Evidence-Based Practice in LTC

To Test or Not to Test

An overview of the role & value of lab tests in long-term care

Steven Levenson, MD, CMD
Multi-Facility Medical Director, Baltimore, MD
Chair, Caring's Editorial Board

by Charles Crecelius, MD, PhD, CMD
Past President, Missouri Association of Long-Term Care Physicians
Medical Director, Delmar Gardens, St. Louis, MO
Member, Caring's Editorial Board

Common Practice

Many laboratory and diagnostic tests are obtained on nursing facility residents and postacute patients. Many nursing facility medical records contain countless pages of lab test results. But only some of these results are clinically significant, requiring a physician decision or a change in the care plan or orders. Sometimes, even if treatment is feasible, there may be a reason (e.g., advance directives) not to do anything.

Yet the purpose of diagnostic testing may be misunderstood, the rationale for ordering tests may be unclear, and the significance of lab test results may frequently be misinterpreted. Treatment decisions are often based on test results, independent of the "big picture" of the patient's condition.

Many nurses call physicians with test results without being prepared to provide a clinical context or previous results for comparison, or may over-react to lab reports of "panic" values. Facility administrators or directors of nursing may demand that nursing staff call physicians immediately with all (or all abnormal) lab test results, regardless of their urgency or clinical significance.

Mistakes and misconceptions about lab testing persist at all levels of long-term care, despite having been identified in the literature for several decades. Staff may pressure physicians to address test results without correlating them with a patient's clinical condition. It's not uncommon for physicians to order significant treatments in asymptomatic patients based on lab test results alone, despite the limited applicability of such results. Examples include positive culture results, low albumin levels, and elevated white blood cell counts (see Caring, "Evidence-Based Practice in LTC," March 2003). It's also not unheard of for other staff to recommend lab tests to physicians (or even to write orders for tests that aren't based on consulting with the physician) as though they were verbal orders.

Drug blood levels are typically reported by labs as "subtherapeutic" even though the patient may be doing well on the current dosage of medication. Staff and physicians may erroneously interpret this to mean that the patient needs a higher dose of a medication, leading to unnecessary dosage increases that result in significant side effects without a detectable clinical advantage to the patient. (See Caring, "Evidence-Based Practice in LTC," December 2003.)

On the other hand, lab tests may not be ordered in situations when they could be helpful: for example, a basic metabolic profile to help assess individuals with an acute change of condition or who may be at risk for fluid and electrolyte imbalance while receiving both diuretics and ACE inhibitors.

Most nursing facilities have some protocols for lab testing. But physicians are not always involved in writing such protocols. Often, the protocols are based on recommendations made by others, such as professional pharmacy or laboratory consultants. They may be derived from published federal guidelines for lab monitoring related to drug regimen reviews, which originated several decades ago. It may be assumed erroneously that such tests are required, when they are just guidelines. The nursing staff, physician, and consultant pharmacist may not tailor these testing recommendations to the patient's specific needs. Thus, unnecessary lab tests may be obtained while important tests that are not mentioned in regulatory guidelines may be overlooked.

The Evidence

Common reasons for ordering lab tests in nursing facilities are to a) screen or test for new or unsuspected conditions or b) monitor chronic conditions and drug effects.

Screening tests can be defined as laboratory tests performed on admission or at a fixed interval without regard to clinical condition or medication use (see Joseph reference in "The Evidence," below). Common screening tests identified in nursing facilities include a complete blood count, chemistry profile, and urinalysis. Other screening tests may include thyroid function tests, blood glucose, chest x-ray, fecal occult blood testing, albumin or prealbumin, and cholesterol.

Tests to monitor chronic conditions or drug treatment include potassium or electrolytes for individuals on diuretics, blood glucose to monitor diabetes, and prothrombin time for those on warfarin. Examples of drug levels include digoxin, phenytoin, theophylline, phenobarbital, and lithium.

Several studies conducted in nursing facilities with a significant physician presence have reported no significant clinical benefit from routine screening tests. A few studies conducted in typical community nursing facilities have suggested some limited benefit in finding previously unidentified abnormalities that may help establish a diagnosis, but few meaningful interventions that ultimately benefit patients. The overall conclusion of these studies through the years appears to be that screening targeted to clinically pertinent individual risks may be beneficial, but the yield of routine screening tests (for example, on admission and annually) is too low to warrant their routine use.

Lab testing is often indicated in long-term care patients with acute changes of condition, and might actually prevent hospitalization. Periodic lab testing is generally useful in helping monitor patients with specific risks or conditions, or who take medications with known risks that cannot be reliably identified by clinical signs and symptoms. According to one study, tests that are more likely to yield some patient benefit include CBC, electrolytes, blood urea nitrogen and creatinine, glucose, serum thyroxine (T4), and urinalysis.

Regardless, lab testing has been identified as more useful when the results are correlated with a patient's clinical condition based on the history and physical exam. The urgency of addressing abnormal lab test results (whether done for screening or for monitoring someone at risk or with a change of condition) depends primarily on how the patient is doing. Few test results are meaningful enough by themselves to warrant urgent attention. Often, abnormal results actually signify improvement, when compared with previous results.

For example, an elevated BUN and creatinine may reflect an improved hydration status when compared to previous, even more abnormal levels. Or an elevated white blood cell count may indicate that a patient is successfully fighting an infection, while a normal or low WBC may be present despite overwhelming infection (see Caring, "Evidence-Based Practice," July 2003). Or a low drug blood level (often reported as "subtherapeutic") may not require action unless an individual's condition is unstable or declining.

In addition, "normal" lab test reference ranges are based on statistical "normality." Normal ranges typically cover two standard deviations (95% of results) above and below the mean. That means that approximately 5% of test results are actually normal for that individual despite falling outside of a lab's so-called normal ranges. Further, normal ranges are mostly derived for relatively healthy younger patients, while those for older, frailer patients may not be well established. Age-related changes in cardiac, renal, and pulmonary function should be--but are not necessarily--considered in interpreting lab results.

Conclusions & Recommendations

The literature about lab testing and interpreting test results in long-term care has remained relatively constant in past two decades. Most of all, laboratory tests are more useful if they are linked to clinical data offered by the history and physical exam. According to Joseph ("Routine laboratory assessment of nursing home patients." J Am Geriatr Soc. 1992; 40(1):100), "Laboratory testing does have a role in the care of nursing home patients, but not as a substitute for the involvement of a primary care provider...in the clinical course of the patient."

Thus, every nursing facility should use evidence and reliable consensus to guide their policies about obtaining, reporting, and acting on lab test results. They should seek physician input--usually through the facility medical director--to develop and implement such policies and practices.

Failing to obtain appropriate clinical correlation of test results, or simply mandating the immediate reporting of all test results, ignores the evidence and is often counterproductive. Tests should not be performed just for regulatory compliance. Even worse, treatments should not be rendered based on abnormal test results just to demonstrate regulatory compliance. Surveyors should not overstep their authority by suggesting that regulations require lab testing to evaluate patients in various circumstances. But practitioners should consider pertinent lab tests to help them monitor key risks and condition changes in their long-term care and postacute patients.

Deciding Whether to Obtain Tests: The reason for getting a test may influence the urgency of acting upon the result. Therefore, before ordering or requesting lab tests, physicians, nurses, and others should identify the reasons for testing (e.g., screening, diagnosis of causes of an acute change of condition). If abnormal results are not likely to lead to some beneficial action for the patient, then the test may be inappropriate. For example, the patient may be terminally ill, the underlying cause may already be known, or the patient may be unable or unwilling to cooperate with further tests or procedures.

As a general rule, routine screening tests for all nursing facility residents are not advisable. They are most useful when the disease is common, silent, and treatable. Don't order them unless they will affect a clinically meaningful action. Before ordering tests, weigh the direct and indirect value and risks of screening, assess the patient's values and preferences, and estimate the patient's life expectancy and risk of dying from the condition.

For individuals with an acute change of condition or decline in mental status, function, or appetite without evident cause, pertinent lab tests may help establish the diagnosis and may prevent unnecessary hospitalization. Examples of worthwhile tests (depending on the clinical picture) might include a CBC, electrolytes, BUN and creatinine, glucose, T4/TSH, and urinalysis.

Some tests are important in monitoring individuals with specific risks or taking certain medications, regardless of whether they are mentioned in regulations. For example, thyroid function tests may not be indicated as a routine annual screen, but may be relevant to those who are receiving medications such as amiodarone or phenytoin that could affect thyroid function.

Routine chest x-rays are not useful, although a chest x-ray may be indicated to screen for tuberculosis or other pulmonary diseases in at-risk individuals. Don't treat x-ray results in the absence of clinical signs and symptoms. For example, an infiltrate on a chest x-ray in an otherwise afebrile, asymptomatic individual does not prove that the individual has pneumonia or should receive antibiotics, without additional clinical correlation.

Don't obtain urinalyses or urine cultures routinely. The high prevalence of bacteriuria in this population is well known. But the consensus against treating asymptomatic bacteriuria (including recommendations from the Centers for Disease Control and Prevention) should be respected, although it continues to be widely ignored. Culture results alone may demonstrate bacteriuria, but do not prove that an infection is present, without additional evidence (see Caring, "Evidence-Based Practice in LTC," March 2003).

Response to Abnormal Results: Every nursing facility should identify specific circumstances for immediate and more routine reporting of test results to physicians. "Abnormal" does not necessarily mean "clinically significant" and does not necessarily warrant treatment. The urgency of communicating with a physician and the means for such communication should be related to the degree of the abnormality and the individual's current condition.

It may be helpful to adopt protocols to guide staff to triage test results--especially for higher risk medications; for example, monitoring prothrombin time to adjust warfarin doses. Regulations and related surveyor guidance should not be used as the primary or sole basis for managing test results.

It is helpful to identify situations when abnormal lab results are sufficiently significant that they should be discussed with a physician regardless of how the patient is doing; for example, markedly low or elevated potassium, sodium, or glucose. It is also helpful to identify situations where abnormal results may not need urgent reporting; for example, when the tests have been ordered to monitor chronic conditions (such as chronic anemia or chronic renal failure) and the results are similar to or better than previous abnormal results.

If a test was ordered as a routine screen or as follow-up in someone with a change of condition, then a nurse should compare the results with previous ones for the same test before notifying a physician. If the results are abnormal but similar to or better than before, and the individual is stable or improving, then routine (nonimmediate) notification may be feasible. If the results are worse or the individual is not stable or improving, the nurse should still assess the patient and review previous results before contacting the physician.

If a test was obtained to monitor a blood level of a medication, then the nurse should assess the individual's current status and compare the level with any previous results. If the drug level is high, the nurse should discuss the situation with a physician before giving any more doses.

If a drug level is low or in the "therapeutic" range and the individual is stable or improving, then routine notification should suffice. If the individual's condition is unstable or declining, the nurse should generally discuss the situation promptly with a physician, regardless of the blood level of the drug.

A practitioner should decide whether an abnormal result warrants treatment. Physicians should not decide to treat--and nurses or other staff should not pressure them to render or change treatments--without considering the clinical significance of the results and the relevance of treatment to a specific patient. Appropriate physician and nurse documentation can show how results relate to an individual's current condition and demonstrate that the significance of test results was considered appropriately--even if it is decided not to change or add treatment.

It is time for physicians and long-term care facility management and staff to modernize their approaches to obtaining, reporting, and responding to lab test results. The related evidence and consensus have remained consistent over many years. Regulations are relevant to clinical decisions, but they cannot be the primary basis for making them. As with other clinical situations, correlation of diagnostic test results with knowledge of the patient is essential for effective, efficient care.

back to top