Generic filters
Filter by Categories
Clinical Practice Research
Ethical Legal and Regulatory Considerations
Planning, Managing, and Publishing Research
Research Design and Method

Health Service Research: What Is It?

An overview of definitions, designs, and methods.

Michael Weiner

DOI: 10.1044/cred-pvid-implscid1p4

The following is a transcript of the presentation video, edited for clarity.

I’m going to talk about health services research and introduce that to you in terms of definition and some examples of what is health services research and what isn’t health services research.

I think you’ll start to see by the end how this is a little bit different from implementation science, but is very much intertwined and interlocked with the aspects of implementation science.

My goals for today are to define how services research, learn about some key concepts, and consider relationship to implementation.


In terms of simplistic terminology, if you can understand what a health service is, you can understand what health services research is. So let me step back and say, “What is a health service?” Perhaps most of you in this room provide health services and don’t need a lot of explanation about that. But even if you provide health services or if you are already conducting health services research I think it’s useful to think about the nature of your work as it may relate to implementation.

These are some definitions according to the World Health Organization. “All services dealing with the diagnosis and treatment of disease or the promotion, maintenance, and restoration of health.” So you may say, oh, I do that.

I think that third bullet there may help you think further about what it means for this to be related to research. “Service provision refers to the way input, such as money, staff, equipment, and drugs are combined to allow the delivery of health interventions.”

So it’s a little bit more than just providing a diagnosis or providing a treatment, but really how you’re organizing the delivery of health care itself that is going to embody the health service.

Here you can see some factors needed to improve access, coverage, and quality. You need availability of services, you need organization, and you need incentives for providers or clinicians.

So now you can imagine what health services research is because it’s the study of health services, but look at it this way. Health services research refers to the ways in which organization, structure, process, health technology, social factors, financing and personal behaviors are related to and influence these other kinds of outcomes shown on the right here: Access to care, quality of care, cost of care, and health itself. I’m going to talk a little bit about those four items in a minute.

You can see that health services research can be done at any level from an individual up to a community, up to a population and beyond.

Academy Health, which is an organization focusing on health services, research, policy, and practice also describes health services research in these terms. What works? For whom? At what cost? Under what circumstances?

You can see from this that it’s starting to touch on implementation a little bit, perhaps, and sometimes the lines are a little bit blurred. But health services research as distinct from implementation isn’t so much emphasizing and focusing on the implementation process and the adoption and the fidelity as much as these other kinds of outcomes.

So it’s, perhaps, a stepping stone or a pathway towards thinking further about implementation.

What is not health services research? This is important, too, distinguishing between different kinds of disciplines because you need to then relate that to the methods and the approaches that you use.

When we talk about basic and clinical research we’re typically not talking about health services research. An efficacy study of a clinical intervention with a clinical outcome is not health services research study. We don’t think of animal studies and bench science studies as health services research studies. In addition, epidemiological studies that may look at the prevalence or incidence of a disease is not a health services research study. However, you could look at the prevalence or incidence of a treatment or a health service itself and then that kind of reframed epidemiological study might actually be considered health services research.

Concept and Methods

Let’s look just for a minute at some aspects of cost, quality, health, and access and it’ll give you a little bit of a flavor for the kinds of issues and outcomes at hand here.

What is access to care? You can see some examples of factors that are important in thinking about access to care. Obviously, supply and demand. The affordability of a health service. The acceptability of a service. And then is it geographically distributed in a way that provides access to people? Patients’ knowledge about a health system is also related to access because if you don’t know about health service or an opportunity it’s very hard to access that unless it happens by chance or somebody brings it to you.

Let me move on to cost. The term cost is very loosely used in some settings. Maybe inappropriately so. And cost can refer to so many different things that it’s helpful to be explicit what you mean when you’re referring to cost.

So think about cost to whom. Who’s incurring the costs? Are they private costs? Are they out-of-pocket costs? Is it a public payer that is actually paying for something? Of course there are other kinds of cost like emotional and psychological costs. So remember to ask yourself cost to whom.

Let me go to quality just for a minute. Again, these are really sound bites, overview issues to get you thinking a little bit more about what all of this means.

What is quality? This might be the hardest one partly because quality means so many different things to different people. Something that is high quality for you might actually not be the most important quality factor for me.
But these are examples. Quality often can refer to safety or include aspects of safety. Aspects of errors. Medical errors, clinical errors, diagnostic errors, therapeutic errors.

Adherence to guidelines represents one aspect of quality. And in the cases where there aren’t guidelines to tell us what to do we think about standards of care, which are more commonly adopted and accepted practices among a clinical community.

So all of these things relate to quality. There are many aspects of quality that go beyond these issues.

Let me go to quality just for a minute. Again, these are really sound bites, overview issues to get you thinking a little bit more about what all of this means.

What is quality? This might be the hardest one partly because quality means so many different things to different people. Something that is high quality for you might actually not be the most important quality factor for me.
But these are examples. Quality often can refer to safety or include aspects of safety. Aspects of errors. Medical errors, clinical errors, diagnostic errors, therapeutic errors.

Adherence to guidelines represents one aspect of quality. And in the cases where there aren’t guidelines And what is health? We could spend many days, I’m sure, just talking about that. Some factors are shown here. Quality of life, well-being. So these are a lot of the more subjective things. Things that might be hard to measure. Qualitative in nature.

We can think of more objective aspects of health, like absence of disease or survival itself. And we could debate about whether these things really represent health or not.

The Research Question Determines the Best Study Design

Let me talk just for a few minutes about possible study designs or study approaches because that may get you thinking about ways that you could start to think further about health services research.

You do need to pick a study design that actually matches the research question that you have. Sounds very obvious, right? But it’s a mistake that has been made many times where the wrong study design is actually selected for a question.

o tell us what to do we think about standards of care, which are more commonly adopted and accepted practices among a clinical community.

So all of these things relate to quality. There are many aspects of quality that go beyond these issues.

We can have observational or interventional studies even in health services research. Many of the studies in the field are observational in nature, but, of course, if I’m trying to introduce a new health service or change a health service or improve it, then I may be looking at an interventional study design.

These are examples are very commonly used study designs in health services research, and I’ll mention a few of these.

A case report or a case series obviously is an observational study design where you are discussing, reporting, characterizing one or more events that have been observed. This is often used to study small numbers of cases and it’s typically not looking to identify statistical associations, but more to record something new to start to understand what it means.

Correlational studies are looking at strengths of relationships among pairs of variables, pairs of factors, and there may be many factors in a correlational study. We’re interested in thinking about outliers and outliers of a correlational study can have an important impact on the interpretation of findings and thinking about whether the distribution of the results are skewed in some ways and don’t represent a normal distribution.

In cross-sectional surveys, we’re looking at slices of time or windows typically where we’re trying to assess exposure in disease in a defined population. And this can be used to assess prevalence. It could be a first step in a cohort study that gives me some initial data and findings. Maybe raises some new research questions.

A case series in a cohort study are similar, but they do have some important differences.

Of course in both a case series and a cohort study you’re looking at a population that has experience, something in common in terms of either an exposure or outcome.

But you can see some of the important differences here. In both a case series and a cohort study there’s an outcome. There is often an exposure. You can think of an exposure as a predictor variable of some kind. It might be a treatment. It might be some other factor in a person’s life. A case series may or may not actually have an exposure.

The basis for sampling is where we have the main difference. In a case series what defines the case series is an outcome and that’s the thing that actually is common for all the participants in a case series.

In a cohort study it’s the opposite. What a cohort has in common, in a traditional cohort study, is that they have a common exposure or a common other factor that has led to an outcome and the outcomes may vary across the cohort. And so most typically we’re interested in understanding the relationship between the exposure, or the predictor, and the outcome.

You have access, perhaps, to many different kinds of data. You may in your current studies be in the position where you’re collecting primary data. You’re doing direct interventions. You’re gathering data about your own clients or patients.

There’s so many other kinds of datasets, it’s important to think about them. There could be administrative data that come from a claims or government sources or payers about your clients. There could be other kinds of clinical data from medical records about your clients or study participants. There could be study registries, public health data, death data, and even databases that have been collected and saved to be reused for multiple research purposes.

In health services research we often use combinations of datasets that we can link with each other to enrich the findings that we can actually generate from a study design.

Examples of Health Services Research

Let me give you a couple of examples of health services research studies from literature. I’m just going to show you snapshots of these. And some of them might not actually be health services research. I just highlighted some phrases that are indicative of a kind of service.

Here’s one that is a survey of otolaryngology services in Central America discussing needs for comprehensive intervention. And, you know, as I’ve just, sort of, gone through some literature and identified examples, in this particular study the scientists or clinicians are looking at variation in the availability of different kinds of clinicians by country, and then differences in the training programs by country.

So this is a good example of health services research because it’s looking at the actual access to services and what the differences are according to different factors.

Here’s another one. Implications of the World Report on Disability for Responding to Communication Disability in Underserved Populations. Again, you’re getting this flavor: services, access to care, maybe some issues with quality.

Another one: Twenty year view of hearing characteristics and audiologic practices before and after newborn hearing screening. Hearing practices and so on.

Another one. Mining and analysis of audiology data to find significant factors associated with tinnitus masker. So this one I can’t tell from looking at the title. Is this really health services research or is it more clinical study looking at clinical factors with outcomes and predictors of outcomes? I think we may not be so sure on just looking at the title.

Regional differences and outcomes after cochlear implants in children. So there, again, maybe I’m not sure. You know, if you’re looking at the efficacy of the implant to improve understanding of speech, I would say that’s not really a health services study. It’s more about clinical research study. Again, a clinical outcome.

I was on the bus yesterday coming over, and one person said, was saying I studied ultrasound to help people understand their placement of their tongue. That’s an example of a clinical research study, not a health service research study. It could become a health services research study if one started to look at some other aspects of that treatment.

And then another person was saying, I help children learn how to construct complex sentences using language cues. Again, that is a clinical study, but it could become a health services research study once you want to look at some other factors like: How much does it cost? Does everybody have the ability to engage in such a intervention or practice? Does it actually help well-being? Does it help other kinds of outcomes?

This one is looking at aspects of-you can see here. Parent participants, support teacher hours, mode of communication, educational placement. In a particular intervention it is comparing geographic areas within a country. East side of the country. West side of the country. This is health services research.

Oh, this one’s about parachute use to prevent death. And if you haven’t seen that one, “We think that everyone might benefit if the most radical protagonists of evidence-based medicine participate in a double-blind, randomized, placebo-controlled crossover trial of the parachute attempts.”

So it’s a good example of a few things. First of all, it’s not health services research. It is clinical research. And second is sample of a study that we couldn’t do. We couldn’t do that intervention, but we could study it. We could look at and observe the natural practice of parachute use. And depending on our outcomes we could turn it into health services study.

So common pitfalls. I do want to touch on this before I end because this is very important.

First of all, think of ascertaining the problems. Think about is your ability to ascertain either a treatment, a population, a disease. Is that reflecting the sample or the population that you are actually interested in? It never will perfectly. And so your question is how close am I? How good is my ascertainment?

Think about missing data. Non-linear data. Outliers. Nowadays, you know, if we’re submitting grant proposals to NIH or AHRQ and other agencies, if we don’t actually talk about how we’re going to handle missing data, it’s probably a major flaw in the study design.

Relating resource utilization to quality. Some people sometimes make the mistake that if I’m coming into my clinic more often as a patient, it means I’m getting better care. That might be true. Or the opposite might be true. It might be that I’m actually really sicker than everybody else and that’s why I need more care. And you can’t tell the relationship between resource utilization and quality unless you dive into it and understand it.

Association does not prove causation. Two things are associated. And maybe one causes the other, but you might not be sure which one causes which. And it might be that neither one causes the other. You have to look for that problem if you want to find out and understand it.

Confounders and bias obviously are very important problems.

Relevance to Implementation Science

For the sake of time I’m going to end here with how is this related to implementation? If you look at the box at the bottom, it contains some factors related to implementation and implementation science, integrating findings into practice, reasons for adoption and effectiveness, applying findings and new settings, these things are really intertwined with the health services factors that are shown here at the top from when I was describing it earlier.

Depending on where you’d like to focus your problem, your research question, it may be a services question. It may be an implementation question. It may be some of both. And I think just like there are hybrid designs with clinical research and implementation science one could imagine a hybrid design with health services research study and implementation science. And many such studies have been designed and conducted.

If you’d like to find a professional society that dedicates their attention fully to this, it’s Academy Health. And they do not just research, but policy and practice.

Key points I mentioned already but they’re summarized here. Think about your study design. Your sampling. Your sensitivity and specificity of ascertainment. Your missing data.

Questions and Discussion

Audience Question

When you look at big databases, people don’t ask about communication problems, like the Medicare beneficiary survey. How do we change that? Because we know that there are lots of people with communication disorders. We know they have a complex constellation of things going on, but the databases don’t let us look at it.

It’s a great question. It’s a very complex question because whether that information is considered valuable for inclusion in that database depends on the purpose of the database from the beginning. In other words, if it’s a database used to account for reimbursement for Medicare beneficiaries and reimbursement is not tied to whether I asked the patient about communication it’s not going to happen, right? And similarly for so many other conditions.

I work in the geriatrics field and there’s so many important measures of geriatric function that are not represented in our day-to-day databases that we have to find other solutions.

One approach is: OK, I’m going to create a new database and I’m going to link it to the one that I need the other data out of.

Another is you go to the parent organization of this database and discuss what the value proposition is and whether it makes sense to try to actually change the way that database is being organized. That may be either difficult or very difficult depending on your circumstance. But it has been done. I don’t want to discourage you from doing it. Just to say put some energy behind it.

A separate Q&A Panel including this presenter is available online: Implementation Science Summit: Panel Discussion with Renee Boothroyd, Michael Weiner, Robert Horner, Kirsten Senturia & Matthew Kreuter


Dekkers, O. M., Egger, M., Altman, D. G. & Vandenbroucke, J. P. (2012). Distinguishing case series from cohort studies. Annals of Internal Medicine, 156, 37–40 [Article] [PubMed]

Lohr, K. N. & Steinwachs, D. M. (2002). Health services research: An evolving definition of the field. Health Services Research, 37(1), 15 [Article]

References: Websites of Interest


World Health Organization. Health Services.

VA Health Services Research and Development Service (HSR&D). What is Health Services Research?.

Michael Weiner
Indiana University School of Medicine

Presented at the Implementation Science Summit: Integrating Research Into Practice in Communication Sciences and Disorders (March 2014). Hosted by the American Speech-Language-Hearing Foundation.

Copyright © 2015 American Speech-Language-Hearing Association

Share on facebook
Share on pinterest
Share on twitter
Share on linkedin