Study design and setting
We carried out a facility-based cross-sectional study for a period of 5 months extending from 1st May 2019 to 30th September 2019 in the 6 health districts (HDs) in Yaoundé. These HDs include: Biyem-assi, Cité Verte, Djoungolo, Efoulan, Nkolbisson and Nkolndongo.
The study variables were the socio-professional characteristics of participants, HF-related, and health system (HS)-related characteristics of participating HFs, and Health Facility and Community Information System Standards. Socio-professional characteristics included age, sex, professional qualification, years of experience, and function.
The independent variables were the HF-related, and health system (HS)-related characteristics : Status of the HF (public/private), Presence of health information unit ((HIU; (Yes/No)), Stable person in charge of statistics/data management (Yes/No), Available functional computer for data management (Yes/No), Stable internet (Yes/No), Availability of call credit (Yes/No), RHIS supervision received (Yes/No), Receiving feedback from hierarchy (Yes/No), Receiving training on HI (Yes/No), and Presence of a performance evaluation plan (PEP, Yes/No). A PEP is a written document that describes the process of carrying out the monitoring and evaluation of the RHIS performance, detailing the “What”, the “How”, and the “Why It Matters” for the RHIS evaluation, as well as exploiting evaluation results for RHIS performance improvement and decision making (19).
The dependent variable was the score of the Health Facility and Community Information System Standards. This variable was defined and classified into domains and subdomains by WHO and MEASURE Evaluation (20) as follows:
- Management and Governance (Policies and Planning, Management, Human Resources)
- Data and Decision Support Needs (Data Needs, Data Standards)
- Data Collection and Processing (Data Collection and Management of Individual Client Data; Collection, Management and Reporting of Aggregated Facility Data; Data quality assurance; Information and Communication Technology (ICT))
- Data Analysis, Dissemination, and Use (Analysis, Dissemination, Data Demand and Use)
The score of all the domains was calculated as follows (23):
i) Proportion of scores for the various subdomains: we summed the score for each item (question) in the subdomain, multiplied by 100 and divided by the maximum score for that subdomain.
ii) Proportion of scores for the domains: we summed up the obtained scores of the subdomains in the given domain, multiplied by 100 and divided by the maximum score for that domain.
iii) Global score for all the domains, we summed up all the obtained scores of all the domains, multiplied by 100 and divided by the maximum global score. The obtained global score was then grouped into good score (scoring 60% and above) and poor score (less than 60%).
Sample size and sampling
A minimum sample size (n) of 106 HF that were visited was obtained using the formula: (21), where Z is the quantile of the normal distribution at 5% level which equals to 1.96, P is the proportion of adequately functioning HFs which is 10% (22), d is the precision= 0.06 (21), and non-response rate of 10%. HFs were selected through a stratified sampling that uses probability proportional to size in each HD. The two stratified variables were HD and HF status (Private, Public).
The recruitment criteria for HFs included functional public and private HFs of the operational level whose consent for participation was obtained. Respondents (one per HF) were either the head, person in charge of statistics/data management, or any other responsible staff, capable to provide the needed responses.
Interviewers were recruited and trained to understand the objectives and the methodology of the study. Data were collected using the WHO/MEASURE Evaluation pre-established Rapid Assessment questionnaire (20) that was slightly modified to include the socio-professional characteristics of respondents, HF and HS-related characteristics. Each question was scored as: 0 (no answer/not applicable); 1 (not present, needs to be developed); 2 (needs a lot of strengthening); 3 (needs some strengthening); and 4 (already present, no action needed).
Statistical data analysis
Data were entered into Microsoft Excel 2013, cleaned and then exported for analyses using IBM-SPSS version 25 (24). Frequencies and percentages (%) were used to describe qualitative variables. Consistency of the RHIS assessment tool in measuring the gaps and weaknesses in the RHIS was measured using the Cronbach’s alpha (α) whose score is comprised between 0 – 1, and was interpreted as (25-26): (i) unacceptable if α < 0.7, (ii) acceptable if 0.7 ≤ α < 0.8, (iii) good if 0.8 ≤ α <0.9, (vi) excellent if α ≥ 0.9. Pearson’s Chi-square test (Fisher exact test where relevant) was used to establish relationships between qualitative variables. Associations were further quantified using unadjusted Odds ratio (OR) for univariable analysis and adjusted Odds ratio (aOR) for multiple logistic regression analysis with 95% confidence interval (CI). The Hosmer-Lemeshow (HL test)’s goodness of fit test was used to assess the adequacy of the multiple logistic regression. A p-value of less than 0.05 was considered statistically significant.
The study received ethical approval CE N0 00786/CRERSHC/2019 from the Regional Research Ethics Committee for Human Health of the Centre and the authorization N0 00756-/AP/MINSANTE/SG/DRSPC/CRERSH from the Regional Delegate of Public Health for the Centre Region. Study procedures were described to participants, during which they were briefly and clearly informed of their voluntary participation in the study; and that refusal to participate would have no negative consequences. Their informed consents were then obtained prior to the interview.