Figure. Number of laboratory tests per encounter. Compared are the encounters with recent off-site tests with those without, before and after the implementation of the health information exchange.
Hebel E, Middleton B, Shubina M, Turchin A. Bridging the Chasm: Effect of Health Information Exchange on Volume of Laboratory Testing. Arch Intern Med. 2012;172(6):517-519. doi:10.1001/archinternmed.2011.2104
Author Affiliations: Harvard Medical School and Clinical Informatics Research and Development, Partners Healthcare Systems, Boston, Massachusetts (Drs Hebel, Middleton, and Turchin); Divisions of General Internal Medicine (Dr Middleton) and Endocrinology (Drs Shubina and Turchin), Brigham and Women's Hospital, Boston.
Sharing of patient information between health care providers, including through health information exchanges (HIEs), has been proposed as one of the essential changes to improve the quality and efficiency of the health care system in the United States.1 It has been estimated that HIEs could decrease health care costs across the country by approximately $78 billion annually.2 Despite numerous potential advantages of HIEs, there are few studies documenting their benefits.3 This lack of objective information might have slowed down their acceptance.4 Studies that demonstrate tangible evidence of benefits provided by HIEs are urgently needed. Provider surveys show that reduction in duplicate testing is one of the most commonly expected benefits.5,6 We therefore investigated whether the introduction of an HIE between 2 academic medical centers was associated with a reduction in volume of laboratory testing.
We conducted a retrospective study to investigate whether the availability of laboratory test results from a nonencounter hospital reduced the number of subsequent laboratory tests at the encounter hospital. The institutional review board at Partners HealthCare approved the study.
All new outpatient consultations at 2 affiliated academic hospitals between January 1, 1999, and December 31, 2004, were studied. Encounters during the year 2000 when an internal HIE was being rolled out were excluded. We also excluded patients hospitalized for 10 days or fewer after the index encounter and patients with tests in both hospitals prior to the encounter.
A single new consultation encounter—the index encounter—served as a unit of analysis. The number of laboratory tests performed until the end of the day of the index encounter at the same institution as the encounter (postencounter on-site tests) served as the primary outcome variable. Presence of laboratory tests performed at either of the institutions during the 7 days prior to the index encounter and whether the encounter took place before (1999) or after (2001-2004) the HIE rollout served as the primary predictor variables.
A multivariable Poisson regression model was used to evaluate the effect of the recent preencounter tests on the number of postencounter on-site laboratory tests while covariates were adjusted for.
We identified 122 771 patients between January 1, 1999, and December 31, 2004. We excluded 5146 patients who were admitted to the hospital 10 days or fewer after the index encounter and 19 patients who had tests in both institutions during the week prior to the encounter. The remaining 117 606 patients were included in the study.
Of the 346 study encounters with recent off-site tests, 44 took place prior to HIE rollout. Among the 117 260 encounters without preceding off-site tests, 21 968 took place prior to HIE rollout. Patients with recent off-site tests had a mean (SD) of 22.07 (30.13) tests prior to the index encounter. Patients without recent off-site tests had a mean (SD) of 1.62 (10.61) tests prior to the index encounter. Most of the encounters without off-site tests (110 110 or 93.9%) did not have any tests in the preceding week.
In univariate analysis, the number of laboratory tests performed after encounters that had recent off-site laboratory tests decreased by 49% after introduction of the HIE (Figure). In multivariable analysis using a Poisson regression model adjusted for the patient demographics, Charlson Comorbidity Index, site of the encounter, season, encounter year, and the number of prior tests at the encounter and nonencounter institutions, the number of tests after the encounters with prior off-site tests decreased by 52.6% (95% CI, 16.6%-73.1%) after the electronic medical record integration (P = .01).
The number of postencounter tests increased by 2.5% for each point increase in the Charlson Comorbidity Index (P < .001), and it rose up to 51.7% with every subsequent year (P < .001). The number of tests decreased by 0.84% for every $10 000 increase in the patient's median household income (P < .001). It was also 9.06% lower for the patients on Medicaid (P < .001) compared with patients with private health insurance.
In this large retrospective study we have demonstrated that the introduction of an internal HIE was associated with a significant decrease in the number of laboratory tests ordered for patients new to the provider when recent laboratory results were available from another institution. Importantly, our results indicate that the reduction in laboratory tests may be as high as 50%. This could potentially translate into significant savings in settings where patients frequently receive care at multiple institutions.
Our research therefore confirms the hypothesis that having access to the patients' laboratory test results influences the decision process in regard to ordering further tests, which supports the predictions of financial savings made in the HIE cost-benefit models.2 Further studies are required to evaluate the impact and direct financial savings associated with sharing other health information, including imaging studies, physician notes, and discharge summaries.
Correspondence: Dr Turchin, Partners HealthCare, 93 Worcester St, Wellesley, MA 02451 (firstname.lastname@example.org).
Author Contributions: Dr Hebel had full access to all the data in the study and takes responsibility for the integrity of the data and accuracy of the data analysis. Study concept and design: Hebel, Middleton, and Turchin. Acquisition of data: Hebel and Turchin. Analysis and interpretation of data: Hebel, Shubina, and Turchin. Drafting of the manuscript: Hebel. Critical revision of the manuscript for important intellectual content: Hebel, Middleton, Shubina, and Turchin. Statistical analysis: Hebel and Shubina. Administrative, technical, and material support: Hebel and Middleton. Study supervision: Middleton and Turchin.
Financial Disclosure: None reported.
Funding/Support: This research was funded in part by scholarship grants from the Fulbright Commission–Institute of International Education (Dr Hebel) and the National Commission for Scientific and Technological Research, Chile.