Glossary & methodology

How the survey was processed, how every metric is computed, and a plain-English definition of every term used in the app.

Methodology

Sample

300 US-based oncology healthcare professionals (HCPs) actively managing breast cancer patients. Roles include oncologists and hematology-oncologists across community and academic settings, with a spread of years practising and monthly patient volume. Source file: Prime - US Breast Cancer Survey Export n=300.xlsx.

Profile tags

Setting - Community or Academic, taken directly from Q3.
Experience - Early-career (≤10 years) or Experienced (>10 years), derived from Q4.
Volume - Low (1–9), Mid (10–25), High (26–50), Very-high (50+) active breast-cancer patients per month, from Q6.
Research-active - Yes if the respondent reports PI, sub-investigator or active trial enrolment (Q7).

How rates and indices are computed

Selection rate = % of respondents in the current slice who picked an option. Multi-select questions can sum above 100%.

Index = (slice selection rate ÷ full-sample selection rate) × 100. 100 means the slice behaves like the average; 130 means 30% over-indexed; 70 means 30% under-indexed. We highlight pink at ≥125 and dim at ≤80.

Likert means are computed on the 1–7 acceptability scale used in the digital cadence question.

Statistics tab

The descriptive statistics (range, mean, median, std dev, variance) shown on the Statistics tab are lifted directly from the workbook's Statistics sheet so the numbers match the source export exactly. Values are computed against numeric response codes (1 = first option, 2 = second option, etc.) and are most meaningful on ordinal questions.

KEE normalisation

Free-text expert names from Q149–Q151 were case-normalised, whitespace-trimmed and fuzzy-merged on last name + first initial. Obvious typos collapse to the most common spelling. Anonymisation (toggle in the header) replaces names with “KEE #n, m mentions” while preserving rank order.

Caveats

Slices with very small n (<10) are noisy - read indices with caution.
No statistical significance testing in v1.
“Other (please specify)” free-text answers are excluded from charts.

Glossary

Index (i): Index vs. total sample. 100 = same rate as the full n=300. 130 = this slice picks it 30% more often than average; 70 = 30% less often.; Calculated as (group selection rate ÷ total sample selection rate) × 100. Highlighted pink when ≥ 125 (over-indexes) and dimmed when ≤ 80 (under-indexes).
n: Number of respondents in the current slice.
Selection rate (%): % of respondents in the slice who selected this option. Multi-select questions can sum above 100% because each respondent can pick several options.
Mean: Arithmetic average of the numeric codes used to record each answer (1 = first option, 2 = second, etc.).
Median: Middle value of the response codes when ordered low to high.
Standard deviation: Typical distance of any individual response from the mean. Low = consensus; high = polarised answers.
Variance: Square of the standard deviation. Same shape, different scale.
Range: Lowest and highest response code observed for this question (which options were ever chosen).
Likert 1–7: Acceptability scale where 1 = strongly disagree / not at all acceptable and 7 = strongly agree / very acceptable.
KEE: Key External Expert. A clinician, academic or KOL named by respondents as influential when interpreting new breast-cancer data.
Co-mention: Two KEEs named by the same respondent. Strong co-mention pairs suggest natural partnership candidates for co-created MedEd.
MSL: Medical Science Liaison - a non-promotional, scientifically trained representative from a pharmaceutical company.
Congress journey: Three discrete moments - before, during and after a major scientific congress - where HCPs consume different formats.
Setting: Practice setting: Community (private practice / community hospital) vs Academic (academic medical centre).
Experience: Early-career (≤10 years practising oncology) vs Experienced (>10 years).
Patient volume: Active breast-cancer patients personally managed per month, bucketed Low / Mid / High / Very-high.
Research-active: Whether the HCP currently acts as PI, sub-investigator or enrols patients into clinical trials.