The authors would like to correct errors in the article cited above, published in the January 2015 issue of JONA.1 The issues fall into 3 categories: (1) typographical, (2) statements needing additional clarification, and (3) misinterpretation/misreporting of source data. Data corrections were incorporated into the cost analyses, and revised tables are provided in this update (excerpt Table 1, Table 4, Table 5).
Typographical Errors
In the Discussion section, the CALNOC performance data are cited, noting that the median patient injury fall rate was 0.5 per 1000 patient days in 2007 and remained at 0.5 in 2013. In fact, the median patient injury fall rate was 0.8 per 1000 patient days, as stated in the Data and Methods section.
Table 1 indicated that the study of Dykes et al2 occurred in 1 hospital, but in fact it took place in 4 hospitals; this has been corrected in the table below (excerpt from Table 1).
Quigley and colleagues'3 study took place on 2 units at 1 Veterans Administration hospital, not at 2 hospitals as originally stated.
Clarification
The introductory section notes that falls are "[horizontal ellipsis]one of the most commonly reported adverse events" and notes that they are reported to occur in 2% to 20% of admissions. The data in Table 1 indicate that the preintervention fall rates averaged 3.16 per 1000 patient days. Average hospital length of stay is 4.9 days, which would equate to a rate of approximately 15.5 falls per 1000 discharges, or 1.55% of discharges. Note that this is lower than the citations in the introduction.
Misinterpretation/Misreporting of Source Data
Weinberg Et Al Data
Falls with injuries were imputed incorrectly from the Weinberg et al article.4 Weinberg et al reported a preintervention fall rate of 3.61 falls per 1000 patient days, and a postintervention fall rate of 1.305 falls per 10 000 patient days. They reported a rate of 0.554 falls with minor injuries and 0.1003 falls with moderate injury per 1000 patient days in the preintervention period, and postintervention rates of 0.2529 minor injury falls and 0.0361 moderate injury falls per 1000 patient days. The original analysis interpreted these injury fall rates as percentages of the baseline fall rates, but they are in fact the fall rates in and of themselves. Thus, the preintervention falls with injury rate should be 0.65 per 1000 patient days (0.554 + 0.1003), rather than 1.74, and the postintervention falls with injury rate should be 0.29 per 1000 patient days (0.2529 + 0.0361), rather than 0.52. The decline in falls with injuries is thus 0.36 per 1000 patient days.
Quigley Et Al Data
The data in the article by Quigley and colleagues3 were misreported. Quigley et al present rates of falls per 1000 in some instances, and rates per 10 000 in other instances. The data extracted were falls per 10 000, but were reported as falls per 1000. Thus, the preintervention fall rate should be 0.16 per 1000 days (1.63 divided by 10), and the postintervention fall rate should be 0.07 per 1000 days (0.67 divided by 10). The change is thus 0.096 (rounded to 0.10).
Revised Cost Analysis
The 2 changes to fall with injury rates reduce the average rate of falls with injury, from 1.69 to 0.41 per 1000 patient days preintervention and from 0.58 to 0.18 per 1000 patient days postintervention. The effect of fall prevention programs on the rate of falls with injury is thus smaller than originally computed (0.23 per 1000 patient days in the revised model vs 1.11 per 1000 patient days in the original article) (excerpt from Table 1).
Lower rates of falls with injury reduce the expected costs of treating falls. As seen in Table 4 errata, the predicted cost of falls preintervention is $62.23, as compared with $115.00 in the original article. The predicted cost of falls postintervention is $30.50 in the revised model, as compared with $47.15 in the original article. Thus, the predicted savings in the revised article are $31.73, which is less than half that computed in the original article ($67.85). With the average cost of prevention at $83.30, the cost increase associated with fall prevention in the revised analysis is $51.57, which is more than 3 times the increase reported in the original article ($15.45).
Table 5 errata present the results for alternative costs of inpatient fall treatment and prevention and for alternate improvements in fall rates. The top panel of Table 5 presents the net savings from investment in fall prevention for each of the lowest, mean, and highest fall treatment and prevention costs. If fall prevention costs are at the minimum level reported in the literature, then prevention is cost saving at any level of cost for caring for patients after fall. However, the savings are smaller than presented in the original article. However, if fall prevention costs are at the average or maximum, then prevention is always cost increasing, and the net costs are larger than presented in the original article. The lower panel of Table 5 errata presents the net savings from investment in fall prevention if the highest preintervention rate was present, and the lowest postintervention rate was achieved. In this case, the cost savings associated with the lowest-cost fall prevention program were larger, and the cost increases were smaller than in the top panel.
Discussion and Implications for Nursing Administration
The data corrections reduced the value of interventions to prevent patient falls, because lower rates of falls with injuries portend lower expenses for hospitals. The lower return to fall prevention programs is largely the result of patient falls being comparatively rare events-occurring less frequently than other adverse events such as pressure ulcers-and falls with injury are even less common.
These revised findings demonstrate that it is essential that nurse leaders carefully assess the value of potential interventions. Nurse leaders can use the framework presented in this article to develop and estimate the cost savings that can be attained by their own fall prevention programs. As this revision demonstrates, differences in rates of patient falls can affect the expected cost savings from patient safety interventions.
References