“Patient Satisfaction” – it’s bad for everyone

In Editorial, Featured, Opinion by Shahbaz Syed5 Comments

Higher patient satisfaction scores are linked to increased health care costs, and higher mortality1.

Pause. Take a second, and re-read that last sentence.

The follow up question is obvious – why would administrators push patient satisfaction if it results in worse outcomes? This post seeks to delve into the harms of targeting patient satisfaction, why it appears to be done, and the importance of speaking to our administrators in order to advocate on behalf of patients and the healthcare system.

What is patient satisfaction?

There is no clear consensus as to the definition of patient satisfaction, but most literature suggests that it is the degree of congruency between the patient’s perceived care and their expectations2. As we continue to advance the field of quality improvement, patient satisfaction is becoming a proxy for the quality of care provided to individuals. There is no general consensus as to how to collect this information, but the vast majority of administrations doing this provide questionnaires to their patients, to assess how satisfied they were with their healthcare visit3.

Show me the evidence!

I can’t. At least, none that exists. Alarmingly, despite it’s inclusion as a growing measure of quality of care, there is no evidence to suggest that patient satisfaction is an effective or useful measurement. A systematic review in the BMJ4 correlated patient experience with clinical safety and effectiveness, and found that the patient experience is positively associated with clinical effectiveness and safety. I would suggest it is fair to assume that the vast majority of us would know/agree this to be true. However, I would argue that patient experience is by no means a correlate to patient satisfaction. Patient experience tracks an individual’s journey through the healthcare system – it focuses on the primary interactions that the patient has during their hospital or clinic visit. Satisfaction, however, relies on the patient’s expectations and perception of appropriate care – something that they may have little to no insight about.

What could go wrong?

To highlight this point, I think it is important to utilize a case example. Imagine that you have a patient who presents with “bronchitis, and requires antibiotics”. First off, there is no convincing evidence that bronchitis is even a real disease (seriously, smokers with tight lungs do certainly get viral URTI’s, this does not equate to a novel disease entity). Secondly, there is excellent evidence that these patient’s do not require antibiotics. However, the patient expectation is that they should receive antibiotics (because their previous GP or walk-in clinic physician did this), despite expertise demonstrating why antibiotics are inappropriate, the patient will be unsatisfied with anything except a prescription. Thereby, the physician has lower patient satisfaction scores, and perceived lower quality of care. In this scenario, the physician is punished by doing the right thing and practicing appropriate/high quality medicine.

To provide some evidence to help further conceptualize this, if we go back to the initial statement in this post, an prospective study found that a focus on patient satisfaction resulted in increased hospital admissions, increased drug and total healthcare expenditures and increased mortality1. In essence, focusing on patient satisfaction encourages physicians to provide the perception of better care, while potentially harming the patient and delivering worse overall care. Patients may often think that “more is better“, and that they would benefit from more investigations or therapies, without appreciating the potential iatrogenic or negative effects from this. By doing more, physicians can provide the illusion of better care, even though a plethora of evidence exists showing that the patient will most likely have worse outcomes.

Patient Expectations

Part of what makes patient satisfaction so difficult, is that it is based upon patient expectations of the type of care they should receive. Patient’s nowadays have so much information within reach (we google everything; from what taco’s to eat to how to remove a splinter), and so patient’s are certainly googling their symptoms and presentations. The problem, however, is the amount of misinformation on the internet5. Resources are consistently inaccurate, peddle to fear mongering or dabble in anecdotal evidence, which provides an unfortunate bias from the patient perspective.

We’ve all seen an example of this – a patient with back pain or a headache requesting an MRI. Physicians scoff at this comment, but if an MRI was an easily obtainable investigation (i.e.: we didn’t need to sell our soul to radiology), lots of patient’s would get their wish. How often do we undertake imaging, or bloodwork simply because a patient has requested it, even when we know the studies will be negative?

When entering the realm of patient satisfaction, hospitals adopt a “customer is always right mentality”, which can be quite a dangerous precedent to be setting in a healthcare environment. Unfortunately, patient’s are not always right, and yet hospitals are being forced to devote entire “complaint departments” to when patient expectations are not met. This places healthcare providers in a difficult situation – attempting to navigate between the most medically appropriate course of action versus the patient’s expectations or satisfaction.

If it’s no good, why is it being done?

There is certainly no evidence to help guide us here, but personally I think it is fairly obvious why patient satisfaction is being used as a proxy for healthcare outcomes, because it’s easy. As previously discussed, the limited literature in this area seems to focus more on patient experience. Patient satisfaction, however, only requires a follow up patient survey with relatively vague questions that can be answered on a 0-5 scale (i.e.: how satisfied were you with your physicians diagnosis, etc). For hospital administrators, it is fairly simple to collect and collate this data. I suspect it is for this reason that patient satisfaction is becoming an proxy for quality of care, because the vast majority of other ways to assess care would be more difficult, time consuming and arduous. For example, an excellent measure would be tracking individual patient outcomes for physicians after their healthcare visit, however, this would be quite difficult and costly. I don’t think it is clearly evident what would make a good quality measure, both from an effectiveness and cost efficiency standpoint, however, the potential harms that may be incurred from utilizing patient satisfaction should make it a rather unappealing option.


The evidence for utilizing patient satisfaction as a proxy for quality of healthcare is sparse, with most literature looking at the patient experience, rather than satisfaction. When patient satisfaction is employed as a quality measure – it results in increased spending, resource utilization and increased mortality. If hospitals continue to push patient satisfaction as a measure of care, the onus is on us to report back to administrators to ensure they’re aware of our concerns and outcomes. Furthermore, we have a responsibility to ensure our patient’s are well educated and informed around their own healthcare, so that we can make appropriate decisions together.


Fenton J, Jerant A, Bertakis K, Franks P. The cost of satisfaction: a national study of patient satisfaction, health care utilization, expenditures, and mortality. Arch Intern Med. 2012;172(5):405-411. [PubMed]
van C, Sixma H, Friele R, Kerssens J, Peters L. Quality of care and patient satisfaction: a review of measuring instruments. Med Care Res Rev. 1995;52(1):109-133. [PubMed]
Shi W, Bao L, Wu J. [Studies on a new tumor marker with monoclonal antibody against human colorectal carcinoma antigen]. Shi Yan Sheng Wu Xue Bao. 1994;27(3):377-381. [PubMed]
Cathal D, Laura L, Derek B. A systematic review of evidence on the links between patient experience and clinical safety and effectiveness. BMJ. 2013;3(1).
Chung M, Oden R, Joyner B, Sims A, Moon R. Safe infant sleep recommendations on the Internet: let’s Google it. J Pediatr. 2012;161(6):1080-1084. [PubMed]
Shahbaz Syed

Shahbaz Syed

FRCPC Emergency Medicine Physician at the University of Ottawa, with a fellowship in Digital Scholarship, and an special interest in rational resource utilization.
Shahbaz Syed
- 6 hours ago
  • Shawn Mondoux

    So how do you propose patient opinion/satisfaction/experience of care figure into healthcare delivery? Should practitioners do only what they feel is right?

    And how much of this literature comes from settings where one hospital ISN’T competing with another for more patient visits? Does your conclusion still apply in a single payer system?

    Lastly, imagine you “google” two restaurants before you head out to dinner.
    One gets 4.5/5 and the other 4.8/5. The difference is not likely to be easily quantified one your visit. In this case, the 4.5 restaurant should not spend a ton of money to get to 4.8. This will be expensive and possibly fruitless. But what if one of the restaurants scores 1.9. Will this change where you go? Should we not have the same availability with selecting our healthcare practitioners and hospitals? Is there not a minimum standard of patient satisfaction we must achieve to be patient-centred (which is only one of 6 domains of quality per the IHI)? And as an extension, if there is a minimum to achieve, should we not at least measure it?

    Implying that the patient experience should not inform the delivery of healthcare may imply that the only people fit to decide on the delivery of care is practitioners. Practitioner-centric medicine is what we have now. How well
    Is our current system at delivering quality care?

    • Shahbaz Syed

      Some insightful thoughts Shawn, thanks!
      I agree, that the patient experience is certainly an important one, and that the practice of paternalistic medicine is not going to situate itself well with the patient experience. However, targeting patient satisfaction as a primary metric (because it is easy), has demonstrated increased harms and resource utilization.
      Using your restaurant example – there are some very famous restaurants with incredible food and dining experience (i.e.: Gordan Ramsay), but interaction with customers may not be the most friendly. Despite this, these places do great as business, because the quality of their product, and food is excellent. Customers are satisfied because the quality that they’re receiving is excellent. Providing good quality, evidence based healthcare should be seen in the same light. Hospitals also have the unique responsibility of being stewards for the healthcare system (within the new CanMED leadership role), as well as minimizing harms to patients – something that we often think very little about.
      I think the confusion around this is certainly demonstrated in the last sentence, that ‘patient experience should not inform the delivery of healthcare’. Patient experience, I think is an important facet of healthcare, but is in glaring contrast to patient satisfaction, and I think this confusion has often lead to people comparing the two with the same lens, but they are very different entities. Patient experience certainly focuses (and should) on patient-centered care, and by no means am I advocating practitioner-centric medicine, I think we need to make informed and shared decisions with our patients. I am, however, arguing that patient satisfaction as a metric has never demonstrated any glaring benefits, and rather has demonstrated harms.
      There may potentially be a better metric, to be honest, I don’t know what that is – to this, I defer to the expertise of some of our great quality improvement guru’s like yourself!

      • Shawn Mondoux

        Hey Shahbaz!!! Always a great debate. If you’ll forgive me another reply, i’ll be in your debt.

        The IHI defines quality in healthcare using the 6 following domains: Safe, Effective, Efficient, Patient-centered, Equitable and Timely. The notion in QI is that if we increase any one of these, without losing ground on the other elements, we have increased quality. Let’s go back to the restaurant example and apply these principles.

        Safe: You would expect not to get sick after you eat your food.
        Effective: Adheres to all the best practices and new culinary trends that demonstrate the best tasting food.
        Efficient: When you order, the process is simple. There isn’t a million options and there is very little food waste.
        Equitable: That everybody that walks through the door is served in the same way.
        Timely: That reservation times are kept that the food arrives at your table in a reasonable timeframe.
        Patient-centred: that ultimately, they cater to the clientele. This could be measured by the customer evaluation of the restaurant. Certainly the customer should be able to offer an evaluation of the meal.

        Although some great restaurants are more minimalist in their customer service, they excel in many other areas. They make exceptional food. Furthermore, I believe their quality would be made better if they had better customer service. I also doubt you would find any Michelin 3 star restaurant with awful customer service.

        We can find a host of parallels that are seen as great institutions in which they may be lacking in a single category. The argument is not that despite this they are sufficient. Instead, it is that they would be made better if they continued to improve.

        Ultimately, an institution’s view on patient-centeredness can often reflect their culture and we often find some institutions which view this concept as essential and others view it as dispensable. My personal view is that, at minimum, we should be using satisfaction as an “alarm” score. That is to say that we don’t care about getting 100% (because good care is not always popular care) but there is a number that is too low, a number that we should always be above. This is for an institution to decide. The same could be said for MD patient satisfaction scores.

        The patient satisfaction score is particularly nefarious in the US as hospitals seek to retain “business” by catering to satisfaction scores (Press Ganey scores). The Doyle study signal is driven largely by American data (no Canadian studies) in a setting of open healthcare competition. These groups are targeting incremental increases in satisfaction scores for business ends. Here the case is made for compromises on other elements of the quality agenda.

        I would posit that in Canada, we are not in a “race to the top” environment. The papers included don’t demonstrate that we shouldn’t care for patient satisfaction scores. They demonstrate that there is no data on satisfaction in our single-payer system. What we need is better data on our performance, including from a satisfaction perspective using the data as a “minimum passable” target rather than a “perfection” target.

        But saying that patient satisfaction doesn’t matter does little for developing systems designed for the end-user.

  • Paul Horvath

    Dr. Syed, thanks for the thought provoking piece! I Particularly appreciate your distinction between patient satisfaction and patient experience. That being said, I feel the need to comment on three points.

    First, while patient experience is clearly a broader concept than patient satisfaction, I don’t think we should discount patient satisfaction as an important and, as you mention, easily measureable component of the larger whole simply because it does not encompass the entirety of a complex concept. We should, however, strive to strengthen our understanding of patient experience and work to develop better metrics.

    Number two, I disagree that obtaining high marks on patient experience mandates catering to a patient’s every last whim. Using your antibiotic example above, while not data driven, I would hypothesize that a physician could be patient centered in their approach and educate patient about the lack of efficacy of antibiotics for a viral condition rather than a) caving into their request with the sole goal of improving patient satisfaction scores or b) taking the less patient centric approach and simply refusing to prescribe without an adequate explantation why.

    Finally, the devil really is in the details. By this I mean rather than scrapping the concept of patient experience entirely, I recommend using our energy to set achievable goals. While even the best of us will occasionally get low marks or even a patient complaint, I suspect we could all intuitively recognize that a physician who is focused on high-quality human interactions with his or her patient should achieve better outcomes than one who pays no attention to the patient experience.

  • Pingback: Can you be satisfied with a longer wait? #WILTW | The Rolobot Rambles()