Ethical Issues in the Design and Implementation of Population Health Programs – free full-text /PMC5834965/ – J Gen Intern Med. – 2018 Mar
When a single doctor is supposed to work both for a “population” and an “individual”, ethical problems are inevitable. I doubt any patient wants to be treated as just a standard member of a “population”.
This is the current problem that pain patients are having: because we need a drug that is also used illicitly (by others), we are treated as though we were members of the “population” of “illicit drug users”.
Instead of receiving treatments that are effective for our pain and tailored to our individual needs, we are literally treated like “people with addiction” instead of “people with pain”.
…specific population health activities may not be in every patient’s best interest in every circumstance, which can create ethical tensions for individual physicians and other health care professionals.
Because individual medical professionals remain committed primarily to the best interests of individual patients, physicians have a unique role to play in ensuring population health supports this ethical obligation.
Using widely recognized principles of medical ethics—nonmaleficence/beneficence, respect for persons, and justice—this article describes the ethical issues that may arise in contemporary population health programs and how to manage them.
The Triple Aim’s goals of
- improved population health,
- reduced health care costs, and an
- improved patient experience of care
motivate health system reform.
I don’t believe this for one minute.
When a corporation with a “moral requirement” to generate profits is running healthcare, its focus, by design, must be financial.
Everything else is aligned to support that goal: patients are considered “consumers”, doctors are “providers”, clinics and labs are “profit centers”, and everything else is “overhead”.
As in all business, the objective is to lower the cost of the “provider/product”, trim “overhead” to the barest of minimums, and encourage the “consumer” to choose your products while charging as much as the market can bear.
Of course, this last principle can only work in a truly “free market”, and healthcare is anything but that.
Instead of simple pricing, the healthcare industry (soon to be the largest employer in the whole country) consists of a bizarre patchwork of charges and payments to and from multiple layers of intermediaries, like insurance, benefits management, and administrative companies, each taking their own chunk of profit.
How can any of this possibly work to benefit the patient (or population for that matter)?
many of its associated concepts, including the need to reduce health care costs and achieve higher value care, apply to medical practice globally. Many physicians now practice in settings where population health programs affect how care is delivered.
The Institute of Medicine has defined population health as
“the health outcomes of a group of individuals,
including the distribution of such outcomes within the group.”
In practice, population health is best understood by analyzing the design and implementation of the discrete population health programs (PHPs) undertaken by health care organizations and others for its sake.
Physicians practicing within PHPs may face ethical tensions regarding their commitment to individual patients and the broader population
Population health can push physicians
- toward traditional public health ethics (where decisions are made with primary concern for large groups of people) and
- away from clinical ethics (where the decision-making locus is the patient-physician relationship, concerned primarily with individual patient welfare).
Table 1: (link to table in separate window)
How Population Health May Place Physicians in a Zone of Ambiguity Regarding Their Roles
Public health ethics Population health ethics Clinical medical ethics Unit of concern Large groups, frequently defined in part by formal city/state/regional/national boundaries Groups of people, usually defined by where care is received, local geography, and/or payer arrangements An individual person, defined by the patient-physician relationship Decision-making locus Public health agency Health care organization, system, or payer Patient-physician relationship Primary animating ethical principle(s) Group welfare, safety, or protection from harm Protection of individual patient welfare (i.e., non-maleficence and beneficence) Secondary constraining ethical principles(s) Liberty rights
Respect for individual autonomy/choice
Paradigmatic ethical issues • Tension between mandatory vaccination or quarantine measures and individual autonomy/choice
• How to determine fair resource allocation during an epidemic
• Use of behavioral economics to “nudge” food choices and the impact on individual choice
• Tension between improving individual or group performance measures and patient choice regarding their health priorities
• Impact of population health activities (e.g., automated reminders, contact from non-care team members) on patient trust in physician
• How to appropriately define and respond to “at risk” patients via predictive analytics (including their effect on disparities)
• Respecting patients’ autonomous choices, even when not in their “medical” best interests (informed consent)
• Physicians’ obligation to recognize individual- and system-level biases the lead to disparities
• Physician engagement in shared-decision making and encouragement of healthy behaviors
A prior Society of General Internal Medicine Ethics Committee analysis examined ethical tensions in performance measurement in the context of pay-for-performance programs.
Incentives, financial or non-financial, add an important ethical dimension to performance measurement.
Whether or not costs or incentives are directly involved, however, ethical issues may arise in PHPs that deserve examination.
ETHICAL ISSUES IN PHPs (population health programs)
Table 2: (link to table in separate window)
Examples of Ethical Issues in Population Health Programs (PHPs) with Example Management Strategies
Ethical questions Ethical values Non-maleficence and beneficence Respect for persons Justice Distributive Procedural Patient level “Is this PHP in my individual best interest?” “Does this PHP protect and enable my choice?” “Are the benefits and burdens of the PHP shared equally across all patients?” “Were patients like me involved in the PHP development?” Example: Improving the rate of colorectal cancer screening in those age 50–75 by colonoscopy, sigmoidoscopy, or FOBT Efforts to meet this metric result in unnecessary screening of patients with limited life expectancy as an unitended consequence (e.g., a 74 year old with a terminal illness) Rather than allowing for discussion of the risks and benefits of colonoscopy, sigmoidoscopy, or FOBT, FOBT is advocated by the PHP as a way to achieve performance quickly Electronic outreach messages via the EMR preferentially reach certain groups, reducing the potential beneft for others (those with limited English proficiency or without electronic access) Patients were not involved in the review of the electronic message or its content Example management strategy Actively monitor, using EMR data, for evidence of over- and under-screening and take action to prevent it Present the risks and benefits of possible recommended actions in a balanced way to enable informed choice (not, e.g., encouraging FOBT just to meet the metric quickly) Ensure messaging is equitably accessible for all groups at the time of initial design (lest these groups be left out, once a goal is met) Design PHPs with input from patient and family advisory councils (PFACs) Physician-level “Will this PHP enable me to fulfill my obligations of beneficience to my patients?” “Do I still feel free to choose and recommend what I believe best for patients?” “Are the benefits and burdens of PHPs equitably distributed among all clinicians?” “Were front-line clinicians involved in PHP development?” Example: Improving A1C control in patients with diabetes using a new default Diabetes Care Order Set The focus on diabetes in EMR clinical decision support distracts a physician from other medical or social issues important to individual patients A pre-filled referral order routes patients to a list of preferred endocrinologists, when the physician and patient believe a different endocrinologist would be best Primary care phyicians, compared to specialists, bear more of the burden for meeting the metric (e.g., time in EMR documentation) A diabetes order set is designed without input from front-line physicians, reducing buy-in (and adherence) to it Example management strategy Evalute for unintended consequences of PHPs, e.g., whether improvement in one area results in lagging performance in others or reduced patient satisfaction in the encounter Design referral processes to preserve patient-physician shared decision-making (e.g., informing both about the rationale behind the preferred list and allow exceptions) Recognize this extra burden by providing adequate time or personnel resources (or, if applicable, shared savings) that result from these efforts Design the PHP with input from the physicians who will use it to improve buy-in, trust, and PHP success Organization-level “Will this PHP actually improve the care we deliver to patients, not just improve measurement or documenation?” “Is the PHP implemented in a way that is respectful? For example is is culturally sensitive and respectful of persons?” “Are we giving special concern to our most vulnerable patients?” “What structures ensure ongoing engagement of physicians and patients in the design and implementation of PHPs?” Example: Reducing hospital readmissions A health care organization reduces readmissions to its hospital by shifting care to the emergency room or observation status A post-discharge home care program is standardized but fails to accommodate patients’ social and cultural differences (e.g., related to family involvement in home care)
A medical chart flag for patients at high risk of readmission using predictive analytics is inadvertently stigmatizing
In a setting where certain patients most in need decline post-discharge home care, an organization excludes these patients from its denominator when it calculates its success rate A post-discharge care management program is designed with patient input, but not from patients who will use the program Example management strategy A readmission reduction program pays careful attention to whether it unintentionally shifted costs when it tracks program outcomes Post-discharge home care programs should be explicitly designed to be respectful of social and cultural differences
Careful language should be used to describe “high risk,” preferably crafted with patient input and with training of health care professionals regarding its meaning
Special attention is given to better understanding why patients decline and developing appropriate ways to reach them Ensure that the patient engagement program that informs program design includes patients representative of the end-user group
In practice, the lines between harmful, non-beneficial, and beneficial effects of PHPs for individual patients may blur.
PHPs could improve a population’s overall health while resulting in unintended non-beneficial or even harmful individual consequences
For instance, implementing a standard colorectal cancer screening metric with an age cutoff of 75 years—as might be done with an electronic health record pop-up reminder—appears to have been associated with underscreening of healthy individuals over age 75 and overscreening of unhealthy individuals under age 75.
Unintentionally, a metric appeared to discourage appropriately individualized clinical decision-making.
Thus, standard recommendations applied rigidly to groups for average benefit may introduce ethical tensions for physicians caring for unique patients.
This is exactly where chronic pain patients find themselves when standard doses are set for all patients.
Physicians should support system defaults that apply standard recommendations, but that application must leave space for individualized risk-benefit determination.