A total of 742 providers were interviewed and 665 women were observed during labor and delivery. Observations included took place at 77 public health facilities, with 9% (61) at the 5 specialized hospitals, 9% (57) at the 5 regional hospitals, 41% (271) at the 27 provincial hospitals and 41% (276) at the 40 district hospitals or comprehensive health centers (CHCs) with five or more births per day. Average caseload across all health facilities sampled was 504 births/month in the year preceding the assessment (see Table 1 for caseload across facility type).
Table 1. Health facility labor and delivery caseload statistics
|
Health facility type
|
n
|
Average caseload (births/month)
|
Caseload range
(births/month)
|
Specialized hospitals
|
5
|
1346
|
988-1659
|
Regional hospitals
|
5
|
1574
|
1167-2157
|
Provincial hospitals*
|
26
|
448
|
76-1186
|
District hospitals and CHCs (5+ births per day)
|
40
|
228
|
142-1233
|
*Caseload data missing for one provincial hospital
|
Quality of care index items were assessed for all 665 labor and delivery observations across the 77 public health facilities (see Table 2 below). Sample sizes vary at different stages of labor as women were admitted to the facility or referred to emergency care at different times. A couple of the QoC index items were widely implemented such that a uterotonic drug was prepared for use in active management of the third stage (AMTSL) of labor for 90% of women, and newborns were immediately dried with a towel in 87% of cases. However, the vast majority of QoC actions were infrequently implemented. Implementation of initial client assessment and examination was low such that providers rarely asked women if they experienced headaches or blurred vision (13%), vaginal bleeding (30%), or took their pulse (31%). Only 33% of women ever received an explanation from a provider of what would happen to them in labor and only 29% of women had their vital signs assessed 15 minutes after birth. There was variation in the level of implementation of QoC index items across health facility type, with regional hospitals generally underperforming.
Table 2. Quality of care index items by health facility type (n=665)
Quality of care index
|
Secondary and Tertiary Hospitals
|
Primary Care Facilities
|
Total average
|
Specialized Hospitals
|
Regional Hospitals
|
Provincial Hospitals
|
District Hospitals and CHCs
|
Initial client assessment and examination
|
n=27
|
n=40
|
n=195
|
n=189
|
n=451
|
1. Asks whether woman has experienced headaches or blurred vision
|
15% (4)
|
2% (1)
|
12% (23)
|
16% (31)
|
13% (59)
|
2. Asks whether woman has experienced vaginal bleeding
|
33% (9)
|
7% (3)
|
29% (56)
|
36% (68)
|
30% (136)
|
3. Takes blood pressure
|
78% (21)
|
30% (12)
|
61% (120)
|
65% (123)
|
61% (276)
|
4. Takes pulse
|
67% (18)
|
27% (11)
|
27% (53)
|
31% (58)
|
31% (140)
|
5. Washes his/her hand before any examination
|
56% (15)
|
42% (17)
|
37% (72)
|
37% (70)
|
39% (174)
|
6. Wears high-level disinfected or sterile gloves for vaginal examination
|
74% (20)
|
47% (19)
|
66% (128)
|
73% (138)
|
67% (305)
|
First stage of labor
|
n=43
|
n=39
|
n=213
|
n=220
|
n=515
|
7. At least once, explains what will happen in labor to the woman and/or her support person
|
35% (15)
|
20% (8)
|
33% (70)
|
35% (77)
|
33% (170)
|
8. Prepares uterotonic drug to use for AMTSL
|
98% (42)
|
79% (31)
|
88% (187)
|
93% (205)
|
90% (465)
|
9. Uses partograph (during labor)
|
74% (32)
|
36% (14)
|
42% (89)
|
60% (133)
|
52% (268)
|
10. Self-inflating ventilation bag (500mL) and face masks (size 0 and size 1) are laid out and ready for use for neonatal resuscitation
|
21% (9)
|
13% (5)
|
27% (58)
|
30% (67)
|
27% (139)
|
Third stage of labor
|
n=61
|
n=57
|
n=271
|
n=273
|
n=662
|
11. Correctly administers uterotonic (timing, dose, route)*
|
52% (32)
|
28% (16)
|
40% (108)
|
37% (100)
|
39% (256)
|
12. Assesses completeness of placenta and membranes
|
56% (34)
|
32% (18)
|
57% (154)
|
49% (135)
|
51% (341)
|
13. Assesses for perineal and vaginal lacerations
|
80% (49)
|
56% (32)
|
75% (203)
|
68% (185)
|
71% (469)
|
Immediate newborn and postpartum care
|
n=61
|
n=56
|
n=263
|
n=272
|
n=652
|
14. Immediately dries baby with towel
|
92% (56)
|
82% (46)
|
87% (230)
|
87% (238)
|
87% (570)
|
15. Places newborn on mother’s abdomen skin-to-skin
|
64% (39)
|
36% (20)
|
47% (124)
|
48% (131)
|
48% (314)
|
16. Ties or clamps cord when pulsations stop, or by 2–3 minutes after birth (not immediately after birth)
|
88% (54)
|
66% (37)
|
70% (183)
|
71% (194)
|
72% (468)
|
17. Takes mother's vital signs 15 minutes after birth
|
25% (15)
|
11% (6)
|
30% (78)
|
34% (92)
|
29% (191)
|
18. Palpates uterus 15 minutes after birth
|
44% (27)
|
20% (11)
|
48% (126)
|
50% (135)
|
46% (299)
|
19. Assists mother to initiate breastfeeding within one hour
|
28% (17)
|
34% (19)
|
34% (89)
|
44% (119)
|
37% (244)
|
* Correct administration of uterotonic included 10 IU of oxytocin intramuscularly within 1 minute of delivery, 600µg of misoprostol orally within 1 minute, or 200µg of ergometrine intramuscularly within 1 minute. Results here are lower than previously published findings from this assessment as Ansari and colleagues examined a subset of these criteria[13].
Quality of care scores were calculated for each labor and delivery observation and averaged across health facility type and group (see Table 3). There was a significant difference across health facility type (p<0.0001) but no significant difference between primary and secondary/tertiary facilities (p=0.115). Of note, no health facility type had an average quality of care score above 56%.
Table 3. Average quality of care score, by health facility type and group (n=665)
|
Number of observations
|
Mean quality of care score (scale 0-1)
|
Standard deviation
|
Secondary and tertiary hospitals
|
389
|
0.50
|
0.20
|
Specialized Hospitals
|
61
|
0.56
|
0.16
|
Regional Hospitals
|
57
|
0.37
|
0.20
|
Provincial Hospitals
|
271
|
0.51
|
0.20
|
District hospital or CHC (5+ births/day)
|
276
|
0.52
|
0.24
|
All facilities
|
665
|
0.51
|
0.22
|
Associations between provider experience at health facilities and quality of care scores were examined (see Table 4). In terms of training, the proportion of providers who received related clinical training in the last three years, including any pre- or in-service labor and delivery training, BEmONC training, and training in newborn resuscitation, was not significantly associated with quality of care scores at the same facility. Only training in essential newborn care was associated, such that as the proportion of providers being trained in essential newborn care increases at a facility, so too does the quality of care score (effect size 7.9% [95% CI=1.1%-14.7%], p=0.023). Proportion of providers trained in various quality improvement endeavors was not significantly associated with quality of care scores however training on respectful maternity care (RMC) was associated. As the proportion of providers trained in RMC at a health facility increases, so too do the quality of care scores at that facility (effect size 9.6% [95% CI=1.0-18.1%], p=0.028). RMC training variables and related index items were not correlated. Similarly, quality of care scores significantly increased with an increasing proportion of providers having received training in gender and human rights (effect size 12.8% [95% CI=2.9-22.7%], p=0.012).
Supervision was another factor explored. There was a significant positive association between the proportion of providers receiving supervision and quality of care scores at health facilities such that facilities with more providers reporting having any or recent supervision also scored higher on the QoC index (effect size 7.6% [95% CI=2.7-12.5%], p=0.002). Facilities with a higher proportion of providers who felt more respected by their supervisors (scale of 1-5, with 5 being most respected, averaged across providers at each facility), tended to have higher quality of care scores (effect size 3.6% [95% CI=0.0-7.3%], p=0.050). Providers who had received supervision were asked about the content of their most recent supervision visit. At facilities where more providers reported that their supervisor had observed their work, quality of care scores were significantly higher (effect size 22.2% [95% CI=14.2-30.2%], p<0.001), however, where more providers reported receiving verbal feedback or discussing problems with supervisors, quality of care scores were significantly lower (see Table 4).
Table 4. Association of maternity provider experience with quality of care scores
|
β
|
SE
|
95% CI
|
p-value
|
Training in last 3 years
|
|
|
|
|
Any pre- or in-service training on labor & delivery
|
-0.0286
|
0.0343
|
(-0.0959 - 0.0387)
|
0.4039
|
BEmONC
|
-0.0376
|
0.0433
|
(-0.1227 - 0.0474)
|
0.3852
|
Essential newborn care
|
0.0791
|
0.0347
|
(0.0109 - 0.1473)
|
0.0231
|
Newborn resuscitation
|
0.0400
|
0.0273
|
(-0.0136 - 0.0935)
|
0.1432
|
Maternal death or near miss review/audit
|
-0.0353
|
0.0540
|
(-0.1413 - 0.0706)
|
0.5129
|
Quality improvement approaches
|
0.0448
|
0.0558
|
(-0.0648 - 0.1545)
|
0.4225
|
HMIS data quality and use
|
0.1110
|
0.0584
|
(-0.0036 - 0.2256)
|
0.0576
|
Respectful maternity care
|
0.0957
|
0.0435
|
(0.0102 - 0.1812)
|
0.0283
|
Gender and human rights
|
0.1278
|
0.0505
|
(0.0286 - 0.2271)
|
0.0116
|
Supervision
|
|
|
|
|
Recent supervision of providers*
|
0.0761
|
0.0249
|
(0.0273 - 0.1249)
|
0.0023
|
Level of respect providers feel from supervisors
|
0.0363
|
0.0185
|
(0.0000 - 0.0727)
|
0.0503
|
During last supervision:**
|
|
|
|
|
Supervisor observed provider’s work
|
0.2222
|
0.0406
|
(0.1425 - 0.3019)
|
0.0000
|
Supervisor gave verbal feedback about work
|
-0.0661
|
0.0309
|
(-0.1269 - -0.0053)
|
0.0331
|
Supervisor discussed problems provider encountered
|
-0.0687
|
0.0316
|
(-0.1306 - -0.0067)
|
0.0299
|
*Categorical variable (never supervised; supervised in last 3 months; supervised more than 3 months ago) averaged across providers interviewed at each facility
**If provider did not report receiving supervision, the activity was categorized as having not occurred
Associations between health facility work environment characteristics and quality of care were also assessed (see Table 5). Availability of routine labor and delivery supplies (p<0.0001) and the ability to sterilize necessary equipment (p<0.0001) were both significantly positively related to quality of care. For each additional element of supplies available, quality of care scores at the same health facility increased by 4.1% (2.9-5.4%). Facilities that had any type of functioning sterilizer available experienced 13.2% (7.5-18.9%) higher quality of care scores than those without. There was a significant negative relationship between facility management of obstetric and newborn complications, as indicated through performance of BEmONC signal functions, and quality of care scores for routine delivery at those facilities (effect size -3.3% [95% CI=-0.7—6.0%], p=0.013). Of interest is the link between various measures of gender equality among providers and the quality of care extended to clients. Providers (all female, per Afghan custom) were asked if they believed they had equal treatment and opportunities as colleagues of the opposite sex across several domains. Facilities where a greater proportion of maternity care providers believed they had gender equality in terms of training (effect size 12.9% [95% CI=7.0-18.7%], p<0.001), time off (effect size 12.1% [95% CI=6.3-17.9%], p<0.001), and work schedule (effect size 10.6% [95% CI=4.3-17.0%], p=0.001), had significantly higher quality of care scores.
Table 5. Association between work environment characteristics and quality of care scores
|
β
|
SE
|
95% CI
|
p-value
|
No. BEmONC signal functions conducted
|
-0.0334
|
0.0135
|
(-0.0598 - -0.0069)
|
0.0134
|
No. routine labor & delivery supplies available
|
0.0415
|
0.0065
|
(0.0286-0.0543)
|
0.0000
|
Any functioning sterilizer
|
0.1320
|
0.0292
|
(0.0747-0.1892)
|
0.0000
|
Daily caseload of deliveries
|
-0.0009
|
0.0005
|
(-0.0020 - 0.0001)
|
0.0650
|
Has a written job description for position
|
0.1909
|
0.0722
|
(0.0492 - 0.3327)
|
0.0084
|
Believes has equal treatment and opportunities as colleagues of the opposite sex in terms of:
|
|
|
|
|
Training
|
0.1288
|
0.0297
|
(0.0704 - 0.1871)
|
0.0000
|
Time off
|
0.1208
|
0.0296
|
(0.0626 - 0.1790)
|
0.0001
|
Work schedule
|
0.1061
|
0.0323
|
(0.0427 - 0.1696)
|
0.0011
|
Workload
|
0.0392
|
0.0298
|
(-0.0192 - 0.0976)
|
0.1880
|
The multivariate linear regression model (Table 6) significantly predicts about 27% of the variability in quality of care scores (F<0.001, R2=0.268). Holding all other variables constant, health facility type is significantly associated with quality of care (p<0.001) such that provincial and regional hospitals perform worse than district hospitals and CHCs while specialized hospitals perform better. A broad spectrum of variables representing provider and work environment characteristics also significantly predict quality of care scores. In particular, having a greater proportion of providers trained in respectful maternity care had a large significant positive effect on quality of care (19.5% [8.3-30.6%]). Feeling there is greater gender equality among providers in terms of training, having more necessary supplies available, having more recent supervision, increased proportion of supervisors observing providers’ work, and having a written job description all had significant positive effects on quality of care scores. Holding all other variables constant, an increasing proportion of providers receiving any training in labor and delivery in the last three years was significantly negatively associated with quality of care (-13.4% [-4.1- -22.7%]). As with the bivariate associations above, increased proportion of supervisors providing verbal feedback continues to have a significant negative effect on quality of care scores.
Table 6. Multivariate linear regression model of quality of care (n=622)
|
β
|
SE
|
95% CI
|
p-value
|
Health facility type (Ref. District hospital or CHC 5+)
|
|
|
|
0.000
|
Provincial hospital
|
-0.036
|
0.018
|
(-0.070 - -0.001)
|
0.042
|
Regional hospital
|
-0.201
|
0.032
|
(-0.264 - -0.137)
|
0.000
|
Specialized hospital
|
0.065
|
0.029
|
(0.008 - 0.123)
|
0.026
|
Any pre- or in-service training on labor and delivery in last 3 years
|
-0.134
|
0.047
|
(-0.227 - -0.041)
|
0.005
|
Respectful maternity care training in last 3 years
|
0.195
|
0.057
|
(0.083 - 0.306)
|
0.001
|
Believes has equal treatment and opportunities as colleagues of the opposite sex in terms of training
|
0.100
|
0.030
|
(0.041 - 0.158)
|
0.001
|
No. BEmONC signal functions
|
-0.027
|
0.014
|
(-0.055 - 0.001)
|
0.060
|
No. routine labor & delivery supplies available
|
0.041
|
0.007
|
(0.028 - 0.054)
|
0.000
|
Recent supervision of providers
|
0.081
|
0.026
|
(0.031 - 0.132)
|
0.002
|
Last time supervised, supervisor gave verbal feedback about work
|
-0.164
|
0.032
|
(-0.227 - -0.102)
|
0.000
|
Last time supervised, supervisor observed provider’s work
|
0.227
|
0.039
|
(0.150 - 0.305)
|
0.000
|
Has a written job description for position
|
0.169
|
0.072
|
(0.028 - 0.309)
|
0.019
|
*Sample size reflects birth observations that had complete data for all variables included in the model