Table 4. Mixed Methods Matrix of Factors Influencing Delivery and “Normalisation” of the CQI Intervention .
| Quantitative | Qualitative | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Participation in CQI (% of All Visits) | Time to First PDSA * Start | Time to First PDSA * Review | Number of Steps to First Observed Improvement (≥10 Percentage Points) § | Guidelines Factors | Individual Health Professional Factors | Patient Factors | Professional Interactions | Incentives and Resources | Capacity for Organisational Change |
| Clinic 1 (September 29, 2015) | |||||||||
| Medium size, rural setting | |||||||||
| Operational manager 39% | 2 days | 20 days | VL:1 step | Staff turnover within clinic CQI team – affected trialability of intervention | Staff not familiar with 2015 eMTCT guidelines | Patients leave clinic prior to HIV care tests due to long queues | Limited sharing of CQI skills between clinic CQI team members and other clinic staff | Staffing shortages | Operational manager authorisation required to implement all CQI activities |
| Professional nurse 61% | |||||||||
| Lay counsellor 64% | Rpt HIV test: 0 steps | ||||||||
| Limited self-efficacy | Patients not contactable for follow-up | Data quality challenges due to lack of communication between staff cadres | Poor documentation of tests and results in medical records | ||||||
| Data quality challenges due to limited understanding of M&E data | Patient not adherent to ART due to lack of food | Operational manager authorisation required to implement all CQI activities | DoH eMTCT monitoring forms not available | ||||||
| No printer cartridge for printing essential clinical and M&E forms | |||||||||
| Landline out of order | |||||||||
| Paper-based results and routine M&E | |||||||||
| Clinic 2 (November 24, 2015) | |||||||||
| Large size, urban setting | |||||||||
| Operational manager 50% | 1 day | 62 days | VL: 4 steps | Staff turnover within clinic CQI team – affected trialability of intervention | Enthusiasm for CQI | Improved patient awareness of VL and voluntary attendance for results follow-up | Limited sharing of CQI skills between clinic CQI team members and other clinic staff | Staffing shortages | Operational manager authorisation required to implement all CQI activities |
| Professional nurse 64% | |||||||||
| Lay counsellor NA** | Rpt HIV test: 1 step | ||||||||
| Limited self-efficacy | Data quality challenges due to lack of communication between staff cadres | Paper-based results and routine M&E | |||||||
| No ownership of improvement activities | Operational manager authorisation required to implement all CQI activities | DoH eMTCT monitoring forms not available | |||||||
| Staff not familiar with 2015 eMTCT guidelines | HIV test kits out of stock | ||||||||
| Clinic 3a (January 26, 2016) ‡ | |||||||||
| Small size, rural setting | |||||||||
| Operational manager NA** | 86 days | 70 days | VL: 1 step | Limited self-efficacy | General clinic patients unwilling to adjust attendance to accommodate ANC patient needs | Data quality challenges due to lack of communication between staff cadres | Staffing shortages | Operational manager authorisation required to implement all CQI activities | |
| Professional nurse 54% | |||||||||
| Lay counsellor NA** | Rpt HIV test: 3 steps | ||||||||
| Data quality challenges due to limited understanding of M&E data | Patients not contactable for follow-up | Operational manager authorisation required to implement all CQI activities | Poor documentation of tests and results in medical records | ||||||
| Staff not familiar with 2015 eMTCT guidelines | VL results delays | ||||||||
| DoH eMTCT monitoring forms not available | |||||||||
| Computer not working | |||||||||
| Paper-based results and routine M&E | |||||||||
| Clinic 3b (January 28, 2016) | |||||||||
| Very small size, rural setting | |||||||||
| Operational manager 38% | 7 days | 5 days | VL: 1 step | Staff not familiar with 2015 eMTCT guidelines | Clinic staff know community members very well due to living in deep rural community | Data quality challenges due to lack of communication between staff cadres | Staffing shortages | Operational manager authorisation required to implement all CQI activities | |
| Professional nurse 24% | |||||||||
| Lay counsellor 55% | Rpt HIV test: 1 step | ||||||||
| Limited self-efficacy | Operational manager authorisation required to implement all CQI activities | Low clinical workload – more time to implement CQI activities | |||||||
| Data quality challenges due to limited understanding of M&E data | Good team spirit | Overcrowding on doctor’s day due to small clinic size | |||||||
| Difficult for lower cadre staff (eg, data capturer, lay counsellor) to feedback CQI information to more senior staff | DoH eMTCT monitoring forms not available | ||||||||
| Poor documentation of tests and results in medical records | |||||||||
| No printer cartridge for printing essential clinical and M&E forms | |||||||||
| Paper-based results and routine M&E | |||||||||
| Clinic 4 (March 17, 2016) † | |||||||||
| Large size, urban setting | |||||||||
| Operational manager 44% | 63 days | 55 days | VL: 2 steps | Staff turnover within clinic CQI team – affected trialability of intervention | Some staff not familiar with 2015 eMTCT guidelines | Good teamwork within clinic CQI team | Staffing shortages | Operational manager authorisation required to implement all CQI activities | |
| Professional nurse 63% | |||||||||
| Lay counsellor 61% | Rpt HIV test: 0 steps | ||||||||
| Data quality challenges due to limited understanding of M&E data | Data quality challenges due to lack of communication between staff cadres | HIV test kits out of stock | |||||||
| Operational manager authorisation required to implement all CQI activities | Lack of space for sorting laboratory results | ||||||||
| DoH eMTCT monitoring forms not available | |||||||||
| Poor documentation of tests and results in medical records | |||||||||
| Paper-based results and routine M&E | |||||||||
| Clinic 5 (May 18, 2016) | |||||||||
| Small size, rural setting | |||||||||
| Operational manager 3% | 20 days | 16 days | VL: 2 steps | Limited self-efficacy | Demanding patients, also attend overnight even for non-emergencies | Limited sharing of CQI skills between clinic CQI team members and other clinic staff | Staffing shortages | Operational manager authorisation required to implement all CQI activities | |
| Professional nurse 59% | |||||||||
| Lay counsellor 41% | Rpt HIV test: 0 steps | ||||||||
| Data quality challenges due to limited understanding of M&E data | Reluctance to queue for clinical consultations during daytime | Incomplete handover of patient tracking processes during periods of annual leave | HIV test kits out of stock | ||||||
| Data quality challenges due to lack of communication between staff cadres | ART out of stock | ||||||||
| Operational manager authorisation required to implement all CQI activities | Poor documentation of tests and results in medical records | ||||||||
| Paper-based results and routine M&E | |||||||||
| Clinic 6 (July 19, 2016) | |||||||||
| Medium size, rural setting | |||||||||
| Operational manager 28% | 7 days | 58 days | VL: 2 steps | Limited self-efficacy | Incomplete handover of patient tracking processes during periods of annual leave | Staffing shortages | Operational manager authorisation required to implement all CQI activities | ||
| Professional nurse 45% | |||||||||
| Lay counsellor 62% | Rpt HIV test: 2 steps | ||||||||
| Data quality challenges due to limited understanding of M&E data | Data quality challenges due to lack of communication between staff cadres | HIV test kits out of stock | |||||||
| Operational manager authorisation required to implement all CQI activities | |||||||||
| Health worker perspectives | |||||||||
| Increased awareness of eMTCT guidelines | Understanding rationale of eMTCT guidelines | Patients start ANC late in pregnancy | Good team work as a result of CQI | Staffing shortages | Resistance to change | ||||
| Increased effort needed to maintain patient tracking notebook as not compatible with M&E registers | Limited self-efficacy -needing operational manager for all decisions | Patients not contactable – cell phone not working | Needing leadership – operational manager to guide services and decisions | CQI interesting; CQI mentors nice people | CQI not sustainable without external mentorship or supervision | ||||
| CQI as an ‘eye opener’ on quality shortfalls | |||||||||
Abbreviations: ANC, antenatal care; ART, antiretroviral therapy; CQI, continuous quality improvement; M&E, monitoring and evaluation; PDSA, Plan-Do-Study-Act cycle; eMTCT, elimination of mother-to-child transmission of HIV; VL, viral load.
Clinic size is based on clinical workload rather than building size. Most participating clinics were in small single-storey buildings; DoH, Department of Health.
Operational manager, professional nurse and lay counsellor participation as a proportion of all CQI visits at each clinic were estimated from attendance registers.
* Time is in calendar days. PDSAs in this table refer to activities directly addressing HIV VL monitoring and/or repeat HIV testing (Figure 1, Change Ideas). General data quality improvement activities including other PDSAs (eg, checks for consistency between source documents) are not included in this table.
**Not applicable as health worker not recruited to clinic CQI team or not working at clinic.
‡ “Gross” staffing shortages noted at this clinic.
† “Extreme” staffing shortages were noted at this clinic which was frequently full. The operational manager was on annual leave at the start of the intervention, and the Acting operational manager was often providing clinical services and unable to attend CQI meetings.
§Steps are counted from the step immediately preceding intervention rollover to the first noted improvement step (regardless of subsequent step trends) – eg, an improvement noted during the intervention rollover step was counted as 1 step to improvement. Clinics which had a decrease or minimal change throughout the post-intervention period were allocated 0 steps. Although the number of steps to first observed improvement is described, there were fluctuations in endpoint achievements with intermittent decline in testing as shown in Figure 2.
Note: Qualitative data reported in this table are based on observations by the CRH CQI mentors and health worker interviews, listed according to the TICD framework. Details of each factor and its likely effect on intervention implementation are provided in Tables S3 and S4.
Quantitative data summarise ‘reach’ of CQI for key health workers, time to first PDSA cycle start and review (proxy for time to intervention uptake), and number of time steps to first observed improvement§ in each endpoint (proxy for delayed intervention effect).
Clinics are listed in order of randomisation with intervention rollover date in brackets.