May 7, 2008 - Researchers are inconsistent about disclosing financial conflicts of interests, and journals are inconsistent in how they use or publish disclosure information, a novel study suggests [1]. The findings call into question the whole purpose and efficacy of a process intended to preserve integrity and eliminate bias in medical literature, investigators say.

"The system of disclosures that we have for the professional literature in medicine is not as extensive as we'd hoped," lead author Dr Kevin P Weinfurt (Duke University, Durham, NC) told heartwire. "I think people are under the impression that disclosures are more consistent and frequent than they actually are."

The results appear online May 6, 2008 in PLoS ONE.

For their study, researchers at Duke University searched for studies about stents published in 2006, then compared prevalence, nature, and consistency of financial disclosures across the papers. They report that of 746 papers by 2985 authors, only 17% had a disclosure statement for any author, and only 6% of authors had disclosure statements. Strikingly, in one-third of the studies, no disclosure statement was provided, despite the fact that one or more authors had disclosed conflicts of interest in other papers.

In an interview, Weinfurt explained that it is difficult to tease out from the analysis whether authors were intentionally failing to disclose a conflict, whether they were unsure what constituted a "relevant" conflict, or whether it was the journals not requesting information or not publishing disclosure information.

"Disclosure is sort of like a new drug that's been proposed to solve a health problem," Weinfurt suggested. "In the journals, we're sort of in a situation where people are using this drug very inconsistently and not at the same dose, and very few people are giving the drug. We're asking the question, how effective is the drug? But the problem is, we're really not seeing a consistent use of the drug so that we can evaluate whether it's working or not."

Lack of Disclosure Policies vs Failure to Disclose?

In their review, Weinfurt et al found that the more renowned medical journals (journals that endorsed International Committee of Medical Journal Editors guidelines or that have a "higher impact factor") were the most likely to include disclosure information. More surprising statistics from the study are below.

Disclosure-Statement Information in Stent-Research Papers in 2006

Occurrence	%
No disclosure statement in article	82.9
Declaration of "no conflicts of interest" in article	12.4
Disclosure of conflict of interest in article	4.7
Proportion of authors with articles with disclosure statements	6
For authors of multiple papers: no disclosure statements vs declarations of no conflicts of interest	64
For authors of multiple papers: no disclosure statements vs specific disclosures	34
For authors of multiple papers: specific disclosures vs no conflict of interest	2

But the numbers don't tell the full story. According to Weinfurt, when he and his colleagues conducted an informal Google search of authors' names who had expressly listed "no conflicts," they found that some were advisory board members for stent manufacturers or consulted for stent companies or for companies that made drugs used during or after stenting. One person had even founded a company that makes stents, yet had not disclosed that information.

While some authors may have intentionally withheld information, Weinfurt gives them the benefit of the doubt.

"Investigators may be getting only very vague directions from the journals and doing their best to comply with them, but perhaps not including information that the public might think is appropriate. If the journal says, please disclose any relative conflicts of interest, and if my article is evaluating the effectiveness of stents, I sincerely might not believe that the consulting work I do for a major manufacturer of medical interventions would be a conflict. So part of the problem might be a lack of awareness on the part of investigators as to what might constitute a conflict and what might be perceived as a conflict."

Of note, just 1.3% of the research papers included in the analysis (623 papers) included information on "author contributions" (the extent to which individual authors contributed to the research and writing of the paper)--something the authors argue is a valuable corollary to understanding conflicts of interest. "If one purpose of financial disclosures is to allow readers to assess the influence of the financial interests on the overall study, presumably readers would also need to know the role played by the authors with the financial interests," they write.

Across the Board

Weinfurt said he and his coauthors chose to look specifically at stents, within a one-year period, partly because of the controversies around drug-eluting stents in the past few years and partly because it is a very active area of scholarship, where people "are really looking to the research to figure out what to do."

The results of stent research will have a major influence on public health, on the multibillion-dollar stent industry, and on products/strategies that are alternatives to stents, he said.

"I think problems with disclosure are across the board, but I think the finances involved with stents and other devices do make them ripe territory for problems with conflicts of interests," he told heartwire.

1. Weinfurt KP, Seils DM, Tzeng JP, et al. Consistency of financial interest disclosures in the biomedical literature: The case of coronary stents. PLoS ONE 2008; 3:e2128. Available at: http://www.plosone.org.

The complete contents of Heartwire, a professional news service of WebMD, can be found at www.theheart.org, a Web site for cardiovascular healthcare professionals.

Background

The study was conducted in the ED of a 426-bed tertiary care hospital in rural North Carolina. A clinical pharmacist was assigned to the ED for consultation with and other assistance to the health care providers during all hours of each shift. The pharmacists' duties in the ED included clinical consultation, patient education, order screening, staff education, and emergency preparedness.

The ED is divided into two units, one for adults and one for children. The adult unit officially has 44 assigned rooms but frequently operates over capacity. The pediatric unit has 13 designated beds. The average combined daily census is 250-270 patients. Usual staffing includes 15-18 registered nurses in the adult unit and 4-6 registered nurses in the pediatric unit; 4 physicians and 2 physician assistants are responsible for both units per shift. The adult unit admits approximately 20% of patients and the pediatric unit less than 5%. The average time in the ED is four hours for patients who are subsequently discharged and not admitted to the hospital. For patients who are admitted, the average time in the ED is seven to eight hours.

Pharmacists in this health system are responsible for reporting all clinical interventions through the computer system (Siemens Corp., Malvern, PA). The use of the Siemen's program was implemented in the hospital in October 2004 and was adopted as an institutionwide medication-order-review system. This program provides pharmacists with the means to document interventions and is accessible only through a personal computer with network access. The program provides data on the number and type of pharmacist recommendations and the percentage of recommendations accepted.

Methods

Definitions

A medication error was defined as any preventable event that may lead to inappropriate medication use or patient harm while the medication is in the control of the health care professional.^[13] Categories of medication errors were omission of dose, omission of information, unauthorized drug, wrong dose, extra dose, wrong route, wrong form, wrong technique, and wrong time.^[14]

Procedures

The study protocol was approved by the hospital's institutional review board. The study was designed as a retrospective review of the charts of all patients, regardless of age, who were admitted to the ED between November 6, 2005, and December 6, 2005 (control group) or between November 6, 2006, and December 6, 2006 (intervention group). For the control group, no pharmacist was present in the ED to check drug orders; for the intervention group, a pharmacist was present. Patient records were selected by convenience sampling by the medical information department. Patients with incomplete charts or no medication orders were excluded.

All medication errors were identified by an independent evaluator who reviewed the medical records, including all progress notes, drug orders, and laboratory test results, using the chart-review techniques described by Flynn et al.^[14] To document the errors, the evaluator used a structured form that was organized and analyzed with Microsoft Access (Microsoft Corp., Redmond, WA).

Each drug order was analyzed for medication errors. A medication error, if generalized to include more than one medication order (as when drug allergy information was missing), was documented only once and not multiple times. After the evaluator identified medication errors, a blinded panel of three pharmacists assessed each one for validity. A majority decision was necessary to validate an error.

Data on the number and types of pharmacist recommendations and the percentage accepted were provided by computer. The ED pharmacists completed a report describing each recommendation that might have led to a change in a medication order.

Statistical Analysis

The study was designed to have an 80% power to detect a 66% relative risk reduction (RRR) in the rate of medication errors historically occurring in at least 10% of ED medication orders.^[6,15] Sample-size calculations showed that 56 chart reviews, or 224 medication orders, were required per study group to find a 66% RRR at the two-tailed significance level of 0.05.^[16] Historical estimates were used for medication error frequency because of a lack of relevant documentation at the health care facility.

The chi-square test was used to compare the groups with respect to sex, race, and medication error rates. Differences in age and number of medications were analyzed with Student's t test and analysis of variance, respectively. Descriptive statistics were used to assess pharmacist recommendations. All statistical analyses were performed with JMP, version 6.0.3 (SAS Institute Inc., Cary, NC).

Results

A total of 490 medication orders written for 198 patients were evaluated for errors. The control group (n = 94) and the intervention group (n = 104) did not differ significantly with respect to age, sex, race, or number of medication orders ( Table 1 ).

A total of 37 and 14 medication errors were identified for the control and intervention groups, respectively. The rate of errors was 16.09 per 100 medication orders for the control group and 5.38 per 100 orders for the intervention group, a 66.6% difference (p = 0.0001).

The pharmacists made 183 recommendations ( Table 2 ). Practitioners accepted 72 (98.6%) of 73 recommendations documented by the pharmacists; 110 recommendations were undocumented as accepted or declined. The most common recommendations involved dosage calculations (29%); inappropriate dosages, drugs, routes, or schedules (26%); order clarifications (16%); and drug allergies (12%).

Discussion

The results of this study suggest that pharmacists assigned to an ED played a role in reducing the rate of medication errors by two thirds. This reduction is consistent with the 66-78% rate reductions observed in previous studies in other hospital settings.^[6,7] Almost all the pharmacists' recommendations were accepted by other health care professionals. The 99% acceptance rate in this study was higher than the acceptance rates reported for the EDs at other institutions but similar to the rates reported for other settings at other institutions.^[6,8,17,18]

The study has several limitations. First, the evaluator was not blinded to the intervention group; this was considered impractical because of the ubiquitous presence of date identifiers in the medical records. However, a structured form with standard definitions of medication errors was used for documentation to increase the objectivity of the analysis. Also, a panel of pharmacists was used to assess the clinical validity of each potential medication error. The panel was blinded to the intervention group, reducing the potential impact of investigator bias. A second limitation is the absence of a concurrent control group. The decision to omit a concurrent control group was based on the lack of a comparable unit without a pharmacist's presence. A third limitation is the use of patient medical records for detection of medication errors. Since direct observation has been shown to be more efficient and accurate than chart review, the rate of medication errors detected may have underrepresented the true rate.^[14] Finally, the study was limited to patients admitted to the ED; the results may not be generalizable to other specialty units.

Conclusion

The rate of medication errors in an ED decreased significantly when pharmacists prospectively reviewed ED medication orders. Virtually all of the pharmacists' recommendations were accepted by other health care providers.

Type of Recommendation	No. Intervention (%)
Dosage calculation	53 (29)
Inappropriate dosage, drug, route, or schedule	48 (26)
Order clarification	29 (16)
Identification of drug allergy	22 (12)
Miscellaneous or unspecified	14 (8)
Approval of nonformulary medication	7 (4)
Identification of duplicate therapy	7 (4)
Clarification of medication history	2 (1)
Identification of drug interaction	1 (1)
Total	183

References

Kohn LT, Corrigan JM, Donaldson MS, eds. To err is human: building a safer health system. Washington, DC: National Academy Press; 1999. Fairbanks RJ, Hays DP, Webster DF et al. Clinical pharmacy services in an emergency department. Am J Health-Syst Pharm. 2004; 61:934-7. Aspden P, Wolcott J, Bootman JL et al. Preventing medication errors: qualitychasm series. Washington, DC: National Academy Press; 2006. Cobaugh DJ, Schneider SM. Medication use in the emergency department: why are we placing patients at risk? Am J Health-Syst Pharm. 2005; 62:1832-3. McCaig LF, Burt CW. National hospital ambulatory medical care survey: 2003 emergency department summary. Adv Data. 2003; 358:1-38. Leape LL, Cullen DJ, Clapp MD et al. Pharmacist participation on physician rounds and adverse drug events in the intensive care unit. JAMA. 1999; 281:267-70. Kucukarslan SN, Peters M, Mlynarek M et al. Pharmacists on rounding teams reduce preventable adverse drug events in hospital general medicine units. Arch Intern Med. 2003; 163:2014-8. Ling JM, Mike LA, Rubin J et al. Documentation of pharmacist interventions in the emergency department. Am J Health-Syst Pharm. 2005; 62:1793-7. McMullin ST, Hennenfent JA, Ritchie DJ et al. A prospective, randomized trial to assess the cost impact of pharmacist-initiated interventions. Arch Intern Med. 1999; 159:2306-9. Crowson K, Collette D, Dang M et al. Transformation of a pharmacy department: impact on pharmacist interventions, error prevention, and cost. Jt Comm J Qual Improv. 2002; 28:324-30. Lada P, Delgado G Jr. Documentation of pharmacists’ interventions in an emergency department and associated cost avoidance. Am J Health-Syst Pharm. 2007; 64:63-8. Leape LL, Bates DW, Cullen DJ et al. Systems analysis of adverse drug events. JAMA. 1995; 274:35-43. National Coordinating Counsel for Medication Error Reporting and Prevention. What is a medication error? www.nccmerp.org/aboutMedErrors.html (accessed 2006 Aug 25). Flynn EA, Barker KN, Pepper GA et al. Comparison of methods for detecting medication errors in 36 hospitals and skilled-nursing facilities. Am J Health-Syst Pharm. 2002; 59:436-47. Barker KN, Flynn EA, Pepper GA et al. Medication errors observed in 36 health care facilities. Arch Intern Med. 2002; 162:1897-903. Brant R. Inference for proportions: comparing two independent samples. Power/sample size calculator. http://newton.stat.ubc.ca/~rollin/stats/ssize/b2.html (accessed 2006 Aug 31). Lee AJ, Boro MS, Knapp KK et al. Clinical and economic outcomes of pharmacist recommendations in a Veterans Affairs medical center. Am J Health-Syst Pharm. 2002; 59:2070-7. Sayles TJ. Documentation of pharmacists’ interventions and associated cost savings. Am J Health-Syst Pharm. 2004; 61:838, 840. Letter.