A multi-country brand tracking study among Health Care Professionals for a leading pharma brand
The brief for the brand tracking study
We were recommended to a leading pharma brand by its research supplier to review a multi-country brand tracking study over 12+ markets among Health Care Providers (HCPs). The survey involved interviews with a sample of four types of HCP with universe estimates derived by local reps. The study was a multilevel study involving observations relating to HCP decisions as well as patient records nominated by HCPs. The goal was a fast turn-around review on the statistical representativeness of HCP and Patient level data, highlighting any sources of bias and sampling error.
The deliverables of the project
The project involved a review of:
- The target setting process and variation in sample distribution between HCP groups from wave to wave.
- The sampling frame, which was an opt-in HCP online panel.
- The unbiasedness of selection of HCPs from the panel and of patients from different types of HCPs.
- Effect on sampling error of patient cluster sampling and weighting.
The steps we took
The initial stages involved gathering data on sample sizes over recent waves and the rationale behind universe estimates. We received a full data set of recent waves to facilitate additional analysis. We also compiled technical documentation collating data from various sources, including the client, their agency, and an independent adviser working on the tracker. An ongoing dialogue was established with all parties as the review progressed.
The brief
To review a multi-country brand tracking study over 12+ markets among Health Care Providers (HCPs).
The goal
The goal was a fast turn-around review on the statistical representativeness of HCP and Patient level data, highlighting any sources of bias and sampling error.
The challenges we faced
The biggest challenge was understanding the survey process, as there was no central technical report. The technical information on sampling, sample frames, and weighting was the domain of various stakeholders. The rationale for some decisions was not well documented. We could compile some of the information needed from multiple sources, but as usual, with this type of review, there were many gaps. Another challenge was making practical recommendations to implement given available sources of information. We also needed to recommend how changes should be managed to avoid suddenly impacting trends.
The outcome
We produced a written report recommending a move towards random probability sampling from the HCP panel. The proposal was to use a two-stage process for patient selection, such that patients across all markets had a fixed and known probability of selection. We raised some practical issues connected to the panel and the sampling processes needed to achieve this. We also recommended how to transition the existing design to assist with smoothing out differences in trends. Recommendations were also made around getting more consistency in target-setting processes and in creating a formal technical report.
Challenges
The biggest challenge was understanding the survey process, as there was no central technical report.
Outcome
We produced a written report recommending a move towards random probability sampling from the HCP panel.