Abstract

Objective

To examine the relationship between workforce capacity and quality of pediatric care in outpatient clinics in Afghanistan.

Design

Annual national performance assessments were conducted between 2005 and 2008 to determine quality of care through patient observations in >600 health facilities, selected by stratified random sampling each year. Other variables measured were health provider capacity, competency and adequacy of support systems.

Setting

Primary care facilities in 29 provinces in Afghanistan.

Participants

Pediatric patients and their caretakers greater than 2400 were selected at random each year.

Main outcome measures

Index of observed quality of care for patient assessment and counseling based on WHO's Integrated Management of Childhood Illness (IMCI) clinical guidelines.

Results

Quality of care improved for all IMCI indices between 2005 and 2008 (IMCI index increased from 43.1 to 56.1; P < 0.001) and was significantly associated with the availability of doctors, IMCI training and knowledge and factors such as provider job satisfaction, availability of clinical guidelines, frequency of supervision and the presence of community councils. There was also a progressive increase in the index summarizing staffing capacity during the study period. Basic health centers increased from 75.6 to 85.5% (P < 0.001), comprehensive health centers increased from 27.9 to 37.9% (P < 0.03) and district hospitals increased from 34.1 to 37.2% (P > 0.05).

Conclusions

Enhancing workforce capacity and competency and ensuring appropriate supervision and systems support mechanisms can contribute to improved quality of care. Although the results indicate sustained improvements over the study period, further research on the mixture of provider skills, competency and factors influencing provider motivation are essential to determine the optimal workforce capacity in Afghanistan.

Introduction

Despite the advances in medical technology and innovation, critical deficits in health workforce pose key constraints in service delivery and compromise quality of care and health outcomes, particularly in fragile health systems [1–3]. The global shortage of doctors, nurses and midwives estimated at 2.4 million escalates to 4.3 million when other allied health workers are included [3]. The movement of health providers from rural to urban settings, public to private sectors and migration to wealthier countries create additional distortions and inequities in health service delivery, especially in low-income countries with high disease burden. Variability in staffing and excessive workloads have been shown to compromise patient safety and quality of care and lead to poor outcomes, but there are limited data from low-income settings [3–7].

WHO and its member countries have launched several strategies to explore mechanisms to retain the health workforce particularly in extreme resource-constrained environments, but there is still a paucity of information on the specific operational systems for implementation in specific country contexts to address the challenge [8]. As in other countries emerging from conflict, in Afghanistan the human resource crisis is characterized by poor working conditions, including minimal financial compensation, inadequate staffing, lack of career development opportunities or other incentives and worsening security and further exacerbated by chronic inadequacies in both public infrastructure and lack of training capacities, resulting in severe deficits in human resources. Despite the enormous investments to revamp the health infrastructure, capacity and service availability [9–12], disturbing trends in workforce migration threaten the gains achieved [13]. According to recent estimates produced by the Global Health Alliance, the Afghanistan health workforce of 1.08 workers per 1000 population is far less than the 2.5 postulated as needed to achieve the Millennium Development Goals (MDGs) [3]. Another major impediment to improving quality and coverage of service delivery is a lack of female providers, who are in high demand, particularly for child and reproductive healthcare.

The Basic Package of Health Services (BPHS) was introduced in Afghanistan in 2004, prioritizing maternal and child health, as the country ranked among the highest for child and maternal mortality, although the recent adjusted estimates indicate substantial improvements [14, 15]. The Integrated Management of Childhood Illness (IMCI) case management algorithm was incorporated as a key policy strategy in the BPHS, as it is a cost-effective measure for addressing the major disease burden in children if implemented effectively [16–20].

Results from the National Health Service Performance Assessments (NHSPA) conducted annually between 2004 and 2008 demonstrated progressive improvements in all six performance domains: patient and community perspectives, staff perspectives, service provision, capacity for service provision, financial systems and overall vision measured by the national balanced scorecard performance system [12]. During this period, efforts were accelerated to train and deploy providers, specifically for IMCI. In studies examining specific determinants of IMCI quality of care, we found improved trends in adherence to IMCI standards in primary health-care facilities [21]. However, hospitals had a lower level of performance [22]. Previous studies have not examined the role of changes in the adequacy of providers on the quality of care. The purpose of this study is to examine how changes in sufficiency of health providers (capacity) at health facilities, and health provider training and knowledge in IMCI (competency) have affected the quality of IMCI care. We also wanted to assess whether key individual provider factors such as gender, cadre and motivation influenced the quality of care, along with other managerial factors, including supervision, the presence of guidelines and community governance.

Methods

The sample for this study was derived from the annual NHSPA evaluations between 2005 and 2008. Farah, Helmand, Zabul, Kandahar and Uruzgan provinces were excluded because of inaccessibility during some of the assessment years, and 29 provinces were included in the analysis. Data from 2004 were not included, as some key variables were not included.

Regional training of survey teams comprising clinicians, nurses and vaccinators was conducted annually for the NHSPA. Up to 25 facilities were selected from each province, employing stratified sampling to include 3 district hospitals (DHs), 7 comprehensive health centers (CHCs) and 15 basic health centers (BHCs). The evaluations comprised case management observations of sick children under five years attending outpatient clinics, followed by exit interviews with their caretakers and interviews of providers. Five children were selected at random from each facility by systematic sampling using a random starting point and a sampling interval based on the reported daily average number of new patients, resulting in a sample of >2400 patients. Subsequently, four providers responsible for clinical management of patients were randomly selected for interviews.

Consistent with the previous research [21], we created an IMCI quality index as a measure of provider performance based on the observation of a comprehensive set of indicators of diagnostic and counseling standards used for the management of ill children in an outpatient setting [23]. The IMCI quality index represents the average provider compliance with specific IMCI standards computed at the facility level for all observed child consultations. In this study, we also included the caretaker's report of weight assessment in the IMCI quality index, as chronic malnutrition in children is of critical public health importance in Afghanistan, and considerable effort has been invested in nutrition programs [24, 25].

Provider competence was measured by the completion of refresher training, IMCI training and practical knowledge of IMCI and selected disease conditions included in the BPHS. The workforce capacity was determined by the degree to which standards for staffing were met at the different levels of primary care facilities providing the BPHS: BHCs, CHCs and DHs. Other facility-level factors considered include the presence of IMCI guidelines, frequency of supervision and the presence of community councils to oversee health facilities. Individual-level factors assessed include the cadre and sex of the health provider, patient's and caretaker's sex and measures of health provider motivation, including factors related to compensation. Previously, the health system included a cadre of assistant doctors (known as Rogh-Tiya-Paal), who received a 4-year training. Although the training has been discontinued, some health facilities still retain these providers.

The response rate of interviewed caretakers and providers was over 98%. A detailed description of the survey instruments and field survey methods is provided elsewhere [21, 26].

Data processing and analysis

We first explored the descriptive profile of the sampled patients, caretakers and providers using univariate analysis and dropped missing values (<5%) as they were confirmed to be missing at random. Applying bivariate analysis with robust standard errors to account for clustering within provinces, we compared various groups of independent variables between 2005 and 2008. Subsequently, we constructed multiple linear regression models using ordinary least squares to compare the IMCI quality index between the groups.

Huber–White robust estimates of standard error were used to account for clustering at the facility level. The analysis was performed using STATA 10.0 (Stata Corp). Post-estimation procedures in STATA were utilized to assess the normality in distribution of residuals by constructing distribution plots, and multicollinearity was assessed by estimating the variance inflation factor [27].

Results

Table 1 describes the characteristics of the sample of patients and providers selected in the study. Approximately one-third of the patients were <1 year of age and the major presenting complaints were IMCI conditions; diarrhea, fever and cough or difficulty in breathing. Over 70% of the children were accompanied by a female caretaker.

Table 1

Descriptive characteristics of patients and caregivers

Characteristics Year of assessment
 
 2005 %
 
2008 % 
Child's age (months) n = 2485 n = 2780 
 2–<12 31.5 29.2 
 12–59 68.5 70.8 
Child's sex n = 2477 n = 2761 
 Male 53.9 52.2 
 Female 46.1 47.8 
Major presenting symptoms n = 2485 n = 2780 
 Diarrhea 48.7 42.9 
 Fever 18.1 15.6 
 Cough/difficult breathing 15.3 20.4 
 Othera 17.9 21.2 
Caretaker of sick child n = 2456 n = 2777 
 Mother 66.8 70.2 
 Father 23.2 18.9 
 Other female caretaker 4.8 5.8 
 Other male caretaker 5.1 5.1 
Characteristics Year of assessment
 
 2005 %
 
2008 % 
Child's age (months) n = 2485 n = 2780 
 2–<12 31.5 29.2 
 12–59 68.5 70.8 
Child's sex n = 2477 n = 2761 
 Male 53.9 52.2 
 Female 46.1 47.8 
Major presenting symptoms n = 2485 n = 2780 
 Diarrhea 48.7 42.9 
 Fever 18.1 15.6 
 Cough/difficult breathing 15.3 20.4 
 Othera 17.9 21.2 
Caretaker of sick child n = 2456 n = 2777 
 Mother 66.8 70.2 
 Father 23.2 18.9 
 Other female caretaker 4.8 5.8 
 Other male caretaker 5.1 5.1 

aSkin infection, pus, injury, ear pain, jaundice, etc.

The overall trends on workforce capacity and IMCI quality of care are illustrated in the tables available in the supplemental files. The tables included in this work are primarily to illustrate changes between years 2005 and 2008, except for Table 4 that includes data for all study years to determine predictors of quality.

An analysis of the adequacy of health providers (Table 2 and Supplementary File Table S2a) indicated a progressive increase in workforce numbers over the study period. During this period, the facilities meeting staffing standards rose significantly at most types of facilities: the percentage of BHCs meeting staffing standards rose from 75.6 to 85.5% (P < 0.001), among CHCs from 27.9 to 37.9% (P < 0.03) and among DHs from 34.1 to 37.2% (P > 0.05.). However, by 2008, 30% of the sampled facilities still did not meet the staffing standards. A subanalysis by cadre indicated a disturbing decline in availability of doctors and assistant doctors from 36.8 to 25.4% by 2008. Although efforts have been made to address the increased demand for female health providers, the BHCs, which are mostly located in rural communities, have a disproportionate ratio of male staff (70%) in comparison to DHs (54%).

Table 2

Provider profile, capacity, competence, supervision and job satisfaction

Provider profile Year of assessment
 
Percent change from 2005 to 2008 
2005
 
2008
 
 n n  
 Provider cadre (interviewed)a 1438  2233   
  Doctor  35.9  24.5 NA 
  Assistant doctor  7.1  1.2  
  Nurse  32.3  17.0  
  Othersb  24.7  57.2  
 Provider cadre (observed)c 2485  2780   
  Doctor  68.0  73.5 NA 
  Assistant doctor  10.9  3.2  
  Nurse  17.7  19.8  
  Othersb  3.4  3.5  
 Provider gender (interviewed)a 1407  2169  NA 
  Males  65.8  64.8  
  Females  34.2  35.2  
 Provider gender (observed)c 2481  2764  NA 
  Male  92.0  95.8  
  Female  8.0  4.2  
Provider adequacy 
 Provider standardsd all facilities 589 56.2 612 67.3 11.1*** 
  BHC (≥2) 344 75.6 379 85.5 9.9** 
  CHC (≥6) 204 27.9 190 37.9 10.0* 
  DH (≥21) 41 34.1 43 37.2 3.1 
 Adequate doctorse 589 36.8 612 25.4 −11.4*** 
 Adequate nurse/midwifef 589 35.1 612 42.1 7.0* 
Training, knowledge, supervision and job satisfaction 
 Facilities with ≥1 providers trained in IMCI 589 49.7 612 64.5 14.8*** 
  BHC 344 43.3 379 61.5 18.2*** 
  CHC 204 60.3 190 70.5 10.2* 
  DH 41 51.2 43 65.1 13.9 
 Providers reporting IMCI training 1433 30.6 2223 24.7 −5.9*** 
  Doctors 515 43.5 548 55.5 12*** 
  Assistant doctors 102 33.3 27 40.7 7.4 
  Nurses 463 24.0 378 35.4 11.4*** 
  Othersb 353 19.5 1270 7.9 −11.6*** 
 Refresher training for providersg 589 87.9 NA NA NA 
  BHC 344 84.9 NA NA NA 
  CHC 204 92.2 NA NA NA 
  DH 41 92.7 NA NA NA 
Knowledge (high)h 1379 86.0 2090 93.6 NA 
 ≥6 external facility supervision visits in past 6 months 535 77.4 574 56.4 −20.9*** 
 ≥1 provider supervised in last past 6 months 1438 96.7 2231 97.4 0.7 
 Provider job satisfaction (high)h 1374 83.6 1763 88.5 4.9*** 
Provider profile Year of assessment
 
Percent change from 2005 to 2008 
2005
 
2008
 
 n n  
 Provider cadre (interviewed)a 1438  2233   
  Doctor  35.9  24.5 NA 
  Assistant doctor  7.1  1.2  
  Nurse  32.3  17.0  
  Othersb  24.7  57.2  
 Provider cadre (observed)c 2485  2780   
  Doctor  68.0  73.5 NA 
  Assistant doctor  10.9  3.2  
  Nurse  17.7  19.8  
  Othersb  3.4  3.5  
 Provider gender (interviewed)a 1407  2169  NA 
  Males  65.8  64.8  
  Females  34.2  35.2  
 Provider gender (observed)c 2481  2764  NA 
  Male  92.0  95.8  
  Female  8.0  4.2  
Provider adequacy 
 Provider standardsd all facilities 589 56.2 612 67.3 11.1*** 
  BHC (≥2) 344 75.6 379 85.5 9.9** 
  CHC (≥6) 204 27.9 190 37.9 10.0* 
  DH (≥21) 41 34.1 43 37.2 3.1 
 Adequate doctorse 589 36.8 612 25.4 −11.4*** 
 Adequate nurse/midwifef 589 35.1 612 42.1 7.0* 
Training, knowledge, supervision and job satisfaction 
 Facilities with ≥1 providers trained in IMCI 589 49.7 612 64.5 14.8*** 
  BHC 344 43.3 379 61.5 18.2*** 
  CHC 204 60.3 190 70.5 10.2* 
  DH 41 51.2 43 65.1 13.9 
 Providers reporting IMCI training 1433 30.6 2223 24.7 −5.9*** 
  Doctors 515 43.5 548 55.5 12*** 
  Assistant doctors 102 33.3 27 40.7 7.4 
  Nurses 463 24.0 378 35.4 11.4*** 
  Othersb 353 19.5 1270 7.9 −11.6*** 
 Refresher training for providersg 589 87.9 NA NA NA 
  BHC 344 84.9 NA NA NA 
  CHC 204 92.2 NA NA NA 
  DH 41 92.7 NA NA NA 
Knowledge (high)h 1379 86.0 2090 93.6 NA 
 ≥6 external facility supervision visits in past 6 months 535 77.4 574 56.4 −20.9*** 
 ≥1 provider supervised in last past 6 months 1438 96.7 2231 97.4 0.7 
 Provider job satisfaction (high)h 1374 83.6 1763 88.5 4.9*** 

Trend data for years 2005–2008 are available in the online supplemental files.

aFacility-level data: interviews with a sample of providers from each facility.

bMidwife, auxiliary midwife, pharmacists, technologists, community health workers supervisor and vaccinators.

cData from case management observations of provider and patients.

dClinical provider adequacy (standards from BPHS 2005: doctors, assistant doctors, nurses and midwives).

eStaffing standard for doctors/assistant doctors (BHC ≥1 (if Nurse or Midwife is absent), CHC ≥2, DH ≥7).

fStaffing standard for nurses or midwives: (BHC ≥2, CHC ≥4 and DH ≥14).

gRefresher training components were not queried in 2008.

hKnowledge and job satisfaction scores: Low = below mean score for 2005; high = mean and above mean score for 2005.

Note: For subsequent years, the cut-offs created for 2005 were used to classify as low and high. Included knowledge of IMCI, expanded program for immunization and reproductive health; difference of difference analysis was not computed as the tests were different in 2008.

*P < 0.05.

**P < 0.001.

***P < 0.0001.

Provider knowledge scores remained at high levels over the study period, but the proportion of providers reporting IMCI and refresher training increased significantly, especially at BHCs (Table 2). However, a disproportionate number of doctors and assistant doctors are targeted for IMCI training when compared with the targeting of mid-level cadres. Other individual factors of quality of care shown in Table 2 indicated that health provider job satisfaction had small but statistically significant improvements. Facility-level factors also indicate that providers reporting at least one supervision visit in the previous 6 months remained at high levels (>90%), but the frequency of external supervision visits diminished significantly (by 20%), over the study period.

The quality of patient assessment and counseling improved significantly for all IMCI index indicators between 2005 and 2008 (P < 0.01), except for the assessment of diarrheal symptoms, which had high levels of compliance in 2005 (Table 3). By 2008, only 32% of children were examined for stridor or wheezing, although measurement of the respiratory rate improved significantly. Improvements were evident in counseling on disease cause, but less than a quarter of the caretakers were informed about potential adverse reactions. Based on caretaker's report, we found a significant improvement from one-fifth to more than half of the children having a weight assessment, indicating a greater emphasis on screening for acute malnutrition at the end of the study period.

Table 3

IMCI index and adherence to IMCI quality of care indicators

 Year or assessment n (%)
 
Percent change from 2005 to 2008 
2005
 
2008
 
n Mean (SD) n Mean (SD) 
IMCI index 2485 43.1 (18.2) 2780 56.1 (24.4) 13.0*** 
 n n  
Provider asked/checked 
 Danger signs      
  Ability to drink or breastfeed 2482 56.6 2775 63.9 7.3*** 
  Vomits everything 2476 60.3 2772 68.2 7.9*** 
 Lethargic/unconsciousness 2472 27.7 2776 38.8 11.1*** 
  Convulsions 2470 12.6 2763 36.6 24.0*** 
Presenting complaints      
 Diarrhea 2485 89.3 2779 89.2 −0.1 
 If diarrhea present 2485 58.2 2779 52.7 −5.5*** 
  Durationa 1447 97.3 1454 95.3 −2.0** 
  Blood in stoola 1447 75.5 1454 77.1 1.6 
  Skin turgora 1447 43.2 1454 58.5 15.3*** 
 Cough/difficult breathing 2485 77.9 2778 80.8 2.9** 
 If cough/difficult breathing present 2485 34.3 2778 39.2 4.9*** 
  Durationb 851 91.0 1084 82.6 −8.4*** 
  Stridor/wheezingb 851 32.3 1084 32.4 0.1 
  Respiratory rateb 851 31.4 1084 53.6 22.2*** 
  Lifted shirtb 851 70.6 1084 77.2 6.6** 
  Used stethoscopeb 851 66.6 1084 72.0 5.4* 
  Fever in the past 24 h 2481 68.3 2780 74.8 6.5*** 
Checked 
 Palm for anemia 2485 11.5 2780 25.9 14.4*** 
 Feet/ankles for edema 2479 11.7 2779 23.6 11.9*** 
 Immunization card 2484 20.9 2771 48.8 27.9*** 
Counseled caretaker on 
 Disease, cause and course 2480 36.1 2773 55.0 18.9*** 
  Home care 2481 73.1 2776 79.5 6.4*** 
  Administer medicines 2461 79.3 2757 85.9 6.6*** 
  Adverse reactions 2448 13.9 2759 21.4 7.5*** 
  Signs for immediate return 2484 31.9 2777 50.9 19.0*** 
  Child weighed (caretaker report) 2483 18.6 2775 54.0 35.4*** 
 Year or assessment n (%)
 
Percent change from 2005 to 2008 
2005
 
2008
 
n Mean (SD) n Mean (SD) 
IMCI index 2485 43.1 (18.2) 2780 56.1 (24.4) 13.0*** 
 n n  
Provider asked/checked 
 Danger signs      
  Ability to drink or breastfeed 2482 56.6 2775 63.9 7.3*** 
  Vomits everything 2476 60.3 2772 68.2 7.9*** 
 Lethargic/unconsciousness 2472 27.7 2776 38.8 11.1*** 
  Convulsions 2470 12.6 2763 36.6 24.0*** 
Presenting complaints      
 Diarrhea 2485 89.3 2779 89.2 −0.1 
 If diarrhea present 2485 58.2 2779 52.7 −5.5*** 
  Durationa 1447 97.3 1454 95.3 −2.0** 
  Blood in stoola 1447 75.5 1454 77.1 1.6 
  Skin turgora 1447 43.2 1454 58.5 15.3*** 
 Cough/difficult breathing 2485 77.9 2778 80.8 2.9** 
 If cough/difficult breathing present 2485 34.3 2778 39.2 4.9*** 
  Durationb 851 91.0 1084 82.6 −8.4*** 
  Stridor/wheezingb 851 32.3 1084 32.4 0.1 
  Respiratory rateb 851 31.4 1084 53.6 22.2*** 
  Lifted shirtb 851 70.6 1084 77.2 6.6** 
  Used stethoscopeb 851 66.6 1084 72.0 5.4* 
  Fever in the past 24 h 2481 68.3 2780 74.8 6.5*** 
Checked 
 Palm for anemia 2485 11.5 2780 25.9 14.4*** 
 Feet/ankles for edema 2479 11.7 2779 23.6 11.9*** 
 Immunization card 2484 20.9 2771 48.8 27.9*** 
Counseled caretaker on 
 Disease, cause and course 2480 36.1 2773 55.0 18.9*** 
  Home care 2481 73.1 2776 79.5 6.4*** 
  Administer medicines 2461 79.3 2757 85.9 6.6*** 
  Adverse reactions 2448 13.9 2759 21.4 7.5*** 
  Signs for immediate return 2484 31.9 2777 50.9 19.0*** 
  Child weighed (caretaker report) 2483 18.6 2775 54.0 35.4*** 

IMCI Index (10 indicators): four danger signs, assessment of cough, diarrhea and fever, weight assessment, immunization and pallor checked and four counseling indicators.

aCases with diarrhea.

bCases with cough.

*P < 0.01.

**P < 0.001.

***P < 0.0001.

A multivariate analysis with all variables that showed a significant association in the bivariate analysis, including year of assessment (Appendices 1 and 2; available online), showed that except for facility type (where providers in CHC illustrated better quality of care), all other variables, including year of assessment, high provider knowledge, training and satisfaction, availability of doctors or assistant doctors, IMCI training, availability of IMCI clinical guidelines, supervision visits, facilities managed by a contracting-in mechanism, functional community councils, patients aged <24 months, those accompanied by a female caretaker and case management performed by doctors or assistant doctors (Table 4) were significant predictors of quality of care. The models only explained 14 and 19% of the variability as illustrated by the R2, raising the possibility that other system level factors, patient co-morbidity, illness severity and the exclusion of other non-IMCI disease conditions that were not included in this model may contribute to some of the variability.

Table 4

Determinants of IMCI quality employing multivariate analysis

IMCI quality indexa 
Patient–provider characteristics 
Overall model fit: Number of observations = 10 075, F(6, 10 068) = 274.74, Prob > F = 0.0000, R2 = 0.1377b, Root MSE = 19.925 
 Coefficient SE P-value 
 Child's age (in months)    
  <12    
  12–59 −3.44 0.43 < 0.001 
 Caretaker's sex    
  Male    
  Female 4.91 0.47 < 0.001 
 Provider cadre    
  Doctors/assistant doctors    
  Other providers −7.57 0.48 < 0.001 
 Facility type    
  BHC    
  CHC 1.60 0.44 < 0.001 
  DH −0.20 0.83 0.807 
 Consultation time (months)    
  <10    
  ≥10 16.19 0.51 < 0.001 
Facility characteristics    
Overall model fit: Number of observations = 2266, F(13, 2252) = 42.92, Prob >F = <0.0001, R2 = 0.1915b, Root MSE = 17.376 
 Year 2005    
  2006 4.85 0.97 < 0.001 
  2007 7.32 1.04 < 0.001 
  2008 11.63 1.28 < 0.001 
 Refresher training during previous year    
  No providers    
  Some or all providers 4.24 1.15 < 0.001 
 IMCI training    
  None trained    
  All or some trained 4.99 0.80 < 0.001 
 Provider's knowledge    
  Low    
  High 3.13 1.60 0.05 
 Provider's satisfaction    
  Low    
  High 4.71 1.27 < 0.001 
 Availability of doctors    
  Not adequate    
  Adequate 2.36 0.86 < 0.001 
 IMCI guidelines    
  Absent    
  Present 5.42 0.86 < 0.001 
 External facility supervision visit in past 6 months    
  <6 visits    
  ≥6 visits 2.02 0.94 < 0.05 
 Contracting type    
  Contracting in    
  Contracting out −7.61 1.60 < 0.001 
  Other −8.62 1.75 < 0.001 
 Community councils    
  Not functional    
  Active 2.68 0.97 < 0.05 
IMCI quality indexa 
Patient–provider characteristics 
Overall model fit: Number of observations = 10 075, F(6, 10 068) = 274.74, Prob > F = 0.0000, R2 = 0.1377b, Root MSE = 19.925 
 Coefficient SE P-value 
 Child's age (in months)    
  <12    
  12–59 −3.44 0.43 < 0.001 
 Caretaker's sex    
  Male    
  Female 4.91 0.47 < 0.001 
 Provider cadre    
  Doctors/assistant doctors    
  Other providers −7.57 0.48 < 0.001 
 Facility type    
  BHC    
  CHC 1.60 0.44 < 0.001 
  DH −0.20 0.83 0.807 
 Consultation time (months)    
  <10    
  ≥10 16.19 0.51 < 0.001 
Facility characteristics    
Overall model fit: Number of observations = 2266, F(13, 2252) = 42.92, Prob >F = <0.0001, R2 = 0.1915b, Root MSE = 17.376 
 Year 2005    
  2006 4.85 0.97 < 0.001 
  2007 7.32 1.04 < 0.001 
  2008 11.63 1.28 < 0.001 
 Refresher training during previous year    
  No providers    
  Some or all providers 4.24 1.15 < 0.001 
 IMCI training    
  None trained    
  All or some trained 4.99 0.80 < 0.001 
 Provider's knowledge    
  Low    
  High 3.13 1.60 0.05 
 Provider's satisfaction    
  Low    
  High 4.71 1.27 < 0.001 
 Availability of doctors    
  Not adequate    
  Adequate 2.36 0.86 < 0.001 
 IMCI guidelines    
  Absent    
  Present 5.42 0.86 < 0.001 
 External facility supervision visit in past 6 months    
  <6 visits    
  ≥6 visits 2.02 0.94 < 0.05 
 Contracting type    
  Contracting in    
  Contracting out −7.61 1.60 < 0.001 
  Other −8.62 1.75 < 0.001 
 Community councils    
  Not functional    
  Active 2.68 0.97 < 0.05 

aQuality of care index was computed for all years from 2005 to 2008 as the dependent variable.

bUnadjusted R2.

Discussion

Investments for enhancing the capacity and quality of health service delivery in Afghanistan have clearly resulted in improved performance trends for health service delivery since the inception of the BPHS in 2004 [12]. However, severe deficits in professional workforce, particularly in remote and insecure areas, pose enormous challenges to sustain the gains achieved in service quality and coverage [13]. Aside from ensuring adequacy of the workforce, appropriate investments are required to improve opportunities for professional development and capacity building to ensure the competency, motivation and retention of health providers delivering care in insecure and complex health-care environments.

Improvements in quality of care

Efforts by the ministry and its partners to augment IMCI performance were evident in the increased proportion of providers reporting IMCI training and improved adherence to standard assessment and counseling protocols. Additional quality improvement efforts are still needed to enhance counseling, especially for potential adverse reactions of medications and signs for immediate return to ensure compliance to treatment. The presence of IMCI guidelines was another significant predictor, demonstrating the importance of distributing reference job aids to ensure compliance to standards. The improved trend in screening children for severe malnutrition is also encouraging, although this was based on caretaker's report. These findings strengthen the rationale for continued investments in IMCI, to effectively manage common pediatric illnesses contributing to major disease burden to achieve the targets for the MDGs. Other predictors of quality in the multivariate modeling were consistent with previous studies in Afghanistan [21, 22] and other IMCI research reporting better quality of care by female providers, younger children and those escorted by female family members [17, 18, 28]. However, few of the previous studies reported on the effect of workforce capacity and quality of care. Research in hospital settings suggest that adequate staffing, especially nurses, is a critical prerequisite to quality of care and patient outcomes [4–7]. In Afghanistan's primary health-care facilities, the adequacy of doctors or assistant doctors was significantly associated with better care for children with IMCI conditions.

Aside from the potential for Hawthorne's bias in program evaluations as a limitation of this study, accuracy of diagnosis and treatment was not obtained using a gold standard reassessment. Health facilities that are inaccessible to the investigators may experience larger deficits in health workforce. The necessary elimination of some of the insecure provinces from the study sample represents another limitation of the study. It seems likely that facilities in these unstable provinces are likely to have poorer infrastructure, lower manpower and access to essential supplies and equipment. Subsequently, the NHSPA attempted to engage trained teachers from the insecure provinces to perform the facility audits.

Addressing the workforce challenge

Shortfalls in pre-service training

Estimates obtained from the Ministry of Public Health indicate a severe shortfall in investments for pre-service training, and this may be further compounded by other phenomena substantiated in some developing countries of internal and external migration and internal maldistribution with a higher density of providers in the urban sector. The demands for health professionals are currently met by 8 medical schools, 8 nursing schools, 8 midwifery schools, 28 community midwifery schools and 6 community nursing education schools that offer pre-service training. The National Health Workforce Plan (2012–2016) proposes to deploy additionally 7000 nurses, 6000 midwives and 20 000 community health workers to address the current deficits and augment the requirements for achieving the 90% coverage goal for the BPHS.

Workforce management

In our study, high provider satisfaction was significantly associated with better adherence to IMCI standards of care, although overall high levels of provider satisfaction in this study may have resulted from inherent factors of a culture of low expectations, courtesy bias and obvious improvements in the capacity of the health system. Several motivational themes have been identified from systematic reviews for supporting performance, especially in underserved communities; financial incentives, professional development opportunities, infrastructure, resource availability, health system management, including regulatory mechanisms and recognition [28, 29]. In a recent study in India, authors demonstrated that satisfaction and motivation varied based on the practice setting, governance factors and other non-monetary variables [30]. In a qualitative study in Kenya and Benin, authors reported that professional conscience and ethos to ensure services and self-esteem were equally strong motivators as the financial drivers of performance [31]. Supportive supervision has been illustrated to be another key motivational factor for provider performance [32] and substantiated in this study from provider feedback. However, external facility supervision visits declined significantly, which may be attributed to worsening security. Systems for accreditation and licensure are also under consideration by various taskforces to better regulate both the private and public sector and develop an appropriate paradigm for health workforce management.

Task shifting and skill mix

Task shifting to lower cadres, especially in an environment with chronic deficits of physicians and nurses to clinical officers, has shown promise in some countries, although the specific mechanism needs to be configured to the health system context [33]. Experiments to benchmark successes achieved in the deployment of community providers to ensure equitable access for primary and preventive healthcare in countries such as Thailand, Malawi, Ethiopia, India and Pakistan [1] are ongoing in Afghanistan. The training of community providers in IMCI has been aggressively pursued by NGOs to enable both preventive and care treatment for children, which may be a cost-effective and feasible option for child health, especially for communities with poor geographical access. However, there is a growing concern for sustaining and regulating this volunteer workforce, especially in insecure environments. The establishment of community councils, shura-e-sehies, to improve accountability and governance of the public health system has accelerated since 2004, and the presence of these councils emerged as one of the predictors of quality of care. This requires further investigation, to explore the potential of community councils to ensure management oversight of health facilities.

Future considerations

Innovations for improved motivation and retention experimented in other countries, such as extension of the retirement age, additional compensation for extended duty hours, task shifting with mid-level cadres and policies for mandatory rural service following medical certification, may be considered to complement the traditional strategies for increased workforce and reducing income disparities between the public and private sector and performance incentives.

Despite the dearth of evidence of workforce financing, migratory patterns, provider attrition and absenteeism and performance of the private sector in Afghanistan, this study provides some evidence of the importance of workforce adequacy, supervision and other factors we studied on care quality for children. Although task forces have been appointed by the Ministry to advocate and explore viable solutions for strengthening workforce capacity, further empirical research on human workforce management is warranted to determine contributory factors that foster provider motivation and performance to ensure efficiency and care quality. The ongoing evaluation of the performance-based financing scheme will provide better insight into the contextual and individual factors that influence provider motivation and performance and inform policy in Afghanistan.

Conclusion

The health workforce in Afghanistan faces some daunting challenges in the backdrop of worsening security conditions and yet continues to provide care, illustrating remarkable resilience, dedication and commitment. Aside from experimenting with successful innovations from other low-resource settings to determine contextually appropriate mechanisms for workforce motivation and retention, concerted efforts to train, deploy and supervise health providers can result in consistent improvements in quality of care.

Supplementary material

Supplementary material is available at International Journal on Quality in Health Care online.

Funding

The evaluation was funded through a contract to the Johns Hopkins University by the Afghanistan Ministry of Public Health.

Acknowledgements

A special thanks to Dr Abdul Wali, Ministry of Public Heath for providing the estimates on health workforce and Paul Ickx, Management Sciences for Health for his insight into study implications. We are grateful to the Ministry of Public Health, Afghanistan and the survey teams from Johns Hopkins University and Indian Institute of Health Management Research for their contributions. We appreciate the thoughtful comments and recommendations from the anonymous reviewers and editorial team. This study was conducted as a Third Party Evaluation Contract with the Government of Afghanistan to Johns Hopkins University and reviewed by the Institutional Review Board at Johns Hopkins University and the Ministry of Public Health Ethical Review Board in Afghanistan.

References

1
World Health Organization
The World Health Report 2006–Working Together for Health
 , 
2006
Geneva
WHO
2
Anand
S
Barnighausen
T
Human resources and health outcomes–a cross country econometric study
Lancet
 , 
2004
, vol. 
364
 (pg. 
1603
-
9
)
3
Chen
L
Evans
T
Anand
S
, et al.  . 
Human resources for health: overcoming the crisis
Lancet
 , 
2004
, vol. 
364
 (pg. 
1984
-
90
)
4
Kane
RL
Shamliyan
TA
Mueller
C
, et al.  . 
The association of registered nurse staffing levels and patient outcomes: systematic review and meta-analysis
Med Care
 , 
2007
, vol. 
45
 (pg. 
1195
-
204
)
5
Lang
TA
Hodge
M
Olson
V
, et al.  . 
Nurse-patient ratios: a systematic review on the effects of nurse staffing on patient, nurse employee, and hospital outcomes
J Nurs Adm
 , 
2004
, vol. 
34
 (pg. 
326
-
37
)
6
Bostick
JE
Rantz
MJ
Flesner
MK
, et al.  . 
Systematic review of studies of staffing and quality in nursing homes
J Am Med Dir Assoc
 , 
2006
, vol. 
7
 (pg. 
366
-
76
)
7
Blustein
J
Borden
WB
Valentine
M
Hospital performance, the local economy, and the local workforce: findings from a US National Longitudinal Study
PLoS Med
 , 
2010
, vol. 
7
 pg. 
e1000297
 
8
Dolea
C
Stormont
L
Shaw
D
, et al.  . 
Increasing Access to Health Workers in Remote and Rural Areas Through Improved Retention
2009
 
9
Reilley
B
Frank
T
Prochnow
T
, et al.  . 
Provision of health care in rural Afghanistan: needs and challenges
Am J Public Health
 , 
2004
, vol. 
94
 (pg. 
1686
-
8
)
10
Waldman
R
Strong
L
Wali
A
Afghanistan's Health System since 2001: Condition Improved, Prognosis Cautiously Optimistic
 , 
2006
Kabul
Afghanistan Evaluation and Research Unit.
11
Sabri
B
Siddiqi
S
Ahmed
AM
, et al.  . 
Towards sustainable delivery of health services in Afghanistan: options for the future
Bull World Health Organ
 , 
2007
, vol. 
85
 (pg. 
712
-
8
)
12
Edward
A
Kumar
B
Kakar
F
, et al.  . 
Configuring balanced scorecards for measuring health system performance: evidence from 5 years’ evaluation in Afghanistan
PLoS Med
 , 
2011
, vol. 
8
 pg. 
e1001066
 
13
Belay
T
Building on Early Gains Challenges and Options for Afghanistan's Health and Nutrition Sector
 , 
2010
Washington
World Bank
14
Ministry of Public Health
A Basic Package for Health Services for Afghanistan
 , 
2005
Kabul
Islamic Republic of Afghanistan
15
Afghan Public Health Institute, Ministry of Public Health (APHI/MoPH) [Afghanistan], Central Statistics Organization (CSO) [Afghanistan], ICF Macro, Indian Institute of Health Management Research (IIHMR) [India], and World Health Organization Regional Office for the Eastern Mediterranean (WHO/EMRO) [Egypt]
Afghanistan Mortality Survey 2010
 , 
2011
Calverton, MD, USA
APHI/MoPH, CSO, ICF Macro, IIHMR and WHO/EMRO
16
Armstrong Schellenberg
J
Bryce
J
de Savigny
D
, et al.  . 
The effect of integrated management of childhood illness on observed quality of care of under-fives in rural Tanzania
Health Policy Plan
 , 
2004
, vol. 
19
 (pg. 
1
-
10
)
17
Adam
T
Amorim
DG
Edwards
SJ
, et al.  . 
Capacity constraints to the adoption of new interventions: consultation time and the integrated management of childhood illness in Brazil
Health Policy Plan
 , 
2005
Suppl. 1
(pg. 
i49
-
57
)
18
Amaral
J
Leite
AJ
Cunha
AJ
, et al.  . 
Impact of IMCI health worker training on routinely collected child health indicators in Northeast Brazil
Health Policy Plan
 , 
2005
, vol. 
20
 
Suppl. 1
(pg. 
i42
-
8
)
19
Gouws
E
Bryce
J
Habicht
JP
, et al.  . 
Improving antimicrobial use among health workers in first-level facilities: results from the multi-country evaluation of the integrated management of childhood illness strategy
Bull World Health Organ
 , 
2004
, vol. 
82
 (pg. 
509
-
15
)
20
Armstrong Schellenberg
JR
Adam
T
Mshinda
H
, et al.  . 
Effectiveness and cost of facility-based Integrated Management of Childhood Illness (IMCI) in Tanzania
Lancet
 , 
2004
, vol. 
364
 (pg. 
1583
-
94
)
21
Edward
A
Dwivedi
V
Mustafa
L
, et al.  . 
Trends in the quality of health care for children aged less than 5 years in Afghanistan, 2004–2006
Bull World Health Organ
 , 
2009
, vol. 
87
 (pg. 
940
-
9
)
22
Lind
A
Edward
A
Bonhoure
P
, et al.  . 
Quality of outpatient hospital care for children under 5 years in Afghanistan
Int J Qual Health Care
 , 
2011
, vol. 
23
 (pg. 
108
-
16
)
23
World Health Organization
Child and Adolescent Health and Development
2011
 
24
Loewenberg
S
Afghanistan's hidden health issue
Lancet
 , 
2009
, vol. 
374
 (pg. 
1487
-
8
)
25
Johnecheck
WA
Holland
DE
Nutritional status in postconflict Afghanistan: evidence from the National Surveillance System Pilot and National Risk And Vulnerability Assessment
Food Nutr Bull
 , 
2007
, vol. 
28
 (pg. 
3
-
17
)
26
Peters
DH
Noor
AA
Singh
LP
, et al.  . 
A balanced scorecard for health services in Afghanistan
Bull World Health Organ
 , 
2007
, vol. 
85
 (pg. 
146
-
51
)
27
Chaterjee
S
Hadi
AS
Price
B
Regression Analysis by Example
 , 
2000
3rd edn
New York
John Wiley and Sons
28
Willis-Shattuck
M
Bidwell
P
Thomas
S
, et al.  . 
Motivation and retention of health workers in developing countries: a systematic review
BMC Health Serv Res
 , 
2008
, vol. 
8
 pg. 
247
 
29
Grobler
L
Marais
BJ
Mabunda
SA
, et al.  . 
Interventions for increasing the proportion of health professionals practising in rural and other underserved areas
Cochrane Database Syst Rev
 , 
2009
pg. 
CD005314
 
30
Peters
DH
Chakraborty
S
Mahapatra
P
, et al.  . 
Job satisfaction and motivation of health workers in public and private sectors: cross-sectional analysis from two Indian states
BMC Human Resources for Health.
 , 
2010
, vol. 
8
 pg. 
27
 
31
Mathauer
I
Imhoff
I
Health worker motivation in Africa: the role of non-financial incentives and human resource management tools
Hum Resour Health
 , 
2006
, vol. 
4
 pg. 
24
 
32
Naimoli
JF
Rowe
AK
Lyaghfouri
A
, et al.  . 
Effect of the integrated management of childhood illness strategy on health care quality in Morocco
Int J Qual Health Care
 , 
2006
, vol. 
18
 (pg. 
134
-
44
)
33
Fulton
BD
Scheffler
RM
Sparkes
SP
, et al.  . 
Health workforce skill mix and task shifting in low income countries: a review of recent evidence
Hum Resour Health
 , 
2011
, vol. 
9
 pg. 
1