The C-reactive Protein (CRP) Point of Care Testing (POCT) Implementation to Adult Patients Presenting With Symptoms of the Acute Respiratory Tract Infection (ARTI) in Eight Primary Health Care (PHC) in Jakarta.

1. Representative sampling strategy and sample size calculation This research is aim to assess whether the implementation of CRP POCT reduces unnecessary antibiotic prescriptions for ARTI without increasing clinical failure by day seven. A total sample size of 2,544 patients is targeted, with an estimated 318 patients per PHC site and intraclass correlation coefficient of 0.002. This sample size is calculated using data from a prior study, where 7.7 percent of patients has clinical failure (proven by re-consultation in day 3) (pC = 0.077), the expected odds ratio comparing the new approach to standard is 1.25, a non-inferiority margin of 15%, equal implementation and standard ratio (k = 1), 80% power, a two-sided α = 0.05 and loss to follow up 10%. The intracluster correlation coefficient (ICC) value of 0.002 used in the sample size calculation was derived from a previous stepped-wedge cluster randomized trial conducted in Tanzania. Although Tanzania and Indonesia are both classified as low- and middle-income countries (LMICs) and share several contextual similarities, such as resource constraints and primary health care-based service delivery, there are notable differences between the two settings, including the distribution of health facilities, workforce availability, and disease burden. As there are currently no published studies from Indonesia employing a similar cluster design in primary care settings, this ICC value was adopted as the most contextually comparable estimate available. The 15% non-inferiority margin in this study represents a safety parameter for the implementation of CRP point-of-care testing. This margin refers to the acceptable difference in the proportion of patients who may require a follow-up visit to another healthcare facility and receive antibiotic treatment within seven days after the initial consultation. A 10-15% margin has been commonly applied in previous non-inferiority studies evaluating the clinical safety of CRP-guided antibiotic reduction strategies, reflecting an acceptable trade-off between reduced antibiotic use and clinical outcomes in mild acute respiratory infections.
2. Study procedures This study is designed as a pragmatic, stepped-wedge cluster design conducted across eight PHC sites in Jakarta, Indonesia. The PHC sites will be the unit of the study. Eight PHC sites will be purposively selected and assigned to either early or delayed implementation groups, forming the clusters in this stepped-wedge design. Given that conventional experimental designs may be perceived by PHC sites stakeholders as unethical in groups that will not receive CRP POCT during project time, we have opted to employ a pragmatic stepped-wedge cluster design. A stepped-wedge trial utilizes a crossover design with repeated observations over time and involves the allocation of clusters to different timepoints at which they begin receiving the implementation. This design allows for both within-cluster and between-cluster comparisons, thereby increasing statistical power and reducing the number of clusters needed compared to a traditional parallel cluster-design.

Eligible facilities are those averaging ≥30 ARTI consultations per week and willing to participate. At the study initiation, none of the clusters will receive implementation (CRP POCT). Patient enrolment, baseline assessments, and case report form-based data collection will be conducted continuously throughout this transition and implementation period. Every two weeks, there will be transitions of the assigned group of PHC sites from standard group to implementation group. Over a total duration of 18 weeks, all clusters eventually receive the implementation, allowing for both within- and between-cluster comparisons.

The order of primary health centers (PHCs) transitioning to the CRP POCT implementation phase (Group 1 to Group 8) will be determined consecutively based on the sequence in which each PHC sites provides formal consent to participate in the implementation, following allocation approval by the Jakarta Provincial Health Office.

According to this design, all PHCs will experience two phases:

1. Standard phase: Patients will be treated in accordance with current standards of care for acute respiratory tract infections without the use of CRP point of care testing (POCT).
2. Implementation phase: Patients will be treated with usual care and with the addition of CRP point of care testing (POCT). The CRP POCT results will be provided to the treating physician as additional information to support clinical decision-making. The decision to prescribe antibiotics will remain based on the clinical decisions.

Data collection occurs at two key points:

1. At the initial consultation (Day 0): Healthcare providers record patient complaints, physical exam findings, any test results (including CRP levels where applicable), diagnoses, treatments prescribed (including antibiotics), and referrals.
2. At follow-up (Day 7): A follow-up assessment will be conducted via telephone to assess clinical status, any further antibiotic use, additional healthcare visits, and perceived recovery.
3. Data management and analysis Direct access will be granted to authorized representatives from host institution for monitoring and/or audit of the study to ensure compliance with regulations. All data will be captured and processed anonymously. Data will be uploaded onto REDCap. Data will be retained for at least 5 years after end of study.

All study outcomes will be analyzed using multilevel logistic regression models to account for clustering at the level of the primary health care PHC sites, with individual patients included as a second random effect when appropriate. Fixed-effect covariates will include adjusting for stratification variables used and for key baseline characteristics identified a priori as potential confounders or found to be imbalanced between study arms, especially among participants with and without available day 7 follow-up data. Covariates will include:

1. Patient-level variables: age, sex, presenting complaint category, and availability of day 7 follow-up data.
2. Facility-level variables: average monthly outpatient attendance.

Outcomes assessed on day 0 will be analyzed in the per-protocol population, defined as participants with complete documentation of final treatment. Day 7 outcomes will be analyzed in both the per-protocol and complete-case populations (i.e., those with available follow-up data), and results will be reported for both unless otherwise specified.

Prespecified subgroup analyses will assess effect modification across sex, age categories, and presenting complaint types. All statistical analyses will be conducted using Stata software.