Customer satisfaction is a useful metric for business. The relationship between a positive customer experience and profit is evident. In the healthcare industry, many argue that patient satisfaction is important for profits and also can measure for quality care. Patients, who are satisfied stick with their provider and talk amongst their family and friends. However, the relationship between satisfaction and healthcare quality is not well understood. Nor is it known whether satisfied patients become healthier people.
Health care is big business in the United States, with the highest per capita spending and greatest percentage of the gross domestic product at 17.9% in 2011. The federal government projects that spending is will reach 20% of the GDP by 2021. On the policy level there are bitter debates about health reform, Medicare, and Medicaid that are amplified during the current presidential election season. Meanwhile, hospitals and independent providers are trying to attract and retain patients into their care.
Many refer to something called the patient satisfaction industry that is driven by pressures to boost the number of patients. Much data derived from Medicare insured patients is readily available for citizen viewing. Private companies have launched websites, where patients can rate their doctor and also view satisfaction scores. In a few months, the government will provide financial incentives to providers, who score well on quality measures that include patient satisfaction. Unfavorable surveys may also lead to bad publicity, but also consequences by potential lawsuits. The pressure to achieve perfect satisfaction scores may lead providers to game the system.
What is patient satisfaction?
In the dialogue about patient satisfaction, it is often described as a measure rather than a distinct phenomenon. For example, the Robert Wood Johnson Foundation defines patient satisfaction:
Patient satisfaction is a measurement designed to obtain reports or ratings from patients about services received from an organization, hospital, physician or health care provider.
Research has developed the meaning of patient satisfaction to be more precisely defined as the patient experience. On a practical level, patient satisfaction is equivalent to the actual measure. Therefore, the quality of the measure, the questions asked, sampling, and response rates shape the results. Measures vary on how well they ask questions related to health services, rather than on the comfort of the waiting room and the flavor of the food. In recent years, there is increasing standardization on patient satisfaction measures.
The most commonly used one developed by the federal government, the Consumer Assessment of Healthcare Providers and Systems (CAHPS). The survey in use for nearly a decade has versions for hospitals and independent practices. These measures contain closed-ended questions specific inpatient and outpatient encounters. The areas of focus include speed of accessing care, provider-patient communication, referral to other care, and follow-up. Additional demographic questions are asked as well as the patient’s perceived health status.
Bias introduced into patient satisfaction measures
These surveys are based on sound research principles such as with randomized sampling, high response rates, and other systems to reduce bias. However, in the real world, bias has been introduced into these objective measures.
Bias can be introduced in a number of ways. One fast and effective method to boost satisfaction scores is to exclude patients with complaints. Press-Ganey is a healthcare consulting company that works with about 40% of U.S. hospitals in gathering patient satisfaction data. The company’s name itself is equated with patient satisfaction ratings, known as the PG score. William Sullivan, DO, JD and Joe DeLucia, DO made an analysis of Press-Ganey’s procedures to collect and report on patient satisfaction data. They found several junctures of bias for Emergency Department patients such as excluding patients transferred elsewhere, admitted to inpatient floors, and patients who leave without being seen. Also, patients who frequently use emergency services are excluded because measurement is supposed to be more than 90 days. Overall, the doctors concluded was that Press-Ganey selects a sample that has a lower acuity.
Non-response bias can also shape the results. These surveys are typically done by mail or by mail and telephone follow-up. While many hospitals are able to achieve the minimum response percentages, not all subgroups are represented. Response rates from Spanish speaking patients are low and studies have found lower satisfaction as well.
This concept is purely from the patient’s point of view and, because they are anonymous, they cannot be directly linked to particular doctor, test, or inpatient stay.
The risks of pleasing the patient
A study published earlier this year in the Archives of Internal Medicine found that patient satisfaction was associated with inpatient hospital stays, greater care costs, increased prescription costs, and increased mortality risk . These findings remained after adjusting for other factors that may increase mortality such as age and health status. The authors of the study note that patient expectations affect doctor behavior, which may explain the increased prescriptions offered. Increased mortality may be related to more elective medical services. Most medical procedures and all medications carry risks. In this study of more than 51,000 patients, the effects of even small increases in risk is likely to be detected.
One ideal of these surveys is to improve patient satisfaction through meaningful change to how a healthcare responds to its sick and worried patients. However, what a patient wants may not be necessarily be healthy or safe for them. Research has found that patients tend to prefer medications and testing, despite their risks. Doctors on the front lines hear directly what patients want. They may want an antibiotic for a virus or a CT scan for a muscle strain. The fee-for-service health care system reimburses per procedures further drives this resource and risk intensive patient care. Kevin Pho, MD argues on his blog that doctors should risk patient ire by saying no to denying habit forming medications, and resisting the fee-for-service model that rewards unnecessary procedures. Theresa Brown, RN in an Op Ed for the New York Times pointedly notes:
Hospitals are not hotels, and although hospital patients may in some ways be informed consumers, they’re predominantly sick, needy people, depending on us, the nurses and doctors, to get them through a very tough physical time 
The implications of physician’s experiencing pressure in pleasing their patients has not been measured in these satisfaction surveys. In fact, there is no data collected about the more objective data on prescriptions, diagnostic tests, or health outcomes. Quality and safety indicators are separate from studies of patient satisfaction.
Satisfaction may obscure safety information
Hospitals, group practices, and individual physicians collect objective data about patient outcomes. There is publicly accessible information on quality indicators, including how hospitals fare in terms of complications and mortality. The federal agency Health and Human Services manages a site Hospital Compare where any person can examine satisfaction; time to care; complications, readmissions, and deaths; use of diagnostic tests; and relative spending. When examining some of the top-rated hospitals in the country, the data reports variable satisfaction and mortality data. Also, much of the mortality data is suppressed by the federal government, leaving the average patient with incomplete and contradictory information.
Conclusion: Toward greater patient health literacy
This era provides an unprecedented amount of data about how patients experience their hospital stays and outpatient visits and how well they live and survive. Using the data to penalize or reward providers without an explicit description of evidenced-based practices only serves to reward quick fix solutions of shaping the sample to artificially elevate scores.