Development and temporal external validation of a prediction model for unplanned readmission in hospitalized patients with cancer and coronary heart disease

Q: What are the key findings of this study?

A four-variable admission model predicted unplanned readmission in patients with cancer and CHD with an optimism-corrected c-index of 0.682 (95% CI 0.599-0.766) and validation c-index of 0.841.

Q: What is the clinical evidence from this study?

Study design: Cohort. Population: Cancer and coronary heart disease (n=438). Intervention: Four-variable admission prediction model (serum sodium, diastolic blood pressure, hemoglobin, and digestive cancer) vs. Padua score and NRS2002. Primary outcome: Unplanned hospital readmission after discharge (OR 2.268 (for digestive cancer vs others in multivariable model); optimism-corrected c-index 0.682, 95% CI 0.599-0.766, p=0.002).

Q: What does this research mean for the field?

A four-variable admission model comprising serum sodium, diastolic blood pressure, hemoglobin, and digestive cancer predicts unplanned readmission in patients with cancer and coronary heart disease with moderate discrimination and adequate calibration. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

Nuoyi Wu; Xinlian Zhang; Mei Yang; Wenyan Yang; WENLI YU; Yuyi Yang; Juan Liu; Qian Wen

doi:10.21203/rs.3.rs-11046775/v1

November 5, 2026Open Access

Development and temporal external validation of a prediction model for unplanned readmission in hospitalized patients with cancer and coronary heart disease

View Full Paper

Ask AI

Bookmark

Share

Key result

Four-variable admission model predicts unplanned readmission in patients with cancer and CHD.

OR 2.268 (for digestive cancer vs others in multivariable model); optimism-corrected c-index 0.682
95% CI 0.599-0.766
P=0.002
n=438

Why the study?

Unplanned readmission is a substantial burden in patients with coexisting cancer and coronary heart disease, yet no risk prediction tool had been developed for this population.

Does a four-variable prediction model using routine admission data accurately predict unplanned readmission compared to existing risk scores (Padua score and NRS2002) in hospitalized patients with coexisting cancer and coronary heart disease?

Key Points

To develop and temporally validate a prediction model for unplanned readmission using routine clinical data available at admission in hospitalized patients with cancer and coronary heart disease.
Conducted a single-center retrospective cohort study of 438 admissions (407 unique patients) discharged between January 2022 and June 2024.
Screened 34 candidate predictors using univariable analysis and LASSO regression with 10-fold cross-validation, refitting selected variables into an unpenalized multivariable logistic model.
Evaluated model discrimination, calibration, and clinical utility, and tested the model without refitting on a temporal validation cohort (n = 118) discharged between October 2023 and June 2024.
The final model retained serum sodium, diastolic blood pressure, hemoglobin, and digestive cancer (likelihood-ratio test χ2 = 16.55, P = 0.002), with digestive cancer as the only statistically significant individual multivariable predictor (OR 2.27, 95% CI 1.08–4.78, P = 0.031).
The apparent and optimism-corrected c-indices were 0.709 and 0.682 (95% CI 0.599–0.766), whereas the temporal validation cohort achieved a c-index of 0.841 (bootstrap 95% CI 0.643–0.985), with adequate calibration in both cohorts (Hosmer-Lemeshow P = 0.932 and P = 0.606).
At the Youden-optimal predicted probability cutoff of 0.089, sensitivity was 0.61 and specificity was 0.75, outperforming both the Padua score (c-index 0.538) and NRS2002 (c-index 0.502).

Study Design

Type

Cohort (n=438)

Multicenter

No

Structured PICO

Does a four-variable prediction model using routine admission data accurately predict unplanned readmission compared to existing risk scores (Padua score and NRS2002) in hospitalized patients with coexisting cancer and coronary heart disease?

P

Population

438 hospital admissions (407 unique patients; median age 70 years, 37.4% female) with confirmed malignant tumor and established coronary heart disease discharged between January 2022 and June 2024 in China.

E

Exposure

Four-variable multivariable logistic regression prediction model and nomogram using routine data available within 24 h of admission: serum sodium, diastolic blood pressure, hemoglobin, and digestive cancer

C

Comparator

Padua prediction score and Nutritional Risk Screening 2002 (NRS2002)

O

Outcome

Unplanned readmission to the same institution after discharge (excluding planned admissions such as scheduled chemotherapy cycles)hard clinical

Main Result

Odds Ratio: 2.268 (95% CI 1.077–4.775)

p-value: p=0.002

A four-variable prediction model using routine admission data (serum sodium, diastolic blood pressure, hemoglobin, and digestive cancer) provides moderate discrimination and adequate calibration for predicting unplanned readmission in hospitalized patients with cancer and coronary heart disease, outperforming existing non-specific risk scores.

Limitations

Single-center retrospective study with a small number of events (33 events, events per variable of 8.25)
Temporal validation cohort contained only 6 events, resulting in wide confidence intervals and imprecise calibration estimates
Excluded death as a competing event and did not capture out-of-hospital deaths or readmissions to other institutions
Lacked data on tumor stage, treatment phase, discharge medications, socioeconomic status, and medication adherence
Included repeated admissions from 26 patients, though sensitivity analysis restricted to first admissions showed concordant results
Single-center study with only 33 events (EPV of 8.25 falls below the recommended threshold of 10), constraining variable selection and lacking potentially informative predictors (tumor stage, treatment phase, discharge medications)
Temporal validation cohort contained only six events, resulting in wide confidence intervals and unstable calibration estimates
Outcome excludes death as a competing event and does not capture out-of-hospital death or admissions to other institutions; lacked data on socioeconomic status, medication adherence, and distance to hospital
26 patients contributed repeated admissions
Used complete-case analysis for the primary model

Ask about this paper

Answers are grounded in this paper and cite their sources.

Implication

Offers moderate readmission risk stratification in cancer-CHD; leaves open need for prospective validation before clinical adoption.

Abstract

Background Unplanned readmission is a substantial burden in patients with cancer, particularly when coronary heart disease (CHD) coexists, yet no risk tool has been developed for this population. We developed and temporally validated a prediction model for unplanned readmission based on routine data available at admission. Methods We conducted a single-center retrospective cohort study of adults with cancer and CHD discharged between January 2022 and June 2024 (438 admissions; 407 unique patients). The outcome was unplanned readmission (33 events, 7.5%). Thirtyfour candidate predictors with < 5% missing data were screened by univariable analysis and least absolute shrinkage and selection operator (LASSO) regression with 10-fold × 5-repeat cross-validation (1-SE rule) and refitted in an unpenalized multivariable logistic model. Discrimination was assessed with the c-index and 1000-bootstrap optimism correction; calibration with the Hosmer-Lemeshow test, calibration slope, and intercept; and clinical utility with decision curve analysis. The model was applied without refitting to a temporally non-overlapping cohort discharged from October 2023 to June 2024 (n = 118; 6 events). Results The final model comprised serum sodium, diastolic blood pressure, hemoglobin, and digestive cancer (overall likelihood-ratio test χ2 = 16.55, P = 0.002). The apparent and optimism-corrected c-indices were 0.709 and 0.682 (95% confidence interval [CI] 0.599–0.766), and the temporal validation c-index was 0.841 (bootstrap 95% CI 0.643–0.985). Calibration was adequate in both cohorts (Hosmer-Lemeshow P = 0.932 and 0.606). At the Youden-optimal threshold (predicted probability 0.089), sensitivity and specificity were 0.61 and 0.75. The model outperformed the Padua score (c-index 0.538) and NRS2002 (0.502). Conclusion A four-variable model using routine admission data provides moderate discrimination with adequate calibration for unplanned readmission in patients with cancer and CHD and remained

Discussion

No takes yet. Share an insight, caveat, or question.

Cite This Study

Wu et al. (2026) conducted a cohort in Cancer and coronary heart disease (n=438). Four-variable admission prediction model (serum sodium, diastolic blood pressure, hemoglobin, and digestive cancer) vs. Padua score and NRS2002 was evaluated on Unplanned hospital readmission after discharge (OR 2.268 (for digestive cancer vs others in multivariable model); optimism-corrected c-index 0.682, 95% CI 0.599-0.766, p=0.002). A four-variable admission model predicted unplanned readmission in patients with cancer and CHD with an optimism-corrected c-index of 0.682 (95% CI 0.599-0.766) and validation c-index of 0.841.

synapsesocial.com/papers/6aee6be671a7594806220d01 https://doi.org/10.21203/rs.3.rs-11046775/v1

View Full Paper

Ask AI

Bookmark

Share

Development and temporal external validation of a prediction model for unplanned readmission in hospitalized patients with cancer and coronary heart disease

Key Points

Study Design

Structured PICO

Main Result

Limitations

Ask about this paper

Implication

Abstract

Discussion

Cite This Study

Development and temporal external validation of a prediction model for unplanned readmission in hospitalized patients with cancer and coronary heart disease

Key Points

Study Design

Structured PICO

Main Result

Limitations

Ask about this paper

Implication

Abstract

Discussion

Cite This Study

Keep reading

Keep reading