Medical Inspection – Health Blog

Data Editing and Outlier Identification

To help ensure the credibility of the results, data used in the CEI were subjected to a series of steps involving screening, edits, and measures to identify outliers (data values that stand out from others in a set).

First, simple discharge counts were obtained from a hospital’s data and were compared with industry standards and logical expectations. For instance, the number of discharges in a month should be fairly uniform, with some degree of seasonal variability. This step was implemented before a hospital data set was accepted for analysis, and any deviations from expectations were explained or corrected before the project team continued to process the data.
buy antibiotics canada

Table 3 Inpatient Data Edits for Clinical Effectiveness Initiative

After the CEI-DRGs had been assigned and patient cost calculation was complete, the data were subjected to a series of record-by-record edits (Table 3). These edits focused on the internal consistency of data in the patient records; for example, if the discharge date occurred before the admission date, the record was excluded from the analysis. Because the CEI focuses on clinical and financial analyses, the clinical and financial data were also compared and tested for consistency; thus, for patients undergoing major surgery, surgery-related costs should be included on the record.

Overall, approximately 2% of patient records were removed from the CEI database for failing one or more edits; two-thirds of these records failed an edit because they lacked accommodation costs (Edit No. 15). Most of these records were for same-day surgery patients who were erroneously included in an inpatient data set.
kamagra oral jelly uk

After the data edits were applied, the project team examined the distribution of costs by CEI-DRG. Records with costs that were considerably higher or lower than those for other patients in a CEI-DRG were “flagged” as outliers and not included in most analyses. Although such patient records might not be erroneous, they are not typical of patient care in their CEI-DRG and would skew any comparisons. About 0.3% of patient records were identified as outliers, mostly high-cost ones.

Table 4 Definitions of Patient Categories

Geographic and Trend Adjustments

Because the CEI study reported in this article included hospitals from throughout the U.S., hospital costs had to be adjusted for geographical differences. The Medicare Wage Index was used to accomplish this step. Northern New Jersey, the site of the pilot hospitals initially used in the study, was used as the standard to which all hospitals were adjusted.

Because data for multiple years were included in the study, the financial analysts standardized all costs to 2002 dollars using the annual Medicare hospital “market basket” cost trend (in 1997 weightings). canadian pharmacy viagra

Patient Categories

To summarize CEI results in a way that was clinically meaningful to physicians, CEI-DRGs were grouped into patient categories, which were defined to match as closely as possible the FDA-approved indications for the use of LMWHs. After the appropriate CEI-DRGs were assigned to these categories, the remaining CEI-DRGs—in which both an UFH and an LMWH were used—were grouped together into additional clinically meaningful patient categories (Table 4).

Table 5 Detailed Pharmacy Data for “Flagged” Drugs

Because the “Medical at-Risk” category included CEI-DRGs that tended to have a long length of stay regardless of body system or disease etiology, some of these CEI-DRGs also appeared in other patient categories. Some CEI-DRGs were not assigned to any patient category.

Drug Categories

The detailed pharmacy data used in the CEI identified all drugs that were provided to the patients. Patients receiving UFHs, LMWHs, or other anticoagulants (Table 5) were flagged. Patients were then assigned to only one “drug category,” based on the hierarchy shown in Table 6.

Calculating Savings

The goal of the financial component of the CEI was to measure the difference in total patient-care costs of one drug (e.g., LMWH) compared with another (e.g., UFH). To accomplish this, the project team calculated the mean cost per discharge for patients receiving each drug. kamagra soft tablets

To provide reasonable comparisons, only CEI-DRGs with at least 10 patients receiving each drug were included in the analysis for an individual hospital. For analyses involving all hospitals pooled together, a threshold of 30 patients was used. As a result, not all patients receiving the study drugs were included in the CEI analysis. For example, if all patients in a CEI-DRG were given one drug but not the other, no financial comparison could be performed; these patients, therefore, were not included in the analysis even though they received one of the drugs of interest.

Table 6 Drug Categories

Preliminary savings per discharge in a CEI-DRG were then calculated as the difference in the mean costs for patients who were treated with each drug. A “savings” could be positive or negative (a loss). These differences were calculated for total cost and for each individual cost center. A two-tail t-test, based on log-transformed total cost, was used to determine the statistical significance of a cost difference within a CEI-DRG.

Because of fixed and step-variable costs, not all of the differences calculated in this manner could realistically be saved by hospitals in the short term. For example, eliminating the need for some laboratory tests by using one drug instead of another did not mean that a laboratory technician job position would be eliminated; instead, some overtime pay might be saved in addition to the costs of reagents and supplies that would not be used. (If enough tests were eliminated, then adding a new technician position might be avoided in the future; however, that would be a long-term savings.)

To reflect this reality, estimated savings were based on the cost differences reduced by a variable cost percentage, an estimate of the marginal costs that could be saved. For the current CEI analyses, a variable cost percentage of 50% was used. buy kamagra

Table 7 presents an example of this calculation. CEI-DRG “A” had 500 patient discharges during the time period; 200 were treated with older drug 1 and 50 were given the newer drug 2. The remaining 250 patients received other drugs or no drugs in this class. The mean cost per case for patients receiving drug 1 was $7,000 (d₁), and the cost for patients receiving drug 2 was $5,000 (d₂). The difference (d₁ – d₂) was $2,000. Reducing this by a variable cost percentage (V) of 50% resulted in an estimated savings for drug 2, compared with drug 1, of $1,000 per case.

Related Posts

• Measuring Clinical Effectiveness: DISCUSSION
• Measuring Clinical Effectiveness: SAMPLE RESULTS part 2
• Measuring Clinical Effectiveness: SAMPLE RESULTS
• Measuring Clinical Effectiveness: METHODOLOGY