Category: Procedure Areas

Innovations in technology have given us better and more powerful medical equipment but have also increased the cost of those devices. A reader recently asked when is it better to purchase medical equipment outright as opposed to pay-per-use financing? As always, the answer is… it depends.

When purchasing a high-cost piece of equipment outright, the hospital (or medical practice) either pays for the entire cost upfront or pays for it in installments. It is like buying a new car – you can either pay cash at the time of purchase or you can finance it over a period of a few years. A second method of acquiring that same piece of high-cost medical equipment is by a pay-per-use contract. In this acquisition model, the manufacturer lends the hospital the equipment for free and instead charges the hospital each time that equipment is used. There are situations when outright purchase is better and there are other situations when pay-per-use is better. For this post, I’ll use the example of a surgical robot but the principles apply to any high-cost piece of medical equipment. The total cost of a surgical robot system varies depending on model, price negotiation, and types of robotic arms used. But we’ll assume a fairly typical $2 million purchase price, $2,000 per operation consumables (eg, robotic arms), and $180,000 annual service contract.

First, create a pro forma

In hospital financing, a pro forma is a document that projects the total costs and total revenue from a proposed new service or piece of equipment purchase over time. Because of the high cost of a surgical robot, no hospital will purchase one just because a surgeon goes to the CEO and says the justification is “Because I want one“. Instead, the hospital is going to want to know whether in the long run the hospital is going to make money or lose money on the surgical robot. However, when I look at a pro forma, it is not just the dollars and cents that are important. There can be non-monetary benefits that can justify purchase of a surgical robot, even if the hospital is not going to make money on the robotic procedures. For example, shortened time in the operating room per surgical procedure, shorter patient recovery vs. non-robotic surgery, fewer surgical complications, improved patient satisfaction, attraction of new patients who prefer to have robotic surgery, etc.

When drafting a pro forma for outright purchase of a medical device such as a surgical robot, you need to look at all of the costs. This includes the purchase price of the equipment, the cost of consumables (such as the surgical arms that can only be used for a fixed number of surgeries), the cost of a service contract, and the cost of personnel. When possible, convert non-monetary factors into dollar-equivalents – for example, determine the cost per hour of an operating room and then include cost savings of a shorter OR procedure time with a robotic versus non-robotic operation. Conversely, if the robotic surgery will take longer than a non-robotic surgery, include this as a cost rather than a cost savings. The type of physicians and the types of procedures will need to be estimated. For example, if the surgical robot will be used by general surgeons, colorectal surgeons, gynecologists, urologists, and ENT surgeons then the relative costs of each of these different types of robotic surgeries needs to be included. Other costs to be factored in can include the cost of any building renovation required to accommodate the new equipment, the cost of training personnel to use the new equipment, the cost of insuring the new equipment, the cost of advertising the availability of the new equipment, etc. The service contract costs must be factored in as well. If a surgical robot is used 18 times per year and the service contract is $180,00 per year, then that works out to an expense of $10,00 per case. On the other hand, if the surgical robot is used 500 times a year, the expense is only $360 per case.

The revenue from a surgical robot will require an estimation of the volume of surgeries performed each year and the expected reimbursement. Reimbursement needs to be stratified by all of the different procedures that will be performed, for example, a robotic cholecystectomy vs. a prostatectomy vs. a hysterectomy. This can be pretty complicated because the reimbursement for an operation such as a cholecystectomy can vary depending on whether it is being paid for by Medicare, commercial insurance, Medicaid, or self-pay (note: self-pay usually translates to no-pay).

Central to a pro forma is the concept of depreciation. This is the expected number of years of life of that piece of equipment before you have to buy a new one. Medical equipment is often depreciated over 5 years. For example, a surgical robot that costs $2 million to purchase and is expected to last 5 years can be depreciated over those 5 years at $400,000 per year. In this example, the pro forma should include tables for each of the 5 years of depreciation. If the hospital projects doing 500 robotic cases per year, then the capital equipment costs would be $800 per case ($400,000 per year ÷ 500 cases per year).

Once you create a pro forma for outright purchase, you then need to create a pro forma for pay-per-use financing. Include all of the costs and revenues you used with the outright purchase pro forma but instead of equipment cost depreciated over the depreciation period (eg, 5 years), include the total annual pay-per-use costs over those 5 years. It is essential to clarify whether servicing the equipment is included in the pay-per-use contract. Usually it is but if the hospital is required to purchase a separate $180,000 per year service contract, then the financial advantages of pay-per-use acquisition can disappear.

Hospitals are unique environments compared to business and manufacturing. Proposals for capital equipment purchases should not be finalized until they have been evaluated and approved by the biomedical engineering department, the infection control department, the staff responsible for structural engineering, radiation safety personnel, etc.

Variables affecting the decision to purchase outright vs. pay-per-use

The decision of whether to buy a piece of medical equipment or pay-per-use is much more complex for a hospital than the decision of whether to acquire a piece of manufacturing equipment by outright purchase or pay-per-use for a factory. There are variables that are inherent in healthcare that do not exist in manufacturing and these variables can have a profound effect on how to pay for a new medical device. Here are some of the variables that the hospital must factor in to the decision-making process.

1. The number of physicians who will use it. If only one surgeon or one physician group will use a piece of equipment, then there is the risk that if that surgeon or group leaves the hospital to practice elsewhere, then the hospital could be stuck with an expensive device that just gathers dust in a closet. Just like it is risky to put your entire retirement investment portfolio in a single stock (as opposed to a mutual fund), it is risky to base the entire pro forma on the equipment’s use by just one physician.
2. The number of specialties that will use it. In the example of a surgical robot, it is far better to buy a robot that will be used by general surgeons, cardiothoracic surgeons, urologists, ENT surgeons, and gynecologists rather than just one of these specialties alone. This ensures that the device can be used every day, Monday through Friday, all year. This overcomes specialist slow downs due to medical conferences, outpatient clinic days, procedure seasonality, etc.
3. The hospital’s cash flow and budget. A $2 million purchase for a surgical robot can wipe out most of a small hospital’s annual new equipment budget. If the hospital lacks sufficient cash to purchase an expensive piece of equipment, then a pay-per-use model may be preferable since there will be an immediate return on investment. Otherwise, it could take several years of use to generate enough revenue to cover the cost of outright purchase.
4. Anticipated volume. If you anticipate a relatively low procedure volume each year, then it could take many years before you recoup your return on investment for the purchase and you’d be better off with the pay-per-use model. The more procedures you can do, the more likely you will be better off purchasing equipment outright.
5. Expectation of a new model in the near future. Medical equipment manufacturers usually keep dates of release of new models of their equipment secret until the last minute (sort of like Apple staying mum about new iPhone models until they are ready to be released). Nevertheless, a little detective work can give you an idea of whether a new version is on the future horizon. If so, you are often better off with pay-per-use initially and holding off on purchasing until the new model comes out. This also holds if you would be happy with the older (current) model since manufacturers will generally discount them to clear out their inventory once a new model comes out. The surgical robot model costing $2 million this year might drop to $1.5 million next year when the next model is released.
6. Commitment by the physicians. I got burned several years ago when our gastroenterologists insisted that we needed to start doing endoscopic ultrasound pancreatic biopsies. We spent a half million dollars on new endoscopic equipment that could only be used for those procedures and also invested in cytology telemedicine so that cytopathologists at another hospital could read the needle aspirates real-time during the procedure. Five years later, the gastroenterologists had not done a single endoscopic ultrasound procedure at our hospital and we basically wasted the money. If there is any uncertainty about whether the doctors will use the equipment, then a per-use model (at least at first) is preferable until the doctors prove that they will actually use it.
7. Procedure payer mix. A manufacturing company can determine the price it will charge for a product and base it’s pro forma on just one sales price. In medicine, however, the hospital gets paid different amounts by different payers for doing any given procedure or service. This means that creation of an accurate pro forma requires the hospital to not only project the total annual volume of procedures to be performed with a new device but also the projected payer mix for those procedures and the financial margin for each payer. Reimbursement from Medicare and Medicaid is fairly easy to project since they are fixed by CMS. However, each commercial insurance company will reimburse different amounts for any given procedure, depending on the hospital’s negotiated contract with that insurance company. Imagine the complexity of a manufacturer who projects that by installing new factory equipment, it can make and sell widgets. But sales contracts dictate that 30% of customers pay $10 per widget, 15% of customers pay $5 per widget, 25% of customers pay $18 per widget, 15% of customers pay $27 per widget, and 15% of customers don’t pay anything and get their widgets for free. In general, per-procedure hospital reimbursement is highest for commercial insurance, a bit lower for Medicare, lower still for Medicaid, and negligible for self-pay. We once had a surgeon who specialized in surgically implanting very expensive medical devices. When we did the initial pro forma, it looked like the hospital would net a small profit each year on the procedures. But after a couple of years, we noticed we were losing tens of thousands of dollars. It turned out that the surgeon was performing implants on commercially insured patients at a private hospital in town and only operating on Medicaid and self-pay patients at our hospital.
8. Non-monetary benefits. The decision of whether to purchase a piece of equipment outright or utilize a pay-per-use financing should not depend solely on the expense vs revenue columns on a pro forma. There can be non-monetary benefits that the hospital may value, even if the new equipment does not increase revenue. These can include attraction of new patients, improved patient satisfaction scores, reduced mortality, reduced complication rates, shortened operative times, etc. It is also important to keep the doctors happy because if the physicians really want to use a new piece of medical equipment and the hospital won’t buy it, those physicians will leave to go practice at another hospital that will buy the equipment. In this case, a pay-per-use acquisition model may allow the hospital to keep the doctors happy while eliminating or at least minimizing financial loses.
9. Connectivity. In an increasingly electronically interconnected world, the ability of medical equipment to connect to the monitors, electronic medical record, scheduling software, and billing software is essential. If there is concern about electronic compatibility, then pay-per-use might be a better option until optimized connectivity issues can be resolved.

The bottom line: its complicated

All too often in hospitals, the person who is the most eloquent, loud, or otherwise persuasive is the one who most heavily influences purchasing decisions. And this person is usually a powerful, silver-tongued physician. The hospital’s best defense against undue influence is the requirement to create a pro forma. This can guide the hospital about whether it is better in both the short-term and the long-term to purchase an expensive piece of equipment outright or utilize a pay-per-use acquisition model. One hospital may find that outright purchase is preferable whereas another hospital in the same town may find that pay-per-use is preferable. An accurate and well thought out pro forma is like a vaccination against future regret. No big-ticket equipment purchase should be put on the hospital’s final annual budget without one.

March 12, 2024

When interpreting pulmonary function tests (PFTs), a patient’s test results are compared to a group of normal people in order to determine that patient’s percent predicted value for each result. Those normal values are based on a person’s age, height, and gender. Until recently, we also used race as one of those variables – but should we?

Height, age, and gender

Height is an important variable in PFT interpretation. Taller people have larger lungs than shorter people. In pulmonary function testing, the overall size of the lungs is measured by the total lung capacity, which is the total volume of air in the lungs. If a patient’s total lung capacity is below the 5th percentile of normal people of a similar height, that patient is considered to have restrictive lung disease. Interstitial lung disease, muscle weakness, pregnancy, and chest wall deformities are some of the common causes of restrictive lung disease. If we did not stratify normal populations by height, then shorter people would be incorrectly diagnosed with restrictive lung disease, which could trigger unnecessary expensive diagnostic testing to determine the cause of the restriction. The graph below shows the relationship between a normal person’s height and their total lung capacity. Normal for a 170 cm (5′ 6″) man is 6.0 liters whereas normal for a 190 cm (6′ 3″) man is 8.0 liters.

Age is another important variable in PFT interpretation. As a child grows older (and also taller), the child’s lung volume increases. But after adulthood, our lung become smaller as we age. For example, the average forced expiratory volume in 1 second (FEV1) at age 30 is 4.4 liters but the average FEV1 at age 70 is 3.2 liters.

We define obstructive lung disease as an FEV1/FVC ratio below the 5th percentile of normal. Once again, we see that what we consider to be normal changes with age. In addition, we see that there are also gender differences in what constitutes normal values. As people age, the normal FEV1/FVC decreases and for any given age, men have a lower FEV1/FVC than women. So, for example, the FEV1/FVC that we would define as obstructive lung disease for a man at age 20 would be 0.73 but for a man at age 80 would be 0.62. A 20-year-old woman would have obstructive lung disease with an FEV1/FVC of less than 0.76 whereas a 20-year-old man would not be considered obstructed until his FEV1/FVC is less than 0.73.

Race and ethnicity

In the past, race was also factored into the determination of normal PFT values. Not for nefarious reasons but for the simple fact that when large numbers of normal people were tested, there are racial differences in the average PFT values after age, gender and height were all accounted for. Last year, the American Thoracic Society recommended eliminating race in PFT interpretation because of a concern that it implied biological differences between people of different races and ethnicities but those changes are actually due to social and environmental factors as well as structural racism. So, should we stop using race and ethnicity in defining normal values in pulmonary function test interpretation?

What is a person’s race, anyway?

My grandmother’s mother was White and her father was Chinese. She was not allowed to attend Atlanta public schools because in the eyes of the school board she was considered Chinese if one of her parents was from China. Genetically, she was just as much White as she was Chinese. Similarly, if you ask 1,000 Americans if Barack Obama is White or Black, 999 of those Americans will say he is Black. His father was from Africa but his mother was Caucasian of Irish ancestry. Like my grandmother, he is genetically just as much White as he is Black. The point is that we assign race labels using social criteria as much as (or more than) true ancestral heritage. When those large numbers of healthy people were tested to determine normal values for PFTs, they were grouped by self-reported race and not by sending gene tests off to 23andMe to determine their genetic ancestry. In the melting pot that is the United States, the vast majority of us have very complex ancestry. Lumping everyone into categories of White, Black, Asian, Hispanic, Native American, etc. is often arbitrary and not very accurate.

But differences do exist…

There are clearly problems fitting many people into one specific race or ethnicity group. But for those people who do self-report themselves belonging to one group or another, there are racial and ethnic differences in normal values. The NHANES III normal values are most commonly used in pulmonary function test interpretation. The NHANES III data set indicates that for a 5′ 9″ 65-year-old man, the average forced vital capacity (FVC) is 4.60 liters for normal Whites, 4.52 liters for normal Hispanics, and 3.87 liters for normal Blacks. There are similar racial differences for women. The normal values were obtained by testing a large number of healthy non-smokers. A valid criticism of the NHANES III data set is that it stratified people into only three groups: White, Black, and Hispanic; other racial and ethnic groups were not included.

A second commonly used data set of normal pulmonary function test values is the Global Lung Initiative (GLI) that stratified people into five groups: Caucasian (Europe, Israel, Australia, USA, Canada, Mexican Americans, Brazil, Chile, Mexico, Uruguay, Venezuela, Algeria, Tunisia), Black (African American), Northeast Asian (North China & South Korea), Southeast Asian (South China, Taiwan, Hong Kong, and Thailand), and other/mixed. Using the GLI data set for a 65-year-old man who is 5′ 9″, the average normal FVC for Caucasian is 4.30 liters, Black 3.63 liters, Northeast Asian 4.13 liters, Southeast Asian 3.82 liters, and other/mixed 3.96 liters.

So, for any given height, gender, and age, normal people who identify as being Southeast Asian or Black have lower lung volumes than those who identify as being Northeast Asian, White, or Hispanic. But that does not mean that race biologically caused those differences. Instead, race and ethnicity may merely correlate with other factors that affect lung volumes.

Factors that correlate with race and ethnicity

When differences exist between people of different races or ethnicities, it does not necessarily mean that those differences are caused by the person’s race – it usually means that those differences are correlated with the person’s race. There is a big difference between causality and correlation. For example, say you are studying the incidence of vitamin D deficiency and you find that Whites in Norway have a higher rate of vitamin D deficiency than Hispanics from Guatemala. Race did not cause the vitamin D deficiency – living at a high latitude with little sunlight caused the vitamin D deficiency. Race merely correlated with vitamin D deficiency because of Hispanics in Guatemala live at a lower latitude than Whites in Norway. Here are some of the factors affecting pulmonary function that correlate with (but are not caused by) race and ethnicity.

Genetics. Lung volumes can be affected by a person’s genes. Just like all of the members of a family might have big ears, all of the members of a family might have big lung volumes. We often define race by skin color. But race is a poor surrogate for genetics and there is no good reason to believe that the genes that determine the amount of pigment in a person’s skin should also dictate the size of a person’s lung volumes.

Socioeconomic status. When English physician John Hutchinson invented the spirometer in 1846, he used it to show differences in lung function between people in different professions, which was a crude estimate of socioeconomic status. There are a lot of reasons why people from lower socioeconomic groups could have lower lung volumes than those from higher socioeconomic groups. Crowded living conditions can lead to more frequent childhood respiratory infections and greater exposure to environmental tobacco smoke. Air pollution in lower class residential areas can affect lung function. Poor nutrition in childhood can have a profound impact on both height and lung function in adulthood. Inadequate treatment of childhood asthma due to limited access to healthcare in childhood can result in lower pulmonary function values in adulthood. When we identify racial and ethnic differences in many health metrics, often what we are really identifying is the socioeconomic differences in those metrics, with race just being a reflection of socioeconomic status.

Maternal health. Maternal smoking during pregnancy, lower birth weight, and premature birth can all affect lung development in infancy. There are significant racial differences in access to maternal healthcare that can impact a child’s lung function.

Obesity. Body weight is not used as a demographic variable in PFT interpretation but obesity can have a profound effect on lung volumes, causing them to be lower. African American adults have a high prevalence of obesity (38.4%) compared with Hispanic American adults (32.6%) and White American adults (28.6%).

Occupation. Certain occupations can affect lung health and thus lung function. Exposure to airborne chemicals, toxins, and dusts can impact lung volumes and flow rates. These are often lower-paying occupations that disproportionately employ workers from minority races and ethnicities.

Altitude. People who live their entire lives at higher altitude have higher lung volumes than those who live at lower altitudes. Studies of inhabitants from Peru, Korea, and Tibet have found that people living at high altitudes have higher values for forced vital capacity (FVC) and forced expiratory volume in 1 second (FEV1). This is presumed to be an adaptive mechanism since people living in high altitudes with lower atmospheric PO2 levels must maintain a higher minute ventilation to maintain normal tissue oxygenation.

The danger of ignoring race and ethnicity in PFT interpretation

One of the arguments against using race and ethnicity in PFT interpretation is that by separating patients into different racial or ethnic groups, we are encouraging health disparities. However, an equal argument can be made that if we do not use race and ethnicity in PFT interpretation, we are actually causing health disparities.

Several years ago, I got a panicked call from a family medicine physician who had gotten a spirometry test that showed a low FVC interpreted as indicative of restrictive lung disease and he was worried he had interstitial lung disease. I obtained a new full set of pulmonary function tests and confirmed that both his FVC and total lung capacity were below the 5th percentile of normal, indicating restrictive lung disease. But he was from India and moved to the U.S. when he was a teenager. Our PFT machine utilized the NHANES data set that did not include a racial designation of Southeast Asian (or even just Asian) so he was compared to normal values for Whites. I told him that people from India normally have lower lung volumes compared to White people from the U.S. and that I believed that he was healthy. However, he was very anxious and ended up getting a high resolution chest CT and a cardiopulmonary exercise test just to prove that he did not have interstitial lung disease.

This case illustrates that if we do not use race and ethnicity in PFT interpretation, then we run the risk of incorrectly labeling many Southeast Asian and Black patients as having restrictive lung disease when in fact, they are normal. In addition, by including all racial and ethnic groups in the calculations of normal values, we end up with lower values for the 5th percentile of normal for White, Hispanic, and Northeast Asian patients. As a result, we can miss restrictive lung disease in these racial and ethnic groups.

PFTs are also used to determine life insurance premiums and suitability for certain occupations. I recently got an email from one of the nurses at our hospital whose son was unable to enter firefighter school because his FEV1 was too low (he is healthy with no known lung disease). Abnormal PFT values can keep a person from entering the military or becoming a commercial pilot. PFTs are used in disability determination, in pulmonary rehabilitation eligibility, and in pre-operative assessment for lung cancer surgery. By eliminating race and ethnicity, we could inadvertently prevent African Americans and Southeast Asians from getting certain jobs or getting needed lung cancer surgery. Similarly, we could make it more difficult for Northeast Asians and Whites to get disability benefits or get into pulmonary rehabilitation.

PFT interpretation is as much art as it is science

When I was a resident, one of my mentors who was a cardiologist told me that the non-cardiologist at a patient’s bedside could interpret that patient’s EKG better than the cardiologist reading that EKG who has never seen the patient. That is because tests such as EKGs are best interpreted in the context of the individual patient’s clinical presentation and the person in the best position to know that clinical presentation is the physician at the beside taking care of that patient (provided that the physician is well-trained in EKG interpretation).

Pulmonary function tests are similar in this way to EKGs. If a patient is 8 months pregnant or has severe scoliosis on physical exam and that patient’s PFTs have a computer interpretation of restrictive lung disease, I am not going to do an extensive work-up for interstitial lung disease because my physical exam shows me that the PFT changes are due to diaphragm limitation by a gravid uterus or to chest wall abnormalities caused by scoliosis. If we do not include race in the demographics entered into the PFT computer and the computer interpretation shows mild restrictive lung disease, I will be less concerned if I know that patient is from India.

Using race and ethnicity in PFT interpretation is a conundrum – we are damned if we do and damned if we don’t. At the workshop that led to the new American Thoracic Society guidelines, 30 out of 33 attendees recommended to eliminate race and ethnicity in PFT interpretation. For this reason, it is likely that in the near future, race/ethnic demographics will not be requested when entering data into PFT machines and those machines will use race-neutral data sets of normal people in the determination of percent predicted values. It will be incumbent on all of us who use pulmonary function tests to ensure that we do not create healthcare disparities in our attempt to eliminate healthcare disparities when race-neutral data sets are used.

February 10, 2024

Each year, hospital quality departments select a group of metrics to monitor for quality performance. It is time for hospitals to focus on adenoma detection rates as one of these metrics. Adenomas are pre-cancerous polyps found during colonoscopy. By removing polyps found on screening colonoscopy, the risk of colon cancer can be substantially reduced.

 

Colon cancer is the 4th most common cancer in the United States and is the 2nd most common cause of cancer death. Colon cancers arise from adenomatous polyps and it takes about 10-15 years for a polyp to develop into a cancer. There are several accepted methods to screen for colon cancer. Colonoscopy has the highest sensitivity for detecting both colon cancers (95%) and large adenomas (95%) compared to other screening methods. The next most sensitive screening methods are CT colonography (92% sensitive for colon cancer and 42% sensitive for large adenomas) and stool DNA testing (84% sensitive for colon cancer and 84% sensitive for large adenomas). An advantage of screening colonoscopy over other screening methods is that if an adenoma is detected, it can be removed during the same procedure whereas if other screening methods are abnormal, a colonoscopy is then necessary for adenoma removal and/or cancer biopsy. Because of its more favorable sensitivity, every 10 year colonoscopy is generally recommended over other screening methods for persons at average risk of developing colon cancer.

The U.S. Preventive Services Taskforce now recommends screening all Americans between ages 45 to 75 years of age for colon cancer. If using screening colonoscopy, this would imply that all Americans with average colon cancer risk should have 4 screening colonoscopies over the course of their life. However, those Americans with a higher risk may require earlier initiation of screening (age 40) and more frequent colonoscopy (every 2-5 years).

Screening colonoscopy versus diagnostic colonoscopy

All colonoscopies are not created equal. The Affordable Care Act requires Medicare and commercial insurance companies to cover the cost of screening for colorectal cancer from age 45 to age 75 and this generally means that the patient does not have to pay for a screening colonoscopy.  If a polyp is found during screening colonoscopy, it is generally removed at the time of that procedure. Most commercial insurance companies will cover the cost of the polypectomy at the time of a screening colonoscopy but Medicare will usually require a co-pay for the polypectomy portion of the screening colonoscopy.

The Affordable Care Act does not require insurance companies to fully cover diagnostic colonoscopies. This is often a source of confusion and frustration for patients who do not understand the differences between screening and diagnostic colonoscopy. A diagnostic colonoscopy is a colonoscopy that is done to investigate abnormal symptoms, tests, prior conditions, or family history. Diagnostic colonoscopies may include out-of-pocket costs for patients (such as co-pays or deductibles). Both screening and diagnostic colonoscopies are coded with the same CPT code. However, the diagnosis code that the physician uses with the CPT code dictates whether the colonoscopy is a screening or diagnostic procedure. The proper diagnosis code for a screening colonoscopy is Z12.11 (encounter for screening for malignant neoplasm of colon).

If the incorrect diagnosis code is linked to the colonoscopy CPT code, then commercial insurance companies may consider the procedure to be a diagnostic colonoscopy in which case, the patient could have a sizable co-pay or the cost of the procedure could be subject to policy deductible limits. Insurance companies have been known to deny full reimbursement for a screening colonoscopy if the physician lists a gastrointestinal diagnosis code (such as GI bleeding), if the patient is at high risk for colon cancer (such as having a family history of colon cancer), or if the patient has a personal history of colon cancer or polyps. If the colonoscopy is being done to follow up on an abnormality found on another colon cancer screening test (such as a stool DNA test or a fecal occult blood test), then that colonoscopy is generally considered to be a diagnostic colonoscopy and is therefore subject to patient co-pays and deductibles.

Because of the significant implications for patient co-pays and deductibles, it is important that patients understand up front the financial implications of choosing any colon cancer screening test other than a colonoscopy since if that initial test is negative, the patient will likely be responsible for some or all of the cost of the subsequent colonoscopy to follow-up the abnormal results. It is also important that physicians and coders select the proper CPT code and diagnosis code in order to prevent the patient from being incorrectly billed for the procedure.

Adenoma detection rates

One of the best measures of the skill of the physician performing screening colonoscopy is the adenoma detection rate (ADR). This is defined as the percentage of asymptomatic patients undergoing initial screening colonoscopy who are found to have an adenomatous polyp during the exam. For example, an ADR of 25% would mean that 1 out of 4 patients are found to have a polyp during their initial screening colonoscopy.

A study published in JAMA this month found that patients undergoing screening colonoscopy by physicians who have a low adenoma detection rate were more likely to be diagnosed with colon cancer later in life. The basis for these finding is that colon cancers arise from adenomatous polyps. Thus by identifying and removing adenomas, subsequent colon cancers can be prevented. However, colonoscopy and identification of polyps is a skill that is very operator-dependent and not all physicians are equally skilled in the procedure.

In the United States, most screening colonoscopies are performed by either gastroenterologists or general surgeons. The American Society of Gastrointestinal Endoscopy recommends that physicians perform 275 colonoscopies during training in order to be credentialed to perform colonoscopy. The Accreditation Council for Graduate Education (ACGME) specifies that general surgery residents should perform 50 colonoscopies by the end of their residency. The ACGME does not specify the minimum number of colonoscopies to be performed during gastroenterology fellowship and instead focuses on achieving competency milestones based on observed procedural skill rather than procedure numbers. A recent study from the journal Clinical Gastroenterology and Hepatology found that the average U.S. male gastroenterology fellow performed 552 colonoscopies and the average female gastroenterology fellow performed 488 colonoscopies during fellowship training. The ACGME does not include a minimum adenoma detection rate as a requirement for either general surgery residency or gastroenterology fellowship. The reality is that some physicians can become competent at colonoscopy after performing fewer procedures whereas other physicians may require considerably more procedures to achieve competency.  Thus, the number of colonoscopies performed during training should not be the sole requirement for colonoscopy credentialing.

A recent study from The University of Florida found that adenoma detection rates improved as gastroenterology fellows progressed through training. At the beginning of training, the adenoma detection rate was 27% and increased to 50% by the end of training. However, adenoma detection rates have not historically been tracked in all gastroenterology fellowship programs or in all general surgery residency programs.

The ideal adenoma detection rate is dependent on the population of patients screened. For example, there are gender differences in the presence of adenomas as well as the risk of death from colon cancer. Men have a 25% higher rate of death from colon cancer than women. Other colon cancer risk factors include older age, African American race, history of kidney transplant, tobacco use, alcohol consumption, diabetes, obesity, and diets heavy in red meats. Thus, the demographics of the patient population undergoing screening at any given hospital as well as by any given physician can affect the percentage of patients who undergo colonoscopy and have an adenoma.

So what is the optimal adenoma detection rate? As a general rule, a minimum ADR of 20% for women and 30% for men undergoing colonoscopy should be expected. For a physician with a mixed-gender patient population, the minimum ADR can be simplified to 25%; this is the minimum ADR recommended by the European Society of Gastrointestinal Endoscopy. For physicians with patient populations at higher risk of colon cancer, these thresholds may need to be increased. For example, a physician who primarily performs screening colonoscopy on African American men over age 65 may need to have a higher ADR threshold than a physician who primarily performs screening colonoscopy on Caucasian American women under age 65.

What should hospitals do with physicians who have a low ADR?

At our hospital, most of the colonoscopies are performed by the university-employed gastroenterologists and surgeons. However, several years ago, we had a number of private practice gastroenterologists who performed procedures in our hospital’s endoscopy unit. One of these gastroenterologists was observed to miss polyps by the nurses assisting him with colonoscopies. This gastroenterologist also had an unusually high number of procedural complications. I required that all of his colonoscopies for several months be proctored by another gastroenterologist who either would be physically present in the room during the procedure or would review a video recording of the entire scope withdrawal through the colon. In this case, the gastroenterologist chose to move on to another hospital rather than be subjected to several months of proctoring.

A disadvantage of retrospective video proctoring is that if the proctor sees a polyp missed by the original physician who performed the colonoscopy, then that patient will need to be called back for a second, diagnostic colonoscopy. Not only does this inconvenience the patient, but diagnostic colonoscopies generally have higher patient co-pays than screening colonoscopies. An alternative to procedural proctoring could include remedial training in colonoscopy, such as the “return to practice” training programs often used for physicians who have been away from the active practice of medicine for an extended period of time. Another option for physicians with an excessively low adenoma detection rate is suspension of screening colonoscopy privileges.

The ultimate goal of adenoma detection rate monitoring is not just to identify physicians who simply should not be performing screening colonoscopy. Instead, the goal should be to get all physicians’ ADRs as high as possible, ideally close to 50%. In addition to setting a minimum ADR for screening colonoscopy privileges, there should be an incentive program in place to encourage physicians to have as high of an ADR as possible. This could take the form of publicly reporting individual physicians’ ADRs, prioritizing physicians with a high ADR on the endoscopy unit schedule, or providing a monetary bonus based on physicians’ ADRs.

In addition to the physician’s own colonoscopy skill, patient factors can contribute to a low ADR. One reason that adenomas may be missed is because of a poor colon preparation. Colon preps are unpleasant for the patient and often difficult for the patient to complete. Many patients with a poor colon prep at the time of their screening colonoscopy will refuse to repeat the prep and return for a second procedure at a later date. Every effort should be made to ensure that the patient has a satisfactory prep the first time. One barrier to adequate colon preps is that the gastroenterologist or surgeon performing the colonoscopy usually does not see the patient in the office prior to performing the colonoscopy and the preps are often ordered by the patient’s primary care physician. Institutional standardization of the specific colon prep to be used can eliminate primary care physicians from ordering a less effective prep. The use of split-dose preps, low fluid volume preps, and hospital-employed “patient navigators” can improve patient compliance with preps and reduce the rate of poorly prepped colons.

Another reason for a low ADR can be inadequate patient sedation. In the past, most colonoscopies were performed by using ‘moderate sedation’, that is, sedative medications administered by the physician performing the colonoscopy that result in the patient being sleepy but arousable. It can be difficult to perform a thorough colon visualization if patients are insufficiently sedated. Because of this, more colonoscopies are now being performed using ‘deep sedation’ that causes the patient to be asleep and less arousable during the procedure. Deep sedation requires a second provider to administer the sedative medication, generally an anesthesiologist or a CRNA. Deep sedation is often preferred for colonoscopies performed on patients with co-morbid medical conditions such as sleep apnea, COPD, or heart failure. It may also be preferred in patients who are highly anxious, have low pain thresholds, or who have been difficult to sedate in the past. Increasing the availability of deep sedation for screening colonoscopies is one tactic that the hospital can take to improve its overall adenoma detection rate.

Everyone wins with a high ADR

Endoscopy units are often sites of ‘turf battles’ between different physicians, different medical groups, and different specialties. A common conflict is between gastroenterologists and general surgeons, both of whom want to have a piece of the screening colonoscopy business. My own philosophy has always been that any physician credentialed to do a procedure should have equal access to the procedure areas where that procedure is performed. In the future, it should not be the physician’s specialty that determines who performs screening colonoscopies, it should be the physicians’ ADR that determines who performs screening colonoscopies.

A high adenoma detection rate means more colon cancers are prevented and that is a win for the patients. A high rate also means that fewer colon cancers are missed and that is a win for physicians who can avoid costly medical malpractice lawsuits for failure to diagnose. And a high rate is a win for the hospital which can avoid negative publicity from missed cancers. It is time for the ADR to be a routine quality measure in our hospitals.

June 18, 2022

An important mission of the hospital is the relief of suffering and that includes relief of pain. Pain management programs are central to fulfilling this mission. The past decade has taught us that you cannot effectively manage pain with an opioid prescription alone. The combination of opioid addiction plus the COVID-19 pandemic has proven to be lethal for many Americans. Last year, there were 100,306 drug overdose deaths in the United States, up by 28.5% from the 78,056 overdose deaths in the previous year. The vast majority of these deaths were caused by natural or synthetic opioids and for many of these people, their addiction started with a pain medication initially prescribed by a doctor.

In the 1990’s, “Pain as the 5th vital sign” was the mantra of pain management services with the implication that physicians were not prescribing enough opioids and that it was our moral duty to prescribe more. The consequence of this campaign was that many of our patients became opioid-dependent. When we realized this, the pendulum swung the other way, with state medical boards restricting the amount and duration of opioid prescriptions that doctors could order. As a result, the supply of prescription opioids fell dramatically and the opioid-dependent population turned to illegal opioids. Coincident with this, inexpensive synthetic fentanyl became readily available on our streets and many Americans died of unintentional overdose due to the unpredictable concentrations of fentanyl in purchased quantities of street drugs.

The good news, is that we have a number of great alternatives to opioid pain medications for both acute and chronic pain management. However, a high-functioning hospital needs to have  more than just one of these pain management services.

What is pain, anyhow?

Pain exists when our peripheral nerves let us know that a part of the body is being injured. This is a great defense mechanism to avoid bodily harm, for example, pain is how we know to pull our hand away when we touch a hot stove. But pain can become pathologic when those pain nerves keep firing even though there is no avoidable injury – for example, the patient with bone metastases from cancer, the patient hospitalized after multiple trauma, the patient with chronic arthritis, or the patient recovering from a knee replacement surgery. In those situations, the pain nerves just keep firing away and there is nothing that the person can do by themself to make those nerves stop.

But there is a lot more to the perception of pain than just signal from a peripheral nerve. There are pain amplifiers that can turn the volume of pain up. The most important of these are fear, anxiety, and depression. Often, the presence of one of these modifiers can convert tolerable pain into intolerable pain.

What pain management services does the hospital need?

Comprehensive pain management does not boil down to having a single pain management service. Hospitals need to have a spectrum of options for treating pain in order to do the most good for the most people. All too often, the physicians or advance practice providers who are proficient with one type of pain management option are not proficient with other options.

1. Acute pain services. These are inpatient providers, frequently anesthesiologists, nurse anesthetists, pharmacists, and/or nurse practitioners. These providers are very good at managing temporary pain, particularly post-operative pain and trauma-related pain. They will have experience in managing pain pumps and in selecting opioid and non-opioid pain medications that are meant to be used for limited numbers of hours or days. A larger hospital can afford to maintain an acute pain service but the low patient volume at a smaller hospital may make an acute pain management service cost-prohibitive. In order to serve our smaller, urban hospital, we created an acute pain telemedicine consultation service with providers located at our larger, tertiary care hospital located on the other side of town.
2. Pain and palliative care services. These are providers who may work in either inpatient or outpatient areas and typically focus more on chronic pain management. They are usually physicians who have completed a palliative medicine fellowship who lead a team that may include nurse practitioners, physician assistants, pharmacists, and social workers. Cancer-related pain and sickle cell anemia-related pain are examples of their clinical focus. Although chronic opioid prescription may be a part of their practice, they will also typically address pain modifiers, such as fear and depression.
3. Interventional pain services. These are physicians who have done fellowship training in interventional pain management and most commonly draw from anesthesiology, physical medicine & rehabilitation, and neurology. Their practice is generally outpatient and many include steroid injections, radiofrequency ablation, intrathecal pumps, sympathetic blocks, peripheral nerve stimulators, and spinal stimulators. They will often interface with outpatient therapies such as physical therapy, aqua therapy, and psychology. The procedures that they perform often require use of the operating room or an imaging area such as a cath lab or interventional radiology lab. Many of their procedures are done using moderate sedation but some may require general anesthesia.
4. Sports medicine. These are family physicians, internists, or pediatricians who have done fellowship training in sports-related injuries and over-use injuries. Despite the name, sports medicine physicians treat many patients who are not athletes. They will often interface with physical therapists, athletic trainers, and orthopedic surgeons when directing specific treatments for injuries accompanied by pain.
5. Complementary and alternative medicine. This includes a wide variety of services such as acupuncture, massage therapy, chiropractic treatments, yoga, and traditional Chinese medicine. Although physicians may be involved in alternative medicine, many of these providers are non-physicians. Many hospital medical directors take a jaded view of alternative medicine. However, these services can often de-amplify pain by reducing anxiety and fear. They can also provide a sense of control to patients with chronic pain that can make pain much more manageable. Regardless of what the hospital medical director may think, if the patient believes that these services work, then they can be beneficial.
6. Inpatient physicians. Hospitalists, surgeons, and anesthesiologists are the first-line of pain management for most inpatients. However, the formal training that they get in pain management is highly variable. Clinical practice guidelines and treatment protocols can be very useful to ensure a hospital-wide standard of practice. Periodic continuing medical education events are also valuable. One of the most important roles of these physicians is to manage pain expectations. If patients are told that they are going to have post-operative pain before they actually have their surgery and they are told that their pain will be manageable with non-steroidal anti-inflammatory drugs and physical therapy, then those patients are less likely to require opioids post-operatively compared to patients who go into surgery unprepared to experience any pain after surgery.
7. Outpatient physicians. Primary care physicians, surgeons, and emergency medicine physicians are the front-line of pain management for most outpatients. Once again, their formal training in pain management can be highly variable and so just as for inpatient physicians, clinical practice guidelines, treatment protocols, and periodic continuing medical education are usually necessary. Most state medical boards have state-specific rules and regulations regarding chronic opioid prescription and  it can be very difficult for the primary care physician to ensure that all of the monitoring and documentation requirements are met. A robust electronic medical record can help with this. But if there is a critical mass of patients receiving chronic opioid medications, an advanced practice provider dedicated to chronic, stable-dose opioid management can be cost-effective.

In addition to pain services that manage a spectrum of conditions, there are also disease-specific specialists needed to manage certain conditions. Migraine (often managed by neurologists) and fibromyalgia (often managed by rheumatologists) are two examples. Having a physician on the hospital medical staff who specializes in these conditions can help avoid primary care physicians ordering opioids out of frustration.

Match the patient with the pain service

Although there is frequently a lot of overlap between different types of pain services, to optimally meet the needs of the most patients, all seven of the above pain services need to be available – if not in each hospital, then at least somewhere in the community. No two patients are exactly alike when it comes to pain tolerance and pain perception. Treating fibromyalgia with chronic opioids just doesn’t work. Nor does bone metastasis pain with physical therapy. We should strive to match the patient’s type of pain with the right type of pain service.

Our natural tendency as humans is to use whatever tool we are familiar with to fix whatever problem we face (“When all you have is a hammer, everything looks like a nail”). When it comes to pain management, be sure that your hospital has a full toolbox.

February 19, 2022

Pulmonary function tests (PFTs) can be very intimidating. A PFT report has many different numbers and for the non-pulmonologist, it is often difficult to know which ones are important and what they mean. As a consequence, many non-pulmonologists just look at the computer-generated interpretation and ignore all of the numbers. The bad news is that the computer-generated PFT interpretations are woefully bad – the computer is pretty good at identifying normal but generally terrible at identifying degrees of abnormal. The good news is that PFT interpretation is actually very easy and in this post, I am going to show you everything you need to know to interpret 99% of the PFTs that you order by looking at just 5 numbers. There are fundamentally four components to a complete PFT: spirometry, lung volumes, diffusing capacity, and the flow-volume loop. We will next look at each of these components.

You can view my OSU MedNet webcast on PFT interpretation that covers more details here.

Spirometry

Spirometry is the most common PFT that you will order. It can either be performed by a respiratory therapist in a pulmonary function lab using an expensive PFT machine (> $50,000) or it can be performed by a nurse in an office setting using a portable spirometer (about $1,000). Spirometry is primarily a measurement of air flow and is used to determine whether or not a patient has obstructive lung disease.

You will see a lot of different numbers on a spirometry report but the only ones that are truly important are the forced vital capacity (FVC), the forced expiratory volume in the first second of exhalation (FEV1), and FEV1/FVC which is the ratio of these two numbers. The other numbers are of minimal importance and you can ignore them. The results of the FVC and FEV1 will be listed both as actual values (in liters of air) and as the percent of predicted. In order to determine the percent of predicted, researchers have measured the FEV1 and FVC of thousands of normal people to determine what a normal value is. “Normal” depends on a person’s age, height, gender, and race so it is important that these demographic elements are correctly entered into the PFT machine so that the computer in the machine picks out the correct normal value based on that patient’s age, height, race, and gender. If the patient’s FEV1 or FVC is lower than the 5th percentile of normals (lower limit of normal), then the PFT machine will flag the result – usually with a different color font or an asterisk by the number.

The ratio between these two numbers, the FEV1/FVC, tells you whether or not the patient is obstructed. If the FEV1/FVC is too low, the patient is obstructed and if the FEV1/FVC is either normal or higher than normal, then the patient is not obstructed.  There are 2 definitions of a “normal” FEV1/FVC that are commonly used. The first (and most common) is the definition used by the American Thoracic Society (ATS) – it is based on a patient’s age because the older we get, the lower a normal person’s FEV1/FVC becomes simply as a result of normal aging. The average FEV1/FVC for a 20-year-old is 87% whereas the average FEV1/FVC for an 84-year-old is 71% and the lower limit of normal for an 84-year-old is 59%. The computer in the PFT machine will flag the FEV1/FVC as being too low based on whatever age is entered in the patient demographics. The second definition of normal FEV1/FVC is used by the Global initiative for chronic Obstructive Lung Disease (GOLD) which uses a flat FEV1/FVC ratio of 70% as being normal for everyone and does not make any adjustments based on a person’s age. Almost all PFT machines will use the ATS method when generating a PFT report but Medicare requires the GOLD method when determining eligibility for enrollment in pulmonary rehabilitation.

If the FEV1/FVC ratio is low for the patient’s age, then the next step is to determine how obstructed they are. This is determined by how low the FEV1 is. IMPORTANT: if the FEV1/FVC ratio is normal, then it does not matter how low the FEV1 is since that patient is not obstructed. The severity of obstruction is usually based on the ATS criteria but once again, the GOLD criteria is different.

February 2023 update: the ERS/ATS just issued a revised severity scoring system and recommend using FEV1 z-scores (z-scores are the number of standard deviations below the mean values). The new severity criteria are:

• Mild obstruction: FEV1 z-score -1.65 to -2.5

• Moderate obstruction: FEV1 z-score -2.52 to -4.0

• Severe obstruction: FEV1 z-score <-4.1

If a patient is found to have obstruction based on a low FEV1/FVC, then sometimes a bronchodilator study is performed by giving the patient an albuterol treatment and then repeating the spirometry a few minutes later. Reversible obstruction is defined as an increase in either the FEV1 or FVC by 12% after albuterol. In addition, the amount of the increase in FEV1 or FVC must be at least 200 ml to qualify as reversible obstruction. If the initial FEV1/FVC is normal, there is no point in doing a bronchodilator study.

February 2023 update: the ERS/ATS just issued revised criteria defining reversible obstruction as an increase of at least 10% in the percent predicted value of either the FEV1 or the FVC after a bronchodilator. 

For the spirometry test result to be accurate, three things must occur: (1) the patient’s age, race, height, and gender must be entered correctly, (2) the patient must be coached correctly by the nurse or respiratory therapist administering the test and the patient must give a good effort, and (3) the physician or advance practice provider must correctly interpret the test.

Steps in spirometry interpretation:

1. Is the patient obstructed? This is determined by whether the FEV1/FVC ratio is low.
2. If they are obstructed, how obstructed are they? This is determined by how low the FEV1 is below normal.
3. If they are obstructed, is the obstruction reversible? This is determined by the amount of increase in the FEV1 and/or FVC after albuterol.

Lung Volumes

Lung volumes cannot be measured using an office spirometer and require testing in a pulmonary function laboratory. This is typically done by having a person sit in a plethysmographic box and performing a series of inhalation and exhalation maneuvers. Once again, the PFT machine will generate a whole bunch of numbers but the only one that is really important is the total lung capacity (TLC). If this number is too low, then the patient is restricted and if it is too high, then the patient is hyperinflated. Just as in spirometry, the TLC percent predicted is based on the patient’s age, height, race, and gender.

Of note, some office spirometry machines will give a computer-generated interpretation of restriction based on the FVC alone. You really cannot diagnose restriction using the FVC alone from spirometry, the diagnosis requires measuring the TLC with lung volumes – the FVC can sometimes be low when the TLC is totally normal. The finding of a low FVC on office spirometry should merely be an indication to send that patient to the pulmonary function lab for a complete set of PFTs that includes lung volume measurements.

If the TLC is larger than the upper limit of normal (generally around 120% of predicted), then the patient is hyperinflated. The only other number that is worth knowing about is the residual volume (RV). If the RV is too high (above the upper limit of normal), then the patient has air-trapping. Patients who are obstructed will often also have hyperinflation or air-trapping. In fact, isolated air-trapping alone can indicate early obstruction, even when the FEV1/FVC ratio is normal.

Diffusing Capacity

The DLCO is the “Diffusing capacity of the Lung for carbon monoxide (CO)”. This is measured by having the patient breath a tiny amount of carbon monoxide gas and then comparing the amount of carbon monoxide in the inhaled gas versus the amount in the exhaled gas. The difference indicates how much of the carbon monoxide was taken up by the body (the carbon monoxide binds to hemoglobin in the bloodstream). The more carbon monoxide is taken up, the easier it is for gases to get from the air a person breaths into the bloodstream. If the amount of carbon monoxide taken up is low, then this is a sign that gases cannot move normally from the lungs into the bloodstream.

The DLCO will be low in diseases such as emphysema and interstitial lung disease. However, the DLCO can also be low if the patient has small lung volumes for any reason. Because of this, the PFT machine will often report the DLCO/VA where the VA stands for “alveolar volume” and is more or less the same thing as the total lung capacity. The DLCO/VA is a better measurement of diffusing capacity in situations such as previous partial lung removal or restriction due to muscle weakness. The DLCO can also be low if the patient is anemic. For this reason, the PFT machine will often report the DLCOcor which is the DLCO corrected for the patient’s hemoglobin level. Whether to use the DLCO, the DLCO/VA, or DLCOcor is a matter of debate. Most of the time, you can just go by the DLCO unless the patient is known to have significant anemia or a prior lung resection.

Flow-Volume Loops

Interpreting the flow-volume loop involves looking at the pattern of the graph of airflow during inhalation and exhalation. The PFT machine will not provide an interpretation of the flow-volume loop so the physician or advance practice provider must do their own interpretation. There is a wealth of information in these graphs but too often, people ignore the flow-volume loop and go straight to the PFT numbers. This is a mistake because sometimes the only clue to a disease is found in the shape of the flow-volume loop. Also, the shape of the loop can be an indication that the patient gave poor effort during the test and the results may not be valid. The figure below shows the elements of a normal flow-volume loop with exhalation occurring during the top part of the loop and inhalation occurring during the bottom part. PEF is the peak expiratory flow. The FEF25% and FEF75% are the flow rates between the first 25% of the volume of exhaled air and the last 25% of the exhaled air. But what is important is the shape of the curve.

The shape of the normal loop should have an early expiratory peak with a gradual drop downward until the patient reaches the maximum amount of air that they can exhale (FVC). The inspiratory portion of the loop on the bottom should be more symmetric and look sort of like a half-circle. If the patient does not give a good effort, then the test is invalid and you should not believe the FVC or FEV1 numbers. In this situation, the shape of the loop will be very irregular as shown below:

Coughing during the test will also invalidate spirometry results and can be identified by a sudden drop in expiration causing a crevice or “double hump” to appear in the top portion of the loop:

There are two conditions that can result in flattening of the flow-volume loop. Vocal cord paralysis will cause flattening of the inspiratory limp but the expiratory limb will appear normal. Tracheostenosis will cause flattening of both the inspiratory and expiratory limbs of the flow-volume loop:

Patients with vocal cord dysfunction will clinically mimic asthma with dyspnea and wheezing. The flow-volume loop is often the first clue that a patient has vocal cord dysfunction. The expiratory limb of the flow-volume loop will appear normal but the inspiratory limb will be choppy and irregular. Vocal cord dysfunction can be confirmed by ordering a videolaryngostroboscopy study that will show that the vocal cords come together too much during inspiration, particular in the anterior portion of the vocal cords (anterior is the top portion of the photos below).

Putting it all together

If y0u have a full set of pulmonary function tests, these are the steps to interpreting the results that will be all you need to do in 99% of the PFTs you order:

1. Is the spirometry test valid? Look at the shape of the flow-volume loop and also confirm that the patient’s demographics were correctly entered.
2. Is the patient obstructed? Look to see if the FEV1/FVC ratio is low.
3. If they are obstructed, how obstructed are they? This is determined by how low the FEV1 is below normal.
4. If they are obstructed, is the obstruction reversible? This is determined by the amount of increase in the FEV1 and/or FVC after albuterol.
5. Is the patient restricted? Restriction is present if the TLC is below the lower limit of normal.
6. Is the diffusing capacity reduced? This is defined as a DLCO that is below the lower limit of normal.

Every type of lung disease will give a particular pattern of PFT results. The following table summarizes the most common lung conditions and their PFT patterns. The cells in the table that are shaded yellow are the key findings that point toward each of these 4 conditions:

When it comes to PFT interpretation, less is more. You really only need to know 5 numbers to interpret the overwhelming majority of PFTs. You can ignore the rest of the data and leave that for us pulmonologists who like to get way down in the weeds of the PFT report.

December 2, 2021

Imagine if a Boeing 777 jet crashed and killed all on board. And then imagine such a crash occurring every day for a year. That is how many Americans die of lung cancer, a disease that is not only preventable (if you don’t smoke) and curable (if found early). More people die of lung cancer than die of colon cancer, breast cancer, and prostate cancer combined.

This year, 235,760 Americans will be diagnosed with lung cancer which accounts for 12.4% of all cancer diagnoses. The 5-year survival rate is only 21.7% and an estimated 131,880 Americans will die of their lung cancer this year. The problem with lung cancer is that it is usually found late, after it has already spread and no longer surgically curable. As a consequence, the 5-year survival of lung cancer is much lower than any other common type of cancer. However, lung cancer screening programs can identify lung cancer at an early stage, when it can still be surgically cured.

Screening for colon cancer and breast cancer is relatively straight forward: at a certain age, everyone starts getting a colonoscopy every 10 years and all women start getting a mammogram every year. Screening for lung cancer is more complicated for two reasons. First, because the criteria for who should or should not undergo screening is more complex and second, because there has to be a process in place for managing all of the abnormalities that are identified on screening tests (most of which are not lung cancer).

It has long been known that screening with regular chest x-rays does not work; x-rays just do not identify lung cancers at an early enough stage. A landmark study in 2011 showed that low-dose chest CT scans not only identify more lung cancers than chest x-rays, but patients who got chest CT scans were 20% less likely to die of lung cancer in the subsequent 6.5 years than those who only got screening chest x-rays:

A more recent study from last year showed that patients who got lung cancer screening chest CTs had a 25% lower risk of dying of lung cancer in the subsequent 10 years:

Clearly we need to be screening patients for lung cancer but only 2-4% of eligible smokers are currently getting screened. So, why aren’t we screening more? The two major barriers are patients and healthcare providers. Patients are often unaware of screening programs, fear a cancer diagnosis, worried about the costs, or simply do not have access to screening. Physicians are often unfamiliar with the screening guidelines, unsure of insurance coverage, lack the time in the office to counsel patients about screening, don’t know what to do about abnormalities found on CT, are skeptical about the efficacy of screening, or are worried about the risks of false positive findings. In February 2015, CMS approved lung cancer screening for Medicare recipients if they met a group of specific criteria. In March 2021, the US Preventive Services Task Force revised the guidelines for screening to now include:

1. Adults age 50 – 80 years old
2. At least a 20 pack-year smoking history
3. Currently smoking or quit within the past 15 years

People who meet all three of these criteria are recommended to undergo an annual low-dose chest CT scan. Screening should continue until the person has quit smoking > 15 years earlier, is no longer willing to undergo curative surgery if a cancer is found, or develops another medical condition that substantially reduces life expectancy. For Medicare coverage, the patient must additionally have no signs/symptoms of lung cancer and screening must include smoking cessation counseling.

One of the issues raised by lung cancer screening is that chest CT scans can pick up a lot of benign abnormalities. In fact, 97% of all abnormalities found on screening chest CTs are not cancer. For this reason, there has to be a process for managing these abnormalities – both for choosing the best way to biopsy those patients who have abnormalities that are more likely to be cancer and for arranging follow up testing for those patients who have abnormalities that are less likely to be cancer. This is where a carefully designed lung cancer screening program can be effective and efficient.

Components of a lung cancer screening program

To be successful, a lung cancer screening program should include a CT scan capable of low-dose chest imaging, a radiologist available to interpret that CT, a clinical provider, and a pulmonologist. Ideally, the screening should be able to be completed within an hour and a half with the patient going to just one location. The entire screening visit should be able to be ordered by the patient’s primary care provider using a single order set. The screening visit should consist of:

1. An initial review of the patient for inclusion criteria
2. An encounter with a clinical provider with experience in pulmonary nodule management and smoking cessation counseling
3. The chest CT scan with radiologist interpretation
4. A second encounter with the clinical provider after radiologist’s CT interpretation is available
5. Ordering of appropriate follow-up testing

Let’s look at each of these steps in detail:

Initial review of patient for inclusion criteria. The US Preventive Services Task Force lung cancer screening criteria have not yet been adopted by all insurance companies. As a result, different insurance companies will have different requirements for screening eligibility. After the primary care provider places an order for lung cancer screening, the order should initially go to a nurse who can check the patient’s insurance and verify that the patient meets the age and smoking history requirements for that specific insurance company. Most insurance plans additionally require that the patient has not had a chest CT for any purpose within the past year. Once the patient’s eligibility is confirmed, then the order can be routed to the screening clinic for scheduling.

Initial encounter with clinical provider. This provider can be a physician or an advance care provider. Given the nature of this encounter, a nurse practitioner or physician assistant is an ideal choice. During this encounter, the screening process is discussed, the patient’s eligibility is confirmed, smoking cessation counseling is given (if the patient is an active smoker), and the CT scan is ordered. There should be “shared decision making” between the provider and the patient so that the patient understands that non-cancerous abnormalities are common but may require additional testing. It should also be confirmed that the patient is willing to undergo biopsy and/or surgery if warranted by the CT findings. Typically, 8-10 patients can be scheduled during a 4-hour clinic time.

The chest CT scan. Ideally, this should be done immediately following and at the same location as the encounter with the clinical provider. The CT machine should be capable of low-dose chest CT protocols. The procedure time for this type of CT scan is less than a minute. The radiation dose of a standard chest CT scan is 7.0 mSv whereas the low-dose chest CT is only 1.5 mSv. To put this in perspective, a chest x-ray is 0.1 mSv and a mammogram is 0.4 mSv. A normal person gets 3.0 mSv in background radiation every year. Ideally, the CT should be interpreted by the radiologist immediately with the results available to the clinical provider.

Second encounter with clinical provider. If the CT scan results are immediately available, then the patient should go directly from the CT scan back to see the clinical provider. If the results are not immediately available, then this second encounter can be done by telephone or telemedicine later that day or the following day. Because most electronic medical records are configured to release radiology reports the same day as the CT scan is performed, there is the potential for patients to see the results before the clinical provider if the second encounter does not happen immediately after the CT scan is performed. This can result in a great deal of anxiety if the patient does not understand the significance of abnormalities noted in the radiology report. For this reason, it is optimal for the clinical provider to be able to discuss and explain the findings as soon as possible following the CT scan. Depending on the radiologist’s report, the clinical provider has several options:

• If the CT scan is normal, a follow up lung cancer screening visit in 1 year can be ordered.
• If the radiologist identifies a nodule or other abnormality and the patient has had a previous chest CT elsewhere in the past, the provider can request those images and arrange a follow-up appointment to compare the findings and determine if they meet radiographic stability criteria.
• If the radiologist identifies a nodule or other abnormality and the patient has NOT had a previous chest CT, then the clinical  provider can order a follow-up chest CT scan and office visit based on the 2017 Fleischner Society guidelines. These guidelines provide recommendations for how soon to perform follow-up CT scans based on whether nodules are solid or subsolid, whether nodules are solitary or multiple, the size of the nodule, and whether the patient has lung cancer risk factors.
• If a finding suspicious for lung cancer is identified, then the provider should have access to a pulmonologist to determine the most appropriate next step. Because a PET scan is most commonly performed prior to biopsy or surgery, this will often entail the clinical provider ordering a PET scan to be followed by consultation with a pulmonologist. Sometimes, the pulmonologist may be able to advise the clinical provider regarding next steps via a telephone consultation. These next steps could include:
  • PET scan
  • Bronchoscopic biopsy
  • CT-guided needle biopsy
  • Surgical biopsy/resection
• If the patient desires more extensive smoking cessation assistance, then the clinical provider can refer the patient to a formal smoking cessation clinic.

Lung cancer screening is more than just ordering a chest CT

Lung cancer screening is a lot more complex than screening for other cancers. To be successful, lung cancer screening requires interdisciplinary coordination, incorporation of smoking cessation, and ability to order follow-up testing. Although some primary care physicians may be able to orchestrate all of these elements themselves, it is far more efficient for hospitals to develop a comprehensive lung cancer screening program with standardized management protocols.

September 30, 2021

Clinical laboratories are certified by CMS using the Clinical Laboratory Improvement Amendments (CLIA). Radiology departments are accredited by certification by the American College of Radiology. There are no certification or accreditation standards for pulmonary function laboratories currently so it falls to each hospital to design its own PFT lab. After being involved in the design of 4 PFT labs over the years, these are a few of the things about lab design that I have learned.

First decide what tests will be performed

The tests that the laboratory will perform will dictate the number of rooms and space required for the lab. The initial design of a pulmonary function lab should specify which types of tests will be performed in each room in order to ensure that each room is large enough for all of the equipment and supplies required for those tests.

 

The most common tests performed in a pulmonary function laboratory are spirometry, lung volumes, and diffusing capacity. These can all be done using an enclosed plethysmograph device that the patient sits inside of, sometimes called a “body box”. Each plethysmograph should be in a separate room. A small hospital or an outpatient physician group practice may only need 1 plethysmograph but most pulmonary function labs will need 2 to 4 plethysmographs, requiring 2 to 4 separate rooms. Spirometry can also be ordered as spirometry pre- and post-bronchodilator. The bronchodilator study does not require special space but usually does require a “Terminal Distributor of Dangerous Drugs License” from the state pharmacy board.

The next most common test is the 6-minute walk test. This is generally performed in a long, straight hallway with distances marked on the floor. The patient walks as fast as comfortable and the number of laps walked in 6 minutes are calculated along with the oxygen saturation during the test. The hallway should be wide enough to accommodate an oxygen tank on wheels and should should be lightly trafficked so that it can be blocked off during the duration the test. A related test is the oxygen titration study. In this test, a patient walks until their oxygen saturation drops below 89% and then supplemental oxygen is applied in increasing flow rates to determine the proper flow rate for that patient’s oxygen prescription. The oxygen titration study can be performed in the same hallway as the 6-minute walk test or can be performed on a treadmill.

The methacholine challenge test is a broncho-provocation test done by having the patient inhale increasing concentrations of methacholine, with spirometry performed after each concentration. In the past, an on-site pharmacy was generally required to perform dilutions of methacholine; however, pre-filled, pre-diluted testing kits are now commercially available, thus obviating the need for an on-site pharmacy. This test can be done in the same room used for one of the plethysmograph boxes. A related test is the eucapnic voluntary hyperventilation test that is used to diagnose exercise-induced bronchospasm.

The cardio-pulmonary exercise test is performed by having a patient ride a stationary bicycle (or sometimes by using a treadmill) while breathing into a metatabolic cart in order to measure values such as minute ventilation and oxygen uptake. This test is generally performed in separate room dedicated to exercise testing but can be performed in a room normally used for plethysmograph testing if the room is large enough to accommodate both the plethysmograph box and the exercise test equipment.

The high-altitude hypoxia simulation test is performed by measuring the patient’s oxygen saturation while breathing a 15% oxygen/85% nitrogen gas mixture from a large medical gas cylinder via a face mask. This test is used to determine if a patient requires supplemental oxygen when flying in a commercial aircraft. Because the only equipment required is the medical gas cylinder, this test can be performed in a room used for plethysmographic testing. However, it is preferable to perform this test in a room with a treadmill (or a stationary bicycle) so that the high-altitude hypoxia simulation test can be combined with an oxygen-titration test as a high altitude hypoxia exercise test in order to determine the oxygen flow rate required when a patient is walking at a high-altitude travel destination (such as Denver).

Arterial blood gases are performed by inserting a needle into the radial artery to withdraw arterial blood. This test is most commonly performed to get direct measurement of the amount of oxygen and carbon dioxide in the blood. Arterial blood gases can also be performed while the patient breaths 100% oxygen in the physiologic shunt study.

Get infection control involved early

Patients who get pulmonary function tests are vulnerable to contagious diseases due to their underlying respiratory compromise as well as due to frequenting taking immunosuppressive medications. In addition, these patients often have respiratory infections that can be transmitted to others. Your infection control department input is crucial to ensure that patients and staff are not at risk of acquiring infections from exposures in the lab.

One of the most important aspects of infection control of respiratory pathogens is the number of air changes in each room per hour. The more air changes per hour (ACH), the faster respiratory pathogens such as tuberculosis or the coronavirus causing COVID-19 are cleared from the breathable air.

The Centers for Disease Control has recommendations for the minimum ACH for each type of hospital room. This can range from a high of 15 ACH for an operating room to 2 ACH for certain storage rooms. An exam room or a hospital inpatient room is recommended to have 6 ACH and a bronchoscopy room is recommended to have 12 ACH. The CDC does not specify the ACH for a pulmonary function laboratory. However, the Veteran’s Administration recommends at least 8 ACH for a room used for plethysmographic testing and at least 10 ACH for a room used for cardiopulmonary exercise testing. In the era of COVID-19, the higher the ACH, the better. If the pulmonary function lab will also do sputum induction for suspected tuberculosis, then a negative airflow room is necessary.

In the past, pulmonary function testing utilized non-disposable mouthpieces, nose clips, and other equipment that required cleaning. This resulted in the requirement to have both a clean and a dirty utility room in the pulmonary function lab. Now, most labs use disposable mouthpieces, nose clips, and supplies so that there is no longer a need for a dirty utility room to avoid clean/dirty equipment conflicts.

The infection control department can also be helpful in room design. For example, selecting anti-microbial materials (such as copper) for door handles and other fixtures. Flooring should be made out of resilient tile with minimal seams. There should be hand washing sinks and wall-mounted hand sanitizer in each room used for diagnostic testing.

Efficiency and flexibility

Patients coming in for pulmonary function testing are often in wheelchairs and are often using supplemental oxygen. Doors to testing rooms need to be wide enough to accommodate the width of a bariatric wheelchair (48 inches). Similarly, diagnostic rooms need to contain bariatric-sized chairs. Because of the impaired mobility of many pulmonary patients, the lab should be located as close to building entrances and elevators as possible.

To optimize staff efficiency, a shared patient registration area that can serve multiple outpatient services is preferred for all but the largest pulmonary function labs. Shared waiting areas can optimize efficient use of building space; however, waiting areas should be designed so that staff can maintain line of sight observation of patients. Similarly, when possible, share resources for linen storage, housekeeping, general storage, waste storage, and staff support areas.

Most pulmonary function labs will require hemoglobin testing as part of the diffusing capacity test. Also, most pulmonary function labs will perform arterial blood gas testing. If these specimens must go to a central clinical chemistry lab, then the PFT lab should be close to that lab (at least within the same building). Most PFTs labs find it easier to perform point-of-care testing for arterial blood gases and finger-stick hemoglobin, however. Regardless of where these tests are run, sharps containers are needed in all diagnostic rooms.

Human needs

In addition to a close-by, adequately-sized waiting area, there needs to be restrooms and a staff break room near the lab (you don’t want your staff eating in the diagnostic area). The interior design should convey the appearance of a healthcare setting. There must be adequate lighting in all rooms and hallways. Be sure to have televisions in waiting areas and wifi access in all public areas. Artwork should be chosen carefully – for example, if there is a sizable Afghanistan war veteran patient population, avoid pictures of desert mountains. Similarly, pictures of happy people doing recreational activities can be depressing to patients confined to wheelchairs or oxygen tanks. Attention to privacy in door and window location can ensure that patients undergoing diagnostic testing cannot be easily seen from the hallway.

If there are exterior windows in the area of the building, it is preferable to locate rooms used for diagnostic testing where there are windows and then use windowless interior rooms for support purposes, break rooms, restrooms, staff offices, etc. Some patients get claustrophobic when enclosed in a plethysmographic box and having an exterior window in the room can lessen that claustrophobia. The plethysmograph box should be positioned so that the patient can see out the window when sitting in the box.

Room acoustics are frequently overlooked when designing the PFT lab. If you have ever stood outside of a room where spirometry is being performed, then you have inevitably heard a PFT technician shouting “Blow, blow, blow, as hard as you can…“. Performing PFTs is a loud process. Include acoustic ceiling tiles and adequately insulated walls in the initial design.

Physical layout

Rooms used for plethysmographic testing should ideally be at least 12 ft x 10 ft in size in order to accommodate the plethysmograph box, a workstation for the PFT technician, a chair, sink, equipment storage, trash can, sharps container, etc. Most plethysmographic boxes are about 7 feet tall so the ceiling height also needs to be considered. For hallway throughput safety, doors should open into the room rather than into the hallway. Data entry keyboards used by the staff should either be on mobile workstations-on-wheels or should be on swing-mounts on a wall but positioned so that the technician is facing the plethysmograph box and so that an opened door does not block the ability of the staff to see the patient in the plethysmograph box. Most plethysmograph boxes are 36 to 42 inches in diameter so having a 48 inch doorway is preferred to be sure you can get the box into the room.

Rooms used for exercise testing generally should be to be at least 12 ft x 20 ft in order to accommodate a treadmill and metabolic cart.

The hallway used for 6-minute walk testing should be adjacent to the diagnostic area. Wall-mounted medical gas outlets in the diagnostic rooms are convenient to support the needs of patients requiring supplemental oxygen but most labs can get by with re-fillable oxygen cylinders. Even if medical gas outlets are available in the diagnostic rooms, portable oxygen cylinders will still be required for tests such as oxygen titration studies; therefore a room dedicated to oxygen cylinder storage is required. Staff charting areas should ideally be in a location where staff can maintain visual observation of patients.

One of the most common mistakes in lab design is failing to plan for future growth. Most PFT labs have seen a steady increase in testing volume over the past 20 years. It is far easier (and less expensive) to expand an existing lab than to either build an entirely new larger lab or build a second satellite lab when the demand for services increases. Having adjacent space that can be readily re-purposed is wise. For example, staff offices adjacent to the lab can be relatively easily moved to a different location in the hospital or clinic building so that those offices can be converted into PFT lab expansion space in the future.

Patients who come in for pulmonary function testing are also frequently coming in to see their pulmonologist or coming in to do pulmonary rehabilitation. The best PFT labs are co-located with pulmonary physician offices and pulmonary rehab areas. Having a “one-stop-shop” for pulmonary patients can improve patient satisfaction and can give the clinic or hospital a competitive edge. Having close proximity to a physician or advance practice provider is also useful in the inevitable situations when patients develop medical conditions during pulmonary function testing or exercise testing.

Planning is key

Most people have a hard time conceptualizing what an architectural plan will look like in real-life. It is a good idea to find a large, open area and tape out the dimensions of the planned rooms on the floor. Then add taped out placements for all of the equipment and furniture as well as the door swing area. Then get input from the PFT technicians, an interior designer, the pulmonologist, and the infection control staff. It is far less expensive to get everything right the first time.

August 15, 2021

As I prepared to write this post and discussed it with my colleagues, there were only two responses: this is a great idea and this is a horrible idea. When it comes to abdicating physician clinical responsibilities to advance practice providers, there just is no more polarizing topic in medicine. And the performance of colonoscopy is the most polarizing of the polarizing.

30 years ago, non-surgical procedures were the realm of the physician. As a resident, I was trained in and expected to be able to perform bone marrow biopsies, central venous catheters, endotracheal intubation, flexible sigmoidoscopy, lumbar punctures, and paracenteses. But these procedures have largely disappeared from the portfolio of today’s residency training programs and are now frequently relegated to non-physician practitioners. Recently, our medical center rolled out a bedside procedure team staffed by nurse practitioners who perform central line placement, thoracentesis, lumbar puncture, and paracentesis. Nurse practitioners do nearly all of the bone marrow biopsies in our hospital. Across the country, nurse midwives perform deliveries, respiratory therapists do endotracheal intubation, and physician assistants place dialysis catheters. One of the last bastions of physician-performed procedures is the screening colonoscopy. But should colonoscopy go the way of central lines and bone marrow biopsies? This was the subject of a recent pro and con article in AGA Perspectives publication of the American Gastroenterological Association.

How Many Screening Colonoscopies Need To Be Done Each Year?

Colon cancer occurs in 4.4% of American men and 4.1% of American women. One-third of people who get colon cancer will die of it. Most colon cancers arise from colon polyps and it takes about 10 years for a polyp to turn into a cancer. 30% of American men and 20% of American women will develop a colon polyp at some time during their lives. Screening colonoscopy can identify colon polyps before they turn into cancer and allow for removal of the polyp, thus preventing colon cancer. The current recommendations are for every American to have screening colonoscopy every 10 years from age 50 through age 75. If a polyp is found, then screening should be increased to every 5 years. If a person has a first-degree relative with colon cancer, then screening should start at age 40 and occur every 5 years. When you do the math it works out that:

1. The average low-risk American should get 3 colonoscopies during a lifetime (66% of Americans).
2. The average American with a polyp should get 6 colonoscopies during a lifetime (22% of Americans).
3. The average American with a family history of colon cancer should get 8 colonoscopies during a lifetime (12% of Americans)

Doing additional math, overall, the average American should get 4.3 colonoscopies during a lifetime. Based on the age demographics from the 2016 U.S. Census, on any given year, 4,235,000 Americans should be getting their first colonoscopy and 17,881,000 Americans should be getting a follow-up colonoscopy. Adding these numbers up, if we were to optimally screen every American for colon cancer, we would need to do a total of 22 million screening colonoscopies every year.

Screening colonoscopy is primarily done by gastroenterologists and general surgeons. Currently in the United States, there are 14,000 gastroenterologists and 25,000 general surgeons. If all screening colonoscopy was done by gastroenterologists, then the average gastroenterologist would need to do 1,571 screening colonoscopies per year to completely cover the needs. If screening colonoscopy is done by both gastroenterologists and general surgeons, then they would need to do 637 per year on average. In reality, these numbers are way over-inflated because there will always be Americans who adamantly refuse to undergo colonoscopy, those who wait 11 years rather than 10 years between regular screening colonoscopies, those who are uninsured, and those who don’t undergo colonoscopy because it is pointless if they are dying of some other disease. The current estimate of the actual numbers in the United States are 14 million screening colonoscopies and 3 million screening flexible sigmoidoscopes.

Medicare alone currently spends $1.8 billion per year on outpatient colonoscopies ($416 million just on professional fees). However, since the majority of screening colonoscopies would be done in patients under age 65 (and thus not covered by Medicare), private insurance companies in the United States pay even more than this.

Are We Meeting All Of The Country’s Needs For Screening Colonoscopy?

If all that a gastroenterologist did was screening colonoscopy, then the demands could be easily covered. However, gastroenterologists (and general surgeons) do far more than just screening colonoscopy. They do endoscopy and colonoscopy for purposes other than screening for colon cancer, they do consults, and they do longitudinal management of patients with gastrointestinal diseases.

The economic law of supply and demand predicts that gastroenterologists will migrate to where screening colonoscopy is highly lucrative. A diagnostic colonoscopy is worth 3.26 work RVUs which equates to about $114 for Medicare. Medicaid will pay considerably less, about $70. The Congressional Budget Office estimates that the average commercial insurance plan pays about 1.75 times more than Medicare for colonoscopy, or about $200 for the physician work component. So, in suburban areas serving patients with commercial insurance, there should be plenty of gastroenterologists to cover the demand for screening colonoscopy whereas in an urban area serving patients with Medicaid, there will likely be insufficient gastroenterologists to cover the demands. The same will be true for rural areas and Veterans Hospitals. So, even if we have enough gastroenterologists (and general surgeons) to meet the overall screening colonoscopy needs in the United States, there will inevitably be geographic pockets of unmet needs that will not be supplied by gastroenterologists (and general surgeons).

Can These Areas Of Unmet Colon Cancer Screening Needs Be Met By Non-Gastroenterologists (And General Surgeons)?

The 2018 Medscape Physician Compensation Report indicates that the average gastroenterologist makes $408,000 per year and the average general surgeon makes $322,000 per year. In comparison, the average physician assistant makes $105,000, the average nurse practitioner makes $107,000 per year and the average nurse anesthetists makes about $169,000 per year. There is not enough data to know what the salary of an NP/PA colonoscopist would make but it is probably in between the average NP/PA and a nurse anesthetists, say about $135,000 per year. Therefore, combining the compensation for the professional services portion of a diagnostic colonoscopy with the average salaries we can determine the number of diagnostic colonoscopies it would take to cover salary (benefits not included).

The next part of this question is whether or not non-physicians can competently perform screening colonoscopy. A meta-analysis from 2014 that pooled results of 24 studies found that polyp detection rates, colon cancer diagnosis rates, and complication rates were similar between nurse practitioner/physician assistants and physicians.

How Many Colonoscopies During Training Are Needed To Achieve Proficiency?

The current recommendations for gastroenterologist training is to do a minimum of 275 supervised colonoscopies. However, one study in 2010 suggested that optimal competency requires 500 supervised colonoscopies; a more recent study in 2016 confirmed the 500 procedure per trainee number. Other studies found competency could be reached after 250 or 275 procedures. The problem with many of these studies is the they did not separate purely screening colonoscopy from colonoscopy with interventions (such as polypectomy). Different professional societies have different recommendations for the number of procedures performed under supervision in order to get hospital privileges.

1. American Academy of Family Physicians: 50 colonoscopies
2. American Board of Surgery: 50 colonoscopies
3. American Society for Gastrointestinal Endoscopy: 275 colonoscopies

So, should your hospital train and utilize physician assistants and nurse practitioners to do colonoscopy?

The answer is probably no if:

1. The hospital is in a location dominated by patients covered by commercial health insurance
2. There is currently enough credentialed physicians to meet colonoscopy demand in a timely fashion
3. There is not a wait time for non-procedural services by the gastroenterologists (outpatient consults, etc.)
4. There is not an unmet need for surgical services by general surgeons

The answer may be yes if:

1. The hospital largely cares for a large number of Medicaid and Medicare patients
2. It is a Veterans Administration hospital
3. There is a long wait time for screening colonoscopy
4. There is a long wait time for gastroenterologist consultation because the gastroenterologists are spending too much time in the endoscopy suite
5. Patients cannot get their hernias repaired and their gall bladders removed because the general surgeons are spending too much time in the endoscopy suite.

What Would I Do If I Was Making Up The Rules?

First, a minimum number of supervised colonoscopies during training would need to be established. Using the ASGE recommendations, this would likely be 275. To win over skeptics, it may need to be as many as 500. A training program for physician assistants or nurse practitioners would be costly because of the additional time required by the gastroenterologist (or general surgeon) to supervise a trainee doing a procedure versus the shorter time it would take himself or herself to do the procedure. Furthermore, it is unlikely that many NPs or PAs would do this time of training for free so there would be the cost of paying their salary during the training period. It would likely take about 3-4 months of fairly intensive, full-time procedural training to achieve the minimum of 275 screening colonoscopies. Once you factor in didactic teaching, non-procedural disease management, and training in other procedures (such as paracentesis), then the entire training program would probably be about a year.

Second, a decision would need to be made about whether the NP or PA will be permitted to do polypectomy unsupervised. If the answer is no, then there would need to be a gastroenterologist (or general surgeon) immediately available in the endoscopy suite who could step in to perform or supervise polypectomy. The only way to make this financially viable would be to have one physician available to two or three NPs or PAs simultaneously.

Third, there would have to be up-front negotiation with payers about reimbursement for screening colonoscopies performed by NPs or PAs. For example, if there is a physician present in the endoscopy suite, can the colonoscopy be billed as “incident to” service or would it be reimbursed as an NP/PA independent service (which generally pays 15% less)?

Fourth, there will need to be a decision made about upper endoscopy. For example, should an NP/PA be able to do routine screening upper endoscopy for patients with Barrett’s esophagus?

What Would My Ideal Program Look Like?

1. I would start off by having 3 trained NPs or PAs doing screening colonoscopies with a gastroenterologist (or general surgeon) present in the endoscopy suite for polypectomies, complication management, or diagnostic questions. Statistically, at any given time, 1 of the 3 NP/PAs will be doing a colonoscopy that will require the physician’s presence.
2. I would create a monitoring office where the video feeds from all 3 individual procedure rooms could be fed into monitors so that the supervising physician could periodically check the progress and findings of the procedures.
3. The NP/PAs would also be trained and credentialed in paracentesis so that they could also be doing those time-intensive paracenteses that gastroenterologists do not want to do.
4. I would create a hospital-paid position of “Director of Endoscopy Training” to provide financial support for the time a gastroenterologist (or general surgeon) spends supervising the 275 screening colonoscopies that the NP/PA in training requires.
5. I would make it financially lucrative for the gastroenterologist (or general surgeon) to supervise the 3 NP/PAs by splitting the work RVUs with the physician in a way that makes supervising screening colonoscopy more attractive than performing it solo themself.
6. During the time that the NP/PAs are not doing colonoscopy, I would have them follow up biopsy results, arrange patient follow-up, complete procedure notes, see inpatient consult follow-ups and, see outpatient return office visits.
7. I would create a peer-review process whereby the recorded videos of the colonoscopy withdrawal by the NP/PA would be reviewed by another of the NP/PAs so that each NP/PA would have 25 procedures reviewed per year for quality purposes.
8. Eventually, I would create a pathway for the NP/PAs to be able to perform polypectomies independently. Because I would anticipate a considerable amount of skepticism about a non-physician colonoscopy program, I would phase in polypectomy a few years after the initial screening colonoscopy program with the requirement that the NP/PAs perform some minimal number of supervised polypectomies (for example, 150).

Probably the biggest barrier will be the professional threat that gastroenterologists may feel to having a non-physician do colonoscopy. For most gastroenterologists, colonoscopy is an integral part of their core identity as a specialist. They take pride in their art and skill of colonoscopy and take pride in the number of lives they save by colon cancers prevented. However, medicine is increasingly requiring a team approach and it may be time for us to consider screening colonoscopy as a team sport rather than as an individual sport.

May 4, 2019

No test in pulmonary medicine is fraught with more confusion and mystery than the cardiopulmonary exercise test, or CPET. This is a test that reports a large number of physiologic measurements during exercise when the patient is asked to exercise to his or her maximum effort. On the one hand, there is a wealth of information in all of those numbers but on the other hand, most physicians don’t know what to do with all of those numbers. As a consequence, the CPET is one of the most under-utilized tests in the hospital. There are several very detailed guides to CPET interpretation, such as the ATS/ACCP Statement on CardiopulmonaryExercise Testing, but these have so much detail that they can be difficult to approach by the average physician. In this post, I will go over a very basic approach to CPET interpretation, albeit one that is fairly oversimplified. I will then cover a second, slightly more advanced approach to CPET interpretation.

Physicians of different specialties use the CPET for different purposes and look at different physiologic measures. I am going to focus on the use of the CPET from the pulmonologist’s vantage point. Although there are many reasons for ordering a CPET, the most common reasons are (1) to determine the cause of a patient’s dyspnea, (2) to determine if a patient can tolerate surgery or lung resection, and (3) to determine whether a patient is disabled.

The test is done by having the patient ride a stationary bicycle (although a treadmill can also be used). The resistance on the bicycle is gradually increased making it harder and harder to pedal and the patient exercises until they can’t go any longer. The patient breathes through a mouthpiece and has a number of monitoring devices that measure:

1. Heart rate
2. EKG
3. Blood pressure
4. Oxygen saturation
5. Respiratory rate
6. Minute ventilation (amount of air that the patient breathes in a minute)
7. Exhaled carbon dioxide concentration
8. Inhaled and exhaled oxygen concentration

From these different measurements, a computer is able to calculate a large number of physiologic variables and can generate a large number of physiologic graphs. This is where the CPET report can get overwhelming for many physicians in that there are so many numbers and graphs that it can seem like you are drowning in data. But there is good news… you can ignore most of the data and still get the information that you need from the CPET most of the time. Here are two ways to approach the CPET, a basic interpretation approach and an advanced interpretation approach.

The Basic Approach

This will answer the questions “Is the exercise impaired?” and “If so, is impairment due to heart disease or lung disease?”.

1. Look at the mVO2 (maximum oxygen uptake). If it is reduced, then there was abnormal impairment to exercise. If it is normal, then the patient can exercise normally and is not impaired; most of the time, this means that you can stop here.
2. Look at the reason for stopping exercise. Most of time, patients will stop because of leg fatigue or shortness of breath. However, if the patient has chest pain, it may be a clue to myocardial ischemia. If the patient had calf pain, it may be a clue that they are limited by claudication. If they fell off the bike or couldn’t keep the mouthpiece in place, then the test is invalid.
3. Look at the heart rate. If the patient reached a maximum predicted heart rate, then exercise was limited by the cardiovascular system. A normal predicted heart rate = (220 – age). Anything > 90% of maximum is considered abnormal. Note: normal people are limited by their cardiovascular system so if the mVO2 is normal (indicating no exercise impairment), then the patient should reach a maximum predicted heart rate. If the patient is taking a beta blocker, they will not reach a maximum heart rate and so the heart rate analysis will be indeterminate.
4. Look at the mVE (maximum minute ventilation). A normal person should have an mVE that is < 75% of predicted, in other words, a normal person’s exercise is never limited by their lungs. Another way of saying this is that we are all born with 25% more lung than we actually need. The predicted maximum minute ventilation can either be calculated by the equation [FEV1 x 40] or by having the patient breath as rapidly as possible while sitting at rest for 30 seconds then multiplying the result x 2 (this is called the maximum voluntary ventilation or MVV). The difference between the predicted maximum minute ventilation and the actual maximum minute ventilation during the CPET is the “ventilatory reserve”. If the mVE is > 85% of predicted (i.e., the ventilatory reserve is < 15%), then the patient is abnormally limited by pulmonary disease.
5. Look at the oxygen saturation. A significant drop is > 4%. The oxygen saturation does not drop when a normal person exercises. If the oxygen saturation falls, then there is likely pulmonary disease.
6. Look at the blood pressure. Normally, the systolic pressure goes up during exercise but the diastolic pressure stays the same. If the diastolic pressure rises, then hypertension could be the cause of the patient’s limitation. Normally, at peak exercise the blood pressure should be < 220/90.
7. Look at the EKG. If there are ischemic changes, then heart disease may be the cause of exercise limitation.

Using the basic approach will usually allow you to stratify patients into 4 categories of exercise limitation: (1) normal, (2) cardiac, (3) pulmonary, or (4) non-cardiac, non-pulmonary.

 

This table is not perfect. For example, many patients with heart disease may reach a maximum predicted heart rate but have a normal diastolic blood pressure and no ischemic changes on the EKG. Similarly, some patients with lung disease may have a maximum minute ventilation > 85% of predicted but do not desaturate. In other words, patients with abnormal cardiovascular or pulmonary limitation do not need to have all of abnormalities listed above, just one of them. Also, the category of “non-cardiac, non-pulmonary” is very broad and can include patients who gave a poor effort on the test, patients with neuromuscular disease, patients with peripheral vascular disease, those who are deconditioned, and those who are obese.

If you are ordering the CPET to determine if a patient is able to undergo lung resection surgery, you can just focus on the maximum oxygen uptake (mVO2). Those patients with a relatively preserved mVO2 can generally undergo lung resection safely. This can be a particular help in those patients with lung cancer and a low FEV1 who might appear to be high risk based on spirometry. However, the cardinal rule of lung cancer treatment is to never miss an opportunity to send a surgically curable patient to a surgeon, so the CPET can sometimes give you the confidence you need to proceed with surgical resection.

If you are ordering the CPET to determine disability, then there are gradations of disability that are used in the AMA’s Guides To The Evaluation Of Impairment that are often accepted by disability granting agencies. Disability can be stratified by disability class or by the percentage of the whole person that is impaired based on the mVO2.

The Advanced Approach

If the basic approach is “CPET 101” (freshman level CPET interpretation), then the advanced approach is “CPET 201” (sophomore level CPET interpretation). There is an even higher level approach to CPET interpretation (“CPET 401”) that is covered by the ATS/ACCP Statement on CardiopulmonaryExercise Testing. The advanced approach relies on analysis of the anaerobic threshold and analysis of the physiologic graphics generated by the CPET.

The anaerobic threshold occurs when a person’s muscles switch from aerobic metabolism to anaerobic metabolism. When this happens, the muscles are not consuming oxygen to produce energy and the muscles start to produce lactic acid. The body keeps producing carbon dioxide so the VCO2 keeps going up, but the body doesn’t take up greater amounts of oxygen so the VO2 stops rising as quickly. The body can only sustain anaerobic metabolism for a short time, however. When the muscles shift into anaerobic metabolism, a number of things happen:

1. The blood lactate level rises. However, serial lactate testing is usually not used in a CPET since it would require blood draws every 30-60 seconds.
2. The production of carbon dioxide rises disproportionate to the consumption of oxygen. This can result in the VCO2 (carbon dioxide production) rising faster than the VO2 (oxygen consumption). In this graph, The VCO2 is plotted on the vertical axis and the VO2 is plotted on the horizontal axis. In normal aerobic metabolism, the relationship between the VCO2 and VO2 should be a straight line. Once the exercising patient shifts into anaerobic metabolism, the VCO2 starts to go up faster than then VO2. This is shown in the graph where the blue dots (that represent individual measurements) start to rise rapidly. This defines the anaerobic threshold, shown by the green line. This is sometimes called “a change in the V-slope”. The change in VCO2 compared to the VO2 can also be seen in this graph of the VO2 and VCO2 plotted against the total work performed. In this case, the VCO2 (blue dots) begins to rise faster than the VO2 (red dots) when the patient switches into anaerobic metabolism result in the two curves crossing at 75 watts of work.
3. The respiratory quotient starts to increase. A normal respiratory quotient (RQ) in a person who is burning up carbohydrates during metabolism is 1.0. If they are metabolizing fat, the RQ is 0.7 and if they are metabolizing protein, the RQ is 0.8. When carbon dioxide is derived from lactic acid, the RQ goes up to a level above 1.0. In this graph, you can see the rise of the RQ (green dots) above the RQ of 1.0 (red dashed line) – this sudden increase in the RQ indicates that the exercising patient just crossed the anaerobic threshold. Sometimes, the RQ (respiratory quotient) will be reported as the RER (respiratory exchange ratio) in a CPET report.
4. The end-tidal oxygen concentration (PETO2) begins to rise. When the muscles switch to anaerobic metabolism, they are no longer primarily consuming oxygen to make energy. Consequently, the lungs extract less oxygen out of the inhaled air. This results in the oxygen in the exhaled air (the PETO2) to start to go up. In this graph, the PETO2 was fairly stable at about 110 mm Hg during aerobic metabolism but when the muscles switched to anaerobic metabolism, the PETO2 began to rise (yellow arrow) at the 10 minute mark of the CPET.
5. The ratio of minute ventilation to oxygen uptake begins to rise. During normal aerobic exercise, this ratio is fairly consistent but when the muscles switch to anaerobic metabolism, there is less oxygen being taken up by the lungs (as we saw with the rise in the PETO2 in the previous paragraph). This results in a change in the ratio of the minute ventilation to the oxygen uptake (VE/VO2) so that as the VO2 goes down in anaerobic metabolism (in the denominator of the VE/VO2 equation), the ratio VE/VO2 goes up. This is shown by the rise in the red dots marked by the yellow arrow at the 13 minute mark of this CPET. In this case, the VE/VCO2 also goes up after anaerobic threshold but to a lesser degree (blue dots).

In the best of possible worlds, all of these markers of anaerobic threshold should happen at the same time. Unfortunately, the anaerobic threshold is not always easy to determine. The graphs never seem to work out like the examples that are usually shown in textbooks. There can be leaks in the mouthpieces and other causes of subtle errors in measurements. Furthermore, anaerobic threshold is not just an all-or-none event for all of the muscles in the body at exactly the same time – some muscle groups go into anaerobic metabolism sooner than others so the anaerobic threshold that we seen on the CPET is really an average of lots of muscle groups all going into anaerobic metabolism at different times. A normal person should reach anaerobic threshold at about 50-60% of their mVO2; if the patient reaches anaerobic threshold at < 40% of the predicted mVO2 then the anaerobic threshold is considered reduced.

Another variable that can be useful is the dead space determination. Normally, to calculate the dead space, arterial blood gases are necessary but these are generally not done during most CPETs. You can get a rough estimate of the dead space during exercise by looking at the VE/VCO2 at anaerobic threshold. A value greater than 34 indicates an abnormally increased dead space. This can be caused by lung diseases, pulmonary hypertension, or heart failure. If the dead space goes up extremely high during exercise, think about pulmonary hypertension.

By first determining if the patient actually exercised to the point of anaerobic metabolism and then determining the anaerobic threshold, you can do a deeper analysis of the CPET for the advanced approach:

1. Look at the mVO2 (maximum oxygen uptake). If it is < 85% of predicted, then there was abnormal impairment to exercise. Most CEPT reports will give the predicted mVO2 based on the patient’s actual body weight. However, if the patient is obese, then it is very helpful to also calculate the predicted mVO2 based on the patient’s ideal body weight. This correction can help you determine if a patient’s low mVO2 is just due to their obesity.
2. Determine if there was respiratory limitation. To assess this, you will need to look at the mVE, the dead space, and the oxygen saturation. There are 3 abnormalities that can indicate abnormal respiratory limitation:
   1. mVE > 85% of the MVV (or > 85% of [FEV1 x 40])
   2. Increased dead space as defined as an increased VE/VCO2 of > 34 at anaerobic threshold (or > 40 at peak mVO2)
   3. Oxygen desaturation of > 4% from baseline.
3. Determine if there was cardiovascular limitation. There are 3 abnormalities that can indicate abnormal cardiovascular limitation:
   1. The mVO2 is low and the patient reaches > 90% of the maximum predicted heart rate (220 – age)
   2. There are ischemic changes on the EKG
   3. There is an abnormal blood pressure response. The could either be drop in the systolic pressure with exercise or an abnormal rise in diastolic blood pressure (>90 mm Hg) with exercise
4. Determine if the anaerobic threshold was reduced. There are many diseases that can cause a reduced anaerobic threshold (defined as an anaerobic threshold occurring at < 40% of the predicted mVO2) and so if there is an isolated reduction in the anaerobic threshold, you will need to search for the cause. Some of the more common causes are:
   1. Hypoxemia during exercise (from lung disease)
   2. Pulmonary vascular disease
   3. Liver failure
   4. Renal failure
   5. Cardiac disease (including ischemia, heart failure, valvular disease, and conduction disease)
   6. Anemia
   7. Peripheral vascular disease
   8. Neuromuscular disease

Using this analysis, the results can point toward several conditions:

 

Here is an example of a patient with combined diastolic heart failure, secondary pulmonary hypertension, coronary artery disease, plus interstitial lung disease. The clinical question was which of his diseases was responsible for his shortness of breath? The anaerobic threshold was determined to occur at 10.09 minutes (best seen on Plot 6 of the graphs). First, look at the mVO2 – in his case, it was reduced at 12.4 ml/kg/min (67% of predicted), so exercise was abnormally impaired. Second, determine if he had an adequate ventilatory reserve – he did not since his mVE was 73.8 L (98% of predicted) – this indicates lung disease as a cause of his exercise limitation. Third, determine if he desaturated – he did not since his saturation at peak exercise was 97%. Fourth, determine if he had an excessive dead space – he did since his VE/VCO2 at anaerobic threshold was 36 (anything over 34 being abnormal) – this is further evidence of lung disease. Fifth, determine if he reached his maximum predicted heart rate – he did not since his heart rate at peak exercise was 93 (62% of normal) indicating that it was not his cardiovascular system that limited his exercise. Sixth, determine if the anaerobic threshold is reduced – his anaerobic threshold occurred at a VO2 of 6.4 ml/kg/min which is a VO2 of 34% of his predicted mVO2 (6.4 ÷ 18.5) and therefore reduced since it is less than the normal threshold of 40% of the predicted mVO2. Seventh, determine whether the blood pressure response is normal – his systolic blood pressure rises normally to 160 but his diastolic blood pressure also rises from 60 to 78, which is abnormal – this probably is related to his diastolic heart failure but since the diastolic blood pressure remained < 90 mm Hg, it was probably not the primary limit to exercise. From the exercise test, we can determine that it is his lungs (in his case, the interstitial lung disease), not his heart that is the cause of his impairment.

As with many tests in medicine, the results are best interpreted in the context of the individual patient. Therefore, the physician who is familiar with the patient’s history and has done a physical examination is in the best position to accurately interpret the CPET. But for those patients who have shortness of breath and you are unsure if it is due to undiagnosed heart disease, undiagnosed lung disease, obesity, or deconditioning, the CPET can often be tremendously helpful. Also, in those patients who have both known heart disease AND known lung disease, the CPET can help determine which of the two diseases are the cause of shortness of breath and exercise limitation.

January 8, 2018