Spinal cord injuries are common in dogs and can result in a variety of clinical signs ranging from spinal hyperesthesia alone to paralysis with loss of nociception.1–6 Despite intense investigation into factors associated with outcome and the likelihood of recurrence in dogs with various spinal cord injuries, only a few studies7–9 have examined the effects of spinal cord injury on quality of life in dogs.
At least part of the reason for the low numbers of studies on the effects of spinal cord injury on quality of life in dogs is the lack of a clear definition of quality of life in animals. McMillan,10–12 for example, has suggested that quality of life represents a multidimensional balance between pleasure states and discomfort, whereas other authors have stressed animal welfare aspects of quality of life and the degree to which mental and physical needs are satisfied.13–15 Similarly, a unified definition of quality of life has been elusive in human medicine, with some authors calling for decreased emphasis on clinician-derived scales and simplification of measurement systems.16
Various attempts have been made to validate measures of quality of life in human patients with spinal cord injuries. Financial security, social interaction, and control over the environment seem to affect quality of life, whereas the effects of lesion severity, dependence on a ventilator, and lesion location are inconsistent.17,18 Recently, the use of experts to devise categories for evaluation of the quality of life in human patients with spinal cord injuries has been questioned, as these individuals are not neutral and may create instruments that merely reflect their own values.17 Furthermore, physician or researcher views on quality of life may differ from patient views.19
A questionnaire for obtaining owner-perceived, weighted quality-of-life assessments in dogs with spinal cord injuries has been described.20 The questionnaire differed from previous questionnaires developed for assessing quality of life in dogs in that owners were required to identify 5 areas or activities they believed had the most influence on their dogs' quality of life, assess their dogs' current status in each of those areas, and provide a weighting for the importance of each area so that a weighted quality-of-life score could be calculated for dogs on an individual basis.20 The purposes of the study reported here were to identify changes in weighted quality-of-life scores obtained with this questionnaire over time in dogs with various spinal cord injuries, determine whether underlying etiology affected weighted quality-of-life scores in dogs with spinal cord injury, and determine whether owner-assigned, weighted quality-of-life scores were associated with veterinarian-assigned Frankel scores for severity of neurologic dysfunction.
Materials and Methods
Study population—One hundred dogs examined because of a spinal cord injury were included in the study. Information on inclusion criteria has been published previously.20 In brief, dogs initially examined at the Texas A&M University Veterinary Medical Teaching Hospital between May 2006 and March 2007 were eligible for inclusion in the study. Owners of dogs included in the study provided written consent. The study protocol was approved by the Texas A&M University Institutional Review Board and the College of Veterinary Medicine Clinical Research Review Committee.
For all dogs enrolled in the study, an etiologic diagnosis was determined on the basis of results of physical and neurologic examination, clinicopathologic testing, advanced diagnostic imaging (eg, myelography, computed tomography, and magnetic resonance imaging), and analysis of CSF. In dogs that did not undergo advanced diagnostic imaging, the etiologic diagnosis was classified as unknown or suspected disk herniation on the basis of clinical data.
Data collection—Information was obtained on signalment, duration of dysfunction, and etiologic diagnosis. In addition, at the time of initial examination, owners were asked to complete a questionnaire20 from which weighted and VAS quality-of-life scores were calculated. Owners were asked to complete the questionnaire a second time during a recheck examination between 4 and 6 weeks after discharge. For those owners who did not schedule a recheck examination, questionnaires were mailed to their home address. Up to 2 reminder postcards were mailed to nonrespondents.
A modified Frankel score of 0 (paraplegia or tetraplegia with no deep nociception), 1 (paraplegia or tetraplegia with no superficial nociception), 2 (paraplegia or tetraplegia with nociception), 3 (nonambulatory paraparesis or tetraparesis), 4 (ambulatory paraparesis or tetraparesis and ataxia), or 5 (spinal hyperesthesia)8,9,21 was assigned at the time of initial and recheck examinations. For dogs that were not returned for a recheck examination, a modified scoring system based on owner responses to the questionnaire was used to assign equivalent scores (0 or 1 = not walking with no ability to move or feel limbs; 2 = not walking and unable to move limbs; 3 = not walking but movement present; 4 = walking with weakness or incoordination; and 5 = walking normally with or without signs of pain).
Quality-of-life questionnaire—The questionnaire used in the study has been described.20 Briefly, each owner was requested to list the 5 areas of life or life activities that he or she believed had the biggest influence on the dog's quality of life and to indicate on a 10-cm VAS labeled worst possible on one end and best possible on the other the dog's current status for each of those 5 areas. The owner was then provided with a small laminated disk that contained 5 moveable slices labeled with the 5 areas and was instructed to manipulate the disk so that the size of each slice represented the importance of each area in relation to the others. The weighting for each slice was recorded as a percentage, with the sum of the weightings totaling 100%. Separately, the owner was asked to indicate on a 10-cm VAS labeled worst possible on one end and best possible on the other his or her perception of the dog's current overall quality of life. Finally, the owner was asked to indicate on a 10-cm VAS labeled poor on one end and excellent on the other how well he or she was coping with the dog's spinal cord injury. From the questionnaire, weighted quality-of-life scores ranging from 0 to 100 and VAS quality-of-life scores ranging from 0 to 10 were calculated.
Follow-up questionnaires sent by mail did not contain the laminated disk for assigning weightings to the 5 areas of importance to quality of life. Instead, written instructions were provided for clients to provide weightings. In addition, the follow-up questionnaire contained an additional question asking respondents to score their dogs' neurologic status since the initial examination as significantly improved, improved, the same, or more severely affected.
Statistical analysis—The Shapiro-Wilk test was used to determine whether data were normally distributed. The Wilcoxon-Mann-Whitney test for independent observations was used to compare weighted quality-of-life scores, VAS quality-of-life scores, and VAS coping scores between ambulatory and nonambulatory dogs with spinal cord injuries and to compare weighted and VAS quality-of-life scores between dogs with acute injuries and those with more chronic injuries. The Wilcoxon signed rank test for paired observations was used to compare VAS coping scores and weighted quality-oflife scores assigned at the time of initial examination and again 4 to 6 weeks later, and the Kruskal-Wallis test for multiple groups was used to compare weighted and VAS quality-of-life scores among dogs grouped on the basis of spinal cord injury type. The Spearman rank correlation method was used to test for correlations between modified and equivalent Frankel scores at the time of initial examination, weighted quality-oflife scores and dog age and weight, VAS quality-of-life scores and dog age and weight, weighted and VAS quality-of-life scores and duration of dog ownership, and weighted quality-of-life scores and Frankel scores at the time of initial examination and at the time of the 4- to 6-week recheck evaluation. All tests were performed as 2-sided tests with standard software.a Values of P < 0.05 were considered significant.
Results
Information on breed, age, and duration of ownership at the time of initial examination for dogs enrolled in the study has been published.20 Median time from injury to initial examination was 3 days (range, 1 to 670 days). Etiologic diagnoses included confirmed or suspected disk herniation (n = 71), unknown (14), fibrocartilaginous embolic myelopathy (5), vertebral column malformation (3), neoplasia (2), vertebral column fracture (1), atlantoaxial subluxation (1), meningomyelitis (1), diskospondylitis (1), and cervical spondylomyelopathy (1).
Not all owners completed the questionnaire or provided a VAS score for overall quality of life at the time of the initial examination and 4 to 6 weeks later. Weighted quality-of-life scores obtained at the time of initial examination were available for 70 dogs, and weighted quality-of-life scores obtained at the time of recheck examination were available for 30 dogs, with weighted quality-of-life scores assigned at both time points available for 24 dogs. Visual analog scale quality-of-life scores obtained at the time of initial examination were available for 90 dogs, and VAS quality-of-life scores obtained at the time of recheck examination were available for 50 dogs, with VAS quality-of-life scores assigned at both time points available for 45 dogs. Visual analog scale quality-of-life scores were available for 67 of the 70 dogs for which a weighted quality-of-life score assigned at initial examination was available and for 29 of the 30 dogs for which a weighted quality-of-life score assigned at recheck examination was available. Both weighted and VAS quality-of-life scores assigned at the time of initial and recheck examinations were available for 21 dogs.
Modified Frankel scores were available for 92 dogs at the time of initial examination and 51 dogs at the time of recheck examination. For 67 dogs, sufficient information was available at the time of initial examination to assign both modified and equivalent Frankel scores. There was a significant (ρ = 0.74; P < 0.001) linear correlation between the 2 sets of scores, although scores agreed for only 38 of the 67 (57%) dogs.
Neither weighted nor VAS quality-of-life score was significantly correlated with age of the dog (ρ = 0.12 and P = 0.3 for weighted quality-of-life score; ρ = 0.009 and P = 0.5 for VAS quality-of-life score) or duration of ownership (ρ = 0.03 and P = 0.8 for weighted quality-of-life score; ρ = −0.03 and P = 0.8 for VAS quality-oflife score). Neither weighted (P = 0.7) nor VAS (P = 0.4) quality-of-life score was significantly different between dogs with acute spinal cord injury (≤ 3 days' duration) and dogs with chronic spinal cord injury (> 3 days' duration; Table 1).
Weighted and VAS quality—of—life scores assigned at the time of initial evaluation and during a follow—up evaluation 4 to 6 weeks later in dogs with various spinal cord injuries.
| Variable | Weighted score | VAS score |
|---|---|---|
| Duration of injury* Acute (≤ 3 days' duration) | 46.9 (2.2−97.2) (n = 35) | 4.6 (0−10) (n = 33) |
| Chronic (> 3 days' duration) | 42.6(5.4−100) (n = 33) | 2.6(0−10) (n = 32) |
| Etiology (initial evaluation)† Disk herniation | 41.4(2.2−100) (n = 49) | 3.8(0−10) (n = 46) |
| Other‡ | 43.4(19.8−97.2) (n = 12) | 2.3(0.5−10) (n = 12) |
| Unknown | 59.4(14.8−96.3) (n = 9) | 4.9 (0.4−7.5) (n = 9) |
| Etiology (recheck evaluation)† Disk herniation | 84.5(47.5−100) (n = 15) | 8.0 (6.0−10) (n = 14) |
| Other‡ | 76.6 (46.3−97.9) (n = 8) | 7.9(5.7−10) (n = 8) |
| Unknown | 70.3 (26.4−95.2) (n = 7) | 8.3 (2.8−9.5) (n = 7) |
Potential weighted quality—of—life scores ranged from 0 to 100; potential VAS quality—of—life scores ranged from 0 to 10. Data are given as median (range).
Weighted and VAS quality—of—life scores were not significantly different between dogs with acute versus chronic injuries.
Weighted and VAS quality—of—life scores did not differ significantly among groups.
Other included dogs with the following diagnoses: fibrocartilaginous embolic myelopathy, vertebral column malformation, neoplasia, vertebral column fracture, atlantoaxial subluxation, meningomyelitis, diskospondylitis, and cervical spon—dylomyelopathy.
At the time of initial examination, weighted quality-of-life scores were significantly (P = < 0.001) higher for ambulatory dogs (median, 59.6; range, 10.8 to 100; n = 36) than for nonambulatory dogs (median, 31.3; range, 2.20 to 96.9; 34). Similarly, VAS quality-of-life scores at initial examination were significantly (P = 0.004) higher for ambulatory dogs (median, 4.8; range, 0.10 to 10; n = 46) than for nonambulatory dogs (median, 2.3; range, 0.0 to 9.9; 44). Weighted quality-of-life scores at initial examination were only weakly correlated with modified Frankel scores (ρ = 0.28; P = 0.03). Weighted (P = 0.45) and VAS (P = 0.5) quality-of-life scores at initial examination were not significantly different between dogs with confirmed or suspected disk herniation, dogs with other spinal cord injuries, and dogs for which the underlying etiology was unknown (Table 1).
Scores reported by owners when they were asked at the time of initial examination how well they were coping with their dog's spinal cord injury ranged from 0 to 10 (median, 3.3; n = 89). Scores were significantly (P = 0.001) higher for owners whose dogs were ambulatory than for owners whose dogs were nonambulatory. Scores reported by owners at the recheck examination (median, 8.9; range, 0 to 10; n = 46) were significantly (P < 0.001) higher than scores reported at the initial examination.
Of the 51 dogs for which initial and follow-up Frankel scores were available, 33 (65%) had an increase in Frankel score ≥ 1 point and 18 (35%) did not have any change in Frankel score. The 18 dogs with no change in Frankel score had scores of 5 (n = 2), 4 (15), or 3 (1). There were 17 dogs with an increase in Frankel score for which initial and recheck weighted quality-of-life scores were available. For this group, weighted quality-of-life scores at the time of recheck examination (median, 85.6; range, 26.4 to 100) were significantly (P = 0.003) higher than initial scores (median, 40.0; range, 2.20 to 96.3). There were 7 dogs without any change in Frankel score for which initial and recheck weighted quality-of-life scores were available. For this group, weighted quality-of-life scores at the time of recheck examination (median, 56.9; range, 46.3 to 97.9) were not significantly (P = 0.09) different from initial scores (median, 42.6; range, 10.8 to 97.2). Frankel scores and weighted quality-of-life scores obtained at the time of recheck examination were significantly correlated (ρ = 0.55; P = 0.002; n = 30).
Discussion
Results of the present study suggested that owner-assigned, weighted quality-of-life scores for dogs with spinal cord injuries did not vary significantly on the basis of underlying etiology (ie, intervertebral disk herniation vs unknown vs other) and were not significantly different between dogs with acute (≤ 3 days' duration) versus chronic (> 3 days' duration) injuries. However, scores were significantly higher for dogs that were ambulatory than for dogs that were nonambulatory. At both the initial and recheck examinations, owner-assigned, weighted quality-of-life scores were correlated with veterinarian-assigned Frankel scores. However, the correlation was weak for scores assigned at the time of initial examination.
The questionnaire used in the present study to obtain weighted quality-of-life scores was based on a questionnaire (the schedule for the evaluation of individual quality of life–direct weighting) used to obtain quality-of-life scores in human patients.22 Owners were not provided specific definitions of quality of life or areas of life or life activities when completing the questionnaire. Rather, all such determinations were made by respondents. This method allowed owners to identify those areas that they believed contributed to quality of life in their own dogs and to assign a degree of importance to each area. Understanding the elements that an owner perceives as helping to determine quality of life in the owner's individual dog may allow clinicians to more effectively discuss the impact of spinal cord injury on perceived quality of life with that person.
Although we attempted to have owners who enrolled in the study complete all elements of the questionnaire, at the time of the initial examination, only 70 of 100 owners provided sufficient information to allow a weighted quality-of-life score to be calculated, and only 90 provided a VAS quality-of-life score. Similarly, only 30 of 56 owners who completed follow-up questionnaires provided sufficient information for a weighted quality-of-life score to be calculated, and only 50 provided a VAS quality-of-life score. There likely were many reasons owners failed to appropriately complete the questionnaire, although the most important probably were the emotional circumstances associated with owning a dog with a spinal cord injury, including an inability to consider that the dog's quality of life may have deteriorated, and confusion associated with the questionnaire, especially with the use of the laminated disk to assign weightings to the areas and activities associated with quality of life.20 Nonresponder bias was a potential limitation of the present study and could have affected our results if individuals who failed to complete specific elements of the questionnaire represented a fairly homogeneous group. Bias could have been introduced, for example, if owners with lower educational levels had more difficulty understanding the laminated disk used to assign weightings and had fundamentally different attitudes regarding quality of life than did owners with more education.
Various VAS scores have been used in veterinary medicine as a means of quantifying quality of life and in general have had high reliability.8–10,23 In the present study, ambulatory dogs had significantly higher VAS quality-of-life scores (median, 4.8) than did nonambulatory dogs (median, 2.3), suggesting that owner-assigned VAS quality-of-life scores were correlated with a crude measure of neurologic function (ambulatory vs nonambulatory). Similarly, weighted quality-of-life scores were significantly higher in ambulatory dogs (median, 59.6) than in nonambulatory dogs (median, 31.3), suggesting a relationship between weighted quality-of-life scores and ambulatory status. However, weighted quality-of-life scores were not strongly correlated with modified Frankel scores. Although it is possible that the low numbers of dogs in each score category contributed to this result, it is also possible that loss of voluntary motor function or nociception in nonambulatory dogs might not have had a substantial impact on owner-assigned weighted quality-of-life scores, especially at the time of the initial evaluation. Some studies17,24,25 involving human patients that were not able to ambulate because of a spinal cord injury also show a lack of correlation between degree of disability and self-rated quality of life, although caution should be taken when trying to compare these results to proxyderived quality-of-life scores for dogs with spinal cord injury. The ranges of weighted quality-of-life scores for nonambulatory (2.20 to 96.9) and ambulatory (10.8 to 100) dogs in the present study further suggested that factors beyond degree of neurologic dysfunction may have affected how owners perceived quality of life in their dogs. The fact that etiology affects prognosis for return to ambulation26 but did not affect weighted quality-of-life scores for dogs with similar degrees of impairment in the present study needs to be considered when interpreting this result. Future studies are needed to better evaluate how weighted quality-of-life scores and modified Frankel scores compare throughout the recovery process. Because the present study was performed at a veterinary teaching hospital with many of the dogs being followed up by referring veterinarians, the ability to do this in the present study was limited.
In the present study, dogs with an increase in Frankel score between the initial and follow-up evaluation had a significant increase in weighted quality-of-life score, whereas for dogs that did not have any change in Frankel score, initial and follow-up weighted quality-of-life scores were not significantly different. Thus, changes in weighted quality-of-life scores may have been a result of neurologic recovery. This comparison was complicated, however, by the fact that modified Frankel scores could not be assigned for dogs that were not reexamined at the veterinary teaching hospital, and equivalent scores had to be assigned instead on the basis of owner reports. Although modified Frankel scores and Frankel equivalent scores were correlated, comparisons involving Frankel equivalent scores need to be interpreted cautiously, as they rely on client-derived data and group neurologically normal dogs with dogs with paraspinal hyperesthesia.
It is also possible that changes in weighted quality-of-life scores between the initial and recheck examinations were related to alterations in the owners' attitudes regarding quality of life. Some studies17,24 of quality of life in human patients with spinal cord injury have found that longer duration of complete injury was associated with higher scores, likely because of adapted expectations in those patients who had had a longer period to contemplate their injury. However, this effect has not been consistent in other studies, and a previous study7 of paraplegic dogs did not show that the duration of time between injury and questionnaire completion affected owner-perceived quality-of-life scores. Our data for follow-up weighted quality-of-life scores need to be interpreted cautiously because the nonresponder rate was 44%. Although reasons for nonresponse vary between studies, socioeconomic factors, participant impressions of the questionnaire, and dissatisfaction are potential sources of nonresponse and may introduce bias.27,28 Furthermore, distribution of follow-up questionnaires by differing means (ie, mail vs provision at a recheck appointment) may have affected our results. It is possible, for example, that individuals who distributed questionnaires unintentionally provided verbal or nonverbal cues to owners regarding their impressions of a dog's recovery or additional instructions for owners who were having difficulty understanding the survey.
Owner ability to cope with a dog's spinal cord injury was assessed to a limited extent as part of the present study. Coping VAS scores were significantly lower for owners of nonambulatory dogs than for owners of ambulatory dogs, suggesting that loss of ambulation produced emotional distress in owners. The inability of many owners to cope with the circumstances surrounding a nonambulatory dog's neurologic status needs to be considered when communicating medical information to these individuals. Coping VAS scores were higher at the time of recheck evaluations likely because most dogs had improved neurologically by this time. However, it is also possible that scores were higher because some owners had emotionally adapted to their dogs' neurologic limitations over time.
Given that the questionnaire used in the present study was validated in the same population of dogs,20 our results have to be interpreted cautiously. Further validation of the questionnaire, including validation by individuals other than the original authors, is needed. Nevertheless, although we strongly support the use of neurologic injury scales as a means of describing outcome in dogs after spinal cord injury, we believe that owner-based assessments of quality of life should also have a role in defining success. Furthermore, a general understanding of factors that owners perceive as being important contributors to quality of life in dogs with spinal cord injuries would be helpful when rendering advice concerning treatment. Although the primary goal of medical or surgical treatment should be to improve neurologic function, this goal should be understood in light of those factors important to each individual client.
ABBREVIATION
| VAS | Visual analog scale |
Intercooled Stata, version 9.2, Stata Corp, College Station, Tex.
References
- 1.
deLahunta A. Veterinary neuroanatomy and clinical neurology. 2nd ed. Philadelphia: WB Saunders Co, 1983.
- 2.
Cerda-Gonzalez S, Olby NJ. Fecal incontinence associated with epidural spinal hematoma and intervertebral disk extrusion in a dog. J Am Vet Med Assoc 2006;228:230–235.
- 3.
Beal MW, Paglia DT & Griffin GM, et al. Ventilatory failure, ventilator management, and outcome in dogs with cervical spinal disorders: 14 cases (1991–1999). J Am Vet Med Assoc 2001;218:1598–1602.
- 4.
Dewey C. A practical guide to canine and feline neurology. Ames, Iowa: Iowa State Press, 2003.
- 5.
Chen AV, Bagley RS & West CL, et al. Fecal incontinence and spinal cord abnormalities in seven dogs. J Am Vet Med Assoc 2005;227:1945–1951.
- 6.
Stiffler KS, Stevenson MA & Sanchez S, et al. Prevalence and characterization of urinary tract infections in dogs with surgically treated type 1 thoracolumbar intervertebral disc extrusion. Vet Surg 2006;35:330–336.
- 7.↑
Bauer M, Glickman N & Glickman L, et al. Follow-up study of owner attitudes toward home care of paraplegic dogs. J Am Vet Med Assoc 1992;200:1809–1816.
- 8.
Levine JM, Levine GJ & Johnson SI, et al. Evaluation of the success of medical management for presumptive thoracolumbar intervertebral disk herniation in dogs. Vet Surg 2007;36:481–490.
- 9.
Levine JM, Levine GJ & Johnson SI, et al. Evaluation of the success of medical management for presumptive cervical intervertebral disk herniation in dogs. Vet Surg 2007;36:491–498.
- 10.
McMillan FD. Quality of life in animals. J Am Vet Med Assoc 2000;216:1904–1910.
- 11.
McMillan FD. Maximizing quality of life in ill animals. J Am Anim Hosp Assoc 2003;39:227–235.
- 12.
McMillan FD. Calls attention to difference between quality of life and health status (lett). J Am Vet Med Assoc 2006;229:26.
- 13.
Wojciechowska JI, Hewson CJ & Stryhn H, et al. Evaluation of a questionnaire regarding nonphysical aspects of quality of life in sick and health dogs. Am J Vet Res 2005;66:1461–1467.
- 14.
Wojciechowska JI, Hewson CJ. Quality-of-life assessment in pet dogs. J Am Vet Med Assoc 2005;226:722–728.
- 15.
Wojciechowska JI, Hewson CJ & Stryhn H, et al. Development of a discriminative questionnaire to assess nonphysical aspects of quality of life of dogs. Am J Vet Res 2005;66:1453–1460.
- 16.↑
Gill TM, Feinstein AR. A critical appraisal of the quality of quality-of-life measurements. JAMA 1994;272:619–626.
- 17.↑
Hammell KW. Exploring quality of life following high spinal cord injury: a review and critique. Spinal Cord 2004;42:491–502.
- 18.
Duggan CH, Dijkers M. Quality of life after spinal cord injury: a qualitative study. Rehabil Psychol 2001;46:3–27.
- 19.↑
Slevin ML, Plant H & Lynch D, et al. Who should measure quality of life, the doctor or the patient? Br J Cancer 1988;57:109–112.
- 20.↑
Budke CM, Levine JM & Kerwin SC, et al. Evaluation of a questionnaire for obtaining owner-perceived, weighted quality-oflife assessments for dogs with spinal cord injuries. J Am Vet Med Assoc 2008;233:925–930.
- 21.
Levine JM, Ruaux CG & Bergman RL, et al. Matrix metalloproteinase-9 activity in the cerebrospinal fluid and serum of dogs with acute spinal cord trauma from intervertebral disk disease. Am J Vet Res 2006;67:283–287.
- 22.↑
Hickey AM, Bury G & O'Boyle CA, et al. A new short form individual quality of life measure (SEIQoL-DW): application in a cohort of individuals with HIV/AIDS. Br Med J (Clin Res Ed) 1996;313:29–33.
- 23.
Mullan S, Main D. Preliminary evaluation of a quality-of-life screening programme for pet dogs. J Small Anim Pract 2007;48:314–322.
- 24.
Westgren N, Levi R. Quality of life and traumatic spinal cord injury. Arch Phys Med Rehabil 1998;79:1433–1439.
- 25.
Fuhrer MJ, Rintala DH & Hart KA, et al. Relationships of life satisfaction to impairment, disability, and handicap among persons with spinal cord injury living in the community. Arch Phys Med Rehabil 1992;73:552–557.
- 26.↑
Dewey CW. Myelopathies: disorders of the spinal cord. In: Dewey CW, ed. A practical guide to canine and feline neurology. Ames, Iowa: Iowa State Press, 2003;277–336.
- 27.
Fejer R, Hartvigsen J & Kyvik KO, et al. The Fumen neck and chest pain study: analyzing non-response bias by using national vital statistic data. Eur J Epidemiol 2006;21:171–180.
- 28.
Kotaniemi JT, Hassi J & Kataja M, et al. Does non-responder bias have a significant effect on the results in a postal questionnaire study? Eur J Epidemiol 2001;17:809–817.
