ABCE: Understanding the costs of and constraints to health service delivery in Zambia

Access, Bottlenecks, Costs, and Equity (ABCE):Understanding the costs of and constraints to health service delivery in Zambia

On behalf of the ABCE research teamInstitute for Health Metrics and Evaluation | University of Zambia

January 2015

Overview

• Overview of the ABCE project in Zambia

• Key findingso Facility capacity and service provisiono Non-HIV patient perspectiveso Efficiency and costs of careo A focus on HIV: service provision and

patient characteristics

• Using ABCE work and findings for policymaking

• Conclusions

Overview of the ABCE project in Zambia

Overview of the ABCE project in ZambiaABCE study design and implementation

• Collaboration between UNZA and IHME

• Primary data collectiono September 2011 – April 2012

• Three main data collection mechanisms:o ABCE Facility Survey

o Clinical chart extractions of HIV-positive patients on ART

o Patient Exit Interview Survey

Overview of the ABCE project in ZambiaABCE Facility Survey

• Primary data collection from a nationally representative sample of 188 facilities

• Collected data on a full range of indicatorso Inputs, finances, outputs, supply-side

constraints and bottlenecks, indicators for HIV care

• Randomly sampled a full range of facility typeso All levels of hospitals, health centers

(urban and rural), health posts, drug stores or pharmacies, and DHMTs

Overview of the ABCE project in ZambiaClinical chart extraction

• Extracted data on HIV-positive patients currently enrolled in ART

• Chart data included patient demographic information, ART initiation characteristics (e.g., CD4 cell count, WHO stage, drug regimen, referral points), and patient outcomes

Overview of the ABCE project in ZambiaPatient Exit Interview Survey

• Over 2,700 structured interviews were conducted with patients after they exited facilities from the ABCE sample.

• Interviewees included patients who sought HIV care and those who presented at facilities for non-HIV services.

• Questions included reasons for the facility visit, satisfaction with services, expenses paid associated with the facility visit, and HIV-specific indicators.

Key findings from the ABCE project in ZambiaFacility capacity and service provision

Facility capacity and service provisionAvailability of health services in 2011-2012

• Relatively high availability of key services across facilities:o 88% provided family planning

o 86% had a formal immunization program

o 74% had HIV/AIDS care

o 75% stocked ACTs for treating malaria

• Other services remained fairly scarce, particularly at lower levels of care:o e.g., nutrition services were available at only 49% of urban health

centers, 38% of rural health centers, and 13% of health posts

Facility capacity and service provisionGaps in reported and full capacity for care

• Many facilities reported providing a given service, but then lacked the full capacity to provide that service.o e.g., stocking out of key childhood immunizations or storing vaccines

outside the recommended temperature range

ServiceFacilities reporting

capacity

Facilities with four key childhood vaccines and

proper storage temperature

Immunization services 86% 53%

Facility capacity and service provisionAvailability of vaccines and capacity for immunization services

• All private hospitals and 88% of health posts stocked all four key childhood vaccines (BCG, measles, OPV, pentavalent).

• After private hospitals, health posts and urban health centers had the lowest levels of any vaccine stock-outs (13% and 15%, respectively).o Pentavalent was the most commonly stocked-out vaccine.

• Among the 22 districts in the ABCE sample, four did not have any facilities with a vaccine stock-out at the time of visit: Chadiza, Chama, Kasempa, and Nyimba.

• Of the facilities that routinely stored vaccines, 28% had refrigerators operating outside the optimal range (2°C to 8°C).o More facilities had temperature readings below 2°C than above 8°C.

Facility capacity and service provisionAvailability of vaccines, by platform, 2011-2012

Facility capacity and service provisionVaccine stock-outs, by platform, 2011-2012

Facility capacity and service provisionVaccine storage temperature for immunization services, 2011-2012

Facility capacity and service provisionAvailability of malaria treatment

• Majority of facilities stocked artemisinin-based combination therapies (ACTs) at the time of facility visit.o ACT availability ranged from 67% of private health centers to 90% of

level 2 and 3 hospitals.

• Across platforms, 10% of facilities had stocked out of ACTs at the time of facility visit.o Private health centers and rural health centers had the highest

proportion of facilities stocking out of ACTs (33% and 14%, respectively).

• Except for private health centers, 10% to 17% of facilities reported never carrying ACTs.o This did not vary much by level of care, facility ownership, or urbanicity.

Facility capacity for service provisionAvailability of ACTs, by platform, 2011-2012

Facility capacity and service provisionAvailability of modern contraceptives

• Condoms, injectables, and oral contraceptives were the most widely available modern contraceptives across facilities.

• Excluding private hospitals, at least 70% of facilities offered condoms and at least one type of female family planning method.o 96% of rural health centers had these two forms of modern

contraceptives.

• Fewer facilities stocked condoms, at least one type of female family planning method, and emergency contraceptives. o Private hospitals and urban health centers had the highest availability of

these three contraceptives (50% of facilities for each platform).

Facility capacity for service provisionAvailability of modern contraceptives, 2011-2012

Facility capacity and service provisionCapacity to test for and treat hypertension

• Most facilities had at least a blood pressure cuff or a type of medication to treat high blood pressure.o Level 1 hospitals and private hospitals had the highest capacity for full

case management, with 92% and 83% of facilities, respectively, stocking antihypertensives and having a blood pressure cuff.

• An urban-rural divide emerged, with 9% of rural health centers and 18% of health posts lacking both hypertension diagnostics and treatment.

• A substantial number of private facilities had the diagnostic equipment for hypertension but lacked antihypertensives –whereas the opposite was more often true for public and NGO-owned facilities (i.e., stocked antihypertensives but lacked a blood pressure cuff).

Facility capacity for service provisionCapacity to test for and treat hypertension, 2011-2012

Facility capacity and service provisionAvailability of and deficiencies in physical capital

• Power supplyo Access to functional electricity was largely divided by location.

Nearly 100% of urban facilities were connected to the energy grid. 38% of rural health centers and 47% of health posts lacked functional electricity.

o 30% of facilities with functional electricity also had a generator.

o About one-third of rural health centers and health posts were powered by solar power, but none of these facilities had a generator.

• Water and sanitationo Access to improved water sources varied by location and level of care.

77% of level 1 hospitals had piped water. 12% of health posts lacked any source of water on or near facility grounds.

o Primary waste systems followed an urban-rural divide.o Nearly all facilities in urban areas had flush toilets.o Two-thirds of rural health centers and health posts had covered pit latrines, but

about 25% of them also had uncovered pit latrines.

• Transportation and communicationo The majority of primary care facilities lacked emergency transportation and

did not have access to a facility-based phone.

Facility capacity and service provisionAvailability of and deficiencies in physical capital, 2011-2012

Facility capacity and service provisionAvailability of equipment across platforms

• Substantial gaps in equipment availability were found across all facility types.o 89% of level 2 and 3 hospitals did not have an ECG machine.

o 63% of urban health posts did not have a microscope.

o 53% of health posts lacked equipment to measure blood pressure.

• Private facilities generally had higher availability of equipment than equivalent facilities in the public sector.

• A subset of facilities had all of the required medical equipment for their level of care (five health posts, one urban health center, and one rural health center).

Facility capacity and service provisionHuman resources for health, 2010

• Non-medical staff accounted for the largest proportion of personnel across most facilities, ranging from 31% at private health centers to 48% at level 2 and 3 hospitals.

• Urban facilities generally had more doctors, clinical officers (COs), nurses, and midwives than rural facilities at the same level of care.o Urban health centers averaged nine nurses per facility.

o Rural health centers averaged two nurses per facility.

• 62% of health centers had at least two skilled medical personnel on staff, which was much higher than a 2009 baseline of 47%.

Facility capacity and service provisionHuman resources for health: hospitals, 2010

Facility capacity and service provisionHuman resources for health: health centers, 2010

Facility capacity and service provisionOutputs, 2006-2010

• Outpatient visits remained relatively stable over time across facilities.o The clear exceptions were health posts and level 2 and 3 hospitals, at

which an average of 10% and 11% annual gains, respectively, in outpatient visits occurred between 2006 and 2010.

• Inpatient visits gradually increased at most facilities.o Rural health centers were the exception, with average inpatient visits

decreasing 19% between 2006 and 2010.

• ART visits rapidly rose across platforms from 2006 to 2010.o There was 279% increase across all facilities.o This growth was particularly evident among urban health centers, at

which average ART visits nearly quadrupled during this time.

Facility capacity and service provisionOutputs: average outpatient visits, by platform, 2006-2010

Facility capacity and service provisionOutputs: average inpatient visits, by platform, 2006-2010

Facility capacity and service provisionOutputs: average ART visits, by platform, 2006-2010

Key findings from the ABCE project in ZambiaNon-HIV patient perspectives

Non-HIV patient perspectives Patient reports of expenses associated with facility visit

• As part of the Patient Exit Interview Survey, patients who did not seek HIV services reported the types of expenses they had in association with the facility visit.

• Zambian policy abolished user fees for primary health services and medical care in rural areas in 2006.

• Based on the ABCE sample, few patients (6%) reported medical expenses associated with visits to level 1 hospitals, rural health centers, and health posts.o Many more patients reported having transportation expenses.

Non-HIV patient perspectives Patient reports of expenses associated with facility visit, 2011-2012

Non-HIV patient perspectives Levels of patient medical expenses

• Of patients who had medical expenses at public and NGO-owned facilities, the majority spent less than 5 kwacha ($1).

• By contrast, most patients who had medical expenses at private facilities spent more than 50 kwacha ($9) for care.

Efficiency and costs of careLevels of patient medical expenses, by facility, 2011-2012

Non-HIV patient perspectives Patient wait times at facilities

• A large portion of patients waited at least one hour before receiving care at public and NGO-owned facilities, while most patients who sought care at private facilities received care within an hour.

• At level 2 and 3 hospitals, 53% of patients spent at least two hours waiting for care. At private hospitals, nearly 60% of patients received care within 30 minutes.

• A greater proportion of patients received care within one hour at urban health centers than rural health centers.

Non-HIV patient perspectives Patient reports of wait times at facilities, by platform, 2011-2012

Non-HIV patient perspectives Patient ratings of facilities

• Overall, patients gave high ratings for care received across platforms.o Private hospitals generally received the highest ratings across all

indicators.

• Patients rated staff interactions highly, especially for provider respectfulness.

• Patients generally gave lower ratings to facility characteristics, particularly for wait time.

Non-HIV patient perspectives Patient overall ratings of facilities, by platform, 2011-2012

Non-HIV patient perspectives Average patient ratings of facility indicators, by platform, 2011-2012

Key findings from the ABCE project in ZambiaEfficiency and costs of care

Efficiency and costs of careEstimating efficiency: Data Envelopment Analysis (DEA)

• DEA: quantifies the relationship between a facility’s resources (medical staff, beds) and its production of services (outpatient visits, inpatient bed-days, births, and ART visits) relative to comparably sized facilities in the ABCE sample.

• Efficiency score: a value between 0% and 100%, reflecting the alignment of facility resources to service production.o 100% = maximum use of facility resources for output production

• Outpatient equivalent visits (OEV): weighting different outputs in a standardized way to allow for direct comparisons across facilities.o Average across facilities:

Inpatient bed-day = 3.7 outpatient visits Birth = 10.6 outpatient visits ART visit = 1.6 outpatient visits

Efficiency and costs of careAverage production of outputs across facilities

• Across platforms, facilities averaged a total of 8 outpatient equivalent visits per medical staff per day, ranging from 3 visits at private hospitals to 16 visits at health posts.

• Outpatient visits accounted for the largest proportion of patient visits experienced per medical staff per day at primary care facilities (health centers and health posts).

• Inpatient bed-days accounted for the largest proportion of patient visits produced per medical staff per day at hospitals.

• Urban health centers recorded the largest volume of ART visits per medical staff per day (0.9, as measured in OEV).

Efficiency and costs of careAverage production of outputs across facilities, 2010

Note: All visits are in outpatient equivalent visits, with an average of one inpatient bed-day equaling 3.7 outpatient visits; one birth equaling 10.6 outpatient visits; and one ART visit equaling 1.6 outpatient visits.

Efficiency and costs of careEfficiency scores varied across and within platforms

• Across all facilities, the average efficiency score was 42%.

• 70% of facilities had an efficiency score at or less than 50%.

• Average efficiency scores were generally higher at public or NGO-owned facilities than at private facilities at the same level of care.

• Tremendous range in efficiency scores among non-private facilities:o At least one facility had an efficiency score of 78% or higher for each platform.o Multiple facilities had efficiency scores close to 0% for each facility type.

• No consistent relationship between urbanicity and efficiency scores:o Urban level 2 and 3 hospitals generally had higher efficiency scores than rural

hospitals.o Rural health centers often had higher efficiency scores than urban health centers.

Efficiency and costs of careEfficiency scores across platforms, 2006-2010

Efficiency and costs of careEstimated potential for expanded service production

• We estimated that facilities had substantial potential for increasing output production, especially among private facilities.

• An average of 13 additional visits, measured in OEV, could be added across facilities, based on observed resources.

• This potential for expanded service production does not reflect the quality of services delivered; it shows the alignment of facility resources and output production.

• 14% of facilities in Zambia had an efficiency score of 80% or higher.o Potential to learn from these facilities to further bolster efficiency.

Efficiency and costs of careEstimated potential for expanded service production, 2010

Efficiency and costs of careCross-country comparison of efficiency

• Zambia showed less potential for expanded service provision, given observed resources, than a subset of other countries in the ABCE project.o Suggests that many Zambian facilities have already increased service

production.

Efficiency and costs of careEstimating costs of care

• Using information produced through DEA, output-specific spending by facilities was divided by outputs produced by each facility.

• All cost data were adjusted for inflation and reported in 2010 Zambian kwacha.o All US dollar estimates were based on the 2010 exchange rate of 5.34

kwacha per $1.

o All estimates were converted to align with the 2013 rebasing of the kwacha by dividing cost estimates by 1,000.

Efficiency and costs of careAverage facility cost per visit, across outputs and by platform

• Facility costs per patient visit varied across platforms and by output type.

• The average facility cost per outpatient visit was generally the least expensive to produce, and births were the most expensive.

• Private hospitals generally spent the most per patient visit produced, whereas health posts and rural health centers generally produced patient visits at the lowest facility cost per output.

• Aside from births, level 1 hospitals had more similar average costs per output to costs estimated for health centers than for other hospitals.

Efficiency and costs of careAverage facility cost per visit, across outputs and by platform, 2010

Efficiency and costs of careCross-country comparison of output costs, 2010

• Zambian facilities averaged the least expensive production cost per inpatient bed-day, and were on the lower end for outpatient visits and births.

Key findings from the ABCE project in ZambiaA focus on HIV: service provision and patient characteristics

HIV service provision and patient characteristicsART regimen at initiation, 2007-2010

• From 2007 to 2010, there was a continued transition away from d4T-based ART regimens toward those with a TDF backbone for ART initiates.• Zambia was one of the first countries in sub-Saharan Africa to adopt TDF-

based therapies as its first line ARV.

• TDF prescription rates varied across facilities, from 11% to 100% in 2010. o Health centers generally had slightly higher proportion of ART patients

initiating on TDF-based regimens than hospitals in 2010.

HIV service provision and patient characteristicsART regimen at initiation, 2007-2010

HIV service provision and patient characteristicsART regimen at initiation, by facility, 2010

HIV service provision and patient characteristicsPatient clinical characteristics at ART initiation: WHO staging

• There was a gradual shift toward ART initiation at earlier stages of disease progression between 2007 and 2010.

• In 2007, 31% of patients initiated at WHO stage 1 or 2. In 2010, 53% began treatment at the same stages.

• There was substantial heterogeneity in ART initiation by WHO stage across facilities in 2010.o In general, level 2 and 3 hospitals saw a greater proportion of ART

patients starting therapy at WHO stage 1 than health centers.

HIV service provision and patient characteristicsWHO stage at initiation, 2007-2010

HIV service provision and patient characteristicsWHO stage at initiation, by facility, 2010

HIV service provision and patient characteristicsPatient clinical characteristics at ART initiation: CD4 cell count

• A greater proportion of ART patients began therapy at higher CD4 cell counts in 2010 than in 2007.o In 2007, 34% of patients initiated at a CD4 cell count of 200 cells/mm3

or higher. In 2010, 45% of patients initiated at this level of CD4.

• Median CD4 cell count increased 31%, from 143 cells/mm3 in 2007 to 187 cells/mm3 in 2010.

• A substantial portion of ART patients still began therapy once they were symptomatic.o About 10% of patients initiated ART with a CD4 cell count less than 50

cells/mm3 from 2007 to 2010.

HIV service provision and patient characteristicsCD4 cell count at initiation, 2007-2010

HIV service provision and patient characteristicsFacility availability of patient clinical information

• Testing rates have remained stable over time, indicating that recordkeeping has increased in parallel with rising ART patient volumes.

• In 2010, a portion of ART initiates still did not receive key tests.o 15% lacked a CD4 cell count

o 3% were not assigned a WHO stage

o 3% did not have a weight measurement

o 22% did not have a height measurement

• Much progress was made in administering follow-up tests during the second year of therapy.

HIV service provision and patient characteristicsFacility availability of patient clinical information

HIV service provision and patient characteristicsHIV patient reports of expenses associated with visit

• As part of the Patient Exit Interview Survey, patients who sought HIV services reported the types of expenses they had in association with their facility visits.

• Zambian national policy stipulates that ART services should be free at public facilities.

• Based on the ABCE sample, very few HIV patients (< 5%) reported any medical expenses associated with visits to public facilities.

• More than 30% of HIV patients experienced some kind of transportation expense at higher-level hospitals and urban health centers.

HIV service provision and patient characteristicsHIV patient reports of expenses associated with visit, 2011-2012

HIV service provision and patient characteristicsHIV patient reports of wait times at facilities

• Overall, HIV patients reported relatively long wait times at facilities and often spent more time waiting than non-HIV patients at similar facilities.

• This was consistently found across platforms:o Level 1 hospitals

43% of HIV patients waited at least two hours.

23% of non-HIV patients waited more than two hours.

o Urban health centers 21% of HIV patients received care within one hour.

41% of non-HIV patients received care within one hour.

HIV service provision and patient characteristicsHIV patient wait times at facilities, by platform, 2011-2012

HIV service provision and patient characteristicsHIV patient ratings of facilities

• Overall, HIV patients gave high ratings for care received across platforms.o Nearly 50% of HIV patients gave level 2 and 3 hospitals a 10 out of 10.

• HIV patients generally gave higher ratings, across facility indicators, than non-HIV patients – except for wait time.

• Like non-HIV patients, HIV patients rated staff interactions highly, especially for provider respectfulness.

• HIV patients gave high ratings of facility privacy, but rated wait time very poorly – especially at hospitals.

HIV service provision and patient characteristicsHIV patient overall ratings of facilities, by platform, 2011-2012

HIV service provision and patient characteristicsAverage HIV patient ratings of facility indicators, by platform, 2011-2012

HIV service provision and patient characteristicsEfficiency scores for facilities providing ART

• Across facilities with ART, the average efficiency score was 49%.

• ART facilities typically had higher levels of efficiency, compared to all facilities in the ABCE sample.

• Potential to expand ART patient volumes, especially among level 1 hospitals.

HIV service provision and patient characteristicsEfficiency scores for facilities providing ART

HIV service provision and patient characteristicsEstimated potential for increased ART visits given resources

• We estimated that many facilities had potential for increasing annual ART visits.

• Given observed facility resources, we estimated that an average of 9,063 additional ART visits could be added, per facility, each year.

• This gain represents a 117% increase in ART visits from the average annual ART visits observed in 2010 (7,727 visits).

HIV service provision and patient characteristicsEstimated potential for increased ART visits given resources

HIV service provision and patient characteristicsCross-country comparison of ART efficiency

• Zambia showed substantial potential for expanded ART provision, given observed resources, and at a greater magnitude than Kenya and Uganda.

HIV service provision and patient characteristicsProjected facility ART costs: analytical approach

• Four streams of data were used to project ART costs:1. Average facility cost per ART visit, excluding ARVs, based on the ABCE

sample2. Recommended number of annual visits for new and established ART

patients, based on 2010 national guidelines3. The ARV regimens of ART patients in 2010 extracted from clinical charts4. The ceiling ARV prices for 2010 published by the Clinton Health Access

Initiative (CHAI)

• Analytical steps for projecting ART costs1. Visit costs: multiplied average facility cost per ART visit, excluding ARVs, by

the recommended number of annual visits observed for new and established ART patients.

2. Total costs: using the relative proportion of TDF-, d4T-, and AZT-based regimens observed for patients, applied the ceiling price for each ARV and added projected ARV costs to estimated visit costs.

HIV service provision and patient characteristicsProjected facility ART costs, 2010

• ARVs accounted for a large portion of projected annual facility costs for ART, but varied across patient types and platforms.o New patients

ARVs accounted for 41% of total projected ART costs at level 2 and 3 hospitals.

ARVs accounted for 73% of total projected ART costs at rural health centers.

o Established patients ARVs accounted for 61% of total projected ART costs at level 2 and 3 hospitals.

ARVs accounted for 86% of total projected ART costs at rural health centers.

• Facility costs for ARVs may be viewed as more stable over time, whereas visit costs associated with ART services are likely to be lower for established patients.o Substantial implications for longer-term ART care and funding sources

HIV service provision and patient characteristicsProjected facility costs for ART, 2010

HIV service provision and patient characteristicsCross-country comparison of ART costs, 2010

• Zambian facilities had higher ART costs than the average facility costs estimated for Kenya and Uganda.

• ARVs accounted for 68% of annual facility costs in Zambia, which is slightly less than Kenya (71%) and Uganda (73%).

Using ABCE work and findings for policymaking

Using ABCE for policymakingIdentifying health system progress and challenges

• Provides policymakers with the evidence to pinpoint areas of success and for improvement as linked to national goals and priorities.

• Enables direct comparisons across facility types and ownership, allowing policymakers to contrast facility capacity in the public sector with that of the private sector.

• Supports the timely use of data to inform policy dialogues.

Using ABCE for policymakingABCE Zambia policy report

http://www.healthdata.org/dcpn/zambia

Conclusions

ConclusionsFacility capacity for service provision

• High availability of a subset of services reflects how Zambia has expanded service availability throughout the country.o Family planning, immunization services, HIV/AIDS care, availability of ACTs.

• Substantial gaps in reported capacity and full capacity to provide services found across all levels of care. o This was particularly pronounced among primary care facilities and for the

management of NCDs.

• Facility infrastructure and the availability of functional electricity, improved water, and sanitation systems largely followed an urban-rural divide.

• All facility types experienced some equipment deficiencies, irrespective of level of care.

• Non-medical staff generally accounted for the largest proportion of facility personnel. Urban facilities generally had more skilled medical personnel than rural facilities.

ConclusionsFacility production of health services

• Average patient volumes generally remained stable or gradually increased over time, whereas ART visits rapidly increased between 2006 and 2010.

• Shortages in human resources and facility overcrowding have been viewed as widespread; in the ABCE sample, most facilities averaged fewer than eight visits per medical staff per day.

• Given observed facility resources, service production could be potentially increased by an additional 13 outpatient equivalent visits, on average, per facility.

• Annual ART visits could potentially increase as well, but at a larger magnitude (a 117% gain).

ConclusionsPatient perspectives

• Among public facilities in rural areas, few patients reported any medical expenses associated with their facility visit.o This reflects Zambia’s prioritization of removing cost barriers to health

services.

• In general, a large portion of patients spent more time waiting at facilities to receive care than the time they spent traveling to the facility.o Given average staffing observed across facilities and patients seen per

medical staff per day, it is unlikely that inadequate human resources are the main driver of these long wait times.

• Patients gave high ratings of facilities, especially HIV patients. o Staff interactions were regularly rated higher than facility characteristics.o Patients gave fairly low ratings of wait time, particularly HIV patients.

ConclusionsFacility costs of care

• Average facility cost per patient visit differed substantially across platforms and types of visits.

• In comparison with a subset of other countries in the ABCE sample, Zambia had the lowest average facility cost per inpatient bed-day in 2010.

• On average, ARVs accounted for a large proportion of ART facility costs, but how much varied based on patient status (new or established).o Projected ART facility costs, including ARVs, were generally higher in

Zambia in comparison with Kenya and Uganda, but ARVs contributed to a smaller portion of overall annual costs in Zambia (68%) than the other two countries (71% and 73%, respectively).

ConclusionsFacility-based provision of ART services

• The shift away from d4T-based ART regimens and toward TDF continued throughout Zambia – a significant success.

• Gradual progress took place for initiating ART patients at earlier stages of disease, for both WHO staging and CD4 cell counts.• However, a portion of patients still began treatment after

becoming symptomatic in 2010.

• Steady improvements were made in collecting ART patient clinical data, but too few did not receive key measures and tests at initiation and during follow-up visits.o Greater investment in ART patient recordkeeping and data collection

ought to be considered.

ConclusionsPriority considerations for future work

• Updated analyses across indicators to assess progress and to identify areas that may require more investment.

• Targeting a broader set of facilities to capture a clearer picture of levels and trends in facility performance.

• Linking estimates of efficiency to quality of the services produced at facilities, as well as other factors. o e.g., expediency with which patients receive care, demand for increased services

• Updated analyses for ART patient characteristics at initiation, to determine more recent uptake of new eligibility guidelines.

• Generating estimates of cost effectiveness based on facility delivery of services and costs of production, and linking to ongoing work on estimating trends in health outcomes and disease burden.

Thank you

http://www.healthdata.org/dcpn/zambia