Tag Archives: wonk time

Unfounded Assumptions and Faulty Logic, Ahoy!

s.e. emailed me an alert to this Atlantic post by Derek Thompson about the potential hidden costs of failing to extend unemployment insurance benefits, with a warning that I would likely ‘enjoy’ it. So I clicked through with trepidation. Rightfully so. I’ll excerpt the relevant piece here so you don’t need to reward this nonsense with pageviews (and if you do, avoid the comments, which are full of “unemployed people are fundamentally lazy” tripe):

Consider this statistic, from Peter Orszag at an event in DC called The Future of American Jobs Part II: The number of applications for the Social Security Disability Insurance has increased from an average of 500,000 per quarter in 2006 to 750,000 in 2010. Why? Well, it’s unlikely that American disabilities themselves have increased by 50 percent in the last four years. It is more likely that healthy Americans discouraged from the awful job market have sought out disability insurance and collected Social Security money even though they’re not actually disabled.

SSDI requires that its recipients be unable to work. Unemployment insurance, on the other hand, requires that its recipients look for work. Like any government program, UI can be gamed, and sometimes it surely is. But the fundamental point holds: If we stop supporting unemployed people with cash, there is a risk that we kick them out of the labor force onto disability insurance, where we pick up the tab by paying for them through SSDI not to work and not to look for work.

WHERE TO EVEN BEGIN, amirite? Let’s start with identifying the huge, glaring factual error in the argument – the assumption that all SSDI requires for eligibility is “that its recipients be unable to work.” You may not be surprised to learn that the long term disability insurance program through the federal government actually requires that the individual’s inability to work be due to a disability, rather than due to a lack of jobs at their skill level, or a lack of jobs that pay living wages, or any number of other reasons that a person would be unable to work. Yes, this disability benefits program actually has requirements about having a disability! And the Social Security Administration requires quite a lot of verification from medical professionals (who it considers more objective and reliable than people themselves) that it independently reviews and assesses before determining whether a person has a disability that qualifies them for SSDI. You cannot just walk into a Social Security office and demand SSDI benefits, no matter how long you’ve been out of work.

Now let’s look at the statistic he is using to lend legitimacy to his argument: “The number of applications for the Social Security Disability Insurance has increased from an average of 500,000 per quarter in 2006 to 750,000 in 2010.” What does this statistic tell us? It does not tell us anything about the trends, if any, of people receiving SSDI – just the number applying for it. Granted there is some cost to Social Security to review and assess these applications, but the main cost Thompson is raising is the amount of money spent on SSDI benefits – about which this statistic tells us absolutely nothing. Even the knowledge that there are 250,000 more SSDI applications per quarter this year than in 2006 tells us nothing about whether the application (or approval) rates are anywhere near the actual underlying prevalence rates.  Even if disability rates have remained perfectly stable, we have no idea whether the 500,000 applications per quarter in 2006 was undercapturing the number of people who were actually eligible for SSDI.

(This built-in, unstated assumption that the starting point in increasing disability benefit rates was the “correct” or desired rate and that any increases are dirty lying cheating fraudulent people is extremely common. You see it in discussions about changes in special education enrollment, rates at which students are identified with learning disabilities, and more. And the assumption that the lower rate correctly reflects the actual prevalence in the population is so ingrained that people, including Thompson, do not even state that they are making such an assumption. It should be obvious to the reader, they imagine, that lower disability benefit rates are right!)

But Thompsons’ main point – that people currently in the workforce may shift to disability benefits if they are unable to find work – gets to the conflation of “disabled” and “unemployable” that I’ve discussed before. There are certainly people who would meet the criteria for SSDI who are employed right now, influenced by any number of factors including the stigma against not working, the difficulty of verifying eligibility for SSDI, and the likely higher income available through work. Unemployment may motivate those people towards SSDI, but so could a bunch of other things, like failure to provide reasonable accommodations at work and employment discrimination against people with disabilities. But a whole lot of people who are unemployed are simply ineligible for SSDI, because they’re unemployed for economic reasons and do not have a disability that would qualify them for SSDI.

A potential takeaway from this kind of thinking is that for reasons including saving money on disability benefits, we should mitigate and eliminate barriers to employment for people with disabilities. But this post concludes that because everyone on unemployment could pull down the higher SSDI benefits if they got the notion, we should support the unemployed. That’s not relevant either to people with disabilities or without them and is fundamentally nonsense.

Language Interpretation and Health Care

As promised, here is a post and a place to discuss policies and procedures about language interpreting in a health care context. Unfortunately, it is going to be a slap dash post as I’m knee deep in a number of things, both professional and personal, including preparing to clean ALL THE THINGS and cook a great number of the things for the upcoming U.S. Thanksgiving.

So. Here are some assorted thoughts and resources on interpretation policy!

My first and most prominent thought is the absolute importance of having interpretation done well. It is essential that the health care provider be able to understand the patient’s reported symptoms and experiences and sensations, their medical history, their concerns, and their desires and goals for the medical treatment or care. It is also essential that the patient be able to understand the instructions and questions of the provider and discuss any potential treatment options, including tests, medications, and therapies. Without accurate and competent translation, the provider cannot understand what is wrong and the patient cannot understand how it will be addressed. While there are times that medical treatment takes place without communication – primarily emergency situations when the patient is unconscious – that is definitely the exception rather than the rule and is not a good model for non-emergency care or even most emergency care. Communication is crucial and should be important to all individuals involved in health care transactions. In other words, bad communication could kill people.  So, this is an enormously important issue, both to people who do not speak English (or whatever the predominant language is where they are) and to people with hearing impairments and d/Deaf people.

Problems with interpretation:

In my experience, there is usually a vast gulf between interpretation policies as written and as implemented in practice. Even when a policy requires use of a trained and certified interpreter during all interactions, with clear instructions on when and how to obtain an interpreter, there will be interactions where no interpreter is present. (See, for example, this recent audit of interpreter use by New York City police finding that officers routinely failed to comply with the existing interpretation policy.) This is in large part because of logistical difficulties with interpreter use that I’ll discuss below and seems to be a feature of nearly every interpreter policy. This kind of built in failure rate makes me inherently nervous about relying on interpreters.

Interpreter policies are logistical nightmares, in large part due to the difficulty of matching language competence to language need that we discussed last week. This is especially true in areas with significant language variation – the court where I worked had several Spanish interpreters and an Armenian, Russian, Farsi, Cantonese, and Tagalog interpreter on call all the time, and still had incidents when a litigant needed interpretation in a language that wasn’t available. This is even more difficult in an emergency room situation, when it is impossible to anticipate when a need will arise and there is no way to just have everyone wait for a bit until an interpreter gets there. This leads to using interpreter phones – where the provider and patient use a handset to speak to an off-site interpreter, which are both expensive and clunky. Even if an interpreter is on staff, finding them in the building, pulling them out of a meeting or their break, and getting them to the desired location to provide interpretation is often a hassle.

There also needs to be an entire infrastructure around training interpreters and measuring and certifying their competence. Language knowledge is not enough for competent interpretation, interpreters must also be trained in interpretation skills and professional standards. There is a lot to well-done interpretation – positioning yourself so the provider and patient can maintain eye contact, knowing when and how to ask clarifying questions, providing concurrent interpretation without interrupting the flow of the conversation. This training and certification requires time and money.

The limited-English-proficient patient has very limited ability to address problems in interpretation policy. To address the failure to provide an interpreter at all, they can provide their own, usually by bringing a friend or family member to do the interpretation. This is a big problem for a number of reasons. First, there’s no way to know the competency of the friend or family member in interpreting, especially with specialized medical vocabulary and concepts. When the doctor tells me that the leg pain I am experiencing could be a rupture of my plantaris tendon, plantar fascitis, or tendonapathy, is there any probability that my sister will be able to interpret that accurately? (Personally, I’d have trouble identifying even what portion of the leg was being discussed if I didn’t have Medline open to get those terms in the first place.)  Second, this is a major confidentiality issue. If I am going to the doctor to discuss genital warts, I am unlikely to be as frank or forthcoming if my brother is my interpreter. This is especially true for medical issues around sexuality, pregnancy, mental health, and injuries from domestic violence, when discussing symptoms and experiences can open the patient to strong stigma effects. This is even more problematic if a child is serving as an interpreter and being expected to interpret discussion about conflicts between his parents or his mother’s suicidal thoughts, which could be dangerous or traumatic for the child.

Overall, I am most concerned about the inherent non-compliance caused by the logistical hassle and the patient’s agency being limited to bringing a friend or family member to interpret, and it is these issues that make me prefer language proficient medical professionals rather than reliance on interpreters. I also think that language proficiency for providers would help with cultural competence issues, but that’s more a side effect.

So – have at it! Interpretation? Language proficient providers? What makes sense? What do we want?

How Can We Get More Bilingual Health Care Providers?

Los Angeles County, where I live, is incredibly diverse, both racially and linguistically. According to the Asian Pacific American Legal Center’s 2008 report on Language Diversity and English Proficiency in LA County (pdf file), more than half of Los Angeles County residents speak a language other than English at home. “The 10 most frequently spoken languages countywide are: English, Spanish, Chinese, Tagalog, Korean, Armenian, Vietnamese, Farsi, Japanese and Russian.” The report lists 39 distinct languages and almost 10,000 residents speak another language not on the list. About 29% of county residents are Limited English Proficient (LEP), which means they have some degree of difficulty communicating in English.

All of these people will have some contact with the health care system at some point in their lives and ideally, at more than one point. And, according to the U.S. Department of Health and Human Services’ Agency for Healthcare Research and Quality, having a health care provider who speaks the same language as the patient has lots of important benefits to patient care (internal cites omitted):

A growing body of literature finds that language concordance between patients and providers (i.e., both speak the patient’s primary language well) results in greater patient understanding, leading to increased satisfaction, better medication adherence, greater understanding of diagnoses and treatment, greater well-being and better functioning for persons with chronic disease, and more health education.

The report goes on to note that only some of these problems can be mitigated by having the conversation interpreted. Using an interpreter can also disrupt the human connection between provider and patient and curtail full discussion. Interpretation, no matter how competent, is somewhat cumbersome and inefficient.  This all means that having the health care provider be bilingual is by far the best solution, both for patients and providers. And often, cultural conceptions of health issues like pain, disability, and mental illness can be very important in understanding how a patient is describing symptoms or experiences, so a native speaker is the best.

The problem, of course, is that not all health care providers are bilingual. This means those who can speak another language are relatively valuable, so they can be more exclusive or take more high-paying jobs. If you are a patient who can afford to pay high rates or has great insurance, you may be able to get a provider who speaks your language, but you’re not guaranteed. And if you are a poorer patient and rely on emergency rooms and county health clinics for care, you just have to hope to get lucky. As I once heard it put, “If you’re a Cambodian therapist, you can basically write your own ticket. You’re not going to work at the County Department of Mental Health.”

So how can we go about getting more bilingual health providers, especially for relatively low-paying jobs to care for low-income patients? There’s no obvious answer. Here are some ideas, and their potential drawbacks:

  • Require all health care providers to become fluent in another language. Providers all go through some training and licensing procedures, so we could build in a language requirement. There are some obvious difficulties – how would we ensure languages were proportionally represented? would we match providers to areas where there was a need for the language they spoke? how do we make sure someone learns, say, Hmong? would the cost of administering all those language proficiency tests be better spent elsewhere in the health care system? It also doesn’t serve our goal of having providers be native speakers.
  • Recruit more native speakers of non-English languages to become health care providers. This could take a lot of different forms – scholarships and incentives for these people to enter training programs or medical schools, reaching out to younger kids to stimulate interest in health care professions, providing tutoring or other support resources, or a number of other methods. Most of these things would take a very long time before they resulted in a change in the makeup of health care providers. It’s also unclear how effective any of these methods are, and how much they cost.
  • One possible solution is always to throw money at it. We could dedicate a lot of funding to paying big salaries for providers who speak other languages. While that would probably work, and relatively quickly, it would cost a lot of money. And would continue to cost a lot of money to maintain. And, most importantly, would not do anything to increase the total number of health care providers who teach non-English languages and if more were induced to enter the profession because of the high salaries, it would cost more and more money over time. To compensate, there would be fewer and fewer health care providers overall, or some other significant effect on the health care system from the significantly shrinking resources.

Personally, I support a little bit of all three. (Equivocation is a policy-maker’s prerogative.) Increasing incentives for health care professionals to know and learn non-English languages, aggressive recruiting for native non-English speakers to become health care professionals, and paying bonuses or other incentives to bilingual providers. To make the best of the current situation, I also support training interpreters and ensuring they’re used appropriately.

Are there other policies you think would help the problem?

The Veil of Ignorance

As some of you who follow me on Twitter know, I spent the weekend in Kansas City with my grandfather, who had experienced an as-yet-to-be-diagnosed health crisis and was in the hospital. (Nothing is dire and he’s expected to make a full recovery and be discharged in the next week or so.) At home in Los Angeles, most of my hospital experience is with Los Angeles County Hospital, an enormous facility that is massively overloaded in trying to meet the health care needs of Los Angeles County’s indigent population. In Kansas City, I was spending time in the Rehabilitation Care wing of what seemed to be a very well funded private hospital. And the differences were monumental.

Los Angeles County Hospital serves about 39,000 inpatients a year, with over 150,000 emergency room visits a year. This is utterly massive, and has led to a lot of complaints about overcrowding, with allegations that emergency room visitors wait an average of 35 hours to be seen – sometimes without even having their vital signs taken. A recent LA Times story showed that likely because of administrative policy hiccups, a patient was admitted to another county hospital and was an inpatient for two days before being assigned a doctor. In another county facility, since closed down, a woman waiting to be seen in the ER bled out and died in the waiting room without being seen. Having spent some time at those hospitals, I can attest that while the relatively new buildings are pretty nice, the old building, in which a lot of care is still provided, feels like a rickety relic of the early 1970s, with sliding metal doors that make the patient rooms feel like drawers in a filing cabinet.

In contrast, this other hospital felt like a palace. There was free valet parking for outpatients. All of the rooms were private and spacious, with room for 9 visitors to fit inside and with a window looking out on trees in gorgeous fall colors. The nurses all knew my grandfather’s name, and his wife’s name, and my name, and my mom’s name. The emergency room in the facility had a gorgeous waiting room looking out into an interior courtyard with fountains and plantings and even a creek running through it. I visited that waiting room often, because that’s where the vending machine with the Diet Coke was, and I never saw more than a handful of people there. Certainly nobody was going to die there without being seen.

Describing the two hospitals like this, I can guess which one we would all choose for ourselves, our loved ones, and our friends. (The nice hospital, just to be clear. I’d rather be at the nice one.) But often when we’re making policies – especially policies for the health care of low-income people – policy makers are not thinking about how they would like to be taken care of, the facilities they would like to be in. They are making policies for other people, policies they know will never apply to them.

That’s why I’m such a big fan of the concept of the “veil of ignorance.” The idea comes from American philosopher John Rawls’ book Theory of Justice, considered an important text in political philosophy.  The veil requires a person to create a standard of justice without knowing what place or value they will have in that society. As Rawls described it:

Among the essential features of this situation is that no one knows his place in society, his class position or social status, nor does any one know his fortune in the distribution of natural assets and abilities, his intelligence, strength and the like. I shall even assume that the parties do not know their conceptions of the good or their special psychological propensities. The principles of justice are chosen behind a veil of ignorance.

I like this idea so much. This would require us to make health care policy without knowing whether we would have the most super expensive fancy health insurance policy possible, or have no health insurance and depend on the indigent care available through the county; without knowing whether we would have a disability or be in perfect health; without knowing any of the categories or identities we would be a part of.

It is, of course, impossible to think about policies without drawing from personal experience, but I believe that policymakers should do their best to assume this veil of ignorance. At the very least, they should consider how their policies will apply to and affect the most disadvantaged person possible, to understand the full scope of potential problems that could be created by the policy. Now if only we could get voters to look at things this way…

Figuring Out the Actual Goal of a Policy

Last time, we talked about three main areas of a policy where things can go wrong: drafting a written policy to match the envisioned goal or mission of the policy (articulation), creating an administration or structure to carry out the policy (implementation), and making sure that people are actually following the policy (enforcement). These three areas are extremely complicated when everyone involved has a common view of what the policy is trying to achieve. When there’s no clear agreement on the intended goal of the policy, things get even more confusing. For an example of that kind of confusion, let’s look at a recent policy proposal in New York State to prohibit the use of food stamps to buy soda. To understand what’s really going on here, we’ll need to take a closer look at a couple of things: what the intended goal and policy of the food stamp program is, what the proposed policy is, and how the two interact.

Let’s start with food stamps. The United States food stamp program started in 1939. Although it currently functions as an anti-poverty program, it started as a subsidy for American farmers administered by the U.S. Department of Agriculture as a way to help American food producers by making it easier for consumers to buy their goods. Originally, consumers bought coupons that could be used for food and were issued additional coupons that could be used for “surplus” food – food that was being produced but not purchased because of the problems in the economy. Even now, changes to the food stamp program take place in congressional “Farm Bills” that primarily focus on agricultural issues.

The food stamp program is also one of the primary anti-poverty programs administered by the federal government. Over time, eligibility for the stamps has been narrowed to include only low-income individuals and families. Unlike some other anti-poverty programs, eligibility is based on more than just an individual’s or household’s income and includes a calculation for how much they are paying for housing costs to determine the amount of income considered available for purchasing food. (You can go through the pre-screening on the federal food stamp website if you want to estimate your eligibility – if you’re from outside the U.S., just pick a random state.)

Although a person is approved for a certain dollar amount of food stamps per month, the benefit is not given to them as cash. Instead, they get that dollar amount of “food stamps” on an electronic debit card that can be spent only on food — it can’t be used for rent, gas money, diapers, clothing, vitamins, medicine, toiletries, or any other non-food items. There are also restrictions on the food that can be bought with food stamps: no hot or pre-prepared food (like a deli sandwich) and no alcohol or cigarettes. These rules get somewhat complicated – pumpkins can be bought if they’re edible, but not if they’re a decorative gourd.

So, given that framework, we can now understand what New York state is proposing: a demonstration project for New York City that would “bar the use of food stamps to buy beverages that contain more sugar than substance — that is, beverages with low nutritional value that contain more than 10 calories per eight-ounce serving. The policy would not apply to milk, milk substitutes (like soy milk, rice milk or powdered milk) or fruit juices without added sugar.” And the stated goal of the policy: “bring[ing] us closer to stemming the wave of obesity and diabetes in New York.”

The question of how much and whether these drinks do or do not contribute to an increase in diabetes and obesity (obesity which may or may not be a health issue) is a whole other question – for the purposes of this analysis, let’s just pretend we live in a world where that connection has actually been proven. Even if we accept that very tenuous connection, a big problem with this proposed policy is both overbroad and underbroad. Policies that are overbroad change more things than they need to in order to reach their intended goal. This policy is overbroad because there are surely people who use food stamps to buy some of the prohibited drinks who are neither obese nor have diabetes – so the policy would be prohibit purchases that do not contribute to the harm we are trying to prevent. It is underbroad partly because of what’s excluded – the sugar in fruit juice or chocolate milk is just as sugary as the sugar in soda. But it’s also underbroad because there are surely people purchasing these drinks with their own money – earned, inherited, or otherwise gotten not through food stamps – to buy these drinks that will contribute to their obesity and/or diabetes.

If we were this convinced that these sugary drinks were the root cause of an obesity epidemic, there would be a lot more effective ways to target this policy – by banning the sale of those drinks in the state, to anyone, using any method of payment. Or selling them only to adults. But this isn’t what’s being proposed. Partly because the soda companies would create a huge stink in protest and partly because the population would complain that they have the right to spend their money how they like and would likely vote out any politician that made such a policy. Instead, this is a “demonstration project” that New York promises the “effects [of which] would be rigorously evaluated.” This is really an experiment to find out if limiting purchases of these drinks makes any difference at all, either to the amount of sodas sold or any eventual health outcomes. And it’s a pretty bad experiment, because even with the tight control we have over how poor people spend their money, they could still buy a Coke with their non-food stamp income.

So what’s the real goal of this proposed food stamp policy? To further restrict the purchasing power of people low-income enough to receive food stamps in order to find out if this idea they have has any effect at all on what they want to try to change: the health of the overall population. And this kind of implementation is not at all unusual – these half-baked policy ideas that don’t have a firm scientific underpinning for their presumed causal effect are often tested out on vulnerable populations that don’t have the political power to resist them. If this was implemented on the entire population of New York City or state, there would be a giant backlash. But the poor aren’t nearly as well organized or politically active, so it’s safe to practice on them, for the good of everyone’s overall “health.”

Further reading:

I Love Policy

It’s true. I am a policy wonk. I am endlessly interested in it. I read about it, think about it, talk about it and … write about it. (As in, what I’m doing right now.) And I do all of this because I think it’s immensely important. Crucially important. Vitally important.

Public policy is how the government – whether local, state, provincial, federal, or any other level – takes action on a particular issue. It covers a whole huge range of potential state actions – allocating and spending money, setting and enforcing professional guidelines and standards, creating agencies and staff, structuring tax incentives, even defining what constitutes criminal behavior. That’s an extremely big category that clearly has an enormous and unparalleled effect on the world.

Public policy not only drives state and governmental actions, it also has enormous influence on private sector actors. Tax policy can encourage specific areas of business, grants can encourage specific methods or practices, and governments both licence and regulate businesses. This combined effect on public and private actors means that to my mind, changing public policy is the quickest and most effective way to change things for a big group of people, all at once.

Policy touches almost everything we do and everything with which we come into contact. Right now, I am sitting on my bed, the mattress of which complies with regulations to prevent it going up in flames. I am wearing a shirt made in the United States by workers subject to minimum wage laws and industrial safety protections. The US shirt manufacturer is protected from competition from international producers by trade tariffs and taxes. The soda I am drinking displays nutritional information pursuant to federal regulations. The internet I am using is regulated by the Federal Communications Commission. Even the air I am breathing is affected by pollution standards and the decisions to grant or deny permits to things like coal processing plants. Even my kitty is included – she’s protected from abuse by criminal statute and, where I live, is protected from declawing. And that’s just scratching the surface of all of the policies surrounding and affecting me right this second.

There are, of course, a lot of other factors and forces that influence how people interact with the world, both as individuals and in groups. There’s huge effects from family, religious, cultural, and ethnic beliefs and traditions. Then there’s a myriad of individual differences – the things a person reads and watches and talks about and is talked to about, for example. I would argue that each of those things could also be influenced by government policy – like how the private movie ratings system created as a reaction to public regulation prevented me from seeing R movies until I was at least 16, or what books my local public library system bought and so were available for me to read. I’d also argue that individual preferences and differences are a lot less important than public policy in determining whether an area has a functioning health care system.

There are obviously a lot of difficulties with public policy. First, it’s mainly done by politicians, so political climate and general popular opinion can limit the range of what policies can achieve. For example, the United States could never have created a government-run nationalized health care system given the current makeup of our decision makers. Second, achieving specific goals through policy can be kind of complicated and difficult – if you were the government and wanted to “fix the education system,” it’s not exactly clear what specific steps would reach that outcome, even if we could agree on what a good education system would look like. Third, the differences between a policy as carefully written down and a policy as actually implemented can be vast, so a great policy may end up being too difficult or complicated or expensive or just impossible to implement, or may end up being significantly watered down.

At the end of the day, though, policy is literally life and death. Whether a mentally ill teenager gets tased or shot by a police officer depends on law enforcement policy, training, and management. Whether a PWD can afford and access the medications and equipment they require to live. Policy determines how and why and for how long and under what circumstances people are institutionalized. Whether and how they are protected from abuse and neglect from caretakers and family. Whether and when and how they have children.

So in the coming weeks, I’ll be writing about policy. Good policy, bad policy, and everything in between. Policy often doesn’t feel as sexy or gripping as a lot of the other topics we discuss here, but I’m hoping you’ll find it as interesting and important as I do.