Hospitals’ web site lists of their interventional pain procedures inadequately reflect the diversity of their actual pain medicine practices

Tyler L. Abarca, Franklin Dexter, Richard H. Epstein, Amy C.S. Pearson

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

Background: Multidisciplinary chronic pain management includes many types of interventional pain procedures. However, navigating the landscape of providers offering such services is challenging. Objective: We investigated whether stakeholders (e.g., patients, referring physicians, hospital administrators, nurses working for insurance companies, and state officials) could accurately judge the diversity of interventional services actually provided based on information gathered from hospital Web sites. Design: This was an observational cohort study. Setting: All 119 nonfederal hospitals in Iowa were included in the study. Methods: We recorded the publicly available data presented on all hospital Web pages related to interventional pain procedures. We counted the listed types of procedures and numbers of pain medicine physicians portrayed. We compared those results with actual performed interventional pain procedures calculated using contemporaneous data from the Iowa Hospital Association. The diversity of types of procedures performed was quantified using the inverse of the Herfindahl index. Results: No pain medicine physician was identified on the Web site for 87.4% of hospitals. Such hospitals accounted for 61.4% of the interventional pain procedures performed statewide. The partial Kendall correlation between the count of types of procedures listed on Web sites and the number of pain medicine physicians, controlling for the performed procedures during the year, was too small to be informative: 0.22 (95% Confidence Interval [CI], 0.07 to 0.38; P = .005). The one-sided upper confidence limit that included 0.50 (i.e., moderate) was the 99.98% limit. The partial correlation between the count of types of procedures listed on Web sites and the actual diversity of types of procedures performed, controlling for the performed procedures during the year, was not statistically significant: 0.12 (95% CI,-0.03 to 0.28; P = .12). The partial Kendall correlation between the number of pain medicine physicians listed on the Web sites and the diversity of types of procedures performed was not significant: 0.03 (95% CI,-0.13 to 0.19; P = .73). Limitations: This study was limited to the state of Iowa, where we found that 38.6% of interventional pain procedures were performed at hospitals with at least one pain medicine physician. The 38.6% is substantively less than the corresponding estimate of 54.2% for the state of Florida. The estimate of 38.6% exceeds the estimate of 30.4% for critical access hospitals in the United States nationwide. Although the heterogeneity is as expected, it shows that findings are likely to be heterogeneous across regions. Conclusions: Stakeholders could not have accurate awareness of the spectrum of services provided by multidisciplinary pain medicine clinics and physicians based on currently reported data, even if they sought it out from publicly available information. Transparency will need to come from pain medicine physicians, at facilities providing the full diversity of services, releasing quantitative data about the services that they provide (e.g., counts by procedure).

Original languageEnglish (US)
Pages (from-to)E723-E730
JournalPain Physician
Volume23
Issue number6
StatePublished - 2020

Keywords

  • Data transparency
  • Diversity
  • Herfindahl index
  • Interventional pain procedures
  • Managerial epidemiology
  • Pain medicine physicians
  • Partial Kendall correlation
  • State Ambulatory Surgery and Services Databases

ASJC Scopus subject areas

  • Anesthesiology and Pain Medicine

Fingerprint

Dive into the research topics of 'Hospitals’ web site lists of their interventional pain procedures inadequately reflect the diversity of their actual pain medicine practices'. Together they form a unique fingerprint.

Cite this