This is fourth of a multi-part series on a patient’s experience managing a chronic condition, in this case hypertension. A diagnosis has been made, and our patient has hopefully followed up and has hopefully been maintained on appropriate therapy (there is a 1 in 3 chance that this is happening). Now it is time for our patient’s health care sponsor (such as his employer) to review the health care benefit.

Click on the image to see it larger size

Patient Story (Frydman)

There is no patient story in this phase. At some point during the year, our patient’s employer will discuss provided health care coverage with a health plan or plan(s) who have set rates for coverage in the coming year. On the diagram, there’s no red dot indicating the presence of data because in many (most?) cases there is not a lot of data to guide this conversation. Many health plans have claims data, to show how many services and what types have been paid for throughout the year. They may not have data about the effectiveness of those services. For example, they may not know what percent of office visits for high blood pressure showed effective control. On the employers’ part, they may not have much data, either. If they are self-insured, they may have similar levels of claims data, but not measures of performance.

Even in health care organizations with advanced electronic medical records, the determination of “% patients with appropriate blood pressure control” may not be done in an automated fashion - a random selection of charts may be used to come up with this percentage. The electronic health record may facilitate the selection and review of charts, but nothing more. This is dependent on the health care environment being studied.

(If there are health plan and providers who would like to inform this part of the story, comments are open)

Clinical and Public Health pearls (Houston-Miller and Eytan)

• High blood pressure is one of the most costly conditions for employers, more than cancer, diabetes, heart disease, and behavioral health conditions. This does not take into account that hypertension is responsible for a significant amount of morbidity among patients with heart disease and diabetes. This post shows the costs of each. The first graph shows the cost per person with the condition. When you average the costs across an entire employed population, the large numbers of patients with hypertension escalates the cost of this condition above all others. For those people interested in the cost profile of chronic conditions to employers, The Center for Studying Health System Change hosted a forum where expert Ron Goetzel, Ph.D. provided an updated look at the data. It is compelling.
• Fewer than 10% of the cases of undetected or uncontrolled hypertension could be associated with lack of health care use. In other words, health plans and employers are already paying for this current state. It does not exist because patients are not getting enough health care.

Comment

Where is the data? and What’s Missing? In this case, there isn’t much data in the conversation. The conversation is around use of services, and in that setting, an assumption is typically made that more services is better. The result is that these stakeholders cannot engage at their potential to ensure that services are as effective as possible.

It is possible that a patient or provider may share data about the effectiveness of their blood pressure control services which are being purchased and paid for by employer and health plan respectively. Blood pressure control is already a HEDIS measure, and is a development Pay for Performance measure in California in 2009.

Next post, the yearly checkback, completing the cycle. Comments welcomed, of course

• Milliman Medical Index - Average family of four at $15,609 medical costs / year. Employee share is up 10 %, the highest cost share is for pharmaceuticals, at 27 %. This ties to the Connectivity for Californians work, because it’s a reminder I need to look up total patient costs for hypertension management.

• Aggressive medical care, conservative vs aggressive - Nice XPlane (xplane.com - one of my favorite design groups) produced images of regional differences in health care costs, from Consumer Reports. It's helpful to see this group champion appropriate care for patients.

As we have been planning a multi-stakeholder pilot to demonstrate improved management of chronic conditions by Californians, this paper was just published, which adds compelling information to the discussion. Talk about interesting timing.

The paper is a compendium of research and information to date on the value of home blood pressure monitoring, which has not been previously integrated into the clinical practice of improving blood pressure control. The impact of poor control is reiterated: high blood pressure as accountable for 27% of total CVD events in women and 37% in men.

Useful Facts

• 35% of hypertensive patients check their BP once a week, 64% of patients own a monitor.
• Arm monitors appear to still be the best choice through validation studies.
• There is often “poor” agreement between patient reported readings and readings stored in the monitors’ memory, with one study showing up to a 20% error rate.
• Patients are at risk for over-diagnosis (”white coat hypertension”) up to 20% of the time, and under-diagnosis (”masked hypertension”) up to 10 % of the time, resulting in over/unnecessary treatment, or undertreatment.
• Blood pressures measured outside of the doctor’s office in many cases is superior in predicting which patients are at risk for organ damage.
• It appears that more study is needed regarding the effectiveness of home blood pressure monitoring to guide treatment and ultimately support better control. As far as I can tell, none of the studies mentioned here assess the impact of coupling improved patient-health system communication (eg secure messaging with physicians or health system) with home blood pressure monitoring.*
• Medicare currently reimburses ambulatory (24 hour) blood pressure monitoring for patients with suspected white coat hypertension, and this monitoring has been shown to be cost-effective in reducing unnecessary treatment.
• Of extreme interest in our planning, this paper refers to studies assessing patient attitudes toward home monitoring, which appear favorable, but maybe less so when out of pocket expense is required - this is a follow-up/countermeasure item.

*A helpful study is under way at Group Health Cooperative that is going to add very useful information to this part of the discussion.

Conclusion

Beyond information about the value of home blood pressure monitoring, there are suggested protocols for integrating this monitoring into practice. This seems like a great springboard to integrate this into patient access to their own clinical information, along with potential connections to the health system and other patients.

Based on the information presented, there seems to be a case for employing “connected” blood pressure monitoring for accurate diagnosis of blood pressure and response to treatment. Given that Medicare already reimburses ambulatory blood pressure monitoring for white coat hypertension, there may also be a case to extend, as a pilot, reimbursement for home monitoring for diagnosis and initial management of blood pressure outside of physician visits. This ties well to the data that most patients with high blood pressure are insured and seeing physicians, with only 35% control, making this approach a worthy alternative.

From a biological plausibility perspective, it makes sense that measuring an ongoing physiological state (average blood pressure throughout the day) in its native environment, over time, has a likelihood of being more accurate than a few point measurements done outside of the environment where people live and work (the doctor’s office).

The opportunity for the proposed project here is to integrate the benefits of home monitoring with a sustainable workflow inside and outside of the health system, using technology available today, to improve patient and family involvement in their care. Of interest, the Agency for Healthcare Quality and Research is promoting the idea of patient involvement in care as a quality and safety improvement strategy for patients. This work could extend the strategy to more stakeholders, including employers and the health system itself.

Conflict of Interest Analysis

I think this should be part of a review of any paper, given the information being published about sponsored research (here’s some examples).

The lead author has a significant relationship with device maker Omron, and has received speakers fees from pharmaceutical manufacturer Boerhinger-Ingelheim and Omron. Another author has received speaker’s fees from Merck and serves in a consultant/advisory board capacity for Pfizer and CV Therapeutics.

There was discussion previously about support to the American Heart Association by device makers.

These associations could result in over-exhuberant promotion of home blood pressure monitoring devices and treatment (i.e. it’s unlikely that a device manufacturer would have an interest in less devices being sold), and need to be taken into account when reviewing this piece. This might be reflected especially in areas where the data is/was equivocal about benefits, yet conclusions are framed in the positive or hopeful.

One of the issues in the discussion of device/medication promotion is that new treatments are compared to placebo instead of to current practice. The information presented here compares the treatment of interest to current practice, which has room for improvement. With that in mind, I think the information here is contributory to the work we’re considering and will be used to update the A3 accordingly.

A Disclosure of My Own

I should point out that I assisted in the planning of the Group Health blood pressure study mentioned above from an operations/informatics perspective, and was not funded under the grant and am not a co-author of that study, which is not connected to this work. I am currently funded by the California Healthcare Foundation.

Comments welcome, of course.

• AHA - Trendwatch - Absenteeism and Presenteeism in the Workplace - The American Hospital Association also weighs in on lost productivity due to chronic illness. Their analysis is based on the Goetzel 2004 paper

• Health and Productivity Among U.S. Workers - Another analysis of the impact on productivity on illness. This is more general, and also refers to the Goetzel study published in 2004

In my fielding of this data to various people, this part of the analysis has been by far the most controversial. Let’s first start with indirect costs added on top of direct costs, from the societal perspective, just for California. By indirect costs, we mean lost time from work (absence and short term disability), presenteeism (impairment while at work, to avoid being absent), and caregiving (21 million working men and women are caregivers in the US)

The indirect cost data is only for employed persons reporting the condition, which means it is a subset of the population (doesn’t include indirect costs for the non-employed, doesn’t include costs for the undiagnosed). The graph above is therefore a little bit off, because the medical cost is for all people, productivity costs just for employment related losses.

Now, a different look at the data, which averages costs across the entire employee population, from the employer perspective.

This is different because it takes into account prevalence, and spreads that cost over all employees, whether they report the condition or not. My explanation for the different way this data looks is that the medical cost per person for hypertensive employees is far less than for heart disease, but there are many more hypertensive patients in a population. In the employee samples used to derive this data, there’s a factor of 2 difference compared to heart disease (12.4 % vs 6.4 %).

The sources for this data are the same as previous charts (formatted for Zotero, below).

I have already been asked, “Ted, how can high blood pressure cause presenteeism at all?” and I welcome the skepticism. I reviewed the study below, which defined the term for our profession, and it includes a combination of employee studies, some done quite well, that ask about employee impairment and absences due to multiple conditions. This includes things like side effects of medications (which are a cornerstone of hypertension therapy). Questions, based on the study, were things like whether an employee performance was reduced by ‘losing concentration, repeating a job, working more slowly than usual, feeling fatigued, or generally “doing nothing”‘. The authors specifically chose tools that measured multiple conditions at once, so that comparisons could be made.

One novel study worth mentioning specifically is one by Bank One, that used administrative and computerized productivity records of its employees to explicitly measure productivity losses, in addition to using a health risk appraisal and claims data to come up with estimates. For hypertension, the estimate was 0.4 % in this one, which was right in the middle.

Based on my reading of the paper, I am accepting the methodology as supportive. As a student of LEAN, though, I know that the facts are best obtained on the factory floor, so my next interest is in working with an employer, and ultimately and employee, who experiences these conditions first hand. And I do mean on the factory floor, rather than the health system.

After creating this post, I realized that my A3 (coming next) has one inaccuracy. Fixing that, posting it soon.

1. An Unhealthy America: The Economic Burden of Chronic Disease: California. Take a look at the methodology here.

2. Goetzel, Ron Z, Stacey R Long, Ronald J Ozminkowski, Kevin Hawkins, Shaohung Wang, and Wendy Lynch. “Health, absence, disability, and presenteeism cost estimates of certain physical and mental health conditions affecting U.S. employers.” Journal of occupational and environmental medicine / American College of Occupational and Environmental Medicine 46, no. 4 (April 2004): 398-412.

Comment away, this can only improve with input.