Critical appraisal of the "one-day-per-percent" rule - A register-based observational study.

Background: The "one-day-per-percent" rule states that for burn patients, one day of hospital stay can be expected for each percentage of burned body surface. This study aimed to test the rule's predictive value.

Methods: The study is a register-based observational study on all burns treated in the Norwegian National Burn Center 2000-22. All children and adults admitted and actively treated in this period were included (2269 patients). Patients receiving palliative care were not included (55). Age when injured, sex, year of admittance, LOS (length of stay), %TBSA (total body surface area), discharged alive/dead, and palliative care: yes/no were extracted from the registry. Quantile regression was used to estimate the quantiles of LOS/%TBSA as a function of %TBSA.

Results: For injuries < 15 %TBSA, the median LOS/%TBSA values were above 1.0, and the ratio increased as the injury size decreased. The median LOS/%TBSA values were close to 1.0 for patients with injuries ≥ 15 %TBSA (Table 2, Figure 2). The quantile regression analysis revealed a considerable spread in estimated values for both adults and children. Children had a lower median LOS/%TBSA than adults and a value below 1.0 for injuries 15-30 %TBSA. When survivors and non-survivors were compared, a higher median LOS/%TBSA was found for survivors with injuries > 60 %TBSA.

Discussion: The "one-day-per-percent" rule does not seem applicable for injuries < 15 %TBSA. Even for more extensive injuries, the large spread in estimated values in different quantiles renders it poor at predicting length of hospital stay. There also seem to be many caveats when it is used to measure an institution's efficacy or when comparing different institutions.

Conclusion: The predictive value of the "one-day-per-percent" rule is deficient when considering individual patients. It does not fit for injuries < 15 %TBSA, and the significant variation for patients with injuries of a similar extent renders it not viable as a forecasting tool.