A Cost Analysis of Spinal Cord Injury Research
Wise Young, Ph.D., M.D.
18 December 2001
Many people are raising funds for spinal cord injury (SCI) research but do not have a clear idea of what the funds are paying for. In the article, I will explain research grants and the cost of SCI research, the funding situation in the field, how to improve the quality and rate of research, and what we are doing at Rutgers University to encourage collaboration.
Most SCI research is funded by grants from government agencies, industry, or private foundations. The National Institutes of Health (NIH) funds over 60% of SCI research in the U.S. The NIH awards three types of grants: fellowship grants which provide up to $100,000/year for 3-5 years, individual research grants (R01) which typically fund $250,000/year for 3-5 years, or a program grant which may fund up to $1 million/year for five years to a group of investigators. State agencies, private foundations, and pharmaceutical companies usually give smaller grants of $50,000-$100,000/year for 1-2 years.
Grant budgets have two categories of costs: direct and indirect costs. "Indirect costs" refer to costs incurred by the institution for the research, as opposed to "direct costs" which are expended by the scientist doing the research. Indirect costs are usually estimated as a percentage of direct costs. Indirect costs vary from institution to institution, ranging from 10% to 90%. For example, at Rutgers University, the current indirect cost rate for federal grants is 57% of direct costs. This is typical for many research universities. Thus, if NIH awards direct costs of $100,000 per year for 5 years, the award will add an indirect cost of $57,000 per year, for a total grant of $785,000 over the five-year period.
Direct costs cover personnel salaries and benefits, equipment, supplies, communication, and other costs.
Indirect costs include institutional costs for the research, i.e. buildings, electricity, heating and cooling, storage, library, and administrative tasks such as accounting, ordering, record-keeping, and reporting required by grants.
NIH funding fueled much of the growth of the medical school and major research universities of the United States over the last 50 years. SCI research must compete for NIH funding along with thousands of other medical conditions. The competition is fierce. Through the 1990's, only 15-20% of NIH grant applications were funded. With the recent increase in NIH funding, the funding rate has improved to 25%.
Direct costs of SCI experiments depend on the type of experiments. For the purposes of this discussion, let us assume that the experiments involve 120 rats that were kept for 24 weeks. Care of the rats will require a full-time technician who manages the animal care and surgery, as well as help with the euthanasia and some of the analyses. A postdoctoral fellow carries out the analyses with the help of a graduate student. The principal investigator supervises all experiments and analyses of data, spending 50% of his or her effort on the project. Let us assume that the rats are to be examined histologically for evidence of regeneration and one gene chip is used per rat to assess gene expression of the injury site. Let us further assume that the investigator already has approximately $500,000 of equipment in the laboratory.
Based on these assumptions, we can estimate a typical SCI budget for a one year period. As illustrated in the following table, the direct costs come to over $360,000 per year. Added to an indirect cost budget of over $200,000, the total direct plus indirect cost budget exceeds $560,000 per year. This suggests that the average cost per rat SCI experiment is on the order of $5,000. This costs of course would be lower if the animals were kept for shorter periods of time and less expensive outcome measures were employed. However, because since the budget has fixed costs such as personnel and equipment, the costs can be at most reduced to $4,000 per rat.
The National Institutes of Health (NIH) funded about $60 million of SCI research in 2000. Private, state, and industry funding probably totaled about $40 million, including clinical trials, suggesting a total U.S. investment of about $100 million into SCI research in 2000. In 1995, total U.S. SCI research probably was no more $60 million ($45 million from NIH and $15 million of private and industry funding). How much SCI research is this investment funding?
Number of active U.S. SCI laboratories and scientists. If the total national funding for SCI research in 2000 is $100 million, 80% of the funding was spent on animal research, and the average laboratory requires $500,000 of research support per year, this suggests that there are about 150 funded U.S. laboratories doing SCI research. If each laboratory has an average of one and a half full-time-equivalent investigators working on SCI, this suggest that there are about 240 funded scientists doing SCI research in the U.S. This number seems to be in the right ballpark. For example, the U.S. Neurotrauma Society has about 500 members, about half of whom probably are spinal injury scientists.
Number of U.S. SCI studies. A Medline search (see table below) indicates a total of 1049 SCI papers published in 1995; about 24% of these were animal studies and 34% were from U.S. groups. In contrast, 1430 SCI papers were published in 2000; about 33% of these were animal studies and 37% were from the U.S. groups. The number of U.S. animal studies doubled from 109 in 1995 to 218 in 2000. The number of papers published by U.S. scientists and clinicians declined in 1997 and 1998, particularly clinical papers. The number of clinical studies fell from 271 in 1995 to a low of 117 in 1998. By 2000, 177 clinical SCI papers were published. Although the 2001 Medline database is not complete, already 182 clinical papers have been published, suggesting that clinical studies are rebounding.
There is usually a 3-year lag time between increasing research funding and an increase in published studies, particularly clinical studies which may take several years to complete. The depressed output of 1997-98 corresponds to a low point of spinal cord injury funding in 1994-95 when increases of the NIH budget was barely keeping up with inflation rates. In 1996, however, Congress resolved to double NIH funding in seven years and steadily raised the NIH budget by about 15% per year, accounting for the rebound of both animal and clinical SCI papers in 1999-2001.
Worldwide publications of SCI studies increased steadily from 1995 to 2000. This may reflect increased investment in spinal cord injury research overseas. Between 1995 and 2000, SCI publications increased from 1049 to 1430, an increase of about 36%. The number of animal studies nearly doubled from 248 to 473. The percentage of animal studies increased from 24% to 33% of total SCI publications. In the U.S., the number of U.S. animal studies doubled from 109 to 218 between 1995-2000, going from 29% of 54% of US SCI publications. In 2000, U.S. animal studies constituted 46% of the world animal SCI studies.
How can we improve the quality and pace of SCI research? One approach is to increase funding. The SCI field has been significantly underfunded compared to the cost of the condition. A second approach is to increase the number of good laboratories doing SCI research, so that they can compete successfully for governmental funds from agencies such as the National Institutes of Health (NIH). A third approach is to improve the efficiency and productivity of the laboratories, so that more research can be done with the same funding. A fourth approach is to remove obstacles that slow down movement of therapies to clinical trial.
Increasing SCI research funding. Research funding comes from four major sources: federal, state, industry, and private. These probably totaled about $100 million in 2000 and should increase by 10-15% per year for the coming 2-3 years. However, this amount is clearly insufficient. The pharmaceutical industry recently reported that it costs $800 million to move a therapy from discovery to market. The cost of caring for traumatic spinal cord injury exceeds $10 billion. The U.S. is investing less than 1% of cost of care into research to develop solutions for SCI.
Increasing the number of competitive SCI laboratories. It is not sufficient to increase research funding when there are not enough good scientists working in the field. For example, as pointed out above, the spinal cord injury field has not been able to take full advantage of the increase in NIH funding because there are not sufficient good spinal cord injury laboratories that can compete successfully for NIH funds. In order to take advantage of the doubling of funding, we must double the number of laboratories in the field. Note that state and private foundation funding are very useful for attracting scientists to enter the field and supporting them until they can compete successfully for federal funding. NIH can help if they established a series of Centers of Excellence to help generate new scientists and laboratories for the field.
Improving efficiency and productivity of the field. Another way to improve the quality and pace of SCI research is to increase the efficiency and production of the laboratories. SCI studies are very laborious and time-consuming. The experiments take a long time and are expensive. If it is possible to develop surrogate measures or models that allow treatments and mechanisms to be studied more efficiently, the productivity of the field can be markedly increased. For example, if regeneration associated genes (RAGs) are identified, it may be possible to screen therapies for changes in gene expression rather than waiting 3-6 months for the regeneration to take place and then go through the time-consuming morphological analyses of the spinal cords to demonstrate regeneration. Likewise, having a good spinal cord injury model that is consistent and reproducible from laboratory to laboratory could markedly reduce the number of experiments that must be carried out in order to demonstrate treatment effects. Finally, it would be useful for laboratories to collaborate and share data, reducing unnecessary duplication and allowing meta-analyses to be carried out on data across studies and centers.
Moving therapies rapidly to clinical trial. In the United States, clinical trials are expensive and time consuming. In a typical clinical trial, each subject costs $30,000-$50,000 and a trial may require 2-3 years to complete. If a company is running the trial, there are additional costs for audits of the data collection and analyses, reporting to the FDA, etc. A small phase 2 trial may cost $5-10 million while a larger pivotal trial may cost $10-$30 million. Such costs are obviously not trivial and must be funded either by a pharmaceutical company or the federal government. Inability to find a sponsor often delays promising therapies for years or even decades. There is also a shortage of clinicians who are trained and inclined to run and participate in clinical trials. Some treatments such as cell transplants require multidisciplinary clinical teams, including neurosurgeons, intensivists, radiologists, physiatrists, nurses, physical therapists, and others. Establishing such a clinical trial team may take years. Finally, recruitment of subjects for clinical trials, particularly trials that involve randomization to placebo treatment, may be difficult and often take a long time. One solution to these problems is for the Federal government to establish a network of SCI clinical trial centers. Having centers that have multidisciplinary teams already set up to do clinical trials, that have access to a sufficient population of willing subjects , and that can care for and evaluate the subjects efficiently will greatly reduce the time needed to organize and execute clinical trials.
To help accelerate therapies from laboratory to clinical trial, the W. M. Keck Center for Collaborative Neuroscience provides several critical services for the field. First, we provide and train people to use the IMPACTOR model of rat spinal cord contusion. Second, we developed and provide the first large-scale and low-cost rat gene chips for SCI research. Third, we are providing stem cells and other cells for transplantation. Fourth, we hold 4 workshops training 60-80 people per year to use the IMPACTOR, do SCI research, and apply the gene chips. Over the past three years, we have trained over 100 laboratories to do SCI research, to do gene expression analyses, and to carry out and evaluate cell transplants and regeneration. Each of these services are described below.
The IMPACTOR. This device is sometimes called the NYU Impactor because it was originally developed at NYU Medical Center. This device precisely monitors and delivers a transient mechanical compression of the rat spinal cord. In addition to controlling and measuring the mechanical perturbation of the spinal cord precisely, the model standardized anesthesia, pre- and post-injury care, and outcome measures. The first well-accepted and standardized rat locomotion scale (the BBB score) was developed for this model. This scale allows locomotor recovery to be quantified reliably. The IMPACTOR model is currently used by about 150 laboratories around the world.
Gene Chips. The Neuroscience Gene Expression Laboratory (NGEL) at the W. M. Keck Center for Collaborative Neuroscience developed two rat gene chips. These chips are glass slides spotted with 5000-10,000 genes. To use the chips, one isolates messenger RNA from tissues, label the RNA with fluorescent tags, and then apply the RNA to the chips. The RNA binds to the DNA spots and the level of fluorescence reflect the amount of RNA expressed by that gene in the tissue. The first chip or NGEL 1.0 is a cDNA chip which has 10,000 genes, half of which are known. The second chip or NGEL 2.0 has 5000 identified genes and uses 70-80 mer synthetic nucleotides that are highly specific. These chips will allow discovery of new genes that may play a role in injury, repair, and regeneration of the spinal cord, as well as pain and spasticity. The chips allow treatments to be evaluated quickly and efficiently in rats.
Cell Bank. Most SCI researchers do not have ready access to rat stem cells. They must establish their own cell culture laboratories and hire people to prepare stem cells and other cell lines for treatment of spinal cord injury. We have a transgenic rat that expresses the green fluorescent protein (GFP) in virtually all cells of the body. When transplanted, these cells can be visualized in tissues without staining. We are establishing a cell bank of well-characterized rat stem and other cell lines that researchers can implant into rat spinal cord without immunosuppression. These include embryonic stem cells, fetal stem cells, neuronally restricted precursors (NRPs), glial restricted precursors (GRPs), oligodendroglial precursors, and olfactory ensheathing glia. Having a standardized and well-characterized source of cells not only saves time and money but also allows laboratories to compare results.
The W. M. Keck Center for Collaborative Neuroscience is offering these three services at cost to collaborators. For example, we provide the IMPACTOR for about $6000 which is the cost of manufacturing and servicing the device, the NGEL chips at $100 each (compared to >$1000 cost of commercial chips), and cell lines for the cost of storage, preparation, and shipping. We train people to use the model, the gene chip, and the cell transplants. These services should substantially reduce the cost and increase efficiency of SCI research. We offer these services to collaborating laboratories with one condition, that the data must be shared with other collaborators. The data are stored on two databases: the NGEL database for gene expression data and the SCICure database for injury parameters, treatments, and outcomes. Collaborators who contribute data will have access to both databases. No such databases are currently available.
A consortium of laboratories that share data has major advantages over the current laissez faire approach where scientists are finding out about results at annual meetings or journals. First, the data will be current, without the 6-8 month delay of journal publications and meetings. Second, laboratories will be able to learn about mistakes from each other so that they do not have to reinvent the wheel. Third, laboratories will be able to find out about negative results, i.e. treatments, doses, or approaches that do not work. Most laboratories do not rapidly report mistakes or treatments that do not work. Fourth, sharing data will allow more rapid consensus concerning the efficacy and safety of therapies for clinical trials. Fifth, the data can be used for meta-analyses, i.e. pooling data together from a series of smaller studies to come out with more convincing trends and . Collaboration will accelerate movement of therapies to clinical trial.
1. Grants. Most research grants pay for the direct costs of the research and indirect costs. The latter are the institutional costs of the research and are often 50-60% of direct cost. The National Institutes of Health (NIH) is the major government agency funding spinal cord injury research. NIH may fund fellowships that are about $100,000 per year for 3-5 years, individual research grants for $250,000 per year for 3-5 years, and program grants that may be as much as $1,000,000 per year for 3-5 years. State, industry, and private grants are smaller, on the order of $50,000-$100,000 per year for 1-2 years. An investigator who receives $500,000 per year of direct and indirect costs would be considered well-funded.
2. Budgets. Direct costs include personnel, equipment, supplies, communication, and other costs. Personnel costs include salary and fringe benefits that are typically 25% of salaries. An active laboratory headed by principal investigator usually includes a postdoctoral fellow, a technician, and a graduate student, costing about $190,000 per year. Most laboratories need about $500,000 of equipment in order to carry out SCI research. Assuming that all the equipment is already available, the laboratory will still need about $35,000 per year to upgrade the hardware and software. Supplies for 120 experiments per year average about $75,000 per year. Communications may cost $10,000 per year. Animal and care costs, and service contracts, will cost about $50,000. This direct cost budget of about $360,000 plus an indirect cost of $240,000 add up to over $500,000 per year, suggesting a cost of $4000-$5000 per experiment.
3. Funding. In 2000, the United States invested about $100 million into SCI research, compared to $60 million in 1995. If each laboratory requires about $500,000 and has about 1.5 full-time-equivalent scientists working, this suggests that $80 million can support about 160 SCI laboratories and perhaps 240 SCI scientists. In 2000, 1430 SCI papers were published with 33% animal studies and 37% from the U.S. In 1995, a total of 1049 SCI papers were published with 24% animal studies and 34% from the U.S. The number of U.S. animal SCI studies doubled from 109 to 218 between 1995 and 2000. However, US clinical SCI studies fell, especially in 1996-97. U.S. scientists and clinicians accounted for 37% of SCI studies and nearly half SCI animal studies in the world in 2000.
4. Research. Funding is a critical bottleneck to SCI research. Despite an overall increase of NIH funding over the past year and increasing interest by industry in funding clinical trials, the total US investment in SCI research is still about $100 million. There are not enough SCI scientists successfully competing for NIH funding. Thus, one strategy is to train more laboratories to do better SCI research. State and private foundation support is very helpful in helping nurture new laboratories to the point where they can compete for NIH funding. A second strategy is to improve the efficiency and productivity of researchers in the field. Because SCI studies are so laborious and time-consuming, providing resources so that individual laboratories do not have to reinvent the wheel, using gene expression as a surrogate measure of therapeutic effects, and forming consortia to share data should accelerate the pace of research. Finally, having a well-organized clinical trial infrastructure will reduce the delay in moving therapies from laboratory to clinical trial.
5. Collaboration. We have attempted to address some of the above obstacles systematically at the W. M. Keck Center for Collaborative Neuroscience. Over the past three years, we have trained over 100 laboratories to use the IMPACTOR, a well-standardized and reliable rat spinal cord injury model that allows investigators all over the world to compare results. We developed the first affordable large-scale rat gene chip to assess injury mechanisms and outcomes. We are developing and will provide green fluorescent protein rat stem cells and other cell lines for transplantation. These cells will reduce the work that laboratories must do to evaluate cell transplants and the cells will standardize transplantation therapy so that results can be compared across laboratories and time. Finally, we are developing a consortium of scientists, clinicians, industry, and members of the SCI community to share data and collaborate.
In summary, SCI research is very laborious, time-consuming, and consequently expensive. Although funding of SCI research has steadily increased over the past five years, with a corresponding increase in the number of studies and therapies being studied, the pace of research can be accelerated with more funding and resources. More and better SCI scientists are required to compete for NIH funds. More reliable SCI models and data-sharing should increase the efficiency and productivity of laboratories. A clinical trial network should reduce some of the delays in moving therapies from laboratory to clinical trial. At Rutgers, we have tried to address some of these problems with novel programs to train more laboratories to do SCI research, to provide state-of-the-art molecular and cellular tools that improve the efficiency and productivity of SCI research, and to promote data-sharing and collaboration in the field.