Archives

  • 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • 2021-03
  • br Comparison of Hospital Rank Using Interval

    2020-08-30


    Comparison of Hospital Rank Using Interval Mortality Time Points
    Risk-adjusted estimates of center rates were used to create hospital rankings at 12 interval mortality time points
    Patients underwent surgical procedure
    • Exclude patients with unknown vital status (Excluded n = 35,342)
    • Exclude patients with missing value for last contact (Excluded n = 412)
    • Exclude patients with missing value for days from surgery (Excluded n = 4,632)
    • Exclude patients that Adalimumab did not undergo surgery of primary site at reporting facility or that underwent local destruction instead of surgical resection (Excluded n = 29,235)
    FIGURE 1. Analytic cohort. NSCLC, Non–small cell lung cancer; NCDB, National Cancer Database.
    574 The Journal of Thoracic and Cardiovascular Surgery c August 2019
    Moore et al Thoracic: Lung Cancer
    TABLE 2. Risk-adjusted 30-day and 90-day mortality results
    30-d Mortality
    90-d Mortality
    Variable
    OR LL UL
    OR
    LL
    UL
    Reference
    Reference
    Sex
    Female
    Reference
    Reference
    Male
    Race
    White
    Reference
    Reference
    Black
    Other
    Comorbidities
    Charlson/Deyo score
    Insurance status
    Not insured
    Reference
    Reference
    Private insurance
    Medicaid
    Medicare
    Other government insurance
    Unknown
    Income
    Median income quartiles
    Year of diagnosis
    Primary site*
    C340—Main bronchus
    Reference
    Reference
    C348—Overlapping
    C349—Not otherwise specified
    Analytic stage
    Reference
    Reference
    I
    II
    IV
    Occult
    Histology
    1—Adenocarcinoma
    Reference
    Reference
    3—Carcinoid
    5—Other
    Grade
    Neoadjuvant therapy
    No
    Reference
    Reference
    Yes
    (Continued) 
    THOR
    The Journal of Thoracic and Cardiovascular Surgery c Volume 158, Number 2 575
    THOR  Thoracic: Lung Cancer Moore et al
    TABLE 2. Continued
    30-d Mortality
    90-d Mortality
    Variable OR LL UL
    OR LL UL
    Facility type
    Community Reference
    Reference
    Surgical procedure
    <1 Lobe removed Reference
    Reference
    Average annual procedure volume (cases)
    1-3 Reference
    Reference
    Facility city size
    Values are presented as odds ratio (OR) and upper limit (UL) and lower limit (LL) of 95% confidence interval. Generalized logistic model with mixed effects were used to generate risk-adjusted models for 30-day mortality and 90-day mortality. Boldface type indicates statistical significance. SCC, Squamous cell cancer; BAC, bronchioloalveolar cancer. *Anatomic location of the tumor. The coding is derived from the ICD-O-3 topographical classification.10
    (30 days, 60 days, 90 days, 120 days, 150 days, 180 days, 210 days, 240 days, 270 days, 300 days, 330 days, and 360 days). Table 3 shows the changes in ranking using only 30-day and 90-day mortality. Eleven out of 30 facilities were in the top 2.5% at 30 days and moved to the middle 95% at 90 days. Similarly, 12 out of 30 facilities that were in the bottom ranking group at 30 days moved up into the middle ranking group at 90 days. To investigate the trend across all interval mortality points, Figure 2 shows the ranking changes between time points. Line density cor-responds to the number of facilities changing ranking groups. The total percentage of facilities changing groups is indicated in text at the bottom of Figure 2. The largest changes were seen between 30-day to 120-day mortality models (2.5%-2.7%).
    DISCUSSION
    Thirty-day mortality is the time point used most often to compare surgical outcomes. In other single-institution
    TABLE 3. Number of hospitals that changed ranking categories (bottom 2.5%, middle 95%, or top 2.5%) according to 3-tier rating system between 30-day and 90-day mortality
    90-d Mortality
    studies or smaller database investigations of lung cancer resection, the mortality rate doubles between 30 days and 90 days.5-7 In Pangaea study, 9007 patients died before the 30-day mortality time point. Between 30 days and 90 days, 8097 patients died. This is reflected in the mortality rate, which was almost 3.0% at 30 days and increased to 5.7% at 90 days. Similar factors were associated with death at 30 days and 90 days and included patient, tumor, and facil-ity characteristics. It is likely that each mortality time point represents a different concept in patient care. Thirty-day mortality is reported largely to assess immediate periopera-tive outcomes and safety. However, given advances in crit-ical care, 60-day or 90-day mortality measurements may be a more appropriate measure of postoperative survival and effectiveness of overall oncologic care.