Measuring the Accuracy of Prognostic Judgments
Bibliographic record
Abstract
Abstract The practical test of a true science is the power it confers of prevision, or of knowing now what will follow hereafter. When we can prognosticate with certainty, medicine will have become a science. —H. Hartshorne A system of medicine Prognosis can be defined as, “a forecasting of the probable course and termination of an illness.” In the clinical practice of oncology the prognosis is important because it may determine the optimal choice of treatment, and because patients often want to know what is going to happen to them in order to prepare themselves for the future. Given that prognostic judgments are used as a basis for important decisions, we need to know how accurate they are. We may not yet be able to achieve the “certainty” in prognosis that Hartshorne envisaged as a characteristic of the scientific medicine, but we should at least be able to measure the predictive value of our prognostic judgments, and take this into account in our decision making. It is generally accepted today that we should not place reliance on the results of diagnostic tests without knowing their accuracy, and the same principle should apply to prognostic judgments. However, while modern textbooks of clinical epidemiology deal thoroughly with methods for measuring the accuracy of diagnostic tests, much less has been written about how to measure the accuracy of prognosis. Although there is a very extensive literature about prognostic factors in cancer, there have been very few reports relating to the accuracy of prognostic judgments in practice in individual cases. In this chapter, we are concerned mainly with the quality of prognostic judgments at the level of the individual patient. This is much more of a challenge than predicting the outcome of groups of cases. Once the average outcome of a specific medical problem has been established in a large group of cases, the average outcome in another large group of similar cases may be predicted with great precision. However, a precise knowledge of the average outcome of an illness may have little or no predictive value in the individual case. For example, if a large group of patients with a specific stage of a specific cancer has been observed to have a 5‐year survival of 50% then we can be confident that if a large group of similar cases is managed in the same way in future, it will also have a 5‐year survival of about 50%. When we try to use this information to predict the outcome in the next case of this illness that we encounter, however, it translates into a probability of 5‐year survival of 0.5; In other words, it leaves us in a state of complete uncertainty. The main objectives of this chapter are to describe methods that can be used to measure the accuracy of prognosis at the level of the individual case, and to review what little is known about the accuracy of prognostic judgments in oncology. We end by considering factors that may limit the degree of “certainty” that we can ever expect to achieve in prognosis in oncology.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.012 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.017 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; both teacher heads agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".