FormalPara Key Points for Decision Makers

Low-value imaging, imaging not affecting the patients’ further care and treatment, occupies resources in healthcare that could be used for high-value services.

The cost of low-value imaging could amount to billions of US dollars per year.

Measures for reducing low-value imaging are needed at all levels of healthcare.

1 Introduction

Diagnostic imaging is an essential part of modern patient management at all levels of healthcare [1]. The use and expenditures of healthcare services, including imaging, are increasing worldwide [1, 2]. According to the WHO, 3.6 billion imaging examinations are conducted each year, and 250 million of these are of children under the age of 15 years [3]. About 84% of examinations worldwide are conventional radiography and fluoroscopy (CR), while 8% are computed tomography (CT), about 4% ultrasound (US), 3% magnetic resonance imaging (MRI), and 1% nuclear medicine (NM) [4,5,6,7].

According to the Organization for Economic Co-operation and Development, 10‒34% of healthcare spending is wasteful and inappropriate [8]. Correspondingly, 20‒50% of imaging examinations have been reported to be inappropriate or of low value [8,9,10,11]. Low-value care is defined as services that provide little or no benefit to patients, have the potential to cause harm to patients, or waste limited healthcare resources [12]. Diagnostic imaging would be of low value when the examination has little or no impact on the management of the individual patient. Accordingly, it should not be confused with a negative examination result that might be valuable for ruling out serious conditions and preventing further health expenditures. From a societal perspective, low-value imaging constitutes increasing costs, while for the patient, it is an unnecessary risk due to exposure to ionizing radiation and/or contrast media [8,9,10,11]. Low-value imaging can be found across all imaging modalities [13], as well as in several patient groups [14], and is recognized as a major problem [1, 15]. Accordingly, recommendations, guidelines, and other measures have been issued to reduce its use. However, these measures often have a low impact on clinical practice as many barriers to reducing low-value services have been identified [16,17,18,19,20,21,22].

Imaging is a costly health service, and the number of examinations is rising [1]. Furthermore, in many countries, there is a lack of radiologists; thus, imaging could be a bottleneck in the healthcare system [1]. Low-value imaging examinations constitute a risk to patients, can reduce access to high-value care or delay crucial care, and result in poorer outcomes representing substantial opportunity costs.

Therefore, it is necessary to reduce the number of inappropriate or low-value examinations to free resources for high-value patient care. To do so, we need to know the resources used for low-value imaging. Accordingly, the objective of this systematic review was to provide an overview of the cost of low-value imaging worldwide.

2 Methods

This systematic review was conducted based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement. The database search was developed in Medline—Ovid and further adapted to Embase-Ovid, Cochrane Library, and Scopus. The search terms were built from medical subject headings for Diagnostic imaging/Radiology, Health service misuse/Medical overuse, and Healthcare cost. Keywords were used for the concepts reduce/avoid and cost reduction. The complete search strategy and findings log can be found in the Online Supplementary Material (OSM) File 1. Searches were carried out in March 2022 with the last search on 15 March 2022. The search period was from 2012 to 2022 for the 10-year period in coherence with the activity of the choosing wisely campaign. Language filters were used to exclude papers written in languages other than English, German, Danish, Norwegian, and Swedish as these are the languages the authors are familiar with. Keywords were used to exclude studies on animals, mass screening, deep learning, other types of waste, or unnecessary care besides imaging.

2.1 Eligibility Criteria

Primary empirical studies assessing the costs of low-value diagnostic imaging were included. Studies designed as randomized controlled trials, non-randomized trials, cohort studies, cross-sectional studies, descriptive studies, cost analysis and cost-effectiveness analysis, and mixed-methods studies were included. The reference lists of relevant systematic reviews and meta-analyses were hand-searched for additional primary studies eligible for inclusion.

2.2 Selection of Records and Methodological Quality Appraisal

The records were archived using the Thomson Reuters EndNote X9.3.3 library, and duplicates were removed. All remaining records were transferred to Rayyan QCRI where additional duplicates were removed. All authors participated in the title and abstract screening using Rayyan. During the screening, records marked “Maybe” were discussed among the authors to agree on whether to include or reject them. All authors contributed to the full-text review and quality assessment of the studies. Due to the different methodologies of the included studies, the JBI critical appraisal tool was used for methodological quality assessment [23]. Any disagreements during abstract or full-text screening were resolved through discussion and consensus. During the full-text screening, reference lists of included articles were hand-searched for relevant articles. Google Scholar was used for hand searching for eligible papers that cited the included studies.

2.3 Data Extraction and Analysis

Data extraction was completed using a standardized summary table consisting of the following categories: author, title and year, country, design/methods, population, clinical setting, clinical indication for imaging, imaging modality, low-value imaging examination, control or comparator, cost/cost reduction, currency, the year costs were calculated, and cost denomination. All authors contributed to the data extraction. In addition, data extraction was discussed by the research team for quality assurance purposes.

Meta-analysis or comparison between studies or countries was not performed as reported costs are calculated differently in different studies, ranging from modeling to direct cost calculation. Just the cost of the index imaging study is included in the current review. For all results, the monetary value was converted to July 2022 US dollars using online converter calculators (https://www.oanda.com/currency-converter/en/?from=EUR&to=USD&amount=1) to provide an overview of costs in general. In studies where the year of cost calculation was not explicitly stated, we assumed the year costs were calculated was the same as the publication year.

3 Results

A total of 5,567 records were identified through database searches. 1,985 duplicates were removed, and 3,582 titles and abstracts were screened (Fig. 1). Hand-searching techniques resulted in the screening of 18 additional full-text records. In total, 166 reports were assessed for eligibility in full text, and of these, 60 records were excluded during full-text screening (OSM File 2 gives an overview of reasons for exclusion). In total, 106 studies were included in this review.

Fig. 1
figure 1

PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram of selection of studies

The quality assessment, based on the JBI critical appraisal tool, resulted in no exclusion based on methodological issues, even though the quality of studies varied. Five types of checklists were used: RCT, cohort study, analytical cross-sectional studies, cost analysis, and quasi-experimental appraisal tools, depending on the design of the appraised study. The result of the quality appraisal is available in OSM File 3.

3.1 Characteristics of Included Studies

Table 1 presents the characteristics of the 106 included studies. Most studies employed a retrospective chart review/cohort design (n = 76), while six used a cross-sectional design. Another six studies were cost-effectiveness or cost analysis studies using modeling. The last 18 studies used different methodologies, from randomized controlled trials (n = 1) to mixed methods (n = 1). Most of the studies were conducted in the USA (n = 70), 17 studies were conducted in European countries, seven in Canada, five in Australia, and seven in other parts of the world, mainly Asian countries. The settings of most included studies were hospitals or medical centers (n = 82), while eight studies used multiple settings, and six were from an emergency department setting. Other settings included primary care/hospice (n = 4), intensive care unit (n = 4), and imaging center (n = 2).

Table 1 Overview of setting and characteristics of the included studies

Sixty-eight studies assessed a single imaging modality, CR (n = 32), MRI (n = 15), CT (n = 11), US (n = 8), and NM (n = 2), while 38 studies assessed the costs of multiple imaging modalities.

3.2 Cost of Low-Value Imaging

Sixteen studies reported on the aggregated costs of low-value imaging or possible annual cost reduction, illustrating the amount of healthcare resources spent on imaging without impact on patient management. An overview of the individual studies is presented in Table 2. The possible cost reduction or costs reported varies between setting and patient groups. However, the possible redistribution of resources would be worth billions of dollars globally.

Table 2 Overview of annual spending on low-value imaging and possible annual savings if reducing low-value imaging in different settings and patient groups

Based on the reported cost per examination of the included studies an overview per modality is presented in Table 3, both in terms of overall cost per modality and with examples of specific patient groups. Costs vary based on cost level in different countries and imaging techniques, for example, in MRI number of sequences and use of contrast media [22, 24, 26,27,28,29,30,31,32,33,34, 37,38,39,40,41,42,43,44, 46, 47, 50,51,52,53, 55,56,57,58, 60, 61, 63,64,65,66,67,68,69,70,71,72,73,74,75, 77,78,79,80,81,82,83,84,85,86,87,88,89,90,91, 95, 96, 99, 102, 104, 106,107,108,109,110, 112,113,114, 117,118,119,120,121,122,123,124,125,126].

Table 3 Cost per imaging examination for low-value imaging per modality with some specific patient groups as examples

Low-value CR imaging in the included studies was mostly routine use of musculoskeletal or chest x-rays (CXR) in pre-/post-operative follow-up examinations. However, mammography, angiography, gastrointestinal imaging, and bone density scans used routinely or in follow-up were also included [22, 24, 26,27,28,29,30,31,32,33,34, 37, 38, 40,41,42,43,44, 46, 47, 50,51,52,53, 91, 99, 102, 104, 109, 110, 113, 119, 122,123,124,125].

For CT and MRI, oncology, orthopedic, cardiovascular, and neurology patients were the dominating patient groups included, with routine examinations or imaging non-compliant with guidelines dominating the findings [55,56,57,58, 60, 61, 63,64,65, 74, 75, 77,78,79,80,81,82,83,84,85,86,87,88, 91, 95, 96, 99, 102, 104, 106,107,108,109,110, 112,113,114, 117,118,119,120,121,122,123,124,125,126].

In US, examinations in oncology and cardiovascular diseases were most common amongst the included studies [66,67,68,69,70,71, 73, 91, 95, 96, 104, 109, 112, 118, 119, 121, 123], while in NM, examinations for oncology patients were most common [89, 90, 95, 104, 108, 110, 111, 118, 120, 122, 123].

3.3 Possible Cost Reduction Through Initiatives in Practice

Four studies, including cost-reduction calculations of interventions to reduce the use of low-value imaging, showed that costs of lower back pain imaging could be reduced by 95% in Australia [115, 116], 73–83% in the USA [128], and up to 16% in Belgium without harm to patients [97]. When introducing an intervention to reduce the routine/daily use of CXR in the ICU, four studies showed a cost reduction of 20‒66% without affecting quality of care [25, 35, 36, 54]. Furthermore, the reduction of preoperative routine CXR reduced costs by 88% in one study [45]. For X-rays of ankle injuries, two studies found that using the guidelines could reduce costs by 7‒49% [48, 93]. One RCT study examining a simplified radiography follow-up regime after stable fracture fixations demonstrated a cost reduction of 65% [49]. Four studies, two in a hospital/Emergency Department setting [100, 105] and two specifically in trauma patients [62, 92] assessed the use of health information exchange to reduce imaging examinations. In trauma patients, a 37‒47% cost reduction was shown, while the changes in cost were small in the overall use of imaging, as some examinations increased in use while others were reduced. Other studies on various patient groups showed 34‒70% cost reduction [59, 94, 127, 128].

4 Discussion

This study is the first systematic review of the costs of low-value imaging globally. Given the universal challenges with resource allocation, this represents a vast opportunity cost and a great potential for providing high-quality services and improving the quality of care. Variation in access and use of imaging is found between and within countries and regions, indicating overuse in some areas and underuse in others. Low-value imaging is most common in areas with easy service access [129,130,131,132,133,134]. Thus, the potential opportunity cost and potential for improvement of the quality of services will vary.

Earlier research has identified specific clinical indications where low-value imaging is more prominent, such as atraumatic pain and routine imaging in minor head injuries, urolithiasis, trauma, thrombosis, and follow-ups [14]. This review demonstrates that some initiatives for reducing low-value imaging yield substantial cost reduction, as costs for low-value lower back pain imaging could be reduced by 95%, and preoperative routine CXR costs could be reduced by 88%. At the same time, other initiatives demonstrated less cost reduction (< 20%). Thus, some types of imaging have a higher potential for reducing costs, or there are differences in the success rate of the chosen intervention. The right combination of imaging to target and type of intervention can free resources for other high-value services, both in imaging and in other parts of the health services, by reducing cascaded unnecessary care [135]. This will improve the quality and efficiency of care and could help reduce healthcare emissions [136].

A systematic review of measures for reducing low-value imaging found that most initiatives focus on implementing guidelines and training referrers to reduce the number of inappropriate referrals to imaging [137], and that most initiatives target musculoskeletal, neuro, and vascular imaging. Several of the initiatives succeeded in reducing the use of imaging, at least for a short time. However, initiatives may need to target other parts of the health services than the referrer alone. To create sustainable change, it will be important to target low-value health service drivers broadly. As shown by Landon et al. [138], drivers of low-value services are complex and can be found at all levels: hospital, provider, and patient. Financial incentives, culture in the medical community, and intensity of care are especially strong drivers. Hence, measures addressing drivers on all levels of healthcare and imaging services are needed to tackle the challenge of low-value imaging.

Furthermore, some examinations are more resource intensive. For example, MRI and NM represent 3% and 1% of the total imaging volume, respectively. However, they have the highest overall costs per examination. This can be explained by the high costs of these examinations and the status and imperative of hi-tech imaging [139, 140]. Therefore, reducing low-value utilization of such examinations may result in great resources freed for high-value care.

4.1 Strengths and Limitations

As with all systematic reviews, this study depends on the content and quality of the included studies. According to our assessment, the quality of the studies varied, and results reported are thus of variable quality as well. In addition, the methodology used in the studies for calculating costs varied greatly. While some studies reported on thorough cost analysis, many reported costs as part of an overall objective, for example, estimates on cost reduction from a measure implemented at the hospital. Moreover, many different sources of cost data were used, such as direct and indirect cost calculations, fees, price lists at hospitals, or reimbursement amounts, permitting a large variation in reported costs. However, cost levels vary between healthcare systems and countries worldwide, which may be mirrored in the included studies. Furthermore, countries with underuse are possibly under-represented in this study but are also much less likely to have high rates of low-value care. Correspondingly, cost estimates are based on studies from countries that have documented higher rates of low-value imaging. The bias in the publications on cost may represent a “natural weighting” of the extent of low-value imaging as there are more publications from areas that are more concerned about low-value imaging and its opportunity cost.

As pointed out in the Methods section (Sect. 2), we estimated the year of cost calculation (publication year) for studies lacking adequate information. This may not be correct. However, the resulting errors may be small, as the yearly cost changes are small. For the proportion of low-value imaging, there may be relevant differences for the various modalities. Further, analysis is not possible based on this data set, however points to knowledge gaps urgently needing to be filled by further research.

Finally, a strength of this study is the use of the costs of specific low-value imaging examinations and not for imaging in general. Thus, the costs of low-value examinations were in all probability identified through this search strategy. In addition, while low-value care is a well-defined area in general, the efforts to reduce low-value healthcare have been hampered by the lack of consensus around a single definition for “value.” For example, imaging can be seen as valuable to calm people’s anxiety, i.e., due to its anxiolytic effect. In this review we have only considered direct, intended, and documented health effects as valuable. Moreover, we have, as has been suggested elsewhere [141, 142], included economic values in assessing low-value care.

5 Conclusions

This review documents the costs of low-value imaging amounts to billions of dollars worldwide and demonstrates vast opportunity costs and great potential to improve the quality and efficiency of health services. This potential will be greatest in areas with high utilization of low-value imaging. Drivers of low-value imaging should be addressed at all levels of healthcare to reduce low-value imaging and expedite access to high-value imaging. Hence, more knowledge is needed on costs, specific drivers, and effective measures for reducing low-value imaging and improving the quality of care.