Introduction

Communication about the diagnosis, treatment, and prognosis of cancer is challenging for both veterinarians and pet owners.¹ Communication can positively or negatively impact owner compliance with treatment plans, overall satisfaction with a clinical experience, and recall of information disseminated during a consultation.^2–4 Owners of pets with cancer value when information is communicated in a variety of formats. Specifically, as an alternative to a verbal explanation, written discharge material is regarded as an important modality for relaying information.⁵ The significance of providing owners with written information about their pet’s health is emphasized by data showing reduced owner compliance when such material is omitted.⁶ Discharge summaries are an example of written material that may enhance owner education and help owners understand and implement care plans following a diagnosis of cancer in their pets.

Readability describes the ease with which a reader can comprehend written text. Readability of written information can be measured with a variety of formulas. The Flesch-Kincaid Grade Level (FKGL) and Flesch Reading Ease (FRE) scores are the most widely verified and implemented readability analyzers used in research.⁷ Both tests evaluate various metrics of written text including sentence length, as measured by the mean number of words in a sentence, and word length, as measured by the mean number of syllables per word. Readability scores generated by such tests can be used to approximate the grade level an individual needs to possess to be able to accurately comprehend the written material.

Most adults in the US read at an eighth-grade level, and 1 in 5 adults reads at or below a fifth-grade level.⁸ To ensure comprehension by most adults, the American Medical Association (AMA) and NIH recommend that medical information intended for the public be written at no higher than a sixth-grade reading level.⁹ Yet the readability of human health-care resources is well above the recommended reading levels, with most materials written at a 10th-grade level.^8,10 With specific reference to discharge summaries, readability exceeds recommendations across various medical subspecialties including oncology, emergency medicine, pediatric medicine, surgery, internal medicine, and ophthalmology.^11–16

Unlike human medicine, no guidelines exist regarding the recommended readability level of animal health information intended for the public. Additionally, limited research suggests uniform poor readability of veterinary health information. Three studies confirmed that the readability of online information specifically designed as educational tools for owners about various animal health issues exceeded AMA and NIH guidelines.^17–19 Another study showed that consent forms for clinical trials conducted within a variety of veterinary specialties did not meet current health literacy recommendations.²⁰ When comparing the readability of online pet allergy information with that of human allergy information, veterinary websites were more difficult to read and written at a higher grade level than human health websites.²¹ To our knowledge, there is no information assessing readability of discharge summaries provided to pet owners in veterinary medicine. Recognizing that pet owners appreciate written information but having no data regarding the readability of discharge summaries represents a void in the opportunity to gauge the veterinary professionals’ success or failure in delivering impactful content.

The objective of the study reported here was to evaluate readability levels of discharge summaries distributed to owners of pets newly diagnosed with cancer by the medical oncology service at a teaching hospital. Our hypothesis was that discharge summaries provided to owners would be written above the recommended sixth-grade level. Additionally, we hypothesized readability would remain difficult regardless of diagnosis or recommended treatment.

Materials and Methods

After institutional review board and hospital board approval was received, the computerized medical records database of the North Carolina State Veterinary Hospital was searched for all new accessions to the medical oncology service between June 2017 and January 2019. From this sample, a random number generator²² was used to select a subset of discharges for analysis. Discharge summaries for patients transferred to the oncology service from another service, not confirmed to be discharged from the service, or that were euthanized or died were excluded. Discharge summaries for patients without a definitive or highly probable oncological diagnosis were also excluded.

Summaries were compiled (Word version 16.20; Microsoft Corp) by directly copying and pasting text exactly as it was written in the medical database. This document was subsequently stripped of all identifying features (owner’s name, pet’s name, attending house officer’s name, student’s name, faculty member’s name, etc). At our institution, discharge summaries are used as both a referral letter to the primary care veterinarian and as a discharge summary for the owner. Therefore, analysis of readability was restricted to the portion of the discharge summary intended for owners. Specifically, this included the assessment section, which is broken down into several subsections including disease information, treatment options (surgery, radiation therapy, chemotherapy, or other systemic treatment, alone or in combination), instructions for at-home care (medications, monitoring, activity, and diet), and the follow-up plan. Information related to the clinical history, physical examination, or results of diagnostic tests was excluded, as it was intended for the primary care veterinarian.

The owner-directed portions of the discharge summaries were analyzed with the use of an online readability calculator²³ that assessed the FKGL and the FRE score, both of which are previously established methods for evaluating readability and considered highly accurate by experts in the field.¹⁷ The FKGL score is calculated to approximate the US grade level, or to suggest the number of years of education, required to understand the material. Material should target at or below the sixth-grade level. The FRE score is rated on a scale from 0 to 100, with 0 to 30 rated as very difficult, 30 to 50 as difficult, 50 to 60 as fairly difficult, 60 to 70 as standard, 70 to 80 as fairly easy, 80 to 90 as easy, and 90 to 100 as very easy.²⁴ In addition to the readability scores, data reported from the readability calculators included the mean number of sentences, total word count, and percentage of words with ≥ 3 syllables.

Statistical analysis

A Kruskal-Wallis test was used to test for differences between different parts of communications. Where a difference was found, all pairwise Wilcoxon rank-sum tests were conducted to detect differences. Summary values are also provided. For portions of the data in which observations were made from different parts of individual records, a paired approach was pursued by use of the Wilcoxon signed-rank test. All statistical analyses were performed with a commercially available statistics program (R version 4.0.3; The R Foundation). Values of P ≤ 0.05 were considered significant.

Results

The initial database search returned 1,088 records. To optimize heterogeneity during sample analysis, 130 discharge summaries were randomly selected from this larger sample. After the application of exclusion criteria, 118 discharge summaries were available for final analysis.

The breakdown of summaries by tumor type was as follows: lymphoma (LSA; n = 25), mast cell tumor (13), soft tissue sarcoma (12), osteosarcoma (OSA; 10), melanoma (7), hemangiosarcoma (6), squamous cell carcinoma (SCC; 6), urothelial carcinoma (5), and apocrine gland adenocarcinoma of the anal sac (3). The remaining 31 discharges represented summaries with 2 or fewer cases per tumor type.

The mean ± SD word count for all summaries was 1,095 ± 496 words (median, 998 words; range, 174 to 2,242 words). The mean ± SD sentence length for all summaries was 60 ± 26.5 sentences (median 58 sentences; range, 13 to 123 sentences). The mean ± SD percentage of words with ≥ 3 syllables was 20 ± 2.5% (median, 20%; range, 14% to 27%).

When the owner-directed portion of the text was analyzed in its entirety, no discharge summary was written at or below the AMA- and NIH-recommended sixth-grade level and the lowest numeric readability achieved was equivalent to an 8.6th-grade level. The mean ± SD FKGL score was 11.9 ± 1.1 (median, 11.9; range, 8.6 to 15.5; target ≤ 6). The mean ± SD FRE score was 43 ± 5.9 (median, 42.7; range, 25.5 to 58.1; target ≥ 60). Ninety-three percent (110/118) of summaries were rated as difficult or very difficult with FRE scores < 50. The mean FKGL and FRE scores for discharge summaries pertaining to new diagnoses of cancer did not differ significantly (P = 0.52 and 0.67, respectively) based on tumor type for those most encountered (LSA [n = 25], mast cell tumor [13], soft tissue sarcoma [12], or OSA [10]; Table 1).

The 3 tumor types with the most difficult readability scores for discharge summaries for affected pets were apocrine gland adenocarcinoma of the anal sac, hemangiosarcoma, and melanoma. The 3 tumor types with the least difficult readability scores for discharge summaries for affected pets were SCC, LSA, and OSA. However, the readability of discharge summaries for pets affected by these latter 3 tumor types still exceeded guidelines, with FKGL and FRE scores ranging from 11.4 to 11.6 and 43.4 to 43.8, respectively. The mean FKGL and FRE scores for discharge summaries pertaining to new diagnoses of cancer did not differ significantly (P = 0.95 and 0.45, respectively) on the basis of treatment type discussed (chemotherapy or other systemic treatment, surgery, or radiation, alone or in combination; Table 2).

Discussion

Results demonstrated that discharge summaries related to a new diagnosis of cancer written at a single teaching hospital in the present study exceeded AMA and NIH guidelines, regardless of tumor type or treatment options presented. This aligned with existing information in the human medical literature regarding readability of discharge summaries provided to patients in various subspecialties.^11–16 Results supported previous work demonstrating poor readability of owner-directed animal health-related materials.^17–20

In people, several features are identified as detrimental to discharge education including a lack of standardization of written material and provision of education materials written above a 10th-grade level.¹³ Specific consequences of discharge instructions that exceed readability recommendations include misunderstanding of instructions for prescription drugs, missed appointments, and an inability to recognize complications related to a diagnosis.^25,26 Poor readability is also associated with decreased comprehension of written materials.^27–29 Inadequate education and insufficient clarity in follow-up plans result in decreased caregiver confidence and can adversely affect patient outcomes.¹³ Conversely, enhanced readability of discharge instructions is associated with fewer phone calls after hospital discharge and a lower readmission rate.³⁰ Although research linking readability with comprehension and patient outcomes in veterinary medicine is lacking, it is reasonable to assume that when owner materials exceed recommended readability levels, patient outcomes could be negatively affected. This could include noncompliance in scheduling follow-up visits, incorrect medication administration, inadequate at-home monitoring, and decreased ability to make informed choices regarding diagnostic and treatment options. Additionally, although the implication of poor readability of discharge information is unknown in veterinary medicine, breakdowns in the exchange of information between owners and veterinarians have been shown to negatively impact owners’ perceptions of clinicians’ motivations.³¹

The AMA, NIH, CDC, and Joint Commission provide published recommendations for health-directed literature.^32–35 These guidelines collectively include strategies to improve readability and overall comprehensibility of text. This includes limiting sentence length to 8 to 10 words, limiting paragraphs to 3 to 5 sentences, and primarily using words with 1 or 2 syllables. In addition to improving readability, they also include guidelines to use bulleted or numbered information, infographics or videos, and bolded key points where possible. Discharge summaries evaluated in the present study clearly did not adhere to these guidelines with a mean 6-fold excess of sentence length and 20% of words with ≥ 3 syllables.

The 3 tumor types with the least difficult readability scores for discharge summaries were SCC, LSA, and OSA, which are common diagnoses in companion animals. Despite numerically more comprehensible scores, the mean readability for discharge summaries for patients with these 3 tumor types exceeded existing human health-care recommendations, with each having had a reading level > 11th grade. Therefore, in our sample, higher readability scores were not restricted to cases involving rare diagnoses or tumor types unfamiliar to attending clinicians, which would have been scenarios that could align with an inherent tendency to be more challenging to explain in written text.

One limitation of the present study was the unique patient and owner population. The readability of discharge summaries from a single subspecialty at a teaching hospital may not reflect those of other subspecialties, hospital types (eg, referral private practice or primary care facility), or health conditions. Further studies are necessary to determine whether other veterinary specialties, primary care veterinarians, or both also provide discharge materials that are written at a level above that of the average American. Additionally, the demographics of owners who presented their pets to the referral hospital of the present study may not have reflected those of a more general population of pet owners. Lower health literacy is associated with factors including older age, male sex, racial or ethnic minority status, and socioeconomic status–related characteristics including lower education and income. These variables were not assessed among pet owners in the present study and could have been expected to influence comprehension of written material.²⁵ Additionally, we did not measure the true literacy rates of pet owners in the present study, which could have fallen above or below the average American. Another consideration was possible repetition among discharge summaries, including information regarding disease behavior and prognosis and description of available treatments. Given that all evaluated discharges were written above the recommended sixth-grade reading level, we believe that this was unlikely to have impacted the results of the present study. Lastly, no individual clinician characteristics (eg, years of clinical experience or familiarity with the concept of readability) were considered as factors in readability levels; however, such individual clinician characteristics could have contributed to bias in our analysis.

As readability metrics are only 1 measure of the comprehensibility of text, future studies may benefit from additional measures of information quality and comprehensibility. More recently, the DISCERN instrument has been used by patients and care providers to critically evaluate the quality of written health materials.³⁶ This rubric asks each user to score the material in various categories such as the reliability of the publication (documentation of sources, relevance, biases) and quality of treatment information (completeness, support for patient-centered care, and risks and benefits of different treatment options). The goal of this tool is to help providers create better health materials and assist patients in evaluating and understanding various materials they may be provided.³⁶ Monitoring patients’ usage of online health information portals has also been used to determine which elements of the medical record are perceived to be of greatest value to the patient.³⁷ Finally, although measuring text readability is helpful for clinicians creating owner-directed materials, the best way to judge whether materials are an effective communication tool is by formatively assessing comprehension of the intended audience. Testing targeted written materials using focus or sample groups with short quizzes or interviews will be the most effective way to gauge text comprehensibility.

The establishment of recommendations for improving comprehension of human health-care information by institutions, such as the AMA, NIH, CDC, and Joint Commission, provides a framework by which health-care professionals can design and gauge the efficacy of their educational materials. Similar recommendations by the AVMA or American Animal Hospital Association may assist veterinary professionals in achieving similar goals. Veterinarians should be aware of strategies to enhance owner education and practice effective communications, with special attention paid to the comprehensive nature of written discharge summaries.