Relationship Between Caregiver Attendance at Wound Clinic Visits and Wound Healing in Patients With Diabetes: A Retrospective Multicenter Evaluation

Diabetes is the Trojan Horse of diseases; a single diagnosis with numerous felonious associations. As a prominent chronic disease in the United States, 34.2 million people are affected and 1.5 million new cases are added annually.^1,2 The proatherogenic impact of diabetes implicates this pathology in the number 1 cause of death, cardiovascular disease.^2-4 The complications of diabetes are numerous and worsen with time.^5-8 Well-known complications are nephropathy, retinopathy, and cardiovascular and peripheral arterial disease (PAD); however, the sequelae of dementia, liver disease, cancers, and pancreatitis are often underappreciated.^9-13 In addition, PAD leads to neuropathy, which may culminate in ulcerations and deep tissue infections that, for many, can result in amputation of the digits and limbs.^9,14-16 A patient with diabetes has a risk of amputation that is 20 to 30 times higher than that of a patient without the disease.¹⁷

Foot complications, predominantly ulcerations, are so significant that the American Diabetes Association (ADA) has published clinical guidelines specifically for the management and prevention of these conditions.^11,12,18 The recommendations detail necessary examinations, special population awareness, and surveillance for high-risk individuals.¹⁹ In contrast to recommendations from other disease-focused societies, the ADA’s clinical guideline does not define the caregiver role or include caregiver education, anticipatory guidance, or tools to assist the patient collaboratively.²⁰

Similar to patients with other serious chronic diseases, such as dementia, seizure disorders, organ transplants, and cancer, clinicians and family members observe that people with diabetes often cannot manage the condition alone.^13,21,22 As witnessed by the author and supported by a qualitative vascular study by Zamani et al (N= 6), patients with diabetes-associated wounds are often overwhelmed.²³ As a potential impact of wound complications, patients may have compromised mobility, may be physically unable to reach the lower limb to apply wound dressings, and may be unable to drive.⁶ Loss of visual acuity, common to the disease, can impair wound dressing techniques and all aspects of diabetes self-care.¹²

As seen in the author’s wound clinic, some patients find personal care assistance essential; individuals fulfilling this role are often referred to as a caregiver, family member, or care partner.²⁴ The impact of the caregiver role as related to the management of diabetes-associated wounds of the lower extremity (DWLE) has not previously been studied.

The purposes of this retrospective descriptive study were to evaluate the group characteristics of patients with DWLE who received treatment at an outpatient wound clinic, and to determine whether a correlation exists between regular caregiver attendance at these appointments and wound healing outcomes.

Design. This retrospective study included patient data from January 1, 2018, through December 31, 2018. The hospital-owned electronic health record (EHR), which documented every wound care encounter, was the source for data retrieval for all eligible patients assigned to the principal investigator at 3 outpatient hospital-based wound center locations.

Patients were scheduled for weekly wound care visits that included regular clinic wound debridement and care. A variety of moist wound dressings and offloading devices selected by the clinician were used. Disease mitigation and management strategies were universally encouraged. A proprietary software was utilized to generate patient lists filtered for DWLE diagnosis codes at each of the participating sites. Study variables were selected based on commonly tracked wound metrics and availability in the EHR. A custom form was developed specifically for data collection for this study. Data were extracted and then deidentified by using a pencil-and-paper retrospective chart review method.

The following inclusion criteria were used: patient age 18 to 100 years; patient sex male or female; confirmed diagnosis of type 1 or type 2 diabetes; and an index wound of the lower extremity (hip to toe) coded as diabetic etiology. Exclusion criteria included diagnosis of PAD with an ankle brachial index (ABI) of 0.5 or less, consistent with acute or chronic limb ischemia; patients with malignancy or inflammatory wounds, including pyoderma gangrenosum, and hospice patients electing noncurative care were also excluded. Patients were considered lost to follow-up and data was not included if the patient notified the staff of relocation or care termination, or if care was initiated and the patient never returned for subsequent treatment, leading to administrative discharge after 3 months' absence.

Data. EHR data abstraction included patient demographics of age, sex, race, body mass index (BMI), and self-reported highest education level. Other clinically relevant information retrieved from the EHR included current tobacco use, concurrent dialysis, ABI, glycosylated hemoglobin (HbA1c), insurance status, and history of nonadherence to treatment. Insurance status was retrieved and abstracted, including if the patient was uninsured or unfunded as noted in the EHR demographics. History of nonadherence was extracted if noted in the patient’s list of diagnostic codes (International Classification of Diseases, 10th revision). The BMI was extracted from the EHR without further computation (this value was computer-generated following input of the patient’s height and weight).

Data extraction also included a numerical count of the total number of wound clinic visits attended during the interval period; this was represented as the Visit Count (VC) and was determined by the number of times the patient presented to the clinic for treatment of the index wound.

The patient EHR contained the data field “Accompanied by” that documents anyone who was present during a given visit. Any relationship entry in this field served as a positive response (documented as the Accompanied Count [AC]) versus the phrase “alone” or “none,” reflecting a negative response. The accompanying caregiver was defined as any individual who had a personal relationship with the patient. As a general clinic practice, an individual or group entity only providing transportation was not considered an accompanying caregiver; consequently, these visits were recorded as the patient attended alone. Examples included public transportation, community or hospital bus, taxi, or emergency services. If the field was left blank (a seldom occurrence), the visit was neither counted nor recognized as positive or negative, thereby equally reducing the total VC and the AC. The total AC was extracted and was divided by the total number of index wound encounters (VC) to arrive at a percentage of accompanied visits (% Accompanied). Based upon this percentage, patients were categorically divided into 2 groups: those accompanied by a caregiver at 50% or more of their clinic visits and those accompanied by a caregiver at less than 50% of their wound care visits.

The Weeks in Service (WIS) count was generated and tabulated automatically by the EHR to account for current cumulative weeks enrolled in clinic services with an open wound; this numerical value was extracted directly without manipulation.

Wound measurements were completed at every wound care encounter, entered into the EHR, and automatically tracked within the system. Data regarding a single index wound were extracted. Length and width were measured in centimeters and entered in the EHR at the bedside with computer-generated calculations for area (cm²) and percent change in size; these calculated data were extracted and utilized for this study. In the rare circumstance a patient had more than 1 wound, the principal investigator made a clinical determination to abstract data only for the most significant or difficult-to-heal wound based on size, location, severity, duration of time open, or threat to limb. If the patient presented with full wound re-epithelialization and closure, the wound was deemed healed, and the patient was discharged from services with self-care and monitoring instructions. Thereafter, the wound was subsequently categorized as Wound Healed, and this determination was noted and extracted. Quantification of laboratory inflammatory markers, imaging findings, and detailed vascular or surgical interventions, though relevant for clinical management, were beyond the scope of this research

Data analysis. Data were collected by pencil-and-paper technique and transferred into a digital spreadsheet. Descriptive statistics were used to determine number, mean, range, and standard deviation. A student’s 2-tailed unpaired t-test was used to determine the statistical significance of the difference in mean scores of continuous variables between the 2 groups. Noncontinuous data were converted numerically followed by t-test application for stastical analysis. Level of significance was determined at P < .05. A Pearson r product-moment correlation coefficient was calculated between the % Accompanied of the 2 groups and 5 wound healing variables of visit count (VC), accompanied count (AC), weeks in service (WIS), size reduced percent (SR), and whether the Wound Healed. to determine the presence and degree of correlation as well as significance. The magnitude of correlation of the Pearson r was defined as r < ±0.09 = no correlation, r = ±0.1 to ±0.29 = weak, r = ±0.3 to ±0.49 = moderate, and r = ±0.5 to ±1.0 = strong. All statistical analyses were performed using Microsoft Excel (version 16.50; Redmond, WA).

Ethical considerations. The institutional review board granted determination of exemption (Exemption Qualification Notice, 4.9.2019) permitting the study, which was externally supervised by the research faculty. Patient consent was obtained at enrollment of services at each facility. There was no cost associated with the study other than investigator time.

One hundred twenty-six (126) total patients were identified for the study; of these, 15 patients met the criteria of hospice enrollment (n = 2) or ABI below threshold (n = 13) and were excluded. Of the remaining 111 patients, 26 were lost to follow-up and not included in the analysis. A total of 85 patients remained. Of those, 30 were accompanied by a caregiver > 50% of the time (Accompanied) and 55 were accompanied < 50% of the time (Unaccompanied)

The demographic characteristics of the sample group and the categorical separation of the 2 subgroups are detailed in Table 1. Overall, the sample group of 85 patients had a mean age of 61 years (range [SD], 28–89 [12.06]) and was mostly male (n = 47; 55%) and White (n = 46; 54%). Health and lifestyle demographics (Table 2) showed 99% (n = 84) of participants had type 2 diabetes; mean HbA1c was 8.55% (range [SD], 5.0–13.9 [2.16]) and mean BMI was 36.7 kg/m² (range [SD], 17.7–66.1 [9.93]). The wound metrics data (Table 3) showed an average starting wound size of 17.39 cm² (range [SD] 0.07–510.5 [56.2]) with a reduction in area of 64% (range [SD], 100 to >+200 [140]). A total of 63 patients experienced full healing of the index wound for an overall Wound Healed rate of 74%. The mean AC was 3.3 (range [SD], 0–23 [4.6]) out of a mean visit count (VC) of 12.28 (range [SD], 1–50[11.12]) for an average of 34% (range [SD], 0–100% [41]) accompanied visits. On average, patients received in-clinic care for 18.7 weeks (WIS; range [SD], 1–118 [24.8]).

The Accompanied group (n = 30) had a mean age of 65 years (range [SD], 28–89 [12.76]), which was significantly higher (P = .03) than that of the 55 patients in the Unaccompanied group (mean, 59 years; range [SD], 31–80 [11.28]). Both groups were majority male patients (60% vs 53%), and no significant difference was seen (P = .52). The Accompanied group had a greater percentage of White participants (77%) whereas only 42% of the Unaccompanied group was White; 49% of Unaccompanied patients were Black (n = 27), 5% were Native American (n = 3), and 4% were Hispanic (n = 2); a stastical signficiance between the groups was seen in this race variable (P = .002). Both the Accompanied and Unaccompanied groups had similar percentages of patients reporting high school as the highest completed education (77% vs 71%), but 11% of the Unaccompanied group reported grade school as their highest education level. The only statiscially significant difference between the groups was seen in the variables of age (P = .03) and race (P = .002).

The mean HbA1c of patients in the Accompanied group was higher than that of their Unaccompanied counterparts (8.9% vs 8.3%), although not demonstrating significance (P = .33). Patients in the Accompanied group had higher rates of tobacco use (43% vs 31%, P = .25), and uninsured status (10% vs 9%, P = .89). The mean BMI of the Accompanied group was 35.5 kg/m² (range [SD], 17.7–62 [11.02]), which was less than that of the Unaccompanied group (37.4 kg/m²; range [SD], 20.8–66.1 [9.30]), but this difference was not statistically significant (P = .42). Rates of dialysis use and nonadherence to treatment were lower in the Accompanied group (10% vs 15% P = .55 and 20% vs 33% P = .21, respectively); however, neither was statically significant.

Patients in the Accompanied group had a mean initial wound size of 11.15 cm² (range [SD], 0.07–111.5 [22]), which was smaller than that of the Unaccompanied group (20.7 cm²; range [SD], 0.07–510.5 [77]) but not to a level of significance (P = .51). The percentage of patients whose index wound healed was higher in the Accompanied than in the Unaccompanied group (77% vs 73%) (P = .69). Mean wound size reduction (SR) was 83% (range [SD], 100 to >+54 [39]) in patients in the Accompanied group compared with 53% (range [SD], 100 to >+200 [171]) in their Unaccompanied counterparts; this difference was not statistically significant (P = .35).

In the Accompanied group, the mean VC was 8.97 (range [SD], 1–23 [6.4]) and WIS was 11.5 (range [SD], 1–35 [9.9]), whereas the means for the Unaccompanied group were 14.09 (range [SD], 2–50 [12.8]) and 22.6 (range [SD], 2–118 [29.4]), respectively. These differences were statistically significant (P = .04 and P = .05, respectively). In the Accompanied group the actual number of accompanied visits (AC) was 7.47 visits (range [SD], 1–23 [5.2]) and the percentage was 87% (range [SD], 52–100 [15]). In the Unaccompanied group, the mean for AC visits was 0.98 (range [SD], 0–7 [1.74]) and the mean percentage was 5% (range [SD], 0–29 [8]). Both differences were statistically significant (P < .001).

A Pearson r correlation coefficient was applied to the extracted data of the 2 subgroups investigating the correlation between % Accompanied and AC, VC, WIS, SR, and Wound Healed variables producing the following results as seen in Table 4. A moderate, negative correlation was seen between % Accompanied and VC (r = -0.38) as well as WIS (r = -0.31) in the Accompanied group; in the Unaccompanied group this correlation was weak and positive VC (r = 0.21) and WIS (r = 0.22) (Table 4). The correlation evaluating Size Reduction in both Accompanied and Unaccompanied status was week and negative (r = -0.13 and r = -0.11 respectively). In the field of AC, the Unaccompanied cohort had a strong positive correlation (r = 0.75), whereas the Accompanied group demonstrated no correlation (r = -0.08) with the % Accompanied. The % Accompanied correlation to the Wound Healed category had a weak positive correlation (r = 0.14) in the Accompanied group and no correlation (r = 0.08) in the Unaccompanied group. The most statistically significant correlations were between the % Accompanied and AC in the Unaccompanied group
(P < .001) and the % Accompanied and VC in the Accompanied group (P = .04).

In this study 30 of 85 patients were accompanied by a caregiver more than 50% of the time, whereas 55 of 85 were not when attending a hospital-based outpatient wound clinic. Age and race were the only statistically significant differences in demographics. The Accompanied group was older (P = .03), had higher rates of tobacco use, and a slightly higher proportion of wounds healed. Differences in HbA1c, healing rates, initial wound size and percent reduction in wound size were not significantly different between the two groups.

The Unaccompanied group was younger (mean, 59 years) and more racially diverse with members spanning all identified racial groups, the majority of whom were black (49%); statistical significance was noted in race (P = .002). While this group displayed several clinical features known to favor wound healing, these patients had the unexpected finding of unfavorable wound healing outcome metrics across the board. Notably, the VC (14.09, P = .04) and WIS (22.6, P = .05) were significantly more than that of the Accompanied group, culminating in an overall lower Wound Healed rate (73% vs 77%).

Persons who were accompanied > 50% of the time had significantly more accompanied visits (P < .001), a lower visit count (P = .04) and fewer WIS (P = .05) than persons who were accompanied < 50% of the time. The Accompanied group had a higher rate of wound SR (83% vs 54%) but the difference was not statistically significant.

Essentially, the patients who were accompanied, though older and with higher rates of tobacco use, demonstrated higher rates of wound SR and overall healing with fewer visits and half the weeks in consecutive time with an open wound. However, it should be noted that their wounds were smaller than the wounds of patients in the Unaccompanied group.

An unexpected finding was the range of the % Accompanied in the Unaccompanied group. This range spanned from 0% to 29% (0–50% potential) demonstrating none were near the cut-off point (45–49%), contributing to clarity between groups. Of those in the Unaccompanied group who occasionally did have a caregiver present in the clinic, the accompanying caregiver attended, at most, fewer than one-third (29%) of their total visits. The percentage of accompanied visits in the Accompanied group spanned from 52% to 100% (50–100% potential), nearly the entire range. Of note, the mean difference in % Accompanied between these 2 groups was extremely large and statistically significant (P < .001), with caregivers attending only 5% of visits in the Unaccompanied group versus 87% of those in the Accompanied group.

There were no statistically significant differences noted between the Accompanied and Unaccompanied groups in the comparisons of BMI, HbA1c, wound healed, wound size, and size reduced. Although the association was weak, patients who were accompanied by their caregiver more frequently had higher wound healing percentage. In the Accompanied group, the correlation between % Accompanied and VC was moderately negative (r = -0.38) and met the level of significance (P = .04). For this study population, the higher the percentage of clinic visits in which the patient was accompanied by a caregiver, the more the corresponding VC was reduced. The exact opposite was seen for the Unaccompanied group with the same variable but with a reversed and lesser magnitude (r = 0.21), indicating a weak positive correlation. The level of significance was not met in the area of % Accompanied with WIS; this was a moderate negative association (r = -0.32) for those with a caregiver and weak positive association (r = 0.22) for those attending alone. As the % Accompanied visits increased, the WIS for the Accompanied group decreased, reinforcing the finding that mean WIS was half that reported in the Unaccompanied patients. As was seen with the findings of the VC, the WIS also had an opposite and weaker correlation with the Unaccompanied group. The correlation between % Accompanied and AC was strongly positive (r = 0.75, P < .001) for the Unaccompanied group; however, this was an expected finding as the AC was a determinant for the % Accompanied equation.

Chronic wounds are burdensome on the patient, the health care system, and the family. An abundance of research exists to support the need for wound care and methods of treatment, especially in patients with diabetes; however, wide gaps exist in the literature related to the roles and benefits of caregivers in relation to wound care.^25-28 As discussed in a literature review by Pittman,²⁹ ample studies exist to support the technical advances in wound care and the impact of the wound on the individual, but there is a paucity of literature pertaining to the chronic wound and its effect on the family. Detailed in this review is the narrow scope of study designs in the existing literature; of the few existing studies, only 1 quantitative caregiver study was found.²⁹ In an editorial examining a National Family Caregivers Association survey published in 2001, the author reports an estimated 54 million Americans had been involved in caregiving in the previous 12 months but “in wound care, we have not done much to factor in the contributions of lay caregivers.”³⁰ This author also cites the paucity of peer-reviewed literature and substantiates the need for wound care professionals to establish relationships with lay caregivers and patients.³⁰

A randomized, unblinded, controlled pilot study by Shields et al³¹ examined the patient-centered communication in a family medicine clinic serving elderly patients (N = 30). This study randomly assigned patients to accompanied or unaccompanied visits and found that companions facilitated the patient–physician discussions by bringing up new topics for the physician to explore. The physician spoke more words when the patient was accompanied, and the presence of a companion did not divert the provider’s attention away from the patient. These authors also mentioned the lack of clinical trials in the literature, noting this was the first randomized study to examine the effect of a third person being present during the medical visit.³¹

In a recent narrative literature review, Messenger et al found that limited research has been conducted in the area of informal caregivers to reveal the true impact on outcomes in diabetic foot ulcers.³² The authors noted that the role of these caregivers in diabetic foot ulcer care has the potential to make a significant difference but is underutilized. The contributions of the caregivers at medical appointments were significant, and the caregivers served as an essential source of information sharing. The review concluded that informal caregivers should be identified early and integrated into the care pathway; however, their role in the management of diabetic foot ulcers and complications remains largely unexplored.³²

To the author’s knowledge, the present study is the first to examine the characteristics of patients attending a wound clinic relatively accompanied versus those who are unaccompanied, and whether a correlation existed in their respective wound healing outcomes. The study design was intended to minimize interobserver variability by enrolling patients under the care of a single clinician (the principal investigator). In an attempt to capture real-world clinical presentations and patient complexity, no exclusions were permitted for uncontrolled glucose, Charcot deformity, previous amputations, current tobacco use, uninsured status, dialysis participation, or any history of osteomyelitis. These exceptions to exclusions is consistent with a prospective, multicenter, open-label clinical trial by Frykberg et al³³ that examined placental membrane allograft response in participants with complex wounds (N = 27) who are frequently eliminated from studies. The Frykberg study highlighted the crucial nature of rapid closure of complex diabetic wounds to avoid amputation as well at the need to study the characteristics of the population present in the clinics.

Diabetes imposes a substantial burden on patients, especially the elderly.¹⁶ As seen in a large multivariant analysis of a national longitudinal study (N = 7443), elderly people who did not have diabetes needed an average of 6.1 hours per week of informal care, whereas those with diabetes managed with oral medications received 10.5 hours of care, and those requiring insulin needed up to 14.4 hours of care per week.³⁴ Caregivers are documented as essential in many areas of serious chronic disease. The National Cancer Institute, in its peer-reviewed Physician Data Query (PDQ) database, describes the caregiver role as a “critical ingredient” to effective cancer management, inclusive of decision-making, treatment planning, and care implementation.²¹ The agency enlists caregiver cooperation at the onset of treatment as part of the “unit” of care spanning across physical, psychosocial, and spiritual domains.²¹

In a 2013 seizure clinic correlation study by Robson and colleagues (N = 48)³⁵ it was noted that companions facilitated the physician–patient discussions by providing additional information, and for that reason patients are routinely invited to bring a seizure witness to the specialty clinics. In addition, companions provided essential diagnostic information, tracked patient symptoms, asked more questions, and assisted the patient in understanding the health care professional’s explanations.³⁵ The Work Group on Alzheimer’s Disease and Other Dementias clinical guideline defines caregivers and family members as a critical source of information and vital for implementing treatment plans.¹³ Therefore, the treatment paradigm has shifted to a patient–caregiver systematic approach, and clinicians are encouraged to form an alliance with the family.¹³

Clinical research regarding the role and impact of caregivers is lacking as it relates to the area of wound care; however, literature for other specialty areas of medicine portrays a caregiver as a beneficial component of the treatment team.

The results of the current study demonstrated favorable findings for patients who attended wound clinic visits with a caregiver, including higher healing rates, a reduced number of visits, less time with an open wound, and increased wound size reduction. Because this was a descriptive and correlation study, causation could not be determined; therefore, the findings are not widely applicable. Future studies are indicated, repeating this framework in different geographical locations to determine if these findings can be replicated. In addition, expanding this design to include additional wound types of pressure or venous etiology may contribute to the literature. Prospective clinical trials may further substantiate the body of evidence to expand understanding of caregiver impact and to determine whether greater effort should be expended to delineate the role of caregivers in wound care guidelines.

This study has limitations related to the caregiver and the patient; it does not consider the quality of the relationship, history of dysfunctional or abusive interactions, or extent of caregiver participation. The patient’s history of illegal drug use, alcohol abuse, mental illness, or the impact of poverty are not considered. The study design has set start and end dates and does not factor in patients who may have initiated treatment in the weeks just before the end date where healing is unexpected in a short period of time. In addition, a patient could have a chronic nonhealing wound that requires enrollment in prolonged wound care services; this situation may impact the WIS and may not strictly coincide with index wound duration. The study does not examine actual wound care cost or comparison of the 2 groups, which may be grounds for future investigation.

This study was a multicenter, retrospective, descriptive, and correlational evaluation of patients with a DWLE who attended a wound center with or without the accompaniment of a caregiver. The only stastically significant demographic difference between the groups was age and race. The Accompanied group demonstrated a higher rate of healing, improved wound size reduction, fewer visit counts, and weeks in service of an open wound. Despite negative predictors of healing, including increased rates of tobacco use, advancing age (statistically significant), and higher mean HbA1c, the Accompanied group demonstrated a higher Wound Healed rate with total VC and WIS nearly half that of the Unaccompanied group. The strongest meaningful correlation was seen in the Accompanied group between % Accompanied and VC, which was found to be a moderate negative correlation and was statistically significant. The % Accompanied statistic correlated to fewer clinic visits for the Accompanied group, whereas the reverse was demonstrated for the Unaccompanied group. The results of this study may encourage additional investigations relative to health care cost savings as VC and WIS variables are indirect measures of cost; each visit incurs a health care billed expense that is in addition to the cost of dressings, medications, and time lost from work in pursuit of wound care. More research is needed to expand the body of evidence for a more specific determination of caregiver impact on wound healing outcomes.

Ms Jessee is a medical student, Oceania School of Medicine, and nurse practitioner, FirstHealth of the Carolinas, Pinehurst, NC.

Address all correspondence to: Carol Jessee, MSN, FNP, Medical Student, FirstHealth of the Carolinas, 6322 Fayetteville Road, Raeford, NC 28376; email: caroljessee@gmail.com.