Reference
Excel export reference
EasyHTA exports Excel workbooks for survival curves, cost-effectiveness assumptions, the cure point and SMR grid, probabilistic sensitivity analysis, and budget impact. This page lists the sheets in each and what they hold, so you know what the export covers before you run one. The column headers are in the workbook itself.
Where the exports are
- Survival curves: the Export Excel button on the curves overview, once at least one endpoint has been fitted.
- Cost-effectiveness: the Export sources & assumptions button on the cost-effectiveness results.
- Cure point and SMR grid: the Export grid (Excel) button on the Sensitivity tab, after running the grid on a Partitioned survival with cure model project.
- PSA: the Export PSA to Excel button on the Sensitivity tab, after running a probabilistic sensitivity analysis.
- Budget impact: the Export to Excel button on the BIA results view, after computing.
The Tables tab of a Partitioned survival project shows the inputs and results as dossier-style tables. Copy table puts one table on the clipboard, and Export to Word or Export to Excel downloads the whole set.
Exports are generated in your browser. The workbook is assembled locally from results already on screen, so nothing is uploaded to produce it. Exporting is never locked: a project that has become read-only, for example after a trial ends, still exports.
Survival-curve workbook
Sheets appear in this order. Sheets marked “per endpoint and arm” repeat once for each fitted combination, for example OS_Intervention and PFS_Control.
| Sheet | Contents |
|---|---|
<Endpoint>_<Arm>_Parametersper endpoint and arm | Fitted parameter point estimates, one row per distribution and parameter. Parameter names follow the flexsurv convention for each distribution, listed in Parametric distributions. |
Variance_Covariance | One block per endpoint, arm and distribution, holding the fitted variance–covariance matrix for PSA. See Variance–covariance matrices. |
Model_Fit_Statistics | Every fit across every endpoint and arm in one table, with the fit statistics and summary survival measures for each. This is the sheet to use when ranking models or reproducing a selection table for a submission. It also lists the parameter count and the number of events, and, for spline models, which are grouped under their own heading, the minimum events between adjacent knots and the knot locations in years. |
Survival_Curves | The reconstructed Kaplan-Meier curve (labelled Observed) and every fitted distribution across the time horizon (labelled Fitted), in long format so you can filter to redraw either. Spline models are included only when selected; the other spline fits are on Model_Fit_Statistics. |
<Endpoint>_<Arm>_IPDper endpoint and arm | The reconstructed patient-level dataset, one row per inferred patient with its time and event indicator. See The Guyot algorithm for what this is and is not. |
Metadata | The provenance record to cite in a submission appendix: analysis type, export timestamp, the endpoints and arm configuration, the time horizon, and the reconstruction method. |
Cure_Model_Extrapolationcure analyses | The composite curve from the Survival curves tab, by month. Each point is labelled with its phase (observed data, parametric extrapolation or background mortality), and the distribution and life table used are listed. A curve that could not be built still gets a row with the reason. |
Time_On_Treatmentwhen time-on-treatment curves exist | Which regimen drugs each time-on-treatment curve (ToT1 to ToT3) bills, so the slot names on the other sheets can be read without the app. |
Sources & Assumptions | The assumptions log from the Survival curves tab: section, assumption, value, source and notes. |
Variance–covariance matrices
The Variance_Covariance sheet holds one block per endpoint, arm and distribution. Each block gives the parameter table, the covariance matrix, and the same matrix expressed as correlations, so a probabilistic sensitivity analysis can draw correlated parameter samples. The matrices are on the estimation scale the fit was made on (log scale for a positive parameter such as shape, scale or rate), so sample on that scale and back-transform afterwards; every block restates this.
A spline model’s block also lists its knot locations, in log years and in years. The knots are fixed by the reconstructed event times rather than estimated, so they have no row in the matrices and no standard error.
A block can read
NOT AVAILABLE. That happens for a fit that produced no usable covariance matrix. Running Reconstruct IPD again restores it for an older analysis; if it still reads NOT AVAILABLE, that distribution cannot seed a correlated PSA. The sheet header counts how many of the fitted distributions have a matrix.Cost-effectiveness assumptions
| Sheet | Contents |
|---|---|
Sources & Assumptions | Every assumption behind the cost-effectiveness model, grouped by section, with its value, the source you recorded and any note. Built for a dossier appendix. |
Cure point and SMR grid workbook
| Sheet | Contents |
|---|---|
Grid Settings | The willingness-to-pay threshold, weighting, base case, extremes and axes, so any cell can be reproduced. |
Grid Cells | One row per cell: cure point, SMR, incremental cost, incremental QALYs, ICER and NMB, with the change from the base case and any flags. Values only, with the currency named in each header. |
PSA workbook
| Sheet | Contents |
|---|---|
PSA Settings | The run settings, exactly as the Sensitivity tab discloses them for this run. |
PSA Iterations | One row per draw, with the incremental cost, incremental QALYs and that draw’s ICER. Reproducible from the seed printed in the sheet header. |
PSA CEAC | The probability each arm is cost-effective at every willingness-to-pay threshold, plus a summary at your reporting threshold with its Monte Carlo error. |
Budget-impact workbook
Currency symbols in the headers follow the currency set on the analysis.
| Sheet | Contents |
|---|---|
Inputs | Every input the calculation used, in sections, with the source you recorded against each one. |
Per-year - Current | The current standard-of-care scenario, with a row per drug and per pathway or programme cost line for each year, plus a year total. Omitted when the analysis has no comparator. |
Per-year - New scenario | The scenario with the new treatment introduced, laid out as above. Named Per-year - Treatment when the analysis has no comparator. |
| Per-weight-category sheets | When weight-based dosing is configured, each weight category also gets its own per-year sheet per scenario, laid out as above. |
Aggregate | Scenario totals across the whole horizon and the net budget impact between them, broken down by payer and by cost line, plus any pathway spend falling due after the time horizon (reported separately, outside the totals). |
Capacity | Appears when a resource line carries a physical unit (for example nurse hours or bed-days): the resource released or consumed in each scenario. Reports quantities, never currency. |
Sources & Assumptions | Every field with its source and any note you attached. |
Charts | The charts from the results view, captured as images and stacked vertically. |
Using the export downstream
- Keep the whole workbook.
Metadata,Sources & AssumptionsandPSA Settingsare what make a figure traceable months later, and they are lost if you paste values into a new sheet. - Filter rather than split.
Model_Fit_StatisticsandSurvival_Curvesare deliberately long-format so a pivot table can slice by endpoint, arm, and distribution. - Re-export after any change. The workbook is a snapshot of the results on screen at the moment you clicked.
LinkedIn