1. How does HIV/SRH service integration impact workload
How does HIV/SRH service integration impact workload? A descriptive analysis from the Integra Initiative in two African settings
Sedona Sweeney1*, Carol Dayo Obure1, Fern Terris-Prestholt1, Vanessa Darsamo, Christine Michaels-Igbokwe1, Esther Muketo3, Zelda Nhlabatsi4, the Integra Research Team, Charlotte Warren2, Susannah Mayhew1, Charlotte Watts1, Anna Vassall1
1 London School of Hygiene & Tropical Medicine 2 Population Council, Washington DC 3 Family Health Options Kenya 4 Family Life Association of Swaziland
Hi, my name is Sedona Sweeney – I’m a research fellow at the London School of Hygiene and Tropical Medicine, and I am part of the economics team for Integra. Today I’ll be presenting some work looking at the impact of integration on resource use, specifically looking at human resources and staff time.

2. Background:
Integration of HIV and SRH services may yield improvements in efficiency
Despite a clear rationale for integration, there is scarce evidence on the costs and potential efficiency gains of integrated service provision
Sweeney, S., et al. "Costs and efficiency of integrating HIV/AIDS services with other health services: a systematic review of evidence and experience." Sexually Transmitted Infections (2011)
This analysis was driven by a clear lack of evidence on the economic impact of integration, which we found during a systematic review we carried out in 2011, published in STI journal.
This analysis was carried out on data from all 40 health facilities involved in the Integra Initiative, with data collected at baseline and endline of the project.

3. To begin, here is a brief conceptual framework of the economic impact of integration. For the purposes of this analysis we focus on the measures of integration included in the ‘structural index’ mentioned by Jon – these include an increased availability of services either at the MCH level or at the facility level, consolidation of human resources, and consolidation of physical resources.
The theoretical economic benefits of integration are associated with two economic concepts surrounding efficiency in health service delivery. First, integration may result in ‘economies of scale’, or cost savings through an increase in the number of services delivered with the same level of resources. Integration can lead to increased output as it can enable staff to offer additional services to clients (for example PITC), or it can make health service users more aware of what is offered or reduce other barriers to use such as stigma - potentially leading to increased demand for services [24].
Second, it is hypothesized that integration can also result in ‘economies of scope’, or reductions of costs through joint production of goods/services. For example, an integrated service delivery model could reduce the number of times necessary to perform basic tasks such as height and weight measurement for a client receiving a number of different services. This would reduce the staff time required per patient, and therefore reduce the costs associated with that time.

4. Methods (1)
‘HR integration’: the provision of multiple services by one staff member
Core MCH services: family planning (FP), post-natal care (PNC), antenatal care (ANC)
Non-core services: STI management (STI), voluntary HIV testing and counselling (VCT), provider-initiated HIV testing and counselling (PITC), cervical cancer screening (CaCx), and HIV treatment and care
HR integration score:
To determine whether these hypothesized benefits can be realised in practice, we needed to develop a measure of HR integration. We define ‘HR integration’ as the provision of multiple services by one staff member. Specifically we define this as the combination of any ‘core’ MCH services (such as family planning, post natal care or antenatal care), with any ‘non-core services’ such as STI management, hiv testing and counselling, cervical cancer screening or hiv treatment and care.
This integration can be realised either through moving services from a stand-alone department to one providing both SRH and HIV services, or through adding services to the basic package offered by a staff member without dropping their pre-existing tasks.
Looking at these services, we developed an integration score calculated as the total number of non-core services available within the MCH unit (so taking a value between one and five), divided by the total staff full-time equivalency allocated to these services. So for example, if a facility had one staff member offering one service it would have an integration score of one, and if it had one staff member offering four services it would have an HR integration score of four.
Total number of non-core services available within MCH unit
Total staff FTE allocated to these services

5. Methods (2)
We estimated the time required to deliver services through a mixed methods approach
Key informant interviews with staff, time sheets and direct observation of services
Process and output data collected from routine monitoring registers
Workload ratio:
Next, we developed a measure of human resource use based on the WHO’s measure Workload Indicators of Staffing Needs or WISN. This involves comparing staff time available to staff time required within a facility.
First we estimated the time required to deliver different services through a mixed methods approach to staff time observation, estimating the typical time taken to deliver different types of services. We used these time observations combined with detailed service statistics to estimate the total staff FTE required to deliver services in each facility.
Our estimates conservatively assume 220 working days per year accounting for national holidays and leave time, and assume 33% of this time to be taken by administrative duties, trainings etc.
For some services including HIV counselling and testing and HIV care and treatment, we also considered the time of technical staff such as lab technicians and lay counsellors.
In order to arrive at the workload ratio, we divided actual staffing levels by this estimate of staff FTE required to deliver services within each facility. A workload ratio greater than one indicates some down time for staff members. As the ratio reaches one, the estimated time taken to deliver outputs is equal to the staff time available for patient visits within a facility, while ratio less than one would indicate that staff are overworked (i.e. the time required to attend patients is greater than the staff time available).
Actual staffing levels (available staff FTE)
Estimated staffing requirements for services delivered
Workload ratio > 1: some down time for staff
Workload ratio < 1: staff are likely overworked

6. Methods (3): Data Analysis
Bivariate categorical analysis of workload ratios in facilities with high and low integration scores, by facility type
‘More integrated’ (top 20% of integration scores)
‘Less integrated’ (bottom 80% of integration scores)
Data analysed using Excel and Stata 13
differences in workload estimates between HR integration category were explored, testing for significance at the p < 0.05 and p < 0.10 levels using Student’s t-tests, assuming unequal variance where applicable
We examined the link between these two measures using a bivariate categorical analysis, both for individual services and at the facility level.
We compared workload between facilities with high integration scores (which we classified as the ‘more integrated’ facilities), with those facilities with lower integration scores which we classified as ‘less integrated’. These classifications were stratified by facility type to avoid any bias associated with other health system factors.

7. Results

8. Improvements in Resource Integration from Baseline to Endline
First we take a look at which of our indicators improved from baseline to endline. This chart shows each of the four indicators of ‘structural integration’, with facility numbers along the x axis. For each indicator, if the bar is there, it represents an improvement from baseline to endline. So for example, facility number 8 on the far left improved in all four indicators, while facility 40 on the far right saw no improvement in any indicaor.
What we see here is that HR was more likely to improve in facilities which also improved other aspects of integration, such as availability of services or physical integration. Similarly, service availability within the MCH unit did not improve in any facility without some other measure also improving.

9. Staff Time Observed per Consultation
Number
Observations
Number of Minutes per Consultation
Mean
Median
95% CI
FP Visit
280
8.26
6.00
7.48 -
9.04
PNC Visit 98
14.43
15.00
12.78
16.09
Ca Cervix Screening Visit 20
7.40
6.50
5.33
9.47
STI Visit 9
11.44
11.00
8.89
14.00
HIV Counselling and Testing Visit 33
9.74
7.00
7.14
12.34
HIV Care and Treatment Visit 62
10.15
5.00
12.81
Other MCH / OPD Visit
172
8.13
4.50
6.78
9.48
Moving to our staff time estimates, here is a summary of the average time taken per consultation for the services observed. The time taken to deliver an HIV/SRH outpatient visit during endline observations varied widely. On average, the service taking the longest time to deliver was PNC (mean minutes), followed by STI (mean minutes). The distributions of time taken for a number of services were heavily left-skewed, accounting for a minority of complicated cases with longer consultation times.

10. Mean Staff FTE Available
Workload Indicators
Number of Facilities
Mean Annual Visits
Mean Staff FTE Available
Mean Workload Ratio
2009
2011
District Hospital ( n = 5)
PNC 1 2
426
926
0.14
0.18
1.42
0.80
HIV Care & Treatment
3,912
10,161
6.45
11.06
12.81
5.81
VCT 5
2,432
6,006
2.26
4.27
6.49
8.95
Health Centre (n = 17) 11 12
378
498
0.36
0.27
2.80
1.83 8 10
7,343
8,187
4.01
6.84
5.96
12.48 9 6
889
1,625
0.87
1.09
11.88
6.25
Public Health Unit (n = 2)
3,193
2,524
0.74
2.70
1.08
5.04
4,878
457
3.46
3.65
4.58
49.75
1,357
1,036
0.42
2.52
2.10
16.89
SRH Clinic (n = 8) 7
160
187
0.08
4.41
468
809
0.82
1.70
13.58
10.44
1,936
5,308
1.59
3.19
4.92
5.00
Sub District Hospital (n = 6) 3
229
1,162
0.12
0.49
1.77
2.69
341
1,636
3.71
15.51
17.82
590
1,717
1.44
1.11
25.63
8.78
We applied these times to service use observed within each facility. As you can see here, on the whole service utilization increased on the whole from baseline to endline.
Staffing levels also generally increased, however this was more pronounced for HIV-related services than for other services. On average, the greatest increases in staff FTE available were observed for HIV care and treatment (from an average of 3.31 FTE at baseline to 5.99 FTE at endline), and VCT services (from 1.6 to 2.55 FTE).
When evaluating workload by service type, the highest average workload ratio was for VCT (11.22) at baseline and for HIV care (14.60) at endline – indicating high excess capacity for these services across all facility types. PNC on average had the lowest ratio at endline (1.93), and STI on average had the lowest workload ratio at baseline (not shown here due to space restrictions).
Several individual services had workload ratios less than one indicating insufficient capacity; this was most common for PNC (in 10 facilities at baseline, and 12 facilities at endline).

11. Estimated Facility-level Workload Ratio at Baseline and Endline
When we examined overall workload at the facility level, including other outpatient and MCH services, the workload ratio was high, but again with a substantial variation by facility. The mean facility-level ratio was 5.67 at baseline (SE 0.81), and 6.53 at endline (SE 0.97). No facilities had a ratio less than one at baseline or at endline, indicating that at the facility level facilities were adequately staffed and that staff should not be overworked.
So we are seeing some indications of workload imbalances within facilities, where some staff are overworked and others are not.

12. Workload and Staffing, by HR Integration Category ‘More / Less Integrated’
‘ Less Integrated’ Facilities (n = 58)
‘More Integrated’ Facilities (n = 22)
Mean
Staff FTE Available
Staff FTE Required
Mean Workload Ratio (low/high estimates)
Mean Workload Ratio
(low/high estimates)
Ca Cervix Screening
0.06
0.02
(8.58 – 15.25)
0.10
0.03
(5.30 – 9.42) FP
0.91
0.42
(2.43 – 2.93)
0.99
0.54
(1.96 – 2.37)
HIV Care and Treatment
3.21
0.87
( )
0.45
(12.57 – 21.52)
PITC
0.33
0.09
(3.67 – 6.34)
0.22
0.14
2.19* (1.73 – 2.98)
PNC
0.37
0.13
(2.22 – 2.80)
0.17
(1.82 – 2.29)
STI
0.05
(2.19 – 3.45)
1.63** (1.33 – 2.09)
VCT
1.10
0.20
(7.96 – 13.76)
1.66
0.31
(6.57 – 11.35)
Other MCH / OPD Service
5.44
1.09
(7.19 – 10.06)
6.19
1.70
(12.59 – 17.61)
Total Facility
10.94
2.62
(4.62 – 7.15)
9.34
2.87
(5.63 – 8.35)
Finally, we compare workload by integration. We classify facilities cross-sectionally into ‘more integrated’and ‘less integrated’ groups - thirteen facilities at baseline and nine at endline were classified into the ‘more integrated’ group (top 20%). The ‘more integrated’ facilities delivered an average of 2.7 services per staff member, while ‘less integrated’ facilities delivered an average of 1.3 services per staff member.
Comparing ‘more integrated’ against ‘less integrated’ facilities reveals that the workload ratio for several individual services appear lower in the ‘more integrated’ group; this was only significant for PITC (t = 1.79, p = 0.078) and STI (t = 2.05, p= 0.047). However the overall facility-level workload ratio was not significantly different between the two groups, indicating that it is resourcing within facilities that is driving these workload increases.
*difference from ‘less integrated’ group significant at the p < 0.10 level (t = 1.79, p = 0.078)
** difference from ‘less integrated group significant at the p< 0.05 level (t = 2.05, p= 0.047)

13. Implications for policy
Integration was not scaled up uniformly; readiness assessment should precede integration policy
HIV/SRH integration may be most influential on staff workload for PITC, PNC and STI services
Some of these increases, in particular increased staffing of HIV-related services, may have come at the cost of reductions of staff available for other services such as PNC, and lead to greater imbalances in staff workload within a facility
However, policy makers should also be careful about overworking staff and assess integration in the broader context of HR planning
So to summarize:
We find some evidence to suggest that there is potential to improve productivity through integration, however with some significant challenges, and the pace of productivity gain slow.
Our results indicate that integration might have the most impact on staff workload for PITC, PNC and STI services.
However in some cases, we saw increased staffing of HIV related services come at the cost of reductions of staff time available for other services such as PNC, leading to greater imbalances in staff workload within a facility.
We recommend that any efforts to implement integration therefore are fully assessed in the broader context of HR planning both within and between facilities to understand the impact on different staff cadres and to minimise displacement effects in order to ensure that neither staff nor patients are negatively impacted by integration policy.

14. Acknowledgements Ministry of Health, Swaziland
Ministries of Health, Kenya
Family Health Options Kenya (FHOK)
Family Life Association of Swaziland (FLAS)
Learn more at:
Support for this study was provided by the Bill & Melinda Gates Foundation. The views expressed herein are those of the author(s) and do not necessarily reflect the official policy or position of the Bill & Melinda Gates Foundation
Thanks to all of our research partners, and to our funders the Bill and Melinda Gates Foundation, and thanks to you for listening.
For a copy of this presentation please visit same.lshtm.ac.uk