How Should We Interpret “Abnormal” Lab Results in Older Patients? Ethical and Practical Challenges of Establishing and Verifying Laboratory Reference Intervals in the United States

Abstract

Laboratory reference intervals (RIs) are useful tools that assist healthcare workers with medical decision-making. With current laboratory establishment and verification techniques, older patients can be systematically excluded from the RI process. This in turn can lead to use of improper RIs that may not accurately reflect the normal physiologic variance of aging. Ethically, the use of improper RIs in older patients conflicts with the principles of justice and beneficence. By excluding older patients from laboratory RIs, any effects in this population’s healthcare outcomes would reflect the intrinsic ageist structure of the current RI system. Ways to address this issue include active inclusion of older patients in establishment studies, or establishment of population specific RIs, though each solution has limitations. Use of artificial intelligence and informatics techniques also could prove useful in this endeavor. It is critical that healthcare providers understand these challenges when treating this population.

Introduction

Laboratory reference intervals (RIs) are vital tools in medical decision-making, enabling healthcare providers to interpret analyte-based test results for diagnosing diseases and monitoring treatment responses. It is of paramount importance that these RIs represent an accurate reflection of the patient population receiving these tests, and the disease states for which they are being tested. By their very nature, RIs are problematic since they must delineate what is considered a “normal” (physiologic) analyte level from what is an “abnormal” (pathologic) level in biological systems where there is often a spectrum between these two states (Usher-Smith et al., 2016).

In the early days of laboratory medicine, before the 1960s, laboratories worked independently to develop their own “normal ranges.” These ranges were derived from poorly defined groups of patients labeled “normal,” and were only applicable to specific patient populations and methods. The modern concept of RIs was introduced in the late 1960s by Gräsbeck and Saris (Timbrell, 2024). Following the introduction of the RI concept, expert panels produced recommendations to improve the principles, terminology, and overall quality associated with RIs (Timbrell, 2024).

From 1987 to 1991, the International Federation of Clinical Chemistry (IFCC) published a series of six papers recommending that each laboratory follow defined procedures to produce its own reference values (Ozarda, 2016). In 2008, the IFCC and Clinical and Laboratory Standards Institute (CLSI) published the C28-A3 guideline, titled “Defining, Establishing, and Verifying Reference Intervals in the Clinical Laboratory,” which provided detailed steps for selecting reference individuals, considering pre-analytical and analytical factors, and analyzing reference values (Clinical and Laboratory Standards Institute, 2010; Ozarda, 2016).

RIs are usually defined as the middle 95% of test results in a “healthy” reference population (Horn & Pesce, 2003). Since RIs must reflect local populations, the gold standard for establishing an RI is that each individual laboratory would conduct a study to capture the RI for the specific population they serve. While more achievable for larger firms, this is often not feasible for smaller community laboratories as the required costs and resources are prohibitive. Instead, these laboratories will go through a process to verify RIs established by studies conducted by larger centers to use in their local population (Katayev et al., 2010; Ozarda et al., 2018). Verification of RIs to a local population is a federal requirement for labs who cannot conduct these larger studies, as required by the 1988 revision of the Clinical Laboratory Improvement Act (CLIA), known as “CLIA ’88” (US Centers for Disease Control & Prevention, 2024). Due to the time and resource requirements of larger verification studies, however, the general preference for routine community lab verification is with a small reference group (Katayev et al., 2010). Typically, the laboratory would analyze approximately 20 samples from healthy subjects deemed to adequately represent the local population (Clinical and Laboratory Standards Institute, 2010). The tested values are then compared to the middle 95% RI established from a larger study.

Verification of RIs can pose both practical and ethical challenges for older patient populations (Ozarda et al., 2018; Schuff-Werner, 2018). Several recent studies have demonstrated clinically significant shifts in the RIs of many laboratory analytes for older patients when compared with younger patients (Adeli et al., 2015a, 2015b; Adeli, Raizman, Chen, et al., 2015; Greene et al., 2015; Marriott et al., 2023; Tran et al., 2025). With persistent challenges related to cost, adequate participant recruitment, and difficulties defining reference groups, older patients can be systematically excluded from the verification process. This exclusion could foreseeably lead to use of improper RIs that do not accurately reflect the normal physiologic variance of an older patient. Ethically, this poses an issue with the principle of justice, as under- or overinterpretation of these values can lead to unnecessary medical costs for additional testing and potentially superfluous treatments (Hardie et al., 2002; Moynihan et al., 2012). In the post-CURES Act era (US Department of Health and Human Services, 2020), patients can receive lab values that fall outside of the reference range prior to clarification from their providers. This can lead to potential unnecessary distress for these patients, and more strain on the patient’s care team explaining why a test is labeled as “abnormal” (van Kuppenveld et al., 2020).

Since nonrepresentative RIs could misguide treating teams, we also argue that any effects in this patient population’s healthcare outcomes would reflect the intrinsic ageist and ableist structure of the current RI system due to the systematic exclusion of older reference patients from laboratory RIs. It is critical that healthcare providers understand the challenges of equitable RIs when treating the older population, not only for proper patient care, but also to raise awareness and affect change at a systemic level.

The objective of this perspective paper is to provide discourse on, and raise awareness of, the challenges of establishment and verification of reference intervals in older patient populations. We will highlight both the practical hurdles and the ethical implications inherent in the current RI system in the United States. Furthermore, we seek to engage in a dialog on potential strategies to enhance equity in RI verification, ensuring that the needs of older adults are adequately met in clinical practice. We aim to feature our discussion on the often overlooked, yet equally important, RI verification process in smaller community laboratories.

Ethical and Practical Challenges of Reference Intervals in an Older Population

The process of verifying RIs depends on two subjective challenges for a reference group:

1. Defining a “healthy” reference population

2. Ensuring the reference population adequately represents local demographics

Defining a healthy reference group can be very difficult, particularly for an older population. The presence of health conditions, use of certain medications, and, in some cases, genetic predisposition can alter lab test results and skew RIs, all of which may affect the older population (Adeli et al., 2015a, 2015b; Adeli, Raizman, Chen, et al., 2015; Huber et al., 2006; Nilsson et al., 2003). It is difficult to disentangle these variables. The process often demands additional resources and a larger participant recruitment pool. These variables are far easier to control for in younger reference populations, making them the more reasonable choice for a small verification study from a resource-constrained, community laboratory.

Despite identification of a healthy cohort being required for verification, what is considered “healthy” is itself not well defined (CLSI, 2010; Katayev et al., 2010). It is entirely reasonable for a laboratory to take measures to reduce undue RI variance from influencing medications or disease states. Not doing so could widen the reference range, potentially reducing the overall usefulness of the metric. Of course, this results in the inevitable trade-off: systematic underrepresentation of the population most affected by this natural variance, namely older patients. Without defining inclusive standards, the RI verification system will continue to disincentivize older patient representation.

Improperly reflective RIs pose a practical patient care challenge. These RIs may be difficult to interpret, especially for physicians not trained in geriatrics. For example, significant changes in luteinizing hormone and testosterone levels were identified in men over 70, suggesting age ought to be accounted for when assessing testicular function (Marriott et al., 2023). Similar patterns were found in testing beta-human chorionic gonadotropin (bHCG) and thyroid stimulating hormone (TSH) levels in older patients (Greene et al., 2015; Tran et al., 2025). Were RIs validated for younger patients employed, the analyte levels could be overinterpreted. Under-interpretation is just as much a risk, as there are tests which appear normal by current reference levels but are likely to be considered abnormal by geriatricians. Examples could include mild changes in sodium, C-reactive protein, or total cholesterol levels (Bourdel-Marchasson et al., 2010).

Improperly reflective RIs also pose ethical challenges. An older patient may already be at risk for overdiagnosis of pathology, which could lead to harms from unnecessary medical costs, additional testing, medications, and possibly invasive procedures (Hardie et al., 2002; Moynihan et al., 2012). RIs that, by unintentional design, possibly overcall the normal physiology of aging for pathology unfairly subject these vulnerable patients to consequences, violating the principle of justice.

The second challenge, ensuring the reference population represents local demographics, adds an additional layer of complexity for justice and equity. This is particularly challenging for diverse communities. Verification with smaller reference groups may be the only realistic strategy available to community labs due to their resource constraints. Unfortunately, a reference group of 20 subjects may not adequately capture the diversity of the local population. Further, patient groups with less access to laboratory services are at increased risk of underrepresentation in RI verification, as this diminished access could create a false impression to the lab of the community demographics.

It is not an uncommon practice for labs to recruit their own in-house laboratory staff as members of their reference group. This practice has several advantages, including the potential bypassing of the need for additional resources or IRB approval to recruit members of the community. Of course, this only compounds the problem of appropriate RIs, as laboratory staff tend not to represent general population demographics (Mulder et al., 2024). Verifying appropriate RIs for a diverse older population only compounds the ethical problem of justice and fairness. Not only do improper RIs disproportionately affect older patients, but those at highest risk are also patients from diverse groups that are already underrepresented in and discriminated against by the United States medical system (Williams et al., 2019).

It is important for healthcare providers to understand the challenges of establishing and verifying RIs in their community. When evaluating an older patient with an abnormal laboratory result, providers must take these issues into account to avoid under- or overinterpretation. If there are concerns regarding the appropriateness of RIs to certain populations, it is important for the lab to communicate this information to the treating providers ahead of time.

Ethical Foundations Supporting the Establishment of Reference Intervals for Older Patient Populations

Older patients make up an increasing percentage of the general population and represent a large proportion of healthcare spending in the United States (Caplan, 2023). Though older patients regularly interface with healthcare professionals, geriatric medical training in the United States is inadequate. Little work has been done to educate providers about which aberrancies in medical tests are meaningful and which could be attributed to the physiology of aging (Gleason et al., 2022; Howe & Witt Sherman, 2006; Mezey et al., 2008; Partnership for Health in Aging Workgroup on Interdisciplinary Team Training in Geriatrics, 2014). Additionally, aging is a universal experience which should not be regarded as a pathological state based solely on physiologic fluctuations in laboratory values.

In the United States, since the passage of the CURES act, patients can receive their laboratory results before their provider has reviewed them (US Department of Health and Human Services, 2020). On the patient-facing laboratory report, abnormal results are often highlighted with bright colors. While some patients know what their laboratory results could mean, not all do. These brightly colored “abnormal” lab test could be distressing. For a patient population that makes up such a large proportion of the cases seen in the hospital and clinic daily, laboratory RIs that are significantly impacted by the physiology of normal aging should be re-evaluated. This re-evaluation should be undertaken to exclude the possibility that these RIs are disproportionately affecting the care of a vulnerable population; as was seen with the calculation of eGFR among Black patients seeking kidney transplantation (Eneanya et al., 2019).

The ethical foundations for more appropriate RIs originate in part from avoiding unnecessary patient distress, while also providing the best care for patients. Based on a principlist analysis, the ethical duty to patient beneficence in laboratory medicine would suggest that, if our current system of reporting RIs in this patient population is not entirely accurate, we would be obligated to improve the system if possible. Through a justice lens, in addition to individual patient injustice, the current system could lead to additional unnecessary systemic costs and could stretch laboratory staff thin.

Modern discourse surrounding society’s views of aging have led to emphasis placed on the subtle or overt ageist and ableist structures inherent to our healthcare system (Langmann & Weßel, 2023; Weßel & Schweda, 2023). Such ethical discourse originated from Black feminist academics and civil rights advocates and has since expanded to become a crucial component in the analysis of structural discrimination in medicine and society (Crenshaw, 1989, 1991). In addition to those proposed previously, we argue that the current structure of RIs would serve as another example of subtle ageism and ableism in our current healthcare system (Banerjee et al., 2021; Kokorelias et al., 2023; Lundebjerg & Medina-Walpole, 2021; Rogers et al., 2015). This necessitates action on the part of laboratory directors and clinicians to help make the current system more equitable for their patients.

Taken together, developing age adjusted RIs is an ethical imperative that deserves systematic analysis and multidisciplinary action. Addressing this issue not only helps provide higher quality care to a patient population that continues to grow in the United States, but would also prevent unnecessary harm, promote system-level justice, and remedy their current systematic exclusion all while respecting the value older patients bring to society.

How Challenges of Establishing Reference Intervals for Older Patients Could Be Addressed

There has been significant conversation in recent years surrounding the role laboratory medicine plays in perpetuating inequities, specifically in underrepresented and marginalized communities. One of the most influential outcomes of these debates was the improvement of the estimated glomerular filtration rate (eGFR) calculation which disproportionately impacted Black patients with chronic kidney disease (Delgado et al., 2021; Eneanya et al., 2019; Inker et al., 2021; Levey et al., 1999). This debate has since expanded to include the sex specific RIs that impact transgender patients on hormone replacement therapy (Allen et al., 2021; Bezuidenhout et al., 2022; Humble et al., 2022).

One potential solution could be promoting active inclusion of older patients within RI establishment and verification studies. As mentioned previously, however, these RIs can be influenced by physiologic, pathologic, and iatrogenic variance that increase with age. Inclusion of older subjects in these studies is likely to widen the interval range, potentially reducing the overall usefulness of the value for other populations. That said, since older patients represent an increasing percentage of the population, they tend to be markedly more represented in healthcare treatment and spending and are more vulnerable than the general population. Given this, it may not be an overall negative to skew RIs to better reflect these patients’ needs. Mixing older and younger patient data when establishing these RIs, however, would primarily serve to capture more obvious abnormalities in analyte levels. Under interpretation of otherwise normal analyte values in younger patients would not be well addressed.

There have been multiple recent works advocating for, and trying to establish, specific RIs for certain populations at risk of improper conventional RIs: transgender patients, older patients, and children to name a few (Adeli et al., 2015a, 2015b; Adeli, Raizman, Chen, et al., 2015; Bezuidenhout et al., 2022; Humble et al., 2022). The obvious advantage to this strategy is the establishment of more reflective RIs for these populations, thereby addressing issues related to appropriate patient care, justice, and equity. Additionally, this solution could bypass the issues of reducing the practicality of RIs for other populations and could better address subtle analyte abnormalities in older populations. Of course, this creates a practical disadvantage of increasing cost and resources for establishment of all these different RIs for distinct populations. These disadvantages are further compounded at the small lab level, as each population category of these new RIs would require their own verification process. Further, this strategy does not necessarily address the issue of verification with a reference population. Smaller labs would still be required to choose a pool of 20 “healthy” representative subjects to serve as reference for the verification. More objective criteria for what would be considered healthy would be needed to fully embrace this potential solution. However, it is not currently clear what form these objective criteria could take. Additionally, it is currently unclear what role subjective criteria, such as patient-reported outcomes and health-related quality of life, could play in identifying patients who could qualify as a potential reference patient. Future studies could focus on elucidating these areas.

In discussing how to address the challenges of geriatric RIs, artificial intelligence (AI) cannot be omitted. The advent of AI and informatics approaches in the clinical laboratory have not only dramatically altered the way clinicians can interact with large amounts of data, but they also have opened the clinical laboratory to real-time quality improvement and whole-system trends in population health. Recently, large international laboratories published data analyzing age-and sex-based RI calculations for complete blood counts and biochemical analytes utilizing large datasets (Adeli et al., 2015a, 2015b; Adeli, Raizman, Chen, et al., 2015; Cevlik et al., 2023). These studies revealed that nearly every analyte measurable with clinical laboratory methods exhibited significant variations across age and between sexes, further emphasizing the importance of incorporating these biological variables into RI calculations. AI would be helpful in sorting through this vast data and perhaps find patterns and variables that could assist in the establishment of population specific RIs, or aid in crafting objective criteria for defining “healthy” subject choice.

It is important to mention that the pathologic effects of aging are significantly impacted by socioeconomic factors that disproportionately affect minority communities. When establishing what laboratory changes are considered “physiologic aging,” it is crucial that social determinates of health and demographic data are analyzed to ensure those conclusions accurately represent the patient population the laboratory serves. If such safeholds are not adopted, this well-intentioned change could reinforce the systematic exclusion of these communities.

Conclusion

Laboratory reference intervals are useful tools that assist providers with medical decision-making. Ensuring robust, accurate RIs is critical to their function. However, there are many ethical and practical challenges for both establishing and verifying RIs that adequately represent older, diverse patient populations. It is important for providers to understand these challenges when evaluating a patient with an “abnormal” laboratory result and take this into account as they process its meaning. It is of equal importance that laboratory clinicians continue to capture their local patient population as accurately as feasible during the verification of RIs, accounting for the possibility of underrepresentation where possible. If there are concerns regarding the appropriateness of RIs to certain populations, it is also important for the lab to communicate this information to the treating providers. There are a number of potential ways to address this issue of representation. Some examples include active inclusion of older patients in establishment studies or establishment of population specific RIs, though each potential solution has limitations. Use of AI and informatics techniques could prove useful in establishment of more appropriate RIs for this population.