Overuse of diagnostic testing in healthcare: a systematic review

Abstract

Background: Overuse of diagnostic testing substantially contributes to healthcare expenses and potentially exposes patients to unnecessary harm. Our objective was to systematically identify and examine studies that assessed the prevalence of diagnostic testing overuse across healthcare settings to estimate the overall prevalence of low-value diagnostic overtesting.

Methods: PubMed, Web of Science and Embase were searched from inception until 18 February 2020 to identify articles published in the English language that examined the prevalence of diagnostic testing overuse using database data. Each of the assessments was categorised as using a patient-indication lens, a patient-population lens or a service lens.

Results: 118 assessments of diagnostic testing overuse, extracted from 35 studies, were included in this study. Most included assessments used a patient-indication lens (n=67, 57%), followed by the service lens (n=27, 23%) and patient-population lens (n=24, 20%). Prevalence estimates of diagnostic testing overuse ranged from 0.09% to 97.5% (median prevalence of assessments using a patient-indication lens: 11.0%, patient-population lens: 2.0% and service lens: 30.7%). The majority of assessments (n=85) reported overuse of diagnostic testing to be below 25%. Overuse of diagnostic imaging tests was most often assessed (n=96). Among the 33 assessments reporting high levels of overuse (≥25%), preoperative testing (n=7) and imaging for uncomplicated low back pain (n=6) were most frequently examined. For assessments of similar diagnostic tests, major variation in the prevalence of overuse was observed. Differences in the definitions of low-value tests used, their operationalisation and assessment methods likely contributed to this observed variation.

Conclusion: Our findings suggest that substantial overuse of diagnostic testing is present with wide variation in overuse. Preoperative testing and imaging for non-specific low back pain are the most frequently identified low-value diagnostic tests. Uniform definitions and assessments are required in order to obtain a more comprehensive understanding of the magnitude of diagnostic testing overuse.

Keywords: health services research; healthcare quality improvement; patient-centred care; quality improvement.

Figure 1

PRISMA Flow-diagram. OECD, Organisation for Economic Co-operation and Development; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Figure 2

Assessment outcomes regarding the prevalence of low-value diagnostic tests for all assessments included in the diagnostic imaging category: (A) cardiac tests, (B) combination, (C) scans, (D) endoscopy, (E) ultrasound and (F) X-ray. Among the included studies, some studies contained multiple assessments undertaken in different cohorts. These assessments are distinguished by the following: (A) assessment performed among a commercially insured population, (B) assessment performed among Medicaid beneficiaries, (C) assessment performed among Medicare beneficiaries, (D) assessment performed using Kaiser Permanente EPIC Electronic Healthcare Records data, (E) assessment performed using data derived from the Oregon Community Health Information Network. BPH, benign prostatic hyperplasia; CDUS, colour duplex ultrasound scan; CKD, chronic kidney disease; ERCP, endoscopic retrograde cholangiography; PSA, prostate-specific antigen; SPECT MPI, single-photon emission CT–myocardial perfusion imaging; TIA, transient Ischaemic attack; URI, upper respiratory infection.

Figure 3

Assessment outcomes regarding the prevalence of low-value diagnostic tests for all assessments included in the other diagnostic tests category: A) laboratory tests and (B) electroencephalography tests. Among the included studies, some studies contained multiple assessments undertaken in different cohorts. These assessments are distinguished by the following: (A) assessment performed among a commercially insured population and (B) assessment performed among Medicaid beneficiaries. CKD, chronic kidney disease; EEG, electroencephalography; PSA, prostate-specific antigen; PTH, parathyroid hormone; T-Hyst, total hysterectomy; TSHR, thyroid-stimulating hormone reflexive testing; V-Pap, vaginal Pap smear.