Mental Illness Training for Long Term Care Staff

Abstract

Objectives

Mental illness is prevalent among nursing home residents, but staff are not well trained to deal with it. This research evaluated an Internet mental illness training program designed for certified direct care workers i.e., Nurse Aides (NAs). Pilot research was also conducted to explore effects and acceptance of the same program with a sample of Licensed Health Professionals (LHPs).

Design

Trial 1: Pre-post randomized treatment and control design for NAs; Trial 2: Quasi-experimental pre-post within-subjects design for LHPs.

Setting

Both studies were conducted on the Internet.

Participants

Trial 1: N=62 NAs; Trial 2: N=16 LHPs

Intervention

Internet-based behavioral skills training and knowledge building, using video modeling with mastery learning instructional design.

Measurements

Video situations testing and assessment of psycho-social constructs associated with behavior change; follow-up interviews with a sample of treatment NAs.

Results

Trial 1: MANCOVA analysis showed positive results (p=.003) for knowledge, attitudes, self efficacy, and behavioral intention, with medium-large effect sizes. The training was well received by the users. Trial 2: Paired t-tests showed significant effects on five of six outcome measures, with medium-large effect sizes, and it was well received by the LHP sample.

Conclusions

Internet training can be an effective approach to help staff work with residents with mental illness. In this research, it showed significant positive effects and was well received by NAs and by LHPs.

Introduction

A total of 3.2 million individuals¹ reside in long term care facilities (LTCs). The reported prevalence of mental disorders among this population varies with the research designs and interests of the researchers,² ranging from 22%³ to 90%.^{4, 5} Depression alone may affect 17%^{6, 7} to 30%,⁸ with less than half of LTC residents being diagnosed and treated.⁶ The Center for Medicare and Medicaid Services¹ reports that 68% of LTC residents have cognitive impairment, with 41% showing moderate-severe impairment on the Cognitive Performance Scale.⁹ Dementia may be linked to challenging behaviors and assaults by residents,^{8, 10–16} and up to 90% of patients with dementia are thought to have psychiatric co-morbidities.^17–19 Comorbidity between dementia and depression, delusions, and hallucinations had the highest correlations with aggressive resident behaviors in one study.²⁰ The percentage of younger individuals (i.e., 21–65 years of age) in LTCs has grown from 11.1% in 1999 to 13.6% in 2008,¹ and they are often undiagnosed and without adequate treatment for mental illness.^21–23 To further complicate potential resident behaviors, 47.9% of LTC residents were reported to have taken nine or more medications the day before the facility interview for the National Nursing Home Survey,³ making them potentially at risk for adverse events due to overuse of medications.²⁴

Despite the high incidence of psychiatric disorders in LTCs, NAs usually receive little or no training to work with dementia or other mental illnesses, ^{2, 23, 25–30} which undermines their capabilities to react appropriately if confronted with mental illness behaviors.^{2, 25, 31, 32} Negative stereotypes^{33, 34} increase the likelihood that staff will fear or avoid residents with mental illness,^{2, 35} which in turn erodes the quality of resident care. As a result, mental illness is too often treated with psychotropic medications or restraint, ^36–38 and behavioral methods are under-utilized by care giving staff.^{37,39, 40}

LHPs (e.g., RNs, LPNs, Physical Therapists, Social Workers) are also likely to have direct contact with LTC residents with mental illnesses. Inadequately trained in both geriatrics and mental illness.^{2, 25, 41–44} LHPs may thus be hard pressed to supervise NAs dealing with behaviors caused by mental illness, or to know how to react themselves when encountering challenging behaviors. Lacking adequate training, LHPs may be subject to the same types of stigmas and ill-suited behavioral reactions as may be shown by inadequately trained NAs.

Previous research has shown the efficacy of interactive training for NAs using interactive computer programs^45–48 and on the Internet.^{46, 49} The research described here was funded by the National Institute on Aging to develop a prototype Internet training program to provide NAs with education and theoretically sound behavioral training to work with LTC residents with psychiatric disorders. It was to be tested in a randomized evaluation. Given the lack of geriatric and mental health preparation of LHPs,^{2, 25, 31, 42, 44} but without resources to conduct a second randomized study, we conducted a separate pilot study to assess the impact on, and acceptance of, the same training website by LHPs.

This study addressed the following research questions:

• As demonstrated by responses to questions following the viewing of video simulations of challenging workplace situations, to what extent will users of an Internet training program demonstrate that they have learned appropriate behavioral and communication techniques for working with residents who display behaviors apparently caused by a mental illness?

• To what extent will users’ knowledge of mental illness be improved by the training?

• To what extent will users’ attitudes, self-efficacy, and behavioral intentions regarding resident behaviors caused by a mental illness positively change as a result of program use?

• How satisfied will users be with an Internet training approach?

• Will nurses and other licensed staff benefit from and accept a training program designed for NAs?

Methods

Overview

To test program efficacy, we conducted “real-world” effectiveness studies⁵⁰ via the Internet with two subject populations:

• Trial 1

  The program was evaluated in a randomized pretest/posttest trial with a sample of the intended target audience: NAs working in LTCs;

• Trial 2

  A convenience sample of LHPs was tested using a within-subjects design as pilot research to examine the potential applicability of the training to other care worker populations.

With the help of LTC consultants, researchers, and mental health therapist, we developed a behaviorally focused Internet training program. Video modeling vignettes, testimonials, and narration supplemented by short on-screen text statements written at the 2nd – 6th grade reading level were designed to emphasize teaching points and create empathy for residents with mental illness. The overreaching philosophy of person-centered care^{36, 37, 51, 52} was emphasized in the instructional design. That is, the instructional model focuses on respect for the values and needs of the resident. NA models in the program are always respectful, attempting to: (a) interpret a situation from the resident’s perspective, (b) verbally empathize with the resident without condemnation, and (c) use knowledge of the resident or the resident's care plan as part of a positively oriented non-punitive redirection strategy. The training focused on helping the resident develop the skills and confidence to deal with whatever behavior was being presented, regardless of what mental illness was causing the behavior.

The Internet Program

Caring Skills: Working with Mental Illness

The first module, guided by a video narrator, introduces the topic of mental health and presents short descriptions of common mental disorders (i.e., anxiety disorders, schizophrenia & psychotic disorders, mood disorders, cognitive disorders, and personality disorders). For each short description, an interested user may click on a “More Details” button to see a more sophisticated article about the characteristics and consequences of that disorder. The reading level of these optional articles is 10^th grade or less.

A second module is designed to dispel common myths about mental illness (e.g., “Residents with mental illness are often violent.”; “Mental illness cannot affect me.”). It presents a total of eight statements about mental illness, which were derived from the NIMH and other mental health organization websites. For each statement, the user is instructed to click on icons indicating whether the statement is a myth or a fact. An animated screen provides feedback about whether the choice was correct or incorrect, while voice-over narration and text reinforce the facts and dispel myths.

The third module uses video vignettes and testimonial stories to demonstrate how NAs can gain information to help build a relationship with their resident (i.e., talking to the resident, to family, and to other staff). It describes, and video-models demonstrate how a positive NA-resident relationship promotes friendly interactions and helps the NA recognize potential behavioral triggers, and how to implement solutions that the resident will accept.

A fourth module presents a behavioral strategy for working with a resident exhibiting potentially problematic behaviors apparently caused by a mental illness (e.g., paranoia, hallucinations). Video vignettes supplemented by narration and text and supportive testimonials from NA video-models demonstrate fundamental techniques using the Assess-Investigate-Do something (AID) approach, which was previously developed by the research team.⁴⁶

A final module gives the program users an opportunity for skill building practice in a mastery learning instructional design. In a new video vignette depicting a problematic resident behavior (i.e., agitated resident hallucinates that bugs are on her), the NA-model reacts with either appropriate (e.g., pauses to assess, friendly greeting) or inappropriate responses (e.g., begins to lose patience, inappropriate redirection attempt). The program user is then asked a series of multiple choice questions about the appropriateness of the NAs behavior. Correct answers are praised, and incorrect answers are remediated with an explanation of why they are incorrect. When all the questions have been answered, the vignette is replayed and incorrectly answered questions are presented again, with the same remediation as before. When all the questions have been answered correctly, or when the re-questioning cycle has been repeated two times, the program moves on, and the user is shown a modeling vignette of how to deal more appropriately with the same situation (e.g., stay calm; show empathy; use knowledge of the resident, and redirect to specific activity that the resident will like).

Procedures

Recruitment procedures for research subjects have been described in detail elsewhere.⁴⁶ In brief, after the protocol was approved by an Institutional Review Board, participants were recruited via Internet advertisements and word of mouth referrals. Potential participants visited an informational website and then were screened on-line to verify their employment in an LTC setting, their job type (e.g., NA, LHP), that they had had recent contact with residents with symptoms of mental illness, and if they had access to a computer with broadband capability and a current email address. The advertising was successful, as demonstrated by 517 hits to the informational web site. A total of 412 people linked to the online screening questionnaire to attempt to qualify for the study.

For Trial 1, qualified NAs were randomized into Treatment or Control conditions. Participants were then emailed a link to an online Informed Consent. Those who agreed to the Consent were linked to an online pretest assessment (T1). After submitting the T1 assessment, Treatment Group participants were linked to the Internet training program and were instructed to view all parts of the program. Treatment Group participants were emailed a link to the posttest online assessment (T2) one business day after viewing at least part of the training program. Control Group participants were emailed the link to the T2 assessment one business day after completing the T1 assessment. The surveys took 20–30 minutes to complete. Once participants submitted both assessments, they were mailed a check for $80. A volunteer sample of Treatment Group NAs participated in telephone interviews after the T2 assessment, and they were then mailed a check for $25.

For Trial 2, a convenience sample of LHPs was recruited on-line and screened-in as described above. They followed the Treatment Group protocol above, and were mailed a check for $80 after completing both assessments. No Control Group of LHPs was recruited.

Participants

As shown in Table 1, the Trial 1 sample of NAs (N=70) was predominantly female (93%), fully employed (77%), had <4 years of NA experience (66%), had at least graduated from high school (99%), and were >35 years of age (57%). Trial 2 participants (N=16) included 5 Registered Nurses, 5 Licensed Practical Nurses, 3 Licensed Social Workers, and 3 Activity Therapists. They were primarily females, tended to work more than 30 hours per week, had at least 6 years of professional experience, and half of them were 46–55 years of age.

Table 1.

Demographic Information Nurse Aides (NA) and Licensed Health Professionals (LHP)

Variable	NAs (n=70)		Treatment (n=36)		Control (n=34)		LHPs (N=16)
	N	%	N	%	N	%	N	%
Age
18–20	2	2.9	2	5.6
21–35	38	54.3	16	44.4	22	64.7	4	25
36–45	18	25.7	10	27.8	8	23.5	4	25
46–55	10	14.3	8	22.2	2	5.9	8	50
Over 55	2	2.9			2	5.9
Gender
Male	5	7.1			5	14.7	1	6.3
Female	65	92.9	36	100	29	85.3	15	93.7
Race
African American	25	35.7	13	36.1	12	35.3
Asian American
Caucasian	43	61.4	22	61.1	21	61.8	16	100
Latino/Hispanic	1	1.4			1	2.9
Other	1	1.4	1	2.8
Education
Grade school or less	1	1.4			1	2.9
Some High School
High School Graduate	13	18.6	9	25.0	4	11.8
Some College	31	44.3	18	50.0	13	38.2	2	12.5
College Graduate	19	27.1	7	19.4	12	35.3	12	75
Graduate/Professional	4	5.7	2	5.6	2	5.9	2	12.5
Trade School	2	2.9			2	5.9
Hours Worked per Week
Less than 10 hours
10 to 20 hours	4	5.7	1	2.8	3	8.8	1	6.3
21 to 30 hours	12	17.1	5	13.9	7	20.6	2	12.5
31 to 40 hours	32	45.7	17	47.2	15	44.1	4	25
40 plus hours	22	31.4	13	36.1	9	26.5	9	56.3
Family Income
Less than $10,000	3	4.3	1	2.9	2	5.9
$10,000 – $20,000	7	10.1	3	8.6	4	11.8	1	6.3
$20,000 – $40,000	31	44.9	15	42.9	16	47.1	5	31.3
$40,000 – $60,000	15	21.7	10	28.6	5	14.7	5	31.3
More than $60,000	13	18.8	6	17.1	7	20.6	5	31.3
NA Experience
2–6 months	3	4.3			3	8.8
7–12 months	1	1.4			1	2.9
1–3 years	20	28.6	11	30.6	9	26.5	1	6.3
4–10 years	35	50.0	20	55.6	15	44.1	2	12.5
More than 10 years	11	15.7	5	13.9	6	17.6	13	81.3

Note: NA = Nursing Aide

Measures

NA-participants (Trial 1) and the LHPs (Trial 2) responded to identical on-screen assessments at T1 and T2. Assessment items were designed to measure situational learning, knowledge, attitude, self efficacy, behavioral intentions and program acceptance.

Situational testing studies involve presenting test-takers with realistic hypothetical scenarios and asking them to identify appropriate responses. This assessment approach is useful when in vitro observations are not practical, and it has been conducted as video situational tests (VSTs) with NAs.^{45, 46} VSTs used in this research, comprised four short video vignettes: VST-1: an agitated and confused resident grabs another resident forcefully; VST-2: resident expresses paranoia and fear; VST-3: resident thinks bugs are crawling on her; and VST-4: a resident who appears to feel down and depressed. Each VST scene was followed by a self-efficacy item stating “If you were faced with the situation just shown, how confident are you that you would know what to do?” Participants rated their responses on a on-screen 5-point Likert scale ranging from “Not at all confident” to “Extremely confident”. Subjects then responded to a single multiple choice question testing knowledge of: what action to take first for VST-1, what to say first for VST-2, what to do or say to stop the paranoid behavior in VST-3, and what to do to change the depressed behavior in VST-4.

A battery of psychosocial assessment items then measured changes in constructs associated with social cognitive theory^{53, 54} and the expanded theory of reasoned action,⁵⁵ including attitudes, self-efficacy, and behavioral intentions. The items addressed attitudes about mental illness, working with agitated residents, and the importance of specific behavioral responses to different types of situations. The participants’ knowledge of mental illness was tested by asking them to differentiate myths from facts about mental illness. These items included 3 response choices (i.e., myth, fact, not sure). Self-efficacy items were presented on a 7-point rating scale (i.e., 1 not at all confident; 7 extremely confident). These items addressed the subject's self-confidence to perform specific behavioral responses when faced with a challenging resident behavior by asking: “If you wanted to, how confident are you in your ability to….?” Behavioral intention items were presented as questions about the subject's intention to perform specific behavior when dealing with a situation related to a mental illness. The items were rated on a 7-point rating scale (i.e., 1 not at all likely; 7 extremely likely), and the subject was asked “With resident behavior that is caused by a mental illness, how likely is it that you will…”

Program acceptance was assessed at T2 for the NA Treatment Group (Trial 1) and the LHPs (Trial 2) with 12 items. A total of 8 agree-disagree Likert-type items (1 strongly disagree to 5 strongly agree) were adapted from Website acceptability measures.^{56, 57} Participants also were asked to rate the training program (1 not at all to 7 extremely positive) in terms of helpfulness, enjoy-ability, recommendability, and satisfaction. An additional single item requested typed-in feedback about the program and suggestions for how to improve it.

Finally, a convenience sample of Trial 1 Treatment Group NA-volunteers was interviewed by telephone after their submission of the T2 assessment. Participants were asked about their work-experience with residents with mental illness, and what they remembered about the training website. Then they were asked about their perceptions of the value of the training to them as they subsequently worked with residents. LHPs were not interviewed due to funding limitations.

Results

Program Effects

Trial 1

One evaluation goal was to examine gains by Treatment subjects (N=34) following program use, compared to Control subjects (N=33) on the targeted theoretical constructs. These constructs included: (a) VST knowledge, (b) VST self-efficacy, (c) knowledge of myths vs. facts, (d) attitudes, (e) self-efficacy, and (f) behavioral intentions. Multivariate analysis of covariance (MANCOVA) comparing the two conditions was conducted on the six posttest outcome measures, with pretest outcome measures included as covariates. An overall multivariate model was tested followed by six univariate models. The multivariate test was significant. Treatment participants were found to have significant and large gains compared to the Control participants, F (6, 54) = 3.86, p = .003, partial eta-square = .30, which is a large effect size. Effect size computations complement inferential statistics (i.e., p-values) by estimating the strength of the relationship of variables in a statistical population; partial eta-square .01 = small, .06 = medium, and .14 = large.

Treatment participants differed significantly from the Control participants on five of the six posttest measures (Table 2). The greatest gains were obtained for attitudes about mental illness (eta-square = .157, large effect size), followed by myths about mental illness (eta-square = .117, moderately large effect size), VST knowledge regarding appropriate action to take based on the video scenario presented (eta-square = .072; medium effect size), intention to perform specific behavioral responses in dealing with mentally ill patients (eta-square = .071, medium effect size), and VST self-efficacy regarding perceived self-confidence about taking appropriate action based on the video scenario presented (eta-square = .064; medium effect size). In sum, significant and medium-to-large effects were obtained on five of the six outcome measures providing strong support for the efficacy of the program.

Table 2.

Pre-Post Descriptive Statistics and ANCOVA Results for the Six Outcome Measures for the NA Sample

Outcome measure / Condition	Pre-Test		Post-Test		Condition Effect
	M	SD	M	SD	F-test	p-value	Eta2
VST knowledge					4.56	.037	.072
Treatment	1.71	1.00	2.38	0.89
Control	1.70	0.88	1.88	0.82
VST self-efficacy					4.01	.050	.064
Treatment	3.82	0.72	4.28	0.56
Control	3.73	0.60	3.99	0.62
Myths					7.84	.007	.117
Treatment	8.38	1.76	9.65	1.25
Control	8.82	1.59	8.94	1.35
Attitudes					11.00	.002	.157
Treatment	4.99	0.40	5.57	0.58
Control	4.92	0.51	5.09	0.51
Self-efficacy					0.38	.541	.006
Treatment	5.95	0.73	6.23	0.68
Control	5.79	0.69	6.02	0.66
Behavioral intention					4.49	.038	.071
Treatment	6.07	0.72	6.37	0.63
Control	5.88	0.66	6.03	0.69

Note. N = 67; 34 Treatment and 33 Control participants. Multivariate F (6, 54) = 3.86, p = .003, partial eta-square = .30. Univariate F-test df = 1, 59. Eta-square of .01, .06, and.14 are considered small, medium, large effect sizes, respectively.⁵⁸

Trial 2

To test the efficacy of the training program among the LHP sample, pre-post paired t-tests were conducted on the six outcome measures. Table 3 presents the pre- and post-training means and standard deviations as well as the paired t-test results and effect sizes (partial point-biserial r)⁵⁹ for the outcome measures; partial point biserial r .14 = small; .36 = medium; and .51 = large As can be seen, significant effects were obtained in the predicted direction for the pre-post comparisons across all outcome measures with the exception of a trend-level effect for VST knowledge. Large effect sizes were obtained for VST self-efficacy, attitudes, general self-efficacy, and behavioral intention; medium effect sizes were obtained for VST knowledge and myths.

Table 3.

Pre- and Post-Training Descriptive Statistics and Paired t-test Results on Outcome Measures for the LHP Sample (N = 16)

Outcome measure	Pre-Training		Post-training		t	p-value	Effect Size
	M	SD	M	SD
VST Knowledge	1.88	1.00	2.44	0.89	1.86	.083	0.43
VST Self-efficacy	3.69	0.74	4.18	0.60	4.44	<.001	0.75
Myths	9.44	0.96	10.06	0.85	2.18	.046	0.49
Attitudes	5.13	0.27	5.80	0.57	5.72	<.001	0.83
Self-efficacy	5.64	0.80	6.17	0.66	4.69	<.001	0.77
Behavioral intention	6.10	0.58	6.35	0.63	2.65	.018	0.56

Note. Effect size is being reported as a partial point-biserial r: small = .14; medium =.36; large = .51.⁵⁹

User acceptance

Participant responses were very favorable regarding the website design (see Table 4). NAs rated the web site at more than 6 on a 7-point Likert scale on satisfaction and related categories. Responses about the usability and credibility of the website were also very favorable. Responses from LHPs were also uniformly positive, although consistently slightly less than those of the NAs.

Table 4.

Website usability and acceptability ratings by NA Treatment Group and LHPs

	NAs	LHPs
Item (1 Negative - 5 Positive)	Mean (SD)	Mean (SD)
“ The website has much that is of interest to me.”	4.44 (.61)	4.31 (.60)
“Using the website was a waste of my time.” [reverse scoring]	1.36 (.64)	1.53 (.63)
“I would imagine that most people would learn to use the website quickly.”	4.40 (.67)	4.06 (.85)
“The website needs more introductory explanation.” [reverse scoring]	1.92 (.81)	2.19 (1.05)
“I like how the website guided me through each page.”	4.31 (.71)	4.19 (.91)
“I wish I could move to different sections of the website more freely.” [reverse scoring]	2.78 (1.07)	2.87 (1.15)
“The video situations in the website are believable.”	4.46 (.58)	4.31 (.70)
“The website is attractive and appealing.”	4.35 (.60)	4.13 (.62)
Satisfaction with website (1 Negative -7 Positive)
How satisfied were you with it	6.06 (1.04)	5.63 (1.15)
How helpful was it?	6.29 (1.27)	5.56 (1.50)
How enjoyable was it to use?	5.84 (1.20)	5.31 (1.40)
Would you recommend it to others?	6.04 (1.26)	5.75 (1.24)

Qualitative comments

The evaluation instrument provided users with an opportunity to offer typed-in comments about improving the program. A total of 26 NAs offered comments which were consistently positive. Twenty four of the NAs had positive opinions about the website’s functionality, including one who stated “My six year old son watched the video with me. He was able to navigate through the website and answer the questions.” Six suggestions for improvements included calls for expanded training content with more video scenarios, and greater freedom to move within the program. Eleven LHPs offered feedback. It included 10 positive statements, and 10 constructive suggestions, mostly about additional features that might be added. No negative comments were received from either group.

Telephone calls to 11 Treatment Group NAs averaged 25 minutes in length and occurred an average of 11 days (SD 9.09) after use of the program. The training program continued to be valued more than two weeks after it was viewed. Four interviewees reported using the A.I.D. strategy at work, and six others mentioned using elements of the A.I.D. approach. Several interviewees indicated that they previously lacked the type of training presented in our Internet training program.

Discussion

This research successfully developed an Internet training program that increased NAs confidence and skills to work with LTC residents with mental illness. In parallel, we believe that it may have reduced the stigma of mental illness (NIMH 2010) for them by improving their factual knowledge and attitudes about mental illness.⁴³ Typed-in comments and subsequent telephone interviews suggested broad acceptance of the program, and that some NAs were using recommended techniques with residents. In sum, results suggest that the training was effective, well received by the NAs, and that recommended techniques were generalizable into the LTC setting. We believe that this type of computerized instruction offers a valuable training adjunct, and that it could be part of each new-employee orientation, thus establishing normative expectations for caregiving behaviors at a LTC.

Program effects on LHPs were also positive. While conclusions should be tempered by the small sample size and non-randomized design, the results were positive with medium to large effect sizes. These results are not unexpected, given the lack of mental health preparation in the professional training curricula of LHPs.^{2, 25, 41, 42, 44} That this Internet training was well received by LHPs is in itself interesting, because the program was designed for less-educated certified staff. Overall, the results suggest that even though designed for certified staff, behavior skills training has potential benefit for, and was valued by LHPs too.

Taken together, this study supports previous research indicating the potential efficacy of Internet training for NAs.^{46, 49, 60, 61} Individualized computerized training is attractive because it is replicable and time efficient, compared to a traditional in-service training model.^{45, 46} Our experience from in-person studies and usability testing has been that the lack of computer experience is not an obstacle for NAs, if the program is designed with the inexperienced users in mind (i.e., no keyboarding required; low reading level). A brief training on use of the computer mouse may be needed initially for inexperienced users. Training with a computer, however, need not be exclusively a single-user experience, because in-service sessions with a computer projector may be a valuable training adjunct for review of specific topics and to facilitate group sharing and discussion. For LHP users, the program potentially could be supplemented with additional content about geriatrics and psychiatric disorders, which could facilitate their staff training capabilities and might qualify for continuing education credit.

While the research presented here is promising, it has limitations. First, follow-up assessments are needed to examine maintenance of training effects. Although self-reports via telephone interviews suggested that the intervention had post-training behavioral effects on some of the program users, longer term follow-up, preferably with in vivo data, would provide more solid evidence of training effects. Second, while study participants were randomized, they participated on the Internet without direct contact with the research team. Consequently, we cannot verify that they met the screening criteria. Third, because this was an Internet study, we speculate that the participants may have included more sophisticated computer users than is representative of the NA population as a whole. Finally, because of the small sample size and use of the within-subjects design, the LHP study results should be viewed with some caution until further research has been conducted.

Conclusions

The results reported here show the potential value of training LTC staff about mental illness, and they support previous research, demonstrating the efficacy of Internet training for direct care givers. A program designed for NAs also can be effective and acceptable to licensed staff. Broad implementation of Internet training for LTC staff may yield enhanced employee’s confidence and skills, as well as improved resident care.