 92%

Diagnostic medical 100% 100% 83%

Laboratory services 100% 96% 100%

Outreach services 0% 15% 88%

LOWEST AVAILABILITY HIGHEST AVAILABILITY

Note: All values represent the percentage of facilities, by platform, that reported offering a given service at least one day during a typical week.

erogeneity across facility types in TN by considering the total number of staff in the context of bed strength (i.e., number of beds in the facility) and patient load (Figure 3). Overall, the most common medical staff at district hospi-tals were nurses (63) followed by paramedical staff (58), non-medical staff (51), and doctors (40), while at lower lev-els, paramedical staff outnumbered doctors and nurses. This is reflection of the differential service offerings be-tween higher- and lower-level facilities. Additionally, higher-level facilities tended to have a greater number of health personnel overall. While a degree of this variation is due to differences in service provision and population size, some of this indicates relative shortages in human re-sources for health.

The greatest number of doctors, nurses, paramedi-cal staff, and non-medical staff are concentrated at the district hospitals (average of 211 total staff). Sub-district hospitals reported the second highest number of per-sonnel; however, the total personnel at these facilities was one-quarter of what was reported by district hospi-tals (average of 52 total staff). Community health centres maintained a smaller body of health workers, an average total of 32, with most of the medical staff being paramed-ical (17). Primary health centres reported, on average, 18 staff in total, most of which were paramedical staff (11). Fi-nally, sub-health centres reported one paramedical staff who performs immunizations, simple outpatient care, and community outreach.

Figure3 Composition of facility personnel, by platform

0 50 100 150 200Number of Staff

Sub Health Centre

Primary Health Centre

Community Health Centre

Sub District Hospital

District Hospital

TN

Doctors Nurses

Para-medical staff Non-medical staff

2120

A B C E I N TA M I L N A D U M A I N F I N D I N G S : H E A LT H FA C I L I T Y P R O F I L E S

NursestodoctorsratioThe ratio of number of nurses to number of doctors

is presented in Figure 4. A ratio greater than 1 indicates that nurses outnumber doctors; for instance, a ratio of 2 indicates that there are two nurses staffed for every one doctor. Alternatively, a ratio lower than 1 indicates that doctors outnumber nurses; for instance, a ratio of 0.5 in-dicates there is one nurse staffed for every two doctors.

District hospitals reported an average ratio of 1.8, indi-cating that they staff more nurses than doctors. However, the ratio reported by various district hospitals ranged from a low of 0.5 to a high outlier of 6.4. All but two sub-district hospitals reported more nurses than doc-tors, with a ratio as high as 4.0 nurses to doctors. There was less heterogeneity among community health centres, with ratios ranging from 0.4 to 2.5. Finally, primary health centres reported a wide range of ratios, from 0.3 to 4.0. The average ratio of nurses to doctors was similar for dis-trict hospitals (1.8), sub-district hospitals (1.7), and primary health centres (1.7).

Nursesanddoctorstoparamedicalandnon-medicalstaffThe ratio of number of nurses and/or doctors to num-

ber of paramedical and/or non-medical staff in 2012 is presented in Figure 5. A ratio greater than 1 indicates that nurses and doctors outnumber para-medical and

Figure4 Ratio of nurses and ANMs to doctors, by platform

Vertical bars represent the platform average ratio.

non-medical personnel; for instance, a ratio of 2 indicates that there are two nurses and/or doctors staffed for every one paramedical/non-medical staff. Alternatively, a ratio lower than 1 indicates that para-medical and/or non-med-ical personnel outnumber nurses and/or doctors.

The average ratio for district hospitals and sub-district hospitals was 1.0, though the range of ratios for dis-trict hospitals (0.7 to 2.0) was slightly narrower than for sub-district hospitals (0.7 to 2.7). Community health cen-tres were more homogenous, reporting an average ratio of 0.4, with facilities reporting ratios that ranged from 0.2 to 0.6. The ratio for primary health centres ranged from 0.1 to 1.3, with an average of 0.4 doctors and nurses to paramedical and non-medical staff.

BedstodoctorsratioThe ratio of number of beds to number of doctors in

2011 is presented in Figure 6. A ratio greater than 1 indi-cates that beds outnumber doctors; for instance, a ratio of 2 indicates that there are two beds for every one doc-tor staffed. Alternatively, a ratio lower than 1 indicates that doctors outnumber beds.

The average ratio of beds to doctors is highest in dis-trict hospitals (9.7), followed by sub-district hospitals (8.0). Community health centres have an average of 5.2 beds per doctor, though four facilities have ratios above 10.0. The average ratio among primary health centres is 2.4,

Figure5Ratio of nurses and doctors to paramedical and non-medical staff, by platform


with a range of 0.3 to 9.0. Two primary health centres re-ported fewer beds than doctors.

BedstonursesratioThe ratio of number of beds to number of nurses in

2011 is presented in Figure 7. A ratio greater than 1 indi-cates that beds outnumber nurses; for instance, a ratio of 2 indicates that there are two beds for every one nurse staffed. Alternatively, a ratio lower than 1 indicates that nurses outnumber beds.

Similar to the ratio of beds to doctors, the ratio of beds to nurses was highest among district hospitals (5.9) and lowest among primary health centres (1.8). While sub-dis-trict hospitals and community health centres had a similar average ratio of beds to nurses (5.0 and 5.2, respectively), the range of ratios was much wider for community health centres (1.0 to 12.3) than for sub-district hospitals (2.3 to 8.2).

In isolation, facility staffing numbers are less meaning-ful without considering a facility’s overall patient volume and production of specific services. For instance, if a fa-cility mostly offers services that do not require a doctor’s administration, failing to achieve the doctor staffing tar-get may be less important than having too few nurses. Further, some facilities may have much smaller patient volumes than others, and thus “achieving” staffing tar-

Figure6Ratio of beds to doctors by platform


gets could leave them with an excess of personnel given patient loads. While an overstaffed facility has a different set of challenges than an understaffed one, each reflects a poor alignment of facility resources and patient needs. To better understand bottlenecks in service delivery and areas to improve costs, it is important to assess a facili-ty’s capacity (inputs) in the context of its patient volume and services (outputs). We further explore these find-ings in the “Efficiency and costs” section. As part of the ABCE project in India, we compare levels of facility-based staffing with the production of different types of health services. In this report, we primarily focus on the delivery of health services by skilled medical personnel, which include doctors, nurses, and other paramedical staff. It is possible that non-medical staff also contribute to ser-vice provision, especially at lower levels of care, but the ABCE project in India is not currently positioned to ana-lyze these scenarios.

InfrastructureandequipmentHealth service provision depends on the availability of

adequate facility infrastructure, equipment, and supplies (physical capital). In this report, we focus on four essen-tial components of physical capital: power supply, water and sanitation, transportation, and medical equipment, with the latter composed of laboratory, imaging, and other medical equipment. Table 7 illustrates the range of

Figure7Ratio of beds to nurses, by platform


0 2 4 6

District Hospital Sub District Hospital

Community Health Centre Primary Health Centre

0 5 10 15 20 25



0 5 10 15 20 25



2322


physical capital, excluding medical equipment, available across platforms.

Power supplyAll hospitals, community health centres, and primary

health centres reported access to a functional electrical supply, and just 8% of sub health centres lacked func-tional electricity (Table 7). One facility reported relying solely on a generator for power. Inadequate access to consistent electric power has substantial implications for health service provision, particularly for the effective storage of medications, vaccines, and blood samples, and these results demonstrate an improvement in the availability of electricity at the lowest platform levels compared to 2007–2008, when 87% of primary health centres and 70% of sub health centres had a regular electric supply.5

Waterandsanitation85% of district hospitals had availability of improved

water sources (functional piped water) and improved sanitation with a functional sewer infrastructure with

5 International Institute for Population Sciences (IIPS). District Level Household and Facility Survey (DLHS-3), 2007-08: India, Tamil Nadu. Mumbai, India: IIPS, 2010.

flush toilets (Table 7). Notably, these figures were higher in sub-district hospitals and community health centres, and even in primary health centres. At the lowest plat-form level, the sub health centre, access to improved water sources and sanitation was significantly lower. Hand disinfectant was broadly available as a supplementary sanitation method at most platform levels, though it was not available at many sub health centres. Among all fa-cilities, 13% reported a severe shortage of water at some point during the year. These findings show a mixture of both notable gains and ongoing needs for facility-based water sources and sanitation practices among both hospi-tals and primary care facilities.

TransportationandcomputersFacility-based transportation and modes of commu-

nication varied across platforms (Table 7). In general, the availability of a vehicle decreased down the levels of the health platform, though more community health centres reported having a vehicle than did sub-district hospitals. Only around one-fifth of primary health centres had any four-wheeled vehicles at all, which means transferring patients under emergency circumstances from these facilities could be fraught with delays and possible com-plications. Community health centres had a relatively high level of transport availability, with 96% having a four-

Table7 Availability of physical capital, by platform




PRIMARY HEALTH CENTRE (PHC)

SUB HEALTH CENTRE (SHC)

Functionalelectricity 100% 100% 100% 100% 92%

Pipedwater 85% 88% 88% 87% 59%

Flushtoilet 85% 96% 96% 91% 65%

Handdisinfectant 92% 96% 96% 91% 43%

Any4-wheeledvehicle 100% 35% 96% 22% NA

Emergency4-wheeledvehicle 69% 23% 58% NA NA

Landlinephone 100% 96% 100% 100% 90%

Computer 100% 100% 100% 98% NA

NA: Not applicable to this platform according to standards.


Note: Values represent the percentage of facilities, by platform, that had a given type of physical capital.

Table8Availability of functional equipment, by platform

DISTRICT HOSPITAL

SUB-DISTRICT HOSPITAL

COMMUNITY HEALTH CENTRE

PRIMARY HEALTH CENTRE

SUB-HEALTH CENTRE

Medicalequipment

Wheelchair 92% 100% 100% 100% NA

Adult scale 92% 88% 92% 91% 96%

Child scale 92% 96% 96% 93% 82%

Blood pressure apparatus 92% 100% 100% 100% 92%

Stethoscope 92% 100% 100% 100% 98%

Light source 85% 100% 92% 96% 59%

Labequipment

Glucometer 69% 42% 92% 89% 45%

Test strips for glucometer 69% 42% 83% 72% 35%

Hematologic counter 85% 42% 21% 13% NA

Blood chemistry analyzer 92% 69% 75% 59% NA

Incubator 92% 58% 13% 7% NA

Centrifuge 92% 88% 96% 74% NA

Microscope 92% 96% 92% 56% NA

Slides 92% 92% 100% 98% 47%

Slide covers 92% 88% 96% 89% 33%

Imagingequipment

X-ray 92% 77% 38% NA NA

ECG 92% 85% 92% NA NA

Ultrasound 92% 96% 88% NA NA

CT scan 38% 4% NA NA NA



Note: Availability of a particular piece of equipment was determined based on facility ownership on the day of visit. Data on the number of items present in a facility were not collected. All values represent the percentage of facilities, by platform, that had a given piece of equipment.

2524


Table9Availability of tests and functional equipment to perform routine antenatal care, by platform

DISTRICT HOSPITAL




SUB HEALTH CENTRE

Testingavailability

Urinalysis 92% 96% 92% 48% 16%

Hemoglobin 92% 100% 96% 65% 14%

Glucometer and test strips 69% 38% 79% 72% 33%

Blood typing 92% 92% 92% 54% NA

Functionalequipment

Blood pressure apparatus 92% 100% 100% 100% 92%

Adult scale 92% 88% 92% 91% 86%

Ultrasound 92% 96% 88% NA NA

Servicesummary

Facilities reporting ANC services 100% 100% 100% 100% 95%

Facilities fully equipped for ANC provision based on above tests and equipment availability

69% 27% 63% 31% 5%



Note: Availability of a given ANC item was determined by its availability at a facility on the day of visit. All values represent the percentage of facilities, by platform that had the given ANC item. The service summary section compares the total percentage of facilities reporting that they provided ANC services with the total percentage of facilities that carried all of the functional equipment to provide ANC services.

wheeled vehicle and 58% having dedicated emergency transportation. Alongside transportation, communication is also a necessary facet of the efficient delivery of health services. The availability of modes of communication was generally high at all facility levels: nearly all facilities re-ported having a landline phone, and computer facilities were widely available across platforms.

EquipmentFor three main types of facility equipment – medical,

lab, and imaging – clear differences emerge across levels of health service provision, with Table 8 summarizing the availability of functional equipment by platform.

We used the WHO’s Service Availability and Readi-ness Assessment (SARA) survey as our guideline for what

types of equipment should be available in hospitals and primary care facilities.6 Table 8 illustrates the distribution of SARA scores across platforms. In general, hospitals had greater availability of medical equipment, and defi-cits in essential equipment availability were found in the lower levels of care. Lacking scales and blood pressure cuffs can severely limit the collection of important patient clinical data; these were generally available, but facilities at all levels reported missing some of these vital pieces of equipment.

Microscopes and corresponding components were largely prevalent among all facilities, except at primary health centres where many reported having slides but al-most half had no microscope to use them with. Additional testing capacity was relatively high in district hospitals but

6 World Health Organization (WHO). Service Availability and Readiness Assessment (SARA) Survey: Core Questionnaire. Geneva, Switzerland: WHO, 2013.

Table10 Availability of blood tests and functional equipment to perform routine delivery care, by platform

DISTRICT HOSPITAL




Testingavailability

Hemoglobin 92% 100% 96% 65%

Glucometer and test strips 69% 38% 79% 72%

Cross-match blood 92% 31% NA NA

Medicalequipment

Blood pressure apparatus 92% 100% 100% 100%

IV catheters 92% 100% 100% 98%

Gowns 92% 96% 100% 91%

Measuring tape 92% 100% 100% 98%

Masks 85% 92% 96% 83%

Sterilization equipment 92% 92% 83% 57%

Adult bag valve mask 92% 88% 96% 59%

Ultrasound 92% 96% 88% NA

Deliveryequipment

Infant scale 85% 92% 96% 94%

Scissors or blade 92% 96% 100% 100%

Needle holder 92% 100% 100% 98%

Speculum 92% 96% 100% 100%

Forceps 92% 92% 92% 85%

Dilation and curettage kit 92% 88% 67% 59%

Neonatal bag valve mask 92% 92% 100% 96%

Vacuum extractor 92% 50% 42% 30%

Incubator 85% 42% 21% 22%

Servicesummary

Facilities reporting delivery services 100% 100% 100% 100%

Facilities fully equipped for delivery services based on above tests and equipment availability 54% 4% 0% 0%



Note: Availability of a given delivery item was determined by its availability at a facility on the day of visit. All values represent the percentage of facilities, by platform, that had the given delivery item. The service summary section compares the total percentage of facilities reporting that they provided routine delivery services with the total percentage of facilities that carried all of the recommended pharmaceuticals and functional equipment to provide routine delivery services.

2726


poorer in sub-district hospitals, community health centres, and primary health centres. There were some exceptions to this trend: while 92% of community health centres and 89% of primary health centres had a glucometer, only 69% of district hospitals and 42% of sub-district hospitals had one. Additionally, blood chemistry analyzers were available in only 69% of sub-district hospitals. This indi-cates limited capacity for addressing non-communicable diseases (NCDs) such as diabetes, for which this equip-ment is necessary. Other essential equipment, including hematologic counters and incubators, were notably missing from community health centres and primary health centres.

District hospitals had good availability of imaging equipment, with the notable exception of CT scans, which were available in 38% of facilities. Sub-district hospitals showed somewhat patchier availability of imaging equip-ment, with 77% reporting the availability of X-ray and only 4% having CT scanners. Community health centres had relatively high availability of essential imaging equip-

ment, with the exception of X-ray, which was available in 38% of facilities.

Overall, these findings demonstrate gradual improve-ments in equipping health facilities with basic medical equipment in TN, as well as the continued challenge of ensuring that these facilities carry the supplies they need to provide a full range of services. Measuring the avail-ability of individual pieces of equipment sheds light on specific deficits, but assessing a health facility’s full stock of necessary or recommended equipment provides a more precise understanding of a facility’s service capacity.

FocusonserviceprovisionFor the production of any given health service, a

health facility requires a complex combination of the ba-sic infrastructure, equipment, and pharmaceuticals, with personnel who are adequately trained to administer nec-essary clinical assessments, tests, and medications. Thus, it is important to consider this intersection of facility re-sources to best understand facility capacity for care. In

Table11Availability of blood tests and functional equipment to perform general surgery, by platform




PRIMARY HEALTH CENTRE (PHC)

Testingavailability

Hemoglobin 92% 100% 96% 65%


Medicalequipment

Blood pressure apparatus 92% 100% 100% 100%

IV catheters 92% 100% 100% 98%

Sterilization equipment 92% 92% 83% 57%

Gowns 92% 96% 100% 91%

Masks 85% 92% 96% 83%

Adult bag valve mask 92% 88% 96% 59%

Surgicalequipment

Scissors 92% 96% 100% 100%

Thermometer 92% 96% 96% 78%

General anesthesia equipment 92% 81% 71% 6%

Scalpel 92% 96% 96% 67%

Suction apparatus 92% 96% 88% 74%

Retractor 92% 88% 75% 28%

Nasogastric tube 92% 92% 88% 57%

Blood storage unit/refrigerator 92% 46% 25% NA

Intubation equipment 92% 81% 67% 31%

Servicesummary

Facilities reporting general surgery services 100% 92% 75% 46%

Facilities fully equipped for general surgery services based on above tests and equipment availability 85% 21% 0% 0%



Note: Availability of a given surgery item was determined by its availability at a facility on the day of visit. All values represent the percentage of facilities, by platform, that had the given surgery item. The service summary section compares the total percentage of facilities reporting that they provided general surgery services with the total percentage of facilities that carried all of the recommended functional equipment to provide general surgery services.

Table12 Availability of laboratory tests, by platform


SUB-DISTRICT HOSPITAL(SDH)

COMMUNITY HEALTH CENTRE(CHC)

PRIMARY HEALTH CENTRE(PHC)

Blood typing 92% 92% 92% 54%


Complete blood count 92% 81% 17% 2%

Hemoglobin 92% 100% 96% 65%

HIV 92% 85% 88% 20%

Liver function 92% 35% 8% NA

Malaria 85% 73% 96% 56%

Renal function 92% 42% 4% 2%

Serum electrolytes 62% 8% 8% NA

Spinal fluid test 46% 4% 0% NA

Syphilis 92% 85% 42% NA

Tuberculosis skin 92% 88% 92% 20%

Urinalysis 92% 96% 92% 48%



Note: Availability of a given test was determined by its availability at a facility on the day of visit. All values represent the percentage of facilities, by platform, that had the given test.

2928


this report, we further examined facility capacity for a subset of specific services – antenatal care, delivery, gen-eral surgery, and laboratory testing. For these analyses of service provision, we only included facilities that re-ported providing the specific service, excluding facilities that were potentially supposed to provide a given service but did not report providing it in the ABCE Facility Sur-vey. Thus, our findings reflect more of a service capacity “ceiling” across platforms, as we are not reporting on the facilities that likely should provide a given service but have indicated otherwise on the ABCE Facility Survey.

AntenatalcareservicesIn TN, according to the National Family Health Sur-

vey-4, 81% of women had at least four antenatal care (ANC) visits during their last pregnancy.7 This figure, though, neither reflects what services were actually pro-vided nor the quality of care received. Through the ABCE Facility Survey, we estimated what proportion of facilities stocked the range of tests and medical equipment to con-duct a routine ANC visit. It is important to note that this list was not exhaustive but represented a number of relevant supplies necessary for the provision of ANC.

7 International Institute for Population Sciences (IIPS). National Family Health Survey (NFHS-4), 2015-2016: Tamil Nadu Factsheet. Mumbai, India: IIPS, 2016.

Figure8 Number of outpatient visits, by platform

Note: Each line represents outpatient visits for an individual facility, with the bold line depicting the average for the platform. Scales are different for each platform.

050

000

1000

0015

0000

2000

00V

isits

2007 2008 2009 2010 2011

OP visits by facility OP visits average

CHC

050

0000

1000

000

1500

000

Vis

its

2007 2008 2009 2010 2011


DH

2000

040

000

6000

080

000

1000

00V

isits

2007 2008 2009 2010 2011


PHC0

2000

0040

0000

6000

0080

0000

Vis

its

2007 2008 2009 2010 2011


SDH

020

0040

0060

0080

0010

000

Vis

its

2007 2008 2009 2010 2011


SHC

Figure9 Number of inpatient visits (excluding deliveries), by platform

Note: Each line represents inpatient visits for an individual facility, with the bold line depicting the average for the platform. Scales are different for each platform.

010

0020

0030

00V

isits

2007 2008 2009 2010 2011

IP visits by facility IP visits average

CHC

050

000

1000

0015

0000

Vis

its

2007 2008 2009 2010 2011


DH

050

010

0015

0020

00V

isits

2007 2008 2009 2010 2011


PHC

050

000

1000

0015

0000

Vis

its

2007 2008 2009 2010 2011


SDH

3130


The availability of tests and functional equipment for ANC is presented in Table 9. While all hospitals, com-munity health centres, and primary health centres in this survey reported providing ANC services, few were ad-equately supplied for care. This discrepancy was most striking with sub-district hospitals, where only 27% of fa-cilities were fully equipped to provide ANC, largely due to the fact that just 38% carried glucometers and test strips. One-third of district hospitals were not fully equipped, again due to the lack of a functional glucometer and strips. There was a paucity of testing availability at primary and sub health centres. In general, however, availability of

functional equipment was fairly high. Across the levels of care, we found a substantial gap

between facility-reported capacity for ANC provision and the fraction of the facilities fully equipped to deliver ANC care. This service-capacity gap meant that many facilities, from district hospitals to the lower levels of care, reported providing ANC but then lacked at least one piece of the functional equipment needed to optimally address the range of patient needs during an ANC visit. Lack of sim-ple tests or material for tests (such as glucometer and test strips) prevented most facilities from being listed as fully equipped to provide ANC services. These findings do not

Figure10 Number of deliveries, by platform

Note: Each line represents deliveries for an individual facility, with the bold line depicting the average for the platform. Scales are different for each platform.

020

040

060

080

010

00D

eliv

erie

s

2007 2008 2009 2010 2011

Deliveries by facility Deliveries average

CHC

050

0010

000

1500

0D

eliv

erie

s

2007 2008 2009 2010 2011


DH

010

020

030

040

050

0D

eliv

erie

s

2007 2008 2009 2010 2011


PHC

010

0020

0030

0040

00D

eliv

erie

s

2007 2008 2009 2010 2011


SDH

Figure11 Number of immunization doses administered, by platform

Note: Each line represents immunization doses for an individual facility, with the bold line depicting the average for the platform. Scales are different for each platform.

050

000

1000

0015

0000

Dos

es a

dmin

iste

red

2007 2008 2009 2010 2011

Immunization doses by facility Immunization doses average

DH

050

0010

000

1500

020

000

Dos

es a

dmin

iste

red

2007 2008 2009 2010 2011


CHC

050

010

0015

0020

0025

00D

oses

adm

inis

tere

d

2007 2008 2009 2010 2011


SHC

010

0000

2000

0030

0000

4000

00D

oses

adm

inis

tere

d

2007 2008 2009 2010 2011


SDH

010

000

2000

030

000

4000

050

000

Dos

es a

dmin

iste

red

2007 2008 2009 2010 2011


PHC

3332


0 20 40 60 80 100Percent (%)

SHC

PHC

CHC

SDH

DH

< 30 min. > 30 min.

0 20 40 60 80 100Percent (%)

SHC

PHC

CHC

SDH

DH

<6 6-78-9 10


suggest that these platforms are entirely unable to pro-vide adequate ANC services; it simply means that the vast majority of facilities did not have the recommended diag-nostics and medical equipment for ANC.

Deliverycareservices99% of deliveries in TN occur in a health facility, and

67% in a public facility.8 Availability of essential equip-ment is necessary for providing high-quality delivery care; these results are presented in Table 10. Availability was generally highest in district hospitals, declining at lower levels with notable gaps among community and primary health centres. While all facility levels offered routine delivery services, no community or primary health cen-tres had all essential tests and equipment available, and only 4% of sub-district hospitals were fully equipped. This number increased to 54% among district hospitals,

8 International Institute for Population Sciences (IIPS). National Family Health Survey (NFHS-4), 2015-2016: Tamil Nadu Factsheet. Mumbai, India: IIPS, 2016.

Figure13Patient wait times at facilities, by platform

Figure14 Patient scores of facilities, by platform

though this is still a notable gap. The availability of in-cubators and vacuum extractors was notably lacking at sub-district hospitals, community health centres, and pri-mary health centres, despite these being essential items for service provision. Cross-match blood tests were also not widely available outside of district hospitals.

This finding is cause for concern, as not having access to adequate delivery equipment can affect both maternal and neonatal outcomes at all levels of care.9,10 Again, we found a substantial gap between the proportion of facil-ities, across platforms, that reported providing routine delivery services and those that were fully equipped for their provision.

GeneralsurgeryservicesAvailability of essential tests and equipment for gen-

eral surgery services are presented in Table 11. There was a lack of cross-match blood tests and blood storage units across all platforms. Essential medical equipment was

9 Nyamtema AS, Urassa DP, van Roosmalen J. Maternal health interventions in resource limited countries: a systematic review of packages, impacts and factors for change. BMC Pregnancy and Childbirth. 2011; 11(30).10 Wall SN, Lee ACC, Carlo W, Goldenberg R, Niermeyer S, Darmstadt GL, et al. Re-ducing intrapartum-related neonatal deaths in low- and middle-income countries

— what works? Seminars in Perinatology. 2010; 34: 395–407.

mostly available across platforms, though availability de-clined at lower-level facilities, particularly with regard to sterilization equipment and adult bag valve masks, which were available in less than two-thirds of sub health cen-tres. Availability of surgical equipment was also relatively high at hospitals, with the exception of blood storage units at sub-district hospitals. There were large gaps in surgical equipment in community health centres and primary health centres, indicating a lack of capacity to provide surgical services. It is also crucial to consider the human resources available to perform surgical proce-dures, as assembling an adequate surgical team is likely to affect patient outcomes. Given the nature of documen-tation of human resources in the records, such data could not be captured, but future work on assessing surgical ca-pacity at health facilities should collect this information.

LaboratorytestingThe availability of laboratory tests is presented in Ta-

ble 12. While all district hospitals and sub-district hospitals offer the range of laboratory services, there were gaps in test availability. Availability was generally high in dis-trict hospitals and decreased at lower facility levels, with particularly large gaps among primary health centres. However, some tests had low availability at all levels. Se-

Table13 Characteristics of patients interviewed after receiving care at facilities

DH SDH CHC PHC SHC TOTAL

Totalpatientsample 422 569 413 603 270 2277

Percentfemale 54% 55% 63% 61% 82% 61%

Patient’sagegroup(years)

<16 12% 11% 10% 8% 4% 9%

16–29 32% 26% 36% 33% 41% 32%

30–39 21% 22% 18% 21% 26% 21%

40–49 17% 18% 15% 16% 13% 16%

>50 19% 23% 21% 23% 17% 21%

Scheduledcaste/scheduledtribe 27% 29% 27% 34% 25% 29%

Otherbackwardscaste 60% 56% 60% 58% 65% 59%

Educationattainment

None 18% 22% 18% 24% 18% 20%

Classes 1 to 5 21% 25% 26% 24% 19% 23%

Classes 6 to 9 29% 27% 27% 21% 27% 26%

Class 10 or higher 33% 26% 29% 31% 36% 30%

DH: District hospital; SDH: Sub-district hospital; CHC: Community health centre; PHC: Primary health centre

Note: Educational attainment refers to the patient’s level of education or the attendant’s educational attainment if the interviewed patient was younger than 18 years old.

Figure12 Patient travel times to facilities, by platform

0 20 40 60 80 100Percent (%)

SHC

PHC

CHC

SDH

DH

< 30 min. > 30 min.

DH: District hospital; SDH: Sub-district hospital; CHC: Community health centre; PHC: Primary health centre; SHC: Sub health centre DH: District hospital; SDH: Sub-district hospital; CHC: Community health centre;

PHC: Primary health centre; SHC: Sub health centreDH: District hospital; SDH: Sub-district hospital; CHC: Community health centre; PHC: Primary health centre; SHC: Sub health centre

Note: Facility ratings were reported along a scale of 0 to 10, with 0 as the worst facility possible and 10 as the best facility possible.

3534


rum electrolyte tests, useful as part of a metabolic panel and to measure symptoms of heart disease and high blood pressure, had low availability in district hospitals (62%), sub-district hospitals (8%), and community health centres (8%). Spinal fluid tests were also rare among fa-cilities, present at only 46% of district hospitals and 4% of sub-district hospitals. Liver and renal function tests were widely available at district hospitals, but lacking from other facility levels. There were striking gaps in the capac-ity to test for infectious diseases at primary health centres, as only 20% reported the availability of HIV or tuberculo-sis tests, and just 56% had tests for malaria.

FacilityoutputsMeasuring a facility’s patient volume and the number

of services delivered, which are known as outputs, is crit-ical to understanding how facility resources align with patient demand for care. Figure 8 illustrates the trends in average outpatient volume across platforms and over

time. The number of outpatient visits by fiscal year, by platform, is presented in Figure 8. In general, the average number of outpatient visits increased slightly over five fis-cal years. Patient volume was highest in district (average of 619,435–658,125 visits per year). Sub-district hospitals reported an average of 200,278–221,487 visits per year, which was nearly triple the number reported by commu-nity health centres (average of 58,034–64,999 visits per year). Primary health centres reported more than 40 times more outpatient visits (average of 37,091–45,806 visits per year) than sub-health centres (average of 752–969 visits per year).

Inpatient visits generally entail more service demands than outpatient visits, including ongoing occupancy of facility resources such as beds .The reported number of

Table14Proportion of patients satisfied with facility visit indicators, by platform

DISTRICT HOSPITAL




SUB HEALTH CENTRE

StaffinteractionsNurse/ANM

Medical provider respectfulness 57% 63% 73% 72% 77%

Clarity of provider explanations 57% 55% 71% 70% 79%

Time to ask questions 48% 56% 69% 66% 73%

Doctor

Medical provider respectfulness 61% 64% 72% 74% 92%

Clarity of provider explanations 58% 64% 70% 72% 58%

Time to ask questions 51% 60% 68% 70% 67%

Facilitycharacteristics

Cleanliness 36% 37% 50% 56% 56%

Privacy 50% 42% 58% 54% 58%


Figure15Availability of prescribed drugs at facility, by platform

0 20 40 60 80 100Percent (%)

SHC

PHC

CHC

SDH

DH

Got none/some of the drugs Got all perscribed drugs

DH: District hospital; SDH: Sub-district hospital; CHC: Community health centre; PHC: Primary health centre; SHC: Sub health centre

FemaleMale

>=40 years16-39 years

Other castesBackwards caste

Any schoolingNo schooling

Not given all prescribed drugsGiven all prescribed drugs

Wait time <30 minWait time >=30 min

DHPHCCHCSDH

0 1 2 3 4Odds Ratio

Figure16Determinants of satisfaction with doctors

Dotted vertical line represents an odds ratio of one. Black points represent the reference groups, which all carry an odds ratio of one. Compared to the referent category, significant odds ratios and 95% confidence intervals are represented with blue points and horizontal lines, respectively. Odds ratios that are not significant are represented by green points, and their 95% confidence intervals with a green horizontal line. Any confidence intervals with an upper bound above 4 were truncated for ease of interpretation.


inpatient visits (other than deliveries) by year are pre-sented in Figure 9. Over time, the average number of inpatient visits have increased for all platforms. District hospitals provided care for an average of 50,332–56,729 inpatient visits per fiscal year. Sub-district hospitals pro-vided care for an average of 7,938–13,711 visits per year, while community health centres provided far fewer vis-its (an average between 717 and 1,018 inpatient visits per year). Primary health centres reported substantially fewer inpatient visits (on average 333–606 visits per year). It is important to note that the ABCE Facility Survey did not capture information on the length of inpatient stays, which is a key indicator to monitor and include in future work.

The reported number of deliveries, by platform and

3736


over time, is presented in Figure 10. District hospitals re-ported an average between 3,291 and 3,940 deliveries in each year of observation, while sub-district hospitals re-ported an average of 505–586 deliveries per year. While many hospitals experienced an increase in the number of deliveries over time, several hospitals reported de-creasing numbers over the five years of observation. Community health centres reported an annual average number of deliveries between 174 and 269. Few deliver-ies were reported in primary health centres (an average of 82–167 deliveries per year). The ratio of deliveries to inpa-tient visits is higher among the lower platforms.

ImmunizationThe number of immunization doses administered over

time, by platform, is presented in Figure 11. The average number of doses administered remained stable over the five fiscal years, with slight declines in sub-district hospi-

tals and community health centres. The highest volume of immunization doses administered was seen in sub-district hospitals, with an average between 25,783 and 30,947 doses per year. District hospitals reported an average be-tween 17,285 and 17,771 doses administered in each year of observation. Community health centres reported pro-viding an average number of doses between 4,808 and 5,815 per year, similar to primary health centres, which reported an average of 4,979–6,460 doses administered per year. Sub health centres reported an average of 752–880 doses per year.

PatientperspectivesA facility’s availability of and capacity to deliver ser-

vices is only half of the health care provision equation; the other half depends upon patients seeking those health services. Many factors can affect patients’ decisions to seek care, ranging from associated visit costs to how pa-

FemaleMale

>=40 years16-39 years

Other castesBackwards caste

Any schoolingNo schooling

Not given all prescribed drugsGiven all prescribed drugs

Wait time <30 minWait time >=30 min

DHSHCPHCCHCSDH

0 1 2 3 4Odds Ratio

Figure17Determinants of satisfaction with nurses

Dotted vertical line represents an odds ratio of one. Black points represent the reference groups, which all carry an odds ratio of one. Compared to the referent category, significant odds ratios and 95% confidence intervals are represented with blue points and horizontal lines, respectively. Odds ratios that are not significant are represented by green points, and their 95% confidence intervals with a green horizontal line. Any confidence intervals with an upper bound above 4 were truncated for ease of interpretation.

DH: District hospital; SDH: Sub-district hospital; CHC: Community health centre; PHC: Primary health centre; SHC: Sub health centre

Table15Input-output model specifications

CATEGORY VARIABLES

Model 1

Inputs Expenditure on personnelExpenditure on pharmaceuticalsAll other expenditure

Outputs Outpatient visitsInpatients visits (excluding deliveries)DeliveriesImmunization visits

Model 2

Inputs Number of bedsNumber of doctorsNumber of nursesNumber of ANMsNumber of paramedical staffNumber of non-medical staff

Outputs Outpatient visitsInpatients visits (excluding deliveries)DeliveriesImmunization visits

tients view the care they receive. These “demand-side” constraints can be more quantifiable (e.g., distance from facility) or intangible (e.g., perceived respectfulness of the health care provider), but each can have the same im-pact on whether patients seek care at particular facilities or have contact with the health system at all.

Using data collected from the Patient Exit Interview Surveys, we examined the characteristics of patients who presented at health facilities and their perspectives on the care they received. Table 13 provides an overview of the interviewed patients (n=2,277) or their attendants at public facilities. Most patients were female (61%), and the majority of patients identified as part of a scheduled caste/scheduled tribe (29%) or other backwards caste (59%). 80% of patients had some education, and all facil-ities saw patients with a range of educational attainment. 41% of patients were under the age of 30.

TravelandwaittimesThe amount of time patients spend traveling to facili-

ties and then waiting for services can substantially affect their care-seeking behaviors. Among the patients who were interviewed, we found that travel time to a facility for care (Figure 12) differed by the platform, with shorter

travel time for patients seeking care at lower-level facili-ties than higher-level. It is important to note that patients only reported on the time spent traveling to facilities, not the time needed for round-trip visits.

Fifty-four percent of patients who went to district hospitals traveled fewer than 30 minutes, 32% traveled between 30 minutes and one hour. At primary health cen-tres these proportions were 78% and 17%, respectively, while at sub health centres nearly all patients traveled for less than 30 minutes. This finding is not unexpected, as these are the closest health facilities for many patients, particularly those in rural areas. It also reflects the fact that many patients travel longer distances to receive the kind of specialized care offered at hospitals.

Wait time is also an important determinant of patient satisfaction. The large majority of patients waited less than 30 minutes to receive care at all platforms (Figure 13), and nearly all patients seeking care at sub health centres (94%) received care within 30 minutes. Wait times were longer at district hospitals (34% of patients waited more than 30 minutes to receive care) and sub-district hospitals (34%). Fewer than 6% of all patients waited more than one hour to receive care.

PatientsatisfactionwithcareWe report primarily on factors associated with patient

satisfaction with provider care and perceived quality of services by patients on medicine availability and hospital infrastructure, as these have been previously identified to be of significance in the patient’s perception of quality of health services in India.11

Ratings of patient satisfaction, based on a scale from zero to 10, with 10 being the highest score, are presented in Figure 14. Overall, patients were satisfied with the care they received and, in general, ratings were higher for higher-level platforms. Few patients (6%) gave a rating of 10, and the majority rated the facility they attended an 8 or 9 (42% of all patients). Among patients seeking care at community health centres, only 9% rated the facil-ity below a 6; among patients seeking care at sub-district hospitals, this proportion is 14%.

Patients were also asked more detailed questions about satisfaction with providers and facility characteris-

11 Rao KD, Peters DH, Bandeen-Roche K. Towards patient-centered health services in India—a scale to measure patient perceptions of quality. International Journal for Quality in Health Care. 2006; 18(6):414-421.

3938


Table17 Average annual cost in INR, by platform, last fiscal year. INR denotes Indian Rupees.

DISTRICT DISTRICT HOSPITAL




District 1 15,383,021 8,511,807 2,668,189

District 2 119,704,784 14,915,623 6,157,839 2,932,572

District 3 16,263,143 5,854,098 9,817,366 4,003,384

District 4 93,398,448 27,209,300 7,036,143 4,214,529

District 5 67,691,752 7,796,161 11,579,260 5,243,899

District 6 33,335,870 14,403,316 4,904,956

District 7 15,156,273 7,368,294 3,144,209

District 8 7,046,398 35,766,028 4,886,277

District 9 127,700,392 26,264,536 7,724,846 4,519,042

District 10 95,598,248 12,327,551 12,638,121 4,589,543

District 11 66,003,392 27,048,824 13,004,020 5,014,610

District 12 17,334,950 7,197,371 13,139,467 4,358,396

District 13 41,088,188 41,899,892 15,192,691 2,964,931

Empty cells were either dropped from analysis due to data availability, or there were no facilities to sample of that platform.

Table16Average and range of inputs and outputs, by platform. INR denotes Indian Rupees.

DISTRICT HOSPITAL




Inputs

Personnel expenditure (INR)56,556,802 (6,492,954-119,503,416)

16,471,282 (1,972,154-64,633,828)

9,403,170 (2,990,895-52,689,648)

3,735,343 (237,240-8,011,006)

Pharmaceutical expenditure (INR)7,343,161 (1,771,439-12,103,611)

1,405,913 (320,046-4,244,187)

236,908 (85,069-596,232)

159,776 (79,998-429,856)

Other expenditure (INR)7,742,626 (134,917-27,240,558)

1,149,703 (112,272-9,471,909)

2,847,552 (39,093-30,170,106)

2433,04 (38,489-1,649,957)

Number of beds343 (84-608) 79 (16-258) 21 (3-37) 4 (1-14)

Number of doctors39 (20-78) 11 (2-41) 5 (1-13) 2 (1-6)

Number of nurses65 (27-163) 15 (2-44) 4 (1-7) 3 (0-5)

Number of paramedical staff64 (16-127) 16 (2-39) 17 (6-30) 11 (3-20)

Number of non-medical staff58 (9-144) 8 (0-27) 5 (1-9) 2 (0-9)

Outputs

Outpatient visits663,351 (225,699-1,144,003)

215,165 (4,103-829,210)

63,947 (17,821-170,765)

40,508 (13,500-77,361)

Inpatient visits (excluding deliveries)

52,311 (7,662-146,668)

10,971 (265-116,702) 908 (34-2,682) 480 (69-2,205)

Deliveries 3,999 (487-14,220) 561 (15-3,874) 237 (33-926) 132 (13-431)

Immunization doses10,846 (0-40,197) 25,856 (0-329,688) 5,356 (0-19,058) 5,874 (0-47,380)

tics (Table 14). Most patients were unsatisfied with facility cleanliness at district hospitals (64%) and sub-district hos-pitals (63%), and dissatisfaction with privacy in these facility types was also high (50% and 58%, respectively). Health centres performed slightly better than hospitals on both these metrics.

Three parameters were assessed to document satis-faction with health providers – being treated respectfully by the provider, clarity of explanation provided by the provider, and that provider gave enough time to ask questions about health problem or treatment – using a 5 point Likert scale, with the highest ratings of good and very good responses combined as satisfied, and rest as not satisfied. Using the three parameters of satisfaction,

a composite satisfaction variable was created separately for doctors and nurses – if a patient reported good/very good for all three parameters, it was categorized as sat-isfied. At district hospitals, sub-district hospitals, and primary health centres, patients receiving care from doc-tors reported slightly higher levels of satisfaction than those receiving care from nurses and ANMs. This trend was reversed in sub health centres, where patients were more satisfied with nurses and auxiliary nurse midwives (ANMs), with the exception of very high satisfaction scores for doctor respectfulness in sub health centres. Generally, satisfaction was higher at health centres than at hospitals.

050

100

150

Exp

endi

ture

in 1

00,0

00 R

upee

s

2007 2008 2009 2010 2011

Personnel

Pharmaceuticals and consumables Other

020

040

060

080

0E

xpen

ditu

re in

100

,000

Rup

ees

2007 2008 2009 2010 2011

Personnel


010

2030

4050

Exp

endi

ture

in 1

00,0

00 R

upee

s

2007 2008 2009 2010 2011

Personnel


050

100

150

200

Exp

endi

ture

in 1

00,0

00 R

upee

s

2007 2008 2009 2010 2011

Personnel


Figure18 Average total and type of expenditure, by platform, 2007-2011

COMMUNITY HEALTH CENTRES PRIMARY HEALTH CENTRES

SUB-DISTRICT HOSPITALSDISTRICT HOSPITALS

4140


Figure19 Average percentage of expenditure type, by platform, 2011

Access to affordable drugs has been interpreted to be part of the right to health. Among 1,620 patients who were prescribed drugs and attempted to obtain those drugs during the visit, 1,552 received all prescribed drugs (Figure 15). This ranged from 99% of patients at sub-dis-trict hospitals and primary health centres to 80% of patients at sub health centres.

Many complex factors affect patient satisfaction with the medical care they receive. Given this, a multivariate logistic regression was conducted in order to determine which patient and facility characteristics were associ-ated with patient satisfaction with both medical doctors (Figure 16) and nurses/ANMs (Figure 17). For each char-acteristic –for example, the age or sex of the patient – the odds ratio (OR) is presented. The OR represents the odds that a patient is satisfied given a particular characteristic, compared to the odds of the patient being satisfied in the absence of that characteristic. An OR and 95% con-fidence interval (CI) greater than 1.0 indicates that there are greater odds of being satisfied with care as compared to the reference group. An OR and 95% CI below 1.0 in-dicates that there are lower odds of being satisfied with care than the reference group.

For example, while the OR for patients under age 40 years being satisfied with care from a doctor is 0.88 (95% CI: 0.66–1.16) as compared to patients age 40 years and older, it is not statistically different from an OR of 1.0

(Figure 16). This means that, considering all other char-acteristics, patients under age 40 are not more or less satisfied with care from doctors than patients 40 years and older. In Figures 16 and 17, ORs that are statistically significant are signified by blue points, with blue horizon-tal bars representing their confidence interval. ORs that are not statistically significant are represented with green points and green confidence bars.

Compared to patients of another group, there was slightly lower satisfaction with doctors for male patients (OR: 0.75, 95% confidence interval [CI]: 0.58–0.98). Con-trolling for all other factors, compared to patients who sought care at district hospitals, patients who sought care at primary health centres were more satisfied with care from doctors (OR: 2.39, 95% CI: 1.10–5.17).

Receipt of all prescribed drugs was associated with higher satisfaction with nurses, as compared to patients who received some or none of the prescribed drugs (Fig-ure 17, OR: 2.43, 95% CI: 1.03–5.70). Compared to patients who sought care at district hospitals, those who sought care at sub health centres (OR: 3.74, 95% CI: 1.29–10.83) and community health centres (OR: 3.35, 95% CI: 1.22–9.22) had higher satisfaction with nurses.

EfficiencyandCostsThe costs of health service provision and the efficiency

with which care is delivered by health facilities go hand-in-hand. An efficient health facility uses resources well, producing a high volume of patient visits and services without straining its resources. Conversely, an ineffi-cient health facility is one where the use of resources is not maximized, leaving usable beds empty or medi-cal staff seeing very few patients per day. We present technical efficiency analysis for district hospitals, sub- district hospitals, community health centres and primary health centres.

AnalyticalapproachAn ensemble model approach was used to quantify

technical efficiency in health facilities, combining results from two approaches – the restricted versions of Data Envelopment Analysis (rDEA) and Stochastic Distance Function (rSDF).12 Based on this analysis, an efficiency

12 Di Giorgio L, Flaxman AD, Moses MW, Fullman N, Hanlon M, Conner RO, et al. Efficiency of Health Care Production in Low-Resource Settings: A Monte-Carlo Simulation to Compare the Performance of Data Envelopment Analysis, Stochastic

0 20 40 60 80 100Percent of Total Expenditure




District Hospital

Personnel

Pharmaceuticals and consumables

Other

Figure20 Outpatient load per staff, by platform

020

0040

0060

00V

isits

2007 2008 2009 2010 2011

OP visits per staff by facilityOP visits per staff average

CHC

2000

3000

4000

5000

6000

Vis

its

2007 2008 2009 2010 2011


DH

1000

2000

3000

4000

5000

6000

Vis

its

2007 2008 2009 2010 2011


PHC

020

0040

0060

0080

0010

000

Vis

its

2007 2008 2009 2010 2011


SDH


Note: each line represents an individual facility, with the bolded line depicting the average for the platform. Scales are different for each platform type.

4342


score was estimated for each facility, capturing a facility’s use of its resources. Relating the outputs to inputs, the rDEA and rSDF approaches compute efficiency scores ranging from 0% to 100%, with a score of 100% indicat-ing that a facility achieved the highest level of production relative to all facilities in that platform.

This approach assesses the relationship between in-puts and outputs to estimate an efficiency score for each facility. Recognizing that each type of input requires a

Distance Functions, and an Ensemble Model. PLOS ONE. 2016; 11(2): e0150570.

different amount of facility resources (e.g., on average, an inpatient visit uses more resources and more com-plex types of equipment and services than an outpatient visit), we applied weight restrictions to rescale each fa-cility’s mixture of inputs and outputs. The incorporation of additional weight restrictions is widely used in order to improve the discrimination of the models. Weight re-strictions are most commonly based upon the judgment about the importance of individual inputs and outputs, or reflect cost or price considerations. The resulting ensem-ble efficiency scores were averaged over five years and between the two input models.

010

0020

0030

0040

0050

00V

isits

2007 2008 2009 2010 2011

IP visits per staff by facilityIP visits per staff average

SDH

050

100

150

Vis

its

2007 2008 2009 2010 2011


PHC

Figure21Inpatient load per staff by platform0

200

400

600

800

Vis

its

2007 2008 2009 2010 2011


DH

020

4060

80V

isits

2007 2008 2009 2010 2011


CHC


Note: each line represents an individual facility, with the bolded line depicting the average for the platform. Scales are different for each platform type.

Figure22Deliveries per staff by platform


Note: Each line represents an individual facility, with the bolded line depicting the average for the platform. Scales are different for each platform type.

010

2030

40D

eliv

erie

s

2007 2008 2009 2010 2011

Deliveries per staff by facilityDeliveries per staff average

SDH

010

2030

40D

eliv

erie

s

2007 2008 2009 2010 2011


PHC

510

1520

2530

Del

iver

ies

2007 2008 2009 2010 2011


DH

010

2030

Del

iver

ies

2007 2008 2009 2010 2011


CHC

For these models, service provision was categorized into outpatient visits, inpatient visits, delivery, and immu-nization. Two input-output specifications were used, with the inputs being different in the two models. The inputs and outputs are listed in Table 15. The detailed data uti-lized for this analysis are documented in the annex. The average and range of inputs and outputs for the variables are presented in Table 16.

CostsofcareTotal expenditure, by district and platform, is pre-

sented in Table 17. In terms of annual total expenditures,

trends in average facility spending varied by platform be-tween 2007 and 2011 (Figure 18). All platforms recorded slightly higher levels of average expenditures in 2011 than in 2007, which appeared to be largely driven by in-creased spending on medical supplies and personnel. Figure 19 shows the average composition of expenditure types across platforms for 2011. Notably, sub-district hos-pitals and PHCs spent a slightly greater proportion of their total expenditures on personnel than other platforms. On the other hand, expenditures on medical supplies were


4544


2040

6080

100



highest the at district hospitals, with other expenditure highest at community health centres.

It is important to note that data availability on the input and output indicators varied across the facilities and plat-forms, with more non-availability for PHCs. Facilities with five years of missing data for any input or output variable were dropped from analysis. In addition, the data were smoothed where necessary based on the trends seen in inputs or outputs for that facility.

To further illustrate the production of outputs per in-puts – in this case, staff – a simple ratio of outpatient visits

Figure24 Range of efficiency scores across platforms

DistrictHospitalMean: 74.4Median: 75.5IQR: 70.0-82.0

PrimaryHealthCentreMean: 64.2Median: 65.2IQR: 59.7-71.7

Sub-districtHospitalMean: 55.7Median: 54.9IQR: 44.6-64.3

CommunityHealthCentreMean: 63.1Median: 61.7IQR: 56.4-70.8

Figure23Immunizations per staff per day by platform0

2040

6080

100

Dos

es a

dmin

iste

red

2007 2008 2009 2010 2011

Immunization doses per staff by facilityImmunization doses per staff average

DH

020

040

060

0D

oses

adm

inis

tere

d

2007 2008 2009 2010 2011


CHC


Note: Each line represents an individual facility, with the bolded line depicting the average for the platform. Scales are different for each platform type.

(Figure 20), inpatient visits (Figure 21), deliveries (Figure 22), and immunization doses (Figure 23) per staff are pre-sented. District hospitals produced an average of 3,183 outpatient visits per staff, though the ratio ranged greatly. The average ratio for sub-district hospitals was 4,476 vis-its per staff, for community health centres 2,198, and for primary health centres 2,488. This gradient differed for inpatient visits, with district hospitals providing 264 in-patient visits per staff, sub-district hospitals providing 246, community health centres providing 30, and primary health centres providing 30. The range of inpatient visits

010

0020

0030

0040

0050

00D

oses

adm

inis

tere

d

2007 2008 2009 2010 2011


SDH

010

0020

0030

00D

oses

adm

inis

tere

d

2007 2008 2009 2010 2011


PHC

Note: Each circle represents the five-year facility average efficiency score; IQR refers to intra-quartile range.

per staff was low for primary health centres, where inpa-tient visits are rare. Overall, as expected, outpatient visits accounted for the overwhelmingly large majority of the patients seen per staff per day across the platforms.

Fewer deliveries were performed per staff than other services, with an average of 17 deliveries per staff in dis-trict hospitals, eight per staff in sub-district hospitals, eight per staff in community health centres, and eight per staff in primary health centres. For immunization, 50 doses were administered per staff in district hospitals, 572 per staff in sub-district hospitals, 191 per staff in com-

munity health centres, and 36 per staff in primary health centres. There was quite a bit of variation of these ratios within a platform and over time, however.

EfficiencyresultsUsing the five fiscal years of data to estimate the effi-

ciency scores for all facilities, two main findings emerged. First, efficiency scores were relatively low across all health facilities, with 74.4% being the highest mean across plat-forms. Second, the range between the facilities with highest and lowest efficiency scores was quite large


4746


within platforms, suggesting that a substantial perfor-mance gap may exist between the average facility and facilities with the highest efficiency scores. Figure 24 depicts this range of facility efficiency scores across plat-forms for TN.

Efficiency by district is presented in Table 18. There is variation in facility efficiency both between and within dis-tricts. Some of the least efficient primary health centres were in the same district as the least efficient sub-district hospitals (for example, District 7). District 14, for example, had the most efficient primary health centre but the least efficient sub-district hospital. While one primary health centre in District 10 was 69% efficient, another was only 37% efficient.

Given observed levels of facility-based resources (beds and personnel), it would appear that many facilities had the capacity to handle much larger patient volumes than they reported. Figure 25 displays this gap in poten-tial efficiency performance across platforms, depicting the possible gains in total service provision that could be achieved if every facility in the ABCE sample operated at optimal efficiency.

We found that all types of facilities could expand their

Figure25Observed and estimated additional visits that could be produced given observed facility resources

Table18District-wise efficiency scores (%), by platform

DISTRICT/PLATFORM

DISTRICT HOSPITAL

SUB DISTRICT HOSPITAL

COMMUNITY HEALTH CENTRE PRIMARY HEALTH CENTRE

1 1 2 1 2 1 2 3 4

District 2 42.8 44.6 70.8 89.6 63.6 71.9 71.7 69.8

District 3 57.4 54.6 64.3 62.0 70.9 77.9 60.4 55.8 80.6

District 4 70.8 63.4 54.9 69.8 61.1 75.8 71.4 59.7 70.8

District 5 65.8 35.8 81.1 60.2 61.4 63.2 79.0 71.8 55.9

District 6 75.5 81.8 54.5 71.0 53.1 62.6 51.8 63.9 53.6

District 7 27.8 45.2 35.6 48.7 48.5 43.0 43.4 43.1

District 8 78.5 77.6 68.3 74.1 73.1 62.6 61.3

District 9 37.1 36.4 50.1 56.4 59.9 62.5 78.8

District 10 84.9 58.8 50.5 61.0 69.5 69.0 37.1 66.3 65.8

District 11 85.0 60.5 63.9 77.0 53.7 67.0 60.2 65.2 67.4

District 12 70.0 72.7 63.9 72.2 65.5 72.8 53.8 66.9 75.5

District 13 78.7 48.5 58.7 54.5 68.9 60.3 68.5 72.1

District 14 82.0 76.9 15.5 66.0 59.6 57.3 86.4 59.3

White cells were either dropped from analysis due to data availability, or there were no more facilities to sample of that platform.

0 1,000 2,000 3,000 4,000 5,000Deliveries




District Hospital

Deliveries

Observed Estimate additional deliveries

0 200000 400000 600000 800000 1.0e+06Outpatient visits




District Hospital

Outpatient visits

Observed Estimate additional visits

0 20,000 40,000 60,000 80,000Inpatient visits




District Hospital

Inpatient visits

Observed Estimate additional visits

DELIVERIES IMMUNIZATION DOSES

INPATIENT VISITSOUTPATIENT VISITS

outputs substantially given their observed resources. Based on our analyses, the highest level of care, district hospitals, had the greatest potential for increasing service provision without expanding current resources. Overall, based on our estimation of efficiency, a large portion of TN health facilities could increase the volume of patients seen and services provided with the resources available to them.

On average, district hospitals could provide 249,706 additional outpatient visits with the same inputs, while primary health centres could see an average of 21,906 additional outpatient visits. Sub-district hospitals could administer an average of 26,727 additional immunization doses with the same inputs if all facilities were efficient.

At the same time, many reports and policy documents emphasize that pronounced deficiencies in human re-sources for health exist across India in the public sector health system, such that “significant [human resources for health] will be required to meet the demand” for health services.13 Our results suggest otherwise, as most facilities

13 Rao M, Rao KD, Kumar AK, Chatterjee M, Sundararaman T. Human resources for health in India. The Lancet. 2011; 377(9765): 587-98.

in the ABCE sample had the potential to bolster service production given their reported staffing of skilled person-nel and physical capital.

These findings provide a data-driven understanding of facility capacity and how health facilities have used their resources in TN; at the same time, they are not without limitations. Efficiency scores quantify the relationship be-tween what a facility has and what it produces, but these measures do not fully explain where inefficiencies orig-inate, why a given facility scores higher than another, or

0 10,000 20,000 30,000 40,000 50,000Immunization doses




District Hospital

Immunization Doses

Observed Estimate additional doses

what levels of efficiency are truly ideal. It is conceivable that always operating at full capacity could actually have negative effects on service provision, such as longer wait times, high rates of staff burnout and turnover, and com-promised quality of care. These factors, as well as less tangible characteristics such as facility management, are all important drivers of health service provision, and fu-ture work should also assess these factors alongside measures of efficiency.


4948

Conclusions and policy implications

To achieve its mission to “expand the reach of health care and establishing universal health coverage,”1 India has strived over the past 10 years to expand and strengthen the public

sector of health care, with a focus on reaching rural areas. The country recognizes disparities and has sought to en-act policies and implement programs to expand access to essential and special services for marginalized groups. Our findings show that these goals are ambitious but at-tainable, if the country focuses on rigorously measuring health facility performance and costs of services across and within levels of care, and if it can align the different dimensions of health service provision to support optimal health system performance.

FacilitycapacityforserviceprovisionOptimal health service delivery, one of the key build-

ing blocks of the health system,2 is linked to facility capacity to provide individuals with the services they need and want. With the appropriate balance of skilled staff and supplies needed to offer both essential and spe-cial health services, a health system has the necessary foundation to deliver quality, equitable health services.

The availability of a subset of services, including rou-tine delivery, antenatal care, general medicine, pharmacy, and laboratory services, was generally high across facil-ity types in Tamil Nadu, reflecting the expansion of these services throughout the state. However, clear differences remain between facility types. Sub-district hospitals nota-bly lack certain essential services: for example, only 69% provide STI/HIV services and 77% provide immunizations, meaning that availability is lower in these facilities than in community health centres. Moreover, substantial gaps were identified between facilities reporting availability of these services and having the full capacity to actually

1 Planning Commission Government of India. Twelfth Five Year Plan (2012-17). New Delhi, India: Government of India, 2012. 2 World Health Organization (WHO). Everybody’s Business: Strengthening health systems to improve health outcomes: WHO’s Framework for Action. Geneva, Swit-zerland: WHO, 2007.

deliver them. While almost all facilities, across platforms, indicated that they provided routine delivery care, only 54% of district hospitals, 4% of sub-district hospitals, and no lower-level facilities had the full stock of medical sup-plies and equipment to optimally provide these services. These gaps were also evident for ANC in all facility types. While 85% of district hospitals were fully equipped to provide general surgery, only 21% of sub-district hospi-tals and no lower-level facilities were fully equipped. In general, district hospitals were well equipped with medi-cal, laboratory, and imaging equipment, with the notable exception of CT scans. The availability of equipment de-clined through the levels of the system, particularly with regard to laboratory equipment and imaging equipment. Closing these gaps and making sure that all facilities are fully equipped to optimally provide essential services warrants further policy consideration.

Chronic diseases (e.g., cardiovascular diseases, men-tal health disorders, diabetes, and cancer) and injuries are the leading causes of death and disability in India, and are projected to increase in their contribution to the burden of disease during the next 25 years.3,4,5 Much of the care for chronic diseases and injuries is provided in the private sector and can be very expensive.45 Many NCD-related services, including cardiology, psychiatry, and chemotherapy, are notably lacking at various levels of care. While 77% of district hospitals provide cardiology and psychiatry services, only 8% provide chemotherapy. The availability of all these services declines markedly at lower facility levels, including sub-district hospitals as well as community health centres. Such gaps in the

3 GBD 2015 Mortality and Causes of Death Collaborators. Global, regional, and national life expectancy, all-cause mortality, and cause-specific mortality for 249 causes of death, 1980–2015: a systematic analysis for the Global Burden of Disease Study 2015. The Lancet. 2016; 388:1459–1544. 4 Patel V, Chatterji S, Chisholm D, Ebrahim S, Gopalakrishna, G, Mathers C et al. Chronic diseases and injuries in India. The Lancet. 2011; 377: 413-28. 5 GBD 2015 DALYs and HALE Collaborators. Global, regional, and national disability-adjusted life-years (DALYs) for 315 diseases and injuries and healthy life expectancy (HALE), 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. The Lancet. 2016 Oct 7; 388:1603–1658

C O N C L U S I O N S A N D P O L I C Y I M P L I C AT I O N S

health system will exacerbate disparities by not dealing appropriately with NCDs while continuing to endeavor to eliminate major infectious diseases like tuberculosis, HIV, and malaria, or to reduce neonatal and infant mortality. Furthermore, there also is a paucity of essential equip-ment for NCD services at lower facility levels, including glucometer/test strips, though district hospitals are gen-erally well-equipped. These findings support the need for immediate action to scale up interventions for chronic diseases through improved public health and primary health care systems that are essential for the implementa-tion of cost-effective interventions.

Recent studies show that India has a severe shortage of human resources for health.6 It has a shortage of qual-ified health workers and the workforce is concentrated in urban areas. In the context of a shortage of qualified health personnel at all levels of the health system, but especially rural areas,7,8,9 results reveal disparate staff-ing patterns between facilities. Hospitals employ a large number of staff. At the lower, community levels, para-medical staff including nurses and ANMs provide the majority of care to patients (based on reported staffing). These staffing patterns are not unexpected based on the hierarchy of care. However, nurses do not have much au-thority or say within the health system, and the resources to train them are still inadequate. A call has been made to the government to urgently address the issues of human resources through a comprehensive national policy for human resources to achieve universal health care in India.

However, it should be noted that despite the shortfall in human resources, the study findings suggest suboptimal efficiency in production of services with the given level of human resources.

Adequate operational infrastructure is essential for the functioning of a facility, which in turn affects the ef-ficiency of service provision. In Tamil Nadu, all hospitals and community and primary health centres and almost all sub health centres had access to functioning electric-ity, and only one facility reported being solely dependent on a generator. This means a higher quality of service

6 Rao M, Rao KD, Kumar AK, Chatterjee M, Sundararaman T. Human resources for health in India. The Lancet. 2011; 377(9765): 587-98.7 Government of India, “Twelfth Five Year Plan (2011-17).”8 Hazarika I. Health Workforce in India: Assessment of Availability, Production and Distribution. WHO South East Asia Journal of Public Health. 2013; 2(2): 106-112. 9 Rao M, Rao KD, Kumar AK, Chatterjee M, Sundararaman T. Human resources for health in India. The Lancet. 2011; 377(9765): 587-98.

provision, as it allows for reliable storage of medications, vaccines, and laboratory samples. Access to piped water was more variable in these types of facilities; it was partic-ularly lacking in sub health centres, with all other facilities showing availability of 85%–88%. Access to flushed toilets was actually lower in district hospitals than in commu-nity and primary health centres, though again, sub health centres recorded the lowest levels overall. That so many facilities did report access to essential resources like water, sanitation, and electricity likely reflects India’s com-mitment10,11 to upgrade all facilities so they meet Indian Public Health Standards. However, the marked discrep-ancies evident between sub centres and other types of facility suggest that there should be a sustained focus on making sure that these resources reach the lowest levels of the health system.

Communication is also an important facet of health service delivery. In general, facilities in Tamil Nadu had good access to phones and computers, which makes for more efficient referrals and coordination. However, ac-cess to four-wheeled vehicles was low at primary health centres. There is scope, then, to address these gaps in or-der to ensure that all patients receive timely emergency and curative care.

FacilityproductionofhealthservicesOverall, the number of outpatient visits by year and

platform was relatively stable over the five years of obser-vation. Volume of outpatient visits was considerably lower at the lower health facilities. The volume of inpatient visits and deliveries increased over the five years of observa-tion for most platforms. The highest volumes of visits were held by district hospitals, followed by sub-district hospi-tals. Facility expenditure is dominated by personnel costs – accounting for, on average, at least 70% of total costs.

Efficiency scores reflect the relationship between facility-based resources and the facility’s total patient volume each year. Average efficiency scores by platform ranged from 55.7% to 74.4%, indicating patient volume could substantially increase with the observed levels of resources and expenditure. Within each platform, there is great variation in the efficiency of health facilities be-tween and within districts. With this information, we

10 Planning Commission Government of India. Eleventh Five Year Plan (2007-12). New Delhi, India: Government of India, 2007. 11 Planning Commission Government of India. Twelfth Five Year Plan (2012-17). New Delhi, India: Government of India, 2012.

5150


estimated that facilities could substantially increase the number of patients seen and services provided, based on their observed levels of medical personnel and resources in 2011. As India seeks to strengthen public sector care to reduce the heavy burden of out-of-pocket expendi-tures,12,13 stakeholders may seek to increase efficiency by providing more services while maintaining personnel, ca-pacity (beds), and expenditure.

Further use of these results requires considering ef-ficiency in the context of several other factors, including quality of care provided, demand for care, and expedi-ency with which patients are seen.

The policy implications of these efficiency results are both numerous and diverse, and they should be viewed with a few caveats. A given facility’s efficiency score cap-tures the relationship between observed patient volume and facility-based resources, but it does not reflect the expediency with which patients are seen, the optimal provision of services, demand for the care received, and equity in provision of services to serve those who are dis-advantaged.14 These are all critical components of health service delivery, and they should be thoroughly consid-ered alongside measures of efficiency. On the other hand, quantifying facility-based levels of efficiency provides a data-driven, rather than strictly anecdotal, understanding of how much TN health facilities could potentially expand service provision without necessarily increasing person-nel or bed capacity in parallel.

CostsofcareAverage facility expenditure per year differed sub-

stantially across platforms. We were unable to estimate the costs of care by type of services (such as outpatients, inpatients, deliveries, immunization, etc.) or by type of disease/condition (such as TB, diabetes, etc.) as such data are not readily available at the facilities. Estimating such costs of care and identifying differences in patient costs across the type of platforms is critical for isolating areas to improve cost-effectiveness and expand less costly ser-vices, especially for hard-to-reach populations.

Nevertheless, these results on expenditures offer in-sights into each state’s health financing landscape, a

12 Ibid.13 Kumar AKS, Chen LC, Choudhury M, Ganju S, Mahajan V, Sinha A et al. Financing health care for all: challenges and opportunities. The Lancet. 2011; 377: 668-79. 14 UNICEF. Narrowing the gaps: The power of investing in the poorest children. New York, NY: UNICEF, 2017.

key component of health system performance in terms of cost to facilities and service production. While these costs do not reflect the quality of care received or the specific services provided for each visit, they can enable a compelling comparison of overall health care expenses across states within India. Future studies should aim to capture information on the quality of services provided, as it is a critical indicator of the likely impact of care on patient outcomes.

PatientperspectivesPatient satisfaction is an important indicator of pa-

tient perception of the quality of services provided by the healthcare sector.15,16 Evaluation of services by pa-tients is important for purposes of monitoring, increasing accountability, recognizing good performance, and adapting patient-centric services, and for utilization of services and compliance with treatment. A major strength of this study is that patient satisfaction was assessed across the various levels of public sector health care in the state.

The public health system in India designed as a refer-ral hierarchical system to provide a continuum of health care, and as a consequence of this, failure at one level can impact the chain of care at another level.17 Although var-ious government initiatives have led to improved basic service delivery at primary care health facilities over the last few years, still a large number of patients directly visit higher-level facilities, leading to over-crowding of those facilities,18 which impacts quality of care as it stretches fa-cility resources in terms of both infrastructure and staff. In addition, the persistent shortage of medical staff in pub-lic facilities only aggravates the crowded condition at these facilities.19

Findings indicate that patients were generally satisfied with the care they received, and ratings and satisfaction were generally higher at lower levels of care. However, many were not satisfied with the cleanliness or privacy

15 Mpinga EK, Chastonay P. Satisfaction of patients: a right to health indicator? Health Policy. 2011; 100(2-3):144-150.16 Baltussen RM, Yé Y, Haddad S, Sauerborn RS. Perceived quality of care of prima-ry health care services in Burkina Faso. Health Policy Plan. 2002; 17: 42-48.17 National Health Mission, Ministry of Health and Family Welfare, Government of India. Framework for Implementation National Health Mission (2012-2017). New Delhi, India: Government of India, 2012. 18 Bajpai V. The Challenges Confronting Public Hospitals in India, Their Origins, and Possible Solutions. Advances in Public Health 2014; 2014: 27. 19 Rao M, Rao KD, Kumar AK, Chatterjee M, Sundararaman T. Human resources for health in India. The Lancet. 2011; 377(9765): 587-98.

C O N C L U S I O N S A N D P O L I C Y I M P L I C AT I O N S

provisions at the facility they visited. In general, satis-faction was higher with doctors than nurses or ANMs at hospitals, but not at all health centres. Holding other factors constant, male patients were less satisfied with their care from doctors, and patients who received all prescribed medication were more satisfied with care from nurses.

Most patients experienced short travel and wait times. Most patients traveled less than 30 minutes to receive care, with patients at lower-level facilities reporting the shortest travel times. Hospitals had the highest propor-tion of patients who had to wait more than 30 minutes to receive care; the lowest proportion of patients waiting more than 30 minutes was at sub health centres. How-ever only 6% of patients waited more than one hour to receive care.

Finally, nearly 20% of patients at sub health centres reported being unable to acquire prescribed drugs. Ensuring that all patients may obtain prescribed medi-cations at the time of their visit should be a priority, as it facilitates adherence and continuity of care.

With the developmental priorities for the government of India clearly highlighting the need to increase user participation in health care service delivery for better ac-countability,20 understanding how patients perceive the quality of the existing public health services, encompass-ing various dimensions of care such as time to receive medical attention, staff behavior, etc., could contribute to developing strategies to improve performance and utili-zation of the public health system.21

HealthinformationsystemThis study was dependent on the data availability at

the facilities for the various inputs and outputs. Because of the vast extent of data that were collected for five fi-nancial years across the facilities, there are several lessons regarding the common bottlenecks within the health in-formation system, both at the facility level and at the state level. In general, there is weak staff capacity for data cap-ture, management, and use (interpretation or planning) at all levels. No system of regular review of data at the facility level that could guide planning or improvement of ser-

20 Planning Commission, Government of India. Faster, sustainable and more inclusive growth: An approach to the Twelfth Five Year Plan. New Delhi, India: Government of India, 2012.21 World Health Organization (WHO). Global Health Observatory Data Repository. Geneva, Switzerland: WHO, 2016.

vice provision was observed.It is not possible to assess the outputs by disease/

condition other than that for deliveries, as data are not captured or collated by disease groups at the facilities. At the higher-level facilities, collation of patients seen at the facilities was not readily available, and it was not possible to assess the level of duplication of patients across the departments. Furthermore, documentation of patients as a new patient or a follow-up patient was neither standard-ized nor practiced across most health facilities. Therefore, data interpretation is possible only in terms of number of visits and not in terms of number of patients.

Data were either incomplete or inaccurate at some fa-cilities for expenditure, patient-related outputs, and staff numbers. In general, the expenditure documentation had the most bottlenecks with these data available across various sources for a given facility. For example, it is not possible to document the expenditures at a given facility without procuring relevant data from the facility, a higher level of facility (block level), district health society, and at times from the state. The most limited capacity was to capture the expenditure on drugs, medical consumables, and supplies.

SummaryThe ABCE project was designed to provide policymak-

ers and funders with new insights into health systems and to drive improvements. We hope these findings will not only prove useful to policymaking in the state, but will also inform broader efforts to mitigate factors that impede the equitable access to or delivery of health services in In-dia. It is with this type of information that the individual building blocks of health system performance, and their critical interaction with each other, can be strengthened. More efforts like the ABCE project in India are needed to continue many of the position trends highlighted in this report and overcome the identified gaps. Analyses that take into account a broader set of the state’s facilities, including private facilities, may offer an even clearer pic-ture of levels and trends in capacity, efficiency, and cost. Continued monitoring of the strength and efficiency of service provision is critical for optimal health system per-formance and the equitable provision of cost-effective interventions throughout the states and in India.

5352

Annex:Facility-specific data for 2007 to 2011utilized for the efficiency analysis Please note that data may be missing for some years across the facilities based on availability of data.DH: District hospital; SDH: Sub-district hospital; CHC: Community health centre; PHC: Primary health centre

FACILITY INFORMATION STAFF AND BEDS OUTPUTS EXPENDITURE

District Platform Year Beds Doctors Nurses Paramedical Nonmedical Outpatient Inpatient Vaccinations Births PersonnelInfrastructure+

utilities

Medical supplies+

pharmaceuticalsAdministration

andtraining Non-medical

2 District Hospital (DH) 2007 415 40 112 68 78 1,066,885 94,463 6,041 4,042 29,562,816 7,720,422 7,214,515 200,645 4,030,000





2 Sub-district Hospital (SDH) 2007 66 10 10 18 10 217,543 2,103 180 210 13,888,862 613,452 758,374 17,870 112,560





2 Community Health Centre (CHC) 2007 6 5 4 17 6 89,348 1,109 4,664 226 7,226,222 636,529 239,339 51,651 26,080





2 Primary Health Centre (PHC) 2007 6 2 3 7 1 33,074 261 3,187 38 2,637,805 152,761 137,629 23,075 7,434










2 Sub-district Hospital (SDH) 2007 40 4 4 11 4 116,122 901 39,916 19 7,359,315 3,383,181 524,919 16,195 24,854

2 Sub-district Hospital (SDH) 2008 40 4 4 11 4 100,529 964 39,501 15 7,673,693 3,389,299 639,163 17,925 46,206

2 Sub-district Hospital (SDH) 2009 40 7 6 14 4 130,361 1,968 50,114 34 9,170,000 3,415,434 848,834 33,101 30,286

5554



utilities

Medical supplies+















2 Primary Health Centre (PHC) 2007 2 2 3 9 1 39,680 308 0 60 462,650 89,709 126,499 3,150 4,500






3 District Hospital (DH) 2008 608 49 90 122 101 857,942 37,637 9,608 3,380 101,142,456 1,929,940 10,920,849 442,343 1,138,279




3 Sub-district Hospital (SDH) 2007 76 12 2 14 13 414,905 7,033 2,454 402 12,768,250 1,238,766 1,401,195 50,910 109,755





3 Community Health Centre (CHC) 2007 6 2 3 13 6 43,860 414 2,222 93 4,584,000 277,946 139,864 19,320 83,950












5756



utilities

Medical supplies+




3 Primary Health Centre (PHC) 2010 4 2 3 5 1 25,981 1,108 1,986 164 2,307,862 277,892 184,551 16,246 54,478


3 Sub-district Hospital (SDH) 2007 30 8 7 10 3 99,379 1,908 1,399 145 8,165,801 116,410 519,126 31,280 115,342














3 Primary Health Centre (PHC) 2011 5 2 3 12 3 43,563 731 760 114 5,186,208 201,080 189,904 42,594 40,648






4 District Hospital (DH) 2007 241 55 40 56 38 428,233 21,010 2,442 3,415 6,492,954 782,133 5,017,705 72,860 530,018

4 District Hospital (DH) 2008 241 55 40 56 38 507,816 25,356 1,659 4,404 6,716,595 7,273,183 6,102,681 75,870 548,866










4 Community Health Centre (CHC) 2008 10 3 3 15 5 73,588 545 8,632 266 6,424,167 185,925 168,401 1,364,437 229,600

4 Community Health Centre (CHC) 2009 10 3 3 15 5 47,687 1,796 8,619 264 6,617,289 343,567 228,738 1,781,316 158,175


4 Community Health Centre (CHC) 2011 10 3 3 15 4 63,307 1,690 3 356 7,057,635 448,082 266,527 1,639,960 141,350


5958



utilities

Medical supplies+







4 Primary Health Centre (PHC) 2007 4 2 3 12 1 21,183 355 11,474 72 3,123,353 158,504 97,971 680 10,635

























5 District Hospital (DH) 2007 445 37 58 72 62 637,483 106,345 0 2,820 49,698,636 14,896,651 6,618,859 180,793 1,966,184


5 District Hospital (DH) 2009 445 39 64 74 65 1,000,795 59,717 0 2,489 64,578,588 15,035,472 7,020,596 348,323 2,067,487

5 District Hospital (DH) 2010 445 43 66 77 64 984,992 42,336 0 2,763 69,529,840 15,872,825 6,622,404 398,281 9,644,161


5 Sub-district Hospital (SDH) 2007 258 19 34 28 17 452,310 20,582 10,881 3,497 30,945,390 4,901,502 2,806,004 76,442 1,527,236





6160



utilities

Medical supplies+



5 Community Health Centre (CHC) 2007 3 4 3 17 4 75,752 590 0 64 7,431,071 39,093 150,449 0 0


5 Community Health Centre (CHC) 2009 3 4 3 17 4 64,978 1,038 0 118 7,925,512 39,093 214,996 0 0

































6 District Hospital (DH) 2007 313 32 41 41 46 523,699 90,483 7,261 3,388 51,823,836 350,155 6,508,595 21,507 3,015,500




6362



utilities

Medical supplies+










6 Community Health Centre (CHC) 2008 30 7 4 16 5 68,569 693 1,673 140 10,907,497 947,220 192,915 18,117 1,342,240



6 Community Health Centre (CHC) 2011 30 7 5 17 6 60,643 742 1,514 97 11,497,524 1,118,100 313,851 26,900 880,277















6 Sub-district Hospital (SDH) 2011 33 5 5 9 2 186,566 2,384 3,393 75 6,323,000 201,649 1,050,608 5,432 22,765

6 Community Health Centre (CHC) 2007 28 4 4 11 4 37,734 826 10,253 205 5,743,907 119,834 152,593 6,863 418













6564



utilities

Medical supplies+





7 District Hospital (DH) 2007 421 15 20 16 8 102,938 9,729 2,648 969 769,012 206,500 4,652,155 20,687 91,229

7 District Hospital (DH) 2008 421 14 20 17 10 111,448 16,848 4,449 1,789 894,529 211,714 5,010,731 13,717 126,034













7 Community Health Centre (CHC) 2011 30 7 3 13 3 23,307 273 994 66 7,605,704 139,703 178,259 15,052 90,922























6766



utilities

Medical supplies+















8 District Hospital (DH) 2011 310 63 48 43 13 1,006,651 113,687 12,107 4,067 60,056,976 332,130 4,643,214 45,961 2,500,000

8 Sub-district Hospital (SDH) 2007 178 23 37 23 19 445,179 64,999 3,813 1,636 11,935,030 117,053 3,331,378 19,850 68,500










8 Primary Health Centre (PHC) 2007 2 2 3 11 1 40,494 101 5,189 56 385,383 24,789 135,061 5,700 8,000















8 Community Health Centre (CHC) 2007 30 2 3 11 3 69,718 332 2,148 182 4,311,264 256,300 204,209 41,821 365

6968



utilities

Medical supplies+







8 Primary Health Centre (PHC) 2007 4 2 4 17 2 41,783 255 7,349 31 2,893,400 60,770 139,480 22,535 780















9 Sub-district Hospital (SDH) 2007 16 3 2 6 0 29,551 265 50 32 1,972,154 241,957 779,926 9,346 9,285

9 Sub-district Hospital (SDH) 2008 16 3 3 6 0 32,614 278 45 36 2,033,145 253,853 846,409 21,664 9,411


9 Sub-district Hospital (SDH) 2010 16 3 3 6 0 36,223 1,075 44 20 2,160,851 266,274 1,000,198 32,045 15,977


9 Community Health Centre (CHC) 2007 28 2 4 22 6 52,335 723 3,623 181 6,403,877 1,859,528 136,881 38,328 3,121,667















7170



utilities

Medical supplies+



9 Sub-district Hospital (SDH) 2007 52 4 10 15 3 100,355 350 48,396 20 8,522,976 110,915 534,725 43,802 120,088









9 Community Health Centre (CHC) 2011 30 4 3 16 6 70,287 1,191 13,590 109 52,689,648 437,342 426,064 245,392 7,664,983











9 Sub Health Centre (SHC) 2007 0 0 0 1 0 2,921 0 648 0 261,564 18,171 4,059 7,480 2,750






10 Sub-district Hospital (SDH) 2007 178 22 40 39 25 331,716 11,282 5,545 2,403 38,793,280 1,307,828 2,558,484 41,701 196,301





10 Community Health Centre (CHC) 2007 30 1 2 10 4 78,649 729 427 97 4,038,615 177,758 205,538 24,300 137,160



10 Community Health Centre (CHC) 2010 30 2 5 12 4 89,786 1,044 770 68 6,391,486 175,465 343,464 25,790 143,980





7372



utilities

Medical supplies+









10 Primary Health Centre (PHC) 2011 2 2 4 10 1 43,867 705 4,422 44 3,440,424 90,182 203,792 84,650 0

































7574



utilities

Medical supplies+










































7776



utilities

Medical supplies+







12 Community Health Centre (CHC) 2007 17 3 3 23 5 74,171 672 0 442 9,125,621 205,930 200,000 1,695,561 5,881,340

12 Community Health Centre (CHC) 2008 17 3 3 23 5 65,153 900 0 609 9,401,736 204,951 224,999 2,693,294 3,022,470

12 Community Health Centre (CHC) 2009 17 3 3 23 5 85,478 668 6,582 602 9,686,391 291,598 305,390 4,519,609 4,793,269

12 Community Health Centre (CHC) 2010 17 3 3 23 5 82,261 1,081 6,510 504 9,979,849 329,937 355,538 4,322,606 6,041,624

12 Community Health Centre (CHC) 2011 17 3 3 23 5 104,202 1,807 7,154 926 10,279,374 349,417 596,232 3,374,426 8,680,173































7978



utilities

Medical supplies+
































13 Sub-district Hospital (SDH) 2011 56 8 7 6 3 119,981 4,245 2,450 26 7,372,932 3,019,179 923,900 440 11,900










8180



utilities

Medical supplies+









14 District Hospital (DH) 2007 84 23 27 16 9 225,699 7,662 6,350 568 35,006,760 123,421 1,771,439 15,325 19,657

14 District Hospital (DH) 2008 84 23 29 16 10 230,310 10,194 7,187 487 35,884,084 122,726 2,004,184 0 12,191
























14 Sub-district Hospital (SDH) 2007 198 40 41 29 27 710,962 24,205 973 2,961 51,990,464 1,671,650 3,198,514 18,617 166,114




14 Sub-district Hospital (SDH) 2011 198 41 44 29 27 829,210 24,608 777 2,354 64,633,828 1,656,912 4,109,763 21,600 204,329




8382



utilities

Medical supplies+















INSTITUTE FOR HEALTH METRICS AND EVALUATION2301 Fifth Ave., Suite 600Seattle, WA 98121USA

TELEPHONE: +1-206-897-2800FAX: +1-206-897-2899EMAIL: [email protected]


A B CE

PUBLIC HEALTH FOUNDATION OF INDIAPlot 47, Sector 44Gurugram, National Capital Region 122002India

TELEPHONE: +91 124 478 1400FAX: +91 124 478 1601EMAIL: [email protected]

health service provision in tamil nadu · ihme institute for health metrics and evaluation iphs...

Documents