This Community is Archived.

This community is no longer active as of December 2018. Thanks to those who posted here and made this information available to others visiting the site.

Should Patient Satisfaction Be a Measure of Quality?

By Thomas Tsai Moderator Emeritus | 15 Jul, 2014

Patient experience has recently come under fire as a quality metric in health care. Critics argue that its inclusion as a key measure of hospital quality is driving hospitals to have the wrong priorities, encouraging them to behave like hotels instead of sites of care delivery. Within the field of medicine, many physicians have argued strongly against the measurement of patient experience, citing it as a distraction from the original goals of clinical care.

What are your thoughts?

Attached resources:



Julie Rosenberg Replied at 2:36 PM, 18 Jul 2014

Thanks for pulling these resources together, Thomas. It's an interesting collection of "evidence." The idea of "patient experience" is so broad, and I imagine it's hard to dissociate the experience of being sick and the path of a particular illness with the care one receives. I also wonder whether patient experience matters more in some fields than others. For example, primary care depends in large part on communication and trust from the patient, neither of which are likely without a positive experience, but radiology is less dependent on these aspects of patient experience. Interested to hear what others think. Do you aim to give your patients a positive experience and is there a way to do this without altering your course of care?

Elizabeth Glaser Replied at 4:38 PM, 20 Jul 2014

These are interesting arguments pro and con for incorporating patient satisfaction measures into considerations of quality and safety
( and probably cost effectiveness, too). Given that there is no clear answer here, perhaps patient satisfaction could be be used as a covariate rather than an outcome measure in the quality and safety equation.There are times when low or high patent satisfaction might cue us into systemic problems which are not being addressed or even considered. I can think of two scenarios when satisfaction information might allow us to better understand certain outcomes.

Using an example from international health: many pregnant women in subSaharan Africa will use outpatient prenatal service but opt to deliver at home or with a traditional birth attendant. This practice risks the life of both women in labor and their newborns. There are many reasons why women do not deliver in hospitals and previous intervention efforts have been aimed at increasing access, reducing fees, and facilitating travel to labor and delivery units. These initiatives have helped increase utilization and decrease maternal and neonatal mortality yet women still may choose to deliver outside a medical facility - why?

The emerging evidence points to women opting out of the system to avoid verbal and physical abuse during labor by medical and nursing staff. In areas where there is relatively low utilization of inpatient maternity services, the use of community, rather than facility based, surveys to determine the kind of services used , the reasons using those services, and satisfaction with services used, can help clarify the issues around low utilization that would not come out if we just survey women who deliver inpatient or woman who deliver at home. In this case access is important , but is it clear that dissatisfaction with services may drive people to opt for unsafe care over disrespectful care.

Another instance where satisfaction might clue us into problems is when a facility or provider has high patient satisfactions scores yet very poor outcomes. What is going on? My bias is to believe that this may represent the dark side of providing care to vulnerable groups. People from marginalized groups or communities might be so grateful to get any services that there is little accountability from the community to change or improve - if one's point of comparison is no care vs some kind of care, how does one know if that care is excellent, good, or poor? And if providers are "nice" then consumers tend to equate that with good care, which is not always the case.

So should we use patient satisfaction measures - yes, but very carefully.


IDA CHAPUMA Replied at 8:32 AM, 21 Jul 2014

Patient satisfaction could be a measure of quality and safety because good
and quality services results in patient satisfaction. Patient satisfaction
may be obtained from suggestions box and/or patient exit interviews which
could be done through a questionnaire.

Derek Ritz Replied at 9:48 AM, 21 Jul 2014

I'm not convinced that good quality services automatically result in high patient satisfaction. Elizabeth's cautions are important -- we definitely should use patient satisfaction measures as one of the dimensions of "quality"... but outcomes are the trump card, in my view. Even in Canada (where I live), there is very low transparency regarding outcomes measures. In the face of better access to data, I believe patients would make stronger use of outcomes as part of their perception of "satisfaction". Without it, though, a pleasant but ineffective care episode may be greeted with satisfaction... but not lead to improved health.

Elizabeth Glaser Replied at 1:05 PM, 21 Jul 2014

That is exactly what I am saying.
Since clinicians and researchers also use the health system, as an experiment, one might consider a recent experience as a consumer:

Use one existing validated instrument ( HCAHPS, etc) to rate the experience.
Does it capture your (dis)satisfaction?
What was the health outcome during the experience?
Does (dis)satisfaction directly or indirectly correlate with the health outcome?

I tend to believe that patient (dis) satisfaction can tell us about outliers or as part of counterfactual models and, as noted before, as one of many factors that may go into producing an outcome.


Julie Rosenberg Replied at 1:47 PM, 21 Jul 2014

A study released yesterday by the Associated Press-NORC Center for Public
Affairs Research based on interviews with over 1000 adults in the US showed
that "Most Americans tend to focus on certain aspects of quality relating
to doctor-patient interactions and doctors’ personality traits, rather than
the effectiveness of the care provided or the patient’s own health
outcomes." The full report is here:

Julie Rosenberg Talbot, MPH
Publications and Curriculum Development Manager
The Global Health Delivery Project at Harvard University
(617) 858-1443 | skype: julied.rosenberg | |

Elizabeth Glaser Replied at 2:18 PM, 21 Jul 2014

Dear Julie,
I am interested to read the report -- thanks for this.


SUNIL JI GARG Replied at 5:41 PM, 21 Jul 2014

In this world of objectification of each and everything, the question of
measuring care
by providing a index score to patient satisfaction is worth a longer debate.

Civilization, Community, Culture and Care, these are all subjective things.
who try to put indexes, labels and markings to these things are actually
reducing the importance
of subjective things. Something built into our inherent systems are
transferred from
people to people with the help of value transfer systems. Society needs to
a value transfer system, rather than objectifying these values.

There is nothing like a perfect 10 in care. There is nothing called a zero
Care of a patient is like putting entire soul into that effort. If I am
well trained
my efforts will definitely be the best in the give circumstances. So let us
rank care with the help of patient satisfaction index. Let the quality
indexes be left
to measurable things. Regarding the need for patient satisfaction, it would
better to talk about, what to introduce in the syllabus of medical service
to let them understand this thing better, rather than starting ranking
things by indexes.

Things can take an ugly turn of high commercialization and hospitals will
do anything
to get a better satisfaction score, rather than actually caring.

Janet A DEWAN PhD CRNA Replied at 6:10 PM, 21 Jul 2014

Compliance with care recommendations is one good marker of health system quality , as well as patient satisfaction. When patients do not or are unable to follow care recommendations the quality is suspect. So, one interesting measure is compliance and the qualitative assessment : why or why not.?

IDA CHAPUMA Replied at 9:08 AM, 22 Jul 2014

Patient compliance means the patient trust their doctor and recommendations made to them by their doctors.

Thomas Tsai Moderator Emeritus Replied at 1:37 PM, 22 Jul 2014

Derek and Elizabeth,

You raise important points. As the health care system in the US moves from a fee-for-service payment (and delivery) model towards episode-based bundled payments, should we also be thinking of patient experience over the course of an episode of care? For example, a patient undergoes an uncomplicated colectomy and is discharged 4 days postop. His/her rating of the operation and hospital experience may be high. What if that patient then develops a complication, say an abscess or wound infection a week later? Would a patient experience metric that spans the entire episode--inpatient,readmissions, post-acute care, etc--tell us more about the quality of a health care system?

Thomas Tsai Moderator Emeritus Replied at 1:38 PM, 22 Jul 2014

Janet and Ida,

Agree that adherence may be an interesting aspect of patient experience/satisfaction. However, there is literature that adherence to medications and treatment plans may be influenced by socioeconomic status and medical literacy. How cab we account for that?

Elizabeth Glaser Replied at 1:58 PM, 22 Jul 2014


Yes, it would seem vital to have to metric span the episode of care.Assessing satisfaction over the entire episode might highlight gaps, especially when case is transitioned from the inpatient to outpatient setting.

As I recall, the CAHPS surveys gather data on either inpatient or ambulatory care experiences but not both over a single patient experience. Perhaps that is changing, though?


Mighty Casey Replied at 11:10 AM, 24 Jul 2014

Helping people-commonly-called-patients understand how to assess their "experience" based on health literacy (including outcomes metrics and EBM), not hotel appurtenances, would be a massive step in the right direction on using end-user (patient) input in QI. In the US, however, that's still problematic due to what I call the hospital arms race. Marble lobbies and fountains are lovely, but how do they improve outcomes?

There are certainly community and population-based strategies that can shift culturally based thinking on this, patient-side, such as the maternity conundra that Elizabeth Glaser points out above. The US patient satisfaction effort relies too much on surveys that don't reach the patient for up to 60 days post-discharge. Why the heck can't that survey be delivered digitally as part of the discharge process? Oh, that's right, that's not how the market-cornerers in this slice (Press-Ganey and HCACHPS) roll.

Mighty Casey Replied at 11:16 AM, 24 Jul 2014

Dangerously low caffeine levels led me to leave out a chunk of the 1st graf above (oy). Should read:
Helping people-commonly-called-patients understand how to assess their "experience" based on health literacy (including outcomes metrics and EBM), not hotel appurtenances, would be a massive step in the right direction on using end-user (patient) input in QI. The perfect teacher(s) in that effort would be hospitals, given their community placement and internal expertise. In the US, however, that's still problematic due to what I call the hospital arms race. Marble lobbies and fountains are lovely, but how do they improve outcomes? And then there's the standard "we don't get paid to do that" [raise literacy] push-back.

William Martinez Moderator Replied at 5:44 PM, 24 Jul 2014

The merits of using patient satisfaction as a quality metric can clearly be debated. One problem is accurately measuring patient satisfaction in a meaningful way. Some of the pitfalls have already been alluded to here. One way of assessing "satisfaction" is by assessing its opposite, "patient complaints." My colleagues at Vanderbilt have been studying unsolicited patient complaints for years. Patient complaints are non-randomly distributed among physicians or clinic units. They tend to cluster around certain physicians or clinical areas relative to others and are associated with increased malpractice risk and poor outcomes. They are likely a marker of dysfunction within healthcare teams though this needs closer examination. Analyzing complaints may be more powerful than surveying satisfaction.

Attached resource:

Asfawesen Gebreyohannes Woldegiorgis Replied at 7:33 AM, 29 Jul 2014

Patient satisfaction as a measure of quality of care has its own limitations, nontheless it is one way which tells us the behavior of the providers, quality of infrastructrue, convenience of process and etc which may not get about them from treatment outcome indicators. Though there are no perfect tools to measure it, It should continue to be one dimension of for measuring quality of care. I think patients and community should be part of the development of tools for assessing patient satisfactions.

Elizabeth Glaser Replied at 10:52 AM, 29 Jul 2014

You points about patient complaints are well taken. I would like to understand about it. .

Do people who are typically seen as vulnerable such as elders, immigrants, poor, or with particular diagnoses such as mental illness , complain about care at the same rate as those in other groups?

Does this method identify systemic problems or is it a better metric to identify outliers?


Ruth Staus Replied at 5:02 PM, 29 Jul 2014

I am a primary care provider who runs two free clinics that serve poor, homeless, elderly, immigrants, and chronically and persistently mentally individuals. It is very rare that we hear any complaints about the care. Our clinics are not the only option for care for these populations in our city.

When I practiced with a large non- profit health care system , that served mostly middle class, upper- middle class, and wealthy patients, complaints were common. A frequent situation was the wealthy woman with a viral upper respiratory infection demanding antibiotics. The provider would explain to the patient that her symptoms were viral in nature and that it would be poor medical practice to prescribe antibiotics that would not help the problem and potentially set her up for serious problems with drug resistance in the future. The patient would fill out a patient satisfaction survey and the provider would receive a poor rating. The provider gave appropriate care based on current standards of care and CDC recommendations.

While I think that patient's experiences with healthcare are important and we need to capture that information, I also think that the average patient is lacking the knowledge base required to make an adequate judgement about the quality and appropriateness of the care provided. Sometimes the care is appropriate and of high quality but the patient still has a bad outcome either due to their own actions( or inaction), or just bad luck.

Mighty Casey Replied at 6:23 PM, 29 Jul 2014

Ruth Staus's comment above is why I will continue to bang my Health Literacy for Ever'body drum relentlessly, everywhere, until that becomes unnecessary. Would be outstanding if the basics of managing health were part of K-12 in the US and elsewhere, but ...

Tom Catron Replied at 7:18 PM, 29 Jul 2014

I think this is an opportunity where science can inform public policy. If the goal of reportable hospital/medical center measures is promote reliable, safe medicine then we should have measures/metrics that have a demonstrated link to quality and outcomes. The evidence linking patient satisfaction to quality indicators is poor (see, e.g., the work of Makary and Manary). Our center has been investigating the utility of patient dissatisfaction (not satisfaction) as measured by unsolicited patient complaints. Patient dissatisfaction undermines patient adherence with medical treatment plans, outcomes of care, and patients’ willingness to stay with a practice. Furthermore, physicians associated with high numbers of complaints are at increased risk for medical malpractice claims. We (and others) have demonstrated a link between patient complaints and medical malpractice experience. We recently completed a study that has that establishes a link between patient complaints and surgical outcomes. In additional, we have used this data to develop a process that has been highly successful in the early identification and intervention of high risk physicians. Thus, it also helps if the metrics are actionable with measurable quality outcomes.

Metrics linked to quality and other observable outcomes have additional utility beyond hospital rankings. Self-regulation is the hallmark of a profession. Policies, procedures, project bundles, evidence based treatments, and checklists are meaningless unless providers are held accountable. Using objective measures to bring awareness to high risk colleagues promotes a culture of safety, reduces medical malpractice claims and reduces patient dissatisfaction.

I look forward to the research in this area of medicine growing and informing public policy. There are contributions to our knowledge popping up in the literature more frequently, making these exciting times for dialogue and action.

Thomas Tsai Moderator Emeritus Replied at 9:12 PM, 30 Jul 2014

I agree wholeheartedly that publicly reported metrics should have a demonstrable link to quality. We have done some work illustrating a an association between hospitals performing well on patient experience measures as well as other measures of quality, include objective outcomes such as mortality (link below). Many of the studies that have shown no relationship between patient experience and quality have unfortunately been limited by selection bias and small sample sizes. The other take home message is that not all measures of patient experience are created equal. A recent review article (link below) has shown that patient experience measures collected using robust survey and sampling methodologies have been meaningful for public reporting and pay for performance programs.

A crucial distinction is the difference between metrics for policy and metrics for actionable outcomes (e.g. quality improvement). I don't think there's a one-size fits all for quality metrics, including patient satisfaction. HCAHPS is useful for public reporting and may be useful for pay-for-performance around patient-centered care but tells you little about how to improve care locally at the individual health care provider level. Other measures such as patient complaints may not be useful for public reporting, but can be immensely helpful for local quality improvement. How we do get this distinction across? Does this distinction actually exist? Looking forward to everyone's thoughts!

Attached resources:

Mighty Casey Replied at 10:25 AM, 31 Jul 2014

{Sage link above yields an error; Annals of Surgery requires a subscription]
As anyone who's ever constructed a survey knows, how you ask and what you ask are critical. Asking about parking doesn't deliver actionable intel on outcome metrics.

Also, not making it easy for clinical staff to report issues around quality makes it easier for HODAD (hands of death and destruction) surgeons and drug diversion to entrench themselves and cause endless trouble.

It's a system-wide issue. Let patients help tune the antenna array that's gathering the signal. I'd love to see survey tools in use as part of the discharge process for real-time feedback.

Tom Catron Replied at 10:51 AM, 31 Jul 2014

Good point, Mighty Casey, about co-worker reporting of events. Our work here at Vanderbilt started with looking at unsolicited patient complaints to identify and address high risk docs. Part of this work, now at over 80 hospitals nationally, includes expanding the number of patient complaints captured and reported in the patient relations database. Some of our sites collect as many as 12,000-18,000 complaints a year. As you can imagine, complaints are not randomly distributed and a small percentage of physicians account for the vast majority of physician complaints (50-60% of physicians at most medical centers get none in a four year period).

Our recent work, along with Stanford and Northshore (Chicago), with co-worker complaints also demonstrates the remarkable consistency of the non-random distribution of complaints (a small percentage of docs account for the vast majority of co-worker complaints; about 80% of physicians have no co-worker complaints). We have a process similar to addressing high patient complaint physicians using a tiered intervention approach and peer messengers to address co-worker complaints. The other interesting finding is that physicians that have the highest number of patient complaints often have no co-worker complaints; and physicians with the highest number of co-worker complaints have no patient complaints. There is of course some overlap with some physicians, yet this finding suggests that patient complaints and co-worker complaints measure different constructs (much as Patient Satisfaction measures a different construct) and collectively these measures reflect what some may refer to as professionalism. Multiple measures are better than single measures in this instance.

Finally, patients and employees are the eyes and ears of our medical centers and provide valuable observations to promote quality and patient safety. We find that both patients and employees are willing to share their observations if there is psychological safety (ie, they will not be punished or retaliated against) and there is value in sharing (ie, the medical center will act upon the information). Our medical center has enacted policies to protect "reporters" and enforces these policies, thus increasing the number of patients and employees willing share their observations. And we act upon the data.

Mighty Casey Replied at 11:49 AM, 31 Jul 2014

Tom, I'm delighted to hear about the Vanderbilt program - I've interviewed a number of hospitalists at Vandy over the last few years as part of my gig producing a podcast series for The Hospitalist magazine - and am doubly delighted that you guys have put patients in the mix as frontline "reporters." Your experience in psych research is showing, in a very good way =)

This Community is Archived.

This community is no longer active as of December 2018. Thanks to those who posted here and made this information available to others visiting the site.