Q: Can you please help me determine the query opportunities and code assignment/sequencing argument related to a patient who was admitted with pneumonia, congestive heart failure, acute respiratory failure, and encephalopathy? I thought that the pneumonia would be the primary and the respiratory failure as secondary as the severity of illness/risk of mortality (SOI/ROM) as well as the MS-DRG would all increase. Yet, I’m getting some push back on this train of thought and I’m not sure where the error in my logic may lay. Any insight you could offer would be much appreciated!
A: This is an interesting and common question. I believe you are asking why would we choose the acute respiratory failure as the principal diagnosis when, if we choose the pneumonia with a secondary diagnosis of acute respiratory failure, we have an MCC and it would provide both higher reimbursement and SOI/ROM. Coders and CDI specialists were once taught that when two or more competing diagnoses are present on admission and they could choose the one providing the highest relative weight/increased reimbursement.
In our CDI Boot Camp we explore the concerns around code sequencing in depth and part of our in-class discussion challenges that traditional thought a bit and perhaps spurs our students to think about the matters differently. The traditional choice, to simply code the highest weighed MS-DRG as the principal diagnosis, often gets challenged by auditors. Which condition, they rightfully ask, actually occasioned the admission? Does a typical patient with pneumonia require an inpatient admission? Not usually. When will the physician discharge this patient; when the pneumonia is resolved, or when the acute respiratory failure is no longer a concern?
My bet is that the physician will send this patient home on antibiotics (treatment for the pneumonia) so the pneumonia is not exactly resolved on discharge is it?
I recently assisted an organization with two DRG validation denials from recovery auditors. The auditors agreed with all coded diagnoses but argued the sequencing choices involved. One such account was acute systolic heart failure and acute respiratory failure. Their argument was the choice for admission was based on the acute respiratory failure, not the heart failure. I could not defend against that logic. Once the patient was able to breath without intervention or assist, she was discharged. She was sent home with adjusted medication for her heart failure continuing and follow up with the cardiologist.
I am not saying you should always sequence one way or another, rather, as a CDI specialist it’s your role to closely review the circumstances of admission and carefully consider which conditions meet the definition of principal diagnosis as “That condition after study that occasioned the admission.”
The Official Guidelines for Coding and Reporting tells us it should be a rare instance that two or more diagnoses qualify as the principal and we believe this guideline is used much more frequently than it should be.
Lastly, just to throw another log on the proverbial fire, depending on the circumstances in your example perhaps the encephalopathy could be the principal admission. Again it would depend on the circumstances of this patient and the treatment rendered. This example is a great one to discuss with your fellow CDI specialists and coders.
Editor’s Note: Let’s call this week’s guest post from “a ghost of ACDIS’ past.” Those who’ve worked in the profession for some time will no doubt remember with some fondness the teachings of founding ACDIS Advisory Board member Robert S. Gold, MD, who sadly passed away in the spring of 2016. The following was a note sent some years before that, encouraging ACDIS and the CDI professionals it represents to advance their efforts beyond traditional CC/MCC capture to ensure complete and accurate medical record.
It so often seems that all CDI programs have been developed based on MS-DRG maximization. Such short-term focus however defeats the goals (and challenges) of healthcare in today’s society. When CDI develops from the Medicare revenue perspective, rifts grow quickly between the coders and the CDI because so many of the teachings don’t line up. Rifts develop between the CDI folks and the docs, too, because docs feel like they are targets and are always wrong. Rifts develop between the CDI team and quality team because CDI queries seem to encourage the reporting of complications even when complications don’t exist. And massive holes exist in the patient’s database because of CC/MCC capture concentration which leaves all of the patient’s chronic conditions unsought after and unreported.
Too many instruments support the old, obsolete concepts and too few encourage people to go beyond their silos. All of the initiatives that are needed to drive us into the future will fail if we don’t expand our expectations beyond these past, financially focused, efforts.
Check out this video from CMS from last year. How has your facility changed its documentation improvement efforts around diabetes and how have the recent ICD-10-CM/PCS coding updates, Official Guidelines for Coding and Reporting, and new recommendations from Coding Clinic affected those efforts?
Let us know in the comment section. It’s a concern lots of CDI specialists struggle with.
by Sharme Brodie, RN, CCDS
The AHA’s Coding Clinic for ICD-10-CM/PCS had both its third and fourth quarter 2016 editions published at the same time this fall. The combined publication is very long—the longest I have ever reviewed—so let’s start by looking at some of the changes to the ICD-10-CM Official Guidelines for Coding and Reporting.
Trouble with ‘with’
As many of you have probably heard, Section A.15 of the new Guidelines states that any time the word “with” appears, either in the Alphabetic Index or in an instructional note not in the Tabular List, the classification presumes a causal relationship and the conditions will be linked even in the absence of provider documentation. Coding Clinic reiterates this fact. For CDI specialists, this may seem like a good thing, but it also means we need to do our homework and make sure all these assumed relationships are truly related. We may find we need more queries to verify these cases with the providers. So be careful and go over each case to make sure it is an accurate representation of what is going on with your patients.
Coding Clinic (p. 119) stresses the new Guidelines found in Section A.19, which state that assignment of a diagnosis code must be based on the provider’s diagnostic statement that the condition exists, not on clinical criteria used by the provider to establish the diagnosis. That’s not to say the clinical criteria don’t matter, but that coders and CDI specialists cannot decide when a condition exists based on whether we feel certain criteria are appropriate for a condition.
According to Coding Clinic (p. 8), coders have questioned whether ICD-10-CM codes for sepsis may be assigned based on the new clinical criteria that were released in February 2016, The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). This Coding Clinic states “coders should never assign a code for sepsis based on clinical definition or criteria or clinical signs alone.”
Editor’s Note: There’s a lot more to cover in these oversized editions of the fall 2016 Coding Clinic, and Brodie explores the highlights in an extended white paper available in the Resources section of the ACDIS website. Brodie is a CDI education specialist for HCPro in Middleton, Massachusetts. Contact her at firstname.lastname@example.org. For information regarding CDI Boot Camps offered by HCPro, visit www.hcprobootcamps.com.
Just because you can measure something doesn’t mean should measure it, Jon Elion, MD, founder and CEO, ChartWise Medical Systems, told ICD-10 Monitor’s Talk Ten Tuesday back in February of 2016. Specifically, he pondered why CDI programs hold onto legacy targets such as physician response rates as a measurement of program success. Instead, measures need to change as the CDI program itself matures and grows.
As CDI programs begin, managers need to track:
- physician query rates (how many queries per record reviewed)
- Physician response rates (how many of those queries receive a response either positive or negative)
- physician agree rates (how many query responses align with the intent of the CDI specialist)
Such measures help CDI programs assess the competency of its staff as well as the engagement of its physicians with program goals. Yet any assessment tool needs to be analyzed itself and the reason behind the data interrogated, Elion indicated.
For example, if the CDI team sends a query to one physician but a different physician responds should that response count towards the first or second physician’s data? What about the physician who responds to all queries, just not in a timely way? What about the physician who answers all their queries but always needs to be queried on the specificity of his or her patient’s heart failure?
Theoretically, as a CDI program matures its physician engagement and response rates should increase throughout the first year and stabilize in the second or third years. While programs may not expect 100% consistent physician response rates they should expect it to hover in the mid-90% range.
Elion doesn’t mean that CDI managers should toss out those traditional measurements altogether but to use them instead to identify potential trouble spots, educational opportunities, and to nip any provider support concerns in the bud early.
Such measures should not be used to penalize physicians, however. “Clinicians who are always on their toes would suffer from a toe walking gait which most closely maps to ICD-10 code R26.89,” he said.
What types of metrics does your CDI program measure and what items to you think should be retired?
by Paul Evans, RHIA, CCS, CCS-P, CCDS, and Anny Yuen
The debate regarding which profession makes the “best” CDI specialist unfortunately continues. Many facilities and consulting firms, initially trained to believe that only nurses could perform the duties of a CDI specialist, continue to propagate such expectations.
Yet we believe other clinicians (e.g., physicians, physician assistants, foreign medical graduates) and nonclinicians (e.g., coders and health information management [HIM] professionals) also perform well in the CDI role with appropriate training.
When considering candidates for an open CDI position, CDI managers need to take a closer look at their initial job descriptions and make sure they accurately reflect not only the current needs of the department and the expanded role CDI specialists need to play, but also changes in industry expectations.
We’ve seen instances where programs take sample roles and responsibilities wholesale, and fail to customize their expectations or include professionals outside nursing. It has long been ACDIS’ stance that facilities should find the candidate best suited to the particular position. ACDIS has long expressed itself as an inclusive organization, welcoming coders, nurses, physicians, case managers, quality staff, and all who are interested in learning more about the value of complete and accurate documentation in the clinical record.
Further, in order to sit for the Certified Clinical Documentation Specialist (CCDS) credential, ACDIS lists out several levels of required education and skills. Among them, professionals must have an associate-level college degree—as the role of CDI specialist requires a high level of cognitive analysis and the integration of significant clinical acumen and awareness of healthcare reimbursement processes.
On the coding side, educational differences between those holding the Certified Coding Specialist, the Registered Health Information Technician, and the Registered Health Information Administrator® credentials are vast and may include formal college credits, anatomy and physiology, pharmacology, and pathophysiology, among other areas. However, because some facilities do not require coders to have advanced degrees, the conventional wisdom often gets reiterated—that coders in general have no clinical training or knowledge.
As HIM professionals engaged in CDI efforts at our facilities and as active members of the ACDIS Advisory Board, we stand to represent those from the coding side of the house who have effectively leveraged their experience to help advance the CDI mission. We have proven that we can per- form duties as CDI specialists and lead successful departments, while promoting the team dynamic between providers and HIM and CDI.
Not all coders can serve as CDI specialists, and neither can all nurses. Being a CDI specialist takes creativity and strong understanding about clinical documentation and indicators. The first step to true collaboration requires a deeper awareness and appreciation of the talents each individual, regardless of professional background, brings to the table.
Editor’s Note: This article originally published in the Sept./Oct. edition of the CDI Journal.
by Susan Browne, RN, MN CCDS
ACDIS has been central to my career as a CDI specialist. From attending the national convention every year, earning my CCDS credential, and hosting three ACDIS Washington State Evergreen Chapter meetings over the past six years (with much help!), I have met individuals from all over the country and nearby communities, including RNs, coders, and physicians with unique beginnings, but merging on the same path, all with the same goals and struggles. ACDIS has provided education, guidance, community, and a “big picture” vision for me.
Since its inception, our Washington State Evergreen Chapter persevered through bursts of enthusiasm and a paucity of hosting venues. We shared dessert potlucks and played CDI Jeopardy. We moved from local speakers to presentations from national consultants such as Berkeley Research Group and a lunch hosted by CDI Search Group. We have evolved as individuals, professionals, and as an organization.
This past year has been a hallmark for us as evidenced by the 40+ attendees at our summer meeting, where six individuals volunteered to step up and take accountability for the future of the Washington State Evergreen Chapter. They have already hosted a second meeting and scheduled a third (which will take place via webinar on January 20), with goals of being inclusive to the entire state and providing online education during the winter months. This new leadership team is enthusiastic and synergistic, eager to maintain momentum, and wanting to share the priceless resource of “networking” with other CDI professionals in our state.
I step away from the leadership group knowing it is in excellent hands and can’t wait to see what the future holds! Thank you for your dedication and hard work, Nora Tiffany, Barbara Anderson, Lillian Dickey, Satyananda Vuddagiri, Andrea Dyson, and Martha E. Khayyat, among other volunteers.
Editor’s note: Browne is the director of clinical documentation improvement at Capital Medical Center in Olympia, Washington. Contact her at Susan.Browne@capellahealth.com. Learn more about the Washington ACDIS Chapter and other networking opportunities under the “Networking & Events” tab of the ACDIS website www.acdis.org.
Q: I wanted to write you, concerning a denial letter I have received, regarding morbid obesity. The letter states:
“Although morbid obesity and BMI were documented, from the information received [we] could not verify documentation that supported obesity required increased monitoring, evaluation, diagnostic/therapeutic treatment or increased length of stay, therefore it does not meet criteria for assignment as an additional diagnosis code assignment.”
It went on to say that the diagnosis code does not meet criteria for reporting on the claim due to the ICD-9-CM Official Guidelines for Coding and Reporting rules for assignment of other (additional) diagnosis, UHDDS’ definations of other diagnoses, and other Guidelines (they cited Section I, General Coding Guidelines and Section III, Reporting Additional Diagnoses).
We fought the denial from two angles. One the physician did recommend weight loss and dietitian was monitoring the patient. Second, Coding Clinic address this question specifically. Coding Clinic, Third Quarter 2011, pp. 3-4. This addresses the question of the clinical significance of obesity. Also, Coding Clinic, Third Quarter 2007, pp. 13-14, addresses the issue of coding of chronic conditions. I am still waiting on an answer, but wanted to see if you would share your opinion regarding this situation.
A: I agree with you fully on the obesity. Coding Clinic states:
“Individuals who are overweight, obese, or morbidly obese are at an increased risk for certain medical conditions when compared to persons of normal weight. Therefore, these conditions are always clinically significant and reportable when documented by the provider.”
This statement is fairly definitive and I would encourage you NOT to make any facility specific policy changes based on this inappropriate audit finding.
Editor’s note: Allen Frady RN-BSN, CCDS, CCS, CDI Education Specialist for BLR Healthcare in Middleton, Massachusetts, answered this question. Contact him at AFrady@hcpro.com. For information regarding CDI Boot Camps visit http://hcmarketplace.com/clinical-doc-improvement-boot-camp-1.
Last year at around this time, the ACDIS Advisory Board released a white paper reviewing the role of CDI specialists in assessing information in the medical record from prior treatments.
Codes cannot be assigned based on previous conditions. However, there’s a gray area clouding whether CDI professionals can pull information forward to clarify a diagnosis being treated during the current episode of care, says Cheryl Ericson, MS, RN, CCDS, CDIP, manager of CDI services at DHG Healthcare, during an ACDIS Radio discussion on the topic.
ACDIS created the white paper as a means to help CDI programs open a dialogue about such concerns within their facilities and to help CDI managers begin to craft policies and procedures around compliant and ethical practices regarding electronic health record interrogations.
In particular, CDI specialists face the dilemma of whether to apply information from prior encounters when querying a physician in order to clarify a diagnosis documented in a current admission or episode of care. The CDI profession is divided on this topic: Some are comfortable referencing the historical information within the query when it clarifies a currently documented condition relevant to the current episode of care; however, others believe this practice violates Uniform Hospital Discharge Data Set (UHDDS) definitions regarding an episode of care, as well as coding guidelines.
The paper reviews overarching guidelines and weighs various references such as reporting additional diagnoses and the definition of the term “encounter,” to help CDI programs begin to assess their own practices.
In Arizona where Judy Schade, RN, MSN, CCM, CCDS, works as a CDI specialist at Mayo Clinic Hospital, the population includes a large
number of “snowbirds,” retirees who travel to warmer climates for the winter. For these patients, information included in the electronic medical record often represents an important link between the current encounter and conditions which may have developed in another setting since their last hospital visit.
“We might not have the most current information so we need to be careful and to ask the provider where additional information may be needed to validate a diagnosis and pull it forward,” Schade says.
“It’s not enough for the physician to say this is a complex patient,” Ericson says. “They have to document it. If someone has hypertension they’re clinically always going to have hypertension. However, we cannot automatically make that assumption in coding that’s why the physician has to document ‘history of,’ or ‘chronic,’ or something else that is affecting this episode of care and the resources directed toward treating it.”
Such information “is so much more accessible” due to extensive use of electronic health records than it was in the past, ACDIS Director Brian Murphy says. CDI specialists need to determine whether looking back in the medical record, or opting not to look back, artificially limits a CDI professional’s ability to capture diagnosis specificity or whether concerns regarding the compliance of such activities are valid.
For example, Schade cautions that CDI specialists could be pulling forward outdated or inaccurate information as well intentioned as they may be. So “partner with different departments to formulate your policies. We’re moving in a different way of looking at things so we really need to carefully examine this process and develop the best practices,” she says.
The white paper walks through some common concerns but also recommends reviewing recommendations from the Joint Commission, CMS, and your own facility’s compliance, IT, and coding policies, for example.
by Rose T. Dunn, MBA, RHIA, CPA, FACHE, FHFMA, CHPS
The coding function would not exist if there were no physician documentation from which to code. The challenge for coders is not so much that there is no physician documentation; it’s that the documentation they have does not provide them with what they need to assign an accurate and specific code.
The reason for this is that the physician is capturing the clinical nuggets he or she needs. “The physician mind is focused on the associated process of evaluating, treating, and managing the health conditions presented by each patient. The chart documentation provided by the physician is all framed in the language of diagnostic phrasing and language, and [that language] certainly is not about codes,” (Insights from the HCC, n.d.).
Many hospitals have implemented clinical documentation programs to coach physicians on the documentation elements required for the hospital to optimize its coding efforts. However, if physicians don’t feel they or their patients benefit from efforts to alter documentation practices, they quickly disengage. Physicians are accustomed to being paid by their evaluation and management level, not their diagnoses.
Thus, the lack of precision of documented conditions, such as pancreatitis without specifying whether acute, idiopathic, alcohol-induced, sclerotic, or not indicating length of laceration or use of anesthetic or even providing details without a diagnosis (e.g., glomerular filtration range provided to indicate stage of kidney disease) is not unexpected, according to Lucyk et. al. from the University of Calgary (Lucyk, Tang, & Quan, 2016). No, it’s not just a United States problem.
At the office
For the physician practice, the superbill, or encounter form, often contributes to incomplete, unspecified, and inaccurate coding. The superbill often lists the most common diagnoses seen in the practice.
Conditions treated may be forced into those listed on the form or not captured at all. If the physician documents the condition, it may be generic (e.g., asthma, without indication of whether it’s intermittent, persistent, mild, moderate, or severe). In preparation for ICD-10, many practices took the convenient route and selected the unspecified code for each of their most common diagnoses. Therefore, the more specific ones are not on the form to be selected.
Even if the physician uses the electronic health record (EHR) for diagnosis selection, when a long list of choices appears for the condition, it is unlikely that the physician will take the 30–60 seconds to glance down the list to find the specific one; worse yet, the condition that previously had a code may no longer have one (e.g., accelerated hypertension). The physician may believe that for the current state, diagnoses may not be needed for reimbursement purposes, but that’s short term thinking.
Helping the physician help us
For the physician practice, if we desire details, we need to push some of the responsibility out to others in the practice. It should start at the front desk when the appointment is made and the health questionnaire is returned from the patient. That is when, at minimum, duration or date of onset can be obtained.
There’s value in using the medical assistant (MA) to help with securing some of the diagnosis details needed for ICD-10 and augmenting the physician’s efforts. MAs are typically members of the physician office team. If certified, these individuals have completed a structured education program with courses in anatomy, medical terminology, coding, and disease processes. Physicians can benefit from the talents of their MAs and possibly in areas other than ICD-10.
They are one of the first clinical team members to speak with the patient, often collecting the patient’s initial history information, capturing specimens for lab tests, and in some states placing, initiating an IV, and administering IV medications. Since organizations are struggling with capturing start and stop times for IV infusions, perhaps the MA may be another option for capturing start and stop times.
Given their understanding of medical terminology, and with an orientation to ICD-10 code requirements, MAs can quiz the patient and capture some of the details often overlooked by physicians. MAs can save physicians time, supplement the physician’s documentation, and help the physician select a more specific code.
If we look at the ICD-10 injury code elements, most of the elements can be captured in whole or in part by the MA in a short interview with the patient:
- What was the injury? The MA can query the patient for this information and capture “upper/lower” and laterality, as well.
- When did it happen? The MA can help the physician establish whether this is an initial encounter for active treatment, whether the patient is in the healing stage, or if the condition is sequela.
- Where did it happen? Knowing the patient fell at home will not get us to the most specific code. We need to know where in the home, and sometimes even need to ask for the type of home.
- What was the patient status and what was the patient doing when the injury happened? If the patient has been bitten by a cat, it may be attributed to a patient status of other, but if the person bitten by the cat was a vet tech when she was holding the cat for the vet to give it an injection, the status leads to an activity for income. Assigning the code for the activity of “holding a cat” would lead to the Y code for animal care.
This example shows us that with a little bit of prodding from the MA, we can get the additional information we need for a specified code. [more]